Bruhn + Supplement PDF

ANALYSIS AND DESIGN
OF
FLIGHT VEHICLE STRUCTURES
TRI-STATE OFFSET COMPANY

ANALYSIS AND DESIGN
OF
FLIGHT VEHICLE STRUCTURES
By
E. F. Bruhn, B. S., M.S., C.E.
Professor (Emeritus) of Aeronautics and Astronautics

Purdue University
Assisted by the following persons and

the chapters and parts they contributed:
DR. R. J. H. BOLLARD, (Chapters A24, A25, A26)

Head, Department of Aeronautics and Astronautics
University of Washington, Seattle, Washington
LLOYD E. HACKMAN, (Chapters C12)

Consultant on Composite Structures Engineering
DR. GEORGE LIANIS, (Chapter A18)

Professor of Aeronautics and Astronautics
Purdue University
WILLIAM F. McCOMBS, (Chapter CIO-Part 2, Chapter Cll-

Part 2, Chapter D3)
DR. A. F. SCHMITT, (Chapters A7, A8, A16, A17, A22)

Litton Systems, Inc.
CLARENCE R. SMITH, (Chapter C13)

Engineering Consultant on Fatigue ofMetals and Structures
DR.JOSEPHA. WOLF,JR., (ChapterA23)

Senior Research Engineer
General Motors Research Laboratories
Copyright 1973 by E. F. Bruhn
All rights reserved. including the right of

reproduction in whole or in part in any form.
Portions previously published

under the following titles
"Analysis and Design of Aircraft Structures"
by E. F. Bruhn
"Analysis and Design of Flight Vehicle Structures"
by E. F. Bruhn
Published and Distributed by

Tri-State Offset Company
Printed in U.S.A.
PREFACE
The 1965 edition of the book "Analysis and Design of Flight Vehicle Structures", contained
over 950 pages. As time passes it is normal to expect expanded or new material to be added to later
editions of a book. Thus it appeared desirable to add some 380 pages of additional material, which
meant that a new edition would contain over 1300 pages or too large for a single volume. A decision
was made to issue the new edition in two volumes. This two volume edition was practically ready
for binding, when a further final decision was made to remain with one volume edition and place the
380 pages of new material referred to above in 3 separate, rather small sized books, Information on
these 3 separate books is given on the 3 pages following the Table of Contents. The reader will note
that these 3 books deal with special subjects in the structural design field and are available.
This new 1973 one volume edition presents 2 major changes from the 1965 edition. Chapter
A23 has been completely revised and expanded by a new co-author, namely, Dr. Joseph A. Wolf,
formerly on the faculty of UCLA and now with the General Motors Research Laboratories. Dr. Wolf
had the cooperation of Dr. A. F. Schmitt, the author of the A23 Chapter in the 1965 edition. The
other major change is the replacement of Chapter C13 on the subject of fatigue in the 1965 edition
by a new chapter on fatigue. The Author of this new Ch~pter C 13 is C. R. Smith, a widely known
authority in the broad field of fatigue of materials and structures.
This author expresses his thanks to the several co-authors for checking over their material. Only
minor corrections have resulted from this check.
A considerable amount of material in various chapters by the author made use of reports, manu-
als, drawings and photographs supplied to the author by many aerospace companies, in particular,
Boeing of Seattle; Douglas Co.; General Dynamics, Fort Worth and San Diego Divisions; Martin
in Denver; North American Aviation, Columbus and Tulsa Divisions; and the Vought Division of
LTV Corp. The author is deeply grateful for this assistance and cooperation.
As a final note, if any reference in the Chapters of this book refers to Volume 2 it should
be discarded by the reader.
June 1973
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
I
TABLE OF CONTENTS
Chapter No.
A1 The Work of the Aerospace Structures Engineer.
STATICALLY DETERMINATE STRUCTURES
(Loads, Reactions, Stresses, Shears, Bending Moments, Deflections)
A2 Equilibrium of Force Systems. Truss Structures. Externally Braced Wings. Landing Gear.
A3 Properties of Sections - Centroids, Moments of Inertia, etc.
A4 General Loads on Aircraft.
A5 Beams - Shear and Moments. Beam - Column Moments.
A6 Torsion - Stresses and Deflections.
A7 Deflections of Structures. Castigliano's Theorem. Virtual Work. Matrix Methods.
THEORY AND METHODS FOR SOLVING STATICALLY

INDETERMINATE STRUCTURES
A8 Statically Indeterminate Structures. Theorem of Least Work. Virtual Work. Matrix Methods.
A9 Bending Moments in Frames and Rings by Elastic Center Method.
Al0 Column Analogy Method.
A11 Continuous Structures - Moment Distribution Method.
A12 Slope Deflection Method.
BEAM BENDING AND SHEAR STRESSES.

MEMBRANE STRESSES. COLUMN AND PLATE INSTABILITY.
A13 Bending Stresses.

A14 Bending Shear Stresses - Solid and Open Sections - Shear Center.
A15 Shear Flow in Closed Thin-Walled Sections.
A16 Membrane Stresses in Pressure Vessels.
A17 Bendi ng of Plates.
A18 Theory of the Instability of Columns and Thin Sheets.
INTRODUCTION TO PRACTICAL AIRCRAFT STRESS ANALYSIS
A19 Introduction to Wing Stress Analysis by Modified Beam Theory.

A20 Introduction to Fuselage Stress Analysis by Modified Beam Theory.
A21 Loads and Stresses on Ribs and Frames.
A22 Analysis of Special Wing Problems. Cutouts. Shear lag. Swept Wing.
A23 Analysis by the "Method of Displacements".
THEORY OF ELASTICITY AND THERMOELASTICITY
A24 The 3-Dimensional Equations of Thermoelasticity.

A25 The 2-Dimensional Equations of Elasticity and Thermoelasticity.
A26 Selected Problems in Elasticity and Thermoelasticity.
TABLE OF CONTENTS Continued
Chapter No.
FLIGHT VEHICLE MATERIALS AND THEIR PROPERTIES
B1 Basic Principles and Definitions.

B2 Mechanical and Physical Properties of Metallic Materials for Flight Vehicle Structures.
STRENGTH OF STRUCTURAL ELEMENTS AND COMPOSITE STRUCTURES
C1 Combined Stresses. Theory of Yield and Ultimate Failure.

C2 Strength of Columns with Stable Cross-Sections.
C3 Yield and Ultimate Strength in Bending.
C4 Strength and Design of Round, Streamline, Oval and Square Tubing in Tension, Compression, Bending,
Torsion and Combined Loadings.
C5 Buckling Strength of Flat Sheet in Compression, Shear, Bending and Under Combined Stress Systems.
C6 Local Buckling Stress for Composite Shapes.
C7 Crippling Strength of Composite Shapes and Sheet-Stiffener Panels in Compression. Column Strength.
C8 Buckling Strength of Monocoque Cylinders.
C9 Buckling Strength of Curved Sheet Panels and Spherical Plates. Ultimate Strength of
Stiffened Curved Sheet Structures.
C10 Design of Metal Beams. Web Shear Resistant (Non-Buckling) Type.
Part 1. Flat Sheet Web with Vertical Stiffeners. Part 2. Other Types of Non-Buckling Webs.
Cll Diagonal Semi-Tension Field Design.
Part 1. Beams with Flat Webs. Part 2. Curved Web Systems.
C12 Sandwich Construction and Design.
C13 Fatigue.
CONNECTIONS AND DESIGN DETAILS
Dl Fittings and Connections. Bolted and Riveted.

D2 Welded Connections.
D3 Some Important Details in Structural Design.
Appendix A Elementary Arithmetical Rules of Matrices.

ANALYSIS & DRSIGN OF FLlGEIT VRlilCLR STRUCTIJRRS - BRUHN
CONTE.VTS
Al THE ~ORK OF THE AEROSPACE RNGllriER
At.l Introduction
A~,2 G~ne[al Organization of an Aircraft Company Engineering Division
Al.) T~e ~o[k of the Structur~s Group
A2 RQUILlBRIUlI OF FORCE SYSTEliS. TRUSS STRUCTIJRRS
A2.1 Introduction
A2.2 Equations of Static Equilibrium
A1.3 Structural Pitting Units for establishing the Force Characteristics
of Direction and Point of Application
A2.4 Symbols for Reacting Fitting Units as used in Problem Solution
A1.5 Statically Determinate and Statically Indeterminate Structures
AZ.6 Examples of Statically Determinate and StaLleally Indeterminate
Struc:u.~s
A2.7 Sxample P.oblem Solutions of Statically Deter~in3te Coplanar
St[uc~ures and Coplanar Loadi~gs
A2.3 Stres~0S in Coplanar Truss Structures under Coplanar Loading
(Stat~cally Determinate/Indetec~~na~e ?)
A19 ~eth()d of Joints
A). i.0 Method of Moments
A2.11 ~et~od of Shears (Sections)
.-\.1. 12 Aircaft Wing Structure. Truss Type with Fa~ric or Plastic Cover
:-1.2,13 Landing Ge1r Structure
A~ . 14 Example Problems of Calc~lacing Reactions and Loads on Members
Of Landi~~ Gear Units
A2.1S Probl>:!ms
A3 PROP~RTIES OF SRCTIONS - CE.NTROIDS. MOH7.V!S 02 INERTIA ETC
.-\.3.1 Intr,)duction
.\] . 1 Centroids, Centre of G~avity
A.' .3 ~oment of Inertia
A3 4 ~oment of Inertia of an Area
A3.S Polar Moment of Inert~a
A3.6 Radius of Gyration
A3.7 Parallel Axis Theorem
A3.7a Mass :-laments of Inertia (Tables of Sections)
AJ .q Product of Inertia
AJ.9 Product of Inertia for Axes of S/~~etr,
A3.1O Parallel Axis Theorem
A3.11 Moments of Inertia ~it~ ~espec: of Inclined fu~es
A3.12 Location of Axes for ~hich Produc~ of Inenia is Ze,o
A3.13 ?rLnci?al AY.es
AJ.14 Section Properties 0: Ty?ical Aircraft Str~ct~ral Sections
A3.;S ?:"oble,:s
GF.NE.RAL LOADS ON AlrtCRAIT
A4. [ :ntr0c·-.:ction
A4. J ~imit O~ Applied Loads. ~es~g~ ~oads
A4,3
A4.4 ~~ighc and Inertia Forces
A4.S Air Forces On Ving
A4.6 Forces on Airplane in F~igh:
A4.7 Load rae tors
A4.8 Design Flight Requirements for Airplane
A4.9 Gust Load Factocs
Ai..10 Il1us~r;Hion of Hain ?light C0nditions. VeLocity-Load Factor
Diag:-am
A4.11 Special FliKht Deslgn Conditions
A4.12 Example Problems Involving Accelerated Motion of Rigid Airplane
A4.13 Effect of Airplane not being a Rigid Body
A4.14 General Conclusions on Influence of Dynamic Loading on Structural
Design of Airplane
A4.15 Problems
B&AHS - SHKA.R AND KOKENTS
A5.1 Introduction
AS.2 StaticallY Determinate and Indeter~inate B€amS
AS. ) Shear and Bending Koment
A5.4 Shear and ~ornent Diagra~s
A5.5 Section of Maximum Bending Homent
A5.6 Relation betveen Shear and Bending ~oment
AS.7 Koment Diagr~s as Hade-up of Parts
A5.8 Forces at a Section in terms of Forces at a previous Station
A5.9 Equations for Curved Beams
AS.10 Torsional Moments
A5.11 Shears and ~oments on ~ing
AS.12 Example Problem of Calculating wing Shears and Moments for One
Unit Load Condition
AS .13 Introduction to Shears and ~oments on Airplane Sody
AS.14 Design Condi,ions and Design ~eights
A5.15 Body Veight and Balance Distribution
A5.16 Load Analysis. Unit Analysis
A5.17 Example ?roble~ Illustrating the Calculation of Shears and
Moments on Fuselage due to Unit Load Condi~ions
A5.18 Unit Analysis for ?uselage Shears and Moments
A5.19 Example of Fuselage Shears and Homents for Landing Conditions
AS.20 Inertia Loads due to Angular A~celeration
AS.21 Solution for Inertia Loads due to Unit 100,000 lbs.in
Pitchi:tg ~Olllent
A5.21 ?rob 1e'1Is
BENDING MOMENTS - BEAM - COLUMN ACTION
,1,.5.23 Introduction
,1,.5.24 General Action of a He~be~ Subjected to Combined Axial and
TraIE':er:;~
Loads
AS.25 Equations for a Compressive Axially Loaded Strut ~ith UniEor~ly
Distributed Side Load
A5.26 Formulae for other Single Span Loadings
AS.27 Moments foe Combinations of various Load Syste~s
AS.28 Example Problems
A5.29 Stresses Above Proportional Limit Stress of Material
A5.JO
A5.31 Beam-Columns in Continuous Structures
A6 TORSION - STRESSES AND DEFLECTIONS

A6.1 Introduction
A6.2 Torsion of Hembers ~ith Circular Cross-sections
A6.3 Transm~ssion of Pover by a Cylindrical Shaft
A6.4 Torsion of ~e~bers ~ith Son-Circular Cross-sections
A6.5 Elastic Hemorlne Analogy
A6.6 Torsion of Open Sections Composed of Thin Plates
A6.7 Torsion of So:id ~on-circular Shapes and Th~ck-~alled
Tubular Shapes
A6.8 Torsion ot Thin-~alled Closed Sections
A6.9 Expression ~or Torsional Homent in terms of Internal Shear
Plo~ Systems for MUltiple Cell Closed Sections
A6.10 Distribution of Torsional Shear Stresses in a Multiple-Cell
Thin-Valled Closed Section
A6.11 Stress Distribution and Angle of T~ist for ~-Cell Thin-Vall
Closed Section
A6.12 Example Problems of Torsional Stresses in Multiple-Cell
Thin-Valled Tubes
A6.13 Example for 3-Cell Tube
A6.14 Torsional Shear Flov in ~ultiple Cell Beams by Method of
Successive Corrections
A6.15 Use of Operations Table to Orbanise Solution by Successive
Corrections
A6.16 Torsion of Thin-~alled Cylinder having Closed Type Stiffeners
A6.17 Effect of End Restraint on Kembers carrying Torsion
A6.18 Example Problem Illustrating Effect of End Restraint on a
Member in Torsion
A6.19 Problems
A7 D8PL&CrIONS OF STRUC11JRES
A7.1 Introduction
A 7.2 ~ork and Strain Energy
A7.3 Strain Energy Expressions for various Loadings
A7.4 Theorems of Virtual ~ork and ~inirnum Potential Energy
A7.5 Theocem of Complimentacy Ene~gy and Castigliano's Theocern
A7.6 Calculations of Stcuctucal Deflections by use of
Casttgliano's TheoLem
A7.7 Calculations of Stcuctural De:lections USing Dummy-Unit
Loads (~ethod of Virtual ~ork)
A7.8 Deflections due to Ther~al Strains
A7.9 Matrix ~ethods in Deflection Calculations
A7.10 Member ?rexibility Coe~:icien:s; Compilation of a Library
A7. 11 Application of Matrix ~~:hods co various StrucCuces
A7.12 Deflections and Angular Chang~ of the Elastic CUrve of
Simp:e Beams by Kohr's Het~od of Elastic Veights
A7.13 Example Problems
A7.14 Deflec:ion of aea~s by ~oment-Area Hethod
A7.15 aeam Fixed End Moments by Ke:hod of Area Moments
A7.16 Truss Deflection by Het~od of Elastic Veights
A7.17 Solution by Example P~oblems
A7.18 Problems
A8 5TAT I CALLY INDE:TRIUUNATR STR UC1.1JRES
A8.00 Introduction
A8.0 The Principle of Super?osition
A8.1 The Statically Indeter~i~ate Problem
A8.2.1 Determination of Reduncant Reactions by Least Vork
A8.2.2 Redundant Stresses by Least _ock
A8.3 Redundant Problems by the Ket~od of Dummy-Unit Loads
A8.4 Example Problems - Trusses ~ith Single Redundancy
A8.5 Trusses ~ith DQuble Redundancy
A8.6 Trusses ~ith Multiple Redundancy
AB.7 Redundant Strllccures with ~emberssubjected to Loadings
in addi~ion to Axial Forces
A8.8 Initial Stresses
A8.9 Thermal Stresses
AS .10 Reduncant Problem Stress Calculations by Matrix Methods
A8.11 Redundant Problem Deflection Calculations by Matrix Methods
'\8.12 Precision and Accllracy in Redu~dant Stress Calculations
AB.ll Thermal St:ess Calcular;ons by Ma:rix ~e~hods
A8.14 Thermal De~:ections by Ma:rix ~erholjs
A9 B&NDING MOKENTS IN FR.AXES AND RINGS (Elastic Centre)
A9.1 Introduction
A9.2 Der£vation of Equations. Unsymmetrical Frames
A9.J Equations for Structure with Symmetry about One Axis
thrOIJgh Elastic Centre
A9.4 Example Problem Solutions. Structures with at least
One Axis of Sy~metry
A9.S Unsymmetlical St~uctures. Example Ptoble~ Solutions
A9.6 Analysis of Frame ~ith Pinned Supports
A9.7 Analysis of Prame virh One Pinned and One Fixed Support
A9.3 Solution of Unsymmetrical Frames using ?rincipal Axes
A9.9 Problems
Ala srATICALLY INDKTIlR!illlAT"E STRucruR£S - COLUMN A!lALOGY METBOD
AlO.1 General
A10.2 Geno::ral gxplanation of Column Analogy Method
A10.3 Frames One Axis of SymmetrJ
~ith
A10.4 Unsy~metrical Frames or Rings
A10.S Example Problem - Unsymmetrical Section
A10.6 Problems
All CONTINUOUS STRUCTURBS - HOKENT DISTRIBUTION MKTEOD
A11. 1 Introduction
A11.1 Definitions and Derivatives of Terms
All. 3 Calculation of Pixed End Moments
All. 4 Stiffness Pactor - Carryover Factor - Derivation of
All. 5 General Expressions for Stiffness and Carry-over Factor in
terms of Fixation Factor at far end of 3~a~
A11.5a 2xample Problems
All .6 General S~mrnJry of Proced~res
A11 . 7 ContillUQUS Beams with Yielding or Deflec~ed Supports
All. 8 ?ixed End Koments due to Support Deflec~iJns
All.'1 2nd Moments for Conti~uous Frame~0rks ~hose Me~bers are
not in a st.raight line. Joint Rotation Only
A11.10 Continuous Structures \{ith Members of Variable Homent
of Inert ia
.4.11.11 rrames vith UnLLo'~ Joint Deflections due to Sidesyay
:\11.12 Effect of Axial Load on Moment Distribut:on
A~ L 13 Fixed End Kcme~ts, Stiffness and Car:y-o,er Factors for
Beam Columns of Constant Section
Al1.1~ Illustrative Proble~s
A11.15 Secondary 3ending Home~ts in Trusses ~it~ig:d Joints
Al1. 16 Structures ¥ith Cur~ed Kenbers
.\11.17 Structures ~ith Curved He~bers
Al1.18 Stiffness and Carry-over Factors for Cuc~ed ~embers
:\11.19 2xample Pcoble~s. Continuous Structures involving
Curved Members
All. 10 Problems
Al1 SPECIAL KETBQDS - SLOPE DEFLECTION KETBOD
A12. 1 General
:\12.2 Derivation of Slope Deflection Equation
A12. 3 Hinged End. Slope Deflect ion Equa t ior.
A12.4 Exampl~ Problems
Al1.S Loaded Continuous Beam ~ich Yielding Supports
A',2.6 Statically Indeterminate F[a~es. JOLnt Rotation only
A12 . 7 Frames ~ith Joint Displace~e~ts
AL2.S 'example Problems of ?rames ~i th Unkno'.l" Jo~nt DisplaCement
A12.9 Comments on Slope Deflection Method
A~2.10 Problems
A'3 BENDING STRESSP.S
Al 3. a Introduction
A! 3 . 1 Location of Neutral Axis
AD.2 Equations for Bending Stress, Homogeneous 3eams, Stresses
below Proportional Limit Stress
AD. :3 Method 1. Stresses for Moments about the Principal Axes
AU.4 Method 2. Stresses ~y use of Neutral Axis for given Plane
of Loading
,
AI]. S Hethod 3. Stresses from Homents, Section Properties and
Distances referred to any pair of Rectangular Axes
through the Centroid of the Section
AI).6 Advantages and Disad~antages of the Three Methods
AI].7 Deflections
AI] .B Illustrative Problems
Al J. 9 Bending Stresses in Beams ~ith ~on-Homogeneous Sections
Stresses .ithin the Elastic Range
All.IO Bending Stresses of Homogeneous Beams Stressed above the
Elastic Li~it Stress Range
Al].l1 Curved Bea~s. Stresses ~ithin the Elastic Range
All.12 Problems
A14 BENDING SHRAR STRESSES - SOLID AND OPEN SECTIONS~ SHEAR CKNTR.&
A14.l Introduction
A14.2 Shear Cen t re
A14.] Derivation of Formula for Flexural Shear Stress
Al4.4 Example Prable~s. Symmetrical Sec~ions. External Shear Loads
act through Shear Centre
A14. S Maximum Shear Stresses for Sim?le Cross-sections
A14.6 Derivation of Flexural Shear F:o,,", Equation. Symmetrical
Beam S02-.:tion
A14.7 Shear Stresses and Shear Centre Eor Beam Sections ~ith One
Ax i S 0 E S:,nllne try
A14,3 Shear Str02SS02S for Unsy~me:rical Bea~ Sections
A14. '1 geams ... ith Constant Shear Flo,,", ~eJs
A14.10 Example Pr,)ble~s for Seams ... ith Constant Shear Flows
bet.een Flange ~ernbers
A14.1l Shear C~ntre Location by uSLng Neutral Axis Method
A14 :2 Proble:1s
Al5 SHEAR FLQtJ IN CLOSW THIN-'ii'ALLEJ SECfIONS
AlS.i IntrodiJc.tion
A1S.2 Singl~Cell Bea~. Symmetrical about One Axis. All Katerial
effec:ive in Resisting Sendi~g Stresses
A15.3 Single Cell ~ T~o flange Beam. Constant Shear Flov Vebs
Al5. ~ Shear Cen:re of a Single Cell - T¥o Fla~ge Beam
Al5,5 Single Cell - Three Flanged Beam. Constant Shear Flo~ Vebs
AlS.6 Shear Centre of Single Cell - Three Flange Beam. Constant
Shear Flo. Vebs
AlS,7 Single Cell - Multiple Flange - One Axis of Symmetry
A1S.8 Single Cell - Unsymmetrical - Multiple Flange
A1S.9 T~o Cell - Kultiple Flange 3eam. Symmetrical about One Axis
A1S.10 Three Cell - Multiple Flange 3eam, Symmetri~al about
One Axi s
AlS.ll Shear Flov in Beam vith Hultiple Cells. Method of Successive
Appr:oxi.na~ions
A1S.12 Example ?c~blem Solution
A15.ll Example Problem Solution
AlS.14 Use of Successive Approximation Method for Multiple
Cell BeaTIS ~hen slJbjected to Combined Bending and
Torsional Loads
A15.1S Shear Flo. in Cellular B~a~s ~ith Vaciable ~oment of Inertia
A15.l6 Determtnation of (~e flexura: Shear Flo~ Distribution by
constdec:ng the changes in ?lang~ ~oads
AIS, :7 Example ?:-JDlell
A 15 .. " Shear flay in Tapered Sheet ?anel
A15.l9 Example ?roblem of Shear FloY in Taper:ed ~ultiple Plange
Single Cell 3eam
A15.20 Problems
A16 KRMBRANP, STRESSES IN PRESSURE YESS~LS
A16.l Introduction
A16.2 Hembrane Equations of Equilibrium. Shells of Revolution under
Rotationally Symmetric Pressure Loadings
A16.3 Applications of Simple Pressure Vessels
A16.4 Displacements, Boundary Conditions and Local Bending in
Thin-Valled Shells
A16. 4 Special Problems in Pressurized Cabin Stress Analysis
A16. 5 Shells of Revolution under Unsymmetrical Loadings
A17 BENDING OF PLATES
A17.1 Introduction
A17.2 plate Bending Equations
A17.3 An Illustrative Plate 3ending Analysis
A17.4 Compilatioo3 of Results for Plate Bending Problems
A17. 5 ~flection Limitations in Plate Analyses
A17.6 Hembrane Action in very Thin Plates
A17.7 Large Deflections in Plates
A17 . 8 Considerations in the Applications of Large Deflection Plate
and Membrane Analyses
A18 THEORY OF THE INSTABIUTf OF COUJXNS AND THIN SHEETS
A18.1 Introduction
A18.2 Combined Bending and Compression of Columns
AlB.] Elasti~ Stability of a Column
A18.4 Failure of Columns by Compression
Ala.S Buckling Loads of Colu~ns ~ith various End Conditions
A18.6 Inelastic Buckling. Int~oduction
A18.7 Per teet Column. Reduced Modulus Theory
A18.8 Imperfect Colu~n. Tangent-Hodulus Theory
THEORY OF Tfffi E.LASTIC INSTABILITY OF THIN SI!EETS
"8.9 IntLoc.uct-'.on
A18.10 PULe Bending of Thin Pla~es
A18.11 The Differential Equation at the Deflection SurEace
A18.12 Strain Ener 6 J in Pure Be~di~g of Plates
A18.13 Bending of R~ctangular Plates
A13.14 Combined Bending and Tension or Compression of Thin plates
A18.15 Strain Energy of Plates due to Edge Compression and Bending
A18.16 Buckling of Rectangular Plates vith va~ious Edge Loads and
Support Conditions
Al3.17 Buckling of Simply Supported Recta~gular Plates Uniformly
Compressed in One Direction
A1B.18 Buckling of Simply Supported Rectangular Plate compressed
in r~o Perpendicular Directions
A18.19 Buckling of Simply Supported Rectangular Plates under
combined Bending and Compression
A18.20 Inelastic Buckling of Thin Sheets
STRF"..NGTH OF COLUMNS VlrS STABLE CROSS-SECTIONS
A18.22 Hethods of Column Failure. Column Equations
AIB.23 Column End Restraint. Fixity Coefficients. Column Effective
Length
A18.24 Design Column Curves for various Haterials
A:a.25 Selection of Materials for 2levated Temperature Conditil)ns
A:8.26 Exampl~ Problem
A18.27 Comprehensive Treatmer.t of Column Strength
A18.28 SDme Mechanical and Physical Properties of a few Typical
Aircraft Haterials
A19 INTRODUCTION TO VLNG STRESS &~YSIS
A19. 1 Typical ~ing Structural A.rangement

A19.2 Factors that Influence .:~g Structural Arrangements
A19.3 Ving Strength Requirements
A19.4 ~ing Stress ~~alaysis ~ethods
A19.5 ~xample. Three Flange - Single Cell Viog
A19.6 Example. Hetal covered ~ing ~ith Single External Brace Strut
A19.7 Single Spar - Cantilever ~ing - Hetal covered
A19.8 Stress Analysis of Thin Skin - Multiple Stringer Cantilever
Ving. Introduction and Assumptions
A19.9 Physical Action of ~ing Section in Resisting External Bending
Forces from Zero to Failing Load
A19. 10 Ultimate Design Strength Requirement
AI9.11 Effective Section at Failing Load
A19.12 Example
'19.13 Bending and Shear Stcess Analysis of Tapered - Multiple
Stringec Cantilever ~ing. Unsynmetrical Beam Method
A19.14 Bending and Shear Stress Analysis of Two Cell Multiple String~r
Tapered Cantilever .ing
A19.15 Bending Strength of Thick Skin
A19.16 Example Proble~
A19.17 Application to Prac,ical Ving Section
A19.13 Shear Lag Influence
A19.19 Application of Shea~ Lag Approximation to Wing vith Cut~out
.-\19.20 Approximate Shear Lag Effect in gea~ Regions vhere Large
Concentrated Loads are Applied
AI9.21 Approximate Shear Lag Effect for sudden change in
Stringer Area
.\:'1.22 Problems
INTRODUCI'ION" TO FUSELAGE STRESS A.."l:AJ,YSIS
..'0. 1 General
. '0.1 Loads. Basic StruC:"Jre
.~.~ O. 3 Stress Analys~s ~e:~ods. Effective Ccoss-section
.<>').4 Example Prob>Jl
.<:0.5 Ultimate 3end:~g S,=e~gth of Fuse~age Section. Example
~ ~O. 6 Shear Flow A..L~:ysi$ for Fuselage Structures
.\20.7 Tapered Cir~u~ar ?uselage vith UnsY~we:rical Stringec Areas
\ 20.8 Disconuities - Shear Lag - Pressurization Stresses -
Combined Stresses
.,20.9 Proble!lls
LOADS AND S"tRRSSES ON RIBS AND FRAKES
.-\21.1 Introduction
A2 ~ • 2 Types of ~ing Const.uction
.J..:': 1 • 3 Dist.ibution of Conce~trated Loads to ~hin Sheet Panels
,... 2 ~ . 4 Rib for Single Cell - Tva Flange Beam
A21.5 Stresses in Rib Eor Three Stringer Single Cell Beam
,\21.6 Stress Analysis of Rib for Single Cell Multiple
Stringer ~ing
,,21.7 Rib ~ads due to Discontinuities in ~ing Skin Covering
,.1,21.8 Viog vith Cut-out subjected to Torsion Example
An.9 ~ing vith Cut-out subjected to Bending and Torsional
Load s - Examp:';·
\.21.10 Fuselage Frames
,-1.21.11 SlJpporting Boundary Forces on Fuselage Frames
.\21.:2 Calculation of Fra~e 30undary Support;ng Fo.ces
,-\21.~) Problems
A21 ANALYSIS OF SPECIAL \lING PROBLEMS

An.l In t roduc t ion
A21.2 Stresses around a Panel Cut-out
A21 • ] Shear Lag Analysis of Box Beams
An .4 Stress Analysis of a Box Beam ~ith a Cut-out
A22.5 Analsis of S~e~t Box Seam
A23 ANALYSIS BY TM KETHOD OF DISPLALEXF1ITS
A23.l Introduction
A23.2 Structural Co-ordinates
A23,J Theoretical Basis of the ~ethod of Displacements
A23.4 Tabulation of £le.ent Properties
A21.5 Transformation of Co-ordinates
A23.6 Calculation of Displacements and Stresses
A23.7 Hethod of Gaussian Slimination
A2l.a Gauss-Seidel Iteration
A2J.9 Calculation of Nodal Forces for Loads Acting betyeen
Nodal Faints
A23.10 Structural Modelling using Finite Ele~ents
A2J.II Delta Vi"g Example Proble.
A23.12 Accuracy, Convergence and Error Bounds
A2J.13 Hanagement of Problem for the Computer
A24 THE J-DIXENSIONAL EQUATIONS OF THEJC~OKLASTICITY
A24.1 Assumptions
.\24,2 Formulation of the )-Dimer,sional Squa.::ions of Thennoelasticity
A24.3 The Equilibrium Equations
A24.4 Strain - Displacement Eql.Ja:!orls
A24.5 Stress - Stratn Relations
A24.6 General gquations of 3-Dimensional The~moelasticity
A24.7 The Compa'ibility Equations
A24.8 Boundary Equations
A24.9 General ~otes
A25 THE 2-DIXENSIONAL EQUATIONS OF ELASTICITY &~ THERHOKLASTICITY

A25, 1 Introduction
A2S.Z Plane St::-~ss
.\25.3 Derivation of Equat:ons
A25.4 Formulation of Plane Sl,ess Prob,eo
A25.5 Plane Strain
.\25.6 Kathe~atical Equali~y of the ?:a~e Stress and Plane
Strain Frobl."
).15. 7 Formula~ion of Plane Strai~ ?r0ble~
.;25.8 Concluding Discussion
.\25.9 Proble~s
A26 SELBCTED PROBLEMS IN ELASTICITY &~ TffER.~OELASTICITY
A26.1 Introduction
A26.2 One-Dimensional Problems
A16.3 Two-Dimensional Problems
Bl FLIGHT VEHICLE KATRRl.ALS AND THEIR PROPERTIES
B1.1 Introduction
B1.1 Failure of Structures
B1.l Geoeral Types of Loading
B1.4 The Static Tension Stress-Strain Diagram
31. 5 The Static Compression Stress-Strain Diagram
B1.6 Tangent Modulus. Secant Hodulu$
B1.7 Elastic - Inelastic Action
B1.8 Ductility
B1.9 Capaci ty to Absorb Enecgy. Res i 1 i ence. Toughness
B1. 10 Poission's Ratio
B1.11 Construction of a Stress-Strain Curve through a given Yield
Stress by using a kno~n Test Stress-Strain Curve
B1. 12 Non-Dimensional Stress-Strain Curves
31. 13 Influence of Temperature on Katerial Properties
81. 14 Creep of Materials
B1. 15 The General Creep Pattern
B1. 16 Stress-Ti~e Design Charts
B1.17 EfEect of Ti:ne of Exposure
B1. 18 Effect of Rapid Rate of Heating
81. 19 Ge~er~l ~ffect of Lo~ Tempera~u~e upon ~aterial Properties
51.20 Fatigue of Materials
gl. 21 Effect of Irnp~ct Loading on Hatecial Properties
B1. 22 Examples of some Results of Impact Testing of Materials
82 KECHANTCAL MID PHYSICAL PROPER7IES OF M.....'""TALLlC MATF.R.IALS FOR
F'LlG!IT VEHICLE STRUC11JRES
Cl cmrnINW STRESSES. THEORY OF nSLl) AND ULTIMATE FAILURE
Cl.i Uniform Stress Condition

C1.2 Shearing Stress on Planes at ~ight Angles
C1.] Simple She~r produces Tensile a~d Compressive Stress~$
C1.4 Principal Stresses
C1.5 Shearing Stresses reulting from ?cincipal Stresses
C1.6 Combined Stress Equations
C1.7 Mohr's Circle for )etermination of ?~incipal Stresses
C1.B Components of St:ess from Principa: Stresses by Mohr's Circle
C1.9 Example Problems
C1. 10 Triaxial or Three-Dimensional Stresses
C1. 11 Principal Strains
C1.12 Elastic Strain Energy
C1.I] Str-uctural Design Philosophy. Li;lli: and Ultimate Loads.
Factors of Safety. Hargin of Safety
Cl,14 Required Strength of Flight Stcvctures
C1. 15 Determination of the Ultimate St~e~g:h of a Structural Kember
under a Combined Loading System. Stress Ratio-Interaction
Curve Hethod
Cl,16 Determination of Yield Strength of a Structural ~embe~ under
a Combined Load System
C 1. 17 The Octahe,jral Shear Stress ~heor;
C1,18 Example Proble!ll
c2 STRE.."tGT8 OF COLUXNS '11TH STABLE CROSS-SECTIONS
C2.1 He~hods of Column Failure, Co:~~n ~~uat~0;lS

C2,2 Column End Restraint. Fixity CoeUidents,
Column Effective Length
C2,] Design Column Curves for various Materials
C2.4 Tangent-Modulus from Ramberg-Osgood Equation
C1.5 Non-Dimensional Column Curves
c2,6 Strength of Col'Jmns ~ith Variable Cross-Section or Koml~nt
of Inertia
C2. 1 Design Column Cur'n"s far Cal'...lr.1ns . . ith Non-Unifor:n erns::; Section
C2.3 Column Fixity Coefficients for use vith Columns vith Elastic
Side Restraints and kno~ Bending Restraint
C1.9 Selection of Materials for Elevated Temperature Conditions
C1. 10 Example Problems
C1.11 Solution vithout using Colu~ Curves
c2.11 Strength of Stepped Column
C1.13 Column Stre~gth vith knovn End Restraining Moment
C2.14 Columns .... ith Elastic Lateral Supports
C2. is Pr-oblems
C3 nELD AND ULTIMATE STRENGTH IN BENDING
CJ.l Illtroduc t ion

C3.1 Basic approach to finding the Bending Strength of Members
CJ.J Bending Strength of a Solid Round Bar
CJ.4 The Cozzone Simplified Procedure
CJ.5 Design Curves for finding the Modulus of Rupture
CJ.6 General Accuracy of Method
CJ.7 Example Problems in finding Bending Strength
CJ.B Complex Bending. Sy~metrical Section. Moment Vector not
Parallel to Principal Axes
CJ.9 Section vith One Axis of Symmetry vith ~oment Vector not
Parallel to Either Axes
C3.10 Unsymme:rical Section .... ith no Axis of Sy:n;netcy
C). 11 Alternate more Exact Method for Complex Bending
C). 12 Strellgth under Combined Bending and Flexural Shear
C3.13 Strength under Combined Bending Flexural Shear and
Axial Compression
C3.14 Further Values of f and f
C4 STR&\lGTll A.'ID DESIGN OF ROUND, S1'RRAMl...INE, OVAL AND SQUARE TUBING
IN TR."ISION, COMPRESSION, B~ING, TORSION A.'ID COKBINED LOADINGS
C4.1 IntroductiQn
C4.2 Design for 7ens:on
C4.) Design for Com?ression
C4.4 Column Forllulae for Round Steel Tubes
C4.5 Column For~ulae for Aluminium Alloy Tubes
C4.6 Column Formulae for Hagnesium Alloys
C4.7 Short Column Equations for Other Materials
C4.3 Column Failure due to Local Failure
C4.9 Design Column Charts
C4.10 Section Properties of Round Tubing
C4.11 General Facts in Tubing Design
C4.11 Effect of ~elding of Steel Tubes upon the Tension and
Column Strength
C4.1J Illustrative Proble~s in Strength Checking and Design of Round
Steel Tubes as Columns and Te~$ion Hemoers
C4.14 Illustrative Problems using Aluminium Alloy and Magnesium
Round Tubes as Column and Tension Me~bels
C4.15 Strength of Streamline Tubing
C4.16 Strength of Oval and Square Shaped Tubes in Compression
ULTIMATE BE.NDING STRENGTH OF ROUND TUBES
C4. 1 7 Charts for find~ng the Modulus of Rupture Stress
C4. ~8 Problems involving the Be~d:ng Strength of Tubes
ULTIMATE TORSIONAL STR.RNGTH OF ROUND TUBES
C4. 19 Torsional Hodulus of Rupture

C4.10 Torsional Modulus of Rupture Curves
C4.21 Problems Illustrating the use of Torsional Modulus of
Rupture Curves
STR.RNG1'H OF ROU1ffi TUB&S Ul'fDE..R COKBINRD LOADING
C4.22 Combilled Bending and Compression

C4.23 rllustrative Problem involving Combined Bending
and COilJpcession
C4.23a Combined Bending and Tension
C4.24 Combined Sending and Torsion
C4.24a Ultimate Strength in Combined Compression, Bendi~g
and Torsion
C4.25 Ultimate Strength in Combined Bending and Flexural Shear
C4.26 Ultimate Strength in Combined Compression, Bending,
Flexural Shear and Torsion
C4.27 Ultimate Strength in Combined Tension and Torsion
C-4.28 Ultimate Str~ngth in Combined Tension, Torsion and
Internal Pressure
C5 BUCKLJNG ~~ OF FLAT SHEET IN COMPRESSION,' SRKAR,
BKNDING AND UNDER cmmnnm STRESS SYSTEMS
CS.l Introduction
CS.2 ~quation Ear Elastic Buckling Strength of Flat Sheet
in Compression
cS.3 Buckling Coefficient
C5.4 ~quation for Inelastic Bending Strength of Plat Sheet
in Compress-lon
CS.S Simple Problems to Illustrate the use of Buckling Curves
C5.6 Cladding Redu~tion Factors
c5. j' 3uckling of Flat RectangulAr Plates under Shear Loads
cS .3 Buckling of Flar Rectangular Plates under Buckling Loads
cS.9 Combined Be~ding and Longitudinal Compression
':S.lO Combined Bending and Shear
CS.ll Combined Shear and Longi:udinal Direct Stress
(Tension or Compression)
CS.12 Combined Compression, Bending and Shear
c5.13 Illustrative Proble~s
C5.14 ?roblems
C6 ~AL BUCKLING STRESS FOR COMPOSITE SaAPES
CS.l Introduction
C6.2 Compressive Buckling Stress for Equal Flanged Elements
C-).3 Coropress:ve 3uckling Stress for Sinple Flange-~eb Elements
CS.4 Design Charts fo~ Local Buckling StreSS€S of some
Composite ~eb-Flange Shapes
c6.S Problems Illustrating the use of Charts
c6.6 Buckling of Stiffened Flat Sheets under Longitudinal
c.ompress::'or.
C7 CRIPPLING STRENGTH OF COMPOSITE SHAPES AND SHEET-STIFFENER
PANELS IN COMPRESSION.
SHEET gPF&CTIYR VIDTffS. COLrn-:N STRL~GTH
C 7.1 Introduction
C7.2 ~ethod 1. The Angle CSeedham) Method
C7 , 3 Des ign CJ:-V<2S
C7.4 ~e(hod 2. F~r St:-ess Crippli~g Ca~cula:10n (Gerard ~e:hod)
C 1.5 Stresses a~d Disp~ac:e~e~,s of riat ?la:es aEter 3uckling
under Condi tions of ~niform End Sho~:~~ing
C 7.6 The Gerard Equations for C~ippling Stress
C 7• 7 Correction for Claddinl
C 7.8 Haximum ~alues for Crippling Stresses
C7.9 Restraint produced by Lip5 and Bulbs
C7.10 Illustrative Problems in Calcula~ing Crippling Stress
C7.11 Introduction to Effective Sheet Vidths
C7.12 Effecti~e Sheet Vidths
C7.12 Efeective ~idth for Shee~ ~irh One Edg~ Free
C7.13 Effective Vidth vhen Sheet and Stiffe~e" have Different
Material Properties
C7.14 Inter-Rivet Buckling Stress
C7. 15 Illustrative ?roblem involv:ng Effective Sheet
C7. 16 Failing Strength of Short Sheet-Stiffener Panels
in Compression
C7.17 Failure by Inter~Rivet Buckling
C7.18 Failure of Short Panels by Sheet Vrinkling
c7.19 ~quation for irinkling Stress
c7.20 Rivet Criterion for Rivet Failure
C7.21 Illustrating Calculation of Short Panel Failing Strength
c7.22 General Design Limitations to prevent Secondary Failure
in Sheet-Stiffener Panel
C7.23 Y Stiffened Sheet Panels
C7.24 Example Problem Y Stiffened Panel
C7.25 Column Curve for H€mbe~s vith Unstable Cross-Sections
C7.26 Methods used for Deter~ining the Column Failing Stress
in the Transition Region
C7.27 Example Problems involving the ?inding of the Column
Strength of Columns ~fth Unstable Cross-Sections
(7.28 Column Strength of Stiffene~ ~ith Effective Sheet
C7.29 Sheet-Stiffener Panels ~ith Relatively 8e3vy Sheet Thickness
C8 BUCKLfNG STR.&..'K...TB OF KONOCOQlJE CYLINDERS
C8.1 Introductlon
C8.2 Buckling of Honoc.oque CirclJ1a~ Cylinde~s unde:- A..:dal
Compression
C8.3 Additiollal Convenient Desig~ Charts for Deter~ining
Compressive Buckling Stc€SS
C8.4 Plasticity Correction
C8.5 Buckling of !'lonocoque Ci~c:Jla~ ~yli:1de:-s under Axial Load and
Internal P:essuce
e3.6 IntrodiJction ~o BIJckli~g of Monocoque Circular Cylinders under
Pure Be.!ding
C8.7 Available Design Cur.es for 3e~ding Based on Ex?erimental
Results
eB.3 B1Jckling Strength of Circular Cylinde~s in Bending vith
Internal ?cessuce
C8.9 External Hydrostatic P~essure
C8.1O 8xternal Radial Pressure
C3.11 Buckling of Honocoque Circular Cylinde:-s under Pure Torsion
C8.12 Buckling under Transverse Shear
C8.13 B1Jckling of Circular Cylinders under Pure Torsion with
Internal Pressure
c8.14 Buckling of Circular Cylincers under Transverse Shear and
Internal Pressure
C3. 1 S Buckling of Circular Cylinde:-s under Combined Load Systems
C8.16 Illustrative P:-oblems for Finding the Buckling Strength
of Circular Honocoque Cylinders
C8.17 BtJckling of Thin-Valled (Honocoque) Conical Shells
C8.18 Allovable Compressive Buc~l:~g Stress for Thin~Valled
Conica.l Shells
C8.19 Additional Desig~ Buckling Curves for Thin-~alled
Conical Shells
c9 BUCKJ.ING STRE.'K:TB OF CURVED SHEET PA.'fELS. ULTIKATE STRENC"1'S
OF ST I PFENED CURVED SIIEET STRUCTlJRES
C9.1 Introduction
C9.2 State of the Theory
c9.3 Compressive Buckling Stress of Curved Sheet Panels
C9.4 Shear Buckling Str~ss of Curved Sheet Panels
c9.5 Buckling Stren&th of Curved Sheet Panels under Combined
Axial Compress ion and Shear
C9.6 Compressive Bl.lckling Stress of Curved Panels vith
Intecnal ?ressure
C9.7 Shear BIJckling Stress of Curved Sheet Panels with
Internal Pressure
C9.3 Example Proble::"s
C9.9 Introduction to the Ulti~ate Strength of SifEened Cylindrical
Structures
C9.10 Types of Instability Failure of Semi-Monocoque Structures
C9. , 1 The Determination of the Stresses in a Stiffened Cylindrical
Structure under External Loads
C9. '2 Panel Instability Strength
C9. 1] Calculation of Gene:al Instability
C9. 14 Buckling of Spherical P~ates under Unifor~ External Pressure
ClO OESrGN OF KETAL BRAMS. W'RB SBRAR RESISTANT (NON~BUClU.ING)

TYPE
C1G.1 Introduction to Flat Sheet 'Jeb 'iith Venical Stiffeners

ClO.2 Flange Design
CIa.} Allovale Flange Design Stresses
C10.4 Use of Stress-Strain Diagrams in Computing Beam Plange
Bending Allowable Design Stresses
C10.S Flange Strength (Crippling)
e1O. , ~eb Strength. Stable ~ebs
CI0.7 lIeb Bending and Shear Stresses
ClO.3 Sh~ar Resistance provided by Sloping Flanges
ClO. ~ E:Efect of Variable !'lament of Inertia on Flexural Shear
Str8sS Distribution
CI0. '.0 Stiffener Size to use ... : ~h Son-3uckling 'Jeb
CIO.:1 Notes on Beam Rivet Desig~
C10.:2 Loads on Rivets Attaching Reinforcing P:ates to Flange Member
C10.~) ~2b Splices
C10.:4 ~xample Rivet ?r0ble~
C10.:5 Sxample Problem. Strength Check of Bea~
C11).:5a Use of Longitudinal Stiffener to Increase Sending Buckling
Stress of ~eb Sheet
C10.l6 Other Types of ~eb Design
ClO."':.7 3eaded 'Jebs
ClO. :3 ~ebs ~ith Round Lightening Boles having ?or~ed Flanges
C1'J.:9 ~ebs vith Round Beaded Flange Lightening Boles and
:ntermediate Vertical Hale Beads
CI0.20 Example Problems
ell DIAGONAL S&XI-i&~SION FIELD DESIGN

Cll. 1 Introduction to Beams ~ith Flae Vehs
Cll.2 Elementa~y App~oxima~e Explanation of Tension Field
Beam Action
Cll. ] Elementary Derivation of App~oximate Tension Field
Beam Formulae
GIL 4 General Vagner Equations for Tens~on F:eld Beams
ell. 5 Modified ~agner Equa~ions for Ijse in Design
Cll.6 Shear carried by the 8ea~ Flanges
C11. ) Shear Load carried by .e~
C11 .8 Beams ~ith Parallel F:a~ges bu~ ~i:h O~lique ~eb Stiffeners
C11. 9 ~i-Jet Loads
Cll. >; ?la.Lge Loads
Cll.:1 LOiiC$ in i.leJ Vertical S,-:ffene~5
Cl1.:2 8eams ~ith Non-Pa~al:e: F:a~ges
C11. :] Sxample Problem
Cll.14 NACA Kethod of Strength Analysts for Se~i-Tension Field
Beams ~ith Plat Vebs
Cl1.15 Englneering Theory of Incomplete Diagonal Tens~on
C11.16 Pormulae for Stress Analysis
ell.: 7 Critical Shear Stress
Cl1.18 LO.'lding Ratio
e11.19 Diilganal T~:1sion F3C~O~ k
Cl1.20 Average and Maximum Stress in Upright or ~eb Stiffener
Cll.21 Angle of Dia&onal Tension
Cl1.22 Allo~able St~esses in Uprights
CII.23 Veb Design
Cl1.24 Rivet Design
Cll.25 Secondary Bending Moments in Flanges
C11.26 Shear Stiffness of ~eb
C11.27 Example Proble~ using NACA Method
C11.28 General Conclusion
C11.29 End Bay 3~fec~s
CII.3D Diagonal Tension in Curved ~eb Systems - Introduction
CII.31 General Discussion
Cll.32 Analysis of Stringer Systems in Diagonal Tension
C11.33 Allo~able Stresses (not Interactions)
Cl1.34 Example P"oble~
C11.35 Longeron Type Syste~
C11.36 Calculating S~resses in Longeron System
Cll.37 Example Problem
Cll. 38 Summary
Cl1.39 Problems
C11.40 Proble;ns
C12 SANDVlCU CONSTRUCTION AND DESIGN
e12. 1 Introdllct ion

el2.2 Basic Desibn P~inciples
C12.3 Nom.~ncla::ljce
C12.4 Basic Sand¥ich Structural ~ropecties
CI2.4.1 Sand~~ch Bending Sti~:ness
C12.4.2 Sand~ich Shear Sti£f~ess
C12.4.3 Core ?coperties
C12.4.4 Coce Jensi ty
C12.4.5 Core Shear Pro~ert£es
C12.4.6 Core ?la~~ise Com~ressiQn Properties
C12.4.7 Core ?lat~ise Tension P=operties
C12.4.8 Comb~~ed Load Strength
C12.4.9 Othe~ Isot~opic Cur,es
C12.5 Bone/comb Flat Panel Failuce ~odes
C12.5.1 Genera:
C12.5.2 Intracell 3uckling
C12.5.2.1 Cottlt'cessions
C12.5.2.2 Shear
C12.5.2.3 Biaxial Loading
C12.5.2.4 Biaxial and Shear Loading
C12.5.3 'tJrinkling
C12.5.3.1 Comp,css ion
C12.5.3.2 In-Plane Shear
e12.5.3.3 Biaxial Loading
e12.5.3.4 Biaxial and Shear Loading
C12.5.4 Sh€ar Crimping
c12.5.5 Gene~al Instability - Uniaxial Compression
C12.5.6 Gene:al Shear Buckling Allo~ables
C12.5.7 Biaxial Com~ression
C12.5.8 Biaxial Co~pression and Shear
C12.6 Pailu~e Modes in Curved Honeycomb ?a~els
C12.6.1 Gene"-3l-
CI2.6.2 Unia;<ial Compression - Curved Panel
C12.6.3 Uniaxial Compresseon (Cylinder)
CI2.6.4 Torsion
el2.6.5 Vnifocttl Ex:ernal La:eral Pressure
C12.6.6 Bending :--lament
C12.6.7 Lateral load and Axial Compression
el2.6.8 Axial Compression, Lateral Load and Torsion
e12.7 Buckling of Flat Panels ~ith Dissimilar Paces
e12.7.1 General
C12.7.2 Uniaxial Compression
C12.8 Core Shear Induced by Loads Applied at an Insert
C12.8.1 General
C12.8.2 Shea" Stress Distribution
C12.9 Corrugated Core Sand~ich Failure Modes
e12.9.1 Genecal
C12.9.2 Corrugated Sand~ich Panel General Buckling Coefficients
C12.10 Design of Sand~ich Structures
e12.10.1 General
C12.10.2 Dimpling of Sand~ich Facings
C12.1O.3 ~rinkling of Sandvich Facings
C12.1O.4 Shear Criwping of Sandvich Panels
C12.10.5 General Buckling of Sandvich Panels
C12.10.5 Design of Flat Sandvich Panels
C12.10.5.1 Detr~ining Facing Thicknesses and Panel Parameters
C12.1O.5.2 Detrmining Facing Thicknesses and Panel Parameters for
Honeycomb Panels in Shear
C12.10.5.3 Detrmining Pacing Thicknesses and Panel Parameters for
Corrugated Panels in Compression
C12.10.5.4 Detrrnining Pacing Thicknesses and Panel Parameters for
Corrugated Panels in Shear
C12.10.6 Design of Sandvich Cylinders
C12.1O.6.1 Detr~ining Facing Thicknesses and Panel Parameters for
Cylinder in Axial Com~ression
C12.1O.6.2 Detr~ining Pacing Thicknesses and Panel Parameters for
Cyll~ders in Torsion
e12.10.6.3 Detr~ining Facing Thick~esses and Panel Parameters for
Cylinders under Lateral ?reSSll[@
C12.l1.1-7 Exa;;;ple Problems
c13 FATIGlJE
e13.1 Introduction
C1 3.2 Stress Concentra~ions
e13.3 Pat~gue Data
C13.3.1 Unnotched Data
C13.3.2 Notched Jata
C13.3.3 Endurance Limi t
C13.3.4 Patigue Testing Kachines
C13.3.5 Test Scatter
C13.3.6 Load-N CUr'.'es
C13.3.7 Elevated Temperature
e13.5 Increasing Fatigue Resistance
e13.5.1 Introduction
C13.5.2 Stress, Strain and Plastic Deformation
e13.5.3 Effect of Residual StresS on Fatigue
e13.5.4 Mechanically Introduced Residual Stresses
e13.5.5 Residual Stresses by Diffusion Rardening
C13.5.6 Increasing Rigidity
e13.5.7 Combined Rigidity and Residual Stress
e13. 5.8 Load Relief at Fasteners
e16.6 Cumulative Fatigue 8amage
C13.6.1 In t roduc t i.on
e13.6.2 Loads
C13.6.3 Predicting LiEe
Ct1.6.4 St:-es5 :::n:eraction
e13.6.5 Fatigue Stcength :ndex
CI3.6.6 SIJmma~j of Life P~oduction
C1 3. 7 Safe :"ife, Fail Slo'-' a,'ld :;'ailsaEe
e13.7.1 Safe Life and Crack Propagation
C13.7.2 Pai lsafe
e13.8 Frequently Overlooked Problem Areas
C13.8.1 Fatigue Failures of Compressive Surfaces
e13.8.2 Heavy Forging Attachments on St~essed Parts
Cl1.8.4 Pastener Sizes
e11.8.5 Acoustic Fatigue
CU.9 r:heck [,ist
C13.10 Concluding Remarks
FITTINGS AND CONNECTIONS. BOLTED ~VD RIVETED
ot. 1 Introduction
D1.1 Economy in Fitting Des~gn
01. ) fitting Design Loads. Kini~um Hargins of SQfety
01. 4 Special or Higher Factors of Safety
D1.5 Aircraft Bolts
D1.6 Aircraft Nuts
D1.7 Bolt Shear, Tension and Bendi~g Strengths
01.8 801ts in Combined Shear and Te~sion
01. 8a Bushings
01. 9 Single Bolt Fitting
01. 10 ~ethods of Failure of Single Bolt Pitting and the Allo~able
Fai ling Loads
D1.11 Lug Strength Analysis under Axial Loading
D1. 12 Lug Strength Analysis under Transverse Loading
01. 1) Lug Strength Analysis under ObI ique Loads
01. 14 Bolt Bending Strength
D1. 15 Illustrative Problems
D1. 16 Bolt Loads for Multiple Bolt Fittillg. Bolt Sizes Different
Concentric Loading
D1. 17 Kultiple Riveted or Bolted Joints subjected to Eccentric Load
01. 13 Types of Rive~s
[)1.19 Rivet Material
Dl.10 Sttength of Ri~ets. ?rotruding R€dd Type
01. 2 1 Strength of Rivets. F~ush Type
D1 .22 Blind Rivets
D1,23 Riveted She~C Splice In~ornation
D1. 24 Illustrative Frob lens
Dl. 25 Rivets in Tension
D1. 26 ~ivet Tension Strengchs
D2 'lEUl8D COmiECTrONS
01. 1 Introduction
J2.2 Gas 'Jelding
n. ] General Notes on the ?ra~tical Design of _elded 10ints
)2.4 General Types o~ 'Je'...ded Steel fitting Units
02.5 Electric Arc ~elding
D1.6 Bffect of Velding on Base ~etal
02.6 qeld-Hetal Allo~able Stress
'J2.7 Allo~able Load for Velded Sea~s
C2.8 Brazing
Li 2 • 9 ~el~~ug of Aluminium Alloys
01.10 Illustrative Proble~s involvin~ Velding
01. 11 Spa t Veld i ng
D2.12 Spot ~elding of Al~~iniurn Alloys
D2.1) Spot Strengths
02 .14 Reduction of Tensile Str~ngth of ?arent Metal due to
Spot 'Jelding
D3 SQKE IMPORTANT DETAl LS IN STRUCruRAL DESIGN
·lJ.1 Introduction
'.1 3. 2 Shear Clips
)3.3 Tension Clips
c, .1. '
LI~.<t .Joggled Ke:-;joes
D3.5 Fi llers
D}.6 CllL-O"rS in 'Jebs or 5kin p;onels
OJ.7 Special Cases of B~a~ Design
D3.3 Structural Skin ?anel Deta~ls
01.9 Additional Important S~~uc:ural Details
\PPfl.NVIX A - ELEMENTARY KATRlI Al).::;ERRAE
SPECIAL BOOKS NOW AVAILABLE
As explained in the preface, instead of changing to a 2 volume edition in order to include 380
pages of new material, it was decided at the last minute so-to-speak, to remain with one volume and
issue this new material on three different subjects as separate books. A description of these 3 books
follows:
ENGINEERING COLUMN ANALYSIS

By
WILLIAM F. McCOMBS
Senior Structures Design Specialist
Vought Aeronautics Div. of LTV Corp., Dallas, Texas
"Engineering Column Analysis" has been written to provide the engineer with practical
methods for analyzing nearly all types of columns and beam-columns, including those on elastic
supports and these loaded into the plastic range of the material. The book has been written in such
a manner that no prior study of elastic stability theory is necessary. The only prerequisite is an
understanding of elementary strength of materials as given in undergraduate engineering courses.
The numerous methods and techniques of analysis are presented for "hand solutions" by the
engineer, but these can, of course, be arranged for analysis by computer if desired. The broad scope
of the book is indicated by the following list of Part titles. References are also provided for further
reading on specific subjects. Much of the material has been used by the author in classes for prac-
ticing engineers in the aircraft industry. It is often surprising to one how easily many seemingly
complicated columns or beam-columns can be analyzed for engineering purposes.
Part Title
A A Survey of Various Column Problems.

B Stability and Critical Loads.
C Columns of Constant Section and Single Span.
D Columns of Variable Section and Single Span.
E Beam-Columns of Constant Section and Single Span.
F Multiple Beam-Columns of Constant Between Supports.
G Multispan Columns of Constant Section Between Supports.
H Analysis of Truss Structures for Stability and Strength.
I Elastic Weight Numerical for Calculations Beam Distortions.
J Columns of Varying Section and Single Span.
K Beam-Columns of Varying Section and Single Span.
L Multispan Columns and Beam-Columns of Varying Section.
M Beam-Columns on Elastic Supports.
N Columns on E"lastic Supports.
o Torsional Buckling of Free Columns (Struts).
P Torsional Buckling of Restrained Columns.
Q Special Topics.
DESIGN AND ANALYSIS OF FILAMENTARY COMPOSITE STRUCTURES
By
Lloyd E. Hackman
Consultant on composite structures engineering.
The Author has presented a basic review of a new structural technology. He has also presented
an insight into the development of the technology as well as the application of the technology in
practical structural elements.
The Author presents the application of the design and analysis methodology from the micro-
mechanics of fiber matrix interaction to the use of composite lamination in various types of struc-
tural elements and methods of joining. The application analysis also includes the description of
failure modes, a statement of failure criteria, and test methods for the experimental evaluation of
composite materials.
The Author demonstrates the use of all analysis methodology and failure criteria with example
problems taken from ten years of experience in the design with composite materials in the aerospace
industry.
TABLE OF CONTENTS
C13.l Introduction
C13.Ll The Development of a Technology
C13.L2 The Material Concept
C13.L3 The Design Approach to Composite Structures
C14.L4 Nomenclature
C13.2 General Relations for Orthotropic Laminates

C13.2.l Elastic Relations
C13.2.2 Strength Relations of Fabric Laminates
C13-2-3 Properties of Various Constructions
C13.2.4 Elastic Properties of Oriented Laminates
C 13.2.5 Elastic Properties of Cross Laminates or Parallel Laminates
C13.2.6 Isotropic Laminates Made with Orthotropic Laminations
C13.3 Relationships of Unidirectional Laminate Materials

C13.3.l Elastic Relationships for laminae
C13.3.2 Strength Relationships for a lamina of a Rectangular Array
C13.3.3 Strength Relationships for a lamina of a Hexagonal Array
C13.3.4 Multi Laminate Properties and Stresses
C13.3.5 Compressive Strength of Unidirectional Laminates
C13.3.6 Effect of voids on Compression Strength
C13.4
C13.4.l Laminate Failure Criterian
C13.4.2 Failure modes and Test Methods
C13.4.3 Nonlinear characteristics of Laminate and Failure Modes
C13.5 Stability of Filamentary Composite Structural Elements

C13.5.l Flat Rectangular Panels
C13.5.2 Honeycomb Sandwich
C13.6 Joining of Composite Materials

C13.6.l Bonded Lap Joints
C13.6.2 Mechanical Fastened Joints
ANALYSIS AND DESIGN
OF
MISSILE STRUCTURES
By
J. I. ORLANDO, Branch Chief
Missile and Space Division
Douglas Aircraft Company
And
J. F. MEYERS, Section Chief
Missile and Space Division
Douglas Aircraft Company
This section represents an elaboration for Prof. E. F. Bruhn's book, "Analysis and Design of
Flight Vehicle Structures." It has been designed to introduce the student and beginning engineer to
the general field of missile structures and to give a limited presentation of the preliminary load,
stress analysis, and structural design practices of typical boost missiles.
The first portion is both analytical and descriptive in nature and serves to acquaint the reader
with current practices in configuration, material usage, structural techniques, and design factors.
The main body, or Parts E1.7 and E1.8, takes a hypothetical multistage vehicle, derives the critical
flight load conditions, and indicates analysis for these loads. The analysis techniques derive directly
from the body of Professor Bruhn's book, "Analysis and Design of Flight Vehicle Structures"; hence
giving the reader examples of practical usage of that data for missile design.
TABLE OF CONTENTS
Part ELI HISTORY, TERMINOLOGY, WEIGHT DISTRIBUTION
Part E1.2 PROPULSION SYSTEM FUNDAMENTALS
Part E1.3 MISSILE AND SPACE VEHICLE SPECIFICATIONS
Part E1.4 MISSILE AND SPACE VEHICLE TYPICAL STRUCTURES
Part E1.5 MATERIALS FOR MISSILES
Part E1.6 STRUCTURAL DESIGN FACTORS
Part E1. 7 MISSILE LOADS ANALYSIS
Part E1.8 MISSILE STRESS ANALYSIS
Part E1.9 CONCLUSIONS
APPENDIX: Reference Design Charts and Material Taken from

"Analysis and Design of Flight Vehicle Structures" by Bruhn.
Launch Escape
----""II
System
Command Module ----T"

APOLLO
SPACECRAFT
1--- S -IVB Stage
Space exploration requires

- - S -II Stage
a tremendous launch veh i-
cle, as illustrated, by the
SATURN V Appollo/Saturn ve hi c Ie.

LAUNCH VEHICLE
Note size of vehicle when
compared to size of a man.
- - S-IC Stage Such large vehicles present
many challenging problems
to the structures engineer.
(Drawing - Courtesy of N.
American Aviation Co.)
APOLLO/SATURN VEHICLE
1959 - 60 BEGINNING OF THE "JET" AGE OF AIR TRANSPORTATION - 550-650 MPH
FIRST FLIGHT OF DOUGLAS DC - 8

MAY 30, 1958
MARTIN "404" TRANSPORT CONYAIR "340" TRANSPORT BOEING "STATOCRUISER"
1928 BOEING MODEL 80-A 1930 FORD TRI-MOTOR 1931 BOEING "MONOMAIL" 1933 LOCKHEED "ELECTRA"
1903 FIRST FLIGHT BY 1912 GLEN. L. MARTIN 1918 STANDARD AIRCRAFT 1926 FOKKER TRANSPORT
WRIGHT BROTHERS IN HIS BIPLANE CORP BIPLANE
L....-_ PHOTOGRAPHS ILLUSTRATING PROGRESS OF AIRCRAFT DESIGN FOR AIR TRANSPORTATION - 40 TO 175 MPH
GENERAL DYNAMICS F-lll AIR FORCE/NAVY
TACTICAL FIGHTER. SPEED 2000 M. P. H.
VARIABLE POSITION WING (1965)
SIX AIR FORCE SUPERSONIC FIGHTERS 1959 N. AMERICAN (1964) N. AMERICAN

SHOWING VARIOUS WING SHAPES AVIATION AVIATION
(Picture from United Aircraft Publication "BEEHIVE" Spring 1958) X-15 HYPERSONIC B-70. SPEED
RESEARCH AIRPLANE 2000 M. P.H. (PLUS)
1930 MARTIN P3M-l NAVY 1933 VOUGHT-NAVY SU-4 1936 GRUMMAN-NAVY SF-l 1939 VOUGHT-SIKORSKY S02U
1908 "WRIGHT" AIRPLANE 1918 CURTIS ':JENNY" 1921 BOEING-ARMY 1929 CURTIS P-6 "HAWK"
FIRST ARMY AIRPLANE ARMORED ATTACK AIRPLANE
SOME PRE-WORLD WAR II AIRCRAFT
I----PHOTOGRAPHS ILLUSTRATING TREMENDOUS PROGRESS OF MILITARY AVIATION--.....
LUNAR ORBITER. Will orbit the moon taking
pictures (1966)
PIONEER. Interplanetary Probing

up to 90 million miles (1965)
-- MARINER. To obtain scientific information

on Mars and Venus (1964-65)
PROJECT GEMINI. Two-man Spacecraft. RANGER. Photographed the Lunar Surface.

(1965) (1964)
The Age for Spacecraft is just dawning. Regardless of the type of Spacecraft,
its creation will involve the work of the structural designer.
Drawing Showing General Layout
of Modern Missile
(Courtesy Douglas Aircraft CO.)

10
27
26
25
24
12
23
22
16
NASA DELTA SPACE VEHICLE
MODEL DSV-3B
DDUG~
1. S-17 (OSO) Spacecraft 7. Retro-System Nitrogen Bottle 12. Fuel Tank 18. Stabilizing Fim 24. Guidance Antenna
2. Fairing 8. BTL 600 Series Guidanee 13. Wiring Tunnels 19. 1st Stage Telemetry 25. 2nd Stage Telemetry 21
3. Spin Table 9. Fuel Tank 14. Telemetry Antenna 20. Anti-vortex BajJle Antenna
20
4. Spin Rocket Holder 10. Helium Sphere 15. Oxidizer Tank 21. Anti-slash Baffles 26. ABL X 248-A5DM Solid
5. 2nd Stage Programmer 11. Aerojet AJIO-H8D 16. Turbopump 22. Interstage Structure Propellant M owr
6. Eledronics Padroge Assembly 2nd Stage Engine 17. Rockeldyne 1st Stage Engine 23. Oxidizer Tank 27. Spaeecraft Attaeh Fitting 19
18
I-
Ql c
en en
c - ."
.;: Z Ql
+-
Vl 0"
--0
a.. I-
0::: ::>
Ou
U ::>
Vll:
= ."
u-=E
0- I -
00 U
I .::
u.. 0
ti:-o
~ .~
0::: Q..
U >...
o:::+-
~ (G
+-
a.. ~
Ot;
0:::2
:1::::
I-
0::: en
0.=
Z ~
~
o
LIFTOFF AT THE CAPE - the huge Saturn V S-IC SEPARATION - first stage drops away, S-IVB IGNITES - $-II stage is jettisoned;
in a 71 /,-million-pound thrust toward the stars. and the Saturn S-II second stage is now ignited. S-IVB hurls Apollo spacecraft into earth orbit.
2 3
TRANSLUNAR INJECTION - S-IVB stage pushes APOLLO - S-IVB SEPARATION - the adapter FREE FLY-AROUND - Apollo craft turns
Apollo spacecraft out of earth's gravity pull. panels are blown free; Apollo craft separates. around in space to dock with the waiting LEM.
4 5 6
~.j/
..."",~
..
..... '\
~
. . . . ..
LEM EXPOSED - two crewmen enter LEM LUNAR ORBIT - Service Module engine THE DESCENT-Apollo croft remains in
for an equipment checkout; S-IVB is left behind. retro-fires to put craft in aO-mile lunar orbit. orbit while two nstronouts guide LEM to the moon.
7 8 9
LUNAR EXPLORATION - during moon stay MOON BLAST-OFF - the LEM employs its LUNAR RENDEZVOUS - the LEM ClstrOllauts
fwo astronauts undertake scientific observations. expended descent stage (JS plotfonn for jift-off. line up ernft for critical dock'm(Jte m(Jneuver.
10 11 12
HOMEWARD BOUND - LEM stays behind in SIM JETTISONED - prior to entry into REENTRY - pilot now orients Command
lunar orbit as three men head Apollo for earth. the earth's atmosphere, S/M is detached and left. Module so base heat shield takes friction heat.
13 14 15
In the tremendous task of landing on and returning

man from the moon, the structures engineer is
faced with many new and challenging design
requirements.
Appollo II with Astronauts Armstrong, Aldrin and

LANDING - CIM's three main chutes de-
ploy to bring the craft back down safely to earth.
Collins landed on the moon on July 20/ 1969.
16 (Photos - Courtesy of N. American Aviation Co.)
CHAPTER Al
THE WORK OF THE
AEROSPACE STRUCTURES ENGINEER
AI. I Introduction. scientific machine and the combined knowledge

and experience of hundreds of engineers and
The first controllable human flight in a scientists working in close cooperation is
heavier than air machine was made by Orville necessary to insure a successful product. Thus
Wright on December 17, 1903, at Kitty HaWk, the engineering division of an aerospace company
North Carolina. It covered a distance of 120 consists of many groups of specialists whose
feet and the duration of flight was twenty specialized training covers all fields of
seconds. Today, this initial flight appears engineering education such as Physics, Chemical
very unimpressive, but it comes into its true and Metallurgical, Mechanical, Electrical and,
perspective of importance when we realize that of course, Aeronautical Engineering.
mankind for centuries has dreamed about doing
or tried to do what the Wright Brothers It so happens that practically all the
accomplished in 1903. aerospace companies publish extensive pamphlets
or brochures explaining the organization of the
The tremendous progress accomplished in the engineering division and the duties and
first 50 years of aviation history, with most responsibilities of the many sections and groups
of it occurring in the last 25 years, is almost and illustrating the tremendous laboratory and
unbelievable, but without doubt, the progress test facilities which the aerospace industry
in the second 50 year period will still be more possesses. It is highly recommended that the
~~believable and fantastic. As this is written student read and stUdy these free publications
in 1964, jet airline transDortation at 600 MPH in order to obtain an early general uncier-
is well established and several types of standing of how the modern flight vehicle is
military aircraft have speeds in the 1200 to conceived, designed and then produced.
2000 MPH range. Preliminary designs of a
supersonic airliner with Mach 3 speed have been In general, the engineering department of
completed and the government is on the verge of an aerospace company can be broken down into six
sponsoring the development of such a flight large rather distinct sections, which in turn
vehicle, thus supersonic air transportation are further divided into specialized groups,
should become co~~on in the early 1970's. The which in turn are further divided into smaller
rapid progress in missile design has ushered working groups of engineers. To illustrate, the
in the Space Age. Already many space vehicles six sections will be listed together with some
have been flown in search of new knowledge of the various groups. This is not a complete
which is needed before successful exploration list, but it should give an idea of the broad
of space such as landings on several planets engineering set-up that is necessary.
can take place. Unfortunately, the rapid
development of the missile and rocket power I. Preliminary Design Section.
has given mankind a flight vehicle which, when
combined with the nuclear bomb, has the awesome II. Technical Analysis Section.
potential to quickly destroy vast regions of
the earth. While no person at present knows (1) Aerodynamics Group
where or what space exploration will lead to, (2) Structures Group
relative to benefits to mankind, we do know that (3) Weight and Balance Control Group
the next great aviation expansion besides (4) Power Plant Analysis Group
supersonic airline transportation will be the (5) Materials and Processes Group
full development and use of vertical take-off (6) Controls Analysis Group
and landing aircraft. Thus persons who will be
living through the second half century of III. Component Design Section.
aviation progress will no doubt witness even
more fantastic progress than occurred in the (1) Structural Design Group
first 50 years of aviation history. (~ing, Body and Control Surfaces)
(2) Systems Design Group
AI. 2 General Organization of an Aircraft Company (All mechanical, hydraulic, electrical
Engineering Division. and thermal installations)
The modern commercial airliner, military IV. Laboratory Tests Section.

airplane, missile and space vehicle is a highly
ALl
AI. 2 THE WORK OF THE AEROSPACE STRUCTURES ENGINEER
(1) Wind Tunnel and Fluid Mechanics Test The final results of the work of this
Labs. group are formal reports giving complete applied
(2) Structural Test Labs. load design criteria, with many graphs and sum-
(3) Propulsion Test Labs. mary tables. The final results may give com-
(4) Electronics Test Labs. plete shear, moment and normal forces referred
(5) Electro-Mechanical Test Labs. to a convenient set of XYZ axes for major air-
(6) Weapons and Controls Test Labs. craft units such as the wing, fuselage, etc.
(7) Analog and Digital Computer Labs.
THE WORK OF STRESS ANALYSIS AND STRENGTH GROUP
V. Flight Test Section.
Essentially the primary job of the stress
VI. Engineering Field Service Section. group is to help specify or determine the kind
of material to use and the thickness, size and
Since this textbook deals with the subject cross-sectional shape of every structural mem-
of structures, it seems appropriate to discuss ber or unit on the airplane or missile, and
in some detail the work of the Structures Group. also to assist in the design of all joints and
For the detailed discussion of the other groups, connections for such members. Safety with light
the student should refer to the various air- weight are the paramount structural design re-
craft company publications. quirements. The stress group must constantly
work closely with the structural Design Section
AI. 3 The Work of the Structures Group in order to evolve the best structural over-all
arrangement. Such factors as power plants,
The structures group, relative to number of built-in fuel tanks, landing gear retracting
engineers, is one of the largest of the many wells, and other large cut-outs can dictate the
groups of engineers that make up Section II, type of wing structure, as for example, a two
the technical analysis section. The structures spar single cell wing, or a multiple spar
group is primarily responsible for the multiple cell wing.
structural integrity (safety) of the airplane.
Safety may depend on sufficient strength or To expedite the initial structural design
sufficient rigidity. This structural integrity stUdies, the stress group must supply initial
must be accompanied with lightest possible structural sizes based on apprOXimate loads.
weight, because any excess weight has detri- The final results of the work by the stress
mental effect upon the performance of aircraft.. group are recorded in elaborate reports which
For example, in a large, long range missile, show how the stresses were calCUlated and how
one pound of unnecessary structural weight may the required member sizes were obtained to carry
add more than 200 Ibs. to the overall weight of these stresses efficiently. The final size of
the missile. a member may be dictated by one or more factors
such as elastic action, inelastic action, ele-
The structures group is usually divided vated temperatures, fatigue, etc. To insure
into sub-groups as follows:- the accuracy of theoretical calculations, the
stress group must have the assistance of the
(1) Applied Loads Calculation Group structures test laboratory in order to obtain
(2) stress Analysis and Strength Group information on which to base allowable design
(3) Dynamics Analysis Group stresses.
(4) Special Projects and Research Group
THE WORK OF THE DYNAMICS ANALYSIS GROUP
THE WORK OF THE APPLIED LOADS GROUP
The Dynamics Analysis Group has rapidly
Before any part of the structure can be expanded in recent years relative to number of
finally proportioned relative to strength or engineers required because supersonic airplanes,
rigidity, the true external loads on the air- missiles and vertical rising aircraft have pre-
craft must be determined. Since critical loads sented many new and complex problems in the
come from many sources, the Loads Group must general field of dynamics. In some aircraft
analyze loads from aerodynamic forces, as well companies the dynamics group is set up as a
as those forces from power plants, aircraft separate group outside the Structures Group.
inertia; control system actuators; launching,
landing and recovery gear; armament, etc. The The engineers in the dynamics group are
effects of the aerodynamic forces are initially responsible for the investigation of vibration
calculated on the assumption that the airplane and shock, aircraft flutter and the establish-
structure is a rigid body. After the aircraft ment of design reqUirements or changes for its
structure is obtained, its true rigidity can control or correction. Aircraft contain dozens
be used to obtain dynamic effects. Results of of mechanical installations. Vibration of any
wind tunnel model tests are usually necessary part of these installations or systems may be
in the application of aerodynamic principles to of such character as to cause faulty operation
load and pressure analysis. or danger of failure and therefore the dynamic
AI. 3
clnracteristics must be changed or modified in groups have a special sUb-group which is work-
order to insure reliable and safe operation. ing on design problems that will be encountered
in the near or distant future as aviation pro-
Tne major structural units of aircraft such gresses. For example, in the Structures Group,
as the wing and fuselage are not rigid bodies. this sub-group might be studying such problems
Thus when a sllilrp air gust strikes a flexible as: (1) how to calculate the thermal stresses
wing in high speed flight, we have a dynamic in the wing structure at su)ersonic speeds;
load situation and the wing will vibrate. The (2) how to stress-analyze a new type of wing
dynami,~ist must determine whsther this vibration strUi;ture; (:3) what type of body structure is
is serious relative to induced stresses on the best for future space travel and what kind of
wing stru~ture. The dynamics group is also materials will be needed, etc.
responsible for the determination of the
stabi Ii ty ami performan,;e of missile antj flight Chart 1 illustrates in general a typical
vehic;le glLldl:J.,;e and control systems. The make-up of the Structures Section of a large
dynamL;s group must work (~onstantly with the aerospace compa:1Y. Chart 2 lists the many
various test laboratories in order to obtain items which the structures engineer must be
reliable values of certain fa:;tors that are con'cerned with in insuring the structural
necessdry in m'lny theoreti'~'ll cal,;ulations. integrity of the flight vehicle. Both Charts
land 2 are from Chance-Vought Structures
Tnt<; WOH!\: OF THS SPECIAL PHOJ~CTS GROUP Design ~lnual and are reproduced with their
permission.
In g:meral, all the various techIlical
STRUCTURES SECTION
CHIEF OF STRUCTURES
STRUCTURES STRUCTURES
PROJECT MATERIALS
ENGINEERS GROUP
LOADS AND
METHODS
UNIT
STRUCTURES
TECHNICAL
GROUP
- NON-METALLIC
~-
UNIT
METALLIC
UNIT
AEROElASTlC RESEARCi:
UNIT AND
DEVELOPMENT
UNIT
AEROElASTIC
DESIGN
AEROELASTIC
RESEARCH
AND
DEVElOPMENT
STRESS
ANALYSIS
UNIT
- STRUCTURES
TEST
LABORATORY
GROUP
STRUCTURES
TEST UNIT
HYDRAULIC
AND POWER
-
PLANT TEST
DESIGN UNIT
WEIGHT
CONTROL
GROUP
DESIGN
ALLOWA8LES
DESIGN
METHODS
Chart 1.
- MACHINE
COMPUTATION
GROUP
Structures Section Organization

Chance- Vought Corp.
AI. 4 THE WORK OF THE AEROSPACE STRUCTURES ENGINEER
INKS TO STRUCTURAL INTEGRITY

•••• • ARE NO BETTER THAN THE WEAKEST LINK
MATERIALS OF
CONSTRUCTION
FASTENERS
WELDING
BONDING
PLATE AND BAR
FORGINGS
STIFFNESS CASTINGS
EXTRUS IONS STRESS
CRITERIA SHEET METAL
FLUTTER SANDWICH
ANALYSIS
PLASTIC LAMINATE SKIN PANELS
CONTROL SYSTEM STABILITY
BEARINGS BEAM ANALYS IS
PANEL FLUTIER-SKIN CONTOURS STRAIN COMPATIBILITY
STRA IN CONCENTRATION
CONTROL SYSTEM DEFLECTIONS
JOINT ANALYSIS
THERMAL EFFECTS BEAR ING ANALYS IS
BULKHEAD ANALYS IS
MECHAN ICAL VIBRAT IONS
FIniNG ANALYSIS
ROLL POWER-D IVERGENCE THERMAL STRESS
MECHANICAL COMPONENTS
AERODYNAMIC CENTER SHIFT
EXPERIMENTAL STRESS ANALYSIS
DYNAMIC RESPONSE
LOADS AND
ENVIROMENT MATERIALS AND
FLI GHT LOAD CR ITER IA QUALITY CONTROL
GROUND LOAD CR ITER IA
FLIGHT LOAD DYNAMICS DUCTILITY
LAUNCHING DYNAMICS STRESS-STRAIN
LANDING DYNAMICS
DYNAMIC RESPONSE HOMOGENEOUS MATER IAL
RECOVERY DYNAMICS RESIDUAL STRESS
FLIGHT LOAD DISTRIBUTIONS
COMPONENT ALLOWABLES HEAT TREAT CONTROL
INERTIAL LOAD DISTRIBUTIONS
YIELDING
FLEX IBIL1TY EFFECTS ANALYSIS FRACTURE
STRESS CORROSION
GROUND LOAD DISTRIBUTIONS
REPEATED LOAD SPECTRUMS UN IT SOLUTIONS STABILITY AT TEMPERATURE
FATIGUE
TEMPERATURE DISTR IBUT IONS INDETERMINATE STRUCTURES SPECIFICATION CONFORMANCE
LOADS FROM THERMAL WEAR, BRINELLING
DEFORMATIONS WINGANALYSIS BLUE PRINT CONFORMANCE
CREEP
PRESSURES-IMPACT TAIL ANALYSIS
DEFLECTIONS
FUSELAGE SHELL ANALYS IS
THERMAL EFFECTS
THERMAL ANALYSIS
STIFFNESS
DEFLECTION ANALYS IS
COMBINED LOADINGS
STIFFNESS
BUCKLING
Chart 2
From Chance- Vought Structures Design Manual
CHAPTER A2
EQUILIBRIUM OF FORCE SYSTEMS. TRUSS STRUCTURES
A2.1 Introduction. The equations of static force system pass through a common point. The
equilibrium must constantly be used by the reSUltant, if any, must therefore be a force
stress analyst and structural designer in ob- and not a moment and thus only 3 equations are
taining unknown forces and reactions or unknown necessary to completely define the condition
internal stresses. They are necessary whether that the resultant must be zero. The equations
the structure or machine be simple or complex. of equilibrium available are therefore:-
The ability to apply these equations is no
doubt best developed by solving many problems. ZF x =0 ZMJ. = oo
This chapter initiates the application of these ZF y =0 or ZM" =o } - - - - -(2.2)
important physical laws to the force and stress
analysis of structures. It is assumed that a
ZFz =0 ZM a =
student has completed the usual college course A combination Oi force and moment equations
in engineering mechanics called statics. to make a total of not more than 3 can be used.
For the moment equations, axes through the point
A2.2 Equations of Static Equilibrium. of concurrency cannot be used since all forces
of the system pass through this point. The
To completely define a force, we must know moment axes need not be the same direction as
its magnitude, direction, and point of applica- the directions used in the force equations but
tion. These facts regarding the force are of course, they could be.
generally referred to as the characteristics of
the force. Sometimes the more general term of EQ.UILIBRIUM OF SPACE PARALLEL-FORCE SYSTEM
line of action or location is used as a force
characteristic in place of point of application In a parallel force system the direction of
designation. all forces is known, but the magnitude and
location of each is unknown. Thus to determine
A force acting in space is completely magnitude, one equation is required and for
defined if we know its components in three location two equations are necessary since the
directions and its moments about 3 axes, for force is not confined to one plane. In general
example Fx , Fy ' Fz and Mx , My and Mz • For the 3 equations commonly used to make the re-
equilibrium or a force system there can be no sultant zero for this type of force system are
resultant force and thus the equations of one force equation and two moment equations.
equilibrium are obtained by equating the force For example, for a space parallel-force system
and moment components to zero. The equations acting in the y direction, the equations of
of static equilibrium for the various types of eqUilibrium would be:
force systems will now be summarized.
ZF y .:: 0, ZMx = 0, ZMz = 0 - -(2.3)
EQ.UILIBRIUM EQ.UATIONS FOR GENERAL
SPACE (NON-COPLANAR) FORCE SYSTEM EQ.UILIBRIUM OF GENERAL CO-PLANAR FORCE SYSTEM
ZFx =0 ZMJ. = 0 In this type of force system all forces lie
ZF y =0 ZM" = 0 } - - - - -(2.1) in one plane and it takes only 3 equations to
ZFz =0 ZM a = 0 determine the magnitude, direction and location
of the reSUltant of such a force system. Either
Thus for a general space force system, force or moment equations can be u~ed, except
there are 6 equations of static equilibrium that a maximum of 2 force equations can be used.
available. Three of these and no more can be For example, for a force system acting in the
force equations. It is often more convenient xy plane, the following combination of equili-
to take the moment axes, 1, 2 and 3, as any set brium equations could be used.
of x, y and z axes. All 6 equations could be
moment equations about 6 different axes. The
force equations are written for 3 mutually ZFx = 0 ZFx = 0 ZF y = 0 ZM z J. = 0
perpendicular axes and need not be the x, y ZF y = 0 or ZMzJ.= 0 or ZMzJ.= 0 or ZMz " = 0 2.4
and z axes.
ZM z= 0 ZMz" = 0 ZMz" = 0 ZM za = 0
EQ.UILIBRIUM OF SPACE CONCURRENT-FORCE SYSTEM (The SUbscripts 1, 2 and 3 refer to different
Concurrent means that all forces of the locations for z axes or moment centers.)
A2.1
A2.2 EQUILIBRIUM OF FORCE SYSTEMS. TRUSS STRUCTURES.
EQUILIBRIUM OF COPLANAR-CONCURRENT-FORCE SYSTD1 reacted by other external forces, commonly

referred to as reactions which hold the known
Since all forces lie in the same plane and forces on the structure in eqUilibrium. Since
also pass through a common point, the magnitude the static equations of equilibrium available
and direction of the resultant of this type of for the various types of force systems are
force system is unknown but the location is limited, the structural engineer resorts to the
knO\Vll since the point of concurrency is on the use of fitting units which establish the
line of action of the resultant. Thus only two direction of an unkno.vn force or its point of
equations of equilibrium are necessary to define application or both, thus decreasing: the number
the resultant and make it zero. The combin- of unknowns to be detennined. The figures
ations available are, which follow illustrate the type of fitting
units employed or other general methods for
ZFx =0 or ZFx =0 or ZF y = 0 or ZMz~=O}2.5 establishing the force characteristics of
direction and point of application.
ZF y =0 ZMz = 0 Zl'1z = 0 Zl'1Z2 = 0
(The z axis or moment center locations must be Ball and Socket Fitting
other than through the point of concurrency)
EQUILIBRIUM OF CO-PLANAR PARALLEL FORCE SYSTEM
Since the direction of all forces in this

type of force system is known and since the
forces all lie in the same plane, it only takes For any space or conlanar force such as P
2 equations to define the magnitude and location and Q acting on the bar, the line of action of
of the resultant of such a force system. Hence, such forces must act throuGh the center of the
there are only 2 equations of eqUilibrium avail- ball if rotation of the bar is prevented. Thus
able for this type of force system, namely, a a ball and socket joint can be used to establish
force and moment equation or two moment or control the direction and line action of a
equations. For example, for forces parallel to force applied to a structure through this type
the y axis and located in the xy plane the of fitting. Since the joint has no rotational
equilibrium equations available would be: - resistance, no couples in any plane can be
applied to it.
=0
ZF y
ZMz - 0
or
Zl'1 z 2 = 0
} - - - - - 2.6 Single Pin Fitting
(The moment centers 1 and 2 cannot be on the

same y axis)
EQUILIBRIUM OF COLINEAR FORCE SYSTEM
A colinear force system is one where all For any force such as P and Q acting in the
forces act along the same line or, in other xy plane, the line of action of such a force
words, the direction and location of the forces must pass through the pin center since the
are known but their magnitudes are unknown; fitting unit cannot resist a moment about a z
thus only magnitude needs to be found to define axis through the pin center. Therefore, for
the resultant of a colinear force system. Thus forces acting in the xy plane, the direction
only one equation of eqUilibrium is available, and line of action are established by the pin
namely joint as illustrated in the figure. Since a
single pin fitting can resist moments about axes
ZF =0 or ZM~ =0 2.7 perpendicular to the pin axis, the direction and
line of action of out-of-plane forces is there-
where moment center 1 is not on the line of fore not established by single pin fitting units.
action of the force system
_(0
A B
0)_
ex
A2. 3 Structural Fitting Units for Establishing the Force
Characteristics of Direction and Point of Application.
If a bar AB has single pin fittings at
To completely define a force in space re- each end, then any force P lying in the xy
quires 6 equations, or 3 equations if the force plane and applied to end B must have a direction
is limited to one plane. In general a structure and line of action coinciding with a line join-
is loaded by known forces and these forces are ing the pin centers at end fittings A and B,
transferred through the structure in some since the fittings cannot resist a moment about
manner of internal stress distribution and then the z axis.
A2.3
Double Pin - Universal Joint Fittings Another general fitting type that is used
to establish the direction of a force or reaction
is illustrated in the figure at the bottom of the
first column. Any reacting force at joint (A)
must be horizontal since the support at (A) is
so designed to provide no vertical resistance.
Cables - Tie Rods
Since single pin fitting units can resist

applied moments about axes normal to the pin
axis, a double pin joint as illustrated above Since a cable or tie rod has negligible
is often used. This fitting unit cannot resist bending resistance, the reaction at joint B on
moments about y or z axes and thus applied the crane structure from the cable must be
forces such as P and Q must have a line of colinear with the cable axis, hence the cable
action and direction such as to pass through establishes the force characteristics of direc-
the center of the fitting unit as illustrated tion and point of application of the reaction
in the figure. The fitting unit can, however, on the truss at point B.
resist a moment about the x axis or in other
words, a universal type of fitting unit can A2.4 Symbols for Reacting Fitting Units as Used in
resist a torsional moment. Problem Solution.
Rollers In solving a structure for reactions,

member stresses, etc., one must know what force
characteristics are unknown and it is common
practice to use simple symbols to indicate what
_A~B_
fitting support or attachment units are to be
used or are assumed to be used in the final
~
design. The following sketch symbols are com-
monly used for coplanar force systems.
Pin
In order to permit structures to move at _----"(0"'-)_ _-'1--
~Pin
Q,
- ,-I
support points, a fitting unit involving the

idea of rollers is often used. For example, the
truss in the figure above is supported by a pin A small circle at the end of a member or on
fitting at (A) which is further attached to a a triangle represents a single pin connection
fitting portion that prevents any horizontal and fixes the point of application of forces
movement of the truss at end (A); however, acting between this unit and a connecting member
the other end (B) is supported by a nest of or structure.
rollers which provide no horizontal resistance
to a horizontal movement of the truss at end (B).
The rollers fix the direction of the reaction
at (B) as perpendicular to the roller bed.
Since the fitting unit is joined to the truss
joint by a pin, the point of application of the
Rx
&t
by-Knife Edge
-&
Rx I
I
I
t
reaction is also known, hence only one force ~Ry +Ry
characteristic, namely magnitude, is unknown
for a roller-pin type of fitting. For the
fitting unit at (A), point of application of the
reaction to the truss is known because of the The above graphical symbols represent a
pin, but direction and magnitude are unknown. reaction in which translation of the attach-
ment point (b) is prevented but rotation of the
Lubricated Slot or Double Roller Type of Fitting attached structure about (b) can take place.
Unit. Thus the reaction is unknown in direction and
magnitude but the point of application is known,
~
"A"
~
namely through point (b). Instead of using
direction as an unknown, it is more convenient
Lubricated Slot t ~ ! Double Ron: to replace the resultant reaction by two com-
4-
ponents at right angles to each other as indi-

I I cated in the sketches.
minate, and the degree of redundancy depends on

the number of unknolVlls beyond that number which
~"'Edg. ~n can be found by the equations of static equili-
t
Rollers ~ Rollers brium. A structure can be statically indeter-
minate with respect to external reactions alone
or to internal stresses alone or to both.
~ +
The additional equations that are needed
to solve a statically indeterminate structure
are obtained by considering the distortion of
The above fitting units usinG rollers fix the structure. This means that the size of all
the direction of the reaction as normal to the members and the material from which members are
roller bed since the fitting unit cannot resist made must be known since distortions must be
a horizontal force through point (b). Hence calculated. In a statically determinate
the direction and point of application of the structure this information on sizes and material
reaction are established and only magnitude is is not required but only the configuration of
unknown. the structure as a whole. Thus design analysis
for statically determinate structures is
straightforward, whereas a general trial and
fixed error procedure is required for design analysis
'-'-1-1-1-\ of statically indeterminate structures.
R/ A2. 6 Examples of Statically Determinate and Statically

Indeterminate Structures.
The first step in analyzing a structure is

to determine whether the structure as presented
The graphical symbol above is used to
represent a rigid support which is attached is statically determinate. If so, the reactions
and internal stresses can be found without know-
rigidly to a connectin~ structure. The re-
ing sizes of members or kind of material. If
action is completely unknown since all 3 force
not statically determinate, the elastic theory
characteristics are unlmown, na,11ely, magnitude, must be applied to obtain additional equations.
direction and point of application. It is con-
The elastic theory is treated in considerable
venient to replace the reaction R by two force
detail in Chapters A7 to A12 inclusive.
components referred to some point (b) plUS the
unknovVll moment M which the resultant reaction R
To help the student become familiar with
caused about point (b) as indicated in the
the problem of determining whether a structure
above sketch. This discussion applies to a
is statically determinate, several example
coplanar structure with all forces in the same
problems will be presented.
plane. For a space structure the reaction
would have 3 further unknowns, namely, Rz ' Mx
Example Problem 1.
and My.
w = 10 lb. lin.
A2. 5 Statically Determinate and Statically Indeterminate
Structures.
A statically determinate structure is one

in which all external reactions and internal
stresses for a given load system can be found
by use of the equations of static equilibrium, C
and a statically indeterminate structure is . Fig. A2.1
Pin
one in which all reactions and internal stresses RC/
cannot be found by using only the equations of
equilibrium. In the structure shown in Fig. 2.1, the
known forces or loads are the distributed loads
A statically determinate structure is one of 10 lb. per inch on member ABD. The reactions
that has just enough external reactions, or at points A and C are unknown. The reaction at
just enough internal members, to make the C has only one unknown characteristic, nam61y,
structure stable under a load system, and if magnitude, because the point of application of RC
one reaction or member is removed, the structure is through the pin center at C and the direction
is reduced to a linkage or a mechanism and is of RC must be parallel to line CB because there
therefore not further capable of resisting the is a pin at the B end of member CB. At
load system. If the structure has more ex- point A the reaction is unknown in direction
ternal reactions or internal members than is and magnitude but the point of application must
necessary for stability of the structure under be through the pin center at A. Thus there are
a given load system it is statically indeter- 2 unknowns at A and one unknown at C or a total
A2.5
of 3. With 3 equations of equilibrium avail- attached to reaction points ABCD.

able for a coplanar force system the structure
is statically determinate. Instead of using an At reaction points A, Band D, the reaction
angle as an unknown at A to find the direction is known in direction and point of application
of the reaction, it is usually more convenient but the magnitude is unknown as indicated by the
to replace the reaction by components at right vector at each support. At point C, the re-
angles to each other as HA and VA in the figure, action is unknown in direction because 2 struts
and thus the 3 unknowns for the structure are 3 enter joint e. Magnitude is also unknown but
magnitudes. point of application is known since the reaction
must pass through C. Thus we have 5 unknowns,
Example Problem 2. namely, RA, RB, RD, Ve and He. For a coplanar
force system we have 3 equilibriu~ equations
P
Pin available and thus the first conclusion might
c be that we have a statically indeterminate
structure to (5-3) = 2 degrees redundant. How-
ever, observation of the structure shows two
internal pins at points E and F which means
that the bending moment at these two points is
zero, thus giving us 2 more equations to use
Fig. A2.3 with the 3 equations of equilibrium. Thus
Fig. A2. 2 drawing free bodies of the structure to left of
pin E and to right of pin F and equating moments
Fig. 2.2 shows a structural frame carrying about each pin to zero, we obtain 2 equations
a known load system P. Due to the pins at which do not include unknowns other than the 5
reaction points A and B the point of application unknowns listed above. The structure is there-
is known for each reaction; however, the magni- fore statically determinate,
tude and direction of each is unknown, making a
total of 4 unknowns with only 3 equations of Example Problem 4.
equilibrium available for a coplanar force
system. At first we might conclude that the
structure is statically indeterminate but we
must realize this structure has an internal pin
at e which means the bending moment at this
point is zero since the pin has no resistance
to rotation. If the entire structure is in e D
)::
~cE "ED
equilibrium, then each part must likewise be
in equilibrium and we can cut out any portion ~A V e
as a free body and apply the equilibrium Fig. A2.5 Fig. A2.6
equations. Fig. 2.3 shows a free body of the
frame to left of pin at e. Taking moments Fig. 2.5 shows a beam AB which carries a
about e and equating to zero gives us a fourth super-structure eED which in turn is SUbjected
equation to use in determining the 4 unknowns, to the known loads P and Q. The question is
HA' VA' VB and HB' The moment equation about e whether the structure is statically determinate,
does not include the unknowns Ve and He since The external unknown reactions for the entire
they have no moment about e because of zero structure are at points A and B. At A due to
arms. As in example problem 1, the reactions the roller type of action, magnitude is the only
at A and B have been replaced by H and V com- unknown characteristic of the reaction since
ponents instead of using an angle (direction) direction and point of application are known.
as an unknown characteristic. The structure is At B, magnitude and direction are unknown but
statically determinate. point of application is known, hence we have 3
unknowns, namely, RA' VB and HB, and with 3
Example Problem 3. equations of equilibrium available we can find
these reactions and therefore the structure is
P P statically determinate with respect to external
(1) (2)
reactions. We now investigate to see if the
internal stresses can be found by statics after
having found the external reactions. Obviously,
e-He the internal stresses will be affected by the
internal reactions at e and D, so we draw a free
tRB ,j.Ve body of the super-structure as illustrated in
Fig. A2.4 Fig. 2.6 and consider the internal forces that
existed at C and D as external reactions, In
Fig. 2.4 shows a straight member 1-2 carrying a the actual structure the members are rigidly
known load system P and supported by 5 struts attached together at point C such as a welded or
multiple bolt connection. This means that all Example Problem 6

three force or reaction characteristics, namely,
magnitude, direction, and point of application,
are unknown; or in other words, 3 unknowns
exist at C. For convenience we will represent
these unknowns by three components as shown in
Fig. 2.6, namely, HC' Vc and Me. At joint D in
Fig. 2.6, the only unknown regarding the re-
action is RD, a magnitude, since the pin at each
end of the member DE establishes the direction
and point of application of the reaction RD. I'
Hence we have 4 unknowns and only 3 equations tVA iVB Fig. A2.11 Fig. A2.12
of equilibrium for the structure in Fig. 2.6, Fig.A2.10
thus the structure is statically indeterminate
With respect to all of the internal stresses. Fig. 2.10 shows a 2 bay truss supported at
The stUdent should observe that internal points A and B and carrying a known load system
stresses between points AC, BD and FE are P, Q. All members of the truss are connected
statically determinate, and thus the statically at their ends by a common pin at each jOint.
indeterminate portion is the structural tri- The reactions at A and B are applied through
angle CEDC. fittings as indicated. The question is whether
the structure is statically determinate or not.
Example Problem 5 Relative to external reactions at A and B the
structure is statically determinate because the
type of support produces only one unknown at A
and two unknowns at B, namely, VA, VB and HB as
shown in Fig. 2.10, and we have 3 equations of
static equilibrium available.
We now investigate to see if we can find

the internal member stresses after having found
the values of the reactions at A and B. Suppose
we cut out joint B as indicated by section 1-1
Figs. 2.7, 2.8 and 2.9 show the same in Fig. 2.10 and draw a free body as shown in
structure carrying the same known load system Fig. 2.11. Since the members of the truss have
P but with different support conditions at pins at each end, the loads in these members
points A and B. The question is whether each must be axial, thus direction and line of action
structure is statically indeterminate and if are known and only magnitude is unknown. In
so, to what degree, that is, what number of Fig. 2.11, HB and VB are known but AB, CB, and
unknowns beyond the equations of statics avail- DB are unknown in magnitude; hence we have 3 un-
able. Since we have a coplanar force system, knowns but only 2 equations of equilibrium for
only 3 equations of statics are available for a coplanar concurrent force system. If we cut
equilibrium of the structure as a whole. through the truss in Fig. 2.10 by the section
2-2 and draw a free body of the lower portion
In the structure in Fig. 2.7, the reaction as shown in Fig. 2.12, we have 4 unknowns,
at A and also at B is unknown in magnitUde and namely, the axial loads in CA, DA, CB, DB but
direction, although their points of application only 3 e~uations of eqUilibrium available for
are known. Hence, 4 unknowns (With only 3 a coplanar force system.
equations of statics available) make the
structure statically indeterminate to the first Suppose we were able to find the stresses
degree. In Fig. 2.8, the reaction at A is a in CA, DA, CB, DB in some manner, and we would
rigid one, thus all 3 characteristics of magni- now proceed to joint D and treat it as a free
tUde, direction, and point of application of body or cut through the upper panel along
the reaction, are unknown. At point B, due to section 4-4 and use the lower portion as a free
pin only 2 unknowns, namely, magnitude and di- body. The same reasoning as used above would
rection, thus making a total of 5 unknowns show us we have one more unknown than the number
with only 3 equations of statics available or of eqUilibrium equations available and thus
the structure is statically indeterminate to we have the truss statically indeterminate to
the second degree. In the structure of Fig. the second degree relative to internal member
2.9, both supports at A and B are rigid. Thus, stresses.
all 3 force characteristics are unknown at each
support or a total of 6 unknowns Which makes Physically, the structure has two more
the structure statically indeterminate to the members than is necessary for the stability of
third degree. the structure under load, as we could leave
out one diagonal member in each truss panel and
A2.7
the structure would be still stable and all mine the axial loads in the members and the re-
member axial stresses could be found by the actions on the spar.
equations of static equilibrium without regard
to their size of cross-section or the kind of Solution: The first thing to decide is whether
material. Adding the second diagonal member the structure is statically determinate. From
in each panel would necessitate knowing the the figure it is observed that the wing spar is
size of all truss members and the kind of supported by five struts. Due to the pins at
material used before member stresSes could be each end of all struts, we have five unknowns,
found, as the additional equations needed must namely, the magnitude of the load in each strut.
come from a consideration involving distortion Direction and location of each strut load is
of the truss. Assume for example, that one known because of the pin at each end of the
diagonal in the upper panel was left out. We struts. We have 3 equations of equilibrium for
would then be able to find the stresses in the the wing spar as a single unit supported by the
members of the upper panel by statics but the 5 struts, thus two more equations are necessary
lower panel would still be statically inde- if the 5 unknown strut loads are to be found.
terminate to 1 degree because of the double It is no~tced that the wing spar includes 2 in-
diagonal system and thus one additional equation ternal single pin connections at points 0 and 0'.
is necessary and would involve a consideration This establishes the fact that the moment of all
of truss distortion. (The solution of static- forces located to one side of the pin must be
ally indeterminate trusses is covered in equal to zero since the single pin fitting can-
Chapter AS.) not resist a moment. Thus we obtain two addi-
tional equations because of the two internal pin
A2. 7 Example Problem Solutions of Statically Determinate fittings and thus we have 5 equations to find 5
Coplanar Structures and Coplanar Loadings. unknowns.
Although a student has taken a course in
statics before taking a beginning course in Fig. 2.15 shows a free body of the wing
spar to the right of hinge fitting at O.
aircraft structures, it is felt that a limited
review of problems involving the application y 2460=82 x 30 1013=(30 + 15)45
of the equations of static equilibrium is quite
justified, particularly if the problems are Lx l;--41"~1"-+20~-2-
possibly somewhat more difficult than most of Xo-6 ----e I
the problems in the usual beginning course in
statics. Since one must use the equations of
l Yo
~~lA
,. 117 Y.
static equilibrium as part of the necessary E A Fig. A2. 15
equations in solving statically indeterminate
structures and since statically indeterminate In order to take moments, the distributed
structures are covered in rather complete detail load on the spar has been replaced by the re-
in other chapters of this book, only limited sultant load on each spar portion, namely, the
space will be given to problems involving total load on the portion acting through the
statics in this chapter. centroid of the distributed load system. The
strut reaction EA at A has been shown in phantom
Example Problem S. as it is more convenient to deal with its com-
ponents YA and XA. The reaction at 0 is un-
Fig. A2.14 shows a much simplified wing known in magnitude and direction and for con-
structure, consisting of a wing spar supported venience we will deal With its components Xo
by lift and cabane struts which tie the wing and YO. The sense assumed is indicated on the
spar to the fuselage structure. The distributed figure.
air load on the wing spar is unsymmetrical about
the center line of the airframe. The wing spar The sense of a force is represented
is made in three units, readily disassembled by graphically by an arrow head on the end of a
using pin fittings at points 0 and 0·. All vector. The correct sense is obtained from the
supporting wing struts have single pin fitting solution of the equations of equilibrium since
units at each end. The problem is to deter- a force or moment must be given a plus or minus
sign in writing the equations. Since the sense
of a force or moment is unknown, it is assumed,
20,*/in. 40,*/in. I 30'*jin
15#/in. and if the algebraic solution of the equilibrium
J.r:rf] f f f f f 1ft I f f f i t t t f T":r"=
equations gives a plus value to the magnitUde
f f.- 45"-+- 82" ---iP'1---60" ---J H~ge then the true sense is as assumed, and opposite
A B 0" A Pin to that assumed if the solution gives a minus
Lift -+- sign. If the unknown forces are axial loads in
Struts 3 " members it is COmmon practice to call tensile
Fuselage stress plus and compressive stress minus, thus
Fig. A2.14 If we assume the sense of an unknown axial load
as tension, the solution of the equilibrium
equations will give a plus value for the magni- To find strut load B'C' take moments about
tude of the unknown if the true stress is point C.
tension and a minus sign will indicate the
assumed tension stresses should be reversed or ZMc =1325 x 65 + 2000 x 40 + (5880 - 4400)
compression, thus giving a consistency of signs. 30 - 1500 x 10 - 993 x 35 - 30 (B'C')
30/33.6 =
0
To find the unknown YA we take moments
about point 0 and equate to zero for equilibrium whence, B'C' = 6000 lb. with sense as shown.
ZMO = - 2460 x 41 1013 x 102 + 82YA =0 To find load in member B'C use equation
Hence YA =
204000/82 2480 lb. The plus sign = ZF y = =
0 1325 + 2000 + 1500 + 993 - 6000
means that the sense as assumed in the figure (30/33.6) - 2720 (30/33.6) - B'C (30/54)
is correct. By geometry XA = 2480 x 117/66 = =0
4400 lb. and the load in strut EA equals
v4400 2 + 2480 2 =
5050 lb. tension or as whence, B'C =-
3535 lb. The minus sign
assumed in the figure. means it acts opposite to that shown in figure
or is compression instead of tension.
To find Xo we use the equilibrium equation The reactions on the spar can now be
ZFX = 0 = Xo - 4400 = 0, whence Xo 4400 lb. = determined and shears, bending moments and
axial loads on the spar could be found. The
To find YO we use, numerical results should be checked for equili-
ZF y = 0 = 2460 + 1013 - 2480 - YO = 0, whence brium of the spar as a whole by taking moments
YO = 993 lb. of all forces about a different moment center
to see if the result is zero.
To check our results for equilibrium we
will take moments of all forces about A to see Example Problem 9.
if they equal zero.
ZMA = 2460 x 41 - 1013 x 20 - 993 x 82 =0 check RB\ f.--- 1~' + 12"----1 ;rRD
On the spar portion O'A', the reactions

are obviously equal to 40/30 times those found
o~B ~ r -Y- T
for portion OA since the external loading is 40 ~'\' Brae I ?; ~12 12"
as compared to 30. \'\ Strut % 12 ~
~ ~\ Brace Strut
Hence A'E' = 6750, XO' = 5880, YO' = 1325 24LS,cP \'\
12
E - -
Fig. 2.16 shows a free body of the center 12"

spar portion with the reactions at 0 and 0' as
found previously. The unknown loads in the
struts have been assumed tension as shown by
the arrows.
2000=50 x40 1500=50 x 30
Fig. A2.17
~
+12"
588.Q.
1---
l-- 20,,--15" r
<£==~~======-:i;;==:Jl--4400
50" 5" 20"-1
f~325 B' °t993

, Fig. 2.17 shows a simplified airplane
,, " ,--,
C'~-- -- -~C landing gear unit with all members and loads
I I Fig.A2.16 confined to one plane. The brace struts are
I I pinned at each end and the support at C is of
1- __ -, the roller type, thus no vertical reaction can
be produced by the support fitting at point C.
The member at C can rotate on the roller but
To find the load in strut BC take moments horizontal movement is prevented. A known load
about B' of 10,000 lb. is applied to axle unit at Ao The
problem is to find the load in the brace struts
2MB' = 1325 x 20 - 2000 x 5 - 1500 x 55 and the reaction at C.
- 993 x 80 + 60 (BC) 30/33.6 =0
Solution:
Whence, BC 2720 lb. with sense as Due to the single pin fitting at each end
assumed. of the brace struts, the reactions at Band D
A2.9
are colinear with the strut axis, thus direction the aluminum alloy tubular truss. Trussed type
and point of application are known for reaction beams composed of closed and open type sections
RB and RD leaving only the magnitude of each as are also frequently used in wing beam construc-
unknown. The roller type fitting at C fixes tion. The stresses or loads in the members of
the direction and point of application of the a truss are commonly referred to as "primary"
reaction RC' leaving magnitude as the only and "secondary" stresses. The stresses which
unknown. Thus there are 3 unknowns RB' Rc and are found under the following assumptions are
RD and with 3 equations of static equilibrium referred to as primary stresses.
available, the structure is statically determi-
nate with respect to external reactions. The (1) The members of the truss are straight,
sense of each of the 3 unknown reactions has weightless and lie in one plane.
been assumed as indicated by the vector. (2) The members of a truss meeting at a
point are considered as joined together by a
To find RD take moments about point B:- common frictionless pin and all member axes in-
2MB =- 10000 sin 30 0 x 36 - 10000 cos 30 0 x 12 tersect at the pin center.
- RD (12/17) 24 0 = (3) All external loads are applied to the
whence, RD = - 16750 lb. Since the result truss only at the joints and in the plane of
comes out with a minus sign, the reaction RD the truss. Thus all loads or stresses produced
has a sense opposite to that shown by the in members are either axial tension or compres-
vector in Fig. 2.17. Since the reaction RD is sion without bending or torsion.
colinear with the line DE because of the pin
ends, the load in the brace strut DE is 16750 Those stresses produced in the truss mem-
lb. compression. In the above moment equation bers due to the non-fulfillment of the above
about B, the reaction RD was resolved into assumptions are referred to as secondary
vertical and horizontal components at point D, stresses. Most steel tubular trusses are welded
and thus only the vertical component which tovether at their ends and in other truss types,
equals (12/17) RD enters into the equation the members are riveted or bolted to~ether.
since the horizontal component has a line of This restraint at the joints may cause second-
action through point B and therefore no moment. ary stresses in some members greater than the
RC does not enter in equation as it has zero primary stresses. Likewise it is common in
moment about B. actual practical design to apply forces to the
truss members between their ends by supporting
To find RB take ZFy =0 many equipment installations on these truss
members. However, regardless of the magnitude
ZFy = 10000 x cos 30 0 + (- 16750)(12/17) + RB of these so-called secondary loads, it is
(24/26.8) = 0 common practice to first find the primary
whence, RB =
3540 lb. Since sign comes stresses under the assumption outlined above.
out plus, the sense is the same as assumed in
the figure. The strut load BF is therefore GENERAL CRITERIA FOR DETERMINING WHETHER
3540 lb. tension, since reaction RB is colinear TRUSS STRUCTURES ARE STATICALLY DETERMINATE
wi th line BF. WITH RESPECT TO INTERNAL STRESSES.
To find RC take ZH =0 The simplest truss that can be constructed

is the triangle which has three members m and
ZH = 10000 sin 30 0 - 3540 (12/26.8) + (- 16750) three joints j. A more elaborate truss consists
(12/17) + RC 0= of additional triangular frames, so arranged
whence, RC =
8407 lb. Result is plus and
that each triangle adds one joint and two mem-
bers. Hence the number of members to insure
therefore assumed sense was correct.
stability under any loading is:
To check the numerical results take =
m 2j - 3 - - - - - - - - - - - - - (2.8)
moments about point A for equilibrium. A truss having fewer members than required
ZMA = 8407 x 36 + 3540 (24/26.8) 12 - 3540 by Eq. (2.8) is in a state of unstable equili-
(12/26.8) 36 + 16750 (12/17) 12 - 16750 (12/17) brium and will collapse except under certain
36 = 303000 + 38100 - 57100 + 142000 - 426000 conditions of loading. The loads in the members
= 0 (check) of a truss With the number of members shown in
equation (2.8) can be found with the available
A2. 8 Stresses in Coplanar Truss Structures Under equations of statics, since the forces in the
Coplanar Loading. members acting at a point intersect at a common
point or form a concurrent force system. For
In aircraft construction, the truss type this type of force system there are two static
of construction is quite common. The most
equilibrium equations available.
common is the tubular steel welded trusses that
make up the fuselage frame, and less frequently, Thus for j number of joints there are 2j
equations available. However three independent 500

equations are necessary to determine the exter- 1000
2 I 2
nal reactions, thus the number of equations
necessary to solve for all the loads in the
L o L1. I u. 2 / d 500
\ / \ L3
members is 2j - 3. Hence if the number of truss -
members is that given by equation (2.8) the 40" o
truss is statically determinate relative to the
~
primary loads in the truss members and the 30
I \
truss is also stable.
Uo U1. 3 U2 3 Fig. A2. 18
If the truss has more members than indi- ~ 30"T 30" --t- 30"-1
cated by equation (2.8) the truss is considered
redundant and statically indeterminate since 1
1000
(1 •
the member loads cannot be found in all the +~
members by the laws of statics. Such redundant L 1.. - fi- 500
structures if the members are properly placed Fig. A2.19 2 \~ 3
+
are stable and will support loads of any U2 1

arrangement.
or tension. In equation (b) the load of 1250
ANALYTICAL METHODS FOR DETERl'1INING in L3 L 2 was substituted as a minus value since
PRIMARY STRESSES IN TRUSS STRUCTURES it was found to act opposite to that shown in
Fig. A2.19. Possibly a better procedure would
In general there are three rather distinct be to change the sense of the arrow in the free
methods or procedures in applying the equations body diagram for any solved members before writ-
of static equilibrium to finding the primary ing further equilibrium equations. We must
stresses in truss type structures. They are proceed to joint L 2 instead of joint U2, as
often referred to as the method of joints, three unknown members still exist at joint U2
moments, and shears. whereas only two at joint L 2• Fig. A2.20 shows
free body of joint L 2 cut out by section 2-2
A2.9 Method of Joints. (see Fig. A2.18). The sense of the unknown
If the truss as a whole is in equilibrium member stress L 2 U2 has been assumed as com-
pression (pushing toward joint) as it is ob-
then each member or joint in the truss must viously acting this way to balance the 500 lb.
likewise be in equilibrium. The forces in the
members at a truss joint intersect in a cammon load.
point, thus the forces on each joint form a 500
concurrent-coplanar force system. The method
l• ~H
of joints consists in cutting out or isolating L1. ... --250
a joint as a free body and applying the laws of
equilibrium for a concurrent force system.
Since only two independent equations are avail-
tL2
U2 Fig.A2.20
able for this type of system only two unknowns
can exist at any joint. Thus the procedure is For equilibrium of joint L 2, ~H and ~V 0 =
to start at the joint where only two unknowns ~V =- 500 +L2U2 = 0, whence, L 2U2 = 500 lb.
exist and continue progressively throughout the Since the sign came out plus, the assumed sense
truss joint ~y joint. To illustrate the method in Fig. A2.20 was correct or compression.
consider the cantilever truss of Fig. A2.18.
From observation there are only two members ~H =250 - L 2L1.= 0, whence L 2L1. = 250 lb.
with internal stresses unknown at joint L8 •
Fig. A2.19 shows a free body of joint L8 • The Next consider joint U2 as a free body cut
stresses in the members L 3 L 2 and L 3 U2 have out by section 3-3 in Fig. A2.18 and drawn as
been assumed as tension, as indicated by the Fig. A2.21. The known member stresses are shown
arrows pUlling away from the joint L3 • with their true sense as previously found. The
two unknown member stresses U2L1. and U2U1. have
The static equations of equilibrium for been assumed as tension.
the forces acting on joint L3 are ~H and ~V = O.
~V = - 1000 - L3 U2 (40/50) =
0 - - - - - - -(a)
whence, L3 U2 =-
1250 lb. Since the sign Fig. A2. 21
came out minus the stress is opposite to that
assumed in Fig. A2.19 or compression.
~ =- 500 -(- 1250)(30/50) - L 2L3 = 0 - -(b) ~v =-500 -1250 (40/50) + U2L1. (40/50) = 0
whence, L2L3 = 250 lb. Since the sign comes whence, U2L1. = 1875 lb. (tension as
out plus, sense is same as assumed in figure assumed. )
A2.11
ZH = (-1250) (30/50)
- 1875 (30/50) - U"U" 0 = The algebraic sign of all unknowns came out
positive, thus the assumed direction as shown
whence, U"U" = - 1875 lb. or opposite in on Fig. A2.22 was correct.
sense to that assumed and therefore compression.
Check results by taking 2MB = 0
Note: The student should continue with succeed-
ing joints. In this example involving a canti- 2MB =
1400 x 150 + 500 x 30 - 500 x 120 - 500 x
lever truss it was not necessary to find the 30 - 1000 x 90 - 1000 x 60 = 0 (Check)
reactions, as it was possible to select ~oint To determine the stress in member F", F" and Fa
La as a j oint involving only two unknowns. In we cut the section 1-1 thru the truss (Fig.
trusses such as illustrated in Fig. A2.22 it is A2.22. Fig. A2.23 shows a free body diagram of
necessary to first find reactions R" or R" which the portion of the truss to the left of this
then provides a joint at the reaction point in-
section.
volving only two unknown forces.
1
:"F~_
500 -_:..---,,---~~-
1400
500
~.
~;,
1400 1
Fig. A2. 23 Fig.A2.24
A2. 10 Method of Moments.
For a coplanar-non-concurrent force system The truss as a whole is in equilibrium.
there are three equations of statics available. Therefore any portion must be in eqUilibrium.
These three equations may be taken as moment In Fig. A2.23 the internal stresses in the mem-
equations about three different points. Fig. bers F", F" and Fa which existed in the truss as
A2.22 shows a typical truss. Let it be re- a whole now are considered external forces in
qUired to find the loads in the members F", F", holding the portion of the truss to the left of
Fa, F 4' F 6 and F 6 ' section 1-1 in equilibrium in combination with
the other loads and reactions. Since the mem-
bers a and b in Fig. A2.23 have not been cut the
Fig. A2. 22 loads in these members remain as internal
10"
stresses and have no influence on the eqUilib-
rium of the portion of the truss shown. Thus
the portion of the truss to left of section 1-1
could be considered as a so11d block as shown
, in Fig. A2.24 without affecting the values of
2
f<-----5 @ 30" = 150" - - - - - - - F", F" and Fa. The method of moments, as the
name implies, involves the operation of taking
moments about a point to find the load in a
The first step in the solution is to find the particular member. Since there are three un-
reactions at points A and B. Due to the roller knowns a moment center must be selected such
type of support at B the only unknown element of that the moment of each of the two unknown
the reaction force at B is magnitude. At point stresses will have zero moment about the selected
A, magnitude and direction of the reaction are moment center, thus leaVing only one unknown
unknown giving a total of three unknowns with force or stress to enter into the equation for
three equations of statics available. For con- moments. For example, to determine load Fa in
venience the unknown reaction at A has been re- Fig. A2.24 we take moments about the inter-
placed by its unknown H and V components. section of forces F" and F" or point O.
Taking moments about point A, Thus ZMO = 1400 x 30 - 18.97 Fa =0
ZMA = 500 x 30 + 1000 x 60 + 1000 x 90 + 500 x 42000
30 + 500 x 120 - 150 VB 0 = Hence Fa = 18.97 = 2215 lb. compression (or
acting as assumed)
Hence VB = 1600 lb.
To find the arm of the force Fa from the
Take ZV 0 = moment center 0 involves a small amount of cal-
ZV = VA - 1000-1000-500-500 + 1600 =0 there- culation, thus in general it is simpler to re-
fore VA = 1400 lb. solve the unknown force into H and V components
at a point on its line of action such that one
Take ZH =0 of these components passes thru the moment
ZH = 500 - HA = 0, therefore HA = 500 lb. center and the arm of the other component can
usually be determined by inspection. Thus in
Fig. A2.25 the force Fa is resolved into its the other two unknowns Fe and Fe lies at infini-
component FaV and FaH at point 0'. Then taking ty. Thus for conditions where two of the 3 cut
members are parallel we have a method of solving
for the web member of the truss commonly re-
ferred to as the method of shears, or the sum-
mation of all the forces normal to the two
parallel unknown chord members must equal zero.
Fig. A2. 25
Since the parallel chord members have no com-
ponent in a direction normal to their line of
action, they do not enter the above equation of
equilibrium.
1000#
2I
Fig. A2. 26 1
1000
:..:..:;j""-----r-----"O...-.;..-
500#
500#
Fe
1400#
Fig. A2. 27
500
moments about point 0 as before:-
Fig.A2.28
ZMo = 1400 x 30 - 20Fa H =0
Referring to Fig. A2.27
whence, FaH = 2100 lb. and therefore
ZV =1400 - 500 - 1000 - F.. (l/y'2:).= 0
Fa =
2100 (31.6/30) = 2215 lb. as pre-
viously obtained. whence F.. = - 141 lb. (tension or opposite
to that assumed in the figure.
The load F:a. can be found by taking moments
about point m, the intersection of forces Fe To find the stress in member F7 • we cut
and Fa (See Fig. A2.23). section 3-3 in Fig. A2.22 and draw a free body
diagram of the left portion in Fig. A2.26.
ZMm = 1400 x 60 + 500 x 30 - 500 x 30 Since F:a. and Fe are horizontal, the member F7
- 30F:a. = 0 must carry the shear on the truss on this section
3-3, hence the name method of shears.
whence, F:a. = 2800 lb. (Tension as assumed)
ZV = 1400 - 500 - 1000 + F = 0 7
To find force Fe by using a moment equation,
we take moments about point (r) the inter- Whence F = 100 lb. (compression as
7 assumed)
section of forces F:a. and F .. (See Fig. A2.26).
To eliminate solving for the perpendicular Note: The stUdent should solve this example il-
distance from point (r) to line of action of lustrating the methods of moments and shears
Fe, we resolve Fe into its H and V components using as a free body the portion of the truss to
at point 0 on its line of action as shown in the right of the cut sections instead of the
Fig. A2.26. left portion as used in these illustrative ex-
amples. In order to solve for the stresses in
ZMr =- 1400 x 30 + 500 x 60 + 60 FeV =0 the members of a truss most advantageously, one
usually makes use of more than one of the above
whence, FeV = 12000/60 = 200 lb. three methods, as each has its advantages for
certain cases or members. It is important to
Therefore Fe = 200 x V2 = 282 lb. com- realize that each is a method of sections and in
pression a great many cases, such as trusses with paral-
lel chords, the stresses can practically be
A2. 11 Method of Shears found mentally without writing down equations of
equilibrium. The following statements in gen-
In Fig. A2.22 to find the stress in member eral are true for parallel chord trusses:
F .. we cut the section 2-2 giving the free body
for the left portion as shown in Fig. A2.27. (1) The vertical component of the stress in
the panel diagonal members equals the vertical
The method of moments is not sufficient to shear (algebraic sum of external forces to one
solve for member F.. because the intersection of side of the panel) on the panel, since the chord
A2.13
members are horizontal and thus have zero verti- since no external vertical load exists at joint
cal component. E. Similarly, by the same reasoning for ~H = 0,
load in DE = O. The load in the diagonal FD
(2) The truss verticals in general resist equals the value on the diagonal of the panel
the vertica~ component of the diagonals plus index triangle or 167 lb. It is tension by
any external loads applied to the end joints of observation since the shear in the panel to the
the vertical. right is up and the vertical component of the
diagonal FD must pull down for equilibrium.
(3) The load in the chord members is due
to the horizontal components of the diagonal Considering Joint F. ~H =- FG - F~ 0, =
members and in general equals the summation of which means that the horizontal component of the
these horizontal components. load in the diagonal DF equals the load in FG,
or is equal to the value of the horizontal side
To illustrate the simplicity of determining in the index triangle or - 133 lb. It is nega-
stresses in the members of a parallel chord tive because the horizontal component of DF
truss, consider the cantilever truss of Fig. pulls on Joint F and therefore FG must push
A2.29 with supporting reactions at points A and against the joint for equilibrium.
J.
Considering Joint D:-
~v =
= DFV + DG = O. But DFV 100 (vertical side
of index triangle)
••• DG =- 100
~H =
DE + DFH - DC =
0, but DE =
0 and DFH =
133 (from index triangle)
DC =
133
Considering Joint G:-
Fig. A2.29
~H=- GH + GF - GCH = O. But GF =-
133, and GCH
= 300 from index triangle in the second panel.
First, compute the length triangles in Hence GH =-
433 lb. Proceeding in this manner,
each panel of the truss as shown by the dashed we obtain the stress in all the members as shown
triangles in each panel. The other triangles in Fig. A2.29. All the equilibrium equations
in each panel are referred to as load or index can be solved mentally and with the calculations
triangles and their sides are directly pro- being done on the slide rule, all member loads
portional to the length triangles. can be written directly on the truss diagram.
The shear load in each panel is first writ- Observation of the results of Fig. A2.29
ten on the vertical side of each index triangle. show that the loads in the truss verticals equal
Thus, in panel EFGD, considering forces to the the values of the vertical sides of the index
right of a vertical section cut thru the panel, load triangle, and the loads in the truss di-
the shear is 100 lb., which is recorded on the agonals equal the values of the index triangle
vertical side of the index triangle. diagonal side and in general the loads in the
top and bottom horizontal truss members equal
For the second panel from the free end, the the summation of the values of the horizontal
shear is 100 + 150 = 250 and for the third panel sides of the index triangles.
100 + 150 + 150 =
400 lb., and in like manner
550 for fourth panel. The reactions at A and J are found when
the above general procedure reaches joints A
The loads in the diagonals as well as their and J. As a check on the work the reactions
horizontal components are directly proportional should be determined treating the truss as a
to the lengths of the diagonal and horizontal whole.
side of the length triangles. 'Thus the load in
diagonal member DF =
100 (50/30) =
167 and for Fig. A2.30 shows the solution for the
member CG = 250 (46.8/30) =
390. The hori- stresses in the members of a simply supported
zontal component of the load in DF =
100 (40/30) Pratt Truss, symmetrically loaded. Since all
= 133 and for CG = 250 (36/30) =
300. These panels have the same width and height, only one
values are shown on the index triangles for length triangle is drawn as shown. Due to
each truss panel as shown in Fig. A2.29. We symmetry, the index triangles are drawn for
start our analysis for the loads in the members panels to only one side of the truss center
of the truss by considering joint E first. line. First, the vertical shear in each panel
is written on the vertical side of each index
Using ~v =0 gives EF = 0 by observation, triangle. Due to the symmetry of the truss and
loading, we know that one half of the external horizontal components of diagonal members =
loads at joints U3 and L 3 is supported at re- 312 + 187.5 =
499.5. Therefore, load in U~U2 =
action R~ and 1/2 at reaction R 2, or shear in - 499.5. Similarly at joint L 2, L2L3 = 312 +
187.5 =
499.5. At joint U2, the horizontal
Length
Fig. A2. 30 components of U~U2 and U2U3 =
499.5 + 62.5 562 =
~
25" Triangle
30" which must be balanced by - 562 in member U2U3 •
39" 50 50 50 50 50
The reaction R 2 equals the value on the
U -499.5 U2 -562 U3
vertical side of our index triangle in the end
panel, or 375. This should be checked using
the truss as a whole and taking moments about
R2 •
If a truss is loaded unsymmetrically, the

reactions should be determined first, after
which the index triangles can be drawn, start-
ing with the end panels, since the panel shear
is then readily calculated.
center panel =
(100 + 50) 1/2 = 75. The verti-
cal shear in panel U~U2L~L2 equals 75 plus the A2.12 Aircraft Wing Structure. Truss Type with Fabric
external loads at U2 and L 2 or a total of 225 or Plastic Cover
and similarly for the end panel shear = 225 +
The metal covered cantilever wing with its
50 + 100 = 375. With these values known, the
better overall aerodynamic efficiency and suf-
other two sides of the index triangles are di-
rectly proportional to the sides of the length ficient torsional rigidity has practically re-
placed the externally braced wing except for low
triangles for each panel, and the results are as
shown in Fig. A2.30. speed commercial or private pilot aircraft as
illustrated by the aircraft in Figs. A2.31 and
32. The wing covering is usually fabric and
The general procedure from this point is to
therefore a drag truss inside the wing is
find the loads in the diagonals, then in the
necessary to resist loads in the drag truss
verticals, and finally in the horizontal chord
direction. Figs. A2.33 and 34 shows the gen-
members.
eral structural layout of such wings. The two
spars or beams are metal or wood. Instead of
The loads in the diagonals are equal to the
using double wires in each drag truss bay, a
values on the hypotenuse of the index triangles.
single diagonal strut capable of taking either
The sense, whether tension or compression, is
tension or compressive loads could be used.
determined by inspection by cutting mental
The external brace struts are stream line tubes.
sections thru the truss and noting the direction
of the external shear load which must be bal-
anced by the vertical component of the diagon-
als.
The loads in the verticals are determined

by the method of joints and the sequence of
joints is so selected that the stress in the
vertical member is the only unknown in the
equation ~V =a for the joint in question.
Thus for joint U3 , ~V =- 50 - U3 L3 =a
or U3 L3=- 50.
For joint U2 , ~V =- 50 - U2L3 V - U2L2 0, = Fig. A2. 31 Piper Tri-Pacer

but U2L3 V ::: 75, the vertical component of U2L3
from index triangle. .'. U2L2 =-
50 - 75 =
- 125. For joint L~, ~V = - 100 + L~U~ 0, =
hence L~U~ = 100.
Since the horizontal chord members receive

their loads at the joints due to horizontal
components of the diagonal members of the truss,
we can start at Lo and add up these horizontal
components to obtain the chord stresses. ThUS,
LoL~ = 312 (from index triangle). L~L2 312 =
from ~H = a for joint L~. At joint U~, the Fig. A2. 32 Champion Traveler
A2.15
Wing Tip (End) Bow in all members of the lift and drag trusses will
be determined. A simplified air loading will be
"f4F~~~~--Plywood Tip Fairing assu~ed, as the purpose of this problem is to
Fig. A2. 33 hive the student practice in solving statically
...---Aileron determinate space truss structures .
Aileron Spar
Compression Rib
ASSUMED AIR LOADING:-
Plywood Leading (1) A constant spanwise lift load of 45
Edge Fairing Drag Wire
Ib/in from hinr;e to strut point and t11en taper-
Dragwire ing to 22.5 Ib/in at the wing tip.
Fitting Anti-Drag Wire
Aileron Rib (2) A forward uniform distributed drag
Aileron Hinge load of 6 Ib/in.
Drag Strut or
Leading Compression Tube
Edge The above airloads represent a high angle
Forming or Plain Rib of attack condition. In this condition a for-
Trailing Edge ward load can be placed on the drag truss as
illustrated in Fig. A2.36. Projecting the air
Wing Hinge Butt Rib

Fitting v- - Beam or Spar , Lift
METAL FAIRING
-
Airstream
Drag
Fig. A2. 36
______ Drag Truss
Center of
Pressure Rear
Beam
lift and drag forces on the drag truss direction,

AIL.E""'"
the forward projection due to the lift is great-
er than the rearward projection due to the air
drar" which difference in our example problem
Fig. A2. 34 has been assumed as 6 Ib/in. In a low angle of
attack the load in the drag truss direction
Example Problem 10. Externally Braced Mono- would act rearward.
plane Wing Structure
SOLUTION:
Fig. A2.35 shows the structural dimensional
diagram of an externally braced monoplane wing. The running loads on the front and rear
The wing is fabric covered between wing beams, beams will be calculated as the first step in
and thus a drag truss composed of struts and the solution., For our flight condition, the
tie rods is necessary to provide strength and center of pressure of the airforces will be
rigidity in the drag direction. The axial loads assumed as shown in Fig. A2.37.
DRAG TRUSS
Drag (4) (3)
- Rear
Reaction ~-----JH------...----~---~~<l---'?t'.----_l4----0>1 T
: Beam
Taken off 1 I D
At (2) I 4 1 '0. L
~ ~ S
Y (2)1 --L-Front
2 -"':1---37. 5 + 37.5+37.5 58.5 ----.j121--- Beam
__U1t :I:
:,--Tl-~,2)
~'S
I Q)
1 __[J(3)
~
~ Po<
....;
~
3 0 Dihedral
T
Wing Chord= 72"
I
1.1% =
22.52"
24.21"
Fig. A2. 37
: ~' ~ : + ~ SF SR The running load on the front beam will be 45 x

lirp a1e : + S I 24.2/36 = 30.26 Ib/in., and the remainder or
17 - -(5' 6
I... ~I
- -- - -- ----ill _I-
-..j uk--- 36
-
---I
(6) 45 - 30.26 =
14.74 Ib/in gives the load on the
Fuselage Fig. A2.35
rear beam.
A2 16 EQUILIBRIUM OF FORCE SYSTEMS TRUSS STRUCTURES
To solve for loads in a truss system by a 4 R.B. 3

method of joints, all loads must be transferred
to the truss joints. The wing beams are sup- Drag Truss
ported at one end by the fuselage and outboard
by the two lift struts. Thus we calculate the
reactions on each beam at the strut and hinge 2;::.B.1 ·B.
3
points due to the running lift load on each

beam.
Front Beam 5
F Front Lift
Truss
A6
R Rear
Lift
Truss
~
w= 30.26#~in.
(2)
114. 5"
-.-.
70. 5" ---I

* 15. 13#/in.
~t~f;;:,::Ir-=-t=-n=J member lengths L and the component ratios then
follow by simple calculation.
R 2 R~
Table A2.1
Taking moments about point (2) Member Sym. V D S L V/L D/L S/L
114.5R~ - 114.5 x 30.26 x 114.5/2 - 15.13 x Front Beam FB 5.99 o 114.34 114.50 .0523 o .9986
70.5 x 149.75 - 15.13 x 35.25 x 138 = O. Rear Beam RB 5.99 o 114.34 114.50 .0523 o .9986
hence Front strut SF 57.99 11 114.34 128.79 .4501 .0854 .8878
R~ = 3770 lb. Rear Strut SR 57.49 o 114.34 128.00 .4486 0 .8930

v = vertical direction,
Take ZV = 0 where V direction is taken normal to
beam D = drag direction,
ZV =- R2 - 3770 + 30.26 x 114.5 +

S = side direction,
L: VV2+ D2 + S2
(30.26 + 15.13)
2 70.5 = 0
hence R2 = 1295 lb. We start the solution of joints by starting

(The student should always check results by with joint (1). Free body sketches of joint (1)
taking moments about point (1) to see if ZM~ are slzetched below. All members are considered
two-force members or having pins at each end,
= 0) thus magnitude is the only umznown character-
Rear Beam istic of each member load. The drag truss mem-
bers coming in to joint (1) are replaced by a
sinGle reaction called D~. After D~ is found,
{w = 14. 74#/in. ;J:!.. its influence in causing loads in drag truss
(4) 1_.-
1
-'_1 11/5,,1 f f '.1§t ;0.f5"~ 2
members can then be found when the drag truss as
a whole is treated. In the joint solution, the
R4 R" drag truss has been assumed parallel to drag
direction which is not quite true from Fig.
A2.35, but the error on member loads is negli-
The rear beam has the same span dimensions but
gible.
the loading is 14.74 Ib/in. Hence beam re-
actions R 4 and R" will be 14.74/30.26 = .4875
JOINT 1 (Equations of Equilibriuti)
times those for front beam.
hence R" = .4875 x 3770 = 1838 lb. 3770
-;J ~ r:
770
R4 = .4875 x 1295 = 631 lb. FB 1)
(1) I D~
S .+V FB
The next step in the solution is the F ~+S
solving for the axial loads in all the members. SF +D
We will use the method of joints and consider
the structure made up of three truss systems ZV = 3770 x .9986 - .0523 FB - .4501 SF =
0 - (1)
as illustrated at the top of the next column, ZS = - 3770x .0523 - .9986 FB - .8878 SF = 0 - (2)
namely, a front lift truss, a rear lift truss ZD =- .0854 SF + D~ = 0 - - - - - - - - - - (3)
and a drag truss. The beams are common to both
lift and drag trusses. Solving equations 1, 2 and 3, we obtain
Table A2.1 gives the V, D and S projections 8513 lb. (compression)

of the lift truss members as determined from = 9333 lb. (tension)
information given in Fig. A2.35. The true = 798 lb. (aft)
A2.17
Joint (3) (Equations of equilibrium) points (2) and (4). In the design of the beam
and fittings at this point, the effect of the
1838 (drag truss actual conditions of eccentricity should of
1
~
838 reaction on
RB course be considered.
7 R
(3)
V-S plane
(3)
RB
SR
" pin (3~)
V-D plane Combined Loads on Drag Truss
~V = 1838 X .9986 - .0523 RB - .4486 SR = 0 -(4) Adding the two load systems of Figs. A2.38
~S =-1838 x .0523 - .9986 RB -.8930 SR 0 -(5) = and A2.39, the total drag truss loading is ob-
~D = D" + 0 = 0 - - - - - - - - - - - - - - -(6) tained as shown in Fig. A2.40. The resulting
member axial stresses are then solved for by the
Solving equations 4, 5 and 6, we obtain method of index stresses (Art. A2.9). The
RB =-
4189 lb. (compression) values are indicated on the truss diagram. It
SR =
4579 lb. (tension) is customary to make one of the fittings attach-
Da = 0 ing wing to fuselage incapable of transferring
drag reaction to fuselage, so that the entire
Fig. A2.38 shows the reactions of the lift drag reaction from wing panel on fuselage is
struts on the drag truss at joints (1) and (3) definitely confined to one point. In this ex-
as found above. ample point (2) has been assumed as point where
drag is resisted. Those drag wires which would
(
4)~
be in compression are assumed out of action.
Ga~~36~--------I+---36~
---1
39.5 37.5 58.5
231 225
(2) 8513+ 281.5 254
Fig. A2.38 798 ~~
Drag Truss Panel Point Loads Due to Air Drag
Load.
It was assumed that the air load components -10,313 8513l -413
in the drag direction were 6 Ib./in. of wing 798
acting forward.
The distributed load of 6 Ib./in. is re-

placed by concentrated loads at the panel points Fuselage Reactions
as shown in Fig. A2.39. Each panel point takes
one half the distributed load to the adjacent As a check on the work as well as to obtain
panel point, except for the two outboard panel reference loads on fuselage from wing structure,
points which are affected by the overhang tip the fuselage reactions will be checked against
portion. the externally applied air loads. Table A2.2
gives the calculations in table form.
Thus the outboard panel point concentration
Table A2.2
of 254 lb. is determined by taking moments about
(3) of the drag load outboard of (3) as follows: Point Member Load V D S
P = 70.5 x 6 x 35.25/58.5 = 254 lb. 2

FB
Drag
-13893
- 1908
- 726 0 -13870
0 -1908 0
Reaction
To simplify the drag truss solution, the draG R2 (Reaction) - 1295 1294 0 67
strut and drag wires in the inboard drag truss RB 1191 62 0 1190
panel have been modified to intersect at hinge 4 R4 (Reaction) - 631 630 0 33
6#/in. 5 FS 9333 4205 798 8290
6 RS 4579 2055 0 4090
118.5 281.5 J254
(4) --, Totals 7520 -1110 - 400
I
I
IIT'Ip Applied Air Loads:

I V component = (3770 + 1295 + 1838 + 631)
----~----=----~C:_._----~- _1 .9986 = 7523 lb. (error 3 lb.)
58.5 ~1 D component = -185 x 6 = lllO lb. (error = 0)
S component = -(3770 + 1295 + 1838 + 631)
Fig. A2.39 .0523 =394 lb. (error 6 lb.)
The wing beams due to the distributed lift Lengths & Directional Components of Cabane Struts
air loads acting upon them, are also sUbjected
to bending loads in addition to the axial loads. Member Sym. V D S L V!L D/L S/L
The wing be~~s thus act as beam-columns. The Front Cabane CF 30 10 27 41.59 .721 .240 .648
subject of beam-column action is treated in Diagonal Cabane 30 30 27 50.17 .597 .597 .538
CD
another chapter of this book. Strut
Rear Cabane CR 29.5 6 27 40.42 .731 .1485 .668
If the wing is covered with metal skin Strut
instead of fabric, the drag truss can be omitted L =Vv 2 + D2 + S2
since the top and bottom skin act as webs of a
beam which has the front and rear beams as its
flange members. The wing is then considered as The air loads on the outer panel are taken
a box beam subjected to combined bending and identical to those in example problem 1. Like-
axial loading. wise the dihedral and direction of the lift
struts SF and SR have been made the same as in
Example Problem 11. 3-Section Externally Braced example problem 1. Therefore the analysis for
Wing. the loads in the outer panel drag and lift trus
trusses is identical to that in problem 1. The
Fig. A2.41 shows a high wing externally solution will be continued assuming the running
braced wing structure. The wing outer panel has lift load on center panel of 45 Ib./in. and a
been made identical to the wing panel of ex~~ple forward drag load of 6 Ib./in.
problem 1. This outer panel attached to the cen
ter panel by single pin fittings at points (2) Solution of Center Panel
and (4). Placing pins at these points make the
structure statically determinate, whereas if the Center Rear Beam
beams were made continuous through all 3 panels,
692# (630+62)=692#
the reactions of the lift and cabane struts on
the wing beams would be statically indeterminate w= 14. 74#/in.
since we would have a 3-span continuous beam
resting on settling supports due to strut de- (4) r-----t-::.,---------.,--:::-:-t--~
formation. The fitting pin at points (2) and 90
(4) can be made eccentric with the neutral axis R s = 1650# R s = 1650#
of the beams, hence very little is gained by Fig. A2. 42
making beams continuous for the purpose of de-
creasing the lateral beam bending moments. For Fig. A2.42 shows the lateral loads on the center
assembly, stowage and shipping it is convenient rear beam. The loads consist of the distributed
to build such a wing in 3 portions. If a air load and the vertical component of the for-
mUltiple bolt fitting is used as points (2) and ces exerted by outer panel on center panel at
(4) to obtain a continuous beam, not much is pin point (4). From Table A2.2 of example prob-
gained because the design reqUirements of the lem 1, this reSUltant V reaction equals 630 +
various governmental agencies specify that the 62 = 692 lb.
wing beams must also be analyzed on the as-
smnption that a mUltiple bolt fitting prOVides The vertical component of the cabane re-
only 50 percent of the full continUity. action at joint (8) equals one half the total
(l)
OJ) beam load due to symmetry of loading or 65 x
~.S
,0:: 14.74 + 692 =
1650 lb.
(8) :(4) --L
Solution of force system at Joint 8
36"
1650
1650
T
CRB
/~
CR S
f8
D~C
~D
ZV - 1650 - .731 CR = 0
whence
CR =1650/.731 =2260 lb. (tension)
ZS = - CRB - 2260 x .668 =0

Fig. A2. 41
A2.19
whence (4) are taken from Table A2.2 of problem 1. The

drag load of 336 lb. at (8) is due to the rear
CRB 1510 lb. (compression) cabane strut, as is likewise the beam axial load
of - 1510 at (8). The axial beam load of
ZD = De - 2260 X .1485 =0 - 2281 lb. at (7) is due to reaction of front
cabane truss. The panel point loads are due to
whence the given running drag load of 6 Ib./in. acting
forward.
De = 336 lb. drag truss reaction
The reaction which holds all these drag
Center Front Beam truss loads in equilibrium is supplied by the
cabane truss at point (7) since the front and
568# (1294 - 726) = 568# diagonal cabane struts intersect to form a rigid
w = 30.26 #/in.
triangle. Thus the drag reaction R equals one
(Ref. , Table A2. 2) half the total drag loads or 2634 lb.
I- 20 --!"7iL-- _ 90 ----1"-'-7)20-1
Solving the truss for the loading of Fig.
R., = 2535# Rr =2535# A2.44 we obtain the member axial loads of Fig.
Fig. A2.43 A2.45.
568 568
Fig. A2.43 shows the V loads on the center front
beam and the resulting V component of the cabane
reaction at joint (7). o o
e:- «>
N I
Solution of force system at Joint 7 I
-15027 -17308 -17308
~
2535
(11)
CFB 7 Fig. A2.45
CD VS Plane Loads in Cabane Struts Due to Drag Reaction at

CF Point 7
ZV = 2535 - .721 CF - .597 CD =0 263~

ZS - - CFB - .648 CF - .538 CD = 0
ZD - - .240 CF + .597 CD =0 kF "".
V-D Plane CD
Solving the three equations, we obtain

ZD - - 2634 - .240 CF + .597 CD =0
CFB =- 2281 (compression)
ZV .721 CF - .p97 CD 0 =
CF -- 2635
Solving for CF and CD, we obtain
CD = 1058
Solution for Loads in Drag Truss Members CF - - 2740 lb. (compression)
CD = 3310 (tension)
Fig. A2.44 shows all the loads applied to
the center panel drag truss. The Sand D re- adding these loads to those previously calcu-
actions from the outer panel at joints (2) and lated for lift loads:
Panel point 336 CF =- 2740 + 2635 =- 105

Drag Load 1
336 I CD = 1058 + 3310 = 4368 lb.
for 6#/in.~ ~ 270 t 60
1157 6f 1 ~5~ ! 1~1~~.J95 ~ 1157

CR = 2260 lb.
Fuselage Reactions
-~~-- As a check on the work the fuselage re-
actions will be checked against the applied
13lli-~~VLj~7
1 t • '228f 2m (2)1
loads.
ponents
Table A2.3 gives the V, D and S com-
of the fuselage reactions.
1908 t iR J R7 t 1908
26341t Fig. A2. 44 2634#
Table A2.3
Point Kember Load V D S
9 Front Strut -105 -76 -25 -68

CF
10 Rear Strut CR 2260 1650 335 1510
Dia. Strut CD 4368 2610 -2610 2356
5 Front Lift Strut 9333 4205 798 8290

SF
6 Rear Lift Strut 4579 2055 0 4090
SR
Totals 10444 -1502 16178
Applied Air Loads

z Fig. A2.46
V component - 7523 (outer panel) + 65 x 45
=
10448 (check)
D component - - 1110 (outer panel) - 65 x
6 - - 1500 (error 2 lb.) ~"/;'O
'!i
~1
~'O
~1
/t: z
'!>~ 504
The total side load on a vertical plane thru
centerline of airplane should equal the S com-
ponent of the applied loads. The applied side 0 '1::J
a:>
loads = - 394 lb. (see problem 1). The air load "'bil"
~
on center panel is vertical and thus has zero S
component. T e ri:
r
36"
) '"
From Table A2.3 for fuselage reactions til
have ZS =
16178. From Fig. A2.45 the load in .9
the front beam at ~ of airplane equals - 17308
and 568 for rear beam. The horizontal component
A B C D E \e It>
c>
A B C D E
of the diagonal drag strut at joints 11 equals T
216 x 45/57.6 = 169 lb. 9"
Then total S components = 16178 - 17308 + ~36" ~

568 + 169 =-
393 lb. which checks the side Fig.46c
component of the applied air loads.
Example Problem 12. Single Spar Truss Plus SOLUTION:

Torsional Truss System.
The total air load on the structure equals
In small wings or control surfaces, fabric the average intensity per square inch times the
is often used as the surface covering. Since surface area or (0.5)(.5)(36 x 84) 756 lb. In =
the fabric cannot provide reliable torsional order to solve a truss system by a method of
resistance, internal structure must be of such joints the distributed load must be replaced by
design as to provide torsional strength. A an equivalent load system acting at the joints
single spar plus a special type of truss system of the structure. Referring to Fig. (d), the
is often used to give a satisfactory structure. total air load on a strip 1 inch wide and 36
Fig. A2.46 illustrates such a type of structure, inches long is 36(0.5)/2 9 lb. and its c.g. =
namely, a trussed single spar AEFN plus a tri- or resultant location is 12 inches from line AE.
angular truss system between the spar and the In Fig. 46a this resultant load of 9 Ib./in. is
trailing edge OS. Fig. A2.46 (a, b, c) shows imagined as acting on an imaginary beam located
the three projections and dimensions. The air along the line 1-1. This running load applied
load on the surface covering of the structure along this line is now replaced by an equivalent
is assumed to be 0.5 Ib./in. 2 intensity at spar force system acting at joints OPQRSEDBCA. The
line and then varying linearly to zero at the reSUlts of this joint distribution are shown by
trailing edge (See Fig. d). the joint loads in Fig. A2.46. To illustrate
how these joint loads were obtained, the calcu-
The problem will be to determine the axial lations for loads at joints ESDR will be given.
loads in all the members of the structure. It
will be assumed that all members are 2 force Fig. A2.48 shows a portion of the structure
members as is usually done in finding the to be considered. For a running load of 9
primary loads in trussed structures. Ib./in., along line 1-1, reactions will be for
A2.21
L;My = (6 + 66 + 72 + 72 + 36) 36 - 36 Oz =0
T
24"
whence Oz = 252 lb. acting down as assumed.
To find Oy take moments about z axis through
point (A).
12"
L;Mz =0 + 36 Oy = 0, Oy =0
-"- To find Ay take moments about x axis through
E
--+-12"-1 Fig.A2.48 point N. The moment of the air loads was pre-
viously calculated as - 31752, hence,
simple beams resting at points 2, 3, 4, 5, etc. L;Mx =- 31752 + 9 Ay = 0, whence, Ay =
The distance between 2-3 is 8 inches. The total 3528 lb.
load on this distance is 8 x 9 = 72 lb. One
half or 36 lb. goes to point (2) and the other To find Ny take L;F y =0
half to point (3). The 36 lb. at (2) is then
replaced by an equivalent force system at E and L;Fy =3528 - Ny = 0, hence Ny = 3528 lb.
8 or (36)/3 =
12 lb. to Sand (36)(2/3) =
24 to To find Nz take L;F z 0 =
E. The distance between points (3) and (4) is
8 inches and the load is 8 x 9 =
72 lb. One ZF z = - 252 + 756 - Nz = 0, hence Nz =
half of this or 36 goes to point (3) and this 504 lb.
added to the previous 36 gives 72 lb. at (3).
The load of 72 is then replaced by an equivalent The reactions are all recorded on Fig.
force system at Sand D, or (72)/3 =
24 lb. to A2.46.
8 and (72)(2/3) =
48 to D. The final load at S
=
is therefore 24 + 12 36 lb. as shown in Fig. Solution of Truss Member Loads
A 2.46. Due to symmetry of the triangle CRD,
one half of the total load on the distance CD For simplicity, the load system on the
goes to points (4) and (5) or (24 x 9)/2 108 = structure will be considered separately as two
lb. The distribution to D is therefore (108) load systems. One system will include only
(2/3) = 72 and (108)/3 =
36 to R. Adding 72 those loads acting along the line AE and the
to the previous load of 48 at D gives a total second load system will be remaining loads
=
load at D 120 lb. as shown in Fig. A2.46. which act along line 08. Since no bending
The 108 lb. at point (5) also gives (108)/3 = moment can be resisted at joint 0, the external
36 to R or a total of 72 lb. at R. The student load along spar AE will be reacted at A and N
should check the distribution to other joints entirely or'in other words, the spare alone
as shown in Fig. A2.46. resists the loads on line AE.
To check the equivalence of the derived Fig. A2.49 shows a diagram of this spar
joint load system with the original air load with its joint external loading. The axial
system, the magnitude and moments of each loads produced by this loading are written on
system must be the same. Adding up the total the truss members. (The student should check
joint loads as shown in Fig. A2.46 gives a total these member loads.)
of 756 lb. which checks the original air load.
The moment of the total air load about an x
axis at left end of structure equals 756 x 42 =
31752 in. lb. The moment of the joint load
system in Fig. A2.46 equals (66 x 12) + (72 x
36) + (72 x 60) + (56 x 84) + 144 (24 + 48) +
(120 x 72) + (24 x 84) =
31752 in. lb. or a 224 224 32
check. The moment of the total air load about
line AE equals 756 x 12 =9072 in. lb. The
Fig. A2.49
moment of the distributed joint loads equals TRIANGULAR TRUSS SYSTEM
(6 + 66 + 72 + 72 + 36)36 =
9072 or a check.
The load system along the trailing edge OS
Calculation of Reactions causes stresses in both the spar truss and the
diagonal truss system. The support fitting at
The structure is supported by single pin point 0 provides a reaction in the Z direction
fittings at points A, Nand 0, with pin axes but no reacting moment about the x axis. Since
parallel to x axis. It will be assumed that the loads on the trailing edge lie on a y axis
the fitting at N takes off the spar load in through 0, it is obvious that all these loads
Z direction. Fig. A2.46 shows the reactions flow to point O. Since the bending strength of
0y, 0z, Ay , Ny, Nz • To find Oz take moments the trailing edge member is negligible, the
about y axis along spar AEFN.
A2 22 EQUILIBRIUM OF FORCE SYSTEMS TRUSS STRUCTURES
load of 36 lb. at Joint S in order to be trans- reaction on the spar but does produce a couple
ferred to point 0 through the diagonal truss force on the spar in the Y direction which pro-
system must follow the path SDRCQBPAO. In like duces compression in the top chord of the spar
manner the load of 72 at R to reach 0 must take truss and tension in the bottom chord.
the path RCQBPAO, etc.
Consider Joint R
Calculation of Loads in Diagonal Truss Members:-
The load to be transferred to truss RCJR
Member z y x L z/L y/L x/L is equal to the 72 lb. at R plus the 36 lb. at
S which comes to joint R from truss DRG.
All Diagonal 4.5 12 36 38.2 .118 .314 .943
Truss Members Hence load in RC = (72 + 36)0.5 x (1/.118)
AO, NO 4.5 0 36 37.5 .120 o .960 457 lb.
Whence RJ =457, CQ =457 and JQ =- 457
Consider Joint S
Joint Q
The triangular truss SEF cannot assist in Load to be transferred to truss QBL =72 +
transferring any portion of the 36 lb. load at
S because the reaction of this truss at EF
72 + 36 =
180 lb.
would put torsion on the spar and the spar has Hence load in QB = (180 x 0.5)(1/.118) =
no appreciable torsional resistance. - 762.
Considering Joint S Whence QL = 762, BP = 762, LP =- 762

as a free body and writing
the equilibrium equations: Joint P
L:Fx =-.943 DS -.943 GS = 0 Load = 180 + 66 = 246

whence, DS =- GS Load in PA = (246 x 0.5)(1/.118) =- 1040
L:F z = 36 + .118 DS - .118 GS = 0
Whence PN = 1040
Consider Joint (A) as a free body.
Subt. DS =- GS and solving for GS, gives
GS = 159 lb. (tension), DS = -159 (compression) L:F x =- 1040 x .943 + .960 AO = 0, AO =
1022 lb.
Consider Joint D In like manner, considering Joint N, gives NO
z
=- 2022 lb.
Ly R~159 Couple Force Reactions on Spar
TY-#:E159 T 159
T
Y D
'E As pointed out previously, the diagonal
torsion truss produces a couple reaction on the
z spar in the y direction. The magnitude of the
force of this couple equals the y component of
Let Ty and Tz be reactions of diagonal the load in the diagonal truss members meeting
truss system on spar truss at Joint D. at a Spar joint. Let Ty equal this reaction
load on the spar.
L:F x =- 159 x .943 + .943 DR = 0, hence DR =
159 lb. At Joint C:-
L:Fz =- 159 x .118 + 159 x .118 - Tz =0 Ty =- (457 + 457).314 =- 287 lb.
whence Tz =
0, which means the diagonal Likewise at Joint J, Ty =287
truss produces no Z reaction or shear load on
spar truss at D. At Joint B:-
L:Fy =- .314 x 159 - .314 x 159 - Ty =0 Ty =- (762 + 762).314 =- 479

whence Ty =- 100 lb.
Likewise at Joint L, Ty =479
At Joint A:-
If joint G is investigated in the same manner,
the results will show that Tz = 0 and Ty = 100. Ty =- (1040 x .314) =- 326
The results at joint D shows that the rear Likewise at Joint N, Ty =326
diagonal truss system produces no shear load
A2.23
These reactions of the torsion truss upon

the spar truss are shown in Fig. A2.50. The
loads in the spar truss members due to this
loading are written adjacent to each truss
member. Adding these member loads to the loads
in Fig. A2.49, we obtain the final spar truss
member loads as shown in Fig. A2.51. Level Landing with Inclined Reactions
Level Landing with Nose Wheel Just

Clear of Ground
Fig. A2. 51
Tail Down Landing

If we add the reactions in Figs. A2.50 and
A2.49, we obtain 3528 and 504 which check the
reactions obtained in Fig. A2.46. Fig. A2.52
A2. 13 Landing Gear Structure

Fig. A2.53 shows photographs of typical
The airplane is both a landborne and air- main gear units and Fig. A2.54 for nose wheel
borne vehicle, and thus a means of operating gear units.
the airplane on the ground must be provided
which means wheels and brakes. Furthermore, The successful design of landing gear for
provision must be made to control the impact present day aircraft is no doubt one of the most
forces involved in landing or in taxiing over difficult problems which is encountered in the
rough ground. This requirement requires a structural layout and strength design of air-
special energy absorption unit in the landing craft. In general, the gear for aerodynamic
gear beyond that energy absorption provided by efficiency must be retracted into the interior
the tires. The landing gear thus includes a of the wing, nacelle or fuselage, thus a re-
so-called shock strut commonly referred to as liable, safe retracting and lowering mechanism
an oleo strut, which is a member composed of system is necessary. The wheels must be braked
two telescoping cylinders. When the strut is and the nose wheel made steerable. The landing
compressed, oil inside the air tight cylinders gear is subjected to relatively large loads,
is forced through an orifice from one cylinder whose magnitudes are several times the gross
to the other and the energy due to the landing weight of the airplane and these large loads
impact is absorbed by the work done in forcing must be carried into the supporting wing or
this oil through the orifice. The orifice can fuselage structure. Since the weight of land-
be so designed as to provide practically a ing gear may amount to around 6 percent of the
uniform resistance over the displacement or weight of the airplane it is evident that high
travel of the oleo strut. strength/weight ratio is a paramount design
requirement of landing gear, as inefficient
An airplane can land safely with the air- structural arrangement and conservative stress
plane in various attitudes at the instant of analysis can add many unnecessary pounds of
ground contact. Fig. A2.52 illustrates the weight to the airplane and thus decrease the
three altitudes of the airplane that are payor useful load.
specified by the government aviation agencies
for design of landing gear. In addition to A2. 14 Example Problems of Calculating Reactions and
these symmetrical unbraked loadings, spec ial Loads on Members of Landing Gear Units
loadings, such as a braked condition, landing In its simplest form, a landing gear could
on one wheel condition, side load on wheel, etc., consist of a single oleo strut acting as a
are required. In other words, a landing gear cantilever beam with its fixed end being the
can be sUbjected to a considerable number of upper end which would be rigidly fastened to
different loadings under the various landing the supporting structure. The lower cylinder
conditions that are encountered in the normal of the oleo strut carries an axle at its lower
operation of an airplane.
McDonnell Aircraft
(Military Airplane)
Douglas DC-7 Air Transport
Douglas DC-8
Jet Airliner
Fig. A2.53 Main Landing Gear Illustrations (One Side)

A2.25
Beechcraft Twin Bonanza
Piper-Apache
Piper Tri-Pacer
Navy F4-J
North American Aviation Co.
Douglas DC -7 Air Transport

Fig. A2. 54 Nose Wheel Gear Installations
end for attaching the wheel and tire. This SOLUTION

cantilever beam is subjected to bending in two
directions, torsion and also axial loads. Since The supporting reactions upon the gear at
the gear is usually made retractable, it is points F, H, and G will be calculated as a
difficult to design a single fitting unit at beginning step. There are six unknowns, namely
the upper end of the oleo strut that will FS, Fy, FD, Hy, HD and G (See Fig. A2.56). With
resist this combination of forces and still 6 equations of static equilibrium available for
permit movement for a simple retracting mechan- a space force system, the reactions can be found
ism. Furthermore, it would be difficult to by statics. Referring to Fig. A2.56:-
provide carry-through supporting wing or fuse-
lage structure for such large concentrated To find FS take ~S =0
load systems.
FS + 0 = 0, hence FS =0
Thus to decrease the magnitude of the
bending moments and also the bending flexibility To find reaction Gy take moments about an S
of the cantilever strut and also to simplify axis through points F, H.
the retracting problem and the carry-through =
ZMs 3119 x 50 - 24 Gy 0 =
structural problem, it is customary to add one
or two braces to the oleo strut. In general, Whence, Gy =
6500 lb. with sense as assumed.
effort is made to make the landing gear (The wheel load of 15000 lb. has been resolved
structure statically determinate by using into Y and D components as indicated in Fig.
specially designed fittings at member ends or A2.55) .
at support points in order to establish the With Gy known, the reaction G equals (6500)
force characteristics of direction and point of (31.8/24) =8610 lb. and similarly the compon-
application. ent GD = (6500)(21/24) 5690 lb. =
Two example problem solutions will be pre-
sented, one dealing with a gear with a single
f6"1--
G 16', /
wheel and the other with a gear involving two !!J]:~ --;
wheels.
Example Problem 13
Fig. A2.55 shows the projections of the

l / .
(f
/
landing gear configuration on the VB and VD

planes. Fig. A2.56 is a space dimensional
diagram. In landing gear analysis it is common
to use Y, D and S as reference axes instead of
the symbols Z, X and Y. This gear unit is
assumed as representing one side of the main
gear on a tricycle type of landing gear system.
The loading assumed corresponds to a condition
of nose wheel up or tail down. (See lower
sketch of Fig. A2.52). The design load on the
wheel is vertical and its magnitUde for this
problem is 15000 lb.
The gear unit is attached to the supporting

structure at points F, H and G. Retraction of
the gear is obtained by rotating gear rearward
and upward about axis through F and H. The
fittings at F and H are designed to resist no
bending moment. Hence, reactions at F and Hare
unknown in magnitUde and direction. Instead of
using the reaction and an angle as uhknowns,
the resultant reaction is replaced by its Y and
D components as shown in Fig. A2.56. The re-
action at G is unknown in magnitUde only since
the pin fitting at each end of member GC fixes
the direction and line of action of the reaction
at G. For convenience in calculations, the
+&r
10063
F
6500
3004
t....5690
G
1~109
reaction G is replaced by its components Gy and
GD. For a side load on the landing gear, the
\,n .t 433
H
reaction in the S direction is taken off at
point F by a specially designed unit. Fig.A2.57
A2.27
~
TE the structure as a whole by taking moments about
t D and V axes through point A.
f'V tri 11
ESTE =1;~1lrc ~MA{D) =- 10063 x 14 + 6500 x 8 + 11109 x 8
=- 140882 + 52000 + 88882 = O(check)
24"
f,. ~~~
I r~ ~ = 5690 x 8 - 433 x 8 - 3004 x 14
+
I'~ ~S
I •
,
BIS=-!.LBI
24.6
= 45520 - 3464 - 42056 = 0 (check)
The next step in the solution will be the
calculation of the forces on the oleo strut
unit. Fig. A2.58 shows a free body of the oleo-
strut-axle unit. The brace members BI and CG
~8l=- are two force members due to the pin at each
I 14672# end, and thus magnitude is the only unknown re-
Fig.A2.58 action characteristic at points Band C. The
fitting at point E between the oleo strut and
the top cross member FH is designed in such a
manner as to resist torsional moments about the
oleo strut axis and to provide D, V and S force
reactions but no moment reactions about D and S
axes. The unknowns are therefore BI, CG, ES '
EV' ED and TE or a total of 6 and therefore
statically determinate. The torsional moment
TE is represented in Fig. A2.58 by a vector
with a double arrow. The vector direction
represents the moment axis and the sense of
rotation of the moment is given by the right
hand rule, namely, with the thumb of the right
hand pointing in the same direction as the
arrows, the curled fingers give the sense of
rotation.
Fig. A2. 59
To find the resisting torsional moment TE
take moments about V axis through E.
To find FV' take moments about a D axis
through point H. ~ME{V) =- 3119 x 8 + TE = 0, hence TE
= 24952 in. lb.
~MH(D) = 16 GV + 14672 x 8 - 22 FV = 0
=16 x 6500 + 14672 x 8 - 22 FV = 0 To find CG take moments about Saxis
through E.
Whence FV = 10063 lb. with sense as as-
sumed. ~ME(S) = 3119 x 50 - (24/31.8) CG x 3 -
24 (21/31.8) CG 0 =
To find HD, take moments about V axis Whence, CG = 8610 lb.
through F.
~MF(V) - - 6 GD - 22 HD + 3119 x 14 0 = This checks the value previously obtained
- - 6 x 5690 - 22 HD + 3119 x 14 when the reaction at G was found to be 8610.
=0 The D and V components of CG thus equal,
Whence, HD = 433 lb. CGD = 8610 (21/31.8) = 5690 lb.
To find FD' take ~D =0 CGV = 8610 (24/31.8) = 6500 lb.
~D =- FDHD + GD - 3119
+ 0 = To find load in brace strut BI, take moments
=- FD433 + 5690 - 3119 = 0
+ about D axis through point E.
Whence, FD 3004 lb. = ~ME{D) =- 14672 x 8 + 3 (BI) 22/24.6 +
24 (BI) 11/24.6 0 =
To find HV take ~V =0 Whence, BI = 8775 lb.
~V =- FV + GV - HV + 14672 =
0 and thUS, BIV = (8775)(22/24.6) = 7840 lb.
=- 10063 + 6500 - HV + 14672 =0 BIS = (8775)(11/24.6) = 3920 lb.
Whence, HV =11109 lb.
To find Es take =0~S
Fig. A2.57 summarizes the reactions as found.
The results will be checked for equilibrium of ~s = Es - 3920 = 0, hence ES = 3920
To find ED talee ZD =0
ZD =5690 - 3119 - ED =0, hence ED =2571
To find EV take ZV 0 =
Zy =-Ey + 14672 - 7840 + 6500 = 0, hence
EV = 13332 lb.
Fig. A2.59 shows a free body of the top
member FH. The unknowns are Fy, FD, FS' Hy and
HD. The loads or reactions as found from the
Lower
analysis of the oleo strut unit are also re- Cylinder
corded on the figure. The equations of
equilibrium for this free body are:- Fig.A2.60
Torque
zs = 0 =- 3920 + 3920 + FS = 0, or FS =0 Link
ZMF(D) =22 Hy -
3920 x 2 - 7840 x 20 -
13332x6=0
=
Whence, Hy 11110 lb. This check value
was obtained previously, and therefore is a
check on our work.
ZMF(y) = 24952- 2571 x 6 - 22 HD =0 The torque to be transferred in our prob-
lem is 24952 in. lb.
whence, HD = 433 lb.
Zy =- Fy + 13332 + 7840 - 11110 =0 The reaction R.. between the two units of
the torque link at point (2), see Fig. A2.60,
whence, Fy = 10063 thus equals 24952/9 = 2773 lb.
ZD - - FD + 2571 + 433 = 0 The reactions R a at the base of the link at
whence, FD =3004 lb. point (3) = 2773 x 8.5/2.75 = 8560 lb. With
these reactions known, the strength design of
Thus working through the free bodies of the link units and the connections could be
the oleo strut and the top member FH, we come made.
out with same reactions at F and H as obtained
when finding these reactions by equilibrium Example Problem 14
equation for the entire landing gear.
The landing gear as illustrated in Fig.
The strength design of the oleo strut unit A2.61 is representative of a main landing gear
and the top member FH could now be carried out which could be attached to the under side of a
because with all loads and reactions on each wing and retract forward and upward about line
member known, axial, bending and torsional AB into a space provided by the lower portion
stresses could now be found. of the power plant nacelle structure. The oleo
strut OE has a sliding attachment at E, which
The loads on the brace struts CG and BI prevents any vertical load to be taken by
are axial, namely, 8610 lb. tension and 8775 member AB at E. However, the fitting at E does
lb. compression respectively, and thus need no transfer shear and torque reactions between the
further calculation to obtain design stresses. oleo strut and member AB. The brace struts
GD, FD and CD are pinned at each end and will
TORQUE LINK be assumed as 2 force members.
The oleo strut consists of two telescoping An airplane level landing condition with
tubes and some means must be provided to trans- unsymmetrical wheel loading has been assumed as
mit torsional moment between the two tubes and shown in Fig. A2.61.
still permit the lower cylinder to move upward
into the upper cylinder. The most common way SOLUTION
of providing this torque transfer is to use a
double-cantilever-nut cracker type of structure. The gear is attached to supporting struc-
Fig. A2.60 illustrates how such a torque length ture at points A, Band C. The reactions at
could be applied to the .oleo strut in our these points will be calcUlated first, treating
problem. the entire gear as a free body. Fig. A2.62
A2.29
r- ---+ ZMA(D) =- 60000 x 9 - 40000 x 29 - 66666
__
24"
AlB x 19 + 38 By 0 =
r'.~ rt~~~ whence, By =
78070 lb.
To find Ay, take ~y =0
28 " Brace
("'"
=-
1st'~D •
Fig.A2.61
~y
=0
78070 + 60000 + 40000 + 66666 - YA
whence, Ay = 88596 lb.

To find BD take moments about Y axis through
point (A).
36" Wheel 2
ZMA(y) =57142 x 19 - 15000 x 9 - 10000 x
15000
Wheel1 __
i 29 - 38 BD 0 =
whence, BD = 17386 lb.
10000 I
Wheel 2 ~ 40000 To find AD take ~D =0

6000 I I
110" o· ~D =- 57142 + 15000 + 10000 + 17386 +
AD =
0
whence AD = 14756 lb.
To check the results take moments about Y
and D axes through point O.
ZMo(y) =5 x 10000 + 14756 x 19 - 17386 x
19 = 0 (check)
~I"lO(D) = 20000 x 10 - 88596 x 19 + 78070 x
19 =0 (check)
REACTIONS ON OLEO STRUT OE
Fig. A2.63 shows a free body of the oleo-

64"
strut OE. The loads applied to the wheels at
;TE
r- 17"+ 17"
;TE
-1
~G___ E--ES~-
E
Fig.A2.62
ED-
<D
<D I DG
,\GV
/
DFV-<y4-
i
571~~ 28" t l/ 28"
shows a space diagram with loads and reactions.
The reactions at A, Band C have been replaced
_0
At
5-
~
DGS- ' I3--r::; DFS
I 'v 1-i.61t
l
by their Y and D components.
41/4-1 1-41/4
To find reaction Cy take moments about an VL
S axis through points AB.
L D
36"
s Fig. A2. 63
~~B
Whence Cy
=-
(15000 + 10000) 64 + 24 Cy = 0
=
66666 lb. With sense as
"0" o~ 25000 ~'O"
assumed in Fig. A2.62. 100,000 lb. 200,00~ --100,000 lb.
t 50000"lb. t 50,000"lb.
~\'%.9~
The reaction at C must have
a line of action along the line
CD since member CD is pinned at 28LS. the axle centerlines have been transferred to
each end, thus the drag compon- 24 D point (0). Thus the total Y load at (0) equals
ent and the load in the strut 60000 + 40000 =
100000 and the total D load
CD follow as a matter of geometry. Hence, equals 15000 + 10000 =
25000. The moment of
CD = 66666 (24/28) =
57142 lb. these forces about Y and D axes through (0) are
CD = 66666 (36.93/28) =
87900 lb. tension =
I"lO(y) (15000 - 10000) 10 50000 in. lb. and =
=
I"lO(D) (60000 - 40000) 10 200000 in.lb. =
To find By take moments about a drag axis These moments are indicated in Fig. A2.63 by
through point (A). the vectors with double arrows. The sense of
the moment is determined by the right hand thumb
and finger rule.

The fitting at point E is designed to
resist a moment about V axis or a torsional
moment on the oleo strut. It also can provide
shear reactions ES and ED but no bending
resistance about S or D axes.
The unknowns are the forces ES' ED' DF,

DG and the moment TE'
Fig. A2. 64
To find TE take moments about axis OE.
ZMOE =- 50000 + TE = 0, whence TE =
50000 in. lb.
ZMA(D) =- 89215 x 2 - 77451 x 36 - 38 BV
To find ES take moments about D axis =0
through point D. whence, BV =78070 lb.
ZMD(D) =200000 - 28 ES = 0, whence To find AV take ZV = 0
ES = 7143 lb.
ZV = 89215 + 77451 - 78070 - AV =0
To find force DFV take moments about D whence, AV 88596 lb. =
axis through point G.
To find BD take moments V axis througb A.
ZMa(D) =200000 - 100000 x 17 - 66666 x
17 + 34 DFV = 0 ZMV(D) =50000 + 32143 x 19 - 38 BD =0
whence, DFV =
77451 lb. Whence, BD 17386 lb. =
Then DFS =77451 (17/28) To find AD take ZD =0
= 47023 lb. 17
and DF = 77451 (32.72/28)

- 90503 lb.
28'1C
~:'~
~D = 17386 - 32143
14757 lb.
+ An = 0, or AD =
These four reactions check the reactions
To find DGV take ZV =0 obtained originally when gear was treated as a
ZV =
100000 - 77451 + 66666 - DGV = 0, free body, thus giving a numerical check on the
or DGV = 89215 calculations.
Then, DGS = 89215 (17/28) = 54164 lb. With the forces on each part of the gear
DG = 89215 (32.73/28) = 104190 lb. known, the parts could be designed for strength
and rigidity. The oleo strut would need a
To find ED take moments about Saxis torsion link as discussed in example problem 13
through point D. and Fig. A2.60.
ZMD(S) =- 25000 x 35 + 28 ED = 0,
ED =32143 lb.
A2.15 Problems
(1) For the structures numbered 1 to 10 deter-
The results will be checked for static mine whether structure is statically deter-
equilibrium of strut. Take moments about D minate with respect to external reactions
axis through point (0). and internal stresses.
ZMO(D) = 200000 + 54164 x 36 - 47023 x
36 - 7143 x 64 =
200000 + 1949904 Pin Pin Pin ~10
20
- 1692828 - 457150 0 (check) = ~ ~M3 ~ ~p ta.
B
Pin B
E'~IO
*
ZMO(S) =32143 x 64 - 57142 x 36 = O(check) (1)
Pm
~
10
-10
(2) ,
A ' (3)
10~
REACTIONS ON TOP MEMBER AB
Fig. A2.64 shows a free body of member AB ~-10 10

(5) If ~O (6)
with the known applied forces as found from Al ~ (4) L.:;AO===~A At.. B
the previous reactions on the oleo strut. A B 10 10
(7) t + 'I --+~

.. t . (8)
~
The unknowns are AD' BD, A and BV' To
find BV take moments about D axIs through A. ADm C
ABC B
A2.31
~ + + + ~
(9)~ ~(10)
A
+ + + B ~B
A
(2) Find the horizontal and vertical components I'

of the reactions on the structures illus- Wing
trated in Figs. 11 to 15.
+0
.J+V Front View
Side View
Fig. 20
20"
..l
(6) Fig. 21 shows the wing structure of an ex-
1-16'4\ (12) ternally braced monoplane. Determine the
axial loads in all members of the lift and
drag trusses for the following loads.
(14) Front beam lift load =
30 lb./in. (upward)
Rear beam lift load =
24 lb./in. (upward)
100 Wing drag load =
8 lb./in. acting aft
PLAN VIEW
(3) Find the axial loads in the members of the Wing Drag Truss Anti-Drag Wires
trussed structures shown in Figs. 16 to 18. Drag Wires
C' B'
.10 ;1.0 t10
-*- T T (16)
15"~'
T 30"
30"
10 -.t..
1. 24 +24 4024"4 36" -j!!.- 45"--l
P.OOO
~"~-:k
.L
--l..
2000
t
I
- 00
10"
(18)
3t".1-Ct
1 0
CR
DO'
I- I 1 30,0-01 '
I I I
!--30"-+15"+15'.....j
~tt-
<rl A,A'
Fuselage
Fig. 21
(4) Determine the axial loads in the members
of the structure in Fig. 19. The members (7)
are pinned to supports at A, Band C.
I--- 5 Bays @ 3&' ---I
--l rl C 0
j.-15'..j.lo.:r- 20"
I
T
30"
·~-DragTrUssLS
I
I ..i. AI t t f 1ft t t t t t t tt 25 lb/in.
I
I
~2 /.in,
• 5 lb/in.
. ...JI"5"
.. #/in.
I
Pin Reactions I BC B
I 1
at A, B, C I 60" Fuse- I
1. lage I
Fig. 19 t100# E
I
~~
100# Fig. 22
10"
(5) Fig. 20 shows a tri-pod frame for hoisting Fig. 22 shows a braced monoplane wing. For the
a propeller for assembly on engine. Fipd given air loading, find axial loads in 11ft and
the loads in the frame for a load of 1000 drag truss members. The drag reaction on drag
lb. on hoist. truss is taken off at point A.
(8) (9)
The fitting at In the
points A and B for landing gear
~_ _A B
E
C the landing gear of Fig. 24,
F the brace
structure in Fig. 23
provides resistance members BC
to V, D and S re- and BF are
24" two force
-1.3000
actions and moments N'
members. The
lb. about D and V axes.
Find the reactions at
30" L D
fitting at E
Wheel~_= W~:'j;~'CD
Fig. 23
10000 lb. 10000 A and B and the load provides re-
in member CD for sistance to
given wheel loading.
, 8"
wheel 2
V, D and S
reactions but
15000 10000 Fig. 24 only moment
resistance
about V axis.
Find reactions at E and loads in members BF and
BC under given wheel loading.
Cessna Aircraft Douglas DC-8 Jet Airliner

Nose Wheel Installation (Model 182) Main Landing Gear Unit
CHAPTER A3
PROPERTIES OF SECTIONS - CENTROIDS
MOMENTS OF INERTIA ETC.
A3.l Introduction . In engineering calculations, Then

two terms, center of gravity and moment of iner-
tia, are constantly being used. Thus, a brief J
Ix = yA dA, I y =J
x ta dA, I z =/rtadA
review of these terms is in order. Where Ix, I y and I z are moments of inertia of the
area about the axes xx, yy and zz respectively.
A3.2 Centroids, Center of Gravity. The cen-
troid of a line, area, ·volume, or mass ts that y
point at which the whole line, area, volume, or
mass may be conceived to be concentrated and have
the same moment With respect to an axis as when
distributed in its true or natural way. This '---t,--_..::.°t<---r--+-t
general relationship can be expressed by the
principle of moments, as follows:
Lines:- XL = .zLx, hence x = ZL;' =f xdL Fig. A3.1
L L
Areas: - XA = Zax, hence x = .zax = J'xdA A3 • .5
Polar MOII.ent of Inertia • In Fig. A3.1, the
A A moment of inertia I z =f r 8dA about the Z axis is
Volumes:- xV = .zVx, hence x = .zVx =J'xdV referred to as the polar moment of inertia and
V --V- can be defined as the moment of inertia of an
area With respect to a point in its surface.
Masses:- XM = .zxm, hence x = .zmx =JXdM Since r ta = x a + y8 (Fig. A3.1)
M M Iz =J(Y8 + x 8) dA = Ix + I y or; the polar moment
If a geometrical figure is symmetrical With re- of inertia is equal to the sum of the moments of
spect to a line or plane, the centroid of the inertia With respect to any two axes in the plane
figure lies in the given line or plane. This is passing thru at
of the area, right angles to each other and
the point of intersection ot the po-
obvious from the fact that the moments of the
parts of the figure on the opposite sides of the lar axis with the plane.
line or plane are numerically equal but of op- The radius of gyration
posite sign. If a figure is symmetrical to two A3.6 Radius of Gyration
of a solid is the distance
lines or planes, the centroid of the figure lies to that point in the solid at from the inertia axis
at the intersection of the two lines or the two tire which, if its en-
planes, and likewise, 11' the figure has 3 planes mass could be concentrated, its moment of
of symmetry, the centroid lies at the intersec- inertia would remain the same.
tion of the 3 planes. ThUS,!r 8dM =faM, where e is the radius of
gyration f
A3.3 Moment of Inertia The term moment of in-
ertia is applied in mechanics to a number of
=
Since, r 8dM I, then I =faM or f =\{f
mathematical expressions which represents sec- By analogy, in the case of an area,
ond moments of areas, volumes and masses, such
as
fy 8dA, f r 8dV, J radM etc.
I =e e =Vf
8A or
A3,4 Moaent of Inertia of an Area As applied A3 •.7

Parallel Axis Theorea In Fig. A3.2 let I y
to an area, the term moment of inertia has no be the moment of inertia of the area referred to
physical significance, but represents a quantity the centroidal axis y-y, and let the moment of
entering into a large number of engineering inertia about axis Y1Y1 be required. Y1Y1 is
problems or calculations. However, it may be parallel to yy. Consider the elementary area dA
considered as a factor which indicates the in- With distance x + d trom Y1Y1.
fluence of the area itself in determining the
total rotating moment of uniformly varying for- Then, I Y1 =~(d + X)8dA
ces applied over an area. = f x 8dA + 2df xdA + d8fdA
Let Fig. A3.1 be any plane area referred to
three coordinate axes, ox, oy and oz; ox and oy
being the plane of area.
Let dA represent an elementary area, with
coordinates x, y, and r as shown.
A3.l
A3.2 CENTROIDS, C ENTER OF GRA VITY, MOMENTS OF INERTIA
y TABLE I
I Area A
Section Properties of Areas
d ---l-- x --<i
I 'j'dA Rectangle Area = bd
t c . g. Fig. A3.2
r-b~
Y=d
:2
bd 3 /12
I
l~-l bd 3
2-~-2
y
3"
The first term,~x2dA, represents the mo-
ment of inertia of the body about its centroidal fJ.-J. .289 d
axis y-y and will be given the symbol I. The t------------------------I
second term is zero because~xdA is zero since Area bh
yy' is the centroidal axis of the body. The 1f
last term, d 2.;rdA = Ad 2 or, area of body times y h/3
the square of the distance between axes yy and
y~y~.
Thus in general, 12 _ 2
I = I + Ad 2 1
This expression states that the amount of
inertia of an area With respect to any axis in bh 3
the plane of the area is equal to the moment of T
inertia of the area with respect to a parallel
centroidal axis plus the product of the area and f~-~ .236 h
the square of the distance between the two axes. ~------------------~--------------------------~
8
I 2 Area bh
Parallel Axis Theorem For Masses. If instead of T~riange 1f
area the mass of the body is considered, the
parallel axis can be written:
1 - -+-2 - ---l-"fI x 2/3 h
b hb 3
I = I + Md 2 , where M refers to the mass of x I~_~ 48
the body. n ~
.204 h
A3.7a Mass Moments of Inertia The product of
the mass of a particle and the square of its
distance from a line or plane is referred to as
the moment of inertia of the mass of the parti-
Trapezoid
r- b1-j Area = d(b + b ~)
Y = d(2b + b~)
2
cle With respect to the line or plane. Hence,

3{b + b~)
I = ZMr2. If the summation can be express-
ed by a definite integral, the expression may be I~_~ d 3 (b2+4bb~+b 2)
written I =~r2 dM 36 (b + b 1 ~
Moments of Inertia of Airplanes. In both flying Area = t(b + c)
and landing conditions the airplane may be sub-
jected to angUlar accelerations. To determine x = b 2 + ct Y= d 2 + at
the magnitude of the accelerations as well as 2{b + c) 2{b+c)
the distribution and magnitude of the mass iner- Ix = ~ [t (d-y) 3+ by3 _ a (y _ t) 3]
tia resisting forces, the moment of inertia of 3
the airplane about the three coordinate axes is
generally required in making a stress analysis I y = ~ [t(b - x) 3 + dx 3 - c(x-t) 3J
:3
of a particular airplane.
The mass moments of inertia of the airplane tan 2¢ = 2 I xy , where I xy = abcdt
about the coordinate X, Y and Z axes through the Iy-I X 4 b+c
center of gravity of the airplane can be expres- 2
I ~_~=Ix sin ¢ + I y cos ¢ + I xy
2
sed as follows: sin 2¢

Ix Zwy2 + Zwz 2 + Z6I x I 2-f' Ix coEf¢ + I y sin 2 ¢ - I xy
Iy Zwx 2 + Zwz 2 + Z6I y sin 2¢
Iz Zwx 2 + Zwy2 + Z6I z
A3.3
::-:-::-:- T_AB_L:.E~l_~c~o~n~t~i~n~u~e~d J Elliptical~~~~~--1

Ring
r"" 1;""
~ y~ ~r:ad~2tld + 2a)
A,A'"
xiW/'+b[1:
x
PI \
L I~
il\ _I
,)v:'
_ I ? ~
dl X
d
'x 0 bd" - aid - 2t)"
12
I y = d(b + a)" - 2a"e - 6ab 2e
tan 20 " Idt-t") Ib"-btl
Ix _ I y
\
I---a--l
-;-= I,~\Y-i
,
)1
Area = .215 a
x = .223 a
2
r- b--->j t For I ~_~ and 1 2- 2 see equa- ~ [f'x~ Circular

tions as given for angle. Fillet
Area - x~y~
V-section -3-
Area A = 2dt~ + at
y = d2t~ + .5t 2a x = 3 x~
A 4:
Ix =
Area - nr 2
Circle
2 -- I = nr 4
- =2
'-E-:~:
~-~
4 y y~
4 '7
1 2- 2 = 5 n r
4: TABLE 2
e~ -~ = r Properties of solids
2
Vol. - nr~ (r = radius)
Ring Area n (r o 2 ri 2) Solid Circ. Cyl.
1~-
M = W (Total wt.)
I~ _~ = n (r o4 - ri4) g
4:
\~. 1 I~ _~ = Mr 2
f ~-~ = V r20 +2 rj2 2"

1 2 -.. = M [r 2
+ (L- r3)]14
Area nr 2
Semi-circle 2" Hollow Circ. Cyl. Vol. = n L(r~2- r2~
\l> I~ _~ = M(r~ 2 + r 22)/2
x~ y= .424r
1
.>. ~ - x
1 Ix = .1098 ~ 1 2 - 2 = M(r~2+ r 2 2 + L2/3)/4
ex
A
= .264 r \.
q=outside
For thin hollow eire. eyl
I~ ~ = Mr 2 •
rea _ n (R 2 _ r2) radius
2 Re~ct.
Prism \.
y =i.. (R 2+ Rr + r) ,approx. y-=2r Vol. - abL, M = W
Semi-circular Ring
3n R+r n ' -~ 2 g2
I
x=.1098(R4-r4)-.283R2~ R + r
(R-r)
\.
I:
~
,.....--y 3
12 I
j 0 ~
..ja~ I
~-~= M (a + b )/12
1 2- 2 = ML2/12
~~3 ~
2
-r----------==~~~~------J
I (
x approx)=.3t(r +R)" when t I" " - ML /3
2
Solid Sphere Vol. = (4 n r")/3
is small r
F------;~~-:---~
Ellipse
Sphere I
about axis
2M r 2/5
=
0
Area - nab
•
I = n- a"b
Ix Thin Hollow Sphere Vol. =4 n
JI 4-
~ :3
(r~3 -r2 " )
x- xE'x = a = radius of sphere
~
rl
2 I about dia. -
2Mr~ 2
3
A3 4 CENTROIDS , C ENTER OF GRAVITY MOMENTS OF INERTIA
TABLE 2 - Continued -- a sin 2a (I-cos a) a)

Ring with Circular
INA
X
r 3 t ~- 4 - a
Section Y M
- = mass
=w per unit volume
x~.x
g of body. TABLE 4
I xx =mn 2Ra2 [If + (5a 2/4)) Centroids of Trapizoidal Areas
y
TABLE 3
2
2 2
I yy = ~ mn Ra (4R 2 + 3a )
2
LLET
b
fl- y +xl--
I-l-
Section Properties of Lines

(t is small in comparison to radius) Ratio a Distance x Distance y
b
,e, = 2 n rt
Circular Arc Area 1.01 0.4992 0.5008
1. 02 0.4984 0.5016
1.03 0.4976 0.5024
I~_~ = n r t
3
1.04 0.4968 0.5032
1.05 0.4960 0.5040
I polar = 2 n r 3 t 1.06 0.4952 0.5048
1.07 0.4944 0.5056
1.08 0.4936 0.5064
ex = .707r 1. 09
1.10
0.4928
0.4920
0.5072
0.5080
E' polar = r 1.11
1.12
0.4912
0.4905
0.5088
0.5095
1.13 0.4898 0.5102
Semi-circle Arc A = n rt 1.14 0.4890 0.5110
- 1.15 0.4883 0.5117
Y = .6366 r 0.5123
~
1.16 0.4877
1.17 0.4870 0.5130
I1.-~ = n r t
3 0.5138
1.18 0.4862
-2- 1.19 0.4855 0.5145
1 ---±-- - 1 1. 20
1. 22
0.4849
0.4835
0.5151
0.5165
1"-2 = .2978 r 3t 1. 24 0.4822 0.5178
1. 26 0.4809 0.5191
1 28 0.4795 0.5205
Quarter-circular Arc Area =n rt 1. 30 0.4782 0.5218
-2- 1. 32 0.4770 0.5230
~'
1. 34 0.4758 0.5242
- 1. 36 0.4746 0.5254
Y = .6366 r 1. 38 0.4733 0.5267
f4--~:
2 1. 40 0.4721 0.5279
I~_~ = n-4-
r t 3
1. 45 0.4693 0.5307
0.5333
1. 50 0.4667
1 1.55 0.4641 0.5359
2 1 1.60 0.4616 0.5384
12 - 2 = .149 r 3 t
,
1. 65 0.4592 0.5408
1. 70 0.4568 0.5432
1. 75 0.4545 0.5455
1. 80 0.4523 0.5477
CIRCULAR ARC 1. 85 0.4502 0.5498
r
1. 90 0.4482 0.5518
NAy 0.5538
Y 1. 95 0.4462
2.00 0.4443 0.5557
2.10 0.4409 0.5591
2.20 0.4375 0.5625
x
2.30 0.4343 0.5657
~
2.40 0.4312 0.5688
2.50 0.4284 0.5716
2 60 0.4259 0.5741
: ·NA 2.70 0.4233 0.5767
Xf - ~
2.80 0.4209 0.5791
I r{MEAN)t y 2.90
3.00
0.4188
0.4168
0.5812
0.5832
-x I I i X 3.20
3 40
0.4128
0.4090
0.5872
0.5910
Y 3.60 0.4060 0.5940
3.80 0.4030 0.5970
Area = art a in Radians 4.00
4.20
0.4000
0.3975
0.6000
0.6025
0.6050
4.40 0.3950
r sin a ,
X= a (Myy =A x = r 2
t sin a) 4.60
4.80
0.3928
0.3908
0.6072
u.t;U1l2
5.00 0.3889 0.6111
5.50 0.3848 0.6152
r3t (a + Si~ 2a) 6.00 0.3810 0.6190
IyY =~ 6 .. 50 o 3778 0.6222
7.00 0.3750 0.6250
7.50 0.3725 0.6275
t (a2 + sin42a _ sina a)
2
INA Y =r 3 8.00
9.00
0.3702
0.3668
0.6298
0.6332
10.00 0.3636 0.6364
y = r (I-COS
a
a) , Mxx = Ay- -- rat (I-cos a)
r3 t
I xx =2 (a _ sin2 2a)
A3.5
Table 4
Item Horizontal Vertical
I~
Weight Arm: x lIoment Arm. y lIoment
No. Name 11'# : Wx : l'y
1 Propeller 180 o in. 0 0 0

2 Engine Group 820 46 37720 0 0
3 Fuselage Group 800 182 145600 4 3200
4 Wing Group 600 158 94800 -18 -10800
5 Hor!. Tail 60 296 17760 8 480
6 Vert. Tail 40 335 13400 26 1040
7 Tail Wheel 50 328 16400 -20 -1000
8 Front Land.Gear 300 115 34500 -30 - 900
9 Pilot 200 165 33000 10 2000
10 Radio 100 240 24000 5 500
I~ Totals 3150 417180 -5480
where Ix, I y , and I z are generally referred to Example Problem 2. Determine the moment of iner-
as the rolling, pitching and yawing moments of tia about the horizontal centroidal axis for the
inertia of the airplane. area shown in Fig. A3.4
w = weight of the items in the airplane Solution. We first find the moment of inertia
x, y and z equal the distances from the about a horizontal reference axis. In this so-
axes thru the center of gravity of the airplane lution, this arbitrary axis has been taken as
and the weights w. The last term in each equa- axis x'x' thru the base as shown. Having this
tion is the summation of the moments of inertia moment of inertia, a transfer to the centroidal
of the various items about their own X, y and Z axis can be made. Table 5 gives the detailed
centroidal axes. calculations for the moment of inertia about
If w is expressed in pounds and the distan- axis x'x'. For simplicity, the cross-section
has been divided into the five parts, namely, A,
ces in inches, the moment of inertia is express- B,e,D, and E.
ed in units of pound-inches squared, which can lex is moment of inertia about centroidal
be converted into slug feet squared by multiply- x axis of the particular part being considered.
ing by 1/32.16 x 144. Distance from axis x'x' to centroidal horizontal
Example Problem 1. Determine the gross weight axis = y = ZAv = 17.97 = 2.91"
center of gravity of the airplane shown in Fig. ZA 6.182
A3.3. The airplane weight has been broken down By parallel axis theorem, we transfer the
into the 10 items or weight groups, With their
individual e.g. locations denoted by the symbol moment of inertia from axis x'x' to centroidal
+. axis xx.
I xx = Ix' x' - Ay 2= 79.47 - 6.18 X 2.91 2 =
Solution. The airplane center of graVity will be 27.2 in·
located with respect to two rectangular axes. In
this example, a vertical axis thru the center-
line of the propeller will be selected as a ref-
Radius of Gyration, Pxx = xx =V27 .2 =2.1" Vl
erence axis for horizontal distances, and the A 6.18
thrust line as a reference axis for vertical dis-
tances. The general expressions to be solved ®
are:-
x- = Zwx = distance to airplane e.g. from
ZW ref. axis S-B
y = ZWy = distance to airplane e.g. from
1iW ref. axis X-X
Table 4 gives the necessary calculations,
whenee
x 417180 = 133.3" aft of t propeller Fig. A3.4 Fig. A3.5
3150 Exampre Problem #3. Determine the moment of in-
y = 5480 -1.74" (below thrust line) ertia of the stringer cross seotion shown in
3150 Fig. A3.5 about the horizontal centroidal axis.
Solution. A horizontal reference axis x'x' is
assumed as shown. The moment of inertia is
Thr~tv--7'''=::;''.. _
first calculated about this axis and then
line transferred to the centroidal axis xx. See
Table 6.
Fig. A3.3
A3.6 CENTROIDS, CENTER OF GRAVITY, MOMl!:NTS OF INERTIA
Portion 3 + 3' (ref. Table 1)

Table 5
y= .375+ 4 x (.0841+ .0725+ .0625) =.•375+
3ft .54
I cx
-<
.... ...
-<
~
-<
_ moment of inertia of
portion about centroida1
= .547"
.~72
-<" x axis. approx. y= .375+§ r1 = .375+ .172=547"

A 0.50 5.5 2.75 15.13 1/12 x .5 x 1 3 •• 04 15.17 n
Al 0.50 5.5 2.75 15.13 • . 04 15.17 lcx= [.1098(.0841+ .0625) + .04x .54-
B 1.00 5.25 5.25 27.50 1/12 x 2 x .5 3 • . 02 27.52 .283 x .0841 x .0625 x .04] = .002375
C 2.12 2.875 6.10 17.56 1/12 x .5 x 4.25 3 = 3.19 20.75 .54
D 0.281 1.00 0.28 0.28 bh 3/36 = .75 4/36 • 0.01 0.29 approx. lcx = .3x .04x .27 8 = .002365 in~
D1 0.281 1.00 0.28 0.28 II • 0.01 0.29
E 1.50 0.375 0.56 0.21 1/12 x 2 x .75 3.0.07 0.28 Problem #4. Determine the moment of inertia of
the flywheel in Fig. A3.5a about axis of rotation.
Sum6.182 17.97 79.47 in 4
The material is aluminum alloy casting (weight =
.1 lb. per cu. inch.)
Table 6
WfJJ
Sec.AA
Portion Area y Ay Ay 2 I cx lxIx' • I cx + Ay 2
1 + I' .06 0 0 o .00281 .00281
2 .0654 .706 .04617 .0326 .00168 .03428

~
3 + 3' .0679 -.547 -.0371 .0203 .000475 .02078 1177/.771 1
Sum .1933 .009 .05787 1"1 I 2" 10"
TU.~~-L
-y = ~Av = .009 = .0465"
~
'i ---.L-
~A .1933 t7l/J7!l I"
-" = .05787 - .1933 x
lXX = lxIx' - Ay 1--3''--1
Fig.A3.5a
(.0465)" = .05745 in~
Solution: The spokes may be treated as slender
Radius of gyration fx-x = round rods and the rim and hub as hollow cyl1nd
ers. (Refer to Table 2)
Detailed explanation of Table 6:- Rim,
Portion 1 oft I' weight W = n(R" - r") x 3x .1
~ bd = (.04 X .75 = n(5"-4") .3=6.48 lb.
3
lcx = 8
) x 2
I = .5W(R"+r")=.5x8.48 (5"+4") =
12 12
173.7 lb. in"
= .00281 in~.
Portion 2 (ref. table 1) W = n(l" - .5") x 3 x .1 = .707#
I = .5x .707 (1 2 + .5 2 ) = 044 lb. In~
y = .375 + 4 R"+Rr+r" = .375 + 4
3n R'l-r 3n Spokes,
(.5"+.5x.54+.54") = .375 -10 331 = .706" Length of spoke = 3"
(.5 + .54) Weight of spoke = 3 x n.25" x .1 =
.588#
By approx. method see Table 1. I of one spoke = WL" /12 + Wr~ .588 x
y = .375 + § r 1 = .375 + §x (.52) = .706" 3"/12 + .588 x 2.5" =
n n
4.10 lb. in"
I for 4 spokes = 4x 4.10 = 16.40
lcx= .1098x (R"+r") t (R+r) - .283 R"+r"t
R + r Total I of wheel = 16.40 + .44 + 173.7 = 190.54
lb. in"
.:: .1098(.2916+ .25) x .04x1.04- I = 190.54/32.16 x 144 = .0411 slug ft~
.283 x .2916 x .25 x .04 Example Problem #5. Moment of Inertia of an
1.04 Airplane.
= .002475- .000793= .001682 in~. To calculate the moment of inertia of an
_3 airplane about the coordinate axes through the
Approx. lcx = .3tr13 = .3x .04x.52 = .00l688in 6 •
gross weight center of gravity, a break down of
the airplane weight and its distribution is
A3.7
necessary, which is available in the weight and axes, the product of inertia of the weight about
balance estimate of the airplane. the reference axes is necessary. Column (14)
Table 6a shows the complete calculation of gives the values about the reference axes. To
the moments of inertia of an airplaneo This transfer the product of inertia to the cog. axes
of the airplane, we make use of the parallel
table is reproduced from N.AoO.A. Technical note
#575, "Estimation of moments of inertia of air- axis theorem. Thus
planes from Design Data." zwxz c . g . =ZWXZ(Ref. axes) -Zwxz =48,857,589-
Explanation of Table 5325.3x1l5.9x77.8=839,253 lb. in.2
Fig. A3.5b shows the reference of planes To reduce all values to slug ft~ mUltiply 1
and axes which were selected. After the moments 32.17
1
of inertia have been determined relative to x-
these axes the values &bout parallel axes 144
through the center of gravity of the airplane
are found by use of the parallel axis theorem. Hence Ix = 3061, I y = 6680, I z = 9096, I xz = 181
Having the inertia properties about the co-
Column (1) of Table 6a gives breakdown of air- ordinate c.g. axes, the moments of inertia bout
plane units or items.
the principal axes are determined in a manner as
Column (2) gives the weight of each itemo explained for areas. (See A3.13).
Columns (3), (4) and (5) give the distance of
the c.g. of the items from the references planes The angle 0 between the X and Z axes and the
or axes. principal axes is given by,
Columns (6) and (7) give the first moments of tan 2 0 = 2I x z = 2 x 181 =.05998 hence lO =
the item weights about the Y' and X' reference I z - Ix 9096-3061
axes.
1 0 43"
Columns (8), (9) and (10) give the moment of
inertia of the item weights about the reference
axes.
Columns (11), (12) and (13) give the moments of
inertia of each item about its own centroidal
axis parallel to the reference axes. Such items
as the fuselage skeleton, wing panels and engine
, \
\
have relatively large values for their centroid- '- \
al moments of inertia. \, \
The last values in Columns (3) and (5) give "g '<II
\
\
\
o
<II
the distances from the reference planes to the <II "

<II
" <II
<II
o
center of gravity of the airplane. " <II \ >--
- <II
It-4 " j---
-
N
- , ' Q)
"
<II "
"
-
a
<II .... "
<II
x c.g. = Zwx >< '"

,...t \><Il:aPo
" ...
<II
"',n
..
-=617.024 (col. 6) =115.9 in. 0.

\
Zw 5325.3 (col. 2) \
\
Zc og. = Zwz= 414.848 (col. 7) = 7708 in.
ZW 5325.3
The last values in columns (8), (9) and
(10) were obtained by use of the parallel axis
theorem, as follows:-
zwx 2 about c.g. of airplane =97,891,595-5325.3
x 115.9 2 = 26,691,5950
Zwz 2= 33,252,035 - 5325 .3 x 77.8 2 = 992, 035
The third from the last value in columns
(11), (12) and (13) give the moments of inertia
of the airplane about the x, y and z axes
through the c.g o of airplane. The values are
obtained as follows:
ly = Zwx 2 + Zwz 2 + Z~Iy = 26,691,595 + 999,035 +
3,120,384 = 30,804,014 lb. in~
Ix = Zwy2 + ZW\32 + Z~Ix = 10,287,522 + 992,023 +
2,899,470 = 14,179,027 Ib o in.2
I z = Zwy 2 + Zwz" + Z~Iz = 10,287,522 + 26,691,595 +
5,157,186 = 42,136,303 lb. in. 2 (REF. N.A.C.A. TECH. NOTE No. 575)
In order to determine the principal inertia Fig. A3.5b
A3.8 CENTROIDS, C ENTER OF GRAVITY, MOMENTS OF INERTIA
Table 6a
CO~PUTATIONS OF MOMENTS OF INERTIA
1. a. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14.
Item Weight x y I: WX WI: wx 2 wy2 w.. a t:.I z WXI:
Center sect10n 108.8 lOa 57 11,098 6,a02 1,131,955 353,491 261,229 261,229 632,563
nose assel!lbly
Center ssction a04.6 lal 57 24,757 11,66a a,995,549 664,745 491,245 491,245 1,411,126
beam, etc.
Center section 84.2 148 55 12,462 4,631 1,844,317 254,705 202,164 33,680 235,844 685,388
ribs, etc.
Flap 2a.0 180 53 3,960 1,166 712,800 61,798 48,598 48,598 209,880
Outer panel nose 104.6 105 166 65 10,983 6,799 1,153,215 2,545,546 441,935 184,514 184,514 713,895
Outer panel beam 155.5 120 156 65 18,672 10,114 2,240,640 3,786,682 657,410 274,478 274,478 1,213,680
Outer panel ribs 89.8 139 156 64 12,482 5,747 1,735,026 2,185,373 367,821 158,407 17,601 176,008 798,861
Ailerons 31.4 172 156 62 5,401 1,947 928,938 764,150 120,702 55,390 55,390 334,850
Horizontal tail 87.1 367 96.7 31,966 8,423 11,731,412 814,463 176,378 176,378 3,091,083
Vertical tail 31.4 352 - 125 11,053 3,925 3,890,586 490,625 110,200 10,174 31 1,381,600
Fueelage skele- 314.0 176 81 55,264 25,434 9,726,464 2,060,154 69,394 1,576,594 1,570,000 4,476,384
ton
Engine mount 40.5 60 80 2,430 3,240 145,800 259,200 5,184 5,184 5,184 194,400
Turtleback 48.5 254 80 12,319 3,880 3,129,026 310,400 15,181 57,861 56,648 985,520
( fairing)
Firewall 11.0 70 80 770 880 53,900 70,400 2,200 1,100 1,100 61,600
Steps 2.0 170 20 70 340 140 57,800 800 9,800 23,800
~.A.C.A. cowling 70.0 50 80 3,500 5,600 175,000 448,000 16,940 15,470 15,470 280,000
Cabin and 66.5 146 - 108 9,709 7,182 1,417,514 775,656 2,394 106,400 108,794 1,048,572
Windshield
Foot troughs 2.0 77 5 68 154 136 11,858 50 ·9,248 10,472
Floor, rear 9.5 al0 66 1,995 6a7 418,950 41,382 608 1,368 1,976 131,670
Wing fillets 18.5 142 20 58 2,627 1,073 373,034 7,400 62,234 18,944 18,944 152,366
Bottom cowling 27.0 140 11 75 3,780 2,025 529,200 3,267 151,875 24,300 24,300 283,500
and side frames
Arresting door 1.3 284 63 369 82 104,853 5,160 5 5 23,260
Tail-woeel pan, 4.0 365 84 1,460 336 532,900 26,224 100 100 122,640
etc.
Side doors 17.0 143 IS 82 2,431 1,394 347,633 5,508 114,308 1,088 28,288 27,200 199,342
Baggage door 1.8 165 63 297 113 49,005 7,144 720 720 18,711
Fabric and dope 13.0 254 80 3,302 1,040 838,708 83,200 4,394 15,509 15,509 264,160
Tail cone 7.5 385 91 2,888 683 1,111,688 62,108 120 120 262,763
Cowling, sta- 12.0 110 95 1,320 1,140 145,200 108,300 1,200 1,200 125,400
tione 1-2
Chassis (re- 232.4 115 54 51 26,726 11,852 3,073,490 677,678 604,472 23,240 23,240 1,363,026
tracted)
Retracting mech- 28.6 110 25 67 3,146 1,916 346,080 17,875 128,385 210,782
anism
Wheels, etc. 91.0 141 54 56 12,831 5,096 1,809,171 265,356 285,376 718,538
Tail wheel 26.0 360 74 9,360 1,924 3,369,600 142,376 26 26 692,640
(Table 6a Fra. N.A.C.A. Tech.

A3.9
The principal moments of inertia are given by will be given the symbol I xy , hence
following equation.
2 2 I xy = I xydA - - - - - - - - - - -(1)
I xp = I:z cos 0 + I z sin 0 - I xz sin 2 0. (see The unit, like that of moment of inertia, is ex-
Art. A3.11) pressed as inches or feet to the 4th power.
I yp = I y Since x and y may be either positive or negative,
the term I xy may be zero or either positive or
I z = Ix sin2 0 + I z cos 2 0 + I xz sin 2 0 negative.
substituting Product of Inertia of a Solid. The product of
I xp = 3061 x (0.9996)2 + 9096 x (0.0300)2 -181x inertia of a solid is the sum of the products
.0599 = 3056 obtained by multiplying the weight of each small
portion in which it may be assumed to be divided
I zp =3061x (0.0300)2 + 9096 x (0.9996)2 +181x by the product of its distances from two of the
.0599 = 9102 three coordinate planes through a given point.
I yp = 6680 Thus With respect to planes X and Y
I xy = I xy dW
A3.7b Problems
I xz = I xz dW
I yz = I yz dW
A3.9 Product of Inertia for Axes of Symmetry.
T:-
• J...
If an area is symmetrical about two rec-
tangular axes, the product of inertia about these
'~
1
axes is zero. This follows from the fact that
J..R symmetrical axes are centroidal x and y axes.
1 10 £.4-
8 1"
If an area is symmetrical about only one of
~ k---- It ---+l T ~ two rectangular axes, the product of inertia,
IXYdA, is zero because for each product xydA for
Fig. A3.6 Fig. A3. 7
an element on one side of the axis of symmetry,
(1) Determine the moment of inertia about there is an equal product of opposite sign for
the horizontal centroidal axis for the beam
section shown in Fig. A3.6. the corresponding element dA on the opposite side
of the axis, thus making the expression lydA
(2) For the section as shown in Fig. A3.7 equal to zero.
calculate the moment of inertia about the cen-
troidal Z and X axes. A3.l0 Parallel Axis Theorem
The theorem states that, "the product of

inertia of an area With respect to any pair of
co-planar rectangular axes is equal to the prod-
uct of inertia of the area With respect to a pair
of parallel centroidal axes plus the product of
the area and the distances of the centroid of the
total area from the given pair.of axes". Or, ex-
pressed as an equation,
I xy ~ IXY + AXY- - - - - - - - - -(2)
(3) Determine the moment of inertia about This equation is readily derivable by re-
the horizontal centroidal axis for the section ferring to Fig. A3.10. YY and XX are centroidal
shown in Fig. A3.8. axes for a given area. YY and XX are parallel
(4) In the beam cross-section of Fig. A3.9 axes passing through point O.
assume that the four corner members are the only
The product of inertia about axes YY and XX
effective material. Calculate the centroidal
moments of inertia about the vertical and hori- is I xy = I(x + x)(y ~ y) dA
zontal axes. = IxydA + xy I dA + x I y dA + Y I xdA
The last two integrals are each equal to
A3.,8 Product of Inertia zero, since lydA and IXdA refer to centroidal
In various engineering problems, particu- axes. Hence, I xy = IxydA + xy IdA, which can be
written in the form of equation (2).
larly those involving the calculation of the
moments of inertia of unsymmetrical sections,
the expression I xy dA is used. This expression
is referred to as the prOduct of inertia of the
area With respect to the rectangular axes x and
y. The term, product of inertia of an area,
A3.l0 CENTROIDS, CENTER OF GRAVITY, MOMENTS OF INERTIA
Y Y By adding equations (3) and (4), we obtain I x1 +

IY1 = Ix + I y - - - - - - - - - - - - - - - - -(5)
or the sum of the moments of inertia of an area
with respect to all pairs of rectangular axes,
thru a common point of intersection, is constant.
A3.l2 Location of Axes for which Product of Inertia
is Zero.
In Fig. A3.11
I X1Y1 =jx y dA=j(xcos0+ysin0)(ycos0-
x siJ 0)dA
= (cos" 0 - s in 2 0 j xydA + cos 0 sin 0 j
Y Fig. A3.10 (y"-x") dA
= I xy cos 20+ ~ (Ix - I y ) sin 2 13
A3.ll Moments of Inertia with Respect of Inclined
Axes 2
Therefore, I X1Y1 is zero when
unsymmetrical beam sections are very com-
mon in aircraft structure, because the airfoil tan 2 13 = 2I xy - - - - - - - - - - - -(6)
shape is generally unsymmetrical. Thus, the Iy_I X
general procedure with such sections is to first
find the moment of inertia about some set of A3.l3 Principal Axes.
rectangular axes and then transfer to other in-
clined axes. Thus, in Fig. A3.11 the moment of In problems involving unsymmetrical bending,
inertia of the area with respect to axis X1X1 is, the moment of an area is frequently used with re-
spect to a certain axis called the principal
Ix 1 =jy 1 "dA= j(ycos0-xsin0)"dA axis. A principal axis of an area is an axis
= cos"0j y"dA+sin"0jx 2 dA-2Sin0 cos 0 about which the moment of inertia of the area is
either greater or less than for any other axis
j xydA passing thru the centroid of the area.
= Ix cos" 0 + I y sin" 0 - 2 I xy sin 0 cos 0 (3) Axes for Which the product of inertia is
zero are principal axes.
and likewise in a similar manner, the following Since the product of inertia is zero about
equation can be derived: symmetrical axes, it follows that symmetrical
I Y1 =I x sin"0 + Iycos"0 + 2 I XY sin 0 cos axes are principal axes.
The angle between a set of rectangular
o- - - - - -(4) centroidal axes and the principal axes is given
by equation (6).
Example Problem 4.
Determine the moment of inertia of the ang-
le as shown in Fig. A3.12 about the principal
axes passing through the centroid. Solution:
x--l----------:=-I--==-------L---..~-x
Reference axes X and Yare assumed as shown in
Fig. A3.12 and the moment of inertia is first
calculated about these axes. Table 8 gives the
calculations. The angle is divided into the
two portions (1) and (2).
Fig. A3.11
Table R
I ICy
Part Area cx
Ix Iy
A x y Ay Ax Ay 2 Ax 2 Axy
.l.xl 5(1)3-= 0019 11 3
1 .375 .75 .125 .0469 .281 .0058 .211 .0351 12 • (4) • 12x4Xl, 5 =.070 .0077 .281
2 .500 .125 1,25 .625 .0625 .7800 .0078 .0781 1 1 3 .9470 .010
12 x4 x2= .167 i2x2X.253 •• 0026
.875 .6719 .3435 .7858 .2188 .1132 .9547 .291
A3.1l
~Icx and Icy = moment of inertia of each portion section has been broken down into 16 stringers
about their own X and Y centroid- as listed in column 1. For the top surface, a
al axes. width of 30 thicknesses of the .032 skin 1s as-
sumed to act with the stringers and a width of
Location of centroidal axes;-
25 thicknesses of the .04 skin (see Col. 3). On
y = ZAy .6719 = .767" the lower surface, the skin half way to adjacent
li .875 stringers is assumed acting with each stringer,
or the entire skin is effective. Column 4 gives
x = ZAx .3435 = .392" the combined area of each stringer wlit and is
ZA .875
considered as concentrated at the centroid of
Transfer moment of inertia and product of iner- the stringer and effective skin. All distances,
tia from reference X and Y axes to parallel x and y, columns 5 and 8, have been scaled from
centroidal axes;- a large drawing.
Ix = Ix - A~?' = .955 - .875 x .767 2 = .440 y
Iy = I y - Ai. 2 = .291 - .875 x .392 2 = .157
I XY = ZAxy - AY:.y = .1132 - .875 x .767 x .392 = .032 Skin
.040 Skin r ef.
- .150
""2 3
Calculate angle between centroidal X and Y axes
and principal axes through centroidal as fol- --,---+_-+-~XRef.
lows;- ~~--.---t~~.- x
o
tan 2 0 = ~ = 2(-.150 = -.30 = 1.06 L9 11
-r-
12 13 16
I y - Ix .157-.440 -.283 .J
ca
2 0 = 46 0 - 40' 0 = 23 0 - 20'
Calculate moments of inertia about centroidal
Fig. A3.13
"/1-
Yp Y y
principal axes as follows;-
I xp = Ix cos 2 0 + I y sin 2 0 - 2I xy sin 0 cos 0 Table 9
11
= .44x .918 2 + .157x .3965 2 -2(-.150)x 1 2 3 4 5 6 7 8 9 10
.3965 x .918 =.504 in~ m

Area
~
m .~
I yp = Ix sin 2 0 + I y cos 2 0 + 2I xy sin 0 cos 0 ..." ..."'''
.~ ...u·.." <
" m m...
~
.-< y Ay Ay2 x Ax Ax 2 lxy
., '""" o" "

... m -Axy
~
b<
= .44x .3965+ .157x .918 2 +2(-.150) x 0 ... ~
.3965 x .918= .092 in. 4 '"1 .11 .03 '" .14 E-«
4.00 0.560 2.240 -33.15 -4.641 153.85 -18.56

2 .11 .03 .14 6.05 0.847 5.124 -29.28 -4.099 120.02 -24.80
3 .30 .OR .38 7.00 2.660 18.620 -24.85 -9.443 234.66 -66.10
4 .13 .04 .17 7.37 1.253 9.234 -21.18 -3.600 76.25 -26.53
5 .13 .04 .17 7.55 1.288 9.687 -16.60 -2.822 46.84 -21. 31
6 .13 .04 .17 7.50 1.275 9.563 -12.60 -2.142 26.99 -16.06
7 .13 .04 .17 7.30 1. 241 9.059 - 8.60 -1. 462 12.57 -10.67
8 .13 .04 .17 6.90 1.171 8.080 - 4.00 -0.680 2.72 - 4.69
9 .24 .05 .29 6.50 1.885 12.252 - 0.35 -0.100 0.04 -00.65
10 .07 .10 .17 -3.30 -0.561 1. 851 -33.25 -5.653 187.96 18.65
11 .07 .10 .17 -4.90 -0.833 4.082 -29.28 -4.978 145.76 24.39
12 .13 .15 .28 -5.95 -1. 666 9.913 -24.85 -6.958 172.90 41.40
13 .11 .20 .31 -7.40 -2.294 16.976 -18.70 -5.797 108.40 42.90
14 .11 .20 .31 -8.13 -2.520 20.487 -12.42 -2.850 47.82 31.30
15 .11 .20 .31 -8.62 -2.672 22.432 - 6.10 -1.891 11. 54 16.30
16 .24 .11 .35 -8.87 -3.094 27 •.444 -00.35 -0.122 0.04 1. 08
Totals 3.70 -1. 465 187.04 -58.2381348.36 -13.35
,(
Location of centroidal axes with respect to ref.
axes, -
y = ZAy = - 1.465 = - .396"
li 3.70
Example Problem 5. x=ZAx=- 58.238 15.74"
Fig. A3.13 shows a typical distributed ZA 3.70
flange - 2 cell - wing beam section. The upper
and lower surface is stiffened by Z and bulb I x =187.04-3.70x .396 2 =186.5 in.4
angle sections. Determine the moment of inertia I y = 1348.36 - 3. 70x 15.74 2 = .431. 7 in.4
of the section about the principal axes.
Solution; I XY = -13.35 - (3.70 x .396 x 15.74) = 36.41in. 4
The properties of the cross-section depend
tan 2 0 = ~ = 2 (-36.41) = -.29696
upon the effective material which can develop
resist1ng axial stresses. The question of ef- Iy-IX 431.7-186.5
fective material is taken up in later chapter. 20=16 0 -32.5', 0=8 0 -16.25'
Table 9 shows the calculations for the moment of 2 2
inertia about the assumed rectangular reference Ixp=Yx cos 0+l y sin 0-2 I XY sin0cos0
= 186.46 x .9896 2 + 431. 7 x .1438 2 -2
axes XX and YY (see Fig. A3.13). The cross-
t36.41x .9896x (-.1438)J = 181.2 in. 4
A3.12 CENTROIDS, CENTER OF GRAVITY, MOMENTS OF INERTIA
I yp =Ix sin 2 0 + I y cOS 2 0 + 2 I xy sin 0 A3.14 Section Properties of Typical Aircraft Structural
Sections.
cos 0 =186.46 x .14382 2 + 431. 7 x .9896 2 + Table A3.10 through A3.15 and Chart A3.1
2 -36.41x,9896x(-.1438) =437in. 4 give the section properties of a few structural
shapes common to aircraft. Use of these tables
will be made in later chapters of this book.
y
y
S=1
Table A3.10 Table A3.11

Properties of Properties of
Zee Sections Channel Sections
NOMINAL DIMENSIONS SECTION ELEMENTS NOMINAL DIMENSIONS SECTION ELEMENTS

AREA
Ixx I Iyy P xx I p yy A I BIT R
AREA
e---- y-
Ixx
I Iyy f' xx 1 f'yy
Inches Square Inches 4 Inches 1-"::"--.JL"":::",......,L"":"'....l.""'::':"-., Square I-::-In'-C-h-+-':;:':I:"'n-Chl.e-s";'4!...:...--+L::I:"'n..Jch~e~s-'-'---1
Inches Inches Inches
.050 .094 .0852 .0035 .0068 .203 .283

.625
.063 .125 .1073 .0042 .0082 .197 .277
.500
.050 .094 .0977 .0042 .0122 .206 .353
.750 .063 .125 .1231 .0049 .0148 .200 .346
.050 .094 .0916 .0060 .0069 .255 .274
.625
.063 .125 .115 .0071 .0083 .248 .268 .875
.625 .050 .094 .104 .0070 .0122 .259 .342
.750 .375
.063 .125 .131 .0084 .0148 .253 .336 .050 .094 .0916 . 146 .0020 .0110 .148 .347
1. 000 .063 .125 .163 .0109 .0367 .259 .475 .500 .063 .115 .148 .0025 .0132 .144 .339
.050 .094 .0977 .0090 .0068 .304 .264
~ =~~'~1'01613r":12:51_'lll3!1~§'11916§§'!001417~c'10115i8~c'11190~!~'3i4i7~
.625
.063 .125 .123 .0110 .0082 .298 .259 .625 .078 .158.201.0055.0183 .187 .341
.750 .050 .094 .110 .0106 .0122 .310 . 3~3
.750 .875 .050 .094 .117 .245 .0065 .0151 .237 .360
.063 .125 .139 .0128 .0148 .304 .326 . 750 ~ .147.247.0079.0184.233.354
1. 000 .063 .125 .170 .0165 .0367 .312 .464 .078 .177.253.0094.0214.230.348
.625
.050 .094 .104 .0130 .0068 .353 .257 ~ .125 .216 .362 .0213 .0276 .314 .358
.063 .125 .131 .0158 .0083 .347 .251 1. 000 . 125 .328.379.0304.0378.304.339
.050 .094 .117 .0151 .0122 .360 .323 ~ .123.140.0026.0181.144.384
.750 .063 .125 .147 .0184 .0148 .354 .317 .500
.078 .148.145.0029.0211.141.377
.875 .078 .125 .177 .0214 .0175 .348 .314 ~ . 139 . 187 .0049 .0216 . 187 .394
.050 .094 .129 .0172 .0197 .365 .391 .625 .078 . 168 . 193 .0058 .0252 . 186 .388
~
.875 .163 .0210 .0241 .359 .385 ~ . 155 .229 .0083 .0250 .231 .402
.078 .197 .0245 .0287 .353 .382 ~ . 187 . 242 .0098 .0294 .229 .393
.125
1. 000 ~
.078
.178
.216
.0236
.0276
.0367
.0439
.364
.358
.454
.451
1. 000 .750 .094 . 125 .220 .248 .0113 .0335 .226 .390
~ . 170 .289 .0128 .0285 .275 .409
.050 .094 .110 .0176 .0068 .400 .249 ~ .206.294.0152.0335.271.403
.625
.063 .125 .139 .0216 .0083 .394 .244 .875 .094 .244.300.0176.0383.269.397
.0'" .094 .123 .0205 .0121 .408 .315 ~ .265.460.0417.0460.397.417
~
.750 .155 .0251 .0147 .403 .308 1. 250 . 125 . 4 0 6 . 4 7 8 . 0 6 0 7 . 0 6 4 3 . 387 .398
1. 000 f----t--'-;.0",7",,8-1 .187 .0294 .0175 .396 .307
~ .170 .0285 .0242 .409 .377 1.125 f-_._50_0---+1_:'"'~~~",,~-1 .125 : :~~ : : :;~
: : ~~;;
: ~~:: ::~ :~~
.875
~
.094
.125 .206
.244
.0335
.0383
.0287
.0332
.403
.397
.373
.369 .625
~ .147 .178 .0051.0285.186.441
.078 . 177 . 184 .0060 .0334 . 184 .434
.186 .0320 .0367 .415 .444 ~ . 197 .232 .0102 .0388 .22R .444
1. 000
.226 .0377 .0439 .408 .441 .750 .094 .232.237.0117.0444.225.438
.625 .125 .147 .0285 .0083 .441 .237 ~ . 181 .278 .0135 .0379 .274 .458
.094 .142 .0304 .0197 .464 .374 ~ .216.283.0158.0441.271.452
~
~:~~~"'~-1 ;~~ ~~~ ~~~; ~~~~ ~~~ :~;
.178 .0374 .0241 .458 .368
.875
1.125 ~
.094
.216
.256
.0441
.0507
.0287
.0332
.452
.445
.364
.361
1.125 f-._8_7_5f-1 . 125 : : : : : :
.125 ~ .236.336.0232.0495.314.458
1. 125 ~ .210 .0467 .0530 .472 ~~ 1. 000 ~ .279.341.0269.0569.310.452
~
.094
.255
.303
.0548
.0632
.0637
.0744
.464
.457
.500
.496
.125 .359.352.0331.0695.303.440
~ .350.508.0661.0757.435.466
.750 ~
.078
.170
.206
.0421
.0497
.0148
.0175
.497
.491
.295
.291
1. 375 .125 .453.519.0832.0930.428.453
.625 .094 . 2 2 0 . 1 8 1 . 0 0 7 1 . 0 4 9 3 . 179 .473
~ .202 .0532 .0366 .514 .428
~ ~~'j2~67~t~'f27~8S~'~01~9~0~~·t06~5~0¢~·t2~67S~·~4~9i3~
1. 250 1. 000 .125
.094 .291 .0729 .0511 .501 .419 .875 . 125 .344.288.0234.0796.261.481
1. 250 ~
.094
.294
.338
.0765
.0886
.0887
.1039
.519
.512
.558
.555
1,250 .078
I .125 .265 .377 .0334 .0698 .355 .513
1. 125 ~ .314.383.0389.0807.352.507
.750 .063 .178 .0562 .0148 .543 .288
.125 .406.393.0484.0994.345.495
~ .210 .0662 .0367 .562 .418
1. 500 .125 .500.560.1106 .1292 .470 .508
1. 375 1. 000 ~
.094
.125
.255
.303
.0794
.0911
.0439
.0511
.558
.549
.415
.411 .625
~ .197 .169 .0063 .0540 .179 .524
.094 .232 .175 .0073 .0621 .177 .517
~~~
.294 .0951 .0887 .569 .549
1. 250 I : ~ .236.263.0169.0705.268.547
.453 .1365 .1308 .549 .537
~ t~'j2~7!9=~.~26g8~~.~0~19~8t:1=~.+0~81N84:~.~2~66~~.
5~4~3~
.750 .078 .0765 .875
.226 .0175 .582 .278
1. 375 1-_ _-+-'-.'"'12"'5, .125 .359 .278 .0242 .1002 .259 .528
1. 500 1. 000 ~
.094 .125
.265
.314
.0961
.1116
.0439
.0511
.602
.596
.407
.403
~ .326.371.0403.1010.352.557
1.125 ~ . 4 2 2 . 3 8 1 . 0 5 0 4 . 1246 .346 .544
1. 250 .094 .361 .1349 .1040 .611 .536
.156 . 5 1 3 . 3 9 0 . 0 5 9 1 . 1451 .339 .532
.750 .078 .245 .1104 .0175 .671 .267
1. 625 .156 .668 .612 .1710 .2035 .506 .55~
1. 750 1. 000 .078 .125 .284 .1377 .0439 .696 .393
~ .244 .168 .0076 .0768 .177 .561
1. 250 .094 .385 .1925 .1039 .707 .520 .625 .125 .313 .178 .0090.0939.170.548
.750 .063 .218 .1272 .014 .7 5 .261
1.500 ~ .125 .375 .269 .0249 .1236 .258 .574
2.000 1. 000 .094 .361 .2194 .0511 .779 .376 .875
.125 .156 . 4 5 5 . 2 7 8 . 0 2 8 9 . 1 4 8 5 . 267 .572
1. 250 .094 .408 .2621 .1040 .801 .505
1.125 .094 .338.360.0414.1233.350.604
1. 500 .125 .594 .3852 .2346 .806 .629
.750 .063 .249 .2172 .0148 .9:14 .244
2.500
.875 ~
.094 .125
.323
.385
.2890
.3383
.0288
.0333
.945
.938
.298
.294
1. 000 ~
.125
.408
.531
.3723
.4707
.0512
.0636
.955
.941
.354
.346
A3 13
r-x-f
&
Table A3. 12 SoT
Ill-
Properties of Extruded Aluminum T~
J1 ~'-,- - ~-~-r
W '
Alloy Equal Leg Angles. (Ref. 1938 Table A3.14
Alcoa Handbook) B Properties of
Unequal Angles
~l
I.r-- R
Dimensions Area Axis XX or YY Axis ZZ x
W t R sq. in. I jO d I f> II V ---L 1
5/8
3/4
3/4
3/32
1/16
3/32
1/8
1/8
1/8
.111
.089
.132
0.004
0.004
0.183
0.220
0.187
0.199
.0015
.0018
0.117
0.142
45°
45° I y
I A~J-
0.006 0.219 0.214 .0026 0.141 45°
3/4 1/8 1/8 .171 0.008 0.217 0.227 .0034 0.141 45° 0 NOMINAL DIMENSIONS SECTION ELEMENTS
1 1/16 1/16 .122 0.012 0.311 0.271 .0048 0.199 45° z AREA
1 3/32 1/8 .178 0.016 0.301 0.276 .0066 0.193 45° ~
u
A B T R -
Square
Y x Ixx Iyy ,0 xx ,oyy
1 1/8 1/8 .234 0.021 0.298 0.290 .0085 0.191 45° Inches Inch Inches 4 Inch
1 3/16 1/8 . 339 0.029 0.293 0.314 .0124 0.192 45° '"
OJ Inches
1-1/4 3/32 3/32 .230 0.033 0.38 0.34 .014 0.24 45° I .625 .500 .050 .063 .0535 .126 .187 .00Il .0020 .146 .193
1-1/4
1-1/4
1/8
3/16
3/16
3/16
.30
.43
0.042
0.059
0.37
0.37
0.35
0.37
.017
.025
0.24
0.24
45°
45°
c+3 .750
.500
.625
.063
.063 .063
.0739
.0818
.120
.162
.243
.222
.0014
.0027
.0041
.0044
.139
.181
.235
.232
.625 .063 .0922 .150 .269 .0029 .0069 .177 .274
1-1/4 1/4 3/16 .56 0.074 0.36 0.40 .032 0.24 45° ~
1-1/2 3/32 3/16 .28 0.058 0.46 0.40 .024 0.30 45° ~6
.875
.750 ~
.125
.125 .100
.184
.190
.2I1
.250
.272
.0049
.0082
.0073
.0123
.221
.2I1
.270
.258
1-1/2 1/8 3/16 .36 0.074 0.45 0.41 .031 0.29
1-1/2 3/16
45° ~ .625 ~
.125
.100 .140
.160
.320 .0029
.0050
.0100
.0170
.170 .31b
.304
3/16 .53 0.107 0.45 0.44 .044 0.29 45° ~ .184 .344 .165
1-1/2 1/4 3/16 .69 0.135 0.44 0.46 .057 0.29 45° ~ I. 000 .750 ~
.125
.125
.108
.200
.178
.199
.299
.322
.0051
.0086
.0106
.0182
.217
.208
.313
.302
1-3/4 3/32 3/32 .32 0.096 0.55 0.47 .039 0.35 45° f-jf-
1-3/4 1/8 3/16 .42 0.121 0.53 0.47 .050 0.34 45° 't2 .875 ~
.094
. I16
.167
.220.
.231
.281
.292
.0079
.0IlO
.01Il
.0155
.262
.257
.310
.305
1-3/4 3/16 3/16 .62 0.174 0.53 0.50 .072 0.34 45° ~ ~ .124 .159 .403 .0054 .0197 .209 .400
.0276
1-3/4 1/4 3/16 .81 0.223 0.52 0.52 .093 0.34 45° ~
.750
~ .179 .170 .416 .0074 .203 .393
2 1/8 1/4 .49 0.18 0.61 0.53 .08 0.40 45° .4- I. 250
.125 .125 .231 .181 .427 · 0092 · 0346 .200 .387
2 3/16 1/4 .72 0.27 0.61 0.56 .11 0.39 45° 4~ I. 000
~
~
.139
.202
.239
.250
.361
.373
· 0124
.0171
· 0217
.0306
.298
.291
.394
.389
2 1/4 1/4 .94 0.34 0.60 0.58 .14 0.39 45° 18 .125 .262 .260 .384 .0215 · 0383 .286 .382
2 5/16 1/4 1.16 0.41 0.60 0.61 .17 0.39 45° ~ .750 ~ .204 .156 .522 .0077 .0463 .194 .476
.125 .264 .167 .534 .0096 .0584 .191 .465
-¥.J-. .228 .0512
4~ ~ .228 .472 .0182 .282 .474
I. 500 I. 000
~ .156 .295 .239 .485 .0226 .0647 .276 .468
~ .156 .361 .249 .495 · 0265 .0772 .271 .463
~ ~ .251 .310 .433 · 0347 .0550 .372 .468
.462
1} I. 250
~
.156
.327
.400
.321
.331
.444
.455
.0436
.0518
· 0698
.0834
.365
.360 .457
.0998
~ I. 000 .125 .327 .222 .591 .0236 .269 .553
~ I. 750
I. 250 .125 .358 .300 .545 .0465 .1078 .360 .549
~
30 I. 500
~
.156
.156 .389
.478
.383
.393
.506
.517
.0776
.0928
· I144
.1377
.447
.441
.542
.537
-----'--
I
Y
1=:U- A
~--r
-T
T
Table A3.15
l~
X X
Properties of
Table A3.13 ~S I-Sections
L--I--x
,..,...--R
Properties of
Tee-Sections _J--~
r-- 1 T R..L
----j""TT
~B/ll---<l B ---.j----r;
g NOMINAL DIMENSIONS
AREA
SECTION ELEMENTS 0
Z
NOMINAL DIMENSIONS
AREA
SECTION ELEMENTS
~ A B T R f------ Y Ixx Iyy t"xx t"yy ~ A B T R - Ixx Iyy Pxx f'yy

U Square U Square
Inches 4 Inches 4
'"
OJ
Inches
Inches
Inch Inches
'"
OJ
Inches
Inches
Inches
c--+-2 I. 250
.625
.750
.050
.063
.094
.125
.094
.127
.121
.160
.0028
· 0055
.0077
.0098
.173
.209
.287
.276 ~
~ .750
I. 250 ~
.078 .125
.207
.250
.0206
.0239
.0193
.0235
.316
.310
.306
.307
.750 .094 .106 .148 .0046 .0103 .209 .312 I. 500 .078 .289 .0284 .04I1 .314 .378
~ I. 000
· 050
.063 .150 .228 .0131 .296 .293
3
m .m · u, .366 .336
r---t- I. 375 I. 000 .078 .183 .235 .0156
· 0129
.0158 .292 .294 ~ I. 375
~ .125 .329 .0425 .0375 .359 .338
~ .125
.216 .242 .0188 .290 .295 f----6
.875
~ I. 000 .094
.094 .240 .327
· 0182
.0345 .0168 .379 .280
I. 625
I. 375
.094
.094
.376
.341
.0497
.0580
.0626
.0375
.364 .408
7 I. 250 ~ .413 .332
r--4--- .750
.750
.063
.094
.141
.204
.141
.159
.0052
.0080
· 0169
.0244
.192
.198
.346
.346 r--+-
9
I. 000 I. 625
I. 750
.094
.094
.125 .388
.4I1
· 0676
.0725
.0626
.0786
.418
.420
.402
.437
---&- I. 000 · 078 .193 .225 · 0160 .0200 .288 .327
~
I. 375 .094 .364 .0970 .0375 .516 .321
--IT- I. 500 I. 000 .094 .125 .228 .232 .0186 .0245 .286 .328
~
I. 625 .4I1 .0626 .524 .390
--rr-
--rr- I. 250
I. 250
.078
· 094
.212
.252
.307
.314
.0301
· 0354
.0206
.0245
.377
.375
.312
.312 ~
13
1.250 1. 875
2.125
· 094
.094
· 094
.125 .458
.505
· I128
.1285
.1442
· 0971
.1423
.530
.534
.460
.531
--IT- I. 250 .125 .326 .327 · 0449 · 0319 .371 .313
~
I. 5uO · u.4 .• I1 .15.1 · 0490 .622 .345
.192 .180 .0IlO .240 .370 .125 .594 .2268 .1023
-¥s- .875 · 078
.125 .295 .198
· 0263
.234 .372
~ I. 500
I. 750
2. 000 .125 .125 .656 .1543
.618
.625
.415
.485
17
.875
1.125 .078 .212 .254
· 0161
· 0228
· 0407
· 0263 .328 .352 ~
17 2.500 .125 .781
.2564
.3156 .3058 .636 .626
--IT- I. 625 1.125 .125 .125 .326· .274 · 0339 · 0408 .322 .364
~ 2. 000 2. 000 ~ .125
.552 .3922 · I183 .843 .463
49 I. 375 · 094 .275 .345 · 0475 · 0314 .415 .338 19 .125 .719 .4946 .1544 .830 .464
r-To- I. 375 .125 .358 .358 · 0605 · 0408 .4I1 .338
~ ~ .607 .6571 .1184 I. 04 .442
r-rr- I. 375 .15b
.094
.437
.240
.370
.179
.0725
.0131
.0499
.0393
.407
.234
.338
.405
21 2.500 2.000 .125 .156 .789
.054
.8315 .1546
.1184
I. 03 .443
~
~
.875
.875 .125 .311
.263
.191
.252
.0164
.0273
· 0512 .230
.322
.406
.386
~
23
3. 000 2. 000
I. 500
-4E--
.078
.156
.851
.542
· •• 23
1. 262
1. 298
.1547
.0413
1.23
1. 22
1. 55
.42
.426
.275
~ 1.125 · 094
.276
· 0393
.0628 .314 .387 ~
~
.156 .418 .0412 .740 .1184 I. 60 .400
~ 1. 125
.125 .287 .333 .0484 .0393 .411 .370 ~ 2.000
.125
.969
I. 888
.1547
~ I. 750 1. 375
I. 375
· 094
.156 .457 .358 .0627 .403 .371 ~
27 4. 000 3.000
.125
.156 1. 52
2.424
4.136 .6586
1.58
1.65
.400
.658
~ · 0741
.0394 .500 ----za
~ 1. 625
1. 625
.094
.125
.310
.405
.420
.434
.0777
.0995 .496
.356
.356 29 3.500 ~
.250
.188 2.00
2.60
5.521
6.968
1. 258
1. 638
1.66
1. 64
.794
.794
~ 1. 625 .156 .497 .446 .1200
· 0513
.0629 .492 .356 -T 4. 000 .313 .250 3.53 9.415 3.037 1. 63 .928
30
A3.14 CENTROIDS, CENTER OF GRAVITY, MOMENTS OF INERTIA
Table A3.16
SECTION PROPERTIES OF TYPICAL AIRCRAFT EXTRUDED SECTIONS
Sect. Dimensions Are~ Prooerties about X-X Properties about y-y

No. A B C t tl t2 R in. Axis Ax is r----=---=-
I xx exx e I yy eyy e'
!~ 1 17/32 13/16 1/8 .040 .050 1/16 .068 .00686 .317 .319 .00150 .148 .112
-.-:C i"- 2 1/2 1 7/32 .050 .050 .050 .113 .01515 .3665 .510 .00155 .117 .110
:t~
90- 3 1/2 7/8 3/16 1/16 1/16 3/32 .1025 .0100 .313 .383 .00162 .125 .116
e t
4 9/16 7/8 3/16 .060 .060 3/32 .1031 .0102 .313 .383 .00215 .114 .130
r y.-J
~ e
k- ~
--J,
L 5 9/16 7/8 3/16 .050 .078 13/32 .1061 .0120 .329 .345 .00302 .165 .133
--~~ --
l_e'~_ -It 6 3/4 1-1/8 3/16 .075 .075 i 1/16 .1514 .0239 .397 .435 .00550 .190 .182
- - -.. _ - -
L~A-.I
_.- -
7 3/4 1-5/16 9/32 3/32 3/32 9/64 .230 .0507 .471 .574 .0080 .187 .171
8 7/8 3/8 1/16 1/16 1/16 .095 .0101 .326 .4375 .00111 .108 .108
'j
9 7/8 3/8 3/32 3/32 3/32 .1375 .0136 .314 .4375 .00148 .104 .118
'l(
i
' t
A~.- I --
e
1._ '_
I I'-
I
-x
,--t\
10
11
12
1
1-1/4
1-13/16 1
1/2
11/16
3/32
3/32
3/16
3/32
3/32
3/16
3/32
3/32
3/32
.1736
.2320
.644
.0240
.0536
.2920
.371
.480
.677
.500
.625
.906
.00376
.00992
.0555
.147
.206
.295
.158
.210
.326
ie'~
1
1 13 2-1/2 1 1/8 1/8 1/8 .581 .525 .950 1.25 .0662 .338 .323
~- 84 14 2-1/4 1-1/8 1/8 3/16 1/8 .666 .524 .888 1.125 .0872 .362 .375
15 1-3/8 11/16 1/8 1/16 1/16 1/16 .137 .00425 .176 .133 .0178 .360 .687
~I- -- I~ A .t1 J 16
17
1-5/8
1-5/8
-- I--
3/4
3/4
----
5/32
- -
5/32
1/16
1-----
1/16
1/16
3/32
1/16
5/64
--
.168
.206
.00580
.00589
.186
.170
.133
.121
.0335
.0458
.446
.474
.812
.812
ey--,~C
rrl- - !L_)( 18 2 1-1/16 3/16 3/32 3/32
1---
3/32 .306 .0252 .287 .201 .0791 .509 1.0
~ I,
~ _.:I-t 19 2 1-1/4 3/16 1/8 3/32 3/32 .359 .0475 .364

,
.300 .0827 .479 1.0
,-""-->Is 20 2 1-5/32 1/4 3/32 1/8 1/8 .378 .0291 .279 I

.217 .1095 • 541 1.0
21 2-5/16 1-1/2 5/16 5/32 5/32 , 5/32 .668 .1075 .401 .338 • 2326 .590 1.156
leF 22 1-1/2 1-1/8 1/16 1/16 1/16 .161 .0186 .342 0255 .0171 .326 .750
I A -----e 23 1-1/2 2 1/8 3/32 1/8 .384 .1562 .641 .658 .0346 • 302 .750
:-T( ~*T" 24 1-3/4 2 9/32 1/8 1/8 .760 .315 .643 .757 .0599 .281 .875
25 1-3/4 2-9/16 1/4 5/32 : 5/32 .918 .600 .809 .959 .0730 .282 .875
I
-L---I- e'
'
~ 26 2 2-1/2 5/32 1/8 i 5/32 .605 .385 .800 .793 .159 .514 1.00
!--
,
11, 27 11/16 1 1/8 .050 .050 .050 3/32 .135 .0218 .404 .500 .0163 .348 .6625
28 13/16 1-1/8 3/16 .060 .060 .050 ! 3/32 .180 .0375 .456 .562 .0295 .405 .7825
,1r.:t;"
I _.~ I
8 ~- -
II---t
-
-- --)\
t...,..-
,
e
! 29
30
5/8
3/4
1-;1/2
2
3/16
1/4
.050
.050
1/16
1/16
1/16
1/16
3/32
3/32
.166
.214
.0603
.1382
.595
.793
.75
1.00
.0151
.0281
.298
.358
.600
.725
~t_ :~;J 31 3/4 2 1/4 1/16 1/16 1/16 3/32 .2378 .1449 .773 1.00 .0282 .341 .719
!Le' .j 32 3/4 2 1/4 1/16 3/32 3/32 3/32 .2855 .1820 .795 1.00 .0365 .356 .719
33 9/16 7/8 .050 .050 1/16 .099 .0120 .349 .437 .0056 .238 .537
LA ~ 34 1/2 7/8 1/16

-
1/16 1/16 .115
- - I------
.0135 .341 .437 .00446 .197 .469
~e'''''
_.--- --
I~
35 9/16 1 1/16 .080 3/32 .151 .0239 .397 .500 .00828 .234 .531
-
I ~ I~-t 36 9/16 1 1/4 .075 .075 3/32 .170 .0384 .475 .625 .00876 .227 .525
B -->;
*rl
e
37 11/16 1-1/4 1/16 .100 3/32 .213 .0550 .508 .625 .0194 .302 .655
J -- - L6==J.
~ ~,
38
39
.825
3/4
.-1-------
1-1/2
2
.075
.072
.120
.072
3/32
1/8
.302
.249
.110
.1446
.604
.757
.750
1.00
.0391
.0180
.360
.267
.787
.714,
40 3/4 2 .102 .102 3/32 .330 .1826 .750 1.00 .0228 .265 .700
- - -
CHART A3.1 PROPERTIES OF FORMED CHANNEL SECTIONS
I"
.. 4 1:L+~j n-:;-1 r":,i TTT'
{~~~H
,.~_"
..,.;C . . . ,.:iIIH
_ 0 . 3 5 1lEI::::
4Nt
. . tI:' .•
- , -,
~r::r:"""""'-
T- ,;,: :':T=: _~
'tifF'
i::t:
< .3
; .1
+, jm:1L
+ll'H· tt+
I 0.30 < +co i· :'tt-: t~
<
""e;j
0
II
+~'lli I
...:
.2
'7t:i
.~=+= .
:-:~~:.1:.:...u
t;
,t
.iE,bF
:en
H'
.8
IIf~
~~!
L:';J+;.;: W=2
0.25 .1 ·:t 3
J. ",.j ,"".;1.
'S'" ::tW= lc.1:"4
~-- ~En- ' i--:::-:. :!::t",',U"j
r :j:::::::11
m'
1_,;,c,~
, . c' I.•..
~".;' r4 . l.'~'.. :s 'e

~3 1 -U~"'i~~ m"~
.~ ,~
ii'S
fl If
, _ ~
:.t~~~:-.~ ~
,c"' ...
4-
',1=
~.:..;
0 20
. ~
_ ='---W
~-
'It ::.1="
--= tr 1::=:."
. ii. :s
L = ..- :':~L: !i:
, ,
N.A. t
1
~"
__,,':
_·~'.:,f'
o-~~
n'
.'
. " ::;F'-,
~
~
0::;i
= r-
:::=:: MultIply A&I------J
values
H
............-L-.,. obtained from curves
== by the folloWIng gauge ~

_ ~
~ 'I
~~_
= t+
1"'-
'-:-::' : ~:=:-=
{nil
"-1
-:t-_:'
~~---'-I
:s""
~
0
~
:.:4 .5
." -==-
0.15 ::;i
_::.' factors:
p t:-::-, .. ""
::;i
.HH'l"''' t I A::,:... R=:j ~
I . 8 . 064
+t+: .072
.051 1. 00
1. 23
1. 375
1. 00:;::
1. 24 "'
}.~=:;::
~ :-r;
,1
: 3'
.4
- I
EfutE±: I"·l=::
4"
. r:'ll
'~'I ~:..r~t :i ; L'; ..
H
,:::h,
;.l-<j'-<4±k~ ..\v=it:·, -1-1'
1::-1
.3
·~-I-I·. 'i::
-+-~:-
1;: ..•:n: ~
0.05
~ n:-~
.2
-;::L"t it
tj,'+'H
:c1Tltr;y' II· ,~:;-i; n::::t:
+ti-:J~'
--<'-·-It
+!~
:j:: '1:" ;1-t-'
Mom. of Inertia: ~m:EtE!EJ .1
+-i:t:tU-_i_U--J:-'..l..1.-:~'::j
"lL;jit:!:::-~ t :.a-=--L~U1i: Area:
TIl
•• j
o ~:
dh
-13fT
1.0 2.5 3.0 3.5 ::tJ:t:tt:;:tt'clL:
1.5 2.D
;;-;-;!n;':l~;;:n..;r 1::I;,:,j""
H - HEIGHT OF CHANNEL .025 >
w
1.5 2.5 3 3.5 4
....
H : HEIGHT OF CHANNEL
'"
A3.16 CENTROIDS CENTER OF GRAVITY MOMENTS OF INERTIA
A3.15 Problems (5) For the beam section in Fig. A3.18,

calculate the moment of inertia about the prin-
cipal axes assuming the four stringers as the
only effective material.
~d' + +(0-1
8
/05"''' /.e 0·8 /,2 T L
10''- '2 0 " "/..
1
Fig. A3.19
~
I
,",2. D," 0.8 /.
Fig. A3.14 Fig. A3.15 /5 +1 8
(1) For the section of Fig. A3.14 determine

the moment of inertia about each of the princi- (6) Fig. A3.19 shows a wing beam section
pal axes Xp and Zp. with a cut-out on the lower surface. Determine
(2) Calculate the moment of inertia of the the moments of inertia about the principal axes
section in Fig. A3.15 about the principal axes. assuming the eight stringers are the only effec-
tive material.
----,:
~
i
a" Trrrll;.,
II " 0• 12,,~L
~
0.12. .020 i-
/,
ng. A3.N.
n
- - - I
0.4-°11
1- r 6 '1"""' @ 6" 36" .1 Fig. A3.20
Fig. A3.16 Fig. A3.17

(7) Fig. A3.20 shows a 3 cell multiple
(3) Fig. A3.16 illustrates a box type beam flan~e beam. The 7 flange members on the upper
section with six longitudinal stringers. De- face of beams have an area of .3 sq. in. each
termine the moment of inertia of the beam sec- and those on the bottom skin 0.2 sq. in. each.
tion about the principal axes for the follow- The bottom skin is .03 inches in thickness.
ing assumptions:- Compute the moments of inertia about the
(a) Assume the beam is bending upward principle axes assuming that the flange members
putting the top portion in compression and the and the bottom skin comprise the effective
lower portion in tension. Therefore, neglect material.
sheet on the top side since it has very little
resistance to compressive stresses. The sheet
on the bottom side is effective since it is in
tension. For simplicity neglect the vertical
webs in the calculations.
(b) Reverse the conditions in (a) thus \
placing top side in tension and lower side in \ Cutout for door
compression.
f--
(4) For the three stringer single cell box
beam section in Fig. A3.17, calculate the mo- Fig. A3.21 /
I
ments of inertia about the principal axes. As-
sume all web or wall material ineffective.
0"
....-./.0
1. 5"
(8) Fig. A3.21 shows the cross-section of
a small fuselage. The dashed line represents
l I
(
!)"
Fig. A3.18
a cut-out in the structure due to a door. As-
sume each of the 13 stringers have an area of
a" 0.1 sq. in. Consider fuselage skin ineffective.
10 .8
Calculate the moment of inertia of the effective
section about the principal axes.
CHAPTER A4
GENERAL LOADS ON AIRCRAFT
A4. 1 Introduction. engers safely, efficiently and comfortably over

Before the structural design of an airplane various distances between airports. On the
can be made, the external loads acting on the other hand the Air Force Fighter type of air-
airplane in flight, landing and take-off con- craft has a job of shooting down enemy aircraft
ditions must be known. The complete determin- or protecting slower friendly aircraft. To do
ation of the air loads on an airplane requires a this job efficiently requires a far different
thorough theoretical knowledge of aerodynamics, configuration as compared to the DC-8 transport.
since modern aircraft fly in sub-sonic, trans- Furthermore the Fighter type airplane must be
sonic and super-sonic speed ranges. Further- maneuvered far more sharply to do its required
more, there is a wide range of wing configur- job as compared to the DC-8 in doing its re-
ations, such as the straight tapered wing, the quired job.
swept wing and the delta wing, and many of In general the magnitUde of the air forces
these wings often include leading and trailing on an airplane depend on the velocity of the
edge devices for promoting better lift or con- airplane and the rate at which this velocity is
trol characteristics. The presence of power changed in magnitude and direction (acceleration).
plant nacelle units, external fuel tanks, etc. The magnitude of the flight acceleration factor
are units that affect the airflow around the may be governed by the capacity of the human
wing and thus affect the magnitude and distri- body to withstand these acceleration inertia
bution of the air forces on the wing. Likewise, forces without injury which is the situation in
the fuselage or airplane body itself influences a fighter type of airplane. On the other hand
the airflow over the wing. The theoretical cal- the maneuvering accelerations for the DC-8 are
culation of the airloads on the airplane is too not dictated by what the human body can With-
large a subject to be covered in a structures stand, but are determined by what is necessary
book and it is customary in college aeronautical to safely transport passengers from one airport
curricula to provide a separate course for this to another.
SUbject. Designing the airplane structure for loads
In most airplane companies the loads on greater than the airplane suffers in the per-
the airplane are determined by a group of en- formance of its reqUired job, obviously will add
gineers assigned to the structures Analysis considerable weight to the airplane and decrease
Section and this group is often referred to as its performance or over-all efficiency relative
the Aircraft Load Calculation group. While the to the job it is designed to do.
work of this group is primarily based on the To particularly insure safety in the air-
use of aerodynamics, it is that phase of aero- transportation, along With uniformity and ef-
dynamics which is concerned with determining ficienc~ of design, the government aeronautical
the magnitude and distribution of the air loads agencies (civil and military) have definite re-
on the airplane so that the airplane structure quirements for the various types of aircraft
can be properly designed to support these air relative to the magnitude of loads to be used in
forces safely and efficiently. The engineering the structural design of aircraft. In referring
department of an airplane company has a distinct in general to these specified aircraft loads two
or separate aerodynamics section, but in general terms are used as follows:-
their responsibility is the use of the SUbject
of aerodynamics to insure or guarantee the per- Limit or Applied Loads.
formance, stability and control of the airplane. The terms limit and applied refer to the
A basic general over-all knowledge of the same loads With the civil agencies (C.A.A.)
loads on aircraft is desirable in the study of using the term limit and the military agencies
aircraft structural theory, and hence this using the term applied.
chapter attempts to give this information. In Limit loads are the maximum loads antici-
a later chapter dealing with wing design, this pated on the airplane during its lifetime of
subject will be further expanded. service.
The airplane structure shall be capable of
A4.2 Limit or Applied Loads. Design Loads.
supporting the limit loads Without suffering
Because an airplane is designed to carry detrimental permanent deformations. At all loads
out a definite job, there result many types of up to the limit loads the deformation of the
aircraft relative to size, configuration and structure shall be such as not to interfere With
performance. For example, a commercial trans- the safe operation of the airplane.
port like the Douglas DC-8 is designed to do a Ultimate or Design Loads.
job of transporting a certain number of pass- These two terms are used in general to mean
A4.1
A4.2 GENERAL LOADS ON AIRCRAFT
the same thing. Ultimate or Design Loads are Thrust.

equal to the limit loads multiplied by a factor (3) Power Plant Loads { Torque.
of safety (F.S.) or
Design Loads = Limit or Applied Loads times F.S. Catapulting.
In general the over-all factor of safety is (4) Take off Loads Assisted take off with
1.5. The government requirements also specify { auxiliary short period
that these design loads be carried by the thrust units.
structure without failure. Hoisting Airplane.
Although aircraft are not supposed to Towing Airplane.
undergo greater loads than the specified limit (5) Special Loads Beaching of Hull type Airplane
loads, a certain amount of reserve strength {
Fuselage Pressurizing.
against complete structural failure of a unit is
necessary in the design of practi8ally any ma-
chine or structure. This is due to many factors (6) Weight and Inertia Loads.
such as:- (1) The approximations involved in In resolving external loads for stress
aerodynamic theory and also structural stress analysis purposes, it is convenient to have a
analysis theory; (2) Variation in physical
properties of materials; (3) Variation in fab- set of reference axes. The reference axes
rication and inspection standards. Possibly XYZ passing through the center of gravity of the
the most important reason for the factors of airplane as illustrated in Fig. A4.0 are those
safety for airplanes is due to the fact that normally used in stress analysis work as well as
practically every airplane is limited to the for aerodynamic calCUlations. For convenience
maximum velocity it can be flown and the maxi- the reference axes are often referred to another
mum acceleration it can be subjected to in origin other than the airplane c.g.
flight or landing. Since these are under the z
control of the pilot it is possible in emerg- y
ency conditions that the limit loads may be
slightly exceeded but with a reserve factor of
safety against failure this exceeding of the
limit load should not prove serious from an
airplane safety standpoint, although it might
cause permanent structural deformations that
might require repair or replacements of small x
y
units or portions of the structure.
Loads due to airplane gusts, are arbitrary Z
in that the gust velocity is assumed. Al- Fig. A4.0
though thiS gust velocity is based on years of
experience in measuring and recording gust A4.4 Weight and Inertia Forces.
forces in flight allover the world, it is qUite The term weight is that constant force, pro-
possible that during the lifetime of an air- portional to its mass, which tends to draw every
plane, turbulent conditions near storm areas or physical body toward the center of the earth.
over mountains or water areas might produce air An airplane in steady flight (uniform velocity)
gust velocities slightly greater than that is acted upon by a system of forces in equilib-
specified in the load requirements, thus the rium, namely, the weight of the airplane, the air
factor of safety insures safety against failure forces on the complete airplane, and the power
if this situation would arise. plant forces. The pilot can change this bal-
anced steady flight condition by changing the
engine power or by operating the surface controls
The broad general category of external to change the direction of the airplane velocity.
loads on conventional aircraft can be broken These unbalanced forces thus cause the airplane
down into such classifications as follows:- to accelerate or de-accelerate.
Due to Airplane Maneuvers. (under Inertia Forces For Motion of Pure Translation
the control of the pilot). of Rigid Bodies.
(1) Air Loads Due to Air Gusts. (not under If the unbalanced forces acting on a rigid
{
control of pilot). body cause only a change in the magnitude of the
velocity of the body, but not its direction, the
\
~~~d~~~e)~ Land. (wheel or motion is called translation, and from basic
Physics, the accelerating force F = Ma, where M
(2) Landing Loads Landing on Water. is the mass of the body or Wig. In Fig. A4.1
the unbalanced force system F causes the rigid
Arresting. (Landing on Air- body to accelerate to the right. Fig. A4.2 shows
craft Carriers). the effect of this unbalanced force in producing
A4.3
at = ra, and an = rw s an acceleration normal to

F = unbalanced Effective Force
external force =Z rna = Ma the flight path at A and directed toward the
!rn~a/rnsa
center of rotation (0). From Newton's Law the
effective forces due to these accelerations
--
r - - - - - - - , Motion
..,.. .!'-~-
Motion are:-
Fn = Mrw s = Mvs;r - - (4)
T
w F = Mra - - - - - - - - - -
(5 )
t
Where w = angular velocity at the point A.
a = angular acceleration at point A.
Fig. A4.1 Fig. A4.2 r = radius of curvature of flight path
at point A.
The inertia forces are equal and opposite
a force on each mass particle of m~a, msa, etc., to these effective forces as indicated in Fig.
thus the total effective force is Zilla = Ma. If A4.3. These inertia forces can then be con-
these effective forces are reversed they are re- sidered as part of the total force system on the
ferred to as inertia forces. The external airplane which is in equilibrium.
forces and the inertia forces therefore form a If the velocity of the airplane along the
force system in equilibrium. path is constant, then at = zero and thus the
From basic Physics, we have the following
relationships for a motion of pure translation inertia force F = 0, leaving only the normal
t
if the acceleration is constant:- inertia force F •
n
=
v - Vo at - - (1) If the angular acceleration is constant,
the following relationships hold.
s = vot + ~ at S (2)
v - v o s = 2as (3) w- We = at - - - - - - - - - - - (6)
Q = wot + ~atS (7)
Where,
WS
_ Wo S = 2aQ - - - - - - (8)
s = distance moved in time t.
where G = angle of rotation in time t.
V o = initial velocity
Wo = initial angular velocity in rad/sec.:
=
v final velocity after time t. w = angular velocity after time t.
In Fig. A4.3 the moment To of the inertia
Inertia Forces on Rotating Rigid Bodies. forces about the center of rotation (0) equals
A common airplane maneuver is a motion Mra(r)= Mrsa. The term Mr 2 is the mass moment
along a curved path in a plane parallel to the of inertia of the airplane about point (0).
XZ plane of the airplane, and generally referred Since an airplane has considerable pitching
to as the pitching plane. A pUll up from steady moment of inertia about its own center of gravity
flight or a pUllout from a dive causes an air- axis, it should be included. Thus by the
plane to follow a curved path. Fig. A4.3 shows parallel axis
an airplane following a curved path. If at
point A the velocity is increasing along its To = loa + I e.ga - - - - - - - - - - - - - (9)
path, the airplane is being sUbjected to two where 1 0 = Mrs and I e.g = moment of inertia of
accelerations, namely, at, tangential to the
curve at point A and equal in magnitude to airplane about Y axis through e.g. of airplane.
Inertia Forces For Pitching Rotation of Airplane
Center of Curvature about Y Axis Through e.g. airplane.
01 Flight
Path
r In flight, an air gust may strike the hori-
\ zontal tail producing a tail force which has a
\ moment about the airplane e.g. In some landing
an=Fws=VsIF conditions the ground or water forces do not
pass through the airplane e.g., thus producing
a moment about the airplane e.g. These moments
cause the airplane to rotate about the Y axis
through the e.g.
Therefore for this effect alone the center
Fig. A4.3 of rotation in Fig. A4.3 is not at (0) but at
the e.g. of airplane, or r = o. Thus Fn and Ft air forces on the wing. For example, consider
the two air pressure intensity diagrams in Figs.
equal zero and thus the only inertia force for A4.6 and A4.7. These distributed force systems
the pure rotation is I e.g. a, (a couple) and can be replaced by their reSUltant (R), whiCh
thus the moment of this inertia couple about the of course must be known in magnitUde, direction
e.g. = Te.g. = I e.g. a. and location. The location is specified by a
term called the center of pressure which is the
As explained before if the inertia forces point where the resultant R intersects the air-
are included with all other applied forces on foil chord line. As the angle of attack is
the airplane, then the airplane is in static changed the reSUltant air force changes in mag-
equilibrium and the problem is handled by the nitUde, direction and center of pressure
static equations for equilibrium. location.
A4.5 Air Forces on Wing.
The wing of an airplane carries the major
portion of the air forces. In level steady
flight the vertical upward force of the air on
the wing, practically equals the weight of the
airplane. The term airfoil is used when re-
ferring to the shape of the cross-section of a
wing. Figs. A4.4 and A4.5 illustrate the air
pressure intensity diagram due to an air-
Fig. A4.6 Fig. A4.7
•
Lift and Drag Components of ReSUltant Air Force.
Instead of dealing with the reSUltant force
R, it is convenient for both aerodynamic and
stress analysis considerations to replace the
resultant by its two components perpendicular
-
and parallel to the airstream. Fig. A4.8 il-
lustrates this resolution into lift and drag
components.
Angle of Attack Angle of Attack
L
= 120 = _ 60
Fig. A4.4 Fig. A4.5 AERODYNAMIC

CENTER
M
.P.
stream flowing around an airfoil shape for both FLIGHT DIRECTION FLIGHT DIRECTION
a positive and negative angle of attack. The
shape and intensity of this diagram is in-
fluenced by many factors, such as the shape of Fig. A4.8 Fig. A4 9
the airfoil itself, as the thiokness to chord
ratio, the camber of the top and bottom sur-
faces etc. A normal wing is attached to a Aerodynamic Center (a.c.). Since an air-
fuselage and it may support external power plane flies at many different angles of attack,
plants, wing tip tanks etc. Furthermore the it means that the center of pressure changes for
normal wing is usually tapered in planform the many flight design conditions. It so hap-
and thickness and may possess leading and pens, that there is one point on the airfoil
trailing slots and flaps to produce high lift that the moment due to the Lift and Drag forces
or control effects. The airflow around the is constant for any angle of attack. This
wing is affected by such factors as listed point is called the aerodynamic center (a.c.)
above and thus wind tunnel tests are usually and its approximate location is at the 25 percent
necessary to obtain a true picture of the air of chord measured from the leading edge. Thus
forces on a wing relative to their chordwise the resultant R can be replaced by a lift and
spanwise distribution. drag force at the aerodynamic center plus a wing
moment Ma.c. as illustrated in Fig. A4.9.
Resultant Air Force. Center of Pressure.
A4. 6 Forces on Airplane in Flight.
It is convenient when dealing With the Fig. A4.l0 illustrates in general the main
balancing or equilibrium of the airplane as a forces on the airplane in an accelerated flight
Whole to deal with the reSUltant of the total condition.
A4.5
tiplying factor by which the forces on the air-

plane in steady flight are multiplied to obtain
e a static system of forces equivalent to the dy-
namic force system acting during the accelera-
tion of the airplane. Fig. A4.l1 illustrates
Z
1
LA= total lift (Wing & Tail)
g'oj --.£.c
I - - ~
Thrust L' Chord . W
~~ ,/:----c.ge ;J' D
Flight P: . e>. Fig. A4. 10
orizontal
T=D
Fig. A4.11
W
=
T engine thrust.
forces in steady horizontal flight. L repre-
L = total wing lift plus fuselage lift.
D = total airplane drag. sents the total airplane lift (wing plus tail).
=
Ma moment of Land D with reference to wing Therefore L = W. Now assume the airplane is ac-
celerated upward along the Z axis. Fig. A4.l2
a.c. (aerodynamic center)
shows the additional inertia force wag/g acting
W = weight of airplane.
I L = inertia force normal to flight path. downward, or opposite to the direction of
acceleration. The total airplane lift L for the
I D = inertia force parallel to flight path. Z
1 = rotation inertia moment. I
m nzL
E = tail load normal to flight path.
For a horizontal constant velocity flight
condition, the inertia forces I , I D, and 1 ~ c.g.o_D'
L m T-
would be zero. For an accelerated flight con-
W
dition involving translation but not angular T=D
Fig. A4.12
acceleration about its own c.g. axis, the
inertia moment 1m would be zero, but I and I ~ ~a
g z
L D
would have values. unaccelerated condition in Fig. A4.11 must be
mUltiplied by a load factor ng to produce static
Equations of Equilibrium For Steady Flight. equilibrium in the g direction.
From Fig. A4.l0 we can write:-
Thus, naL
..
- W- ~g a g = 0
~Fx = 0, D + W sin G-T cos 13 =0

Since L = W
~F
g.= 0, L - Wcos G+T sin f3 - E =0
Hence n = I + ag
ZMy = 0, - Ma - La - Db + Tc cos f3 + Ee =0 g g
An airplane can of course be accelerated along

Equations bf Equilibrium in Accelerated Flight. the X axis as well as the g axis. Thus in
~F = 0, D + Wsin G - T cos J3 I = 0 Fig. A4.l3 the magnitude of the engine thrust T
x D is greater than the airplane Drag D, which
nzL
~Fg = 0, L-W cos G + T sin ~ I -
L - E =0
ZMy = 0, - Ma - La - Db + Tc cos f3 + Ee + I m = 0
Forces - Plus is up and toward tail. j:--e.g. D
Signs used: Moment - Clockwise is positive. W

Fig. A4.13 T is greater than D
{ Distances from c.g. to force -
Plus is up and toward tail. twag z
A4. 7 Load Factors. causes the airplane to accelerate forward. It is
The term load factor normally given the convenient to express the inertia force in the X
symbol (n) can~defined as the numerical mul- direction in terms of the load factor n and the
x
• The bar through letter Z has no significance. Same mean-
ing without bar.
weight W of the airplane, hence a definite maximum flight velocity. For com-
mercial airplanes the velocity is limited to a
n W ~ a (See Fig. A14.13) reasonable glide speed which is sufficient to
x g x take care of reasonable flight operations.
0,whence T-D-nxW = 0 A4. 9 Gust Load Factors.

When a sharp edge gust strikes the airplane
Hence n T-D in a direction normal to the thrust line (X
x W axis), a sudden change takes place in the wing
angle of attack with no sudden change in airplane
Therefore the loads on the airplane can be dis- speed. The normal force coefficient (C Z ) can
cussed in terms of load factors. The applied or A
limit load factors are the maximum load factors be assumed to vary linearly With the angle of
that might occur during the service of the par- attack. Thus in Fig. (a), let point (B) repre-
ticular airplane. These loads as discussed in sent th~ normal airplane force coefficient C
Art. A4.2 must be taken by the airplane struc- zA
ture without appreciable permanent deformation. necessary to maintain level flight With a
The design load factors are equal to the velocity V and point (C) the value of C after
limit load factors multiplied by the factor of zA
safety, and these design loads must be carried a sharp edged gust KU has caused a sudden change
by the structure without rupture or collapse, or ~a in the angle of attack without change in V.
in other words, complete failure. The total increase in the airplane load in the Z
direction can therefore be expressed by the
A4.8 Design Flight Requirements for Airplane. ratio C at B.
The Civil and Military Aeronautics Author- zA
ities issue requirements which specify the From Fig. (b) for small angles, ~a = KU/V
design conditions for the various classifications and from Fig. (a), ~CZ = m ~a, where m = the
of airplanes. Generally speaking, any airplane A
flight altitude can be defined by stating the slope of the airplane normal force curve
existing values of load factors (acceleration)
and the airspeed (or more properly the dynamic per radian).
pressure).
The accelerations on the airplane are pro-
duced from two causes, namely, maneuvers and air
gusts. The accelerations due to maneuvers are
subject to the control of the pilot who can
manipulate the controls so as not to exceed a
certain acceleration. In highly maneuverable KU~
military airplanes, an accelerometer is in-
V a
cluded in the cockpit instruments as a gUide to
limit the acceleration factor. For commercial Fig. b Fig. a
airplanes the maneuver factors are made high
enough to safely take care of any maneuvers that The load factor increment due to the gust
would be required in the necessary flight opera- KU can then be expressed
tions of the particular type of airplane. These
limiting maneuver factors are based on years of an ~CZ
A = (KUm) (pV"S)= KUVSm _ (A)
operating experience and have given satisfactory
results from a safety standpoint without pen-· c;:-
A
V 2W 575 W
alizing the airplane from a weight design con- Where
sideration.
The accelerations due to the airplane U = gust velocity in ft./sec.
striking an air gust are not under the control
of the pilot since it depends on the direction K*= gust correction factor depending on
and velocity of the air gust. From much ac- wing loading (Curves for K are provided
cumulated data obtained by installing accelero- by Civil Aeronautics Authorities).
meters in commercial and military aircraft and V = indicated air speed in miles per hour.
flying them in all types of weather and loca-
tions, it has been found that a gust velocity of S = wing area in sq. ft.
30 ft. per second appears sufficient. W = gross weight of airplane.
The speed or velocity of the airplane like-
wise affects the loads on the airplane. The * NACA Technical Note 2964 (June 1953), proposes that the
higher the velocity the higher the aerodynamic alleviation factor K should be replaced by a gust factor,
Kg = 0.88 Mg/5. 3 + Mg). In this expression Mg is the
wing moment. Furthermore the gust acceler- airplane mass ratio or mass parameter, 2 W/apcgs, in
ations increase with airplane velocity. Thus, it which c is the mean geometric chord in feet and g the ac-
is customary to limit the particular airplane to celeration due to gravity.
A4.7
If U is taken as 30 ft./sec. and m as the would give an acceleration ~ess than the lim-
change in C with respect to angle of attack ited values given by lines AB and CD.
zA The positive and negative gust accelera-
in absolute units per degree, equation (A) re- tions due to a 30 ft./sec. gust normal to flight
duced to the following path are shown on Fig. A4.l4. In this example
diagram, a positive gust is not critical within
3KmV (B) the restricted velocity of the airplane since
lin = W/S - - - - - - - the gust lines intersect the line BD below the
line AB. For a negative gust, the gust load
Therefore the gust load factor n when air- factor becomes critical at velocities between F
plane is flying in horizontal attitude equals and D with a maximum acceleration as given by
point E.
3KmV For airplanes which have a relatively low
n = 1 + -- - - - - - - - - - - - - - - - (C)
- W/S required maneuver factor the gust accelerations
may be critical for both positive and negative
and when airplane is in a vertical attitude accelerations. Examination of the gust equation
indicates that the most lightly loaded condition
n = -+ ~ - - - - - - - - - - - - - - - - (D)
(smallest gross weight) produces the highest
W/S gust load factor, thus involving only partial
pay load, fuel, etc.
A4.10 lllustration of Main Flight Conditions. On the diagram, the points A and B corre-
Velocity-Load Factor Diagram. spond in general to what is referred to as high
As indicated before the main design flight angle of attack (H,A,A,) and low angle of attack
conditions for an airplane can be given by (L.A.A.) respectively, and points C and D the
stating the limiting values of the acceleration inverted (H.A.A.) and (L.•A.A.+ conditions re-
and speed and in addition the maximum value of spectively.
the applied gust velocity. As an illustration, Generally speaking, if the airplane is de-
the design loading requirements for a certain signed for the air loads produced by the veloc-
airplane could be stated as follows: "The ity and acceleration conditions at points A, B,
proposed airplane shall be designed for applied E, F, and C, it should be safe from a structural
positive and negative accelerations of + 6.0g strength standpoint if flown within the specified
and -3.5g respectively at all speeds from that limits regarding velocity and acceleration.
corresponding to CL up to 1.4 times the Basically, the flight condition require-
max. ments of the Civil Aeronautics Authority, Army,
maximum level flight speed. Furthermore, the and Navy are based on consideration of specified
airplane shall withstand any applied loads due velocities and accelerations and a consideration
to a 30 ft./sec. gust acting in any direction of gusts. Thus a student understanding the basic
up to the restricted speed of 1.4 times the discussion above should have no difficulty un-
maximum level flight speed. A design factor of derstanding the design requirements of these
safety of 1.5 shall be used on these applied three government agencies.
loads If • For stress analysis purposes, all speeds
In graphical form these design require- are expressed as indicated air speeds. The
ments can be represented by plotting load fac- "indicated" air speed is defined as the speed
tor and velocity to obtain a diagram which is Which would be indicated by a perfect air-speed
generally referred to as the Velocity-accelera- indicator, that is, one that would indicate
tion diagram. The results of the above speci- true air speed at sea level under standard at-
fication would be similar to that of Fig. A4.l4. mospheric conditions. The relation between the
Thus, the lines AB and CD represent the re- actual air speed Va and the indicated air speed
stricted positive and negative maneuver load
factors which are limited to speeds inside line Vi is given by the equation
BD which is taken as 1.4 times the maximum
level flight speed in this illustration. These
restricted maneuver lines are terminated at
points A and C by their intersection with the
Vi ='
Vf§.Po Va
maximum CL values of the airplane. At speeds where

between A and B, the pilot must be careful not
to exceed the maneuver accelerations, since in Vi = indicated airspeed
general, it would be possible for him to man-
ipulate the controls to exceed these values. Va = actual airspeed
At speeds below A and C, there need be no care
of the pilot as far as loads on the airplane Po = standard air denSity at sea level
are concerned since a maneuver producing C •
Imax Pa = density of air in which Va is attained
the loads on the wing should be checked for

cases where the engines are attached to the wing
and are located forward of the leading edge.
In cases where the landing gear is attached
to wing or when the fuel and engines are carried
in and on the wing, the loads produced on the
wing structure in a landing condition may be
critical for some portions of the wing structure
inboard of landing gear and engine attachment
points.
A4. 12 Example Problems Involving Accelerated Motion of
Rigid Airplane.
As previously explained, it is general
practice to place the airplane under accelerated
conditions of motion into a condition of static
equilibrium by adding the inertia forces to the
applied force system acting on the airplane. It
is usually assumed that the airplane is a rigid
body. Several example problems will be pre-
sented to illustrate this general procedure.
Example Problem 1
Fig. A4.15 illustrates an airplane landing
on a Navy aircraft carrier and being arrested by
a cable pUll T on the airplane arresting hook.
If the airplane ·weight is 12,000 lbs. and the
airplane is given a constant acceleration of 3.5g
(112.7 ft/sec 2 ) , find the hook pUll T, the wheel
A4.11 Special Flight Design Conditions. reaction R, and the distance (d) between the line
There are many other flight conditions of action of the hook pUll and the airplane e.g.
which may be critical for certain portions of If the landing velocity is 60 M.P.H. What is the
the wing or fuselage structure. Most airplanes stopping distance.
are equipped with flaps, to decrease the land- W = 12000 lb.
ing speed and such flaps are lowered at speeds
at least twice that of the minimum landing
speed. Since the flapped airfoil has different
values for the magnitude and location of the
airfoil characteristics, the wing structure
must be checked for all possible flap conditions
within the specified requirement relative to
maximum speed at which the flaps may be oper-
R Fig. A4.15
ated. Generally speaking, the flap conditions
will effect only the wing portion inboard of the
flap and it is usually only critical for the Solution: -
rear beam web or shear wall and for the top and
bottom walls of the torsion box. This is due to On contact of the airplane with the arrest-
the fact that the deflection flap moves the ing cable, the airplane is decelerated to the
center of pressure considerably aft thus pro- right relative to Fig. A4.15. The motion is pure
ducing more shear load on the rear shear wall as translation horizontally. The inertia force is
well as torsional moment on the conventional
cantilever box metal beam. Ma = ~g a = (12000)
g'
3 5g = 42000 lb •
The airplane must likewise be investigated
for aileron conditions. Operation of the ailer- The inertia force acts opposite to the direction
ons produce a different air load on each side of acceleration, hence to the left as shown in
of the airplane wing which produces an angular Fig. A4.15.
rolling acceleration of the airplane. further- The unknown forces T and R can now be solved
more, the deflected ailerons change the mag- for by using the static equations of equilibrium.
nitude and location of the airfoil character-
istics, thus calculations must be carried out to ~FX = -42000 + T cos 10° 0
determine whether the loads in the aileron con-
ditions are more critical than those for the hence, T = 42700 lb.
normal flight conditions.
For angular acceleration resulting from -12000 + R - 42700 x sin 10n o
pitching moments due to air gusts on the tail,
A4.9
hence, R = 19420 lb. hence, Rs = 29800 lb. (up)

To find the distance (d) take moments about c.g.
~F~ = 29800 - 9000 + R1 = 0
of airplane,
hence, R1 = _ 20800 lb. (acting down)
L:M
c.g. 19420 x 24 - 42700 d = 0
hence, The velocity at end of catapult track can
d = 10.9 in. be found from the following equation
Landing velocity Vo = 60 M.P.H. = 88 ft/sec.
VS - 0 = 2 x 96.6 x 35
or
SUbt: - v = 82 ft/sec. = 56 M.P.H.
hence stopping distance s = 34.4 ft. Example Problem 3

Assume that the transport airplane as il-
Example Problem 2 lustrated in Fig. A4.17 has just touched down in
An airplane equipped with float is cata-
landing and that a braking force of 35000 lb. on
the rear Wheels is being applied to bring the
pulted into the air from a Navy Cruiser as il-
lustrated in Fig, A4.16. The catapulting force airplane to rest. The landing horizontal veloc-
P gives the airplane a constant horIzontal ac- ity is 85 M.P.H. (125 ft/sec). Neglecting air
forces on the airplane and assuming the propeller
celeration of 3g(96.6 ft/sec S ) . The gross forces are zero, what are the ground reactions
weight of airplane 9000 lb. and the catapult R1 and Rs • What is the landing run distance With
track is 35 ft. long. Find the catapulting
the constant braking force?
force P and the reactions R1 and Rs from the
catapult car. The engine thrust is 900 lb.
What is airplane velocity at end of track run? W = 100,000 lb.
9000 lb.
,:l-~_~ThrustLine __
r-
9'
c'g:--Ma -1- _
A
38'
Fig. A4.17
Solution: -
The airplane is being decelerated horizon-
Fig. A4.16
tally hence the inertia force through the air-
plane c.g. acts toward the front of the airplane.
Solution: - Since the braking force is given we can solVe
for the deceleration factor by the eqUilibrium
The forces will be determined just after equation,
the beginning of the catapult run, where the
car velocity is small, and thus the lift on the ~FX = 35000 - Max = 0
airplane wing and the airplane drag can be
neglected. hence, Ma = 35000
x
Horizontal inertia force acting toward the air- or W
plane tail equals, (-)
g x
a = 35000
whence
Ma = (9000) 3.0g 27000 lb. 35000 / s
g ax = (100000) 32.2 = 11.27 ft sec
From statics: -
To find landing run (s),
4Fx = -900 - P + 27000 = 0, hence P = 26100 lb.
VS - Vo s = 2 axs
To find Rs take moments about point A,
o - 125 s = 2 (-11.27) s
ZM = 9000 x 55 + 27000 x 78 - 900 x 83 - 85R s hence,
A = 0 8 = 695 ft.
To find Rs take moments about point (A) Airplane Load Factor = AirPl~e Lift
ZMA = 100,000 x 21 - 35000 x 9 + 38 Rs 0
111600 - 4100
Rs = 47000 lb. (2 wheels) = 14000 7.7
ZF~ = 47000 - 100,000 + Example Problem 5

R~ = 0
Assume the airplane as used in example
R~ = 53000 lb. problem 4 is in the same attitude as used in
that example prOblem. Now the airplane is
Example Problem 4 further maneuvered by the pilot suddenly pUSh-
The airplane in Fig. A4.l8 weighs 14,000 lb. ing the control stiCk forward so as to give
It is flying horizontally at a velocity of 500 the airplane a Pitching acceleration of
M.P.H. (733 ft/sec) when the pilot pulls it up- 4 rad/sec 2 •
ward into a curved path with a radius of curva- (a) Find the inertia forces and the tail load T,
ture of 2500 ft. Assume the engine thrust and assuming the lift force on the wing does not
airplane drag equal, opposite and colinear With change.
each other (not shown on Fig. A4.l8).
(b) Find the forces on the jet engine which
Find: - weighs 1500 lb. and whose c.g. location is
(a) Acceleration of airplane in Z di- shown in Fig. A4.19.
rection
(b) Wing Lift (L) and Tail (T) forces Assume moment of inertia I y (pitching) of the
(c) Airplane Load factor. airplane equals 300,000 lb. sec 2 • in.
93500 lb.
T T
111,600 lb.
210"
-~:"£"'J£!$~~~_--=:=..J Engine Thrust
14000
Fig. A4.18 Fig. A4.19
Solution: - Solution: -
s
s
V 733
Acceleration a~ - r - 2500 = 215 ft/sec
2 Fig. A4.l9 shows a free body of the air-
plane with the lift and inertia forces as found
or 214.5/32.2 = 6.68g (upward). in Problem 4.
The inertia force normal to the flight path The additional inertia force due to the
and acting down equals angular acceleration a = 4 rad/sec 2 • equals,
Ma = (14~00) 6.68g = 93520 lb. I a = 300000 x 4 = 1,200,000 in. lb.

z y
which acts clockwise or counter to the direction
Placing this force on the airplane through the of angular acceleration.
e.g. promotes static equilibrium, hence to find The airplane is now in static equilibrium
tail load T take moments about wing aerody- and to find the tail load T take moments about
namic center (c.p.) airplane c.g.
ZMc.p. =- (14000 + 93520) 8 + 210 T =0 ZMC • g • = 1,200,000 - 111600 x 8 - 218 T o
hence
T = 4100 lb. (down)
T = 1409 lb.
To find Wing Lift (L) use
To find Ma~ take,
ZF z = - 4100 - 14000 - 93520 + L = 0
ZF~ = 111600 14000 + 1409 - Ma 0
~
L = 111600 lb.
Ma = 99009 lb.
~
A4.11
hence, 99009) Find:

ag = ( 14000 g = 7.1 g ft/sec .. . (a) The inertia forces on the air-
plane.
The c.g. of the engine is 58 inches aft of the (b) The resultant load on the pilot
airplane c.g. as shown in Fig. A4.19. The whose weight is 180 lb. and whose
force on the engine will be its own weight of location is shown in Fig. A4.20.
1500 lb., and the inertia forces due to a g and
a.
Pilot e.g. ~
~40-'i'.+..d'~"r-J~-:-:~_--1_----.-----~ ~
Inertia force due to a g equals,
Mag = (15g00) 7.1g = 10630 lb.

I. ~n"
Inertia force due to angular acceleration a
equals, / I I , \ \ I I I \
~oo
Mra = 32.2 x 12 x 50 x 4 = 776 lb. (down ) 319000 lb. Fig. A4.20
Then the resultant force on the engine equals Solution: -

1500 + 10630 + 7'76 = 12906 lb. (down) The wing lift will be neglected in this
example problem.
Note if the engine had been forward of the air- The inertia forces on the airplane are
plane c.g., the inertia force of 776 lb. would
act upward instead of downward. forces Max and Ma g and the couple I c.g. a.
In calculating the inertia force on a To find Max take,
certain airplane item due to angular acceler-
ation,the equation F = Mra assumes that the
particular item had negligible mass moment of ZF x = 100,000 - Ma x 0
inert i.a about its own centroidal Y axis. In or
the caSe of a large item this centroidal mass Ma 100,000 lb.
moment of inertia may be appreciable and should x
be included in the I y of airplane. hence, 100,000
ax = (100,000) g = Ig ~
M 100,000
Then to find the inertia force for such
an item the equation F = Mra should be modified To find Mag take,
to be
= where ZF ~ = 300,000 - 100,000 - Ma~ =°

F (I
c.g. a)/r
r = distance or arm from airplane c.g. to c.g.

Ma g = 200,000 lb.
hence
of item. 200,000
( 200,000 )g = 2g t
100,000
I
c.g. = mass moment of inertia of item about M
airplane c.g. equals 1 0 + Mr" ,
where 1 0 is mass moment of inertia of To find the inertia couple I a, take moments
item about its own centroidal Y axis. about airplane c.g., c.g.
F = inertia force in Ibs. normal to radius r. ZM 100,000 x 120 300,000 x 84
c.g.
Example Problem 6 + I a = 0
Fig. A4.20 shows a large transport air- c.g.
plane Whose gross weight is 100,000 lb. The
airplane pitching mass moment of inertia I
c,g,
a = 37,200,000 lb.
I y = 40,000,000 lb. sec". in.
. - 37,200,000
The airplane is making a level landing t
hence angular accelera,lon a- 40 , 000 , 000
With nose wheel slightly off ground. The re-
action on the rear Wheels is 319,000 lb. in- 0.93 rad/sec".
clined at such an angle to give a drag com-
ponent of 100,000 lb. and a vertical component
of 300,000 lb.
Calculations of resultant load on pilot: - are for design loads, which in general are 1.5
times the applied loads. It would not be correct
x- = 372" to say that the wing deflections under the ap-
plied loads for these two High Angle of attack
conditions would be 2/3 the deflections shown in
the photograph since under the design loads a
considerable portion of the wing would be stressed
beyond the elastic limit of the material or into
the plastic range where the stiffness modulus is
0.93
180* Fig. A4.21
Fig. A4.21 shows the airplane c.g. accelerations

The forces on the pilot consist of the
pilots weight of 180 lb. and the various inertia
forces as indicated in the figure.
Max = e~O) Ig = 180#
Ma~ = e~o~ 2.0g = 360#

The inertia force due to the angular ac-
celeration a acts normal to the radius arm
between the airplane c.g. and the pilot. For
convenience this normal force will be replaced
by its ~ and x components.
180
= 32.2 x 12 x 40 x 0.93 = 17 lb.
-
F~ = Mxa = 32.2180x 12 x 372 x 0.93 = 161 lb.
Total force in x direction on pilot equals
180 - 17 = 163 lb. Fig. A4.21
Total force in ~ direction =

considerably less than the modulus of elasti-
360 + 180 - 161 = 379 lb. city, hence the deflections under the applied
loads would be somewhat less than 2/3 those
Hence Resultant force R, equals shown in the photograph. ThiS photograph thus
indicates very strikingly that a wing structure
is far from being a rigid body.
Static loads are loads which are gradually
A4.13 Effect of Airplane Not Being a aigid Body. applied and cause no appreciable shock or vi-
The example problems of Art. A4.12 as- bration of structure. On high speed aircraft,
sume that the airplane is a rigid body (suffers air gusts, flight maneuvers and landing re-
no structural deformation). On the basis of actions are applied quite rapidly and thus can
this assumption the applied loads on the air- be classed as dynamic loads. Therefore when
plane in either flight or landing conditions these dynamic loads strike a fleXible (non-
are placed in eqUilibrium with the inertia rigid) airplane cantilever Wing, a rather large
forces which occur due to the acceleration of wing deflection is produced and the wing tends
the airplane. It is obVious that an airplane to Vibrate. This Vibration therefore causes
structure like any other structure is not a additional accelerations of the mass units of
rigid body, particularly a cantilever wing which the wing which means ad.ditional inertia forces
undergoes rather large bending deflections in on the wing. Furthermore if the time rate of
both flight and landing conditions. Figure application of the external applied forces
A4.21 shows a composite photograph taken of a approaches the natural bending frequencies of
test wing for the Boeing B-47 airplane. The the wing, the Vibration excited can produce
maximum upward and downward deflections shown large additional wing stresses.
A4.13
Up until World War II practically all air- for applying the load on the wing when striking
planes were assumed as rigid bodies for struc- the air gust.
tural design purposes. During the war failure NACA Technical Note 2424 reports the flight
of aircraft occured under load conditions which test results on a twin-engine Martin transport
the conventional design procedure based on rigid airplane. Strain gages were placed at various
body analysis, indicated satisfactory or safe points on the wing structure, and strains were
stresses. The failures were no doubt due to read for various gust conditions for Which the
dynamic overstress because the airplane is not normal airplane accelerations were also recorded.
a rigid body. Then slow pUll-up maneuvers were run to give
Furthermore, airplane design progress has similar airplane normal accelerations. The wing
resulted in thin wings and relatively large had a natural frequency of 3.8 cps and the air-
wing spans, and in many cases these wings carry plane speed was 250 M.P.H. Two of the con-
concentrated masses, such as, power plants, clusions given in this report are: - (1) The
bombs, wing tip fuel tanks etc,. Thus the bending strains per unit normal acceleration
flexibility of wings have increased which means under air gusts were approximately 20 percent
the natural bending frequencies have decreased. higher than those of slow pull-Ups for all mea-
This fact together With the fact that airplane suring positions and flight conditions of the
speeds have greatly increased and thus cause tests, and (2) The dynamic component of the wing
air gust loads to be applied more rapidly, or bending strains appeared to be due primarily to
the loading is becoming more dynamic in char- excitation of the fundamental wing bending mode.
acter and thus the overall load effect on the These results thus indicate that air gusts
wing structure is appreciable and cannot be apply a air load more rapidly to a wing than a
neglected in the strength design of the wing. maneuver load giving the same airplane normal
acceleration for a commercial transport type of
General Dynamic Effect of Air Forces on airplane, and thus the dynamic strain effect on
Wing Loads. the wing is more pronounced for gust conditions.
Figs. A4.23, 24 and 25 show results of dy-
The critical airloads on an airplane are namic effect of air gusts on a large wing as de-
caused by maneuvering the airplane by the pilot termined by Bisplinghoff*. The results in these
or in striking a transverse air gust. A trans- figures show that dynamic effects tend to con-
port airplane does not have to be designed for siderably increase wing forces on some portions
sharp maneuvers producing high airplane accel- of the wing and decrease it on other portions.
erations in its job of transporting passengers,
thus the time of applying the maneuver loads is Fig. A4.23
considerably more than a fighter type airplane -600
pUlling up sharply from high speeds. .r Comparative Shear
1\ Distribution
Fig. A4.22 shows the result of a pull-Up :; -400
maneuver on the Douglas D.C.3 airplane at 180
M.P.H. relative to load factor versus time of
CIl
~
", ',- ~ -\
" r-, - - Dynamic Analysis
- - - Rigid Airplane
-200 Analysis
application of load. As indicated the peak I'
load of load factor 3.25 was obtained at the -.......::
No..
end of one second of time. o .2 .4 .6 .8 1.0
- / I'-. I'\. Fig. A4.24

-30
/ "" Fig. A4.22 ....s::

CIl -20
"
~
Comparative Bending
Moment Distribution
.5 1 1.5 2
6
0
)1
"- ~
-10
Pull-up of DC-3 Airplane at 180 mph. I%l ~ "-
f""-.. "-
The author estimates the natural frequency of 0
.2 .4 .6 .8 1.0
the D.C.3 wing to be around 10 to 15 cycles per
second, thus a loading time of 1 second against Fig. A4.25
a time of 1/10 or 1/15 for half a wing deflec- ~ c. Comparative Distri-
-3
tion cycle indicates that dynamic overstress ~ bution of Torque
sqould not be appreciable. In general, it can
be said that dynamic over-stress under maneu-
-2
~C' -1 -- - tl/l About Elastic Axis.
S
~ General Data:-
vering loads on transport airplanes is not as ~ 0 '" '" I
,/"",'
-- ~
Wing Span = 189 ft.
great as from other conditions such as air Eo-< 1 '" Gross wt. = 184000 lb.
t....
gusts or landing. 2 Airlane Vel. =260 mph.
o .2 .4 .6 .8 .10
Dynamic Effect of Air Gusts. Fraction of Semi-Span
The higher the air gust velocity and the * Report on an Investigation on Stresses in Aircraft Struc-
higher the airplane velocity, the less the time tures under Dynamic Loading. M. I. T. Publication.
It has also been found that landing loads w = 15000 lb.

applied through the conventional landing gear or
by water pressure on a flying boat are applied
rapid enough to be classed as dynamic loads and
such loads applied to wings of large span pro-
duce dynamic stresses which cannot be neglected
in the safe design of such structures.
A4. 14 General Conclusions on Influence of Dynamic I \
Loading on Structural Design of Airplane.

The advent of the turbo-jet and the rocket r-- 90"
Fig. A4.26
type engines has opened up a range of possible
airplane airspeeds hardly dreamed of only a few zontal deceleration and the stopping distance
years ago, and already trans-sonic and super- for the airplane?
sonic speed airplanes are a common development.
From an aerodynamic standpoint such speeds have (3). The flying patrol boat in Fig. A4.27
dictated a thin airfoil section which has thus makes a water landing with the resultant bottom
promoted a high density wing. ThuS for air- water pressure of 250,000 lb. as shown in the
planes with appreciable wing spans like Mili- figure. Assume lift and tail loads as shown.
tary bombers and near future jet commercial The pitching moment of inertia of the airplane
transports, which usually carry large concen- is 10 million lb. sec." in. Determine the air-
trated masses on the wing such as engines, fuel plane pitching acceleration. What is the total
tanks etc., the assumption that the airplane is load on the crew member who weighs 200 lb. and
a rigid body is not sufficiently accurate enough is located in a seat at the rear end of the
because the dynamic stresses are appreciab~e. hUll?
The calculation of the dynamic loading on 1000 lb.
the wing requires that the mass and stiffness Fig. A4. 27 25000 lb.
distribution of the wing structure be known.
Since these factors are not known when the
structural design of a wing is started, the
20" --I L .
g. of Crew
----r.....::'....----..:::::.=.--e;;-?IG " Member
20
general procedure in design would be to first -""",,~~ -x----
base the design on the assumption that the wing
is a rigid body plus correction factors based on 500"~
past design experience or available research in- W = 50000
formation to approximately take care of the in-
fluence of the elastic wing on the airplane
aerodynamic characteristics and the build up
dynamic inertia forces. With the wing thus in- (4). The jet-plane in Fig. A4.28 is diving at
itially designed by this procedure, it then can a speed of 600 M.P.H. when pilot starts a 8g
be checked by a complete dynamic analysis and pull-out. Weight of airplane is 16,000 lb.
modif~ed as the results dictate and then re- Assume that engine thrust and total airplane
calculated for the modified elastic wing. This drag are equal, opposite and colinear.
procedure is now practical because of the avail-
ability of high speed computors. (a) Find radius of flight path at start of
pUll-out.
A4. 15 PROBLEMS. (b) Find inertia force in Z direction.
(c) Find lift L and tail load T.
(1). The airplane in Fig. A4.26 is being
launched from the deck of an aircraft carrier by
the cable pUll T which gives the airplane a for-
ward acceleration of 3.25g. The gross weight of
the airplane is 15,000 lb.
;:,
(a) Find the tension load T in the launching '"
.....
cable, and the wheel reactions R1 and
R...
(b) If the flying speed is 75 M.P.H., what
launching distance is required and the
launching time t?
(2). Assume the airplane of Fig. A4.26 is
landing at 75 M.P.H. on a runway and brakes are Fig. A4.28
applied to the rear wheels equal to 04 of the
vertical rear wheel reaction. What is the hori-
CHAPTER AS
BEAMS - SHEAR AND MOMENTS
A5.1 Introduction. P = 100 lb.

In general, a structural member that sup- 30" ---i-10"--..J
ports loads perpendicular to its longitudional al I
axis is referred to as a beam. The structure of
aircraft provides excellent examples of beam
units, such as the wing and fuselage. Very 25 a~15"
seldom do bending forces act alone on a major 75 lb.
aircraft structural unit, but are accompanied by Fig. A5.1
g
axial and torsional forces. However, the bend- 100=P
ing forces and the resulting beam stresses due 10 -1 :9
to bending of the beam are usually of primary It)
importance in the design of the beam structure. C<I

II
>"'1
A5.2 Statically Determinate and Statically Indeterminate Fig. A5. 2 a Fig.A5.3
Beams. 75 75
A beam can be considered as subjected to
known applied loads and unknown supporting re-
actions. If the distribution of the applied 100 = 1
known loads to the supporting reactions can be
determined from the conditions of static equil- :9~ CTI~100=P
ibrium alone, namely, the summation of forces
and moments equal zero, then the beam is con-
sidered as a statically determinate beam. How-
klH
la
::e 75
T
d >"'
la 75
ever, if the distribution of the known applied Fig. A5.4 Fig. A5.5
loads to the supporting beam reactions is in-
fluenced by the behavior of the beam material right side portion as a free body in equilibrium
during the loading, then the supporting reactions as shown in Fig. A5.2. For static equilibrium,
cannot be found by the statical equilibrium ~V, ~H and ZM must equal zero for all forces and
equations alone, and the beam is classified as a moments acting on this beam portion. Consider-
statically indeterminate beam. To solve such a ing ~V = 0 in Fig. A5.3: -
beam, other conditions of fact based on the
beam deformations must be used in combination ~V = 75 - 100 =- 25 lb. (1)
With the static equilibrium equations.
thUS, under the forces shown, the force system
A5.3 SheaL_and Bending Moment. is unbalanced in the V direction, and therefore
A given beam is subjected to a certain ap- an internal resisting force Vi equal to 25 lb.
plied known loading. The beam reactions to hold must have existed on section a-a to produce
the beam in static equilibrium are then calcu- equilibrium of forces in the V direction. Fig.
lated by the necessary equations of static equi- A5.3 shows the resisting shear force, Vi =
librium, namely: - 25 Ib, which must exist for equilibrium.
~V =0, or the algebraic summation of all verti- Considering ZM =
0 in Fig. A5.3, take
cal forces equals zero. moments about some point 0 on section a-a,
~H = 0, or the algebraic summation of all hori- ZM o =- 75 x 15 + 100 x 5 =- 625 in.lb.(2)
zontal forces equals zero.
ZM = 0, or the algebraic summation of all the or an unbalanced moment of - 625 tends to ro-
moments equals zero. tate the portion of the beam about section a-a.
A counteracting resisting moment M = 625 must
With the entire beam in static equilibrium, exist on section a-a to provide equilibrium.
it follows that every portion of the beam must Fig. A5.4 shows the free body with the Vi and
likewise be in static equilibrium. Now consider
the beam in Fig. A5.1. The known applied load Mi acting.
Now ~H must equal zero. The external
of P = 100 lb. is held in equilibrium by the two forces as well as the internal resisting shear
reactions of 25 and 75 Ibs. as shown and are
calcUlated from simple statics. (Beam weight is Vi have no horizontal components. Therefore,
neglected in this problem). Now consider the the internal forces producing the resisting
beam as cut at section a-a and consider the moment Mi must be such as to have no horizontal
A5.1
A5.2 BEAMS -- SHEAR AND MOMENTS
unbalanced force, which means that the resisting rOOO lb. 3°~1
1,500 lb. lb.
moment Mi in the form of a couple, as shown in ,1 ~,3 4,5 6,7 8,
Fig. A5.5, or Mi =
Cd or Td and T must equal C
~ 5'
6 '7
~
to make ,ZH .. O.
~
2 '3
The tendency of the loads and reactions 4' 5' 3'
acting on a beam to shear or move one portion of RA = 1110 lb. RB = 690 lb.
a beam up or down relative to the adjacent por-
tion of the beam is called the External Vertical Fig. A5.6
Shear, or commonly referred to as the beam Vert-
ical Shear and is represented by the term V. Calculations for Shear Diagram: -
From equation (1), the Vertical Shear at We start at the left end of the beam.
any section of a beam can be defined as the al- Considering a section just to the right of the
gebraic sum bf all the forces and reactions 500 lb. load, or section 1-1, and considering
acting to one side of the section at which the the portion to the left of the section, the
shear is desired. If the portion of the beam to Vertical Shear at 1-1 = ,ZV =..;; 500 (negative,
the left of the section tends to move up rela- down on left.)
tive to the right portion, the sign of the + 610 lb. + 610 lb.
Vertical Shear is taken as positive shear and
negative if the tendency is opposite. Or in ~OOlb.
other words, if the algebraic sum of the forces
~
is up on the left or down on the right side,
then the Vertical Shear is positive, and nega- _ 500 500 lb. - 390 lb. - 390 lb.
tive for down on the left and up on the right.
Fig. A5.7 (Shear Diagram)
From equation (2), the Bending Moment at
any section of a beam can be defined as the al- ~in.lb.
gebraic sum of the moments of all the forces
acting to either side of the section about the
section. If this bending moment tends to pro-
duce compression (Shortening) of the upper fib-
ers and tension (stretching) of the lower fibers
of the beam, the bending moment is classed as a
positive bending moment, and negative for the
reverse condition.
A5.4 Shear and Moment Diagrams.
In aircraft design, a large proportion of
the beams are tapered in depth and section, and
also carry a variable distributed load. Thus,
to design or check the various sections of such
beams, it is necessary to have a complete pic-
ture as to the value of the vertical shear and
bending moment at all sections along the beam.
If these values are plotted as ordinates from a
base line, the resulting curves are referred to
as Shear and Moment diagrams. A few example
Shear and Moment diagrams will be plotted, to
refresh the' students knowledge regarding these
diagrams.
Example Problem 1-
Draw a shear and bending moment diagram for
the beam shown in Fig. A5.6. Neglect the weight
of the beam.
In general, the first step is to determine
the reactions.
To find RB, take moments about point A.
ZMA =-4 x 500 + 1000 x 5 + 300 x 13 - 10RB = 0
hence RB .. 690 lb.
,ZV=- 500 + RA - 1000 - 300 + 690= 0
hence RA = 1110 lb.
A5.3
Section 8-8: Vx =
270 - lOx, and hence, the shear de-
ZV =- 300 (positive shear). creases at aconstant rate of 10 Ib./in. from 270
at A to 180 at C.
Fig. A5.7 shows the plotted values on the shear The vertical shear at section D, just to
diagram. the right of load is,
Calculation of the Moment Diagram. VD =
ZVleft =270 - 10 x 9 - 120 =60 up,
Start at section 1-1, and consider the or positive.
forces to the left only:
ZM =-500 x 0 0 = I--
9" -l120l/=P .
C.D w=10lb./m.
Since sections 2-2 and 3-3 are only a differen- At!!l!}!ll!)llli
tial distance apart, assume a section just above • 36
RA and consider the forces on the left side only: RA=270 RB=210
ZM =-500 x 4 =-
2000 in. lb. (Negative
270 lb. Fig. A5.9
VC=180 lb.
moment, because of tension in the top fibers).
Consider the section under the 1000 in. lb. load~ VD=60 lb.
ZM to left =-
500 x 9 + 1110 x 5 = 1050 in.
lb. (positive moment, compressing the top
t= 15"
f--6"
Shear Diagram -210 lb.
fibers).
Check by considering the forces to the right:
flg~
.n ....
o
in
0
....
.... ....
ZM right = 300 x 8 - 690 x 5 = - 1050 in. lb. ~~
Next, consider a section over RB:

ZM right =300 x 3 =
900 in. lb. (Negative The vertical shear between points D and B, when
moment, tension in top fibers). x is the distance of any section between D and
B from A:
At Section 8-8:
ZM right =300 x 0 0 = VDB =270 - 120 - lOx - (1)
Fig. A5.8 shows the plotted values. At point B, x =36:
From the above results it may be noticed hence VB =
270 - 120 - 10 x 36 210 lb., =-
that when the bending moment is obtained from which checks the reaction RB'
the forces that lie to the left of any section, Since the Vertical Shear decreases at a
the bending moment is positive when it is clock- rate of 10 Ib/in. from D to B, it will be 6"
wise. If obtained from the forces to the right, from D to a point where the shear is zero, since
it is positive, when counter-clockwise. The the shear at D is 60 lb.
student should sketch in the approximate shape This point could alsG be located by equa-
of the deflected structure and determine the ting equation (1) to zero and solving for x as
signs from whether tension or compression eXists follows:
in the upper and lower fibers.
o = 270 - 120 - lOx, or x = 150
10
15" fram =
Example Problem 2 A.
Calculate and draw the shear and moment If the shear diagram has passed through
diagrams for the beam and loading as shown in zero under the concentrated load, then the
Fig. A5.9. method of equating the shear equation to zero
First, determine the reactions, RA and RB: - and solving for x could not be used, thus in
general, it is best to draw a shear diagram to
ZMA = 36 x 10 x 18 + 120 x 9 - 36RB = O. find when shear is zero. Fig. A5.10 shows the
hence RB = 210 lb. plotted shear diagram.
ZV =-120 - 36 x 10 + 210 + RA O.=
hence RA =270 lb. Moment Diagram: -
At section A just to the right of reaction
Shear Diagram: - RA the bending moment, considering the forces to
The vertical shear just to the right of the the left, is zero, since the arm of RA is zero.
reaction at A is equal to 270 up, or positive. The bending moment at any section between
This is plotted as line AE in Fig. A5.10. The A and C, at a distance x from the left reaction
vertical shear at section C just to the left of RA , is,
the load and considering the forces to the left
of the section = 270 - 9 x 10 = 180 lb. up, or (2)
positive. The vertical shear for any section
between A and C at a distance x from A is: In equation (2), x can not be greater than
9.
JX 2 Vdx
The equation for the bending moment between
D and B (x greater than 9) is M2
Mx =RAx - P (X-9) -
wx 2
2""" (3 ) J M1
dM '"
x1
Thus, the area of the shear diagram between any
:: 270 x - 120 (x-9) _ 1~2 two points equals the change in bending moment
between these two points.
= 1080 + 150 x - 5x 2 - - - (4) To illustrate this relationship, consider
the shear diagram in example problem 2 (Fig.
At section C, x :: 9n , substitute in equation (4) A5.10). The change in bending moment between
the left reaction RA and the load is equal to
Me = =
1080 + 150 x 9 - 5 X 9 2 2025 in. lb. the area of the shear diagram between these two
(positive, compression in top fibers). points, or
At the point of zero shear, x :: 15 n • 270 + 180
2 x 9 :: 2025 in. lb. Since the bending
=
M 1080 + 150 x 15 - 5 x 15 2 :: 2205
moment at the left support is zero, this change
Thus, by sUbstituting in equation (2) and therefore equals the true moment at a section
(4) the moment diagram as plotted in Fig. A5.11 under the load P.
is obtained. Adding to this the area of the small tri-
angle between point D and the point of zero
A5.5 Section of Maximum Bending Moment. 60
The general expression for the bending mo- shear, or 1f x 6 = 180, we obtain 2205 in. lb.
ment on the beam of example problem 2 is from as the maximum moment. This can be checked by
equation (3); taking the area of the shear diagram between
the point of zero shear and point B =
2~0 x 21 :: 2205 in. lb.
Now, the value of x that will make MX a
maximum or minimum is the value that will make Example Problem 3.
the first derivative of Mx with respect to x Fig. A5.12 illustrates a landing gear oleo
equal to zero, or strut ADEO braced by struts BD and CEo A land-
ing ground load of 15000 lb. is applied through
the wheel axle as shown. Let it be required to
(5 ) find the axial load in all members and the shear
and bendIng moment diagram for the oleo strut.
Therefore, the value of x that will make Mx a
maximum or minimum may be found from the equation
RA - P - wx :: 0 HBT
B
But, observation of this equation indicates 10"
that the term Ra - P - wx is the shear for the
section at a distance x from the left reaction.
Therefore, where the shear is zero, the bending
moment is maximum. Thus, the shear diagram
---t 16"
which shows where the shear is zero is a con- Fig. A5.12
venient medium for locating the points of maxi-
___L
Pins at Points B,
mum bending moment. e, E. No Vertical 15000 lb. 1
Resistance at Point 16"
A5.6 Relation Between Shear and Bending Moment. A.
Equation (5) can also be written ~ :: V, n,-----'L
"0"
since the right hand portion of equation (5) is
equal to the shear.
Hence, dM :: Vdx - - - - - - - - - - - - (6) SOLUTION: -
Which means that the difference dM between To find Vc take moments about point B,
the bending moments at two sections that are a 2MB =- 15000 x 0.5 x 42 + 15000 x 0.866 x 5.77
+ 20.77 Vc :: O.
distance dx apart, is equal to the area Vdx
under the shear curve between the two sections. hence, Vc :: 11550 lb.
Thus, for two sections x 1 and x 2, The axial load in member CEtherefore equals
A5.5
11550/cos 30 ° or 13330 lb. :. He = 11550 x 15/26

= 6660 r::===l--- 500
To find HA take moments about point D,
ZMD =
11550 x 15 - 6660 x 10 - 15000 x 0.5 x 32
+ 10 HA O.= ';'-----I.~---'-'___._------'-'---~>--HB
hence, HA =
13340 lb.
To find VB take ~FV =0,
~FV = 15000 cos 30° + 11550 - VB = O. Fig. A5.14
hence, VB = 24550 lb.
The axial load in member BD therefore equals
24550/cos 30° =
28360 lb. (compression). The SOLillION: -
reaction HB therefore equals 28360 x sin 30 =
Calculations of reactions at A and B: -
14180 lb. To find VB take moments about point A,
Fig. A5.13 shows the oleo strut as a free
body With the reactions at A, D and E as calcu- ZMA =-
500 x 7 - 500 x 6 + 1000 x 20 + 1000 x
lated. Fig. A5.13 also shows the axial load, sin 45° x 10 + 1000 cos 45° x 2 - 22 VB O. =
vertical shear and bending moment diagrams. hence, VB = 999.3 lb. (up).
The bending moments due to applied loads
Without regard to bending deformation of the To find VA take ~V = 0,
beam are usually referred to as the primary ~V = 999.3 - 1000 - 1000 sin 45° 500 + VA = O.
bending moments. If a member carries axial hence, VA = 1207.8 lb. (up).
loads additional bending moments will be pro-
duced due to the axial loads times the lateral To find HB take ~ = 0,
deflection of the beam, and these bending mo- ~H =- 500+ 1000 cos 45° - HB 0, hence =
ments are usually referred to as secondary bend- HB = 207.l.
ing moments. (Arts. A23-30 covers the calcula- With the exception of the 1000 lb. load at 45°,
tion of secondary moments). all loads are applied to brackets which in
turn are fastened to the beam. Therefore the
next step is to find the reaction of the loaded
brackets at the beam centerline support points.
The load at E and the reaction HB at B will be
also referred to beam centerline.
Fig. A5.15 (a,b,c,d) show the cantilever
brackets as free bodies. The reactions at the
base of these cantilevers will be determined.
These reactions reversed will then be the applied
~
Axial Load
-13000 lb. Diagram loads to the beam at points C, D, F and E.
W/00/& '--_-=-=Compression
13340
~"""",,_~';::am
1
T5
r}'"
~ ~OO, ~<I. <: >
2 " ; f J 500
~-840 8" 8" I /'J) 5"
-7500 lb.
Bending
M
= 3000
HC=50
C=500
L HO=500
MO=4000
H -500
F-
L
E
~'10'1.1
\\"E'"
VF=1000 VE=7071
~Oiagram Moment
Fig. a Fig. b Fig. c
Fig. A5.15
Fig. d
Fig. A5. 13 -120000 -133440 in. lb.
Example Problem 4. For bracket at e, to find He take ~ 0, =

Fig. A5.14 shows a beam loaded with both or obviously He = 500 lb. In like manner use
transverse and longitudional loads. This beam ~V = 0 to find V = 500 lb. To find Me take
loading is typical of interior beams in the air- e
plane fuselage which support all kinds of fixed
moments about point C. ZMC =-
500 x 2 + 500 x
equipment. The reactions for the beam are at
8 - Me =
0, hence Me =
3000 in. lb. The student
points A and B. Required: - Shear and bending should check the reactions at the base of the
moment diagrams. cantilever brackets at D and F (See Fig. b,c).
The load of 1000 at 45° and applied at point E" have maximum peak moments without the Vertical
will be referred to point E the centerline of Shear passing through zero. To illustrate
beam. Fig. d shows the reaction at E due to the this fact, consider the beam of Fig. A5.20,
load at E'. The reaction at B should also be namely, a simple supported beam with an ex-
referred to the beam centerline. Fig. A5.16 ternally applied couple moment of 10 in. lb.
shows the beam with the applied loads at points magnitude at point C the center point of the
C D E' F and B'. Figs. A5.17, 18 and 19 show beam. The shear and bending moment diagrams
the axial load, vertical shear and bending mo- are as indicated and a maximum bending moment
ment diagrams under the beam loading of Fig. occurs at C but the shear diagram does not pass
A5.16. through zero.
707.1 1000
MC=10"#
Fig. A5.16 ."I~ISh,,",Dia.

. Bending
-500 lb. -707.1 Axial Load Dia. Flg. A5.20 -5"# Moment Dia.
Fig. A5. l7IL _11__

L-.Q.0 -='-:=20::::7=:':=:J1
707. 8,-,l=b,,-. _ A couple is two equal and opposite forces
O. 7 Shear Dia. not in the same straight line. Let it be as-
Fl'.A5.18.d sumed that the 10 in. lb. couple is made up of
4285
I
.999.J-r 5788.7
forces equal to 100 lb. each and an arm between
them of 0.1 inch as illustrated in Fig. A5.21.
Bending ____ 1¥..:..~.A.1'_' _

Moment Dia.
-207.1
------'-----t 1
1-1 lb.
Fig.A5.19
1CI ======u Shear Dia.
Fig. A5.21 99
The shear diagram is determined in the same
manner as explained before. The applied exter- The shear diagram is as shown in Fig. A5.21 and
nal couples do not enter into the vertical shear now passes through zero under each of the couple
calculations. The bending moment diagram can be forces. Thus if we assume the couple moment has
calculated by taking the algebraic sum of all a dx arm the shear to the right of C is one lb.
couples and moments of all forces lying to the and then changes to some unknown negative value
one side of a particular section. If it is de- and then back to one lb. positive as the dist-
sired to use the area of the shear diagram to ance dx is covered ih going to the left. Thus
obtain the bending moments, it is necessary to the shear goes to zero twice in the region of
add the couple moments to the shear areas to ob- point C.
tain the true bending moment. For example, the
bending moment just to the left of point E will A5.7 Moment Diagrams as Made up of Parts.
be equal in magnitude to the area of the shear In calculating the deflection of statically
diagram between C and E plus the sum of all ap- determinate beams (See Chapter A7) and solving
plied couple moments between C and E but not in- statically indeterminate structures (See Chapter
cluding that at E. A8), the area under the bending moment curve is
To illustrate the calculations are: - required, thus it is often convenient to treat
(- 500 x 5) + (707.8 x 10) 4578 in. lb. (from = each load and reaction as a separate acting
force and draw the moment diagram for each force.
area of shear diagram).
(3000 - 4000) =-
1000 in. lb. (from sum of The true bending moment at a particular point
will then equal the algebraic summation of the
couple moments).
Thus bending moment at E .. 4578 - 1000 = ordinates of all the various moment curves at
3578 in. lb. left this particular point or adding the various
separate moment diagrams will give the true
The bending moment at Eright will equal that at bending moment diagram. Figs. A5.22 and A5.23
Eleft plus the couple moment at E or 3578 + 707 illustrate the drawing of the bending moment
:: 4285 in. lb. diagram in parts. In these examples, we start
from the left end and proceed to the right end
The student should realize that when couple and draw the moment curve for each force as
moments are applied to a beam it is possible to though the beam was a cantilever with the fixed
A5.7
at B and thus leaving only 3 unknown elements

P=lO* of the reaction at B. Fig. A5.25 shows the
I .w=10 lb/in. bending moment curves for each load acting sep-
L-J:-Lu * «...LLt arately on this cantilever' frame. Fig. A5.26
15" r-2''f--10'' -.:r shows the true bending moment as the summation
R~=10 lb. Re=100 R~=62 R =48 e
D~:!:!:ue!:..lo:to~P=====:::J of the various moment curves of Fig. A5.25.
D~1500 - J- 120 As another solution of this fixed ended
in. lb.
Due to --=:::::::::::::-====--::--
-, 000 Duet~620 frame, one could assume the statically deter-
P~ minate modification as a frame pinned at A and
Due to Pe -=::El500 Due to w -==;;::::::j 500 pinned with rollers at B as illustrated in Fig.
=wx e/e - A5.27. This assumed structure is statically
500 ~ determinate because there Pe =10
...............
Final Moment Dia.
............ ~~~ are only 3 unknown elements,
namely the magnitude and di- P" =10
Final Moment Curve
rection of the reaction at A
Fig. A5. 22 Fig. A5. 23 and the magnitude of the re- P~=10
action at B. For conven-
support at the right end. The final bending ience the reaction at A is
moment curve for the true given beam then equals resolved into two magnitudes
the sum of these separate diagrams as illustra- as H and V components. The B=20
ted in the figures. reactions VA' HA and VB can
then be found by statics and
STATIC MOMENT CURVES IN SOLVING STATICALLY the results are shown on Fig. A5.27. Fig. A5.2B
shows the bending moment diagram on this frame
INDETERMINATE STRUCTURES due to each load or reaction acting separately,
The usual procedure in solving a statically starting at A and going clockwise to B. Fig.
indeterminate structure is to first make the A5.29 shows the true bending moment diagram as
structure statically determinate by removing the the summation of the separate diagrams.
necessary redundant or unknown reactions and -60, -=:J,__~~-Due to P e
then calculating the deflection of this assumed
statically determinate structure as one step in ,",~e
v
\ojJ:--Du~
A
-100
the overall solution of the problem (See Chapter
AS). In the solution of such structures it is
likewise convenient to treat the bending moment
diagram as made up of parts. To illustrate,
Fig. A5.24 shows a loaded rectangular frame
fixed at points A and B. The reactions at both
Due to P e
P e =lO* -=:::::::::J -50
f-5' 5,,-1 Due to P~
-601 1-60
r Fig. A5. 28
i1nli~O~
P,,=lO*
Final Bending
-60 Moment Diagram
-6,~ -+....!£i Free 60 -50 90 (Tension on inside
of frame is posi-
Fig. A5.25 tive moment)
A5. 8 Forces at a Section in Terms of Forces at a

100 Previous Station.
STRAIGHT BEAMS.
points A and B are unknown in magnitude, di- Aircraft structures present many beams
rection and location, or each reaction has 3 un- which carry a varying distributed load. Mini-
known elements or a total of 6 unknowns for the mum structural weight is of paramount importance
two reactions. With 3 static equilibrium in aircraft structural design thus it is de-
equations available, the structure is statically sirable to have the complete bending moment
indeterminate to the third degree. Fig. A5.25 diagram for the structure so that each portion
illustrates one manner in which the structure of the structure can be proportioned effic-
can be made statically determinate, by freeing iently. To decrease the amount of numerical
the end A to make a bent cantilever beam fixed work required in obtaining the complete shear
and bending moment diagrams it usually saves

time to express the shear and moment at a given
station in terms of the shear and moment at a
previous station plus the effect of any loads
lying between these two stations. To illustrate,
Fig. A5.30 shows a cantilever beam carrying a
considerable number of transverse loads F of dif-
ferent magnitudes. Fig. A5.31 shows a free body
Q
V F~-"
~~ t- a ~v Fig. A5. 32 Fig. A5. 33
M~c~nM"
(1) (2)
Fig. A5.31
Then from Fig. A5.33 we can write for the
resultant forces and moment at point (2) at
of the beam portion between stations 1 and 2. station 2: -
The Vertical Shear V~ at station 1 equals the
summation of the forces to the left of station 1 V" =V~ + F~-"
and M~ the bending moment at station 1 equals
the algebraic sum of the moments of all forces
lying to left of station 1 about station 1.
Now considering station 2: - The Vertical
Shear Va = V~ + F~_", or stated in words, the
Shear V" equals the Shear at the previous sta- Having the resultant forces and moments for a
tion 1 plus the algebraic sum of all forces F given point on a given station, it is usually
lying between stations 1 and 2. Again consider- necessary in finding beam stresses to resolve
ing Fig. A5.31, the bending moment M" at station the forces into components normal and parallel
2 can be written, M" =
M~ + V~d + F~_"a, or
to the beam cross-section and also transfer
their location to a point on the neutral axis
stated in words, the bending moment Ma at sta-
tion 2 is. equal to the bending moment M~ at a of the beam cross-section.
previous station 1, plus the Shear V at the For example Fig. A5.34 shows the resultant
previous station 1 times the arm d, the dist-
ance between stations 1 and 2 plus the moments
of all forces lying between stations 1 and 2
about station 2.
A5.9 Equations for Curved Beams.
Many structural beams carry both longitud-
ional and transverse loads and also the beams
may be made of straight elements to form a frame
~ S
Mo=M~-Ne
or all beam elements may be curved to form a Fig.A5.34 Fig.A5.35 Fig.A5.36
curved frame or ring. For example the airplane
fuselage ring is a curved beam subjected to
forces of varying magnitude and direction along
its boundary due to the action of the fuselage
skin forces on the frame. Since the complete
bending moment diagram is usually desirable, it forces and moment at point 1 of a beam cross-
is desirable to minimize the amount of numerical section. They can be resolved into a normal
work in obtaining the complete shear and bending force N and a shear force S plus a moment M~ as
moment values. Fig. A5.32 shows a curved beam shown in Fig. A5.35 where,
loaded with a number of different vertical loads
F and horizontal loads Q. Fig. A5.33 shows the N = H cos a + V sin a
beam portion 1-2 cut out as a free body. H~ S =V cos a - H sin a
represents the resultant horizontal force at
station 1 and equals the algebraic summation of Later on when the beam section is being de-
all the Q forces to the left of station 1. V~ signed it may be found that the neutral axis
represents the resultant vertical force at
station 1 and equals the sum of all F forces to
lies at point °
instead of point 1. Fig.
A5.36 shows the forces and moments referred to
left of station 1, and M~ equals the bending point 0, With Mo being equal to M~ - Ne.
moment about point (1) on station 1 due to the
moments of all forces lying to the left of
point 1.
A5.9
AS. 10 Torsional Moments.

The loads which cause only bending of a
beam are located so that their line Of action
passes through the flexural axis of the beam.
Quite often, the loading on a beam does not act
through the flexural axis of the beam and thus
the beam undergoes both bending and twisting.
The moments which cause the twisting action are
usually referred to as torsional moments. The
airplane wing is an excellent example of a beam
structure that is subjected to combined bending
and torsion. Since the center of pressure of
the airfoil forces changes with angle of attack,
and since there are many flight conditions it is
impossible to eliminate torsional moments under
all conditions of flight and landing. For the
fuselage, the Vertical tail surfaces is norm-
ally located above the fuselage and thus a load
on this tail unit causes combined bending and
twisting of the fuselage.
Fig. A5.34 illustrates a cantilever tube
being subjected to a load P acting at point A on
a fitting attached to the tube end. The flex-
velocity With the maneuvering limited to certain

Fig. A5.34 Fig. A5. 35 maximum accelerations. These limiting acceler-
ations are usually specified With reference to
ural axis coincides with the tUbe centerline, or the X Y Z axes of the airplane. Since the di-
axis 1-1. Fig. A5.35 shows the load P being rections of the lift and drag forces change With
moved to the point (0) on the tube axis 1-1, angle of attack it is simpler and convenient in
however the original force P had a moment about stress analysis to resolve all forces With ref-
(0) equal to Pr, thus the moment Pr must be erence to the X Y Z axes which remain fixed in
added to the load P acting at (0) if the force direction relative to the airplane.
system at point (0) is to be equivalent to the As a time saving element in wing stress
original force P at point A. The force P acting analysis, it is customary to make unit load an-
through (0) causes bending without twist and the alysis for wing shears and moments. The wing
moment Pr causes twisting only. shears and moments for any design condition
For the resolution of moments into various then follows as a matter of simple proportion
resultann planes of action, the student should and addition. For example it is customary: -
refer to any textbook on statics.
(1) To assume a total arbitrary unit load act-
A5.11 ShearS-and Moments on Wing. ing on the wing in the Z direction through
Arts. A4.5 and A4.6 of Chapter A4 discusses the aerodynamic center of the airfoil
the airloads on the wing and the equilibrium of section and distributed spanwise according
the airplane as a whole in flight. As explain- to that of the CL or lift coefficient.
ed, it is customary to replace the distributed (2) A similar total load as in (1) but acting
air forces on an airfoil by two resultant in the X direction.
forces, namely, lift and drag forces acting
through the aerodynamic center of the airfoil (3) To assume a total unit wing load acting in
plus a wing moment. The airflow around a wing the Z direction through the aerodynamic
is not uniform in the spanwise direction, thus center and distributed spanwise according
the airfoil force coefficients CL, CD and CM to that of the CD or drag coefficient.
vary spanwise along the wing. Fig. A5.36 shows (4) Same as (3) but acting in the X direction.
a typical spanwise variation of the CL and CD
force coefficients in terms of a uniform span- (5) To assume a unit total wing moment and
wise variation CL and CD' distributed spanwise according to that of
Any particular type of airplane is designed the Cm_ or moment coefficient.
~·c
to carry out a certain job or duty and to do The above unit load conditions are for con-
that job requires a certain maximum airplane ditions of acceleration in translation of the
AS. 10 BEAMS -- SHEAR AND MOMENTS
airplane as a rigid body. Unit load analyses ratio or the spanwise variation of the lift
are also made for angular accelerations of the coeff!cient CL in terms of a uniform distribu-
airplane which can also occur in flight and tion CL • In this example we have taken this
landing maneuvers. ratio as unity since we have no wind tunnel or
The subject of the calculation of loads on aerodynamic calculations for this wing relative
the airplane is far too large to cover in a to the spanwise distribution of the lift force
structures book. This subject is usually cover- coefficient. In an actual problem involving an
ed in a separate course in most aeronautical airplane a curve such as that given in Fig.
curricula after a student has had initial A5.36 would be available and the values to
courses in aerodynamics and structures. To il- place in Column 3 of Table A5.1 would be read
lustrate the type of problem that is encountered from such a curve. Column (2) gives the wing
in the calculation of the applied loads on the chord length at each station. Column (4)
airplane, simplified problems concerning the gives the wing running load per inch of span
wing and fuselage will be given. at each station point. Since a total unit
load of 17760 lb. was assumed acting on the
AS.12 Example Problem of Calculating Wing Shears and half wing and since the wing area is 17760 sq.
Moments for One Unit Load Condition. in., the running load per inch at any station
Fig. A5.37 shows the half wing planform of equals the wing chord length at that station.
a cantilever wing. Fig. A5.38 shows a wing In order to find shears and moments at the
section at station O. The reference Y axis has various station points, the distributed load is
been taken as the 40 percent chord line which now broken down into concentrated loads which
happens to be a straight line in this particular are equal to the distributed load on a strip
wing layout. and this concentrated strip load is taken as
acting through the center of gravity of this
distributed strip load. Columns 5, 6, and 7
show the calculations for determining" the
(~pz) strip loads. Column 8 shows the lo-
cation of the ~Pz load which is at the centroid
of a trapezoidal distributed load whose end
values are given in Column (4). In determin-
ing these centroid locations it is convenient
to use Table A3.4 of Chapter A3.
The values of the shear Vz and the mo-
ment Mx at each station are calculated by the
method explained in Art. A5.8. Columns 9, 10,
11 and 12 of Table A5.1 give the calculations.
For example, the value of Mx = 9884 in Col-
~ IZ Ref. Axis umn (12) for station 220-equals 2436, the Mx
moment at the previous station in Column (12)
~S% s%l--....-r40% line plUS 4908 in Column (10) which is the shear
aerodYll!!,mic I at the previous station (230) times the dist-
center --a.c.- 0 -"Ref. - ance 10 inches plus the moment 2540 in Column
I I?/
I
96"
Ref. Axis Fig. AS. 38
(9) due to the strip load between stations
230 and 220, which gives a total of 9884 the
value in Column (12).
The strip loads ~Pz act through the
The total wing area is 17760 sq. in. For aerodynamic center (a.c.) of each airfoil strip.
convenience a total unit distributed load of Column (13) and (14) give the x arms which is
17760 Ibs. will be assumed acting on the half the distance from the a.c. to the reference Y
wing and acting upward in the Z direction and axis. (See Fig. A5.38). Column 15 gives the
through the airfoil aerodynamic center. The M moment for each strip load and Column 16
spanwise distribution of this load will be ac- the My moment at the various stations which
cording to the (C L ) lift coefficient spanwise
distribution. For simplicity in this example equals the summation of the strip moments as
it will be assumed constant. one progresses from station 240 to zero.
Table A5.1 shows the calculations in table Fig. A5.39 shows the results at station
form for determining the (Vz ) the wing shear in (0) as taken from Table A5.1.
the Z direction, the bending moment Mx or mo- i51
ment about the X axis and My the moment about
the Y axis for a number of stations between the V z=17760/
sJ¢iy;~_o--_x
wing tip station 240 and the centerline station
O. x
Column 1 of the table shows the number of
stations selected. Column 3 shows the CL/
CL CMy=214071 Fig. AS. 39
A5.11
TABLE A5.1
CALCULATION OF WING SHEAR Vz AND WING MOMENTS M" AND My
DUE TO TOTAL UNIT DISTRIDUTED HALF WING LOAD OF 17760 LBS.,
ACTING UPWARD IN Z DIRECTION AND APPLIED AT AERODYNAMIC
CENTERS OF WING SECTIONS. (See Figs. A5.37, 38 for Wing Layout)
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
......
:g -aU .s
8~"
0
c: ;g 8
.
0
CL .s .s 'c:" N "
I:! + 0
... ~'"....
;-
a:a
e~
... r.. ...
..c:
~
I CL
Ratio,
og~
O'C:
..::I ..c:",•
§
~
~
iii
.c
d-
'" 0
I:
p,
<l~
og...:
It>
'"
uo.
0-
"'/li
-a
g
00
-a
g
'"0" ~~
"t" . ",u
u- '"
.9'
" u
1;;3
...."
Q>
. -a H 0
NS
p,-
g '" '" '" Assumed OJ)
::rr C> ~~~ ~g <l~
....'0
• Q)
00 0.;':
c:u~ ..::I bIlu.c '" > - c:
Q", ..::I~ 8 N "
-ct:" :al:.. " '"
°:alli
0I~
N8iO
.5l;
~12i
J B,s
" u
-c..c: Unity. .!'LS..::I
§"-:'
~
"-c ·t II IJ'" 0. ...
:':p'-c
<1
:;<l ~-a
<lu N
'0
• II
N
• ~'S ai-c ·a
H
N
.,..
•
II 8 ....
".
... Gl os:: <J II cIj
~N
P, >< '0 u P,'0 ~fI)"O
~i5~ UU ~~~ <.3
'" 01
<llll ~ II ,,'a~
"
N~
> II <lg ::l!u+

II
lC~><
... "" >01
~cIj oS >e <lg ::l!IIU
240 48 1.0 48 0 0 0 0 7.2 0

48.54 5.0 242.7 2.49 7.28
235 49.09 1.0 49.09 242.7 0 605 605 7.36 1767 1767
49.63 5.0 248.1 2.49 7.44
230 50.18 1.0 50.18 490.8 1213 618 2436 7.53 1846 3613
51. 27 10.00 512.7 4.97 7.69
220 52.36 1.0 52.36 1003.5 4908 2540 9884 7.85 3940 7553
54.00 15.00 810.0 7.43 8.09
205 55.64 1.0 55.64 1813.5 15052 6020 30956 8.34 6540 14093
57 27 15.00 859 0 7 43 8.59
190 58.92 1.0 58.92 2672.5 27020 6490 64466 8.85 7390 21483
60.56 15.00 908.4 7.43 9.09
175 62.20 1.0 62.20 3580.9 40087 6760 111313 9.33 8260 29743
63.84 15.00 957.6 7.43 9.57
160 65.48 1.0 65.48 4538.5 53713 7100 172126 9.81 9170 38913
67.13 15.00 1007.0 7.43 10.06
145 68.78 1.0 68.78 5545.5 68077 7480 247683 10.31 10100 49013
70.39 15.00 1055.5 7.43 10.55
130 72.00 1.0 72.00 6601. 0 83782 7840 338705 10.80 11100 60113
73.62 15.00 1104.3 7.43 11. 04
115 75.23 1.0 75.23 7705.3 99015 8200 445920 11.28 11480 71593
76.89 15.00 1153.4 7.43 11. 54
100 78.55 1.0 78.55 1118587 115579 8570 570060 11.80 13270 84863
80.18 15.00 1202.7 7.43 12.04
85 81. 81 1.0 81. 81 10061 132880 8940 702949 12.28 14480 99343
83.46 15.00 1252.2 7.43 12.52
70 85.10 1.0 85.10 11313 150915 9310 863174 12.76 15600 114943
86.73 15.00 1300.9 7.43 13.01
55 88.35 1.0 88.35 12614 169700 9660 1042534 13.27 16930 131873
89.98 15.00 1349.7 7.42 13.49
40 91. 62 1.0 91.62 13964 189210 10000 1241744 13.71 18200 150073
93.81 20.00 187/1.3 9.92 14.05
20 96.00 1.0 96.00 15840 <179<180 18tiOO 1539624 14.40 26350 176423
96.00 20.00 1920.0 10.00 14.40
0 96.00 1.0 96.00 17760 316806 1920Q 1875630 14.40 27648 214071
Sum = 17760 Checks Total Limit Load Assumed on Half Wing.
When the time comes to deSign the structural forces are the major forces. For the body load
make-up of a cross-section to withstand these analysis the direct air pressures are secondary,
applied shears and moments, the structural de- the major forces being of a concentrated nature
signer may Wish to refer the forces to another in the form of loads or reactions from units
Y axis as for example one that passes through attached to the body, as the power plant, wing,
the shear center of the given section. This landing gear, tail, etc. In addition, since the
transfer of a force system With reference to an- body usually serves as the load carrying medium,
other set of axes presents no difficulty. important forces are produced on the body in re-
sisting the inertia forces of the weight of the
SHEARS AND MOMENTS ON AIRPLANE BODY interior equipment, installations, pay load etc.
A5.13 Introduction.
As in the case of the wing, a large part
The body of an airplane acts essentially as of the load analysis can be made without much
a beam and in some conditions of flight or land- consideration as to the structural analysis of
the body. The load analysis of an airplane body
ing as a beam column which may be also subjected involves a large amount of calculation, and
to twisting or torsional forces. Thus to design
thus the treatment in this chapter must be of a
an airplane body requires a complete picture of simplified nature, and is presented chiefly for
the shearing, bending, twisting and axial forces the purpose of showing the student in general
which may be encountered in flight or landing.
In the load analysis for wings, the direct air how the problem of load analysis for an air-
plane body is approached.
A5.14 Design Conditions and Design Weights. A5.15 Body Weight and Balance Distribution.
The airplane body must be designed to with- The resisting inertia forces due to the
stand all loads from specified flight conditions dead weight of the body and its contents plays
for both maneuver and gust conditions. Since an important part in the load analysis for the
accelerations due to air gusts vary inversely airplane body. When the initIal aerodynamic
as the airplane weight, it customary to analyze and general layout and arrangement of the air-
or check the body for a light load condition for plane is made, it is neceBsary that a complete
flight conditions. In general, the design weight and balance estimate of the airplane be
weights are specified by the government agen- made. This estimate is usually made by an en-
cies. For landing conditions, however, the gineer from the weight control section of the
normal gross weight is used since it would be engineering department who has had experience
more critical than a lightly loaded condition. in estimatinp, the weight and distribution of
The general design conditions which are airplane units. This estimate which is pre-
usually investigated in the design of the body sented in report form gives the weights and
are as follows: (e.g.) locations of all major airplane units
Flight Conditions: or installations as well as for many of the
minor units which make up these major airplane
H.A.A. (High angle of attack) assemblies or installations. This weight and
L.A.A. (Low angle of attack) balance report forms the basis for the dead
LL.A.A. (Inverted low angle of attack) weight inertia load analysis which forms an
I.H.A.A. (Inverted high angle of attack) important part in the load analysis of the air-
plane body. The use of this weight and balance
The above conditions generally assume only
translational acceleration. In addition, it is estimate will be illustrated in the example
sometimes specified that the forces due to a problem to follow later.
certain angular acceleration of the airplane A5.16 Load Analysis. Unit Analysis.
about the airplane e.g. must be considered. Due to the many design conditions such as
The body is usually required to withstand
special tail loads both symmetrical and unsym- those listed in Art. A5.14, the general pro-
metrical which may be produced by air gusts, cedure in the load analysis of an airplane body
engine forces, etc. Also, the body should be is to base it on a series of unit analyses.
checked for forces due to unsymmetrical air The loads for any particular design condition
loads on the wing. then follows as a certain combination of the
unit results with the proper multiplying fac-
Landing Conditions: tors. A simplified example problem follows
In general, the body is investigated for the which illustrates this unit method of approach.
following landing conditions. The detailed re- A5.17 Example Problem lllustrating the Calculation of
qUirements for each condition are given in the Shears and Moments on Fuselage Due to Unit
government specifications for both military and Load Conditions.
commercial airplanes.
Fig. A5.40 and A5.41 shows a layout of the
Landplanes: Level landing. airplane body to b e used in this example prob-
Level landing With side load. lem. It happens to be the body of an actual
Three point landing. airplane and the wing used in the previous ex-
Three point landing with ample problem was the wing that went with the
ground loop. airplane.
-+!~
Nose over or turn over
condition.
.~ 1:,1,y-'~--1~ne"=~
Arresting. (Usually for only
Navy Carrier based air- 1
planes) • N'
QI
I 11'2"
Seaplanes or Boats: C.l
c:
Step landing With and Without ~ II
angular acceleration.
Bow landing. ~J, Fig.A5.40
-.1
!1
Stern landing. I~.-------- 27'2"
Two wave landing.
Beaching conditions.
Catapulting conditions (Navy
airplanes).
SpecIal Conditions or Forces:
Towing of airplane.
I
Front View I
Body supercharging. I I
~47., ' Fig. A5.41 I
20' - 0 " ' - - - - - - _ .
A5.13
Table A5.2 gives the Weight and Balance TABLE A5.3

estimate for the total airplane. This table is
usually formulated by the Weight and Balance WEIGHT AND BALANCE OF AmPLANE LESS WING GROUP
AND INSTALLATIONS IN AND ON WING
Section of the engineering department and it is
necessary to have this information before the { z equal arm from thrust line. + is up.
Ref. Axes: x is distance from z Ref. Axis 5" forward of
airplane load analysis can be made. propeller. • is aft.
Horiz. Hortz. Vert. Vert.

Item Name Weight Arm Mom. Arm Mom.
TABLE A5.2 w
>. x wx z wz
'6.
AmPLANE WEIGHT AND BALANCE S 1 Powerplant group 1100 19 20900 0 0
rzl
2 Fuselage group 350 113.5 39700 1 350
Vert. (Z) arms measured from thrust line
~
Reference
Axis {(+ is up)
Horiz. (X) arms measured from Z axis 5"
forward prop. t (+ is aft)
-
4
5
6
8
Tail group
Surface controls
Electrical system
Tail Wheel group
110
85
130
35
287
127
61
306
31550
10800
7930
10700
24
-14
4
-10
2640
-1190
520
- 350
9 Furnishings 220 116 25520 5 1100
Item
Name W=
Horiz.
Arm
Boriz.
Moment
Vert.
Arm
Vert.
Mom. ~
..:l
10 Radio
Weight empty =
125
2155
181 22600
169700
10
~
4320
No. Wt. iI (X) (in.) wx (Z) wz
i 11
12
Pilot
Student
~
200 151
99
30200
19800
4
4
800
800
1
2
Power Plant
Fuselage Group
1100
350
19
113.5
20900
39700
0
1
0
350
'"
::> Gross weight = ~ 2555 ~ 219700 ~ -mo
3 Wing Group 750 97 72750 -18 -13500 Empty x • 169700 • 78 . 7" Z = 4320 = 2 . 00"
4 Tail Group 110 287 31550 24 2640 2I55 2155
5 Surface Controls 85 127 10800 -14 - 1190
6 Electrical System 130 61 7930 4 520 With useful load x = 219700
2555
= 86 . 0" Z = 5920
2555
• 2 32"
•
7 Chassis Front 235 70 16450 -52 -12200
8 Tail Wheel Group 35 306 10700 -10 - 350
9 Furnishings 220 116 25520 5 1100
10 Radio 125 101 22600 10 1250
Weight empty = 3150 258900 -21380 weight distribution should study the inboard
11 Pilot 200 151 30200 4 800 profile drawing of the airplane Which shows the
12 Student 200 99 19800 4 800
13 Fuel System 760 89 67600 -27 -20500 general arrangement of all the installations and
Gross weight = 4300 376500 -40280 equipment. Furthermore, he should stUdy the
Calculation of C. G. locations: overall structural arrangement as to its possi-
Gross wt. x = 376500/4300 • 98.5" aft of Ref. Axis ble influence on fuselage weight distribution.
z = -40280/4300 = 9.4" below thrust line
The whole process involves considerable common
sense if a good apprOXimation to the weight dis-
SOLUTION: tribution is to be obtained. Fortunately the
large dead weight loads, such as the power
WEIGHT AND BALANCE OF BODY ITEMS. plant, tail, etc. are definitely located, thus
WEIGHT DISTRIBUTION. small errors in the distribution of the minor
distributed weights does not change the over-
Table A5.3 gives the weight and balance all shears and moments an appreciable amount.
calculations for all items attached to fuselage In order to obtain reasonable accuracy, the
or carried in the fuselage, except the wing and fuselage or body is divided into a series of
items attached to the wing as the front landing stations or sections. In Fig. A5.42, the sec-
gear and the fuel. tions selected are designed as stations which
In order to obtain a close approximation to represent the distance from the Z reference
the true shears and moments on the fuselage due axis. The general problem is to distribute the
to the dead weight inertia loads, it is neces- concentrated loads as shown in Fig. A5.43 into
sary to distribute the weights of the various an equivalent system acting at the various
items as given in Table A5.3. Fig. A5.42 shows fuselage station points.
a side view of the airplane with the center of Obviously, if a weight item from Table A5.3,
gravity locations of the weight items of Table represents a concentrated load such as a pilot,
A5.3 indicated by the (+) signs. In the various stUdent, radio, etc., the weight can be dis-
design conditions, the direction of the weight tributed to adjacent station points inversely
inertia forces changes, thus it is convenient as the distance of the weight (e.g.) from these
and customary to resolve the inertia forces into adjacent stations. However, for a weight item
X and Z components. ThUS, in Fig. A5.43, the such as the fuselage structure (Item 2 of Table
weights as given in Table A5.3 are assumed act- A5.3) whose e.g. location causes it to fall be-
ing in the Z direction through their (e.g.) lo- tween stations 80 and 120 of Fig. A5.43, it
cations. The loads as shown would not give a would obViously be wrong to distribute this
true picture as to the shears and moments along weight only to the two adjacent stations since
the fuselage, thus these loads should be dis- the weight of 350# is for the entire fuselage.
tributed in a manner which should simulate the This weight item of 350# should thus be dis-
actual weight distribution. In most weight and tributed to all station points. The controlling
balance reports, the weight items are broken requirement on this distribution is that the
down into considerable more detail than that moment of the distributed system about the ref-
shown in Table A5.3, which makes the weight dis- erence axes must equal the moment of the orig-
tribution more eVident. The person making the inal weight about the same axes. Fig. A5.44
WEIGHT DISTRIBUTION TO FUSELAGE STATIONS
2 STATION
o 11 50 80 120 170 200 230 011 120 170 200 230 260 290
I
I I 1 I 'I I I
I
I I I I I
~ (11 1[ _ ' ,66 ~1 Li-2~2:9'~~~1111~tl00'_-,--I~~~

• - -+ - ._- - --+- + - - - + - - - - -
I
noo
____
~--,-------'--..I
200 ~O
-.---'....:..---.-~350~.---
200....!5
I I
_ ---
I
~ +5
I
I
Fig. A5.42. Location of weight items of Table A5. 3. Fig. A5.46. Weight items from Table A5. 3 acting in X direction.
o 11 50 80 120 170 200 230 290 011 50 80 120 170 200 230 290
I I I
I I I I 'I I I I I i I I I
I I I I I I I I 1.,.--1_-'-...1 I I
-~ -~l----J --fg-k
...-f
...-f
M
...-f
Oan
N en
lit>co "" ...
il9
t -
I
... 0
L4"
h(ff ~,; L
8
65 ~52
-
-1-30 LI0 .£..:10-<:6
----r6,-; rp.- fJ.,,~,,-
I I
•5 U
I
I
I I I I I II! I I I I
Fig. A5.43. Weight items of Table A5. 3 acting in Z direction. Fig. A5. 47. Vertical distribution of fuselage dead weight.
011 120 170 200 230 290 50 80 120 170 200 230 260 290 315
I I ,
I I I I I I I I I I I I
I I I .,.--1_-'-...1 I I
I I
--l·---r--t· -+-!
1
I I
"
9# 520#" 312#" 30#" 10'# 10"# 3"# 0#
19# 80# 78# 65# 30# 9# .&80
3 # W4#
--e-65# -.~5# t-30~10~1~i;)-0#
I I I I I
I I ! I I I I
Fig. A5.44. Results of fuselage weight distribution to stations. Fig. A5.48. Fuselage weight referred to X axis plus couples.
120 170 200 230 290 011 50 80 120 170 200 230 260 290 315
I I
I I I I I I I I I I I I I
.;r---'I_--'-......I I I I I I.;r---'I_--'-......I I I I
19 8 758 682 1650 490 10 27 --·-2j4?
:3 -C..,.388G-307-(~409-(?-311G-76G-I0 (-+21 (1):]~O
I I I I
I I I! I I I
Fig. A5.45. Final weight distribution to station points. Fig. A5.49. Final weight distribution in X direction referred
to X axis plus proper couples.
A5 15
shows how the dead weight of 350* was distribu- manner as to have the same resultant e.g. 10-
ted to the various station points considering cation as the original weight system which is
the weights to be acting in the Z direction. illustrated in Fig. A5.46. Fig. A5.47 illus-
Table A5.4 shows the results of this sta- trates how the fuselage weight distributed
tion point weight distribution for the weight system as shown in Fig. A5.44 is distributed
items of Table A5.3. The values in the hori- in the vertical direction at the various
zontal rows opposite each weight item shows the station points so that the moment of this sys-
distribution to the various fuselage stations. tem about the X axis is equal to that of the
The summation of the weights in each vertical original fuselage weight of 350#. For con-
column at each station point as given in the venience, these distributed fuselage weights
third horizontal row from the bottom of the can be transferred to the X axis plus a moment
table gives the final station point weight. as shown in Fig. A5.48.
These weights are Shown in Fig. A5.45 for Table A5.4 shows the vertical distribution
weights acting in the Z direction. The moment of the various items at the various station
of each total station load about the Z axis is points. The bottom horizontal row gives the
given in the second horizontal row from the moment about the X axes of the loads at each
bottom of Table A5.4. The summation of the station point, Which equals the individual
moments in this row must equal the total WX loads times their Z distances. The summation
moments of Table A5.3 or 219700"#. This check of the values in this horizontal row must equal
is shown in the last vertical column of Table the total wz moment of Table A5.3. This check
A5.4. is shown at the bottom of the last vertical
The distributed system must also be distri- column. Fig. A5.49 shows the results as given
buted in the Z or vertical direction in such a in Table A5.4 for the weight distribution in
the X direction.
TABLE A5.4
PANEL POINT WEIGHT DISTRIBUTION

(Monocoque Type of Fuselage)
Station No. = distance x from Ref. z Axis
Item Wt. 11 50 80 120 170 200 230 260 290 315

Name
No. W w z w z w z w z w z w z w z w z w z w z
,
.
I Powerplant group 1100 874 0 226 0 1100
Fuselae::e group "00 .0 -4 ',. -0 00 0 0 0 ou >u >u 0 '.0 "00
4 a . eroup
5 I Surface contro s
10
85
'" ,4 "9 '4 110
73 - 4 - 4 85
6 Electrical s stern 130 82 4 48 4 130
•
9
Ta' whee l 2:roup
Furnishiruzs
"5
220 76 5 100 44 5
13 -10 22 -10
220
10 Radio 125 79 10 46 10 125
II Pilot 00 76 4 124 4 200
12 Student 200 105 4 95 4 200
Total w 2555 893 .02 388 .02 307 2.47 409 1. 67 311 5.3 76 6.45 10 1.0 21 13lO6 118 19.1 22 -10 2555
Horizontal Moment = Z wx = 9820 19390 24550 49100 52800 15200 2300 5460 34200 6920 219700
Vertiea! Moment = Z wz= 19 8 758 682 1650 490 10 274 2249 - 220 5920
x : distance of station from z reference axis which is 5" forward of propeller.
z : distance above or below thrust line or x axis.
A5.18 Unit Analysis for Fuselage Shears and Moments. (1) Unit acceleration or load factor in Z
Since there are many flight and landing direction and acting up.
conditions, considerable time can be saved if a
unit analysis is made for the fuselage shears, (2) Unit acceleration or load factor in X
axial and bending forces. The design values in direction and acting forward.
general then follow as a summation of the values (3) Unit tail load normal to X axis acting
in the unit analysis times a proper multiplica- down.
tion factor.
The loads on the fuselage in general con- Unit analyses are also usually carried out
sists of tail loads, engine loads, wing re- for engine thrust and engine torque, side load
actions, landing gear reactions if attached to on tail and angular acceleration, but to keep
fuselage and inertia forces due to the airplane the example calculations from becoming too
acceleration which may be due to both transla- lengthy only the above 3 unit conditions will
tional and angular acceleration of the airplane. be carried out in detail. The others will be
For simpl1city, these loads can be resolved into discussed in detail in later paragraphs.
components parallel to the Z and X axes.
To illustrate the unit analysis procedure, Solution for Unit Load Factor in Z Direction.
a unit analysis for our example problem will be
carried out for the following unit conditions: Fig. A5.50 shows the dead weight loads
acting in the Z direction as taken from Table

TABLE A5.5
A5.4 or Fig. A5.45. The wing is attached to the
fuselage at stations 73 and 116 as shown on Fig. FUSELAGE SHEARS AND MOMENTS FOR ONE
A5.50. The fittings at these points are assumed LOAD FACTOR IN Z DffiECTION
as designed to cause all the drag or reaction in (Dead Weight Acting Down)
the X direction to be taken off entirely at the
front fitting on station 73. 1 2 3 4 5 6
To place the fuselage in equilibrium, the Load or V = shear !:J.x = Dist. !:J.M Moment
wing reaction will be calculated: Sta. Reaction =Z w between = V!:J.x M*
No. w (lbs. ) stations (in. lbs.) (in. lbs.)
ZFx = 0, RH + 0 = 0, hence RH =0 o o
315 ~ o o
L11s tation 0 = 219700 - 116 RR - 73 RF = 0 - - (A) - 22 22
(Note: 219700 from Table A5.3)
I-----I----I-----+-- 25 --+----+-----1
290 +_ o 22
- 550 - 550
ZFz = - 2555 + RF + RR = 0 - - - - - - - -- (B ) - 118 140
1-----1----+-----+- 30 -+---+-------1
260 + o 140
- 4200 - 4750
Solving equations (A) and (B) for RF and - 21 161
o 161
230 ~ - 10 171 - 4830 - 9580
RF =1780 lb., RR = 775 lb. 1----+---+----+- 30 - - f - - - - + - - - - /
o 171
200 ~ -
- 5130
76
-14710
247
Table A5.5 gives the calculations for the 1 - - - - 4 - - - + - - - - - + - - 30 --+----+-----1
fuselage shears and bending moments at the var- o 247
ious station points. 170 ~ - 311 558
- 7410 -22120
I----+---+----+-- 50 -+----1------1
Sta.O 11 50 80 120 170 200 230 260 290 31,5 120 ~
o 558
-27900 -50020
- 409 967
I:
I
:
I
i
I
I I
I
I
I
I I
I
:
o 967
4--+----+-----1
116 ~
I I 1 1
- 3868 -53888
I I I,A'\ I \ I I) 775 192
I ~V \ 22)
+--7619 118 ~ o
I.
I
893 388 307
-~-+-----t--J:: -~.
17"
409
20"
311
~- +-
21
.+-- '- 80 : - 307
192
499 - 6912 -60800
.1 ' - - ~_ • -----_0::--_-"-
1------1----1-----+--- 7 --+----+-----1
~I -l-- --,~
73 ~
o 499
- 3493 -64290
... RH=O 1-" 43" --I, 1780 -1281
P=:I RF=1780 RR=775 Fig. A5.50 1---+---+----+- 23 - - f - - - - + - - - - /
I o -1281
STA.73 STA.116 50 ~ +29463
- 388
-34827 -893
1-----1----+-----+- 39 --+----+-----1
Solution for Unit Load Factor in X Direction. o - 893
11 ~ - 893 o +34827 o
Fig. A5.51 shows the panel point dead (3) Up on left and down on right side of a section is
weight distribution for loads acting in the X positive shear.
direction and aft, as taken from Table A5.4 or (6) Tension in upper fuselage portion is negative bending
Fig. A5.40. To place the fuselage in equili- moment.
brium the wing reactions at points (A) and (B) (1) + refers to aft side of station.
will be calculated. - refers to forward side of station.
* M = M at previous station in col. 6 plus !:J.M in col. 5.
ZFx = 2555 - RH = 0, hence RH = 2555 lb. (forward)
Take moments about point (A)
11 50 80 120 172 200 230 260 290 315
rnA = 2555 x 17 + 5920 - 43 RR = 0, oI I I I I I I
, I
I
hence RR m 1147.8 (up) III I I
I
I :
!,,~A<--'-I----+--.\
I I
I I
I I
I
I
I
(5920 equals the sum of the couples from Table 19 I Y8758 682 1650 490 10 274 2Z49
A5.4. ," (~93~,,~388(-307 (-409 (-311(-76 ~10 (-21 C-1lB~22
~ ----fB.
H=2555 220
ZF z =
1147.8 - RF 0, = I ,'!::--A
43" -tOj' ---
hence RF =
1147.8 lb. (down) i RF=1147.'8 IRR=1147.8 Fig. A5.51
STA.73 STA.116
Table A5.6 gives the calculations for the shears,
moments and axial loads for the loading of Fig.
A5.5l.
A5.17
TABLE A5.6 TABLE A5.7

FUSELAGE SHEARS, MOMENTS" AXIAL LOADS FOR
ONE LOAD FACTOR IN X DIRECTION FUSELAGE SHEARS & MOMENTS FOR UNIT
Inertia Loads Acting Aft HORIZONTAL TAIL LOAD IN Z DIRECTION
I 2 3 4 5 6 7 8 9 (Load Acting Down)
Wx = W = Px = ~ Wx V = ~ W lIX =
sta.
No.
load
inX
z
load in
Axial Shear
.z lIMI
Couple
= Dist.
Moment between
lIM2 :
VlIx
M 1 2 3 4 5 6
Z Dir.
Dir. Load Station
0 0
Load or V = shear lIX = dist. LIM Moment
315 + 0 0
22 0
22
0
220 0
220 sta. Reaction =~ w between
= VlIx in. lbs.
25 w lbs. lbs. stations
290 + 0 22 0 220
0 0 0
118 140 - 2249 - 2029
30 + 0 0 0
315 0
260 + 0
21 0
140
161
0
-
0
274 0
- 2029
- 2303
- 0 0 0
30
25
230 + 0 161 0 - 2303 + 0 0 0
0 0
- 0 290 0
10 171 10
30
- 2313 - 0 0 0
0 171 0 - 2313 12.5
200 + 0 0 0
76 247 - 490 - 2803 + 0 0 0
272.5 0
170 + 0 247 0
30
- 2803
- -100 100 0
311
0
558 0
- 1650
0
- 4453 17.5
50 + 0 100 - 1750
260 - 1750
120 + 0
- 409 0
558
967 0
-
0
682 0
- 4453
- 5135
- 0 100 - 1750
4 30
116+ 0 967 - 5135 + 0 100 - 4750
0 230 - 3000
0 1147.8 967 -1147.8 0
36
- 5135 - 0 100 - 4750
0 0 967 -1147.8 41320 36185 30
80 +
307 0 1274 -1147.8 - 758 0 35427 + 0 100
- 3000
- 7750
200
73 + 0 0 1274 -1147.8 0
7
8035 43462
- 0 100 - 7750
2555 -1147.8 -1281 0 -43435 0 27 30
23 + 0 100 -10750
170 - 3000
50 :
388
0
0
-1281
- 893 0
8
0 0
27
19 - 0 100
50
-10750
39
11 + 893
0
- 893
0
19
0
19 + 0 100 -15750
120 - 5000
. -
0 0 0 0
0 100 -15750
: 17 x (2555) : 43435 4
to refers to aft side of station + 0 100 -16150
Col. 1 - refers to forward side of station 116
- 475.6 -375.6 - 400 -16150
Px is plus for tension in -fuselage.
(Col. 9) M = M at previous station in Col. 9. plus II MI of col. 8 plus 36
lIM2 of col. 8 + 0 -375.6 - 2629
80 13521
- 0 -375.6 - 2629
7
Solution for Unit Horizontal Tail Load Acting Down. + 0 -375.6 0
73 2629
- -375.6 0 0
The fuselaga shears and moments will be 23
computed for a unit tail load of 100 lb. on the + 0 0 0
50
tail acting in the Z direction, with balancing - 0 0 0
39
reactions at the wing attachment points. The + 0 0 0
center of pressure on the horizontal tail is at 11
- 0 0 0
station 277.5. Fig. A5.52 shows the fuselage
loading. To find wing reactions at (A) and (B): (Col. 6) =M =M at previous station in Col. 6 plus
1I M in Col. 5.
ZMA = 100 x (277.5 - 73) - 43 RR = 0,
hence RR = 475.6# (up)
CALCULATION OF APPLIED FUSELAGE SHEARS,
- - 100 + 475.6 - RF - 0,
~Fz
MOMENTS AND AXIAL LOADS FOR A SPECIFIC
hence RF = 375.6# (down)
FLIGHT CONDITION.
Table A5.7 gives the detailed calculations
for the shears and moments at the various sta- Using the reSUlts in Tables A5.5, A5.6, and
tion points. A5.7, the applied shears and moments for a given
o 11 50 80 120 170 200 230 flight condition follow as a matter of propor-
II I I , I I I tion and addition. To illustrate, the applied
I I values for one flight condition will be given.
I' I I I : It will be assumed that the aerodynamic
II L.-r--t-------:--.,.I I I
calculations for this airplane for the (H.A.A.)
high angle of attack condition gave the follow-
ing reSUlts, which the student will have to
accept without knowledge of how they were
obtained.
RF=375.6 RR=475.6
Fig. A5.52
STA.73 STA.116
Applied load factor in Z direction = - 6.0

down
Applied load factor in X direction = 1.333 aft.
Applied tail load = 110 lb. up. -1--- --

I ~C.G.
Thus with the load factors in the Z and X
directions and the tail load known, Table A5.8
can be filled in as illustrated. In a similar Fig. A5.53
manner the values for other flight conditions
can be found, the only difference being a new --------
set of multiplying factors since the applied
loads would be different.
TABLE A5.8
APPLIED FUSELAGE SHEARS, MOMENTS & AXIAL LOADS FOR

FLIGHT CONDITION I. (H.A.A.) The resistance to these X and Z components
I 2 3 4 5 6 7 8 9 10
of the ground reaction R is prOVided by the in-
ertia forces of the airplane in the X and Z di-
V z for V z for V z for App. M for Z M for X M App. Axial
Z Load X Load Tail Shear Load Load for Mom. Load rections.
Sta. Factor Factor Load Vz Factor Factor Tail
Load
M Px
Tables A5.5 and A5.6 show the fuselage
(Ibs. ) (Ibs. ) in. Ibs. in. lbs. in. Ihs. in. lbs. (Ibs. ) shears, moments and axial loads for inertia
315 .. 0 0 0 0 0 0 0 0 0 loads due to one load factor in the Z and X
132 0 0 132 0 294 0 0 29.3
290 t 132 0 0 132 - 3300 294 0 - 3006 29.3 directions respectively. Thus to obtain the
- 840 0 0 840 - 3300 - 2710 0 - 6010 186.8
fuselage forces for this given landing condition,
260 t 840 0 0 840 - 28500 - 2710 1925 - 29285 J86.8
966
966
0
0
-110
-110
856
856
- 28500
- 57480
- 3075
- 3075
1925
5220
- 29650
- 55335
215
215
it is only necessary to mUltiply the values in
230 +
- 1026 0 -110 916 - 57480 - 3090 5220 - 52030 228 these two tables by the proper factor and add
1026 0 -110 910 - 88260 - 0090 8540 - 82810 zz8
200 •
- 1482 0 -110 1372 - 88260 - 3740 8540 - 83460 330 the results.
170 +
-
1482
3348
0
0
-110
-110
1372
3238
-132720
-132720
- 3740
- 6085
11820
11820
-124640
-126985
330
745
Thus fuselage forces due to vertical load
120 + 3348 0 -110 3238 -300120 - 6085 17320 -288855 745 factor of 7 would equal 7 times the values in
- 5802 0 -110 5692 -300120 - 6850 17320 -288650 1290
columns (3) of Table A5.5 to obtain shear and 7
116 .. 5802 0 -110 5692 -323328 - 6850 17770 -312408 1290
- 1152 -1530 413 - 413 -323328 - 6850 17770 -312408 1290
times column 6 to obtain bending moment.
80 • 1152 -1530 413 - 413 -364800 48200 2910 -313690 1290
2994 -1530 413 1877 -364800 47250 2910 -314640 1700 Likewise the forces due to the 2.98 load
73 • 2994 -1530 413 1877 -385740 57900 0 -327840 1700
7686 0 0 -7686 -385740 36 0 -385704 -1710 factor in X direction would equal (-2.98) times
-n86 0 - 686 -208962 36 -208920 -1710
50 •
-5358
0
0 0 -5358 -208962 25
0
0 -208937 -1190 the values in columns (4), (5) and (9) of
II t -5358
0
0
0
0
0
-5358
0
0
0
25
0
0
0
25
0
-1190
0
Table A5.6 to obtain axial loads, shears and
bending moments respectively.
NOTES: (Moments & axial loads are referred to the thrust line).
The final or true forces would be the
Col. 2 - 6 x values in column 3 of Table A5.5.
algebraic sum of these results.
L
Col. 3 - 1. 333 x values in column 5 of Table A5.6.
Col. 4 - -1.10 x values in column 3 of Table' AS. 7.
Col. 5 - column (1) .. column (2) .. column (3).
Col. 6 - 6 x values in column 6 of Table A5.5. Landing with Angular Accelerati~g
Col. 7 - 1. 333 x values in column 9 of Table AS.6.
Col. 8
Col. 9
-
-
1.10 x values in column 6 of Table A5.7.
columns (6) • (7) • (8).
In a level landing condition, it is some-
Col.IO - 1. 333 x values in column (4) of Table A5.6. times specified that the horizontal component
of the ground reaction must be a certain pro-
A5.19 Example of Fuselage Shears and Moments for
portion of the vertical component, which causes
Landing Conditions. the line of action of the ground reaction R in
Fig. A5.53 illustrates the airplane in a Fig. A5.53 to not pass through the c.g. of the
level landing condition. The ground reaction is airplane, which creates an external pitching
assumed to pass the center of landing gear wheel moment on the airplane. This moment is us-
and e.g. of airplane. The fuselage shears, mo- ually balanced by the inertia forces due to
ments and axial loads are reqUired when the the angular acceleration produced by the un-
vertical ultimate load factor is 7. (Gross balanced moment about the C.g. The shears and
weight = 4300#). moments on the fuselage due to t his external
moment could be found as explained in Art.
SOLUTION: A5.20.
The vertical or Z component of the ground A5.20 Inertia Loads Due to Angular Acceleration.
reaction R is specified as 7 load factors which In some of the flying conditions, it is
equals 7 x 4300 =
30100.. One half of this is sometimes specified that the airplane must be
acting on each wheel. subjected to an angUlar acceleration as well as
The horizontal or X component of R is 30100 translational acceleration. This angular ac-
tan 23° =
425 x 30100 12800# and acting aft. = cele~ation of the airplane produces inertia
The horizontal load factor on airplane forces which must be calculated if the air-
equals 12800/4300 = 2.98. plane is to be treated as a body in static
A5.19
equilibrium. In some cases, a tail load due to

a gust on the tail is specified which produces a Fz = My w Xc (1)
Iy
moment about the airplane e.g. which produces
angular acceleration of the airplane. In cer-
tain landing conditions, the ground forces do Fx =My
Iy
w zc - - - - - - - - - - - - - - - (2)
not pass through the airplane e.g. thus pro-
ducing a moment about the e.g. which for stress From Table A5.9, I y = 16097600
analysis purposes is balanced by inertia forces.
Moment of Inertia of Airplane My was assumed as 100,000
The calculation of the moment of inertia of hence
an airplane about the center of gravity axes
was explained on page A3.5 of Chapter A3. A Fz
= 16097600
100000
w Xc
= .00621 w Xc
detailed example solution was given in detail in
Table 6A of Chapter A3. The general equations
for the moments of inertia of the airplane about Fx = .00621 w Zc
the reference axes are:
where zc and Xc are the z and x distances of
Iy =Z wx 8
+ Z WZ8 + Z ~ Iy weight w to the airplane e.g.
Columns No. 9 and 10 of Table A5.9 gives
Ix =Z wy8 + Z wz 2 + Z ~ Ix the values of these inertia components. Fig.
A5.54 shows these inertia loads applied to the
I z =Z wy8 + Z wx 8 + Z ~ Iz fuselage. The reactions at wing attachment
points should be computed and then a table of
The last term in each of the above equa- fuselage shears, moments and axial loads should
tions represents the moment of inertia of each be made up. This unit table could then be used
weight item about its own centroidal axes par- for all conditions involving angular acceler-
allel to the reference axes. ation of the airplane.
It should be realized that the inertia
A5. 21 Solution for Inertia Loads Due to Unit 100,000 In. resisting loads in Table A5.9 are only approxi-
Lbs. Pitching Moment. mately, since the moment of inertia neglects
To illustrate the general procedure of de- the centroidal moment of inertia of the big
termining the balancing inertia loads when the items, such as the power plant, wing, etc. The
airplane is subjected to an unbalanced moment example is only for the purpose)of illustrating
about the e.g., an analysis will be made for a the general procedure of determining the inertia
unit 100,000 in. lb. moment. Table A5.9 gives resisting loads due to angular acceleration.
the necessary calculations. The same general procedure can be followed in
considering unbalanced external moments about
From kinetics: the Z and X axes, commonly referred to as yaw-
ing and rolling moments.
Pitching angular acceleration a =My
Iy
(rad/sec. 2 )
TABLE A5.9
where BALANCING INERTIA FORCES FOR UNIT 100,000 IN. LB. MOMENT
ABOUT Y AXIS THROUGH AIRPLANE C. G. (PITCHING MOMENT)
= unbalanced external pitching moment 1 2 3 4 5 6 7 8 9 10
about e.g. of airplane.
Fuselage Sla. 2 ly • Inert. Inert.
2
Iy = pitching moment of inertia of air- Panel sta.
No.
WI.
w
Arm
Z
Arm
X
Arm
Zc
Arm
Xc
Xc t Zc
W(Z~ • X~)
Force Force
Fz Fx
plane about airplane e.g. = Zwr 2 11 893 .02 11 9.42 - 76.5 5949 5300000 -428 -52.1
50 388 .02 50 9.42 - 37. 5 1494 580000 - 90.6 -22.6
80 307 2.47 80 11.87 - 7.5 197 60500 - 14.3 -22.6
The tangential inertia force F for a mass 120 409 1.67 120 11.07 32.5 1182 484000 82.5 -28.2
wig due to an angular acceleration a equals, 170
200
311
76
5.30
8.45
170
200
14.70 82.5
15.85 112.50
7016
12950
2180000
985000
159.0
53.2
-28.4
- 7.5
230 10 1.00 230 10.40 142.50 20480 204600 8.9 - 0.6
=-wgra' =-My
260 21 13.06 260 22.46 172.50 30270 635000 22.5 - 2.9
F but a Iy
g 290
315
118 19.10
22 -10.0
290
315
28.50 202.50
- 0.6 227.50
31085
51800
3670000
1140000
148.3
31. 2
-20.9
- 0.1
Wing 750 -18 97 - 8.6 9.50 164 123000 44.3 40.2

hence Front Chassis 235 -52
Fuel 760 -27
70
89
-42.6
-17.6
- 17.50
1. 50
2120
312
498000
237500
- 25.8
7.1
62.3
83.2
My
F =r y w r, where r is the distance from Iy ' 16097600
Columns (2), (3), &: (4) Taken from Table A5.2.

Column (5) Zc = Z - 9.4 J See Table A5.2 for c. g.
the weight w to the airplane e.g. Column (6) Xc • X - 87.5 location
Column (9) F z •• 00621 WXc
Column (10) F • . 00621 wZc
It is convenient to treat the inertia force
F as resolved into two components Fx and Fz •
hence,
170 230
III. Fig. A5.62 shows the plan form of a canti-
11 50 80 120
I I I I lever wing. Assume a constant normal
I I I distributed load on the surface equal to
I I
I I I 50 lb./sq.ft. Write expressions for shear
and bending moment on wing and find values
52. at 25, 100, 150 and 200 inches from end.
426
~O"
I
I f..g
100"
IX:_--"p=la:::n=fo=-r-m---- .~ 30" Fig. A5. 62
Fig. A5.54 L L -- _\20"
A5.22 Problems
I
1 - < - .- 200"
- - - .
(I) Draw the shear, bending moment and axial

load diagrams for loaded structures in
Figs. 55 to 60 c ' T
90"
Fig. A5.63
Planform
(5
o
IX: Leading Edge
150"
x
1. 01 t 1 f f t t f~ Relative Span-
~O"~ wise Distribution
Fig. A5.63 shows plan form of a cantilever wing.

The total distributed air load normal to surface
is 10000 lb, The relative spanwise distribution
is shown. Take center of pressure at 24 percent
400 (57)
of chord from leading edge. Divide wing into
~60o 10 inch width strips and calculate Vz , Mx and
~c:=;;:=:======±t====:;:::'J- ~ My. and plot curves for same.
;: -r"r 12"-+'" --t3~ 500 lb.

IV. Fig. A5,64 shows an externally braced
monoplane wing. Take an average wing lift
load of 90 lb./sq.ft. normal to wing with
t ~;00 lb. (58) 100. (59) center of pressure at 27 percent of the
~b + 100 To 100 chord from leading edge of wing and calcu-
~ kIt" ~ -100 6"
late and draw the front and rear beam pri-
mary shears and bending moment diagrams.
-\5" f-.- 15"----1 L...J.0''--1
1
1001 hl::o-i~l.~t;oo
20 100
r1~~~------~~~~~---
n
·n
" 44 T
~'/l"I\~" ]~." 1" 60 27 PLAN VIEW
r-lO"-1
I1 k!Iinge Front Beam
60a 60b 60 c
4 '------ --J
/. 100" 'I' 50" ----1-

(II) Draw bending moment diagram for structures Pin eFt'B:::.~R=I.=B==. =
and loading in Fig. A5.61, abc.
t400#
II Fig. A5. 64
30~M8" 100
i 11
(a) 16"
L
1--16"~
Fig. A5. 61
A5.21
BENDING MOMENTS - BEAM - COLUMN ACTION*
A5.23 Introduction deflections will gradually converge and the memb-

er will reach a state of equilibrium. These
A beam-column is a member subjected to
transverse loads or end moments plus axial loads. secondary bending moments could be found by suc-
cessive steps by the various deflection princi-
The transverse loading, or end moments, produces
bending moments which, in turn, produce lateral ples given in Chapter A7. However, for prismatic
beams this convergency can be expressed as a
bending deflection of the member. The axial mathematical series and thus save much time over
loads produce secondary bending moments due to
the axial load times this lateral deflection. the above successive step method. For members of
variable moment of inertia, the secondary moments
Compressive axial loads tend to increase the will usually have to be found by successive steps.
primary transverse bending moments, where as If the end loads P are tension, they will
tensile axial loads tend to decrease them. tend to decrease the primary moments; thUS, in
Beam-column members are qUite common in general, the case of axial compression is more
airplane structures. For example, the beams of important in practical design, since buckling
externally braced wing and tail surfaces are and instability enter into the problem.
typical examples, the air loads producing trans-
verse beam loads and the struts introducing ax-
ial beam loads. In landing gears, one member is A5. 25 Equations for a Compressive Axially Loaded
Strut with Uniformly Distributed Side Load.
usually subjected to large bending and axial
loads. In tubular fuselage trusses, lateral Fig. A5.66 shows a prismatic beam of length
loads due to installations supported on members L subjected to a concentric compressive load P
between truss joints produce beam-column action. and a uniformly transverse distributed load W,
In general, beam column members in airplane With the beam supported laterally at each end,
structures are comparatively long and slender and with end restraining moments M1 and M2 • It
compared to those in buildings and bridges; is assumed that the general conditions for the
thUS, the secondary bending moments due to the beam theory hold, namely; that plane ~ections
axial loads are frequently of considerable pro- remain plane after bending; that stress is pro-
portion and need to be considered in the design portional to strain in both tension and compres-
of the members. sion.
This chapter deale briefly on the theory At any point a distance x from the beam end,
of single span beam-column members. A summary the moment expression is,
of equations and design tables is included to-
wx 2
gether With examples of their use. The informa- M== M1 + (M2 -L M1 ) x - wLx ""2 +""2 - py - - - - - (A5 .1 )
tion in this chapter is used frequently in other
chapters where practical analysis and design of From applied mechanics, we know that
beam-column members is considered. For a com- M - EI d 2 y
pleted and comprehensive treatment of beam-col- - dx 2 therefore, differentiating equa-
umn theory and derivation of equations, see tion (A5.1) tWice with respect to x gives
Niles and Newell-"Airplane structures".
2
A5.2' General Action of a Member Subjected to d M P
- 2 + - M == w - - - - - - - - (A5.2)
Combined Axial and Transverse Loads. dx EI
Sub-figure a of Fig. A5.65 shows a member w w Fig. A5.65
subjected to transverse loads Wand axial com- !---.t! p(a)
I - - L-------j
r-
pressive loads P. The transverse loads W pro- P 'l-
duce a primary bending distribution on the memb- ~J..'J.ernont Benc!ins Ml F:3X
(b) p~\" ' • , f ,
I M2p
•
er as shOwn in Fig. b. This bending will pro-
duce a transverse deflection curve as illustrat- p-r<-:::.-------J.-=..::-'I-P(c) 112-~ w~Run-
I ~r;"'!'J Derledie~ L -T I mng Load
ed in Fig. c. The end loads P now produce an
additional secondary bending moment due to the ......-z7Z7~ . ~d) ~ElaStiC Curve
st ec.on Cl!J Benda.,
end load P times the deflection 0 , or the bend- 6a t\om""t p - -- - - - - - - -p
ing moment diagram of Fig. d. This first sec- P-~-P(e) 1
ondary moment distribution produces the addi- , Def/~til", due t~15t S"CDndQ'y Moment Fig. A5.66
tional lateral deflection curve of Fig. e and
For simplification, let j =Vl'; hence EI
'lIEf. P 1
j2'
the end load P will again produce further bend-
ing moments due to this deflection. If the ax- which, substituted in (A5.2), gives
ial load is not too large, these successive
d2 M 1
* Also discussed in Chapter C2, Part E. dx2 + j2 M == W
A5.22 BEAM - - COLUMNS
The solution of this differential equation gives. 1 M M wLx wx 2

Y =P (M 1 + ra ~ 1 X - "2 +"""'2
j j
M=C 1 sin + Cra cos + wjra - - - - (A5.3)
L
where C1 and Cra are constants of integration and Dra - D1 cos j x
sin x and cos x are the limits of an infinite L sin j
j j sin j
series of variable x. When x = 0, M = M1 and (A5.7a)
j
when x = L, M = Mra , therefore: The slope of the elastic curve at any point
is given by the first derivative of equation
C - Mra - wjra M1 - wjra (A5.7a)
sin L tan L
1 (Mra - ~ wL ~)
1 -
j j i = P --L- - 2" +WX -
C1 x Cra
TCOS j + j sin jJ- (A5.8)
= Mra - wjra - (M 1 - wjra) cos L
j
A5. 26 Formulas for Other Single Span Loadings
sin L
j In investigating other transverse loadings
and Cra = M1 - wjra for a single span carrying axial compression, it
is found that the expression for bending moment
Let D1 = M1 - wjra and Dra = Mra - wjra. Then, in the span always takes the form:
sUbstituting in equation (A5.3),
x x
Dra - D1 cos L M = C1 sin j + Cra cos j + f(w) - - - (A5.9)
M ---::-:;-::--;;---,,-j sin x + D1 cos X + wj& (A5.4)
where f(w) is a term which does not include the
sin L j j axial load P or the end moments M1 and Mra . The
j
expressions for f(w), C1 and Cra depend on the
To find the location of the maximum moment, dif- type of the transverse load.
ferentiate equation (A5.3) and equate to zero. Table A5.I gives the value of these 3
terms for types of transverse loading on a single
1M _ o _ C1 cos X Cra sin ~ span Which are frequently encountered in airplane
dx- - j j-j j
structures. The Table also gives equations for
whence the point of maximum bending moment and its mag-
L
nitUde.
tan ~ = C1 :; Dra - D1 cos j Table A5.II is a table of sines, cosines,
-- - - - - (A5.5) and tangents for L/j in radians which is more
j Cra D i L convenient to use than the usual type of trigo-
1 S n '3
nometric tables. This table is based on values
The value of X must fall within x = 0 to x = L, given in Appendix I of Air Corps Information
otherwise M1 or Mra is the maximum value. Circular #493. The ~ difference have been added
The value of the maximum span moment can be to facilitate rapid use of the tables.
found by substituting the value from equation For single span beams, the critical value
tA5.5) in (A5.4), which gives of L/j is n; that is, if the axial compressive
load is such that the term L/j = n, the center
~ = ~ + wjra - - - - - - - - - (A5.6) region of the beam will tend to deflect until the
cos x combined stresses equal the failing stress of the
j material.
The moment M at any point x along the span can
also be written:
A5. 27 Moments for Combinations of the Various Load
M= D1 r(tan !In . sin~} cos ~J + wjra - (A5. 7) Systems as Given in Table A5. I, Margins of
l j j j Safety. Accuracy of Calculations.
where xm refers to the value of x where the span
moment is maximum, or equation (A5.5). Since it The principle of superposition does not ap-
is customary to locate the pOint of maximum span ply to a beam-column, because the sum of the
bending moment and its value before investigat- bending moments due to the transverse loads and
ing' other span points, the value of tan !In is the axial loads acting separately are not the
j same as the moments when they act simultaneously.
known from equation (A5.5) and thus is available In combining several transverse load systems
to use in equation (A5.7) for finding moments at With their accompanying axial loads, the principle
other pOints along the span. of superposition can be said to app~y if each
If the equation for the beam deflection is transverse loading is used with the total axial
desired, it can be found by substituting the load for the systems Which are being combined.
value of M from equation (A5.3) in equation ThUS, in Table A5.I, to find the moments for
(A5.1), which gives:* several combined loadings, add the values of C1 ,
* Also see Chapter C2, Part E, Eq. (21). Cra and f(w) for the several lo~dings and use
A5.23
Table A5. I
Values of Terms C v C2, and f(w) in Equation
M· Cl sin~ + C2 cos~+ f(w)

j j
Single Span - Axial Compression - Uniform Section
Eq. for Point of Max Eq. for Max. Span

Loading f(w) Bending Moment Bending Moment
Equal End Moments. No
I ~-k Side Load
'f;::x~ L
Unequal End Moments. No
.r
M2 M 1 tan-h.
2j o
Tan~
x=L
2"
= M2 - Ml cos..!:..
M Side Load M j j
Mmax = Ml
cosx
---r-~
II o
sin L M1 sin..!:.
T j
T
p1;x::::j L
Uniform Side Load. No
(-r-1)]
iPLU ~ l-p
End Moments w#/in. wj2 [cos
ill x: .5 L
P t ttttf sink
j
Uniform Side Load Plus D2 - Dl cos L/j

End Moments Tan~: D2 - Dl cos.!=..
sm tIJ j j Mmax=~+wj2
~1#=d f ~f t f f t tf~2
IV where cos x
D 1 sin..!:.
Dl = Ml - wj2 i
D2· M2 - wj2 i
Concentrated Side Load. No

End Moments
x< a, - Wj sin b Tan xlj = g~
T
v
r-a-r-bi sin L
j
o o
Pt=§ L fP x> a, + Wi sin a

tan L
J
- Wi sin a
T
o
T
Triangular Loading. No (NOTE A)
End Moments ----.---rT" J. o To obtain Maximum Moment, compute
VI ~TTII w#/in. moment at 3 or 4 points in span. Draw a smooth
l-=x==:--l~~==ffP
p-;:::' curve thru plotted results.
L---
Triangular Loading. No End
1. Moments wj2(l-E (See Note A)
VII
w#/~~ L
Couple Loading (Clockwise) x < a, - m cos b o o (See Note A)

T
sin L
Vill
T
x> a, - m cos a m cos a o
T i
tan L
i
w or W is positive when upward.

M is positive when it tends to cause compression on the upper
fibers of the beam at the section being considered.
Reference: ACIC #493; Niles, Airplane Design; Newell and Niles

Airplane structures
For Table of many other loadings, see NACA T. M. 985.

TABLE A5.n
NATURAL SINES, COSINES, AND TANGENTS OF ANGLES IN RADIANS
L/i in 6L
Radians
Sin L /j S£ L I COS L Ii Cos Ii Tan L /j Ta~ L /j L/i in
J Radians
0.00 0.00000 1. 00000 0.00000 0.50 0.47943 0.87758 0.54630

0.01000 0.00005 0.01000 0.00875 0.00484 0.01306
0.01 0.01000 0.99995 0.01000 0.51 0.48818 0.87274 0.55936
0.01000 0.00015 0.01000 0.00870 0.00492 0.01320
0.02 0.02000 0.99980 0.02000 0.52 0.49688 0.86782 0.57256
0.01000 0.00025 0.01000 0.00865 0.00501 0.01335 -j
0.03 0.03000 0.99955 0.03000 0.53 O. 50553 O. 86281 0.58591
0.00999 0.00035 0.01002 0.00861 0.00510 0.01352
0.04 0.03999 0.99920 0.04002 0.54 0.51414 0.85771 0.59943
O. 0999 0.00045 0.01002 0.00855 0.00619 0.01367
0.05 0.04998 0.99875 0.05004 0.55 0.52269 0.85252 0.61310
O. 0998 0.00055 0.01003 0.00850 0.00526 0.01385
0.06 0.05996 0.99820 0.06007 0.56 0.53119 0.84726 0.62695
0.00998 0.00065 0.01005 0.00844 0.00536 0.01402
0.07 0.06994 0.99755 0.07012 0.57 0.53963 0.84190 0.64097
0.00997 0.00075 0.01005 0.00839 0.00544 0.01420
0.0 0.07991 0.99680 0.08017 0.58 o. 54802 0.83646 0.65517
0.00997 0.00085 0.01007 0.00834 0.00552 o 0 438
0.09 0.08988 0.99595 0.09024 0.59 0.55636 0.83094 0.66955
0.00995 0.00095 0.01010 0.00828 0.00560 0.01459
0.10 0.09983 0.99500 0.10034 0.60 0.56464 0.82534 0.68414
0.00995 0.00104 0.01011 0.00823 0.00569 0.01478
0.11 O. 10978 0.99396 O. 11045 0.61 0.57287 0.81965 0.69892
0.00993 0.00115 0.01012 0.00817 0.00677 0.01499
-.'''' ~ r --:---;-" ; = -~~-iT' " ~ .... "': ....
(l
1.....
0_74~ ~ .67429-;;; __ - Oc"l3
.
O. 00l3:r"=-
= - -.0,681:641"" O. 73169 _ ..
..
+t--J_""",=-/I!
. 9220: ==
0. 01959-·
::;-
~0~6
----
iii
(J.00961L
0:00963
03l6639:.
0.96377
0:002li2C=
0: 00271
i 0.2'1670"
0.01080
0.77
889:F-
a:-=28=
0.007~
!! 69614" .. O. 00714~
----0=8 _
0.71791
0.01998 0.28 0.'27636 0.96106 • 0,-28756 . 0,70328 0; 71091

- 0.00959 0.00282 0.01085 0.00707-
0.02040 0.29 0.28ii95" 0.095824 _9841 O. 79~ 0.71035 0.70385
0.01093 0.00701-
0.02082 0.30 0.29552 0.95504' 0.30934 0.80 0.71736 0: 69671
0.00954 0.00301 0.01098 0.00693
0.02125 0.31 0.30506 0.95233 0.32032 0.81 0.72429 0.68950
0.00951 0.00309 0.01107 0.00686
0.02172 0.32 0.31457 0.94924 0.33139 0.82 0.73116 0.68222
0.00947 0.00320 0.01114 0.00678
0.02220 0.33 0.32404 0.94604 0.34253 0.83 0.73793 0.67488
0.00945 0.00329 0.01121 0.00671
0.02271 0.34 O. 33349 0.94275 0.35374 0.84 0.74464 0.66746
0.00941 0.00338 0.01129 0.00664
0.02321 0.35 0.34290 0.93937 0.36503 0.85 0.75128 0.65998
0.00937 0.00347 0.01137 0.00656
0.02378 0.36 0.35227 0.93590 0.37640 0.86 0.75784 0.65244
0.00935 0.00357 0.01146 0.00649
0.02434 0.37 0.36162 0.93233 0.38786 0.87 0.76433 0.64483
0.00930 0.00367 0.01155 0.00641
0.02493 0.38 0.37092 0.92866 0.39941 0.88 0.77074 0.63715
0.00927 0.00375 0.01164 0.00633
0.02556 0.39 0.38019 0.92491 0.41106 0.89 0.77707 0.62941
0.00923 0.00385 0.01174 0.00626
0.02621 0.40 0.38942 0.92106 0.42279 0.90 0.78333 0.62161
0.00919 0.00394 0.01184 0.00617
0.02689 0.41 0.39861 0.91712 0.43463 0.91 0.78950 0.61375
0.00915 0.00403 0.01194 0.00610
0.02762 0.42 0.40776 0.91309 0.44657 0.92 0.79560 0.60582
0.00911 0.00412 0.01205 0.00602
0.02835
r---'-O,-.4"'3'----_+_ ~O,-.4",1,-,6~8~7-+-nO.-'0"'0"'9"'07,,-r---0.,-.~9~08,-9,-7---jho •.noon4"2"'2e-t_-,-0.,-4,-,5~8,-62"-i~on.-'0"1"2"16rl 0.93 0.80162
0.00594
0.59783
0.02915 0.44 0.42694 0.90475 0.47078 0.94 0.80756 0.58979

0.00903 0.00430 0.01228 0.00586
0.02998 0.45 0.43497 0.90045 0.48306 0.95 0.81342 0.58168
0.00898 0.00440 0.01239 0.00577
0.03084 0.46 0.44395 0.89605 0.49545 0.96 0.81919 0.57352
0.00894 0.00448 0.01251 0.00570
0.03176 0.47 0.45289 0.89157 0.50796 0.97 0.82489 0.56530
0.00889 0.00458 0.01265 0.00561
0.03272 0.48 0.46178 0.88699 0.52061 0.98 0.83050 0.55702
0.00885 0.00466 0.01278 0.00553
0.03373 0.49 0.47063 0.88233 O. 53339 0.99 O. 83603 O. 54869
0.00880 0.00476 0.01301 0.00544
0.50 0.47943 0.87768 0.54630 1. 00 0.84147 0.54030
TABLE A5.II
L/j in Sin L/j llL Cos L /j llL II L L/j in II L L\L L\L

Radians Sin I j Cos I j Tan L /j Tan I j Radians
Sin L /j Sin I j Cos L /j Cos I j Tan L /j Tan I j
I. 00 0.84147 0.54030 I. 55741 1. 50 0.99749 0.07074 14.10142

0.00536 0.00644 0.03480 0.00066 0.00998 2.32669
I. 01 0.84683 0.53186 I. 59221 1. 51 0.99815 0.06076 16.42811
0.00528 0.00649 0.03592 0.00056 0.00999 3.2415
I. 02 0.85211 0.52337 I. 62813 1. 52 0.99871 0.05077 19.6696
0.00519 0.00855 0.03712 0.00046 0.00998 4.8290
I. 03 0.85730 0.51482 I. 66525 1. 53 0.99917 0.04079 24.4986
0.0 510 0.00860 0.03836 0.00036 0.01000 7.9527
I. 04 0.86240 0.50622 I. 70361 1. 54 0.99953 0.03079 32.4513
0.00502 0.00865 0.03971 0.00025 0.01000 15.6290
I. 05 0.86742 0.49767 I. 74332 I. 55 0.99978 0.02079 48.0803
0.00494 0.00870 0.04110 0.00018 0.0999 44.5435
I. 06 0.87236 0.48887 I. 78442 1. 56 0.99996 0.01080 92.6238
0.00484 0.00875 0.04261 0.00004 0.01000 1182.42
I. 07 0.87720 0.48012 I. 82703 1. 57 1. 00000 0.00080 1275.04
0.00476 0.00879 0.04419 0.00004 0.01000 1383.70
I. 08 8196 0.47133 I. 87122 1. 58 0.99996 -0.00920 -108.661
0.00467 0.00884 0.04588 0.00014 0.01000 56.593
I. 09 0.88665 0.46249 1.91710 1. 59 0.99982 -0.01920 -52.0676
.004"" 0.00889 0.04766 0.00025 0.01000 17.8347
1.10 0.89121 0.45360 I. 96476 1. 60 0.99957 -0.02920 -34.2329
0.00449 0.00894 0.04958 0.00034 0.00999 8.7379
1.11 0.89570 0.44466 2.01434 1. 61 0.99923 -0.03919 -25.4950
0.00440 0.00898 0.05161 0.00044 0.00999 5.1877
1.12 0.90010 0.43568 2.06595 1. 62 0.99879 -0.04918 -20.3073
0.00431 0.00902 0.05380 0.00054 0.00999 3.4361
1.13 0.90441 0.42666 2.11975 1. 63 0.99825 -0.05917 -16.8712
.00422 0.00907 0.05613 0.00065 0.00998 2.4442
1.14 0.90863 0.41759 2. 17588 1. 64 0.99760 -0.06915 -14.4270
0.00413 0.00910 0.05861 0.00073 0.00997 1. 8278
1.15 0.91276 0.40849 2.23449 1. 65 0.99687 -0.07912 -12.59926
0.00404 0.00915 0.06131 0.00085 0.00997 1. 41867
1.16 0.91680 0.39934 2.29580 1. 66 0.99602 -0.08909 -11.18059
0.00395 0.00919 0.06418 0.00094 0.00995 1. 33335
1.17 0.92075 0.39018 2.35998 1. 67 0.99508 -0.09904 -10.04724
0.00386 0.00923 0.06728 0.00104 0.00995 0.92648
1.18 0.92461 0.38092 2.42726 1. 68 0.99404 -0.10899 - 9.12076
0.00376 0.00926 0.07064 0.00114 0.00993 0.77151
1.19 0.92837 0.37166 2.49790 1. 69 0.99290 -0.11892 - 8.34925
0.00367 0.00930 0.07425 0.00123 0.00992 0.65265
1. 20 0.93204 0.36236 2.57216 1. 70 0.99167 -0. 12884 - 7.69660
0.00358 0.00934 0.07818 0.00134 0.00991 O. 55937
1. 21 0.93562 0.35302 2.65033 1.71 0.99033 -0.13875 - 7.13723
0.00348 0.00937 0.08243 0.00144 0.00990 0.48478
I. 22 0,93910 0.34365 2.73276 1.72 0.98889 -0.14866 - 6.65245
0.00339 0.00941 0.08706 0.00153 0.00988 0.42436
1. 23 0.94249 0.33424 2.81982 1. 73 0.98736 -0. 15853 - 6.22809
0.00329 0.00944 0.09212 0.00164 0.00987 0.37456
1. 24 0.94578 0.32480 2.91194 1. 74 0.98572 -0.16840 - 5.85353
0.00320 0.00948 0.09763 0.00173 0.00985 0.33316
1. 25 0.94898 0.31532 3.00957 I. 75 0.98399 -0.17825 - 5.52037
0.00311 0.00950 O. 10371 0.00184 0.00983 0.29828
1. 26 0.95209 0.30582 3.11328 1. 76 0.98215 -0.18806 - 5.22209
0.00301 0.00954 0.11035 0.00192 0.00981 0.26869
1. 27 0.95510 0.29628 3.22363 1. 77 0.98023 -0.19789 - 4.98340
0.00292 0.00956 0.11772 0.00204 0.00979 0.24330
1. 28 0.95802 0.28872 3.34135 I. 78 0.97819 -0.20768 - 4.71010
0.00282 0.00960 O. 12586 0.00212 0.00977 0.22144
1. 29 0.96084 0.27712 3.46721 1. 79 0.97607 -0.21745 - 4.48866
0.00272 0.00962 0.13489 0.00222 0.00975 0.20230
I. 30 0.96356 0.26760 3.60210 1. 80 0.97385 -0.22720 - 4.28627
0.00262 0.00965 0.14498 0.00233 0.00973 0.18577
1. 31 0.96618 0.25785 3.74708 1. 81 0.97152 -0.23693 - 4.10050
0.00254 0.00967 0.15627 0.00241 0.00970 0.17113
1. 32 0.96872 0.24818 3.90335 1. 82 0.96911 -0.24663 - 3.92937
0.00243 0.00970 0.16896 0.00252 0.00968 0.15819
1. 33 0.97115 0.23848 4.07231 I. 83 0.96659 -0.25631 - 3.77118
0.00233 0.00973 0.18331 0.00261 0.00965 0.14668
1. 34 0.97348 0.22875 4.25562 1. "4 0.96"98 -0.26596 - 3.62450
0.00224 0.00974 0.19961 0.00271 0.00963 0.13644
1. 35 0.97572 0.21901 4.45523 1. 85 0.96127 -0.27559 - 3.48806
0.00214 0.00977 0.21021 0.0 280 0.00900 U.12723
1. 36 0.97786 O. 20924 4.67344 1. 86 0.95847 -0.28519 - 3.36083
0.00205 0.00979 0.23962 0.00290 0.00957 0.11895
I. 37 0.97991 0.19945 4.91306 1. 87 0.95557 -0.29476 - 3.24188
0.00194 0.00981 0.2 43" 0.00300 0.00954 .1114"
1. 38 0.98185 O. 18964 5. 17744 1. 88 0.95257 -0.30430 - 3.13039
0.00185 0.00983 0.29326 0.00308 0.00951 0.10473
1. 39 0.98370 0.1 981 5.47069 1. 89 0.94949 -0.31381 - 3.02556
0.00175 0.00984 0.32719 0.00319 0.00948 0.09"56
1. 40 0.98545 0.16997 5.79788 1. 90 0.94630 -0.32329 - 2.92710
0.00165 0.00987 0.36749 0.00328 0.00945 0.09296
1. 41 0.98710 0.16010 6.16537 I. 91 0.94302 -0.33274 - 2.83414
0.00155 0.00987 0.41573 0.00338 0.00941 0.08784
1. 42 0.98865 0.15023 6.58110 1. 92 0.93964 -0.34215 - 2.74630
0.00145 0.00990 0.47436 0.00346 0.00938 0.08314
1. 43 0.99010 0.14033 7.05546 1. 93 0.93618 -0.35163 - 2.66316
0.00136 0.00991 0.54636 0.00356 0.00934 0.07883
1. 44 0.99146 0.13042 7.60182 I. 94 0.93252 -0.36087 - 2.58433
0.00125 0.00992 0.63628 0.00366 0.00931 0.07486
1. 45 0.99271 0.12050 8.23810 1. 95 0.92896 -0.37018 - 2.50947
0.00116 0.00993 0.75052 0.00375 0.00927 0.07119
1. 46 0.99387 O. 11057 8.98862 1. 96 0.92521 -0.37945 - 2.43828
0.00105 0.00994 0.89878 0.003B4 0.00923 0.06779
1. 47 0.99492 0.10063 9.88740 I. 97 0.92137 -0. 3886B - 2.37049
0.00096 0.00996 1. 09698 0.00393 0.00920 0.06467
1. 48 0.99588 0.09067 10.98338 1. 98 0.91744 -0. 397BB - 2.305B2
0.00086 0.00996 1. 36653 0.00403 0.00916 0.06534
1. 49 0.99674 0.08771 12.34991 1. 99 0.91341 -0.40705 - 2.2404B
0.00075 0.00997 I. 75151 0.00411 0.00912 0.05544
1. 50 0.99749 0.07074 14.10142 2.00 0.90930 -0.41615 - 2. 18504
TABLE A5.II
L/j in
Radians
Sin L/. Si~ L/. Cos L/. co~L/ Tan L/. Ta~ L/ L/j in
Radians
Sin L/. Si~ L/ Cos L/
j
Co~ L/. Tan L/
j
Ta~ L/.
J J J J J J J J J J
2.00 0.90930 -0.41615 - 2.18504 2.50 0.59847 -0.80114 - 0.74703

0.00421 0.00907 0.05651 0.00804 0.00596 0.01548
2.01 0.90509 -0.42522 - 2.12853 2.51 0.59043 -0.80709 - 0.73165
0.00430 0.00903 0.05416 0.00810 0.00586 0.01523
2.02 0.90079 -0.43425 - 2.07437 2.52 0.58233 -0.81295 - 0.71632
0.00438 0.00898 0.05195 0.00818 0.00578 0.01503
2.03 0.89641 -0.44323 - 2.02242 2.53 O. 57417 -0.81873 - 0.70129
0.00448 0.00895 0.04990 0.00821 0.00571 0.01482
2.04 0.89193 -0.45218 - I. "7252 2.54 0.505"0 -v. 0'444 - V.OOO"
0.00457 0.00889 0.04766 0.00827 0.00561 0.01461
2.05 0.88736 -0.46107 - I. 92456 2.55 0.55769 -0.83005 - 0.67186
0.00466 0.00885 0.04615 0.00833 0.00554 0.01442
2.06 0.88270 -0.46992 - I. 87841 2.56 0.54936 -0.83559 - 0.65744
0.00473 0.00881 0.04445 0.00649 0.00545 0.01422
2.07 0.87797 -0.47873 - I. 83396 2.57 0.54097 -0.84104 - 0.64322
0.00484 0.00875 0.04284 0.00844 0.00537 0.01405
2.08 0.87313 -0.48748 - 1.79112 2.58 0.53253 - 0.84641 - 0.62917
0.00491 0.00871 0.04135 0.00848 0.00528 0.01387
2.09 0.86822 -0.49 19 - I. 74977 2.59 0.52405 -0.85169 - 0.61530
0.00503 0.00866 0.03993 0.00855 0.00520 0.01370
2.10 0.86319 -0.50485 - I. 70984 2.60 0.51550 -0.85689 - 0.60160
0.00507 0.00860 0.03857 0.00859 0.00511 0.01354
2.11 0.85812 -0.51345 - I. 67127 2.61 0.50591 -0.86200 - 0.58806
0.00518 0.00856 0.03732 0.00664 0.00503 0.01338
2.12 0.85294 -0.52201 - I. 63395 2.62 0.49827 -0.86703 - 0.57468
0.00526 0.00850 0.03610 0.00870 0.00494 0.01323
2.13 0.84768 -0.53051 - I. 59785 2.63 0.48957 -0.87197 - 0.56145
0.00535 0.00845 0.03498 0.00875 0.00485 0.01308
2.14 0.84233 -0.63896 - I. 56287 2.64 0.48082 -0.87682 - 0.54837
0.00543 0.00840 0.03389 0.00878 0.00476 0.01293
2.15 0.83690 -0.54736 I. 52898 2.65 0.47204 -0.88158 - 0.53544
0.00552 0.00834 0.03288 0.00885 0.004ti8 0.OH8V
2.16 0.83138 -0.55570 - I. 49610 2.66 0.46319 -0.88626 - 0.52264 I
0.00559 0.00828 0.03191 0.00888 0.00459 o. 12ti7
2.17 0.82579 -0.56398 - I. 46419 2.67 0.45431 -0.89085 - 0.50997
0.00569 0.00823 0.03098 0.00893 0.00449 0.01253
2.18 0.82010 -0.57221 - I. 43321 2.68 0.44538 -0.89534 - 0.49744
0.00576 0.00818 0.03011 0.00898 0.00441 0.01242
2.19 0.81434 -0.58039 - 1.40310 2.09 0.4364 -0.89975 - 0.48502
-0.00585 0.00811 0.02928 0.00902 0.00432 0.01229
2.20 0.80849 -0.58850 - I. 37382 2.70 0.42738 -0.90407 - 0.47273
0.00591 0.00806 0.02848 0.00907 0.00423 0.01218
2.21 0.80258 -0.59656 - I. 34534 2.71 0.41831 -0.90830 - 0.46055
0.00601 0.00799 0.02773 0.00909 0.00414 0.01206
2.22 0.79657 -0.60455 - I. 31761 2.72 0.40922 -0.91244 - 0.44849
0.00609 0.00794 0.02701 0.00915 0.00404 0.01196
2.23 0.79048 -0.61249 - I. 29060 2.73 0.40007 -0.91648 - 0.43653
0.00616 0.00787 0.02631 0.00918 0.00396 0.01186
2.24 0.78432 -0.62036 - I. 26429 2.74 0.39089 -0.92044 - 0.42467
0.00625 0.00781 0.02566 0.00922 0.00386 0.01175
2.25 0.77807 -0.62817 - I. 23863 2.75 0.38167 -0.92430 - 0.41292
0.00632 0.00775 0.02503 0.00927 0.00377 0.01166
2.26 0.77175 -0.63592 - I. 21360 2.76 O. 37240 -0.92807 - 0.40126
0.00639 0.00769 0.02444 0.00930 0.00358 0.01156
2.27 0.76536 -0.64361 - I. 18916 2.77 0.36310 -0.93175 - 0.38970
0.00648 O. 00762 0.02385 0.00933 0.00368 0.01148
2.28 0.75888 -0.65123 - I. 16531 2.78 0.35377 -0.93533 - 0.37822
0.00 55 0.0075 23 2 O. 3
2.29 0.75233 -0.65879 - I. 14199 2.79 0.34440 -0.93883 - 0.36684
f-g.00662 0.00749 u.02278 0.00941 0.0033" v. v1,,1
2.30 0.74571 -0.66628 - I. 11921 2.80 0.33499 -0.94222 - 0.35553
0.00669 0.00742 0.02227 0.00944 0.00331 0.01122
2.31 0.73902 -0.67370 - I. 09694 2.81 0.32555 -0.94503 - 0.34431
0.00678 0.00736 0.02180 0.00947 0.00320 0.01115
2.32 0.73224 -0.68104 - I. 07514 2.82 0.31608 -0.94873 - 0.33316
0.00685 0.00728 u.02133 0.00950 0.00312 0.01107
2.33 0.72539 -0.68834 - I. 05381 2.83 0.30658 -0.95185 - 0.32209
0.00692 0.00722 0.02089 0.00954 0.00302 0.01100
2.34 0.71847 -0.69556 - I. 03292 2.84 0.29704 -0.95487 - 0.31109
0.00699 0.00715 0.02045 0.00956 0.00292 0.01095
2.35 0.71148 -0.70271 - I. 01247 2.85 0.28748 -0.95779 - 0.30014
0.00707 0.00708 0.02005 0.00960 0.00282 0.01086
2.36 0.70441 -0.70979 - 0.99242 2.86 0.27788 -0.96061 - O. 28928
0.00713 0.00701 0.01966 0.00961 0.00274 0.01081
2.37 0.69728 -0.71680 - 0.97276 2.87 0.26827 -0.96335 - 0.27847
0.00721 0.00694 0.01927 0.00965 0.00263 0.01074
2.38 0.69007 -0.72374 - 0.95349 2.88 0.25862 -0.96598 - 0.26773
0.00726 0.00686 0.01892 0.00967 0.00254 0.01069
2.39 0.68281 -0. 73060 - 0.93457 2.89 0.24895 -0.96852 - 0.25704
0.00733 0.00679 0.01856 0.00970 0.00244 o 01063
2.40 0.67547 -0.73739 - 0.91602 2.90 0.23925 -0.97096 - 0.24641
. u.Ou741 0.00 72 u.01823 0.00973 0.00234 0.01058
2.41 0.66806 -0.74411 - 0.89779 2.91 0.22952 -0.97330 - 0.23583
0.00748 0.00685 0.01790 0.00973 0.00225 0.01054
2.42 0.66058 -0.75076 - 0.87989 2.92 0.21979 -0.97555 - 0.22529
0.00754 0.00656 0.01759 0.00977 0.00215 0.01048
2.43 0.65304 -0.75732 - 0.86230 2.93 0.21002 -0.97770 - 0.21481
0.00760 0.00651 0.01728 0.00979 0.00205 0.01044
2.44 0.64544 -0.76383 - 0.84502 2.94 0.20023 -0.97975 - 0.20437
0.00767 0.00640 0.01701 0.00981 0.00195 0.01040
2.45 0.63777 -0.77023 - 0.82801 2.95 0.19042 -0.98170 - 0.19397
0.00774 0.00634 o 01671 o 0098 o 00186 o0 0 5
2.46 0.63003 -0.77657 - 0.81130 2.96 0.18060 -0.98356 - 0.18362
0.00779 0.00626 0.01645 0.00984 0.00175 0.01032
2.47 0.62224 -0.78283 - 0.79485 2.97 0.17076 -0.98531 - 0.17330
0.00786 0.00718 0.01619 0.00987 0.00166 0.01029
2.48 0.61438 -0.78901 - 0.77866 2.98 0.16089 -0 98697 - 0 16301
0.00792 0.00611 0.01594 0.00988 0.00156 0.01025
'.4" U.OU04O -0.79512 - 0.76272 2.99 0.15101 -0.98853 - 0.15276
0.00799 .00602 0.0156" 0.00"09 0.00146 0.01022
2.50 0.598'7 -0.80114 - 0.74703 3. uv .ld1. - . "8999 - 0.14254
TABLE A5.n
NATUBAL SINES, COS Ins , AIm TAl(GEl(TS OF ANGLES IX BADUJ(S
in L:.. ~ ~
L/ L/
j Bin L/ Sin Cos L/ Cos L/ Tan L/ Tan L/
Radians j j j j j j
3.00 0.14112 -0.98999 -0.14254

0.00991 0.00136 0.01019
3.01 0.13121 -0.99135 -0.13235
0.00992 0.00127 0.01016
3.02 0.12129 -0.99262 -0.12219
0.00993 0.00116 0.01013
3.03 0.11136 -0.99378 -0.11206
0.00994 0.00106 0.01011
3.04 0.10142 -0.99484 -0.10195
0.00996 0.00097 0.01010
3.05 0.09146 -0.99581 -0.09185
0.00996 0.00086 0.01008
3.06 0.08150 -0.99667 -0.08177
0.00997 0.00077 0.01006
3.07 0.07153 -0.99744 -0.0717l
0.00998 0.00066 0.01004
3.08 0.06155 -0.99810 -0.06167
0.00999 0.00057 0.01003
3.09 0.05156 -0.99867 -0.05164
0.00997 0.00046 0.01002
3.10 0.04159 -0.99913 -0.04162
0.01000 0.00037 0. 01001.
3.11 0.03159 -0.99950 -0.03161
0.00999 0.00027 0.01001
3.12 0.02160 -0.99977 -0.02160
0.01000 0.00016 0.01000
3.13 0.01160 -0.99993 -0.01160
o nlnno o 00007 0.01000
3.14 0.00160 -1.00000 -0.00160
0.01001 0.00003 0.01001
3.15 -0.00841 -0.99997 0.00841
0.01000 0.00014 0.01000
3.16 -0.01841 -0.99984 0.01841
0.00999 0.00023 o.nlrJnn
3.17 -0.02840 -0.99960 0.02841
0.01000 0.00034 0.01002
3.18 -0.03840 -0.99926 0.03843
0.00999 0.00043 0.01002
3.19 -0.04839 -0.99883 0.04845
0.00999 0.00083 0.01003
3.20 -0.05838 -0.99830 0.05848
0.00998 0.00064 0.01004
3.21 -0.06836 -0.99766 0.06852
0.00997 0.00073 0.01005
3.22 -0.07833 -0.99693 0.07857
0.00996 0.00084 0.01007
3.23 -0.08829 -0.99609 0.08864
0.00996 0.00093 0.01009
3.24 -0.09825 -0.99516 0.09873
0.00995 0.00103 0.01010
3.25 -0.10820 -0.99413 0.10883
A5.28 BEAM -- COLUMNS
these values in the general expression for M as of the beam, or 4420 x .75 = 3315"# positive be-
given at the top of the Table. cause it produces compression in the top fibers.
In a beam-column member, the bending mom- The moment at (2) due to the cantilever overhang
ents do not vary directly as the load is increa equals (20 + 10) 36 x 16 = 8640"#. Fig. A5.68
ed. Thus, the student should realize that marg- 2
ins of safety based on direct proportion of mom- shows the beam portion between points (1) and (2)
ents to loads are incorrect and lie on the un- as a free body.
safe side. From Art. A5.25, we have the following pre-
It is recommended that four significant cise equations for a beam carrying a transverse
figures be used in computations, making use of uniform distributed load With end compressive
the so-called precise equations, since the re- loads.
sults in many cases involve small differences
x L
between large numbers. tan J = D" - D~ cos J - - - - - - - - -(A)
A5.28 Example Problems* D~ sin L
Example Problem #1 J
Fig. A5.67 illustrates a typical upper, and
outer panel wing beam of a biplane. Let it be M
-iIIaX
= ~
cos x + Wj" - - - - - - - - - - -(B)
required to determine the maximum negative }
bending moment between points (1) and (2), gen-
erally referred to as the maximum span moment. Evaluating terms for substitution in these equa-
To obtain the true bending moments on the beam, tions, we obtain,
the axial beam load as well as the end moments
M~ 3315"#
at (1) and (2) are necessary since they influ-
ence the deflection of the beam. M" 8640"#
Solution:- p = 4420# compression
To obtain the horizontal component Th of I = 10 in 6 given and assumed constant
the lift strut load, we take moments about the throughout the span.
j =V-.3XJ~;X 10 = v'2941 = 54.23

wj"= 20 x 2941 = 58820
D~ = M~ - wj" 3315 - 58820 - 55505
D" = M" - Wj" 8640 - 58820 = - 50180
1500#=Reaction
from Strut to
Lower Wing ~
!:j = 54.23 = 1.844
From Table A5.II sin J = .96290 and cos J =- .26981

Fig. A5.67 Substituting in equation (A)
x L
.--w= 20#/in. tan } = D" - D~ cos}
II' 8640"#
4420#-fi 1t Tt t Tt t r r tIll tIt }1=4420# D~
L
J
3315"#--'"' t L=100:":'''---~1 /
sin
50180 - (-55505 x - :26981) -65156 = 1.2192
940#
-55505 x .96290 =5344I
Fig. A5.68 j = tan-" 1.2192 = .88383
hinge point at the lert end •.
Hence, x = .88383 x 54.23 = 48", which equals the
2Mb = - 2000 x 50 - 540 x 116 distance from the left end of the beam to the
- 1500 x 100 + 70.75 Th = 0 point of maximum span moment.
hence
~
Th = 4420# cos x + wJ~ , cos xJ = .63419 from Table
The axial compressive load induced by the lift } A5.I1
strut at point (2) then equals - 4420#.
Taking ~H = 0 for the load system of Fig. Hence
55505
:63419 + 58820 -_ - 28,700 "#
A5.67 gives P = - Th = 4420#. The end moment on ~
the beam at (1) equals the end load times the
eccentricity of the hinge from the neutral axis To obtain an idea as to the magnitude of the
secondary bending moment, that i8, the moment due
*Additional examples are presented in Chapter C2, Part E, to the axial load times the lateral beam deflec-
with stress analysis. tion, the primary bending moment at a point 48"
A5.29
from the left end will be computed. Substituting values of C~ and C.. and f(W) from
Table A5. I in the above equations:
M.a =3315 + 48 x 20 x 24 - 940 x 48 = -18765"#
Thus the secondary bending moment equals M= (M.. - M~ cos L/j) sin x/j + M cos x/j
sin L!j ~
- 28700 + 18765 = - 9935"# which is a large per-
centage of the primary moment. The transverse
deflection of the beam at the point of max. span
moment then equals - 9935 = 2.25 inches upward.
- 4420
20 000 '"
'~I- > - - - - - L = 41. 762 -j
..
-
,,2 1/2 - 083 Steel Tube 1~02000,000'"
1t
Bending Moment at any Point Along Span

®
Fig. A5.70
Let the moment at a point 10" from point
(2) be required. In this case, x"= 100 - 10 = 90 But, M~ = 0 in our problem, hence,
M= nJ(tan
L
x: . sin j)+ cos j] + wj" (Ref.
A5.7)
Eq. M = M..sin x/j
sin L/j
x 90 x 8
j =.. @. =.. /29 x 10 x .46 =1/667 = 25 826
J = 54:23 = 1.6596, sin J .99605 VlP V 20000 •
cos ~j - .08867 L/j =41.762=1617 - - - sin L/j = .99892
25.826 •
tan ~ =1.2192 = value for x at the point of
maximum bending moment x = L/2 = 20.881
Hence, x/j -- 20.881
M = - 55505 [<1.2192 x .99605) +-.08867 ] + 25.80 -- .8085 - - - sin x/j = .72327
58820 = - 3664"# SUbstituting in the above equations for M
Example Problem #2 M = 36000 x .72327 = 26066"#
Fig. A5.69 shows a simplified landing gear .99892
structure carrying a vertical load of 12000# on This compares with a primary moment of
the axle. Member ABC is continuous thru Band
pinned at C. Let it be required to determine the 36000/2 = 18000"#. The deflection at the mid-
point of BC = 26066 - 18000 = .403 in.
bending moment at the midpoint of member BC and
its lateral deflection due to the 12000# verti- 20000
cal design load. The maximum moment is given by the equation:
~ax = s~n L and it occurs at x = ~j (See
~---Z6"~C J Table A5.I)
-=-------..---
862# A5.29 Stresses Above Proportional Limit Stress of
Material. *
The equations as presented in this chapter

assume that E is constant or in other words the
stresses are within the elastic range. In air-
craft structural design the applied or limit
loads must be taken without suffering permanent
deformation, hence E is constant under such
loads. However the aircraft structure must
take the design loads which equal the limit
loads times a factor of safety (usually 1.5)
Fig. A5.69 without failure. In many cases structural
failure will occur under stresses in the plastic
range where the material stiffness is less and
Line of action of DB goes through E not constant.
A good approximation for an effective
Solution:-
Solving for reactions at C by statiCS, we modulus E' is obtained as follows:-
obtain the axial load in BC = -20000. The
bending moment at B due to 3" eccentricity ot (1) Compute Fc =piA for the given number.
the wheel load = 3 x 12000 = 36000"#.
Fig. A5.70 shows a free body of portion BC (2) With this value of Fe enter the basic
of member ABC. From Table A5. I column curve diagram for the given material (for
=
end fixity C 1) and find value of L'/p cor-
M = C~ sin j + C.. cos j + f(w) responding to the stress Fc '
* Also see Chapter C2, Part E.
A5.30 BEAM -- COLUMNS
(3) Using these values of L' /p and Fc , (4) Determine the bending moment at the
compute centerline of the beam-columns shown in Fig.
AIO.IO Assume EI =
64,000,000 lb. in. sq.
E' =!iL(L'
n p )a a
(4) Then j =(E~I)4/a

Basic column curves for various materials
are given in another chapter of this book.
A5.30 Problems*
Fig. A5.75
500# (5) For the beam-column in Fig. A5.75
~35" --I W=30#/in. calculate the bending moment of the centerline
60 _ ~_ _ _ l!oo~ f t f t t +t ~O# I = 10 in:
=
of the member. Assume E 1,300,000 psi. and
I.. 50"
Fig. A5.71
·1 r 't-
30
300#
60 "---1
Fig. A5.72
(1) Fig. A5.71 shows a 1-1/2 - .065 steel
tUbe sUbjected to both end and lateral loads.
Determine the maximum bending moment on the 5000#~
=] ,t
r 10#/in.
f , fTTTT1EO#/in~
~ 5000#
tUbe. Compare the result with the bending
moment due to the side load only. E 29 X 10 8 = 9000 f-- 5 0 " + 50" ---j 18000"#
psi. I of tUbe =
.075 in.· Compute lateral 500*
deflection at point of maximum bending moment. Fig. A5.76
Area tUbe = 0.293 sq. in.
(6) For the beam-column loading in Fig.
(2) The beam column member in Fig. A5.72 is A5.76, calculate bending moment at center
made of 24ST aluminum alloy. Calculate and pOInt of beam. Take E - 1,200,000 psi and
plot a curve of the bending moments on the I = 10 in!
member. Also plot bending moment due to
lateral loads only. E 10.3 x 10 8 psi. I = = A5.31 Beam-Columns in Continuous Structures.
5.0 in.· The secondary moments in a particular
(3) Determine the maximum bending moment member due to beam-column action also effect or
for the wood wing beam and loading of Fig. influence the deflections in adjacent members
of a continuous structure. This rather involved
AIO.9. I of beam section 17 in.· E 1.3 X = = problem can be handled qUite simply and rapidly
10 8 • by the moment distribution method as explained
and illustrated in Arts. All.12 to 15 of
---96"~40"~
W = 25#/in.
Chapter All and in Part F of Chapter C2.
A numerical procedure for such beam-

t1 t 1ft f f l t t f t t t t t t columns having a varying EI and/or a varying
p-r F ~~ ~Beam I axial load is presented and illustrated in
Part K and L of Chapter C2.
~\ Beam-columns on elastic supports are
T Fig. A5.73
discussed in Part M of Chapter C2.
1000# 2000#
30
arJo"
1-48"~96"-+56'~
1~
-~t"1 ' f f f t , , 1 , , f f f t , 1 t Jt,'l.OOO
,- 10#/in. 0"1\
10000#
• 200" 1
Fig. A5.74
CHAPTER A6
TORSION. - STRESSES AND DEFLECTIONS
A6. 1 Introduction. sumed that any radial line undergoes angular

Problems involving torsion are common in displacement only, or OB remains straight when
aircraft structures. The metal covered airplane moving to OB' .
wing and fuselage are basically thin-walled The unit shearing strain in a distance L
tubular structures and are subjected to large equals,
torsional moments in certain flight and landing
conditions. The various mechanical control sys- BB' r G
tems in an airplane often contain units of E=Aj3=r:,-
various cross-sectional shapes which are sub-
jected to torsional forces under operating con- Let G equal modulus of rigidity of the
ditions, hence a knowledge of torsional stresses material and let ~ equal the unit shearing
and distortions of members is necessary in air- stress at the extreme fiber on the cross-sec-
craft structural design. tion.
A6.2. Torsion of Members with Circular Cross Sections. Hence, ~ = E G = r GG _ (1)
L
The following conditions are assumed in the
derivation of the equations for torsional
stresses and distortions: - In Fig. A6.2 let ~P equal the unit shear-
ing stress on a circular strip dA at a distance
(1) The member is a circular, solid or hollow p from O. Then
round cylinder.
(2) Sections remain circular after application = -P LG -G
of torque.
(3) Diameters remain straight after tWisting of The moment of the shearing stress on the
section. circular strip dA about 0 the axis of the bar
(4) Material is homogeneous, isotropic and is equal to,
elastic. p"GGdA
dM = ~p pdA = ---L- and thus the total in-
(5) The applied loads lie in a plane or planes
perpendicular to the axis of the shaft or ternal torsional resisting moment is,
cylinder. p
L M = J Gp"GdA
into L
For equilibrium, the internal resisting

moment equals the external torsional moment T,
and since GG/L is a constant, we can write,
A6.1
T = Mint. = G~ ]:P"dA = G~J (2)
where J =
polar moment of inertia of the shaft
cross section and equals twice the moment of in-
ertia about a diameter.
Fig. A6.2
L _- _~
From equa t1. on (1) GG r
Fig. A6.1 shows a straight cylindrical bar Hence, T = ~ (3 )

subjected to two equal but opposite torsional r
couples. The bar twists and each section is or ~ =T-r (4)
subjected to a shearing stress. Assuming the J
left end as stationary relative to the rest of also from equation (2), solving for the tWist G,
the bar a line AB on the surface will move to
AB' under these shearing stresses and this ro-
tation at any section will be proportional to
G =GJ
TL (5 )
the distance from the fixed support. It is as- (G is measured in radians).
A6.1
A6.2 TORSION
A6.3 Transmission of Power by a Cylindrical Shaft. Problem 2.

The work done by a tWisting couple T in Fig. A6.4 illustrates an aileron control
moving through an angular displacement is equal surface, consisting of a circular torque tube
to the product of the magnitude of the couple (1-1/4 - .049 in Size) supported on three
and the angular displacement in radians. If the hinge brackets and With the control rod fitting
angular displacement is one revolution, the work attached to the torque tube above the center
done equals 2 n T. If T is expressed in inch- support bracket. Find the maximum torsional
pounds and N is the angular velocity in revolu- shearing stress in the tube if the air load on
tions per minute, then the horsepower trans- the aileron is as indicated in Fig. A6.4, and
mitted by a rotating shaft may be written, also compute the angle of tWist of tUbe be-
tween horn section and end of aileron.
2 n N T
H.P. = 396000 (6)
Average load on surface =40#/0'
where 396000 represents inch pounds of work of Ii -.049 Tube(Al. Alloy)
one horsepower for one minute. Equation (6) ==-L__~=- -=- ~~ --=== =-=
++- -~--+----
-='=-:..::=!JF=
may be written:
T = H.P. x 396000 = 63025 H.P.

2nN N
(7)
24"
24" ~5"L
Fig. A6.4
EXAMPLE PROBLEMS.
Problem 1- SOLUTION:
Fig. A6.3 shows a conventional control
stick-torque tUbe operating unit. For a side The airload on the surface tends to rotate
load of 150 Ibs. on stick grip, determine the the aileron around the torque tUbe, but move-
shearing stress on aileron torque tUbe and the ment is prevented or created by a control rod
angle of twist between points A and B. attached to the torque tUbe over the center
supporting bracket.
SOLUTION: The total load on a strip of aileron one
inch wide =
40(15 x 1/144) 4.16 lb. =
Torsional moment on tube AB due to side Let w equal intensity of loading per inch
stick force of 150i = 150 x 26 = 3900 in. lb. of aileron span at the leading edge point of the
aileron surface, (see pressure diagram in
The resistance to this torque is provided by the
r-
Fig. A6.4).
150#
Then 3w + (0.5 w)12 = 4.16
24 ST Al Alloy
ontrol Stick { G = 3,800,000 psi
hence w = 0.463 Ib./in.
The total load P1 forward of the center-
Oo
N I~---- 28" -----l~ line of torque tube = 0.463 x 3 = 1.389 lb. and
P 2 the load on aileron portion aft of hinge line
1 -~~~
Bearing
Elev. Control Wire
A tl~ x .058 Tube

j
24 St. Al Al
356#
Ai eronOperating TUbe~
= 0.463 x 0.5 x 12 = 2.778 lb.
Fig. A6.3
aileron operating system attached to aileron
horn and the horn pull equals 3900/11 = 356 lb.
The polar moment of inertia of a It - 0.058 The torsional moment per running inch of
round tube equals 0.1368 in 4 • torque tUbe: = - 1.389 x 1.5 + 2.778 x 4 = 9~0
in. lb. Hence, the maximum torque, which occurs
Maximum Shearing stress = ~ = Tr/J = at the center of the aileron, equals 9.0 x 29 =
(3900 x 0.75)/0.1368 = 21400 psi. 261 in. lb.
The angular twist of the tube between
points A and B equals
261 x 0.625 = 2450 psi.
0.06678
Q =GJ
TL = 3900 x 28
3,800,000 x 0.1368
= 0.21 radians (J 0.06678 in 4 .)
or 12 degrees. Since the tUbe section is constant and the torque

AG.3
varies directly as the distance from the end of twists of non-circular cross-sections is given
the ailerons, the angle of twist Q can be com- in Art. A6.6.
puted by using the average torque as acting on
entire length of the tube to one side of horn or AG. 5 Elastic Membrane Analogy.
=
a distant L 29", hence The shape of a warped cross-section of a
non-circular cross-section in torsion is
Q = TL __-::::----;=2~6~1_:::x~2:..:9..",...._=_=_=:__5 needed in the analysis by the theory of elas-
GJ 2 x 3800000 x 0.06678 7.3 - 0.86 ticity, and as a result only a few shapes such
degrees as rectangles, elipses, triangles, etc., have
AG.4 Torsion of Members with Non-Circular Cross- been solved by the theoretical approach. How-
Sections. ever, a close approximation can be made ex-
The formulas derived in Art. A6.2 cannot be perimentally for almost any shape of cross-
used for non-circular shapes since the assump- section by the use of the membrane analogy.
tions made do not hold. In a circular shaft It was pointed out by Prandtl that the
sUbjected to pure torsion, the shearing stress equation of torsion of a bar and the equation
distribution is as indicated in Fig. A6.5, for the deflection of a membrane subjected to
namely, the maximum shearing stress is located uniform pressure have the same form. Thus if
at the most remote fiber from the centerline an elastic membrane is stretched over an open-
axis of the bar and is perpendicular to the ing which has the same shape as the cross-
radius to the stressed point. At a given dist- section of the bar being considered and then if
ance from the axis of rotation the shear stress the membrane is deflected by subjecting it to a
slight difference of pressure on the two sides,
the resulting deflected shape of the membrane
provides certain quantities which can be mea-
sured experimentally and then used in the
theoretical equations. However, possibly the
main advantage of the membrane theory is, that
Fig. AG.5 it provides a method of visualizing to a
considerable degree of accuracy how the stress
conditions vary over a complicated cross-section
of a bar in torsion.
The membrane analogy provides the follow-
Fig. AG.7
ing relationships between the deflected mem-
brane and the twisted bar.
(1) Lines of equal deflection on the membrane
'L max.
(contour lines) correspond to shearing
stress lines of the twisted bar.
is constant in both directions as illustrated in
Fig. A6.5, which means that ends of segments of (2) The tangent to a contour line at any point
the bar as it twists remain parallel to each on the membrane surface gives the direction
other or in other words the bar sections do not of the resultant shear stress at the corre-
warp out of their plane when the bar twists. sponding point on the cross-section of the
If the conditions of Fig. A6.5 are applied bar being tWisted.
to the rectangular bar of Fig. A6.6, the most (3) The maximum slope of the deflected membrane
stressed fibers will be at the corners and the at any point, with respect to the edge
stress will be directed as shown. The stress support plane is equal in magnitude to the
would then have a component normal to the sur- shear stress at the corresponding point on
face as well as along the surface and this is the cross-section of the tWisted bar.
not true. The theory of elasticity shows that
the maximum shear stress occurs at the center- (4) The applied torsion on the tWisted bar is
line of the long sides as illustrated in Fig. proportional to twice the volume included
A6.6 and that the stress at the corners is zero. between the deflected membrane and a plane
Thus when a rectangular bar tWists, the shear through the supporting edges.
stresses are not constant at the same distances
from the axis of rotation and thus the ends of To illustrate, consider a bar With a
segments cut through the bar would not remain rectangular cross-section as indicated in Fig.
parallel to each other when the bar tWists or in A6.8. Over an openin~ of the same shape we
other words, warping of the section out of its stretch a thin membrane and deflect it normal
plane takes place. Fig. A6.7 illustrates this to the cross-section by a small uniform pres-
action in a twisted rectangular bar. The ends sure. Equal deflection contour lines for this
of the bar are warped or suffer distortion deflected membrane will take the shape as il-
normal to the original unstressed plane of the lustrated in Fig. A6.9. These contour lines
bar ends. which correspond to direction of shearing
Further discussion and a summary of equa- stress in the tWisted bar are nearly circular
tions for determining the shear stresses and near the center region of the bar, but tend
A6.4 TORSION
to take the shape of the bar boundary as the From Table A6.1 it is noticed that for
boundary is approached. Fig. A6.9a shows a sec- large values of bit, the values of t he con-
tion through the contour lines or the deflected stants is 1/3, and thus for such narrow rec-
membrane along the lines 1-1, 2-2 and 3-3 of Fig. tangles, equations (6) and (7) reduce to,
A6.9. It is obvious that the slopes of the de-
flected surface along line 1-1 will be greater 3 T (8)
than along lines 2-2 or 3-3. From this we can 't"MAX = bta
conclude that the shear stress at any point on
line 1-1 will be greater than the shear stress _ 3 T
(9)
for corresponding points on lines 2-2 and 3-3. Q - bPG
The maximum slope and therefore the maximum
a. Although equations (8) and (9) have been
derived for a narrow rectangular shape, they
i, can be applied to an approximate analysis of
shapes made up of thin rectangular members
I such as illustrated in Fig. A6.10. The more
f--._.-+-.-.-
generous the fillet or corner radius, the
smaller the stress concentration at these jun-
ctions and therefore the more accuracy of these
approximate formulas. Thus for a section made
up of a continuous plate such as illustrated in
Fig. A6.8 Fig. A6.9
Slope Slope
2~ 3 ~
Fig. a
Fig. A6.9a b can be taken as centerline length for above type
of sections
stress will occur at the ends of line 1-1. The I-- b 3 ---->I .1: t
~
slope of the deflected membrane will be zero at a T3
the center of the membrane and at the four ba
corners, and thus the shear stress at these .!..
points will be zero. I---b ~ -->I 'Et~
A6.6 Torsion of Open Sections Composed of Thin Plates. Fig. A6.10
Members having cross-sections made up of
narrow or thin rectangular elements are some- Fig. (a) of Fig. A6.10, the width b can be taken
times used in aircraft structures to carry tor- as the total length of the cross-section. For
sional loads such as the angle, channel, and Tee sections such as the tee and H section in Fig.
Shapes. A6.10, the polar moment of inertia J can be
For a bar of rectangular cross-section of taken as ~ bt 3/3. Thus for the tee section of
width b and thickness t a mathematical elasticity Fig. A6.10:
analysis gives the following equations for maxi-
mum shearing stress and the angle of tWist per T 3 T
unit length. Q = GJ = G ~ bt 3
(6 ) =G 3 T
(b~t~ + bat~)
For the maximum shearing stress on leg b 1

Q =~ b ~3 G in radians - - (7)
_Tt 1 _3Tt 1 _ 3Tt~ ()
't"b 1 - - J - - ~ - b1t~" + bata" 10
Values of a and ~ are given in Table A6.1.
and for the plate b a ,
Tts 3 Tts -.(11)
= =
't"b s -J-- b 1t1 3 + bats"
If t 1 = t s = t, then
TABLE A6.1
CONSTANTS (l AND <Il

bit 1. 00 1. 50 1. 75 2.00 12. 50 3.00 4 6 8 10 co Q =G t 3
3 T
(b 1 + b s )
(12 )
(l 0.208 0.231 0.239 0.24610.258 0.267 0.282 0.299 0.307 0.313 0.333
<Il 0.141 0.196 0.214 0.22910.249 0.283 0.281 0.299 0.307'0.313 0.333
1" = 3 T (13)
t S (b 1 + b s )
A6.5
EXAMPLE PROBLEM SHOWING TORSIONAL STIFFNESS OF A6.7 Torsion of Solid Non-Circular Shapes and Thick-
Walled Tubular Shapes.
CLOSED THIN WALLED TUBE COMPARED
Table A6.3 summarizes the formulas for
TO OPEN OR SLOTTED TUBE. torsional deflection and stress for a few
Fig. A6.11a shows a 1 inch diameter tUbe shapes. These formulas are ~ased on the as-
with .035 wall thickness, and Fig. A6.11b shows sumption that the cross-sections are free to
the same tube but with a cut in the wall making warp (no end restraints). Material is homo-
it an open section. geneous and stresses are within the elastic
range.
For the round tube J 1 = 0.02474 in·.
A6.8 Torsion of Thin-Walled Closed Sections.
The structure of aircraft wings, fuselages
For open tube J s = 31 x 3.14 x .035
.§
= 0.000045 and control surfaces are essentially thin-walled
tubes of one or more cells. Flight and landing
loads often produce torsional forces on these
major structural units, thus the determination
of the torsional stress and deformation of such
structures plays an important part in aircraft
structural analysis and design.
Fig. A6.12 shows a portion of a thin-
walled cylindrical tUbe which is under a pure
Fig. A6.lla Fig. A6.11b torsional moment. There are no end restraints
on the tUbe or in other words the tube ends and
Let Q1 equal tWist of closed tUbe and Qs tube cross-sections are free to warp out of
equal tWist of open tube. The twist, will then their plane.
be inversely proportional to J since Q = J~ •
Therefore the closed tube is J1/J S 0.02474/ =
0.000045 = 550 times as stiff as the open tube.
This result shows why open sections are not ef-
ficient torsional members relative to
torsional deflection. i dS t-; F=qds
e
T
TABLE A6.3
FORMULAS FOR TORSIONAL DEFLECTION AND STRESS
= KG = twist In radians per Inch of length.

T = Torsional Moment (In. lb.).
c§ o dA,
~
q constant
G = Modulus of Rigidity.
K (1n4) From Table. Fig. A6.12 Fig. A6.13
FORMULA FOR
SECTION K
SHEAR STRESS
SOLID ELLIPITICAL Let qa be the shear force intensity at point
'0
(a) on the cross-section and qb that at point
n a3 b3 ";MAX =~(at ends (b) •
K = a2+b2 nab2
of minor axis). Now consider the segment a a ~ b of the tube
wall as shown in Fig. A6.12 as a free body. The
applied shear force intensity along the segment
SOLID SQUARE.
edges parallel to the y axis will be given the
values Qa y and qb y as shown in Fig. A6.12. For
T
D1
SOLID RECTANGLE.
K • 0.141a4 ";MAX = 0.208a3 (at
midpoint of each side).
a plate in pure shear the shearing stress at a
point in one plane equals the stress in a plane
at right angles to the first plane, hence
~ = qa and qb = qb •
Y Y
";MAX = T(3a - 1. 8b) Since the tube sections are free to warp
DI
1- 2a -l
8aZbZ
midpoint of longside. there can be no longitudional stresses on the
tube wall. Considering the equilibrium of the
segment in the Y direction,
SOLID TRIANGLE. ~Fy =0 = Qa~ - qb~ = 0, hence QaY = qb y and
6~ K = 1. 73a4
80
";MAX = -¥-
a
at midpoint of side.
therefore Qa =
qb or in other words the shear
force intensity around the tube wall is con-
stant. The shear stress at any point ~ q/t. =
If the wall thickness t changes the shear stress
For an extensive list of formulas fOr many shapes both solid and hollow,
refer to book, "Formulas For Stress and Strain" by Roark, 1954 Edition.
A6.6 TORSION
changes but the shear force q does not change, The elastic strain energy dU stored in this
or element therefore equals,
dU = qds 0
2
The product ~ is generally referred to as
the shear flow and is given the symbol q. The However the shear strain 0 can be written,
name shear flow possibly came from the fact that
the equation ~t = constant, resembles the equa- o =.!G =...1.- but q =-.!.
tion of continuity of fluid flow qS = constant Gt' 2A
where q is the flow velocity and S the tube
cross-sectiona~ area.
We will now take moments of the shear flow hence
q on the tube cross-section about some point (0).
T"
In Fig. A6.13 the force dF on the wall element
dF =
qds. Its arm from the assumed moment cen- or U =f 8A"Gt ds, the integral j
ter (0) is h. Thus the moment of dF about (0)
is qdsh. However, ds times h is twice the area is the line integral around the periphery of
of the shaded triangle in Fig. A6.13. the tube. From Chapter A7 from Castigliano's
Hence the torsional moment dT of the force theorem,
on the element ds equals,
Q =
au = ~ _T_ ds _ _ _ _ _ _ _ _ _ _ _ (17)
dT • qhds .. 2qdA
aT 4A"Gt r
since all values except t are constant, equa-
and thus for the total torque for the entire tion (17) can be written,
shear flow around the tUbe wall equals,
Q = 4AT,.G f~ --------- (18)
T = fA 2qdA and since q is constant and since T =2 qA, then also,
Q = 2~G f ~s - - - - - - - (19)
T = q2A (14)
where Q is angle of twist in radians per unit
or length of one inch of tube. For a tube length
of L
q
T
= 2A (15) Q = ~A~ f dtS - - - - - - - - - - - - - - - ( 20 )
where A is the enclosed area of the mean periph- A6.9 Expression for Torsional Moment in Terms of
ery of the tube wall. Internal Shear Flow Systems for Multiple Cell
The shear stress ~ at any point on the tube Closed Sections.
wall is equal to q, the shear force per inch of Fig. A6.16 shows the internal shear flow
wall divided by the area of this one inch length pattern for a 2-cell thin-walled tUbe, when
or 1 x t or the tUbe is subjected to an external torque.
q~, q" and q3 represent the shear load per inch
T on the three different portions of the cell
.~ • q/t .. 2At (16 )
walls.
For equilibrium of shear forces at the
TUBE TWIST junction point of the interior web With the out-
Consider a small element cut from the tube side wall, we know that
wall and treated as a free body in Fig. A6.14, - - - - - - (21)
with ds in the plane of the tUbe cross-section m
and a unit length parallel to the tube axis.
Under the shearing strains the plate element
,10-- 1 ---+I
qdS~~2
t"
t ~
4~~2 Fig. A6.16

Fig.A6.14 Fig. A6.15 n
deforms as illustrated in Fig. A6.15, that is, Choose any moment axis such as point (0).
the face a-a moves with respect to face 2-2 a Referring back to Fig. A6.13, we found that the
distance O. The force on edge a-a equals q ds moment of a constant shear force q acting along
and it moves through a distance o. a wall length ds about a point (0) was equal in
A6.7
magnitude to twice the area of the geometrical inch in the web 1-2 and (q2 - q3) for web 2-3.
shape formed by radii from the moment center to For equilibrium, the torsional moment of
the ends of the wall element ds times the shear the internal shear system must equal the ex-
flow q. ternal torque on the tube at this particular
Let To = moment of shear flow about point section. Thus, from the conclusions of article
(0). Then from Fig. A6.16, A6.9, we can write:
.T = 2q1A~ + 2q2A2 + 2q3A3 - - - - - (24)
(22) For elastic continuity, the twist of each cell

must be equal, or Q~ = Q2 = Q3'
But from equation (21), q3 = q~ - q2' From equation (19), the angular tWist of a cell
SUbstituting the value of q3 in (22) is
Q = ~ ij ds or
2AG t
2GQ = .s.A !j ds
t
(25)
- (23)
where A~ = area of cell (1) and A2 = area of ThUS, for each cell of a multiple cell struc-
cell (2). Therefore, the moment of the internal
shear system of a multiple cell tUbe carrying
ture an expression tf
d: can be written and
equated to the constant value 2GQ. Let au,
pure torsional shear stresses is equal to the
sum of twice the enclosed area of each cell
times the shear load per inch which exists in
represent a line integral f d: for cell wall
1-0, and a~2' a 20 , a 23 and a 30 the line in-
the outside wall of that cell. (Note: The web
mn is referred to as an inside wall of either
cell) •
tegrals f
~ for the other outside wall and
interior web portions of the 3-cell tube. Let
A6.10 Distribution of Torsional Shear Stresses in a clockwise direction of wall shear stresses in
Multiple-Cell Thin-Walled Closed Section. any cell be positive in sign. Now, substituting
Angle of Twist. in Equation (25), we have:
= 2GQ - - - - - - - - - - - - - - - - - - (27)
Fig. A6.17
Equations (24, 26, 27 and 28) are sufficient to

Fig. A6.17 shows in general the internal determine the true values of q~, q2' q3 and Q.
shear flow pattern on a 3-cell tube produced by ThUS, to determine the torsional stress
a pure torque load on the tube. The cells are distribution in a multiple cell structure, we
numbered (1), (2) and (3), and the area outside write equation (25) for each cell and these
the tube is designated as cell (0). Thus, to equations together with the general torque
designate the outside wall of cell (1), we re- equation, similar to equation (24), provides
fer to it as lying between cells 1-0; for the sufficient conditions for the solution of the
outside wall of cell 2, as 2-0; and for the web shear stresses and the angle of tWist.
between cells (1) and (2) as 1-2, etc.
q~ = shear load per inch = ~~t~ in the out-
A6. 11 Stress Distribution and Angle of Twist for 2-C ell
side wall of cell (1), where ~~ equals the unit Thin-Wall Closed Section.
stress and t~ = wall thickness. Likewise, For a two cell tUbe, the equations can be
q2 = ~2t2 and q3 = ~3t3 = shear load per inch in simplified to give the values of q~, q2 and Q
outside walls of cells (2) and (3) respectively. directly. For tubes With more than two cells,
For equilibrium of shear forces at the junction the equations become too complicated, and thus
points of interior webs with the outside walls, the equations should be solved simultaneously.
we have (q~ - q2) equal to the shear load per Equations for two-cell tube (Fig. A6.18): -
A6.8 TORSION
Cell 2
- - - - (29)
2GQ =38~.4 [- (q2 - q~) 335 - 1735 q",J =
.865 q~ - 5.34 q2 - - - - - - - - - - (34)
Fig. A6.18
Equating (33) and (34)
- 14.195 q~ + 8.505 q2 =0 - - - - - - - - (35)
q = ~ [ a 0 ~A 2 + a u A T J _ _ _ _ (30) The summation of the external and internal re-

'" 2 a2oA~" + a~2A2 + aO~A22J sisting torque must equal zero.
I = 25.25"
Q =GTJ - - - - - - - - - - - - - - - - - - (32)
where A = A~ + A2 . Cell ® . 050"

I = 25.30"
A6.12 Example Problems of Torsional Stresses in
Multiple-Cell-Thin-Walled Tubes. . 032"
Fig. A6.19
Example 1 - Torsional Stresses in Un-symmetrical Two-
Cell-Tube.
Fig. A6.19 shows a typical 2-cell tubular 83450 - 2 x 105.8 q~ - 2 x 387.4 q2 = 0 - {36)
section as formed by a conventional airfoil
shape, and having one interior web. An external Solving equations (35) and (36), q~ = 55.6i/in.
applied torque T of 83450 in. lb. is assumed and q2 =
92.5#/in. Since results come out
acting as shown. The internal shear resisting positive, the assumed direction of counter-
pattern is required. clockwise was correct for q~ and q2 or true
signs are q~ = - 55.6 and q2 = - 92.5.
Calculation of Cell Constants
Cell areas; - A~ = 105.8 sq. in.
q~2 =- 55.6 + 92.5 =36.9#/in. (as viewed
from Cell 1).
A", = 387.4 sq. in.
A = 493.2 sq. in. Fig. A6.20 shows the resulting shear pat-
Line integrals a = f ~s ; -
tern. The angUlar twist of the complete cell
can be found by SUbstituting values of q~ and
q2 in either equations (33) or (34), since twist
a~o =26.9
•025
= 1075'' a
u
- 13.4 - 335
- .04 -
of each cell must be the same and equal to
twist of tube as a whole •
=~
25.25 + 15.7 25.3
.05 + .032
=1735 Fig.A6.20
Solution by equating angular tWist of each cell.

General equation 2GQ =
flow of q is posit1ve.
if d:. Clockwise
Cell 1 Subt. in general equation
2GQ = 1;5.8 [-q~ x 1075 + (-q~ + q2) 335J =

SOLUTION BY SUBSTITUTING IN EQUATIONS (29) & (30)
- 13.33 q~ + 3.165 q", - - - - - - - - - - (33)
A6.9
= .002456 T
1735 x 105.8 + 335 x 493.2 T J

[ 1735 X 105.8" + 335 X 493.2 2 + 1075 X 387.4 2 2
= .000665 = .000665 = 55.6t/in.
T X 83450 2 2
=4 [ 916.7 x 100 + 333.3 X 200 + 866.7 X 100
2
J
q
2
= 1. r- a 1o A2 + aaA
2 La 20A1 2 + a 12 A2 + a 1o A2 2
=1 T
J
866.7 x 333.3 + 333.3 x 916.7 + 916.7 x 866.7
= 89.76
1075 X 387.4 + 335 X 493.2 ] T Q =...!.. = x ~9. 76 = .01116 ~ (rad.) per unit
[ 1735 X 105.8 2 + 335 X 493.2 2 + 1075 X 387.4 2 2 JG G length of cell.
= .001107 T = .001107 X 83450 = 92.5./in. A6.13 Example 3 - ':Chree-Cell-Tube.

Fig. A6.22 shows a thin-walled tUbular
Example Problem 2. section composed of three cells. The internal
shear flow pattern will be determined in re-
Determine the torsional shear stresses in sisting the external torque of 100,000"1 as
the symmetrical 2 cell section of Fig. A6.21
shown.
when sUbjected to a torque T. Neglect any re-
sistance of stringers in resisting torsional
moment.
~ Fig.A6.21
SOLUTION:
10"
/ .03" '\ .03"
~~~~~:::J~
Calculation of
terms .04"
Area of cells: -
2
A1 = 100 A2 100
SOLUTION:
A - A1 + A2 200
Line integrals a =
f ~s : -
Calculation of cell constants
Cell areas: -
100 A3 = 100
a 10 -- .05
10 + 20 = 866.7
.03
Line integrals a =J!~t .
- 10 - 333 3
- .03 - . nx 10
a 12 = 10 = 200
2 x .025 =
a 10 = 629
.05
a 20 -- .03
20 + .1 0 = 916.7 _ 20
04 a20 - .03 = 667 a23 = ~= 333
.03
Solution of Equations from Article A6.11: -
a30 = .03
20 + ~
.04
= 917
Equating the external torque to the internal re-
sisting torque: -
_1 r. 916.7 X 100 + 333.3 X 200 J
q1 -2 L916.7 X 100 2 + 333.3 X 200 2 + 866.7 X 1002 T
Subst i tut ing:
= .002540 T
78.6 q1 + 200 q2 + 200 q3 - 100,000 =0 - (37)
Writing the expression for the angular tWist

of each cell: -
Cell (1)
If- 866.7 X 100 + 333.3 X 200 1 1
-2L916.7 X 100 2 + 333.3 X 200" + 866.7 x 100"J T 2GQ = A [q1 a 10 + (q1 - q1) aaJ
1
A6.10 TORSION
Substituting: A6.14 Torsional Shear Flow in Multiple Cell Beams by

Method of * Successive Corrections.
The trend in airplane wing structural de-
1 sign particularly in high speed airplanes 1s
2aG = 39 .3 [629 q1. + 200 q1. - 200 q ..J - - - (3S)
toward mUltiple cell arrangement as illustrated
in Fig. A6.24, namely a wing cross-section
Cell (2) made up of'a relatively large number of cells.
Fig. A6.24
Substituting:
With one unknown shear flow q for each cell,
the solution by the previous equations becomes
2aG = 1;0 [200 q .. - 200 q1. + 667 q .. + 333 q .. - qUite laborious.
The method of successive approximations
provides a simple, rapid method for finding the
333 qaJ - - - - - - - - - - - - - - - - - (39) shear flow in multiple cells under pure torsion.
EXPLANATION OF SUCCESSIVE CORRECTION METHOD.

Cell (3) Consider a two cell tube as shown in Fig.
a. To begin with assume each cell as acting
independently, and
SUbject cell (1) to
Substituting:
such a shear flow q1.
as to make aG1. 1.
From equation
= (2)
cen\T
(19) we can write,
2GG = 1;0 [ 333 qa - 333 q .. + 917 qaJ - - - (40) aG =~ ~ ds
2AJt
Solving equations (37) to (40), we obtain, Now assume aG = 1,
then Fig. a
q1. = 143.4#/in.
_ 2A
q .. = 234.1#!in.
qa = 20S.S#/in. q-fct;
q .. q1. = 90.7!l=/in.
Since practically all cellular aircraft
q.. - qa = 25:3#/in. beams have wall and web panels of constant
thickness for each particular unit, the term
tI; ds L
j t for simplicity will be written ~t' where
L equals the length of a wall or web panel and
t its thickness. Thus we can write,
Fig.A6.23
2 A
q= -------------- (41)
L
~cell t
234.1#/in.
Therefore assuming aG1. =
1 for cell (1) of
Fig. a, we can write from equation (41): -
Fig. A6.23 shows the resulting internal
shear flow pattern. The angle of tWist, if de- * Based on Paper, "Numerical Transformation Procedures
sired, can be found by SUbstituting values of for Shear Flow Calculations" by S. U. Benscoter.
shear flows in any of the equations (3S) to (40). Journal, Institute of Aeronautical Sciences, Aug. 1946.
A6.11
2 x 89.3
25.7 + 10 -_ 178.6 _- 0 • 212 lb. I in.
(1-2) is sUbjected to a shear flow q,. =- .109#
842 lin. (counterclockwise with respect to cell (1)
.04 .05 and therefore negative), in addition to the
shear flow q~ =
.212 of cell (1). The negative
Fig. b. shows the results. shear flow q,. =-
.109 on web 1-2 decreases the
twist of cell (1) as calculated above With the
t-- q = .109
resulting value for GQ 1 =
0.875 instead of 1.0
as started With.
~ Thus in order to make GQ 1 = 1 again, we
f (2) i will have to add a constant shear flow q~ to
cell (1) which will cancel the negative twist
f due to q" acting on web (1-2). Since we are
considering only cell (1) we can compare cell
Fig. b Fig. c wall strains instead of cell tWist since in
equation (41), the term 2A is constant.
In a similar manner assume cell (2) sub- Thus adding a constant shear flow qi to
jected to a shear flow q" to make GQ,. = 1, Then cancel influences of q,. on web 1-2, we can write:
q,.
2A"
=Z L
2 x 39.3
lQ. + 15.7 = 7~i; = 0.109#/in.
qi (z ~) cell (1) - q" (~) web 1-2 =0
t .05 .03 hence
Fig. c shows the results.

Now assume the two cells are joined to- - - - - - - - - (42)
gether with the interior web (1-2) as a common
part of both cells.
See Fig. d. The in- q = .109
terior web is now Substituting values in equation (42)
sUbjected to a re-
sultant shear flow
of q~ - q"
.109) =
=
(.212 -
.103t/in.
Obviously this change
q~I -- q" t .~ J= 200
25.7 + ~ 842 q" = .237 q"
.04 .05
of shear flow on the Fig. d
interior web will cause
the cell tWist to be different for each cell in- Thus to make GQ~ equal to 1 we must correct
stead of the same when the cells were considered the shear flow in cell (1) by adding a constant
acting separately. To verify this conclusion shear flow equal to .237 times the shear flow q"
the tWist measured by the term GQ will be com- in cell (2) which equals .237 x .109 = .0258~/in.
puted for each cell. Since this shear flow is in terms of the shear
flow q" of the adjacent cell it will be referred
Cell (1), to as a correction carryover shear flow, and
will consist of a carryover correction factor
times q".
c: 2 x 189 •3 [.212 x 2~oZ + (.212 - .109) .10~J = Thus the carryover factor from cell (2) to
cell (1) may be written as
= 0.875
C.O.F. (2 to 1) which equals
Cell (2), GQ" =_1_
2A.,
Z q~ =
t
= 2 x 139.3 [.109
15 .237 as found above in substitution in equation
x ..03
7 - (.212 - .109) .05
10J = (42) •
Now consider cell (2) in Fig. d. In
bringing up and attaching cell (1), the common
= 0.466 web (2-1) is subjected to a shear flow of q~ =
- 0.212*/in. (counterc~ockwise as viewed from
Since Q1 must equal 9" if the cross-section cell 2 and therefore negative). This additional
is not to distort from its original shape, it is shear flow changes GQ" tWist of cell (2) to a
evident that the above shear flows are not the relative value of 0.4375 instead of 1.0 (see
true ones when the two cells act together as a previous GQ" calculations). Therefore to make
unit. GQ,. equal to 1.0 again, a corrective constant
Now consider cell (1) in Fig. d. In bring- shear flow q~ must be added to cell (2) to
ing up and attaching cell (2) the common web
A6. 12 TOR S ION
cancel the tWist effect of q1 0.212 on web = Considering Cell (2)

(2-1). Therefore we can write,
Q~ =q~ (C.O.F.) 1 to 2 = .0258 x .277 =
qk (Z ¥) cell (2) - q1 (t) web (2-1) = 0 .00717#/in.
hence Fig. f shows the resulting second set of

corrective constant shear flows for each cell.
Since our corrective
qk = q1 (i) web (2-1)
- - - - - - - - - - (43)
shear flows are
rapidly getting q'~ = • 00717
(z i) cell (2) smaller, the con-
tinuation of the
:;-- -
= .0139
-
'q' " 'f

Substituting in equation (43): - process depends on \ (1) I (2) }
the degree of ac-
curacy we wish for ~--
~
+t #
q~ =q1 (~~g) = .277 q1 = .277 x .212 = .0587:!Vin. the final results. Fig. f
Suppose we a dd one
Thus the carryover factor from cell (1) to more set of cor-
cell (2) in terms of q1 to make Gg" = 1 again can rective constant shear flows q~ and q~, Using
be written the carryover correction factors previously
found we obtain,
C.O.F. (1 to 2)
= (4-) web (2-1) = 200 = .277 qT = .237 x .00717 = .0017~/in.
(z i) cell (2)
723
q~ = .277 x .0139 = .00385i/in.
Fig. e shows the constant shear flow qk and Fig. g shows the results.
qi that were added to make Gg =
1 for each cell.
However these cor- The final or re-
rective shear flows q'" =. 0587 sulting cell shear
were added assuming flows then equal the
the cells were again original shear flows
independent of each plus all corrective
other or did not cell shear flows, or
have the common web
(1-2) • Thus in Fig. e q1 (final) = q1 + qi +
Fig. g
bringing the cells q~ + qT
together again the interior web is subjected to
be resultant shear flow of qi - q~. In other
words if we were to add the shear flows of Fig. e
q "(final) = qo~ + q~ + q~ + q~
~ ~ ~
to those of Fig. d, we would not have Gg 1 and Gg"

equal to 1. The resulting values would be closer Fig. h shows :;;.-_ __ ,,= . 1787
the final results.
to 1.0 than were found for the shear flow system
of Fig. d. To check the
final tWist of each
1q1=.2534 ~t ~
Considering Fig. e, we will now add a second
set of corrective shear flows q~ and q~ to cells
cell the value Gg t~ (1) W~
(1) and (2) respectively to make Gg 1 and Gg" = 1
for cells acting independently.
Considering cell (1), and proceeding with
will be computed for
each cell using the
q values in Fig. h.
---Fig. h
same reasoning as before, Cell (1)
q~ ~ ~1 cell (1) - qk (~j web 1-2 =0 @1 = 2 x 189.3·~

[%34 x 25.7 10 1
.04 + .0747 x .05J = .997
Hence
Gg" = 2 x 139.3 [ .1787 x 15.7 10J = .997

.03 - .0747 x .05
q~ = --r=-"'.->----'--- .0587 x .237 = .0139#/in.
A6.15 Use of Operations Table to Organize Solution by
Successive Corrections.
or Operations Table 1 arranges the calculations
so that the steps dealing with the corrective
q~ = (C.O.F.) 2 to 1 times q~ = .237 x .0587 shear flows can be carried out rapidly and with
=.0139 a minimum of thought.
A6.13
(~) 10
OPERATIONS TABLE 1
Cell 1 Cell 2
web 1-2 .05
(1-2) = = 10 + 30 = .40
C.O.
1
2
C.O. Factor
Assumed a
.277
.212
.237
109
~ (~) cell 2 .05 0.1
3
4
C.O. a
C.O. a
.0258
.0139
= .0587
.00717
10
- (~) web 2-1
.n017 - - 00~8"
6 Fna .<::>34 lin
7 T = 2A
Total T
45.2"# 14.1"# C.O.
(2-1) = = .: = .25
~ (~)
8 59.3"fI
cell 1 .05
Explanation of Table 1 Line 2 gives the shear flow q in each cell,
Line 1 gives the carryover factors for when it is assumed each cell is acting separ-
each cell, computed as explained before. Line 2 ately and is sUbjected to a unit rate of twist
gives the necessary constant shear flow q in or GQ '" 1.
each cell to give unit rate of twist to each The calculations for the q values are as
cell acting independently. (GQ = 1). Line 3 follows: -
gives the first set of constant corrective shear
flows to add to each cell. The corrective q re- For cell (1)
ferred to as the carryover q or C.O.q in the
table consists of the q in the adjacent cell 2 x 100
40 = .25
times the C.O. factor of that cell.
Thus .237 x .109 = .0258 is carryover from .05
cell (2) to cell (1) and .277 x .212 = .0587 is For cell (2)
carryover from cell (1) to cell (2).
Line 4 gives the second set of corrective
carryover shear flows, namely .277 x .0258 = 2 x 100 "'.4
.00717 to cell (2) and .237 x .0587 - .0139 to 10 + ~
.05 0.1
cell (1). Line 5 repeats the corrective carry
over process once more. Line 6 gives the final A1 and As equal cell area of cells (1) and (2)
q values which equal the original q plus all respectively. L = length of wall element and t
carryover q values. its corresponding thickness.
Example Problem 1 (2 cells) In order not to start out with decimal values
Determine the internal shear flow system of q, the values above will arbitrarily be
for the two cell tUbe in Fig. A6.25 when sub- multiplied by 100 to make q1 =
25#/in. and
jected to a torque of 20,000 in. lbs. qs = 40~/in. Since we want only relative
I-- 10" - - \ - - 10" --l values of terms this is permissible. These
values are shown in line 2 of Table (2). The
T
10" .05
.05
Cell
1 .05
0.10
Cell
2 O. 10
corrective carryover process proper is started
in line 3 of table (2). In cell (1) the amount
carried over of q1 =
25 to cell (2) equals the
C.O. factor times 25 or 0.4 x 25 10 which is =
written along the vertical line under cell (2).
1 .05 O. 10
Fig. A6.25
Likewise in cell (2) the amount of qs - 40 that
is carried over to cell (1) equals the C.O.
factor times qs or .25 x 40 =
10, which is
written along the vertical line lmder cell (1).
OPERATIONS TABLE 2
The second set of corrective carryover
constant shear flows are given in line 4 of
Cell 1 Cell 2
Table (2), thus, .4 of the q1 = 10 = 4.0 is
1 C.O. Factor 0.4 0.25 carried over to cell (2) and .25 x value of
2
3
Assumed a
C.O.
25 0
10 ~
40 0
10 q1 = 10 = 2.5 is carried over to cell (1). Line
4 C.O. 2.5 4 5 repeats the process namely 0.4 x 2.5 = 1 to
5
6
7
C.O.
C.O
CO
1.0
o 25
o 10
1
u.4 cell (2) and .25 x 4 =
1.0 to cell (1).
8 F1na "ll.ll:> :>i
This process of carrying over values of q
9 2Aa = " 7770 is continued until the values are small or
11100
10 T Total 18870
11 a far T=20 000 41 2 58 9
negligible. Line 8 gives the final q values in
each cell as the summation of the assumed q
Explanation of solution as given in Table (2); value plus all carryover values of q. Thus in
cell (1) q 38.85#/in. and for cell (2)
Line 1 gives the values of the carryover q = 55.5. Line 9 gives the torque that these
factors. The values are calculated as follows: values of cell shear flow can produce.
C.O. factor cell (1) to cell (2); T = 2Aq
A6 14 TORSION
For cell (1) T = 2 x 100 x 38.85 = 7770 in. lb. OPERATIONS TABLE 3
For cell (2) T 2 x 100 x 55.5 =
11100 in. lb. = Cell 1 Cell 2 Ce.U 3
Line 10 gives the sum of the above two values 1 COFactor 23 .21 .284 .288
2 Assumed q 25.30-... ..- 18.45 _ .--10.70
which equals 18870 in. lb. 3 C 0 3 88....- ~ 5.82 3.u.. ..,;; 5.25
The original requirement of the problem was 4 C 0 1 87 - ... 0.894 1 51-- 2.53
5 CO. 0.505 0.430 0.73 0.685
the shear flow system for a torque of 20,000"#. 6 C.O. 0.244 0.116 0.197 0.330
Therefore the required q values follow by direct 7 CO. 0.066 0.056 0.095 O.UI'II
proportion, whence I'

9
1":
2Aa
na = to ;j.
9200
,I',
6020
.1. ;j,
2215
l!l.oll
10 T Total = 9200 + 6020 + 2215 = 17435"#

20000 11 q - for T =100,000 183 180 112.5
q1 = 18870 x 38.85 = 41.2#/in.
20000 Explanation of solution as given in Table 3: -

qs = 18870 x 55.5 = 58.9#/in.
Cell (1) Cell (2) Cell (3)
These values are shown in line 11 of Table
2. Check on tWist of cells under final q values. A1 - 144 As = 96 A", = 56. 5
~l=.= 1140 L L
The relative total strain around each cell
t ~t= 1040 ~ t = 1055
1 L
boundary is given by the term A ~ q t for the
cell. Shear flow q for GQ = 1 for each cell acting in-
Thus for cell (1) dependently: -
Cell (1) q 2A 1 288

= L: l=. = 1140 = .253
1
A1 ~ q ~ = 1~0 [41.2 t.30~) + (41.2 - 58.9) .10~J= t
_ 2A s _ 192 _
= 212 Cell (2) q - ~ L - 1040 - .1845
t
For cell (2),
_ 2A", _ 113
Cell (3) q
- ~ L - 1055 = .107
l
As ~ q i = 1~0 [(58.9 - 41.2) .~~ + 58.9 x (~.Ol)J t
To avoid small numbers these values of q

= 212 are multiplied by 100 and entered on line 2 of
the table. Calculation of carry factors as
Thus both cells have the same tWist. given in line 1 of Table 3.
In the above calculations q1 and qs act Cell (1) to (2)
clockwise in each cell, hence the shear flow on
the interior common web is the difference of the
two q values. (f) web 1-2
Example Problem 2. Three cells

C.O. (1-2) -
~
(L)
t cell 2
The three cell structure in Fig. A6.26 is
sUbjected to an external torque of - 100,000 in. Cell (2) to (1)
lb. Determine the internal resisting shear flow
pattern.
C.O.
_(~i web 2-1 _ 240 _
(2-1) -TI4Q- .21
1 2 " - + - 8"-1 L: (~) cell 1
T .04 .032
.04
Cell (2) to (3)
12" .04 Cell 1 .05 Cell 2 Cell 3

=(~) web 2-3 _ 300
C.O.
(2-3) - 1055 = .284
~ (~) cell 3
1 "--_-----'-.-"'0.....
4 -'---_.:..:.0"'3'-'=2'-----L_____
Fig. A6.26
A6.15
Cell (3) to (2) as given in Table 4. Line 10 shows the cor-

rection of q values to develop a resisting
_(-fJ web 2-3 - 300 - 288
torque of 100,000"#. The multiplying factor is
100,000/80630.
C. 0 • (3-2) - ( L) - 1040 - •
Z t cell 2 A6.16 Torsion of Thin-Walled Cylinder HaVing Closed
Type Stiffeners.
The balance of the solution or procedure in The airplane thin-walled structure usually
Table 3 is the same as explained for Problem 1 contains lon~itudinal stiffeners spaced around
in Table (2). It should be noticed that cell the outer walls as illustrated in Fi~s. A6.28
(2) being between two cells receives carryover and A6.29.
q values from both adjacent cells and these two
values are added together before being distribu-
ted or carried over again to a djacent cells.
For example consider line 3 in Table (3) and
cell (2). The q value 5.82 representing .23 x
25.3 is brought over from cell (l) and the q
value 3.08 = .288 x 10.70 from cell (3). These
8
Open Type Stiffener
2
Closed Type Stiffener
two values are added together or 5.82 + 3.08 =
8.90. The carryover q to cell (1) is then .21 For the open type stiffener as illustrated in
x 8.90 = 1.87 and to cell (3) is .284 x 8.90 = Fig. A6.28, the torsional rigidity of the in-
2.53. dividual stiffeners as compared t 0 the torsional
In line 8, the final q in cell (2) equals rigidity of the thin-walled cell is so small
the original q of 18.45 plUS all carryover q to be negligible. However a closed type
values from each adjacent cell. stiffener is essentially a small sized tUbe and
Line 10 in Table (3) shows the total torque its stiffness is much greater than an open
developed by the reSUltant internal shear flow section of similar size. Thus a cell With
is 17435"#. Since the problem was to find the closed type stiffeners attached to its outer
shear flow system for a torque of - 100,000 in. walls could be handled as a mUltiple cell struc-
lb., the values of q in line 8 must be mUltipli- ture, with each stiffener acting as a cell with
ed by the factor 100,000/17435. a common wall with the outside surrounding cell.
Line 11 shows the final q values. Since in general the stiffness provided by the
stiffeners is comparatively small compared to
Example Problem 3. Four cells. the over-all cell, the approximate simplified
Determine the internal shear flow system procedure as given in NACA T.N. 542 by Kuhn can
for the four cell structure in Fig. A6.27 when be used to usually give sufficient accuracy. In
subjected to a torsional moment of - 100,000 in. this approximate method, the thin-walled tube
lb. and closed stiffeners are converted or trans-
+- -J..--
~
" ~ 20 20 40 ·1 formed into a single thin-walled tube by modi-
fying the closed stiffeners by either one of the
00 625 (1600)
04 .04 .032 following procedures: -
.06 .05 . 0 4 . 025
16" 1 o.5~ 1300j ~ 12601 8 14001
(1) Replace each closed stiffener by a doubler
I~~ ~ eN M plate having an effective thickness
~ e (1600) te=tSTs/d, and calculate ¢ dS/t with these
(500) (625) doubler plates in place. The enclosed area
Cell 1 Cell 2 Cell 3 Cell 4 of the torsion tube still remains (A) or
the same. See Fig. A6.30.
Fig. A6.27
OPERATIONS TABLE 4
Cell 1 Cell 2 Cell 3 Cell 4
1 C 0 F 173 298 .153 .181 .078 .164

2 0 22.4 __ 38.8_ __ 28.4_ ......20.50 Original
3 C 0 11 6~ ::--3 88 5.14"'- 5.94 3.36~ ~2.22 ~ffener
4 CO 2.69 k ' -2.00 1.68'" ~1.38 0.36- ..... 0.73
5 C.O 1 10 0.47 0.32 0.57 o 12 0.13
R o 15 o 19 o 12 o 08 o 02 o 15 Transformation
3 .":> :>: .l>U 4U .";j "".l>;j Transformation by Procedure (2)
8 :lAo 7620 31560 22550 18900 by Procedure (1)
9 Total T = 80630 I---- d ---...\
10 q 47.1 65.3 49.9 29.41 7 d --l bK
I I ,
In Fig. A6.27 the values in the rectangles lte=tsK+t st ~
represent the cell areas. The values in ( )
represent the L/t values for the particular wall Fig. A6. 30
or web. After studying example problems 1 and 2 Fig. A6.31
one shOUld have no trouble checking the values
A6.16 TORSION
(2) Replace the skin over each stiffener by a

"liner" in the stiffener having a thickness
te =tSK dis. (See Fig. A6.31.) The en-
closed area (A) of the cell now equals the
original area less that area cross-latched
in Fig. A6.3l. Partly lateral shear.
and partly torsional shear ~"",",~/
Procedure (1) slightly overestimates and
procedure (2) slightly underestimates the stiff-
ness effect of the stiffeners.
The corner members of a stiffened cell are
usually open or solid sections and thus their
torsional resistance can be simply added to the
torsional stiffness of the thin-walled over-all
cell.
A6.17 Effect of End Restraint on Members Carrying
Torsion. Fig. A6.32
The equations derived in the previous part
of this chapter assumed that cross-sections
throughout the length of the torsion members
were free to warp out of their plane and thus
there could be no stresses normal to the cross-
sections. In actual practical structures re- A6.18 Example Problem lllustrating Effect of End
straint against this free warping of sections Restraint on a Member in Torsion.
is however often present. For example, the Fig. A6.33 shows an I-beam subjected to a
airplane cantilever wing from its attachment to torsional moment T at its free end. The prob-
a rather rigid fuselage structure is restrained lem will be to determine what proportion of the
against warping at the wing-fuselage attachment torque T is taken by the flanges in bending and
point. Another example of restraint is a heavy what proportion by pure shear, at two different
wing bulkhead such as those carrying a landing sections, namely 10 inches and 40 inches from
gear or power plant reaction. The flanges of the fixed end of the I-beam.
these heavy bulkheads often possess considerable 7"-''-\t'-..l....!;/- Fixed
lateral bending stiffness, hence they tend to End
prevent warping of the wing cross-section.
Since only torsional forces are being considered \---1. 75''-1
here as being applied to the member, the stress- F
1
l" ..
~ .;:: 3
es produced normal to the cross-section of the
member namely, tension and compression must add 1 11
up zero for equilibrium. Thus the applied tor- 3.5"

bw=3.3" II h = 3.4".... I
l~'f:::::::~ =f
que is carried by pure torsion action of the
member and part by the longitudional stresses l'
II
II
II F
1 0""
..
normal to the member cross-sections. The per- '" ."
centage of the total torque carried by each Fig. A6. 33 Fig. A6. 34 Fig. A6. 35
action depends on the dimensions and shape of Fig. A6.34 shows the torque dividing into two
the cross-section and the length of the member. parts, namely the couple force F-F formed by
Fig. A6.32 illustrates the distortion of an
open section, namely, a channel section sUbject- bending of the flanges of the I-beam and the
pure shearing stress system on the cross-
ed to a pure torsional force T at its free end
section. Fig. A6.35 shows the tWisting of the
and fixed at the other or supporting end. Near
the fixed end the applied torque is practically section through a distance 0.
all resisted by the lateral bending of the top The solution will consist in computing the
and bottom legs of the channel acting as short angle of tWist Q under the two stress con-
cantilever beams, thus forming the couple with H ditions and equating them.
forces as illustrated in the Figure. Near the Let TB be the proportion of the total
free end of the member, these top and bottom torque T carried by the flanges in bending
legs are now very long cantilever beams and thus forming the couple F-F in Fig. A6.34. From
their bending rigidity is small and thus the Fig. A6.35, the angle of twist can be written
pure torsional rigidity of the section in this
region is greater than the bending rigidity of
the channel legs. QB =_0_
0.5h
=3F EI
L = § TB
3 h£
3 L3
EI
0.5h
A6.17
Note: The deflection of a cantilever beam with The shaft rotates at constant speed. The dif-
a load F at its end equals FL 3 /3EI, and I the ference in belt pullan two sides of a pulley
moment of inertia of a rectangle about its cen- are shown on the figure. Calculate the maximum
ter axis = tb 3 /12. torsional shearing stress in the 1-3/4 inch
shaft between pUlleys (1) and (2) and between
hence (2) and (3).
(2) A 1/2 HP. motor operating at 1000 RPM
Now let Tt be the portion of the total torque rotates a 3/4-.035 aluminum alloy torque tube
carried by the member in pure torsion. The ap- 30 inches long which drives the gear mechanism
proximate solution for open sections composed of for operating a wing flap. Determine the maxi-
rectangular elements as given in Art. A6.6, mum torsional stress in the torque shaft under
equation (12) will be used. full power and RPM. Find the angular deflection
of shaft in the 30 inch length. Polar moment of
=
Tn
fi = 3 Tt L inertia of tube .01 in. Modulus of rigidity
t Gt 3 (b + b w) =
G 3800,000 psi.
in. lb.
Equating fiB to fit , we can write
TB = 3 E h2 b3 II
Tt 8L 2 Gt 2 (2b + bw) I / I
I
\ / I
SUbstituting values when L = 10 inches. " '1-----
I /,
I /
I
TB = 3 x 10.5 X 10 6 x 3.4 x 1.75 = __- - r/ .......
/
I
Tt 8 X 10 2 x 3.7 X 10 6 x 0.1 2 (2 x 1.75 + 3.3)
= 9.65
Fig. A6. 36
Therefore, percent of total torsional T
taken by bending of flanges equals,
(3) In the cellular section of Fig. A6.36
TB TB determine the torsional shear flow in resisting
TB + Tt (100) = TB + TB (100) = 90.5 percent. the external torque of 60000 in. lb. Web and
9.65 wall thickness are given on the figure. Assume
the tube is 100 in. long and find the torsional
If we consider the section 40 inches from deflection. Material is aluminum alloy. (G =
the fixed end, then L = 40 inches. Thus if 40 2 3800,000 ps 1. )
is placed in the above substitution instead of (4) In Fig. A6.36 remove the interior .035
10 2 the results for TB/Tt would be 0.602 and web and compute torsional shear flow and de-
the percent of the torque carried by the flanges flections.
in bending would be 36 as compared to 90.5 per-
cent at L = 10 inches from support. Thus in T = 100,000 in. lb.
general the effect of the end restraint decreases
rapidly with increasing value of L.
The effect of end restraint on thin walled
n
tubes With longitudional stiffeners is a more
involved problem and cannot be handled in such
a simplified manner. This problem is considered
in other Chapters.
A6.19 Problems.
T
12"
Fig. A6.37
I- 12" -+-8"--.j
(5) In the 3-cell structure of Fig. A6.37
Fig. A6.36 determine the internal resisting shear flow due
to external torque of 100,000 in. lb. For a
(1) In Fig. A6.36 pUlley (1) is the driving length of 100 inches calculate twist of cellular
pulley and (2) and (3) are the driven pulleys, structure if G is assumed 3,800,000 psi.
A6.18 TORSION
(6) Remove the .05 interior web of Fig. I----- 14" 'I' 14"----1
A6.37 and calculate shear flow and twist.
.04 .05
.03
.04
.03
.03 10"
T
(a)
(7) Remove both interior webs of Fig.
A6.37 and calculate shear flow and twist. .03 .03 -:L
1-6"-+8" +-10" --!-10"--I
(8) Each of the cellular structures in
Fig. A6.38 is subjected to a torsional moment
T .03 .03 .03 .03
10" .04 04 .035 .035 .03 .0 .03 (b)
of 120,000 in.lbs. Using the method of
successive approximation calculate resisting
shear flow pattern.
1 .03 .03 .03 .03
All interior webs = . 05 thickness

Top skin . 084" thickness
Bottom skin . 064" thickness
Fig. A6.38
The big helicopters of the future will be used in many important industrial and military opera-
tions. The helicopter presents many challenging problems for the structures engineer.
(Sketches from United Aircraft Corp. Publication "BEE-HIVE", Jan. 1958. Sikorsky Helicopters)
CHAPTER A7
DEFLECTIONS OF STRUCTURES
ALFRED F. SCHMITT'1}
A7.1 Introduction.
Calculations of structural deflections are P
important for two reasons: -
(1) A knowledge of the load-deformation Work =L" Pi
I~I
6 Oi
characteristics of the airplane is of primary in the limit as
importance in studies of the influence of
structural flexibility upon airplane perform- 66 -0
ance.
Work =1,0 P dO
(2) Calculations of deflections are neces-
sary in solving for the internal load distri-
butions of complex redundant structures.
Fig. A7.1
The elastic deflection of a structure
under load is the cumulative result of the curve is a straight line whose equation is
strain deformation of the individual elements P = k6 and the strain energy is readily dom-
composing the structure. As such, one method puted to be
of solution for the total deflection might in-
k62 p2
volve a vectorial addition of these individual U 2 or, equally, U = 2k'
contributions. The involved geometry of most
practical structures makes such an approach
A7.3 Strain Energy Expressions for Various Loadings.
prohibitively difficult.
For complex structures the more popular STRAIN ENERGY OF TENSION
techniques are analytical rather than vectorial.
They deal directly with quantities which are not A tensile load S acting at the end of a
themselves deflections but from which deflec- uniform bar of length L, cross sectional area
tions may be obtained by sUitable operations. A and elastic modulus E causes a deformation
The methods employed herein for deflection cal- 6 = SL/AE. Hence S = A~ 6 and
cUlations are analytical in nature.
A7.2 Work and Strain Energy. rO _ AE 6 2
U = Jo Sd6 - L 2" (1)
Work as defined in mechanics is the prod-
uct of force times distance. If the force Alternately,
varies over the distance then the work is com-
puted by the integral calculus. Thus the work
done by a varying force P in deforming a bOdy (2 )
an amount 6 is
Work = ~ Pd6 and is represented Equations (1) and (2) are equivalent ex-
by the area under the load deformation (P-6) pressions for the strain energy in a uniform bar,
curve as shown in Fig. A7.1. the former expressing U in terms of the de-
If the deformed body is perfectly elastic formation and the latter expressing U in~erms
the energy stored in the body may be completely of load. The second form of expression is more
recovered, the body unloading along the same convenient for general usage and succeeding
P-6 curve followed for increasing load. This strain energy formulas will be put in this form.
er.ergy is called the elastic strain energy of For a tensile bar having non-uniform prop-
deformation (hereafter the strain energy, for erties (varying AE), or for which the axial load
brevity) and is denoted by the symbol U. Thus S varies, the strain energy is computed by the
U = ~ Pd6. Should the body be linearly calculus. Thus the energy in'a differential
element of length dx is given by eq. (2) as
elastic (as are most bodies dealt with in
structural analysis) then the load-deformation S2dx
dU = 2AE where Sand AE
are average values over the length of element.

A7.1
A7.2 DEFLECTIONS OF STRUCTURES
The total energy in the bar is obtained by

summing over the length of the bar.
U = n2q~C
2AE
fL (_ 1 sin ~
)2 dx =
2 o
U =J dU =J S dx
2AE 2(25)2 (40)2 _
9.87(2)(10 x 106)(.227)(40) - 0.0920 in lbs.
Example Problem 1
A linearly tapered aluminum bar is under STRAIN ENERGY OF FLEXURE
an axial load of 1500 lbs. '3.S tJn.ovrr, '.n Fig. A7.2.
Find U. A uniform beam of length L under the action
of a pure couple undergoes an angular rotation
Ao proportional to the couple. Thus, from elemen-
TUTx
L40" ~
A(x) = A o (l _X)
Ao= 2 m
.8~
tary strength of materials
t 1500# Fig. A7.2
Solution: where the constant EI, the product of elastic

modulus by section moment of inertia, is called
From statics, the internal load S -1500'* the lib-ending st iffness" •
at every section. Since the rotation Q builds up linearly
with M the strain energy stored is
U = fL S2dx = (40(_1500)2dx
o 2AE Jo 4(1 _ ~)l0 U = !2 MQ
80 X 10 6 ( .... x/40
or
2(-1500)2 d1: r
= 10 6 }o 2-1: = 4.5 ln 2 inch lbs.
introduced
for conven-
ience)
(4)
Note that although P was a negative (com-

For a beam having non-uniform properties
pressive) load the strain energy remained
positive. (varying EI) and/or for which M varies along
the beam, the strain energy is computed by the
calculus. From eq. (4) the strain energy in a
Example Problem 2
beam element of differential length is
Find U in a uniform bar under a running
load
roc
q = qo cos 2L (Fig. A7.3).
Hence summing over the complete beam to get the

total strain energy one has
L • 40"
2
A • 2 in (5)
6
E • lOxl0 pB1
'1. • 25 lb/in
0 ~xample Problem 3
For the beam of Fig. A7.4 derive the strain
energy expression as a function of Mo.
Fig'. A7.3
/'M C rEI = constant

Solution: (! !
The equation for S (x) was found by statics. I--L~
Fig. A7.4
Solution:
The bending moment diagram (Fig. A7-4a) was

= qo 2L sin rocJL
n 2L
= qo 2L (1 -sin
n 2L
roc) found and an analytic expression written for M.
x
SUbstitution into eq. (3) gave •
A7.3
(6 )
where the constant GJ, the product of the elastic

modulus in shear by the cross section polar mom-
x ent of inert ia, is called the "tors i onal st iff-
ness".
Fig. A7.4a
Since the tWist ¢
builds up linearly with
T the strain energy stored is
Then
or
lT 2 L (7 )
U = 2 GJ
For a shaft of non-uniform properties and
varyinG loading one has
Example Problem 4
Determine the strain energy of flexure of
the beam of Fig. A7.5. Neglect shear strain u = ~
2
STGJdX
2
__ (8)
energy.
In passing it is worth remarking that one
often encounters the group symbol "GJ" in use
rEI = Constant for the torsional stiffness of a non-circular
shaft or beam such as an aircraft wing. In
such a case the torsional stiffness has not been
computed literally as G x J, but rather as-de-
L fined by eq. (6), viz. the ratio of torque to
"2 rate of twist.
Fig. A7.5
Solution: Example Problem 5
For a certain flight condition the torque
The bending moment diagram was drawn first on an airplane wing due to aerodynamic loading
(Fig. A7.5a) and analytic expressions were is given as shown graphically in Fig. A7.6. The
written for M. variation of torsional stiffness GJ is given in
like manner. Find the strain energy stored.
M~ ~~
~(X)' FOX O<X>~
1. 01"0::::--_ _
Fig. A7.5a
Fig. A7.6
Inspection of the diagram revealed that o:-__ -'-_--.:+_
L . -_ _- ' - -_ _- ' - -_ _
the energy of flexure in the right half of the y=o Y=L

beam must be identical with that of the left Root Tip
half. Hence
Solution:
A numerical integration of eq. (8) was
performed using Simpson's rule. The work is
shown in tabular form in table A7.1.
Values of T and GJ for selected wing
stations were taken from the graphs provided
STRAIN ENERGY OF TORSION and were entered in the table. For convenience
in handling the numerical work all the variables
A uniform circular shaft carrying a torque were treated in non-dimensional form, eq. (8)
T experiences a total twist in a length L pro- being changed as follows
portional to the torque. Thus from elementary
strength of materials
where strain energy in a beam under the action of a

T == T/TR, ,.=y/L transverse shear loading V(lbs)*. For this
purpose V = ,. x dy x t and t x dy is called A,
GJ == GJ/GJR the beam cross sectional area. Hence
The subscript R denotes "root" value (y=o). V"dx

The coefficients for Simpson's rule appearing in
dU = 2AG ' and
column (6) were taken from the expression for
odd n therefore the elastic strain energy of shear for
the entire beam is given by
I = ~; (f o + 4f~ + 2f" + 4f 3 + - - - -
- - - - - - - - - - - - - - - - (10)
- - - - + 15 f + 3f + ~f )
4 n-2 n-l 4 n
Example Problem 6
Determine the strain energy of shear in
TABLE A7.1 the beam of Example Problem 4.
(T)"
,.
STA. T (T)" GJ
GJ
Coef£. (T)~
ro- Coeff . Solution:
0 1.0 1.0 1.0 1.0 .0667 .0667 The shear load diagram was drawn first
.2 .98 . 96 .76 1. 26 .267 .336 from the right end •
.4
.6
.8
.94
.67
.29
.88
.45
.084
.54
.37
.24
1. 63
1. 22
.35
.133
.250
.200
.216
.305
.070
r- L
2 t L
2~-o
P
1.0 0 0 .14 0 .0833 0.0 po ..... x 1111111 I-II ~AG= CONSTANT

Z = .994 V(x) rh"""I""n,r-r-, r-, . -Po
Therefore the strain energy was
2U = LTR" x
GJR
.994
Then
1
U = 2AG x 2 x r:
= Po "L
2
Po'OX
2AG
STRAIN ENERGY OF SHEAR
Eq. (9) may be used to compute the shear
A rectangular element "dx" by "dy" of strain energy in a thin sheet. The element
thickness "t" into the page under the action of ~ x dy is visualized as but one of many in the
uniform shear stress,. (psi) is shown in Fig. sheet and the total energy is obtained by
A7.6a. summing. Thus
From elementary strength
of materials the angle of
shear strain r is pro-
U = ~ JJ,."t gx dy - - - - - - - - - - - - (11)
portional to the shear
stress ,. as Here a double integration is required, the
summation being carried out in both directions
'll' = ~
G
where G is the over the sheet. In dealing With thin sheet the
use of the shear flow q == ~ is often convenient
material elastic modulus so that eq. (11) rewrites
Fig. A7.6a
in shear. For the dis-
placements as shown in
the sketch only the down
load on the left hand side does any work. (In
U = ~ S J q" ~t dy - - - - - - - - - - - - (11a )
general all four sides move, but if the motion A very important special case occurs when a
is referred to axes lying along two adjacent homogenous sheet of constant thickness is
sides of the element, as was done here, the de- analyzed assuming q is constant everywhere. In
rivation is simplified). This load is equal to this case one has
q"
,. x dy x t and moves an amount ~ x dx. Then
1
U = 2GtJf dxdy = q"s
2Gt - - - - - - - - - - - (12)
dU = 2 ,.n dx dy where S = JJ dxdy is the surface area of the
1
sheet.
= 2 (9)
* The sketch is visualized as a side view of an element of

Eq. (9) may be used to compute the shear length dx taken from a beam of total heIght dy.
A7.5
Example Problem 7 A7.4 The Theorems of Virtual WorK and Minimum

Find the shear strain energy stored in a Potential Ene rgy.
cantilever box beam of uniform rectangular An important relationship between load and
cross section under the action of torque T. deformation stems directly from the definitions
(Fig. A7.8). of work and strain energy. Consider Fig. A7.9(a).
t::. U =PM
t::.U
Limit Mj .. 0,
dU = PdO
1L-------l.LL-
0
M~ f.-
Fig. A7.8
Fig. A7.9a
Solution: Thus
The shear flow was assumed given by dU :: p (14)

Bredt's formula (Ref. Chapter A-6) d6
T T In words,
q :: 2A :: 2bc , "The rate or change of strain energy with
respecttod8flection is equal ~ the assoCiirted
constant around the perimeter of any section. "load". -
Then Eq. (14) and the above quotation are state-
ments of the Theorem of Virtual Work. The
reader may find this theorem stated quite dif-
ferently in the literature on rigid body me-
chanics but should be able to satisfy himself
that the expressions are nevertheless compatible.
A useful restatement of the above theorem
THE TOTAL ELASTIC STRAIN is obtained by rewriting eq. (14) as
ENERGY OF A STRUCTURE
dU - Pdo :: 0
The strain energy by its definition is
always a positive quantity. It also is a It is next argued that if the change in dis-
scalar quantity (one having magnitude but not placement do is sufficiently small the load P
direction) and hence the total energy of a remains sensibly constant and hence
composite structure, having a variety of ele-
ments under various loadings, is readily found dU - d(Po) :: 0
as a simple sum.
d(U-Po) :: 0 - - - - - - - (15)
The quantity U - Po is called the total

potential of the system and eq. (15), resembling
as it does the mathematical condition for the
+
2
JV dx +
2AG
II q
2
dxdy _ - -
2Gt
(13) minimum value of a function, is said to be a
statement of the Theorem of Minimum Potential.
From the-r0regoing it;is clear that the
The integral symbols and cornmon use ,of "x" Theorem of Minimum Potential is a restatement of
an an index of integration should not be taken the Theorem of Virtual Work.
too literally. It is probably best to read In struCtural analysis the most important
these terms as "the sum of so and so over the uses of these theorems are made in problems con-
structure" rather than "the integral of", for cerning buckling instability and other non-
quite often the terms are formed as simple linearities. No applications will be made at
sums without resort to the calculus. The cal- this point.
culus is only used as an aid in some applica-
tions. A7.5 The Theorem of Complementary Energy and
It is seldom that all the terms of eq. Castigliano' s Theorem.
(13) need be employed in a calculation. Many
of the loadings, if actually present, may be of Again in the case of an elastic body, ex-
a localized or of a secondary nature and their amination of the area above the load-deformation
energy contribution may be neglected. curve shows that increments in this area (called
A7.6 DEFL EC TIONS OF ST R UC TUR ES
the complementary energy,U*), are related to the Force (lbs.) Translation (inches)
load and deformation by (see Fig. A7.9(b). Moment (in. Ibs.) Rotation (radians)
Torque (in. Ibs.) Rotation (radians)
Pressure (lbs/in") Volume (in 3 )
L h/-'-/""'/""7"//..,,-L--,---,--:-c-;~ Shear Flow (lbs/in) Area (in")
T dU* = odP Any generalizations of the meanings of
liP
"force" and "deflection" are possible only so
long as the units are such that their product
.£ .-------'--0 Fig. A7.9b yields the units of strain energy (in. Ibs) •
Once again for emphasis it is repeated that,
while the complementary nature of eqs. (14) and
(17) are clearly eVident, the use of eq. (17)
dU* (Castigliano's Theorem) is restricted to linearly
dP = a - - - - - - - - - ~ - - - - - - (16) elastic structures. A brief example will serve
for illustration of the possible pitfalls.
This is the Theorem of Complementary Energy. The strain energy stored in an initially
Now for the linearly elastic body a very straight uniform column under an axial load P
important theorem follows since (Fig. A7.9c) when deflected into a half sine wave is
dU = dU* so that
P~U* \
~ II U
\ Y = Yo sin~; M = PY
L
Y Consider Flexural Energy Only
~ a =SM"dx = P"Y~L
U 2EI 4EI
Fig. A7.9c ** a = l.J(~)"

2 dx
dx = Yon"
4L
.. U = P"oL"
J12EI
dU x Fig. A7.10
dP = a - - - - - - - - - - - - - - - - - - (17)
where a is the end shortening due to bowing.
In words, Because the deflections grow rapidly as P ap-
"The rate of change of strain energy with proaches the critical (buckling) load the
respectto load 1s equal to the assocra:ted-- problem is non-linear. The details of the
deflection"-.- - -- calculation of U are given with Fig. A7.10.
Eq. (17) and the above quotation are state- Now according to the Theorem of Virtual
ments of Castigliano's Theorem. Work (eq. 14)
For a body under the simultaneous action of dU = P
several loads the theorem is written so as to do
apply individually to each load and its associ- but
ated deflection, thus*: dU P"L"
do = n 2 EI
au = ai (17a)
Therefore
aP i
or
The partial derivative sign in eq. (17a) n"EI
indicates that the increment in strain energy is P = r::;r
due to a small change in the particular load P ,
all other loads held constant. i (Euler formula for uniform column). The
Note that by "load" and "deflection" may correct result.
be meant: Application of Castigliano's Theorem,
eq. (17), leads to the erroneous result:
Load Associated
Deflection
dU = a
dP
dU 2PoL"
dP = n2EI = <5
* The proof of the theorem for the case of multiple loads is
generally formulated more rigorously, appeal to a simple
diagram such as Fig. A7.9c being less effective. See, ** See Art. A17. 6, Chapter A-17 for detailed
for example, "Theory of Elasticity" by S. Timoshenko. derivation of this equation.
A7.7
P (incorrect) Then _6
2U = 43.26 x 10 p" LB FT
Moral: Do not use castigliano's theorem for "dU -6

non-linear problems. Up = dp = 43.26 x 10 P ft.
Fortunately the above restriction upon ·che
use of Castigliano's theorem is not a very Note that in this problem the only use made of
severe one, the majority of every day structural the calculus was in the differentiation. A
problems being linear. Castigliano's theorem simple sum was used to form U, as per eq. (2).
is quite useful in performing deflection calcu-
lations and a variety of applications will be Example Problem 9
made in the following sections. Derive Bredt's formula for the rate of
twist of a hollow, closed, thin-walled tube:
A7.6 Calculations of Structural Deflections by Use of
Castigliano' s Theorem. ds Ref. chap. A6.
As the examples of Art. A7.3 have illus- t
trated,the strain energy of a structure can be
expressed as a function of the external loadings The strain energy is stored in shear ac-
provided the internal load or stress distribu-
tion is calculable. Having the strain energy so cording to - 1
U -2
II
q "dxdy
or
expressed the deflections at points of load ap-
plication can be determined with the aid of eq. This is the only energy stored, secondary ef-
(17), Castigliano's Theorem. fects neglected. Over a given cross section q
In the examples to follow the deflections is a constant and is given by Bredt's equation
of a variety of statically determinate struc- T
q = 2A ' where A is the enclosed area of the tube
tures are computed. Methods of handling redun-
dant structures are considered in subsequent (Ref. Chap. A.5).
articles.
Solution:
Example Problem 8
Find the vertical deflection at the point Since the twist per unit length was desired
of load application of the crane of Fig. A7.11. the strain energy per unit length only was
Cross sectional areas are given on each written. ThUS, assuming y in the axial direction,
member. The stranded cables have effective no integration was made with respect to y. The
modulii of 13.5 x 10 6 psi. E = 29,000,000 for integration in the remaining direction was to be
other members. C carried out around the perimeter of the tube and
so the index was changed from "x" to the more ap-
T
60'
propriate "s". Hence (compare with Example Prob. 7)
,_l Fig. A7.11

The symbol ~ means the integration is carried out
around the tUbe perimeter. Therefore
dU T.r. ds
Q = dT = 4A"G t t
Solution:
USE OF FICTITIOUS LOADS
The strain energy considered here was that
due to axial loading in each of the four mem- In the following example the desired deflec-
bers. The load distribution was obtained from tion is at the free end of the bar where no load
statics and the energy calculation was made in is applied. A fictitious load will be added for
tabular form as follows: purposes of the calculation. The rate of change
of strain energy with respect to this fictitious
load will be found after which the load will be
TABLE A7.2 set equal to zero. This technique gives the de-
sired result in as much as the deflection is
MEMBER AE X 10"S S"L X 10 6 equal to the rate of change of strain energy
S LBS L FT
LBS AE with respect to the-load and such a rate eXists
OA -1. 50 P 40.0 136 0.66 P" even though the load itself be zero.
AB 2.50 P 50.0 11.8 26.48 P
AC 1. 58 P 63.0 11. 8 13.33 P" Example Problem 10
OC -2.12 P 84.5 136 2.79 P" Compute the axial motion of the free end of
Z = 43.26 P" the tapered bar of Fig. A7.12.
-fJ.
q - J
constant J
A = Ao (1- x)
2L
q
Ao
L
= 40#/in.
= 4 in."
= 40 in.
erally are easier to evaluate.
Example Problem 11
Find the deflection at point B of the beam
j E = 10 X 10 6 psi of Fig. A7.13.
+Pl.
Ix Fig.A7.12
lP EI constant
t~i~L~J
3 3 3
Solution:
Fig. A7.13
After addition of the fictitious end load
R the axial load from statics was found to be
Solution:
S(x) = R + q (L - x)
A fictitious load Pl. was added at point B
Hence, since loadings other than tensile are of and the bending moment diagram was drawn in two
a secondary nature parts.
2
= ~ JL SAEdx - See Eq. 3 O<X<L/3
U 2
o Mp
= ~ JL [R + q (L - X)J2 dx O<Y<L/3
2 0 AoE (1 -:L) MPl. O<:?;<L/3
Fig. A7.13a
+ ~ JL(L - x)2dx Then neglecting the energy of shear as being

2A o E (1 _ ~) small*
o 2L
Before evaluating the integrals it was ob-
served that the steps to follow, in which U was
to be differentiated with respect to P, and the
subsequent setting of P, = 0 would drop out both ..
the first and last terms. Hence only the second
term was evaluated. + ~r r3[~12L - 3Y) + P; (L + 3Y) ] dy
U = X + 2L" Rq (1 - In 2)
AoE
+ X
+ -.l:.- JL/
3[(P + 2Pl.) (L d~ •
2EI 9
Then o
Differentiation under the integral sign with

° = dRdU = 2L"q
AoE
(1 _ In 2) respect to Pl. gave
L/3
DIFFERENTIATION UNDER THE INTEGRAL SIGN °B = i~l. lr J 0 [(2P +
An important labor savings may be had in
ir I:/3~12L ~'(L
the calculation of deflections by Castigliano's
theorem. + - 3y) + + 3y)](L ; 3y) dy
In the strain energy integrals arising in
this class of problems, the load Pi' with respect
to which the deflection is to be found, acts as
an independent parameter in the integral. Pro-
+ E~J:/3[p
; 2P'IL _ 30)] 21L ; 30) do .
vided certain requirements for continuity of the
functions are met - and they invariably are in
these problems - the differentiation with re- * For beams of usual high length - to - depth ratios the shear
strain energy is small compared to the energy of flexure.
spect to P may be carried out before the inte- Neglecting the shear energy is equivalent to neglecting the
gration is made. The resulting integrals gen- shear deflection contribution. (see p. A7. 14)
A7.9
The fictitious load P~, having served its pur- The total loadings
pose, was set equal to zero before completing were
the work.
Plan View
2P jL/3x2dx
-.l.... T=T~+T2= Tocos g + PR
°B - EI • 9 o
(1 - cos g)
Lf3 Thus
1 P
+ EI 81
f0 (2L - 3y)(L + 3y) dy Fig. A7.14b
1
U = 2EI
-
PL"
7
-- + 2GJ
1
Je:t/2[Tocos 2
g + PR(l - COS g)J Rd g
486 EI o
(Note use of "Rdg" for the length of a dif-
Example Problem 12 ferential beam element instead of "dx"). Dif-
Fig. A7.14 shows a ferentiating under the integral sign
cantilever round rod of
diameter D formed in a
quarter circle and acted
upon by a torque To.
Find the vertical move-
ment of the free end.
Solution:
R2
+ GJ Jo
rn / 2[ToCOS g + PR(l - cos g)J(l - cos g) dg.
Fig. A7.14a shows Putting P, the fictitious load equal to zero and
the vector resolution of Plan View integrating gave
the applied torque To on
beam elements. T~ (g) = M~
~U _ To R (4 - n - EnI)
2
Tocos g and the moment T~
OVERT = ~P - L i
M~ (g) = To sin g. Ap- GJ
plication of a fictitious
vertical load P (down) at Since J = 21 and G ~ E/2 . 6 the deflection was
the point of desired de- negative (UP).
flection gave the loadings Fig. A7.14a
shown in Fig. A7.14b.
A7.7 Calculation of Structural Deflections by the Method of Dummy-Unit Loads (Method of Virtual Loads).
The strict application of the calculus to Castigliano's theorem as in Art. A7.6, leads to a
number of cumbersome tecl111iques ill-suited to the solution of large complex structures. A more
flexible approach, readily adapted to improved "book keeping" techniques is the Method of Dwnmy-
Unit Loads developed independently by J. C. Maxwell (1864) and O. Z. Mohr (1874)-.---- - -----
--- That the equations for the Method of Dwnmy-Unit Loads may be derived in a number of ways is
attested to by the great variety of names applied to this method in the literature ®. Presented
below are two derivations of the equations stemming from different viewpoints. One derivation ob-
tains the equations by a reinterpretation of the symbols of Castigliano's theorem - essentially an
appeal to the concepts of strain energy. The other derivation uses the principles of rigid body
mechanics. Based as they are upon a common set of consistent assumptions, all the methods must, of
course, yield the same result.
Derivation of Equations for the Method of Dwnmy-Unit Loads

(Virtual Loads)
I - From Castigliano's Theorem II - From the Principle of Virtual Work

Beginning with the general expression for When a system of forces whose resultant is
strain energy, eq. (13) zero (a system in equilibrium) is displaced a
®Variously called the Maxwell-Mohr Method, Method of Virtual Velocities, Method of Virtual Work, Method of Auxilliary
Loads, Dummy Loads, Method of Work, etc.
I Cont'd. II Cont'd.
U= ~
2
JSAEdx
2 + ~ JMEIdx
2
2
~
+ 2 JTGJdx +
2
e t c. small amount without disturbing the equilibrium

balance, the work done is zero - obviously,
differentiate under the integral sign with since zero resultant force moving through a
respect to Pi tu obtain distance does zero work.
6
i
=
AE
j S~~idXEI
+ J M~dX + - - - - - - - etc.
Consider now the set of equilibrium forces
applied to the truss of Fig. A7.15(b). The set
may be divided into two parts: the "external
system", consisting of the unit load applied at
Consider the symbols the point whose deflection is desired and the
~S ~M ~T two reactions fixing the line of reference, and
c'lP 'JP' aP ' - - - etc. the "internal system" consisting of the axial
iii loads acting on the truss members to produce
Each. of these is the "rate of change of so equilibrium. These latter are denoted by the
-and-so with respect to Pi" or "how much so- symbol "u". This set of forces is considered
and-so changes when Pi changes a unit a:nount" OR small enough so as not to. affect the actual be-
Eq,UALLY, "the so-and-so loading for a unit load havior of the structure during subsequent ap-
Pi"· plication of a real set of major loads. This
Thus, to compute these partial derivative unit load set is present solely for mathematical
terms one need only compute the internal load- reasons and is called a "virtual loading" or
ings due to a unit load (the virtual load) ap- "dummy loading".
plied at the point of desired deflection. For Assume now that the structure undergoes a
example, the term c\Mjc\p' could be computed in deformatIon due to application of a set of real
either of the two ways "shown in Fig. A7.15a. loads, the virtual loads "going along for t~
ride". Each member of the structure suffers a
RATE METHOD UNIT METHOD deformation denoted by ~ ®. The virtual load-
ing system, being in equilibrium (zero resultant)
by hypothesis, does work ("virtual work") equal
to zero. Or, considering the subdivision of the
virtual loading system, tue work done by the
external virtual load must equal that absorbed
Fig. A7.15a by the internal virtual loads. The work done by
the external virtual forces is equal to one
pound times the deflection at joint C, the re-
actions R:>. and R2 not moving. That is
M = P:>.x m = dummy loading
External Virtual Work = l~ x 6
c
eHl The internal virtual work is the sum over the
aP:>. = x
structure of the products of the member virtual
Likewise, dsAoP c , where Pc is a load (real loads u by the member distortions~. That is,
or fictitious) applied at joint c of Fig. A7.15b, Internal Virtual Work = Z u.~ .
is given by the loadings for the unit load ap-
plied as shown. Then equating these works,
In practice the use of the unit load is 1 x 6 = Z u·~ .
most convenient. Using the notation c . ..
If the deformations ~ are the result of
u == ~Sp ,m ==!!':L.- t ==~ elastic strains due to real member loads 3 then
i 3P i ' aP
i ~ = 3L/AE for each member and one has
_ av - = aq 3L
6c = Z u AE
v = -c\- q
Pi - c)P i
for the unit loadings, the deflection equation The argument given above may be extended
becomes qUickly to include the internal virtual work of
6 = Sudx +
i AE
J
Mmdx + Ttdx
EI
J
GJ
virtual bending moments (m), torsion loads (t),
J
shear loads (v), and shear flows (q) doing work
during deformations due to real moments (M),
+ ~+ J JJ
q qG~dY - - - - - - - - - (18) torques (T), shear loads (V~nd shear flows
(q). The general expression becomes
6
i
J3udx
=
AE
JMmdx
EI
+JTtdx +
GJ
+ f Vvdx 11 q q dxdy - - -
AG
+
Gt
- - (18)
Af-~::....a...:..~~~~--I.....:...\!~4B
r. 5=R:>. 't. 5=R 2 @ Note that the deformations are not restricted to those due
to elastic strains only. They may be the result of elastic
"un loads due to a unit (virtual) load. or inelastic strains, temperature strains or misalign-
ment corrections.
i
• m
A7.11
In applying eq. (18) the labor of a deflec-

tion calculation divides conveniently into
several steps:
i. Calculation of the real (actual) load
distribution (S,M,T, etc.)
ii. Calculation of the unit (virtual) load
distribution (u,m,t, etc.) due to a unit
(virtual) load applied at the point of desired "u)." loads "u,," loads
deflection and reacted at the reference point(s).
Fig. A7. l6a Fig. A7.l6b
iii. Calculation of flexibilities,

TABLE A7.3
1 1
AE,' Er' etc. MEM. Lin
AE x 10- 0 L x 10 0
S U. Ua
Su.L x 10. SuaL x 100
lbs AI lri/lb AElnch AElnch
iV. Summation and/or Integration.
AB 30 4.785 6.270 10,500 1.5 0 98.75 0
Here again note the general nature of the 3.074
Be 30 9.759 2,250 0 0 0 0
terms "load" and "deflection". (See p. A7.6)
The following examples apply the method of CD 30 3.074 9,759 2,250 0 0 0 0
dummy-unit loads to the determination of ab- EF 30 5.365 5.591 -5,250 -0.75 1 22.01 -29.36
solute and relative deflections, both rotation FG 30 5.365 5.591 -5,250 -0.75 1 22.01 -29.36
and translation.
GB 30 3.48 8.621 0 0 1 0 0
Example Problem 13 BE 50 10.15 4.926 -8,750 -1.25 0 53.88 0
The pin-jointed truss of Fig. A7.16 is BG 50 3.48 14.368 5,000 1. 25 0 89.80 0
acted upon by the external loading system shown.
DG 50 5.365 9.320 -3,750 0 0 0 0
The member loads are given on the figure. Mem-
ber properties are given in Table A7.3. Find BF 40 3.074 13.012 2,000 0 0 0 0
the vertical movement of joint G and the hori- CG 40 3.074 13.012 -1,000 0 0 0 0
zontal movement of joint H.
DB 40 3.074 13.012 2,000 0 0 0 0
Z = 286.4 Z =-58.72
Answers: 0G = 0.286"
VER
0K_ = -.0587" (the negative sign means

--HOR
the joint moves to the LEFT since the unit load
was drawn to the RIGHT in Fig. A7.16b).
Example Problem 14
Fig. A7.l6
For the truss of Fig. A7.16 find the fol-
Solution: lowing relative displacements of joints:
c) the movement of joint C in the direction
Only the energy of axial loadings in the of a diagonal line joining C and F.
members was considered. Unit (virtual) loads
were applied successively at joints G and H as d) the movement of joint G relative to a line
shown in Figs. A7.16a and A7.16b. All Sand u joining points F and H.
loadings were entered in Table A7.3 and the
Relative deflections are determined by
applying unit (Virtual) loads at the points
calculation for °= f Sudx
~ was carried out by where the deflections are desired and by support-
ing such unit load systems at the reference
forming the sum of SUL/AE terms for the members points of the motion. Thus, for solution to
part (c) a unit load system was applied as
of the truss, i.e. ° = L: SuL
AE shown in Fig. A7.16c and for the solution of
me
A7 12 DEFLECTIONS OF STRUCTURES
part (d) the system of unit loads of Fig. A7.16d Rotations, both absolute and relative are
was used. Table A7.4 completes the solution, determined by applying unit (virtual) couples
the real loads and member flexibilities to the member or portion of structure whose
rotation is desired. The unit couple is resist-
Ui;) being the same as for example problem 13. ed by reactions placed on the line of reference
for the rotation. Thus Figs. A7.16e and A7.16f
0 -.6 C 0 0 -.375 -.375 show the unit (virtual) loadings for par ts (e)
and (f) respectively. Table A7.5 completes the
t ~V I :\Vli F
~
R=l#
/
/
-.6
{o
I
0
0
.5
FOG
~U'
1# • .5
H
calculation, the real loads and member flexibil-
ities
problem 13 •
(L/
AE )
being the same as for example
1/50# 1/50# R=1/40#

"u,," loads "U4" loads R=1/40# .025 .025 . 025~D 0 B~.025C 0
-/~~J(°II\Vo
Fig. A7.16c Fig. A7.16d
R=V4o# -.025 -.025 G'-1/50# / /G R=1/4 0#

1 50#
TABLE A7.4 "us" loads "u.s" loads
~x 10· Fig. A7.16e Fig. A7.16f

MEM. AE in/lb Sib. u3 u. ~ xl0 3 Su.L X 103
(See Table inch """"AE inch
A7.3)
AB 6.270 10,500 0 - 0 0 0
BC 9.759 2,250 -.6 -.375 -13.17 -8.23
CD 9.759 2,250 0 -.375 0 -8.23

TABLE A7.5
EF' 5.591 -5,250 0 0 0 0
~x108
17.61
FG 5.591 -5,250 -.6 0 0
MEM.
AE in/lb
Sib
Su.L x 103
u"inch-> ueinch-l. AE ~XI03
(See Table
GH 8.621 0 0 0 0 0 A7.3)
BE 4.926 -8,750 0 0 0 0 AE 6.270 10,500 .025 0 1. 65 0
BG 14.368 5,000 1.0 .625 71. 84 44.9 BC 9.759 2,250 .025 -.025 .55 -0.55
DG ,9.320 -3,750 0 .625 0 -21. 8 CD 9.759 2,250 .025 0 .55 0

BF 13.012 2,000 -.8 -.50 -20.82 -13.0 EF 5.591 -5,250 -.025 0 +.73 0
CG 13.012 -1,000 -.8 0 +10.41 0 FG 5.591 -5,250 -.025 0 +.73 0
DH 13.012 2,000 0 -.50 0 -13.0 GH 8.621 0 0 0 0 0
L: = 65.87 L: =-19.36 BE 4.926 -8,750 0 0 0 0
BG 14.368 5,000 0 .015 0 1.08
DG 9.320 -3,750 -.015 0 .52 0
BF 13.012 2,000 0 0 0 0
CG 13.012 -1,000 0 0 0 0
Therefore the movement of joint c towards
joint F was {j = .06587 inches and the motion of DH 13.012 2,000 0 0 0 0
joint G relative to a line between F and H was L: = 4.73 L: = 0.53

(j = -.0194 inches, the negative sign indicating
an upward movement.
Example Problem 15
For the truss of Fig. A7.16 determine
e) the absolute rotation of member DG Therefore the absolute rotation of member
f) the rotation of member BG relative to DG was gDG = .00473 radians and the rotation of
-
CG = .00053 radians
member CG. BG relative to CG was gBG _
L -
..
A7.13
Example Problem 16 points C, D, E, the attachment points of an

Find the vertical deflection of point C for aileron or a flap. The wing beam deflection
the cantilever beam of Fig. A7.17 carrying a bends the aileron or flap structure by applying
concentrated load P at its end. Also find slope a load at D thru aileron supporting bracket. To
of elastic curve at C. know this force the deflection of the wing beam
at D relative to line CE must be known.
Solution:
Solution:
Origin at B: -
With origin at B
M = 30 x • ~ = 15 2
x
M= - Px (Fig. A7.17)
w • '0 Ib./in
For virtual loading (Fig. A7.17a)
m = 0, for x<b
m = -1 (x - b), for x>b
Hence Mrndx - - Px(-x + b)dx = (Px 2 - Pbx)dx
Whence 0c P
--
EI
fL
b
(x 2 -
P [ ~"_ bX
bX)dx = --
EI 3 2 b
L
2
J
m 0, x < 20
m -.5 (x - 20) = -.5x + 10, when x = 20 to 50
m -.5 (x - 20) + 1 (x -50) = .5x - 40, when
If b = zero, then 0B = PL"/3EI x = 50 to 80
~A
C x-jB °D(CE) -- J~
Mrndx
t== a L
..I..
b 4p Fig. A7.17
Ih 50
h 80
r C
hf
B Fig. A7.17a
= --
EI 20
15 x 2 (-.5x + 10)dx + EI
1
50
15 x 2
(.5x - 40) dx
r-
1A C
-' uni 1:.
B
Fig. A7.17b
150x~l50 + ~ [7.5 x" _ 600x" l80
ccuple =1. [-7.5 x.. +
EI 4 3 J20 EI 4 3 jso
11.72 + 6.25 + .300 - .400 +
For virtual loading see Fig. A7.17b 76.8 - 102.4 - 11.72 + 25.0]
m = 0, x<b, m =- 1, x>b
17.9 X 10 6
Hence Mrndx = (-Px) (-1) dx = Px dx EI
Therefore deflection of point D relative to line
J joining CE is down because result comes out nega-
:
e =
C
Mrndx
EI
=E (L xdx
EI ) b
tive and therefore opposite to direction of vir-
tual load.
[I:j ;x-~ n *~
PL 2 Example Problem
If b = 0, BB = 2EI h rm:1
,. '\ f'
Example Problem 17
For the uniformly loaded cantilever beam M.
of Fig. A7.18, find the deflection of point D
relative to the line joining points C and E on
1 ~D Mdiag'l _ rodiag. _
lf R... diXg.j
the elastic curve of the beam. This is repre- L w~ wLf
R=1/L unit R=1/L
sentative of a practical problem in aeronautics, 2" 2"
(a) (b) (c) couple
in that AB might represent a rear wing beam and
Fig. A7.19 Fig. A7.20
Find the horizontal deflection of point C
for the frame and loading of Fig. A7.19. Also dy = J dx and ~ = A~s' where A = cross sectional
angular deflection of C with respect to line CD. area of beam at section and Es = modulus of
rigidity~ and assuming that the shearing stress
Solution:
V
A is uniform over the cross-section.
Fig. b shows the static moment curve for
the given loading and Fig. c the moment diagram Vvdx
Therefore 1 x 0 = ~.Then the total deflec-
for the virtual loading of a unit horizontal s
load applied at C and resisted at D. tion for the shear slips of all elements of the
beam equals
F< - JEI'
Mindx • m= h
f
uC(H) - L Vvdx _______ (a)
o total = ~
o S
hence
where V is the shear at any section due to given
L(WLx - ~2) hdx = ~(WLx2 _WX") J!. JL
loads. v = shear at any section due to unit
o
J2 EI L 4 6. EI 0 hypothetical load at the point where the deflec-
tion is wanted and acting in the desired direction
of the deflection. The reactions to the hypo-
1 whL" thetical unit load fix the line of reference for
12 EI the deflection.
A is the cross-sectional area and Es the
To find angular deflection at C apply a unit
imaginary couple at C with reactions at C and D. modulus of rigidity. Equation (a ) Is slightly
F.1g. A7.20 shows the virtual m diagram. in error as the shearing stress is not uniform
over the cross-section, e.g. being parabolic for
EI
o
2JL
a rectangular section. However, the average
shearing stress gives close results.
For a uniform load of w per unit length, the
center deflection on a si~ply supported beam is: -
°center =2 J~ Vvdx
AE s
=2 J~ (wL 2
WX) !dx
2
AE s 0'
o wL 2 0
=
8AE s
Linear Deflection of Beams Due to Shear by Virtual Work.
Generally speaking, shear deflections in For bending deflection for a simply supported
beams are small compared to those due to bending beam uniformly loaded the center deflection is
except for comparatively short beams and there-
fore are usually neglected in deflection calcu- 5 wL 4
lations. A close approximation is sometimes 384EI
made by using a modulus of elasticity slightly Hence
less than that for bending and using the bending wL 2
deflection equations.
The expression for shear deflection of a
beam is derived from the same reasoning as in
8AE s
5wL 4 24 using Es = .4E In:r,
r = radius of gyration.
previous derivations. The virtual work equa- 384EI.
tion for the hypothetical unit load system for a 1
shear detrusion dy (Fig. A7.21) considering only For I beams and channels r is approximately 2 d
dx elastic is 1 x o=vdy where v is shear on
section due to unit hypothetical load at point d
and for rectangular sections r = V12 (d = depth)
0, and dy is the shear detrusion of the element
dx due to any given load system or any other d
cause. In aircraft structures a ratio of L is
1
seldom greater than 12 •
f/_. EiL~
.=.J .' . - =ft 0 Thus the shearing deflection in percent of
Fig. A7.21 11# d

the bending deflection equals 4.1N for a L ratio
of l~ for I-beam sections and 1.4 percent for
(JL
* Sometimes "G", instead. See p. A7.4 rectangular sections.

A7.15
Example Problem 19 t = torsional moment at any section due to

a unit virtual couple acting at section
50lf lOO{f
where angle of tWist is desired and
tlon+1o"-i A
Unit.
Load
~f:========:::ilA
!
acting in the plane of the desired de-
flection. (inch lbs/inch lb)
Example Problem
Fig. A7.22
Example Problem 20
Find the vertical deflection of free end A Fig. A7.23 shows a cantilever landing gear
due to shear deformation for beam of Fig. A7.22 strut-axle unit ABC lying in XY plane. A load
assuming shearing stress uniform over cross- of 1000# is applied to axle at point A normal to
section, and AE constant. XY plane. Find the deflection of point A normal
s to XY plane. Assume strut and axle are tubular
= I Vvdx and of constant section.
°A AEs
Solution:
V 100# for x = o to 10
V = 150 for x 10 to 20 The loading shown causes both bending and
tWisting of the strut axle unit. First find
v = 1 for x = o to 20 bending and torsional moments on axle and strut
hence due to 1000# load.
1 flO 20
AE
s 0
100·1 dx
10
+ -1-
AEs f10
150·1 dx
20
1 2500
= =
AEs [lOOxJ + Ais [15Ox] AEs
o 10
Method_of Virtual Work Applied to Torsion of Cylindrical Bars.
11#
B A
The angle of twist of a circular shaft due Fig. A7.24
to a torsional moment may be found by similar
reasoning as used in previous articles for find- Ttdx
ing deflection due to bending or shear forces. o= fMmdx +
EI EJ
The resulting expressions are: - s
o = J TEtsdJX - - - - - - - - - - - - - (A )
Member AB M = 1000 x, (for x = o to 3)
T = 0
Member BC MBC 3000 sin 20° + 1000 x,
(for x = o to 36)
g = JTis~ - - - - - - - - - - - - - - (B)
TBC 3000 cos 20° constant
between Band C.
In equation (A), for translation deflections,
T = twisting moment at any section due to Now apply a unit 1# force at A normal to xy
applied tWisting forces. plane as shown in Fig. A7.24 and find bending
t = torsional moment at any section due to and torsional moments due to this 1# force.
a virtual unit 1 lb. force applied at
the point where deflection is wanted
and applied in the direction of the Member AB
desired displacement. (in lbs/lb)
Es shearing modulus of elasticity for the m 1. x x, (for x o to 3)
material. (also "G")
J = polar moment of inertia of the circular t = 0
cross-section.
In equation (B), for rotational deflections, Member BC
g = angle in twist at any section due to
the applied twisting moments in planes m 3 sin 20° + 1. x (for x = 0 to 36)
perpendicular to the shaft axis.
Angle in radians. t = 3 cos 20° constant between Band C.
ET
SUbt. The shear flows shown on the (nearly) horizontal

edges of the web panels are average values. Fig.
A7.27 is an exploded view of the beam showing
(5 = J Mmdx +
EI
JTtdx
E J s
the unit (virtual) loads.
= i, J: 1000 x . X <Ix + i, J:
6
KlOOO X + 1026)
I
(x + 1.026)J dx + E JJ36(2820) (2.82) dx
s 0
Virtual loading. Fig. A7.27

= Ell ( [333 x 3] 3 + [333 x 3 + 1026 x 2 + 1050
o Since both axial loads and shear flows
were considered, the form of deflection
1 equation used was
EI (16,92500) +
q q dxdy
Gt
_1_ (286200)
EsJ
Integrations in the flanges were made
assuming linear load variations. Such an in-
Note: A practical landing gear strut would integration carried out over a uniform flange of
volve a tapered or reinforced section involving length L whose real load varies from Si to Sj
a variable I and J and the integration would and whose virtual load varies from ui to Uj
have to be done graphically or numerically.
yields
Example Problem 21
For the thin-web aluminum beam of Fig.
A7.25 determine the deflection at point Gunder .'.
the loading shown. Stringer section areas are
given on the figure,
I
A r A=. 15 B cA=. 15 C cA=.15 D The integrations in the trapezoidal sheet
~
panels were made using the shear flows on the
L
200321l '=.032: . '=.032 T
t=. A=. 08 I A=. 08
A=.05"'
I
-L
(nearly) horizontal sides as average values,
assumed constant over the panel. With this
ElF
j.:- 20"
G
----0+--20" --I-- 20"
H+
-I
P=180041 simplification JJ qAV q~~ dxdy = qAV qAV ;t
Fig. A7.25
where S is the panel area.
Solution: The calculation was completed in Table
A7.6.
It was assumed that the webs did not buckle TABLE A7. II
(10)
and carried shear only. Fig. A7.26 is an ex- (1) (2) (3) (4) (5) (6) (7) (8) (9)
~ ~ (8)X(9)"2f~~XI0.
ploded view of the beam showing the internal Ul Uj iEx1o+e (2)x{6)+(3)x(7)
real loads carried as determined by statics: AB 3.00 P 2.184 P 2.00 l.09 1.697 1.393 13.34 8.133 P lOB,;) P
2.184 P 1.201 P 1.09 0 .'127 .363 13.34 2.024 P 27.0 P

BC
CD 1.201 P 0 0 0 13.34
~d~I~'D
-3.00 P _2.164 P -2.00 -1.09 -1.697 -1.393 13.34 8.133 P 108.5 P
EF
- .727 - .363 13.34 2.024 P 27.0 P

Fa -2.184 P -1.201 P -1.09 0
GH -1.201 P 0 0 0
DB 0 0
ca 1 0 .67 .33
3.OP BF o 0
\ Z 271.0 pj2
3. OP-r' 2. 184P 1. 201P 1

P
Wob
Member
A-B-E-F .04097 P ,0454 363 8.14 5.80

NOTE:
E" 10,000,000
G '" 3,600,000
Fig. A7.26 B-C-F~G .0491 P .0546 350 8.14 7.64 The denominatorofterma in
cola. 6 &. 7 should be 6 instead
C~D-G-H .0600 P 01' 3. To correct this error,
* The equations of statics for tapered beam webs are Z ==13.44 P
total sum of col. 10 Is divided
by2.
derived in Art. AI5.18, Ch. A-15.
..,
!
A7.l7
/ T
II1111I11111111111111
_3 I
.147 (1800) 10 in = .265 in. ~ I
:::
A7.8 Deflections Due to Thermal strains.
As noted in the "virtual work derivation"
of the dummy-unit load deflection equations, the L
a
real internal strains of the structure may be
due to any cause including thermal effects.
Fig. A7.29
Hence, provided the temperature distribution and
thermal properties of a structure are known, the
dummy-unit load method provides a ready means Solution:
for computing thermal deflections.
The axial deformation of a differential
Example Problem 22 element of the upper flange (subscript u) was
Find the axial movement at the free end of assumed given by ~ = a T dx where a was the
...., a uniform bar due to heat application to the
material thermal coefficient of expansion.
fixed end, resulting in the steady state tem-
perature distribution shown in Fig. A7.28. As- The lower flange, having received no heating
sume material properties are not functions of underwent no expansion.
Inasmuch as a thermal expansion is uniform in all
temperature. directions no shear strain can occur on a material element.
Hence no shear strain occurs in the web. The apparent
anomaly here - that web elements appear to undergo shear
T temperature above
ambient tempera- deformations ll' = a~ (Fig. A7. 29b) - is explained as
T(X)= To (l-tanh ~X) ture follows: The temperature varies linearly over the beam
K = an empirical con- depth. The various horizontal beam "fibers" thus undergo
axial deformations which vary linearly also in the manner
stant depending of Fig. A7. 29b giving the apparent shear deformation. No
upon thermal prop- virtual work is done during this web deformation since no
erties and rate of axial virtual stresses are carried in the web.
Fig. A7.28 heat addition. With the addition of a unit (Virtual)
load to the free end, the Virtual loadings ob-
tained in the flanges were:
u L --
= - x (= - u )
Solution:
U h L
The thermal coefficient of expansion of the
rod material was a. Hence a rod element of Then the deflection equation was
length dx experienced a thermal deformation
~ = a • T • dx. Application of a unit load at
the bar end gave u = 1. Therefore
Ju . J~( l-tanh ~)
. \;
6 a . T . dx a To dx
1 Example Problem 24
= a ToL [1 -R (In cosh K)J The first step in computing the thermal
stresses in a closed ring (3 times indeterminate)
involves cutting the ring to make it statically
Example Problem 23 determinate and finding the relative movement of
The idealized two-flange cantilever beam the two cut faces.
of Fig. A7.29a undergoes rapid heating of the Fig. A7.30a shows a uniform circular ring
upper flange to a temperature T, uniform span- whose inside surface is heated to a temperature
wise, above that of the lower flange. Deter- T above the outside surface. The temperature is
mine the resulting displacement of the free end. constant around the circumference and is assumed
to vary linearly over the depth of the cross

section. Find the relative movement of the cut
o
surfaces shown in Fig. A7.30b.
Remarks:
In the three elementary examples given
above no stresses were developed inasmuch as the
idealizations yielded statically determinate
structures which, with no loads applied, can
have no stresses. Indeterminate structures are
treated in Chapter A.S.
A7.9 Matrix Methods in Deflection Calculations.
a b Introduction. There is much to recommend the use of
Fig. A7.30 matrix methods~ for the handling of the quantity
of data arising in the solutions of stress and
Solution: deflection calculations of complex structures:
The data is presented in a form suitable for use
An element of the beam of length Rd~ is in the routine calculatory procedures of high
shown in Fig. A7.30c. Due to the linear tem- speed digital computers; a fleXibility of opera-
perature variation an angular change dG = tion is present which permits the solution of ad-
RaT ditional related problems by a simple expansion
~ d~ occurred in the element. The change in
of the program; The notation itself suggests new
length of the midline (centroid) of the section and improved methods both of theoretical ap-
RaT d'/'.
was I::" = 2 th
Unit redundant loads were approach and work division.
The methods and notations employed here and
plied at the cut surface as shown in Fig. A7.30b later are essentially those presented by Wehle
giving the following unit loadings around the and Lansing@ in adapting the Method of Dummy-
ring. Unit Loads to matriX notation. Other appropriate
From unit redundant couple (X) references are listed in the bibliography.
(m positive if it tends to open

1
BAS IS OF METHOD
the ring). Assume the structure to be analyzed has been
U (= axial load) = o (positive of tensile)
idealized into a truss-like assembly of rodS,
x bars, tubes and panels (sheets) upon which are
From unit redundant axial (horizontal) load (y) acting the external loads applied as concentrated
loads Pm or Pn , each with a different numerical
my = -R (1 - cos ~)
uy = cos ~
From unit redundant shear (vertical) load (g)
mg = -R sin ¢
ug - sin ~
The deflection equation by the dummy-unit load

method is (a) (b)
5 = Ju • I::" + J m • dG Fig. A7.31. Idealization into an

assembly of bars and panels.
Then
5 = 0·1::" +
2n RaT
1· --- d
~ RaT
= 2n---
X
J J o
h h
2n RaT 2n RaT
5y = :os ~ 2 d¢ + 0(-) R(l-cos ~) h d~
J ~ For the reader not familiar with the elementary arithmetic
rules of matrix operations employed here, a short appen-
dix has been included in the present edition.
f @ L. B. Wehle Jr. and Warner Lansing, A Method for Re-
ducing the Analysis of Complex Redundlillt"""'StrUctures to a
2n R2 aT (negative indicating movement to RoutineProcedure, Journ. of Aero. Sciences, 19, October
h the right) ~ -
A7.19
subscript. Thus the system of Fig. A7.31a is THE STRAIN ENERGY

idealized into that of Fig. A7.31b. It is next desired to write the strain
With the above idealization an improved energy as a function of the q's. Continuing the
scheme may be employed to systematize the com- illustrative example, write
putation of deflection calculations. The fol-
lowing steps summarize the procedure which is
discussed in detail in succeeding sections. U= ~ JL'M"dx + ~ JL"M"d Y
2 EI~ 2 EI..
I. A set of internal generalized forces, o 0
denoted by qi or qj (i, j are different numeri-
cal subscripts), is used to describe the inter-
nal stress distribution. The q's may represent
axial loads, moments, shears, etc. In con-
junction with a set of member flexibility coef-
ficients, aij, the q's are employed to express L"ydy 1 .. jL "dy
the strain energy U. aij ~ives the displacement
of point! per unit force at point 1.*
II. Equilibrium conditions are used to
+ 2 x
1
2 q .. q3
Observe that each of the integral terms in

I
0 EI .. + 2 q 3 Er.. 0
relate the internal generalized forces qi, qj to the above expression is a property of the struc-
the external applied loads, Pm or Pn • With this tural element (variation of EI) and of the nature
relationship the strain energy expression ob- of the associated generalized force (exponent on
tained in I, above is then transformed to give variable). Introducing the notation
U as a function of the P's.
III. Castigliano's Theorem is used to
compute deflections.
CHOICE OF GENERALIZED FORCES

a~ ~ -
r"
0
xdx
Eh
a .... - r'Y"dY
0
EI..
= r"ydy
Consider for example the problem of writing
the strain energy (of flexure) of the stepped a .. 3 EI.. 0
a 33 r"dy
0
EI ..
cantilever beam of Fig. A7.32a, assuming ex-
ternal loads are to be applied as transverse the strain energy becomes
point loads at A and B. The set of internal
generalized forces of Fig. A7.32b will com- 1 q" a .... + 2 x 1 q .. q3 a ..
pletely determine the bending moment distribu- U = 12 q
..
~
a~~ + 2
.. 2 3
tion in the beam elements and hence the strain

energy. Set (b) then is a satisfactory choice 1
of generalized forces. + 2 q" a 33 - - (19)
3
It should be pointed out that set (b) is
not a unique set. Other satisfactory choices Equation (19) is an expression for U which
(not an exhaustive display) are shown in Figs. could have been written immediately from physi-
A7.32c, d and e. The final selection may be cal considerations. Each coefficient aij is the
made for convenience or personal taste.
Note that only as many generalized forces displacement at point i per unit change in force
are used per element as are required to deter- at point j. This identity is easily seen by
mine the significant loadings in that element. applying Castigliano's Theorem to eq. (19).
With this interpretation the first term in eq.
(19), representing the strain energy in the
outer beam portion, is written by analogy to
~Eh A EI~ B
{I I} 1 S"L)
L"~L~x~(a)
(I eq. (2) of Art. A7.3 ( U = 2 AE • The remain-
q3 q .. q~ (c)
ing three terms, representing the energy stored
~
qj)--,
q3 (b)
I
~q~ q3
II Ir" q~1,
(d)
in the inner beam segment by q .. and q3' are like-
wise readily written, with proper account taken
for the cross influence of one force upon an-
other (the "a .. 3 q .. q3" term).
M = q;>,x O<x<L~
M = q3 + q ..y O<y<L ..
q13~ It.. , q~l~
(e)
Note that
= = a ji - - - - - - - - (20)
Fig. A7.32. Some possible choices
of generalized forces.
Hence
a (Maxwell's Reciprocal Theorem)
'*("Relative displacements in the individual member") ji
F
The general form of strain energy express-

ion is (expanding eq. (19) by induction)
UtN
I
I
I
-I
I
I
I
{2 tJ.
t. --+-
_I
I I I
I
+q4qJ.
, u4J. + - - - - - - - - - - - - - _ -)
I
I LJ. ---1
I
I
1
I I Fig. A7.32f
I J
I
I I
T qNqJ. uNJ. + - - - - - - - - - - + qN qN aNN
In matrix notation
In matrix notation this equation is written

(see appendix)
~2
q,21. U2 2 ------:
I I
I
I
I
I I
(21) Symbolically these relationships are written
I
I I
, I I
I
~;:------- ~
I
I
qN - - - - - - - - - - - - (22)
or, more concisely, The matrix [Gim] is called the ~it load
distribution" inasmuch as anyone column of
[Gim] , say the mth column, gives the values of

In the matrix [U ij ] many (if not most) of
the elements are zero. In the specific example, the generalized forces (the q's) for a unit
eq. (21) would be written value of load Pm' all other external loads
zero.
o 0 THE STRAIN ENERGY IN TERMS OF APPLIED LOADS

o If eq. (22) and its transpose are used to sUb-
stitute for the qi'S in eq. (21) one gets
o q3
The problem of computing and tabulating

various Uij's is considered in detail later. In the notation here i and j are used inter-
changeably as are m and n. Also [Omi] is the
RELATING THE INTERNAL GENERALIZED transpose of [Gim] , i.e. interchange of sub-
FORCES TO THE EXTERNAL APPLIED LOADS
scripts denotes transposition (see appendix).
For the statically determinate structures If the matrix triple product in eq. (23) is
considered in this chapter the internal forces formed and defined as
are related to the external loads by use of the
equations of static equilibrium. A set of
linear equations results. ThUS, in the specific
example considered, if PJ. and P2 are the ex-
ternal loads applied as in Fig. A7.32f, then by
statics (refer to Fig. A7.32b) then
A7.21
plication itself yields a summation to complete

(25) the calculation.
From this last discussion it is clear that
eq. (26a) also may be derived by formulating the
Eq. (25) expresses the strain energy as a Dummy-Unit Load equations (Art. A7.7) in matrix
function of the external applied loads. notation.
In the specific example being used To illustrate the application of the matrix
methods presented thus far, a brief and element-
ary example is worked with those tools already
developed.
Example Problem 25
Determine the influence coefficient'matrix
for transverse forces to be applied to the uni-
DEFLECTIONS BY CASTIGLIANO'S THEOREM form cantilever of Fig. A7.33 at the three points
Application of Castigliano's theorem to indicated.
eq. (25),
,® ,®®
,
(25) tbl,b+bJ
3 ~ 3 3
c:::::::=:J~ q~
~~)q"
yields qs
==~)q
q4 6
(26)
Fig. A7.33
Solution:
The steps in passing to eq. (26) may be
demonstrated readily by writing out eq. (25) The choice and numbering of generalized
for, say, a set of three loads (m, n = 1,2,3), forces are shown on the figure. These forces
differentiating successively with respect to were placed so that previously derived express-
PJ., Ps and P" and then re-collecting in matrix ions for the a's could be used. The following
form. member flexibility coefficients were computed.
The matrix ~mnJ gives the deflection at Note that the only non-zero coefficients of
mixed SUbscripts (i not equal to j) are those
the external points "m" for unit values of the for loads common to an element.
loads P and is therefore, by definition, the
n
l
matrix of influence coefficients. L/~sdx L"
COMPARISON WITH DUMMY-UNIT LOADS EQUATIONS a~~ = lEI = 81EI = a
0 S2 =a 44
It is instructive to write eq. (26) out as This expression was adopted from that de-
veloped for a transverse shear force on a canti-
...
.).. lever beam segment in the preceding illustrative
example.
and cmapare the expression with a typical term

from the dummy-unit load method equations, say
6 = Z u~. In the matrix equation (26a) the
AE This expression is for a couple on the end
[GmiJ is the unit (virtual) matrix correspond- of a cantilever segment (of length L/3).
ing to the symbol "u" in the simple sum. The
L/3
[a ij ] are the member flexibilities correspond- ydy LS
as" =a 46
= "EI = 18EI
J
0
ing to ~.
AE
The matrix product [! jn] 1Pn \
This expression is for the cross influence
gives the member load distriuutions due to the of a couple and a shear load on a cantilever
real applied loads, hence these are the "S" segment. Collecting in matrix form,
loads. Finally the operation of matrix multi-
-
•--.III-r...iiiiiiiiiiiii ................= =
~-----,-_ .
1
as per eq. (26).
27 0 0 0 0 The matrix [!mnJ is seen to be symmetric
1 1 about the main diagonal as it must be: from
0 0 0
27 6L Maxwell's reciprocal theorem A = A (see
[U ij] L 3
3EI 0
1
6L V
1
0 0 eq. (20)). mn nm
1 1 A7.10 Member Flexibility Coefficients:

0 0 0 27 6L Compilation of a Library.
1 1
0 0 0 Several member fleXibility coefficients are
6L L2 derived below for various members and loadings.
A more comprehensive listing is available in the
The unit load matrix [GimJ was computed by paper by Wehle and Lansing referenced earlier.
successively applying unit loads at points 1, 2 BARS

and 3 and computing the values of the q's by The energy in a uniform bar under varying
statics. axial force (Fig. A7.34) is
1 0 0
L/3
1 1 0 u = _1
2AE
I[L q
i
+
[GimJ 0 0 o
1 1 1
2L13 L/3 Then referring to eq. (20), Fig. A7.34
0
Note that the first column of [GimJ gives the

~2U
= ~q 2
=.l.. J(l-~L)" dx
AE
i 0
values of the q's obtained for a unit load at

point "I" with no other loads applied. The
second column gives the q's for a unit load at
point "2" only, and so forth.
Finally,
and,
1
27 0 0 0 0 1 0 0
~
3
1 L/3 1 2L/~ 0 -L
27 -L
6L 0 0 1 1 0
~rr;j
L
= 3EI 0 1 OIL 13 0 lL C" 0 0
L/3
0 0 L
o 0 0 1 0 1 1
6AE
o 0 0
276
1 1 1
1 1 2L L/ 0
An equally likely choice of generalized
o 0 0
6L L" 3 3 forces for the above case is shown in Fig.
A7.34a. The strain energy is (x measured from
free end)
Multiplying (see appendix),
[~= Fig. A7.34a
_ L
= 1...JLdx - AE
Should the deflections be desired at the AE
o
three points one forms
A7.23
In the case of tapered members the

coefficients are determined by evaluating in-
tegrals of the form u = 2~' ~ I q, + qj x J <Ix
and from which

_ I JLf(X)dx
Qij - E 0 A(x) L Fig. A7.34e
EI
NOTE: The coeff-
Such a quadrature can always be made in these icients for bars and
problems. For the linearly tapered bar the beams are directly
results may be obtained as functions of the end analogous (compare
area ratios. Thus, Wehle and Lansing give several cases) so
that for tap ere d
beams one uses the
L results for tapered
Leii - 3A.E ¢11
qj --1' ,., ,,J~ qi 1
SHEAR PANELS
bars with EI in place
of AE.
L
A'/
J -- L - ' Ai
Qij 6A.E ¢!j For the rectangular shear panel with a
1 uniform shear flow qi on all edges (Fig. A7.34f)
Fig. A7.34b
Qjj
L
3AjE . ~jj
S surface area
Fig. A7.34f
The trapezoidal shear
panel (Fig. A7.34g) is treated approximately by
using the average shear flow on the non-parallel
sides as though it were constant throughout the
sheet. Thus
S surface area
S
Fig. A7.34c Q11 = Gt
Area varies linearly
Since by statics qj
Fig. A7.34g
h"
qi~' one could use qj as
. an alternate choice of generalized force and
BEAMS
The energy in the uniform beam of Fig.
, jj = (hh:'J\" GtS .
Q
A7.34d is given by
TORSION BAR
A uniform shaft under torque qi has strain
u = 2~' It q
,
energy
ql"L L
Fig. A7.34d U = 2GJ Then Q11 = GJ •
Then
A7.11 Application of Matrix Methods to Various
a11 structures.
Example Problem 26
The tubular steel truss of Fig. A7.35 is to
be analyzed for vertical deflections at points E
and F under several load conditions in which
vertical loads are to be applied to all joints
An alternate choice of generalized forces excepting A and D. The cross sectional areas of
for the beam of Fig. A7.34d is shown in Fig. tUbe member~ are given on the figure.
A7.34e. Set up the matriX form of expression for the
20" c*I
T
Constant Axial
f'R 1 2 3 4
.217
3'1 0
C>
. ./~ ~ 15"
./. I
-L
Then
-=-
Load} qi-qj
at i - e(ii+ O(jj+
1 0 1.0 0 1.0
. F 2 0 0 0 0
o<.ij+<:>(ji
_ L/AE 3 0 -1.0 0 0
Fig. A7.35 Fig. A7.35a 4 0 1.25 0 1.25
deflections at points E and F. 5 0 -1.03 0 -1.03
Solution: 6 -1.0 - .75 0 - .75
The member flexibility coefficient for a 7 1.13 .848 1.13 .848
uniform bar under constant axial load is
L/AE • Fig. A7.36a gives the numbering scheme 8 -.825 -1.65 -.825 -1.65
applied to the members and the q's (these being
one and the same, since q is constant in a given 9 .20 0.40 .20 0.40
member). Fig. A7.36b shows the numbering scheme
adopted for the external loading points.
Then the matrix triple product
~mn]
1 2
= Emi] [a ij ]
[SI3J 4 3
@jn]
9 6
5 was formed giving, per eq. (26),
8
6~ 440 389 257 389 P~
Fig. A7.36a Fig. A7.36b
62 1
389 927 252 789 P2
=-
Member flexibility coefficients were col- 1I 3 E 257 252 257 252 P3
lected in matrix form as
6.. 389 789 252 789 p ..
inches
~ 1 2 3 4 5 6 7 8 9
1 92.2 0 0 0 0 0 0 0 0 The results here give the deflections of

0 0 0 0 0 0 0
all four points. Since only the deflections of
2 0 92.2
points 3 and 4 were desired the first two rows
3 0 0 138 0 0 0 0 0 0
4 0 0 0 146 0 0 0 0 a of ~mn] may be dropped out. The same result
5 a a 0 a 55.7 a 0 a 0
6 a 0 0 a a 183 a 0 0
could have been achieved by leaving out the
7
8
9
0
a
0
0
0
a
0
0
a
a
a
0
a
a
a
0
0
0
165
0
0
a
55.7
0
0
0
229
first two rows of ~mi] 0 he transpose of ~im J)
The matrix form of equation above is useful
in organizing the computation of deflections for
In the case of a pin jointed truss, where a number of different loading conditions. ThUS,
only a single generalized force is required to should there be several different sets of ex-
describe the strain energy per member, the ternal loads P , corresponding to various load-
n
matrix of member flexibility coefficients is a ing conditions, each set is placed in column
diagonal matrix as above. form giving the loads as the rectangular matrix
Unit load distributions were obtained by
placing unit loads successively at external
loading points 1, 2, 3 and 4 (Fig. A7.36b). ~nk] , k different numerical subscripts for the
The results were collected in matrix form as
load conditions. The matrix product
A7.25
(26b) a .... = LAB~J

now gives the deflections at each point (m) for
the various load conditions (k). a 77 LB~
GJ
Example Problem 27
Deflections at points 1, 2, 3 and 4 of the
truss of Fig. A7.35 are desired for the follow- a 66 = a 6B = LBc;I
ing loading conditions: ;lEI
Condition No. P~ P lO P" P4 ( see Fig. Collected in matrix form (noting that
1 2500 2000 800 450 A7.36b) GJ = EI )
:"1750 1.3
2 -1200 -800 -2100
3 1800 1470 -1200 -1100
1 2 3 4 5 6 7 8
Solution: 1 9.0 0 0 0 0 0 0 0
2 0 3.9 0 0 0 0 0 0
The matrix product formed per eq. (26b) was 3 0 0 9.0 0 0 0 0 0
set up as 4 0 0 0 15,600 648 0 0 0
5 0 0 0 648 36.0 0 0 0
440 389 257 389 2500 -1200 1800 6 0 0 0 0 0 36.0 0 648
7 0 0 0 0 Q 0 46.8 0
389 927 252 789 2000 - 800 1470
~mkJ = E
1 8 0 0 0 0 0 648 0 15,600
257 252 257 252 800 -2100 -1200
389 789 252 789 450 -1750 -1100
The unit load distributions ~i~ were obtained

Example Problem 28
For the landing gear unit of Example Prob- by applying unit external applied loads, numbered
lem 20, Fig. A7.23 find the matrix of influence and directed as in Fig. A7.37b.
coefficients relating deflections due to lift .---- ,...... -
and drag loads acting at point A and torque
about the axle A-B. I~ 1 2 3
1 1.0 0 0
Solution:
2 0 1.0 0
The structure was divided into elements
3 0 0 1.0
and the set of internal generalized forces ap-
plied as shown in Fig. A7.37a. (Torques and 4 .342 0 0
moments are shown vectorially by R.H. rules).
Axial stresses were neglected in C-B. 5 3 0 0
6 0 -.937 1.026
7 0 .342 2.811
8 0 0 1.0
'-- -
At this point the engineer may consider the
problem as solved, for the remaining computation
is a routine operation:
Example Problem 29
The following member flexibility coefficients The beam of example problem 21 is to be re-
were determined solved by the matrix methods presented herein.
Influence coefficients for points F, G and H The matrix triple product

are to be found.
Solution:
Fig. A7.38 shows the choice and numbering
of generalized forces.
completes the calculation.
/ ' (coeff's
Q 33 Q~3 + Q ~3/'
Example Problem 30
= coll'td
qe \ ' fr.two
Deflections of statically indeterminate
ljfdTof
structures often may be computed successfully
by the methods of this chapter provided that
some auxiliary means is employed to obtain an
approximation to the true internal force dis-
tribution. The exact internal force distribu-
~ r, rq~ tion is not necessarily required in making de-
i
, , q7 r , q4 flection calculations inasmuch as such a calcu-
lation amounts to an integration over the
Fig. A7.38 structure - an operation which tends to average
out any errors. Thus one may use the engineering
theory of bending (E.T.B.), experimental data,
previous experience, etc. to obtain reasonable
No forces were shown applied to the lower flange estimates of the internal force distribution for
elements as these were known to ue equal to unit loadings.
those of the upper flange due to symmetry. In the following problem the matrix of in-
Entries were made for a ij in matrix form as be- fluence coefficients is determined for a single
cell, three-bay box beam (3 times indeterminate)
low. Entries for a 33 and a ee were quadrupled as by using the E.T.B.
these occur in two identical members each on 1QQ Fig. A7.39a shows an idealized doubly
and bottom. Entries for a 3e , aee and Gee were symmetric single cell cantilever box beam having
doubled. (See Art. A7.10 for coefficient three bays. Determine the matrix of influence
formulae. ) coefficients for the six point net indicated.
2 3 4 5 7 8 9
1 10
2 2085
3 17.8 4.45 10"
L
tl~ = 10-~ 4
5
6.95
2350'
6 4.45 17.8 4.45
7 7.64
8 2600
9 4.45 8.89
Unit load values were obtained as in Fig. Stringer Areas

A7.27, considered to be external loading num- Constant = • 33 in 2
ber "two". Similar diagrams were drawn for unit

loads at points "one" (H) and "three" (F).(ef. Fig,
A 7-25)
Fig. A7.39a
i 1 2 3
1 l.0 0 0 Solution:
2 .0667 0 0
Fig. A7.39b is an exploded view of the beam
3 l.201 0 0 showing the placement and numbering of the in-
@i~
4 0 l.0 0 ternal generalized forces. Note that only the
upper side of the beam was numbered, the lower
5 .0540 .0600 0 side being identical by symmetry.
6 2.184 l.09 0 Member flexibility coefficients were com-
7 0 0 l.0 puted by the formulas of Art. A7.10 and entered
in matriX form as below. Note that all entries
8 .0447 .0496 .0545 for which there were corresponding loads on the
9 3.00 2.00 l.00 lower surface of the beam were dOUbled. By
this means the total strain energy of the beam
A7 27
4
flow (q = ~) due to the torque developed in
transferring the load to one side.
The matrix triple product
completes the solution.
A7.12 Deflections and Angular Change of the Elastic Curve

of Simple Beams by the "Method of Elastic Weights"
(Mohr's Method).
Fig. A7.39b
In the calculation of structural deflec-
was accounted tor. Note also that entries for tions there occur many steps involving simple
U44 , U55 , u ee and u~o.~o were re-doubled as integral properties of elementary functions.
each of these q's act on two (identical) mem- The Method of Elastic Weights (and the Area
bers. Moment Method to follow in Art. A7.l4) owes its
popularity in large measure to the fact that it
enables the analyst to write down many of these
8 1 2 3 4 5 6 7 6 9 10 11 12 13 14 15 integral properties almost by inspection, rely-
1 1520 ing as it does upon the analyst's familiarity
2 32,000
3 1520 With the properties of simple geometric figures
4 ~~ ~~
5 80.96 20.24
For finding the deflection of a point on a
6
7
4510
46,800
simply supported beam relative to a line joining
8 4510 the supports, the Method of Elastic Weights
9 20.24 80.96 20.24
10 20.24 80.96 20.24 states:
11 8000
12
13
62,000
8000
The deflection at point A on the elastic
14 20.24 40.48 curve ~ §:. simple beam is equal to the bending
15 20.24 40.48
M
moment at ! due to the Ef diagram acting as §:.
-
distributed beam load.
Note: VOID SPACES INDICATE ZEROS.
Spelled out in steps:
Unit load distributions were obtained for i-The ~ diagram is drawn just as it
successive applications of unit loads to points
one through six (Fig. A7.39a). The internal occurs due to the applied beam load
forces predicted by the E.T.B, for a load ii - This diagram is visualized as being
through the shear center (center of beam, due to the loading on a second beam (the conjugate
symmetry) were superposed on the uniform shear beam) supported at the points of reference for
the deflection desired
~ 1 2 3 4 5 6 iii - The bending moment in this conjugate
beam is found at the station where the deflec-
1 0.3 0.1
tion of the original beam was desired. This
2 0.1 -0.1
bending moment is equal to the desired
El~ =
3 0.1 0.3 deflection.
4 2 2
5 2 2 To prove the theorem, consider the dummy-
6 .0875 .0125 0.15 0.05 unit load (virtual work) equation
7 .0375 -.0375 0.05 -0.05
8 .0125 .0875 0.05 0.15
9 2.67 2.67 1.33 1.33
10 2.67 2.67 1.33 1.33 This expression equates the external virtual
11 .0423 .0023 .0713 .0179 0.100 .0333 work done by a unit load, applied at a point de
12 .020 -.020 .0267 -.0267 .0333 -.0333
fleeting an amount 0, to the internal virtual
work on a beam element experiencing an angular
13 .0023 .0423 .0179 .0713 .0333 .100
14 3 3 2 2 1 1 change d Q =Mdx
EI • The sum (integral) of such
15 3 3 2 2 1 1 expressions throughout a beam gives th~ total
deflection at the point (cf. eq. 18). We now In Fig. f consider Mdx as a load on a simply
show that the deflection expression, using the EI
above equation, is the same as the bending supported beam and determine the bending moment
moment expression for a simple beam loaded by at pOints band c due to Mdx acting at point a.
an "elastic weight" ~~ • L _ MLdx EI
2 - 8EI
In Fig. A7-40, the loading of (a) produces
the real moments of (b). Consider the deflec- L _ MLdx
tions of points Band C due to the angular 4; - 16EI
change lMdx
[ [ in a beam element at A" (Fig. A7-40c ) • These values of the beam bending moments at
points band c are identical to the deflections
at band c by the virtual work equations. The
moment diagram m for a unit load at band c
(Figs. d and e) is numerically precisely the same
as the influence line for moment at points b
and c.
{-h-+.--h-.J.---h--+--hj
4 4 4 4 Therefore deflections of a simple beam can
(a) be determined by considering the M curve as an
EI
~
-->l I--dx
imaginary beam loading. The bending moment at
any point due to this M loading equals the de-
(b) EI
flection of the beam under the given loads.
Mdx
EIr?! Likewise it is easily proved that the angu-
lar change at any section of a simply supported
(c) beam is equal to the shear at that section due to
the M diagram acting as a beam load.
+ - EI
~
(d) mb, moment diagram for unit
A7.13 Example Problems
load acting downward at point B Example Problem 31. Find the vertical
3 deflection and slope of points a and b for beam
;~
and loading shown in Fig. A7.41. The lower Fig.
shows the moment diagram for load P acting at
center of a simple beam.
(e) m c , unit load at point C
P
Mdx
~EI
f
3 Mdx
(f)
t
1Mdx a b
EI Constant
4EI 4EI t_------ -----_of' P

L
Fig. A7.40 P 2
"2
Fig. A7.41
For a unit load at point b, Fig. d shows the m
diagram. The value of m at the midpoint of dx
(point A) = L/8. Hence
Mdx L MLdx: PL lil PL lil
5b = EI . 8 = 8 EI 16 16
For deflection of point c, draw m diagram for a Deflection at point a equals bending mom-
unit load at C (see Fig. e). Value of m on ele- ent due to M diagram as a load divided by EI.
ment dx: L (See lower Fig. of Fig. A7.41)
16
PL2 L PL2 L) 1 11 PL3
Hence Mdx L MLdx 5a = ( 16 4 - 64 . 12 EI = 768 EI
5c
EI 16 16EI PL2 L PL2 L) 1 _ 1 PL3
5b = ( 16 2 - 16 . 6 EI - 48 EI
* For simplicity the points A, Band C were placed at the one- The angUlar change of any point equals the
quarter span points. The reader may satisfy himself with
the general character of the proof by substituting xA, xB and
shear due to WEI diagram as a load.
XC for the point locations and then following through the PL2 PL2) 1 3 PL 2
argument once again. Cla =( 16 - 64 EI = 64 EI
A7.29
o (Slope is horizont- I<L Airplane

al or no change , 1--50"- ""-;)0"- te30"
from original di-
rection of beam
axis.) ..:- Rear Beam
1 vAileron E F~
Example Problem 32. Det3rmin3 th'3 deflection . A B C
of a simple beam loaded uniformly as shown in " Front Beam

Fig. A7.42. The bending moment expression for
a uniform load M = wlx - wx 2 or parabolic as 7
2""" 2""" Plan View 1/2 Wmg
shown in Fig. A7.42a. The deflection at mid-
point equals the bending moment due to M dia- ,....-Air Load 25#/m. . r 0"1 3
gram as a load. Ii f f f f iff f f f f If 1f i i III 1f f f

I... Load on Rear Beam
rl1 iff 1 f nli'i
J
15#/in.
/ w/in.
l:'
wL
11111L If
rr 111 wL
Fig. A7.43
Solution:- Due to the beam variable moment of in-

2 -2- ertia the beam length between A and C will be
diVided into 10 equal strips of 10 inches each.
Fig. A7.42 The bending moment M at the midpoint of each will
be calculated. The elastic weight for each strip
will equal Mds, where ds = 10" and I the moment
-1-
of inertia at midpoint of the strip. These elas-
tic loads are then considered as loads on an im-
aginary beam of length AC and simply supported at
A and C. The bending moment on this imaginary
-.l WL3 --.L WL3 = Area beam at point B will equal the deflection of B
24 24
with respect to line joining AC.
Fig. A7.42a The bending moment at C = 15 x 30 x 15 +
10 x 15 x 10 = 8250"#
The shear load at C = (15 + 25) x 30 = 600#
_(1 3 L 1 3 3
°center - 24 wL . 2 - 24 wL • 8 2 EI
L) I -2-
Bending moment expression between points C
5 wL4 and A equals, M = 8250 + 600X + 12.5x2 , where
384 E'I x = 0 to 100.
Table A7.6 gives the detailed calculations
Ucenter =(2~ WL
3
- 2~ 3
WL ) Ell =0 for the strip elastic loads. The I values as-
sumed are typical values for a aluminum alloy
Slope at supports = the reaction beam carrying the given load. The modulus of
elasticity E = 10 x 106 is constant and thus can
Example Problem 33. be omitted until the final calculations. The
Fig. A7.43 shows the plan View of one-half figure below the table shows the elastic loads
of a cantilever Wing. The aileron is supported on the imaginary beam.
on brackets at points D, E and F With self-align- Table A7.6
ing bearings. The brackets are attached to the Strip ds II= I Elastic
wing rear beam at points A, B,and C. When the No. in. moment at mid- load
at midpoint lids
wing bends under the air load the aileron must point -I-
likewise bend since it is connected to wing at
I 10 11563 5.5 21000
three points. In the design of the aileron beam 2 10 20063 6.5 30900
and similarly for cases of wibg flaps this de- 3 10 31063 7.5 41450
4 10 44580 8.5 52300
flection produces critical bending moments. As-
suming that the running load distributed to the
rear beam as the wing bends as a unit is as shown
in the Fig., find the deflection of point B With
5
6
7
8
9
10
10
10
10
10
10
10
60580
79050
100050
123550
149550
9.5
12.0
16.0
20.0
24.0
1781_50._L-.-.. 28~_
1 63850
65940
62530
61770
62320
63600
respect to straight line joining points A and C,
which will be the deflection of E with respect to o
o
o 0
o 0
line joining D and F if bracket deflection is <0
t<)
Ol
o
0
rl
neglected. The moment of inertia of the rear <0 t<) C\l
beam between A and C varies as indicated in

the Table A7.6
Bending moment at point B due to above elastic A7.14 Deflections of Beams by Moment Area Method*
loading = 7,100,000 •.. deflection at B relative For certain types of beam problems the meth-
to line AC ~ 7,100,000 = .71 inch od of moment areas has advantages and this meth-
E-10,000,000 od is frequently used in routine analysis.
Example Problem 34 Angular Change Principle. Fig. A7.44 shows a
Fig. A7.43a shows a section of a cantilev- cantilever beam. Let it be required to deter-
er wing sea plane. The wing beams are attached mine the angular change of the elastic line be-
to the hull at points A and B. Due to wing tween the points Aand B due to any given loading.
loads the wing will deflect vertically relative From the equation of virtual work, we have
to attachment points AB. Thus installations
such as piping, controls, etc., must be so lo-
l\.
cated as not to interfere with the wing deflec-
tions between A and B. For illustrative pur-
poses a simplified loading has been assumed
as shown in the figure. EI has been assumed as
constant Whereas the practical case would in-
\
volve variable I. For the given loading deter-
mine the deflection of point C With respect to
the support points A and B. Also determine the
vertical deflection of the tip Aoints D and E.
o 0 \
o 0 o
o o
t- 400" 9- 300'1 ~,
\
C'<I
~
If)
Fig. A7.43a
D EI Constant E
0 1
01
Fig. A7.43b
\ C
01
0"1 Fig. A7.43c
q<1
.....267'1
o r-----'------'----''''H....1IIi>l'---.L.---L.----. o
C'<I C'<I
q< q<
Fig. A7.43d
M
~
M
~
III
~ ~
III
Solution:- Fig. A7.43b shows the bending moment III I
diagram for the given wing loading. To find the III
deflection of C normal to line joining AB we ~~ Distributed Beam Load
treat the moment diagram as a load on a imaginary
beam of length AB and simply supported at A and B
(See Fig. A7.43c.) The deflection of C is equal
numerically to the bending moment on this ficti-
cious beam.
Hence Elo c = 25920 x 40 - 25920 x 20
or 0c - 518000
-~
To find the tip deflection, we place the elastic Fig. A7.44

loads (area of moment diagram) on an imaginary
beam simply supported at the tip D and E (See
Fig. A7.43d). The bending moment on this imagi-
nary beam at points A or B will equal numerically (A Mdx Where m is the moment at any
the deflection of these points With respect to ab =)B Iff m, section, distant x from B due
the tip points D and E and since points A and B to unit hypothetical couple
actually do not move this deflection will be the applied at B. But m = unity
movement of the tip points With respect to the at all points between Band A.
beam support points. Therefore aB =(A Mdx
Bending moment at A = 193420 x 700 - 40000 x )B EI
433 - 127500 x 124 = 102200000. . .• 0tip = (*Moment area first developed by Prof. C. E. Greene and
102,200,000 published in 1874. )
EI
A7.31
Referring to Fig. A7.44, this expression repre- Hence (

sents the area of the M diagram between points PL 2 2 ) 1 PL3
°B = - ~ . 3 L EI =- 3EI
EI
Band A. Thus the first principle:- "The Example Problem 36
change in slope of the elastic line of a beam Fig. A7.46 illustrates the same simplified
between any two points A and B is numerically wing and loading as used in example problem 34.
equal to the area of the ~ diagram between Find the deflection of point C normal to line
EI joining the support points A and B. Also find
these two points." the deflection of the tip points D and E relative
to support points A and B.
Deflection Principle
In Fig. A7.44 determine the deflection of o
o
point B normal to tangent of elastic curve at A. If>
In Fig. A7.44 this deflection would be vertical

since tangent to elastic line at A is horizontal. 400"
From virtual work expression 0B = I A Mdx

1ff m,
B Fig. A7. 46
where m is the moment at any section A distance
x from B due to a unit hypothetical vertical
load acting at B. Hence m = l.x = x for any
point between Band A.
Hence -fA Mdx
°B - B 1ff' x This
exnression represents the 1st moment of the
M diagram about a vertical thru B. Thus the Solution:-
EI Due to symmetry of loading, the tangent to
deflection principle of the moment area method the deflected elastic line at the center line of
can be stated as follows:- "The deflection of a airplane is horizontal. Therefore, we will find
point A on the elastic line of a beam in bending the deflection of points A or B away from the
normal to the tangent of the elastic line at a horizontal tangent of the deflected beam at point
point B is equal numerically to the statical mo- C which is equivalent to vertical deflection of
ment of the M area between points "A" and "B" C with respect to line AB.
EI Thus to find vertical deflection of A With
about point A". respect to horizontal tangent at C take moments
of the M diagram as a load between points A and
Illustrative Problems
Example Problem 35. Determine the slope EI
C about point A.
and vertical deflection at the free end B of the Whence
cantilever beam shown in Fig. A7.45. EI is con-
stant. (area) (arm)
°A 1
1000 (tangent at C) = EI (650 + 646 )
2 40 x 20
E; (518400) = deflection of C normal to AB.

Fig. A7.45
To find the vertical deflection of the tip point
...l D With respect to line AB, first find deflection
D;<~~~00.0L2'?zz2~ of D with respect to horizontal tangent at C and
2
PL subtract deflection of a With respect to tangent
--2-
at C.
Solution:- The moment diagram for given load is
triangular as shown in Fig. A7.45. Since the ~ (respect to tangent at C)
1000 1 (40000 x 267
= EI +
beam is fixed at A, the elastic line at A is
horizontal or slope is zero. Therefore true 127500 x 576 + 25920 x 720) = iI (102700,000)
slope at B equals angular change between A and B
which equals area Of moment diagram between A (see Fig. A7.46 for areas and arms of M/EI dia-
and B divided by EI. gram). subtracting the deflection of A with re-
spect to C as found above we obtain
Hence 1 PL2
uB = (-PL. L/2) EI = - 2EI ~
1000 ( respect to line AB) -_ EI
1
(102,700,000 -
The vertical deflection at B is equal to the 1st
moment of the moment diagram about point B di- 518400) = E; (102,180,000)
Vided by EI, since tangent to elastic curve at A
is horizontal due to fixed support.
A7.15 Beam Fixed End Moments by Method of Area MA =- Pab 2 and MB =- Pba 2 /L 2 where b = (L - a)
Moments --rr
From the two principles of area moments as To find the fixed end moments for a beam with
given in Art. A7.14, it is eVident that the de- variable moment of inertia use the M/I diagrams
flection and slope of the elastic curve depend in place of the moment diagrams.
on the amount of bending moment area and its lo- Example Problem 37
cation or its center of gravity.
Fig. A7.48 shows a fix-ended beam carrying
Fig. A7.47 shows a beam fixed at the ends two concentrated loads. Fin~ the fixed-end
and carrying a single load P as shown. The bend
ing moment shown in (c) can be considered as moments MA and MB'
made up of two parts, namely that for a load P
acting on a simply supported beam which gives J100#
the triangular diagram with value pa (L-a)/L MA (a)
for the moment at the load point, and secondly ~ 9"+-12"+9" ./
a trapezoidal moment diagram of negative sign
with values of MA and MB and of such magnitude
as to make the slope of the beam elastic curve
zero or horizontal at the support points A and (b)
B, since the beam is considered fixed at A and B.
The end moments MA and MB are statically
indeterminate, however, with the use of the two Fig. A7.48
moment area principles they are easily determin-
ed. In Fig. b the slope of elastic curve at A
and B is zero or horizontal, thus the change in
slope between A and B is zero. By the 1st (c)
, i
/...-a--r
/: A
P
EI Constant
.--- L
Solution:- Fig. b shows the static moment diagram
A assuming the beam simply supported at A and B.
~-----~B (b)
For simplicity in finding areas and taking moments
~astic Curve
of the moment areas the moment diagram has been
Fig. A7. 47 diVided into the 4 simple shapes as shown. The
~a~-a) ':?~ centroid of each portion is shown together with
the area which is shown as a concentrated load at
~_~i (c) the centroids.
Fig. C shows the moment diagrams due to un-
I known moments MA and MB. The area of these tri-
angles is shown as a concentrated load at the
principle of area moments, this means that the centroids.
algebraic sum of the moment areas between A and Since the change in slope of the elastic
B equal zero. Hence in Fig. c curve between A and B is zero, the area of these
(-MA -Mb) L+ (Pa (L-a) ~ moment diagrams must equal zero, hence
0 - - - - - - -(A) 5265 + 14040 + 2160 + 6885 + 15MA + 15MB = 0
2 L' 2
or
In Fig. b the deflection of B away from a tang- 15MA + 15MB + 28350 = 0 - - - - - - - - - - (1 )
ent to elastic curve at A is zero, and also de-
flection of A away from tangent to elastic curve The deflection of point A away from tangent to
at B is zero. elastic curve at B is zero, therefore the first
Thus by moment area principle, the moment moment of the moment diagrams about point A
of the moment diagrams of Fig. C about points A, equals zero. Hence,
or B is zero. 5265 x 6 + 15 x 14040 + 17 x 2160 + 24 x 6885 + 150MA +
300 MB = 0 or 150 MA + 300 MB + 444600 = 0- - - -(2)
Taking moments about point A:- Solving equations (1) and (2), we obtain
MA = - 816 in. lb.
mA = Pa 2(L_a) .§g+Pa(L-a) 2x (a + L-a)+~ x ~ + MB = - 1074 in. Ibs.
2L 3 2L 3 23
With the end moments known, the deflection or
MBLx2 L = 0 - - - - - - - - - - - - - - - (B) slope of any point on the elastic curve between
23 A and B can be found by use of the 2 principles
of area moments.
SolVing equations A and B for MA and MB
A7.33
A7.l6 Truss Deflection by Method of Elastic

Weights d
If the deflection of several or all the (1)

joints of a trussed structure are required, the
method of elastic weights may save considerable
/' --
/
L-'- - - - - ~al--""tL---"o----<"-k--!i'---+------\g
time over the method of virtual work used in 1--. d 6 Panels @ p
previous articles of this chapter. The method Bar bc (2)
in general consists of finding the magnitUde and ~ t ~ection Diagram
location of the elastic weight for each member uLlL = 4pLlLbc~ -fo~··LlL Shortening
of a truss due to a strain from a given truss -r- ~ - LI L of Member bc
~
loading or condition and applying these elastic j ~ Elastic Load on Imaginary Beam
weights as concentrated loads on an imaginary f~ LlL t ilLbc

3r bc 3r (3)
beam. The bending moment on this imaginary beam
due to this elastic loading equals numerically ~-6
L~-
"--Moment Curve (4)
the deflection of the given truss structure. for Elastic Loads
Consider the truss of diaGram (1) of Fig.
A7.49. Diagram (2) shows the deflection curve Bar ck LlL(d+6p)
for the truss for a ~L shortening of member bc,
all other members considered rigid. This de- /1:\ 3r~ (5)
flection diagram can be determined by the vir-
tual work ex pression 0 = u~L. Thus for deflec-
~ection Curve
for LI L Lengthening
tion of joint 0, apply a uhit vertical load act- of Member ck
ing down at joint O. The stress m in bar he due
ilL d 2P
1'1 p El stic Loads on
to this unit load = 2 2P 4P Therefore o k LI Ld Imaginary (6)
:3 r 3r LI~(d+-fu» 6IJr Beam
/1\.~
~
The deflection at other
lower chord joints could
be found in a similar / ~oment Curve (7)
manner by placing a unit load at these joints. for Elastic Loads
Diagram (2) shows the resulting deflection curve. Fig. A7.49
This diagram is plainly the influence line for
stress in bar bc multiplied by ~Lbc' Table A 7. 7 Equations for Elastic Weights
Diagram (3) shows an imaginary beam loaded Elastic-Weight for Chord Members
with an elastic load ~Lbc acting along a verti- (See Member ab)
r Upper Chord Lower Chord
cal line thru joint 0, the moment center for ob- w=LlL
taining the stress in bar bc. The beam reac-
~i
a b
tions for this elastic loading are also given.
Diagram (4) shows the beam bending moment dia- '~
gram due to the elastic load at point O. It is
noticed that this moment diagram is identical to
l~
w=-
the deflection diagram for the truss as shown in h LlL
diagram (2).
The elastic weight of a member is therefore
equal to the member deformation divided by the
arm r to its moment center. If this elastic
load is applied to an imaginary beam correspond-
ing to the truss lower chord, the bending moment
on this imaginary beam will equal to the true
m z\ PiA
a
o LlL
w=-
b
r
I
w=~L
truss deflection. b ~ h~
~
Diagram 5, 6 and 7 of Fig. A7.49 gives a h~~
similar study and the results for a ~L lengthen-
ing of member CK. The stress moment center for ~
w=-
this diagonal member lies at point 0', which lies r
outside the truss. The elastic weight ~L at w= Member Axial Deformation -LlL
-
q Perpendicular Arm to Moment Center r
point 0' can be replaced by an equivalent system The moment center 0 of a chord member is the
at points 0 and k on the imaginary beam as shown intersection of the other two members cut by the
in Diagram (6). These elastic loads produce a section used in determining the load in that memb-
bending moment diagram (Diagram 7) identical to er by the method of moments
the deflection diagram of diagram (5). The sign of the elastic weight wfor a chord
Table A7.7 gives a summary of the equations member is plus i f it tends to produce downward de-
for the elastic weights of truss chord and web flection of its point of application. Thus for a
members together with their location and sign. simple truss compression in top chord or tension
in botton chord produces downward or posi tive
elastic weight
A7 34 DEFLECTIONS OF STRUCTURES
1000 1000 1000 1000

Al -5630 ·I~ -7500
Table A7. 7 WEB DIAGONAL MEMBERS -563010 lc
(continued) (See Member ab)
Fig. A7.50
this step are given in the figure, the stresses
being wri tten adjacent to each member. The next
step or steps is to compute the member elastic
weights, their location and their sense or dir-
~ ~
ection. Tables A7.8 and A7.9 gives these calcu-
lations. Table A7:8
Chord Member Elastic Weights
tlL tlL lIembe r Length Area Load PL ~ tiL Arm Elastic Point of
P=Q=-r P=Q=r L A P Ai r wt. appl1ca-
Ea29xl0 6 w
r
=tiL tion
joint
AB 30 .2172 -5630 -.0268 24 • QOl117 b

For a truss diagonal member the elastic Be 30 .2172 -7500 -.0357 24 .001487 c
weights P & Q have opposite signs and are as- CD 30 .2172 -5630 -.0268 24 .0011p d
ab 30 .1198 3120 .0269 24 .001117 A
sumed to be directed toward each other or bc 30 .1198 6870 .0592 24 .002465 B
away according as the member is in compres- cd 30 .1198 6870 •• 0592 24 .002465 C
de 30 .1198 3120 .0269 24 .001117 D
slon or tension. In fig. a, P is greater
than Q and P is located at the end of the di-
agonal nearest the moment center O. Downward
elastic weights are plus.
Table A7.9 Web Member Elastic Weights.
TABLE A7.7 (CONTINUED)
Truss Verticals
.,. .<: ..
+'
.....
+'
./,
(See Member ab)
a .
+'
il., .,=... ....
.,<
Load
P
tiL =
PL
Ai rl
P •
~
rl
...
'"' r2
Q -
tI L
1'2 '"'
... <'" '"'"
~
:01 < '" <
.... ......
aA 28.25 .242 -5880 -.0236 12.75 -.00185 a 12.75 .00185 A
PI I
Ab
bB
.
.146
.146
4710 .0314 ""
-2355 -.0157
-.00246
-.00123.... A
b
.00246
.00123
b
B
d
P=Q=T
tlL Bc
cC
....
.093
.093
1177 .0123
1177 • 0123 ....
-.000965
-.000965
B
C ....
.000965
.000965
c
c
Cd
Dd
..
• 146
• 146.
-2355 -.0157
4710 . 0314
-.00123
-.00246..
..
d
D
....
.00123
.00246
C
d
Al~lh
De • 242 -5580 • 0236 -.00185 e .00185 D
Fig. A7.5l shows the elastic weights obtained

from Tables 8 and 9 applied to an imaginary beam
1 r b-d'~
P <'Q
whose span equals that of the given truss.
P = h~ tlL Q= tl L
These elastic weights are the algebraic sum of
h2 d d the elastic weights acting at each truss joint.
The elastic weight P acts at foot of verti- 1:'- 1:'- 1:'- 1:'- 1:'-
L[,) 0 rl 0 L[,)
cal and downward if vertical is in tension. co c;J; c;J; co
~
L[,) 1:'- 1:'-
rl 0 N N c;J; N N rl
Q acts opposite to P at far end of chord memb-
er cut by index section 1-1 used in finding
0
0
. . 0
0
0
0
. .
0
0
0
0
0
0
0
0
0
0
0
0
. Fig. A7.5l
stress in ab by method of sections.
a 11
A
1 1 1 1 1
b B c C d D e
A7.l7 Solution of Example Probleas. ~ 8 spaces at 15" _ _
The method of elastic weights as applied to Ra = .005412 Re = .005412
truss deflection can be best explained by the
solution of several simple typical trusses. The deflection at any joint equals the bending
Example Problem 38 moment on the imaginary beam of Fig. A7.51.
Fig. A7.50 shows a simply supported truss Defl. at A = (.005412 + .00185)15 = .007262 x 15 =
symmetrically loaded. Since the axial deforma- .109"
tions in all the members must be found, the Defl. at b = .109 + (.007262 - .000507)15 = .109 +
.006755 x 15 = .209"
first step is to find the loads in all the memb- Defl. at B = .209 + (.006755 - .002347)15 = .209 +
ers due to the given loading. The results of .004408 x 15 = .275"
A7.35
Den. at c = .275+ (.004408- .0027)15 = .301" beam loaded with the elastic weights from Table
The slope of the elastic curve at the truss joint A7.ll. Table A7.l2 gives the calculation for the
points equals the vertical shear at these points joint deflections.
for the beam of Fig. A7.51.
Example Problem 39
Find the vertical deflection of the joints
of the Pratt truss as shown in Fig. A7.52. The
member deformationsAL for each member due to the
given loading are written adjacent to each memb-
er. Table A7.l0 gives the calculation of member f - - - - - - 5 @ 25" = 125"
elastic weights. Fig. A7.53 shows the imaginary
beam loaded With the elastic weights from Table Fig. A7.54
A7.l0. The deflections are equal numerically to
Table A7.11
the bending moments on ..t his beam.
Elastic Weight of Chord Members
0b .01855 x 25 = .465"
0B = .465 + .053 (AL in Bar Bb) = .518" lIember t:. L r w -ll
r
Joint
0c .01855 x 50 - .00387 x 25 - .833"
0c = .833 + .031 (AL of Cc) = .864 AK
DC
.0825
-.055
15.0
17.17
.00550
.00320
B
K
0D .01855 x 75 - .00387 x 50 - .00623 x KJ .096 20.0 .00480 C
25 = 1.03" CD -.069 21.14 .00322 J
J1 .106 22.50 .00471 D
DE -.0736 21.14 .00348 1
1H .113 20.0 .00565 E
.00437
J EE
HG
-.075
.108
17.17
15.00 .00720
H
F
A,r=~~-';';---'-'~~~L...:..='"-----::-~11 G
Elastic Weight of Web lIembers
lIember t:.L rl P -M Joint r2 Q = t:.L Joint

6 @ 25" = 150" rl r2
AB -.0432 9.60 -.00450 A 9.60 .00450 B

Fig. A7.52 BK .072 9.60 -.00748 B 11. 20 .00643 K
KC -.0635 9.27 -.00685 K 10.60 .00600 C
CJ .085 10.60 -.00802 C 11.26 .00755 J
JD .0138 10.33 .00134 J 10.93 -.00126 D
Table A7.10 ID .0496 10.33 .00480 1 10.93 -.00454 D
E1 .044 10.60 -.00415 E 11.26 .00891 1
Elastic Weight Chord lIembers HE -.0254 9.27 -.00274 H 10.60 .00240 E
FH .0985 9.60 .01028 F 11.20 .00878 H
lIember t:. L r "=~ Joint GF -.0585 9.60 -.00610 G 9.60 .00610 F
r
AB .061 30 .00203 b
0 0
Be .061 30 .00203 b It) -<
CD .066 30 .0022 c '<t' co
0 0
bc -.083 30 .00277 C q q
Cd -.091 30 .00304 D
A G
Elastic Weight of Web lIembers B K C J D E H F
ex> '<t'
0 a>
P Q I:-
-<
q Fig. A7.55
...q
0
lIember t:. L rl P =A1... Joint r2 Q • t:.L Joint

rl r2
Ab .124 19.2 -.00648 A 19.2 .00648 b Table A7.12
Bb .053 cP 00
bC .128 19.2 -.00667 b 19.2 .00667 C
cC -.031 25 -.00124 C 25 .00124 D lIoment = 12.5 x
CD .080 19.2 -.00417 c 19.2 .00417 D Panel Panel Shear Z;Shear Shear = de- Point
Dd 0 flection
0 0 A
ex>
'<t'
co
I:-
ex>
C')
...
C')
co
0
a>
co
...
C')
co
I:-
ex>
C')
ex>
'<t'
co
AB
BK
.02158
.01!106
.02158
.04064
.27
.508
B
K
0 0 0 -< 0 0 0 KC .01628 .05692
q q .710 C
q q 0 CJ .01350
ql JD .0 13!1
.07042 .881 J
"ts...
j I I I t .07181 .900 D
G
~i
D1 .00248
B c D E F 1B -.00971
.07429 .930 1
...
-< -<
EH
HF
-.U1361
-.024
.06458
.05097
.026"5
.81
.638
.336
E
H
q q FG .02704 F
Fig. A7.53 .000u,," 0 G
.. error
Example Problem 40
Find the vertical joint deflections for the Example Problem 41
unsymmetrically loaded truss of Fig. A7.54. The Fig. A7.56 shows a simply supported truss
AL deformations for all members are given on the with cantilever overhang on each end. This sim-
Figure. Table A7.ll gives the calculation of plified truss is representative of a cantilever
the elastic weights, their signs a nd points of wing beam the fuselage attachment points being
application. Fig. A7.55 shows the imaginary
at e and en. The ~L deformation in each truss lever over-hang portion of the truss relative to
member due to the given external loading is giv- support points e and e' .
en on the figure. The complete truss elastic Since the cantilever portion is not fixed at
loading will be determined. With the elastic e since the restraint is determined by the truss
loading known the truss deflections from various between e' and e, this fact must be taken into
reference lines are readily determined. account in loading the cantilever portion. The
reactions on the beam of Fig. A7.58 represent
the slope at e due to the elastic loading between
e and e'. This elastic reaction in acting in the
reverse direction is therefore applied as a load
to the imaginary beam between e and a as shown in
A7.56 Fig. A7.59.
~ ~ ~ ~ to M eo') o:t'N '"
M ~
oM _ 0
r j. r j. I..• '"f I'
U':lNo:t'NMNQ'l1 C'> C'>
0 0000 0_ o 0
o
~ ~I ~
00 00 0000
·1 . r
L.!... I
,~,
a' Al b l B' c' C' d' D' e' E' f
Elastic Weight Loading
Fig. A7.57
Table A7.13
Elastic Weights of Chord Members R = .00312 R = .00312

tiL Fig. A7.58
Member r w = tIL Apply at
r Joint
In finding deflections this overhang elas-
ab .080 20 .0040 A
AB -.070 20 .0035 tically loaded portion is considered as fixed at
b
be
Be
.074
kO
-.060
20 .0037 a and free at e. The bending moment at any point
B
.0030 c
cd .070
20 .0035 on this beam equals the magnitude of the vertical
C
CD
de
-.056
20
• 074
20
.0028
.0037
deflection
d at that point .
D
DE -.056
20 .0028 eThus to find the def~ection of the truss end
ef .072
20 .0036 (JOint a) we find the bending moment at point a
E
EE -.048
20 .0024 f
of the imaginary beam of Fig. A7.59.
Elastic Weights of Web Members
Hence
deflection at a = ZMa (calling counterclock-
Member tiL rl P =M
rl
apply
at
r2 Q = M. apply
r2 at
wise positive). = (.01922 - .00312) 80 + .00248 x
joint joint 70 + .00333 x 60 + .00239 x 50 + .00468 x 40 +
aA -.031 8.95 -.00346 a 8.95 .00346 A
.00234 x 30 + .00543 x 20 - .00149 x 10 = 2.13"
Ab .080 8.95 -.00895 A 8.95 .00895 b upward. Due to symmetry of truss and loading of
bB -.063 8.95 -.00702 b 8.95 .00702 B
Be .075 8.95 -.00838 B 8.95 .00838 c
the truss we know the slope of the elastic curve
cC -.060 8.95 -.00670 c 8.95 .00670 C at the center line of the truss is horizontal or
Cd .070 8.95 .00781 C 8.95 .00781 d
dD -.065 8.95 -.00728 d 8.95 .00728 D zero. Thus to find the deflection of any point
De .076 8.95 -.00850 D 8.95 .00850 e with reference to Joint f we can make use of the
eE .095 12.0 .00792 e 12.0 -.00792 E
Ef .024 12.0 .0020 E 12.0 -.0020 f deflection principle of the moment area method.
Thus in Fig. A7.60 the vertical deflection of any
point for example joint a, relative to Joint f
Table A7.l3 gives the calculations for the mag- equals the moment of all elastic loaas between
nitude of the member elastic weightb. The signs a and f about a.
and also the joint locations for locating the
elastic loads are also given. Combining alge-
braically the elastic weights for each joint from
Table A7.l3 the beam elastic loading as shown in
Fig. A7.57 is obtained.
Let it first be required to determine the R = .00312
vertical deflection of Joint f relative to the (From Span ee') Fig. A7.59
truss support points at e and e' •
To determine the deflections of the truss ZMa = 2.077" (student should make calcula-
between the supports e a d e' it is only neces- tions). Previously the deflection of f with re-
sary to consider the elastic weight loading be- spect to e was found to be -.0586". Thus de-
tween these points. Fig. A7.58 shows the portion flection of g with respect to point e = 2.077 +
of the imaginary beam of Fig. A7.57 between these .0586 = 2.135" which checks value found above.
points. The deflection at f relative to line ee' Let it be required to find the deflection of
is equal to the bending moment at f for the por- joint c relative to a line connecting joints b
tion of the imaginary beam between joints e and and d. C'> "" <Xl M '" <Xl ~ M '" co
e' and simply supported at these points. ~I M C'> ~ M M co M ~ ~ ~
~ C'> "'o
......l C ' \ 1 C " ) C ' \ I - . : j 4 C ' \ 1 l , r , ) ' I " " " " l M
Hence deflection at f = - .00312 x 30 + g C'> g g g'l g.) g'l g g g g
.00232 x 15 = .0586" (upward since minus is up). 'IL _-----'·I'______~____'1'______'.L._,--_ _-L._ _.L·!_..J.·!_.....II_---'·!
Find the vertical deflection of the canti- if E e D d C eBb A a
1<1.. Fig. A7.60

A7.37
For this problem we need only to consider (7) For the truss in Fig. A7.66 calculate
the elastic loads between points band d as the deflection of joint C along the direction CEo
loads on a simple beam supported at band d E = 30,000,000 psi.
(See Fig. A7.61) The deflection at C with re-
spect to a line bd of the deflected truss
equals the bending moment at point c for the
loaded beam of Fig. A7.61.
hence
0c = .004743 x 20 - .00239 x 10 =.07 inches
'" 00 '<t'
C") '" C")
f 8i ~j ~j i b
A E
d Fig. A7.67 Fig. A7.68
C c B
Rd=.004743 Rb=.004667 (8) For the truss in Fig. A7.67, find the
vertical and horizontal displacement of joints C
Fig. A7.61 and D. Take area of all members carrying tension
A7.l8 Problems as 2 sq. in. each and those carrying compression
as 5 sq. in. each. E = 30,000,000 psi.
(9) For the truss in Fig. A7.68, determine
the horizontal displacement of points C and B.
E = 28,000,000 psi.
5000 5000
A +B tC D
20" ~OOO#
~t
T
30"
-.L G 4 ~ 4 tE
5060 5000 1-
Fig. A7.64
Fig. A7.69 Fig. A7.70
(1) Find vertical and horizontal deflec- (10) For the truss in Fig. A7.69, find the
tion of joint B for the structure in Fig. A7.62. vertical deflection of joint D. Depth of truss
Area of AB = 0.2 sq. in. and BC = 0.3. E = 180". Width of each panel is 180". The area of
10,000,000 psi. each truss member is indicated by the number on
(2) For the truss in Fig. A7.63, calculate each bar in the figure. E = 30,000,000 psi. Al-
the vertical deflection of joint C. Use AE so calculate the angular rotation of bar DE.
for each member equal to 2 x 107 • (11) For the truss in Fig. A7.70, calculate
(3) For the truss of Fig. A7.64 determine the vertical and horizontal displacement of
the horizontal deflection of joint E. Area of joints A and B. Assume the cross-sectional area
each truss member = 1 sq. in., E = 10,000,000 for members in tension as 1 sq. in. each and
psi. those in compression as 2 sq. in. E = 10,300,000
(4) Determine the vertical deflection of psi.
joint E of the truss in Fig. A7.64 (12) For the truss in Fig. A7.70 calculate
(5) Determine the deflection of joint D the angular rotation of member AB under the given
normal to a line joining joint CE of the truss truss loading.
in Fig. A7.64. 1000# 5000
100'~ t i A====;<:~:¢==~=E
A
10" B
t 1 10' tA
10' rIO'
o.... <=>
o
.... Fig. A7.71 Fig. A7.72
1°" C ~ (13) For the beam in Fig. A7.71 determine the

,~ deflection at points A and B using method of elas-
,bo'
600# E
tic weights. Also determine the slopes of the
elastic curve at these points. Take E = 1,000,000
Fig. A7.66
Fig. A7.65 psi and I = 1296 in. 4
(14) For the beam in Fig. A7.72 find the de-
(6) calculate the vertical displacement of flection at points A and E. Also the slope of the
joint C for the truss in Fig. A7.65 due to the elastic curve at point C. Assume EI equals to
load at joint B. Members a, b, c and h have
areas of 20 sq. in. each. Members d, e, f, g 5,000,000 Ip in. sq.
and i have areas of 2 sq. in. each. E =
30,000,000 psi.
500#
29#/in.
~~=~~==C:==~l~O,~,-t\E
Fig.A7.73 1000#
Reaction
Fig.A7.81 Fig.A7.82
(15) Fig. A7.73 illustrates the airloads on (23) In Fig. A7.81 find the vertical move-
a flap beams ABCDE. The flap beams is supported ment and the angular rotation of point A. Take
at Band D and a horn load of 500# is applied at EI = 12,000,000.
C. The beam is made from a 1"-.049 aluminum (24) Determine the vertical deflection of
alloy round tUbe. I = .01659 in4 ; E = 10,300,00 point A for the structure in Fig. A7.82. EI =
psi. Compute the deflection at points C and E 14,000,000.
and the slope of the elastic curve at point B. 100#
100# 100 100 100 100# 100
1"-3" 1""3" 1:-3 -r- 3 --l F 9" ± 9 " ± 9 " i 't1 it
.~:l:_=======O!;:C Fig. A7.83
t C D t B A ~ 20"
Fig. A7.74 Fig. A7.75

(16) For the beam of Fig. A7.74 determine V
the deflections at points C and D in terms of EI Fig. A7.84 L Drag
which is constant. Also determine slopes of the
elastic curve at these same points.
(17) For the cantilever beam of Fig. A7.75
determine the deflections and slopes of the
elastic curve at points A and B. Take EI as
constant. Express results in terms of EI. Front View Side View
100# 100 100 (25) The cantilever beam of Fig. A7.83 is

100# 400#
loaded normal to the plane of the paper by the
MA,,~ ~6" ±6"~~ two loads of 100# each as shown. Find the de-
;A C D ~B RB -EMA flection of point A normal to the plane of the
Fig. A7.76 Fig. A7.77 paper by the method of virtual work. The rec-
tangular moment of inertia for the tube is
(18) For the loaded beam in Fig. A7.76 de- 0.0277 in4 • E = 29,000,000.
termine the value of the fixed end moments MA (26) The cantilever landing gear strut in
and MB' EI is constant. Also find the deflec- Fig. A7.84 is subjected to the load of 500# in
tion at points C and D in terms of EI. the drag direction at point A and also a torsion-
(19) In Fig. A7.77 determine the magnitude al moment of 2000 in. lb. at A as shown. De-
of the fixed end moment MA and the simple sup- termine the displacement of point A in the drag
port RB' direction. The tube size for portion CB is 2"-
.083 and for portion BA, 2"-.065 round tUbe.
p Material is steel with E = 29,000,000 psi.
L1 p p
t
Fig. A7. 78 Fig.A7.79 Fig.A7.80
A&
(27) Using the matrix equation 2U = LqiJ
[ aij] {qj} compute the strain energy in the
truss of Fig. A7.63 (Problem 2). The member
flexibility coefficient for a member under
uniform axial load is L/AE (see Fig. A7.35a).
Ans. U 22.4 lb. in. =
(20) In Fig. A7.78 EI is constant throughout. (28) Using matriX equation (23) compute
Calculate the vertical deflection and the angu- the strain energy in the beam of Fig. A7.71.
lar rotation of point A. Note: the choice of generalized forces should
(21) For the curved beam in Fig. A7.79 find be made so as to permit computation of the
the vertical deflection and the angular rota- member flexibility coefficients by the equations
tion of point A. Take EI as constant. of p. A7.19. Ans. U = 3533 lb. in.
(22) For the loaded curved beam of Fig. (29) Re-solve the problem of example
A7.80, determine the vertical deflection and the problem 25 for a stepped cantilever beam whose
angular rotation of the point A. Take EI as I doubles at point "2" and doubles again at "3".
constant. (HeaViest section at built-in end.)
A7.39
(30) For the truss of Fig. A7 85 determine
T
the influence coefficient matrix relating
vertical deflections due to loads P 1 , Pa, P a
and P 4 applied as shown. Member areas are
shown on the figure.
B c t = .051"
A CD 1.0 I]J G = 3.85 x 106 psi 20"
1~ ~
10"
.-
I.
Ai = 7.5 x10' Ibs
30"
1 \AE = 'xlO'
lbs.
500#
Fig. A7.86
Fig. A7.85
Answer.
(34) Find the influence coefficients re-
lating deflections at points 1 and 2 of the
1
["67
44.95
44.95
106.56
44.95
99.9 ~.~
38.28
simply supported beam of Fig. A7 87. Use
matrix methods.
~mnJ = E 44.95 99.9 99.9 38.28
33.28 38.28 38.28 33.28 ~ 20" - -.... r.-
DD 1,
20" ./ 20"-l
(31) For the truss of problem (30) deter- t=.040"

mine which of the following two loading con- G=3.85 x 106 1
ditions produces the greatest deflection of (Typ.) ~-!--
point 4, (All loads in pounds).
(i) ~ ®J ~ AE=3xf061bS
AE=8 x 106 lbs
Condition No. P1 Pa P4
1 1000 500 800 400 Fig. A7.87
2 300 700 400 600

Ans.
(32) Determine the matrix of influence 7.10~

coefficients relating drag load (positive aft),
braking torque (positive nose up) and moment 12.3~
in the V-S plane (positive right wing dOwn) as
applied to the free end of the gear strut
assembly of Problem 26.
Note to student: It will be highly instructive to re-work
Answer. problems 33 and 34 using the alternate choice of generalized
forces in the stringers from those used in your first solution.
See p. A7. 22 for alternate generalized forces on a stringer.
-50.7
6.97 References for Chapters A7, A8.
o
TEXT BOOKS ON STRUCTURAL THEORY
"Advanced Mechanics of Materials", F. Seely and

(33) Find the deflection of the load J. O. Smith, 2nd Ed., John Wiley, New York.
applied to the cantilever panel of Fig. A7.86.
(Assume the web does not buckle). Use matrix "Advanced Strength of Materials", J. P. Den
notation. Ans. 6 =
7.94 X 10-a inches. Hartog, McGraw-Hill, N. Y.
"Theory of Elasticity", S. Timoshenko, McGraw-
Hill, N. Y.
A7.4{) DEFLECTIONS OF STRUCTURES
TEXT BOOKS WITH MATRIX APPLICATIONS Wehle, L. and Lansing, W., A Method for Reducing
the Analysis of Complex RedUn~tructures to
"Elementary Matrices", R. A. Frazer, W. J. a-Routine Procedure, Journ. of Aero. Sci.
Duncan and A. R. Collar, Cambridge University VoL 19, 1952.
Press. Langefors, B., Analysis of Elastic Structures by
"Introduction to the Study of Aircraft Vibration Matrix Transformation with Special Regard to
and Flutter", R. Scanlan and R. Rosenbaum, Semi Monocoque Structur~ Journ. of Aero. Sci.
Mac Millan, New York. 19, 1952.
Langefors, B., Matrix Methods for Redundant
Structures, Journ. of Aero. Scr:- VoL 20, 1953.
TECHNICAL PAPERS Falkenheiner, H. Systematic Analysis of Redundant
Elastic Structures by Means of Matrix Calculus,
Benscoter, S. U., "The Partitioning of Matrices Journ. of Aero. SCi~ 20, 1953.-----
!P Structural AnalySIS", Journ. of AppL Mechs.
VoL 15, 1948. Argyris, J. and Kelsey, S., Energy Theorems in
Structural Analysis, Aircraft Engineering,
Oct. 1954, et. seq.
Many problems involving calculation of deflections are encountered in the structural
design of a large modern airplane such as the Douglas DC-8.

CHAPTER As
STATICALLY INDETERMINATE STRUCTURES
ALFRED F. SCHMITT
A8. 00 Introduction. tics are individually useful in forming the bases

A statically indeterminate (redundant) for methods of solution.
problem is one in which the equations of static [[J There are more members in the structure
equilibrium are not sufficient to determine the than are required to support the applied loads.
internal stress distribution. Additional re- If n members may be removed (cut) While leaving
lationships between displacements must be a stable structure the original structure is said
written to permit a solution. to be "n-times redundant".
The "Theory of Elasticity" shows that all
structures are statically indeterminate when-- -COROLLARY-
analyzed in minute detail. The engineer how-
ever, is often able to make a number of as- In an n-times redundant structure the mag-
sumptions and coarse approximations which render nitude of the forces in n members may be assigned
the problem determinate. In addition, auxiliary arbitrarily while establishing stresses in eqUi-
aids are available such as the Engineering librium with the applied loads. ThUS, in Fig.
A8.1 (a singly redundant structure), the internal
Theory of Bending (W;)
and the constant-shear- force distribution of (a) is in eqUilibrium with
flow rules of thumb (q = T/2A) (see Chaps. A-5, the external loads for any and all values of X,
A-6 and A-13 through A-15). While these latter the force in member BD.
are certainly not laws of "statics", the en-
gineer employs them often enough so that prob-
@~ c~ B(8J.707XC
lems in which they are used to obtain stress
distributions are often thought of as being ~ q,"" +~_ ;0:# >< ><
"determinate". ~. X~.- gN b b
to: to:
It is frequently the case in aircraft
_A D -""'l"Ir------f',D A D
structural analysis that, in view of the re- 2000# • X 2000# . .707X.
quiremeLts for efficient design, one cannot ob- +2000# f2000# +2000# 2000#
tain a determinate problem without sacrificing (a) (b) (c)
necessary accuracy. The Theory of Elasticity
assures the existence of a sufficient number of
auxiliary conditions to permit a solution in Fig. A8.1
such cases.
Singly redundant stress distribution, (a) consisting of a stress
This chapter employs extensions of the
methods of Chapter A-7 to effect the solution in static equilibrium with the applied loads, (b), with one
of typical redundant problems. Special methods
of handling particular structural configurations zero-resultant stress distribution, (c), superposed.
are shown in later chapters.
A8.0 The Principle of Superposition.
The general principle of superposition Only the system (b) is actually required to
states that the resultant effect of a group of equilibrate the external loads (corresponding to
loadings or causes acting simultaneously is X = 0). Note that the system (c) has zero ex-
equal to the algebraic sum of the effects acting ternal resultant.
separately. The principle is restricted to the
condition that the resultant effect of the U[J Of all the possible stress (force) dis-
several loadings or causes varies as a linear tributions satisfying static equilibrium the
function. ThUS, the principle does not apply one correct solution is that one which results
when the member material is stressed above the ~kinematicallY possible strains (displace-
proportional limit or when the member stresses ments), i.e. retains continuity of the struc-
are dependent upon member deflections or de- ture.
formations, as, for example, the beam-column, a
member carrying bending and axial loads at the ThUS, for example, there are an infinite
same time. number of bending moment distributions satis-
fying static equilibrium in Fig. A8.l (d) since
AB.1 The Statically Indeterminate Problem. Mo can assume any value. Of these, only one
Several characteristics (and interpreta- will result in the zero deflection of the right
tions thereof) of the statically indeterminate hand beam tip necessary to maintain structural
problem may be pointed out. These characteris- continuity with the support at that point.
-A8.1
A8.2 STATICALLY INDETERMINATE STRUCTURES
the Theorem of Least Work. In words; "the rate

of change of :strain energy with respect to a
fixed redundant reaction is zero".
M A8. 2. 1 Determination of Redundant Reactions by
MO-I~2+PL/4 Least Work.
Example Problem A
By way of illustration, the problem posed
Fig. A8.1d by Fig. A8.2 was carried to completion. The
Singly redundant beam with root bending
bending moment was given by (x, y, s measured
from the left ends of the three beam divisions)
moment Me undetermined by statics.
M (500 + R ) x
x
-COROLLARY- (500 + Rx ) L
M 2 + (Rx -500) y
If n member loads have been assigned ar-
bitrarily while establishing equilibrium with
the external loads, relative movements of the M = Rx (L - :6)
elements will result, violating continuity at n
points. n zero-resultant stress (force) dis- Then
tributions may then be superposed to reduce the
relative motions to zero. The resulting stress
distribution is the correct one. U :: 1
2 fM dx
Er:: 1
2EI
2
f L/
2
2
0 (500+Rx )2X 1x
A8. 2 The Theorem of Least Work.
_1_fL[/~500+RX)L
A theorem extremely useful in the solution
of redundant problems may be obtained from + + (Rx - 500)y]2 dy
Castigliano's Theorem. Consider first the 2EI 2
problem of redundant reactions such as in a o
2~' EX"(L - g)"do

beam over three supports (Fig. A8.2). One of
the reactions cannot be obtained by statics.
+
EI = constant
r Differentiating under the integral sign, (see
b 1. p. A7.8)
t---
500 + R x
L --4--
500-2R
L ---t Rx
L/2
x
~ :: l (500 + Rx) x 2dx
Fig. A8.2
A singly redundant beam with one reaction given an ar-
aR x EI
Jo
bitrary value (R x )'
If the unknown reaction (say that on the far

right) is given a symbol, Rx , then the remaining
+ i, Jf1500 ; RxlL + (Rx - 500)Y](~ + Y) dy
reactions and the bending moments may be de-
termined from statics. The strain energy U may
then be written as a function of Rx , i.e.,
U = f (R x )' Next form
+ ~ r~x
EI
(L - :6) 2 ds
o
This is the deflection at Rx due to Rx • But

~
aRx
0:: (500 + Rx ) roL/~2
J
dx
this must also be zero, since the support is

rigid. Hence
au = 0 - - - - _ - (1) + (500 ; RxlL p~ + Y) dy +
r
aRx
r:(~
Eq. (1) is true for all redundant reactions
occuring at fixed supports. Because it corre-
sponds to the mathematical condition for the (Rx - 5001 + Y) dy + Rx L - Wdo
minimum of a function, eq. (1) is said to state
AB.3
Completion of the computation gave
Rx =- 1500
~ Ibs. =- 93.8 Ibs.,
the negative sign indicating that Rx was down.

au
aMJt = °=
I[ L M-+2PL-M]
ML + -"R
o
LX
3L
-!.
3L
dx
Example Problem B
Determine the redundant fixed end moments
for the beam of Fig. A8.2(a).
0= ML )~ -:L) dx
Fig. AB.2a
~ + 2PL ~)L
A doubly redundant beam with two reactions given + 3L - 0 X (1 - ;L) dx
two arbitrary values.
Solution: 2LY ML + PL - ~ 12~2

+MR -dy+ -dy
The redundant end moments were designated
as ML and MR for the left and right beam ends
J
o
3L 3L
0
3L
respectively and were taken positive as shown.

~ j
The moment equations for the two beam portions
(x from left end, y from right) were
M = ~ +
~ + 2PL - ML
3L
x O<x<L
o = ML
I LX
o
- dx +
~
+ 2PL - M LX 2
L
~
0
-
~
dx
°{y<2L
Then
u
1
= 2EI
Jf,~ML + r~+ 2PL
3L
- ML
X
J2
dx
Evaluating the integrals and solVing simultane-
ously gave
~ -- ~ PL
+-1. M + P -
L L -"R
T'Lj2 dy 9
2EI 2
3L ~ = - PL
9
Differentiating under the integral sign (see AB. 2. 2 Redundant Stresses by Least Work.
remarks on p. A7-8) The Theorem of Least Work may be applied to
the problem of determining redundant member
forces within a statically indeterminate structure.
au Thus, in an n-times redundant structure if the
aML redundant member forces are assigned symbols X,
Y, g, - - - etc., the values which these forces
must assume for continuity of the structure are
such that the displacements associated with these
forces (the discontinuities) must be zero.
Hence, by an argument parallel to that used for
redundant reactions, one writes,
au
ax == 0 30 \2 2
- 58.3 X) x dx
au o - - - - - - - - (2)
ay
1
+--
2EI
etc. AC
In words, "the rate of change of strain energy

with respect to the redundant forces is zero". 1
+--
Eqs. (2), like eq. (1), are statements of the 2EI
CD
Theorem of Least Work. They provide n equations
for the n-times redundant structure. The simul-
taneous solution of these equations yields the ObViously there was no need to consider the
desired solution of the problem. energy in CD as its loading did not depend upon
X and hence could not enter the problem. Dif-
Example Problem C ferentiating under the integral sign
The cantilever beam and cable system of
Fig. A8.3(a) is singly redundant. Find the au _ 58.3 X
member loadings by use of the Least Work ax - EA
AB
Theorem.
~ X) \5: 2 dx
A
30 (1000 -
58.3 EI BC 0
30
+ (~)2
58.3
_X_
EI AC
J0
y2 dy 0
Fig. A8.3
A singly redundant structure with one member force given an
arbitrary value (X).
58.3 X + 11,032 X + 6620 X
Solution: EA EI EI
AB BC AC
The tensile load in the cable was treated
as the redundant load and was given the symbol
X (Fig. A8.3(b)). The strain energies con-
sidered were those of flexure in portions AC,
CD and BC and that of tension in the cable AB. Putting
Energies due to axial forces in the beam port-
ions were considered negligible.
AAB 0.025 in 2
The bending moment in BC (origin at B) was
l\c = (1000 - 5~~3 XJ x
In AC, (origin at A): lAC == 10.0 in 4
50 gave
MAC == 58.3 X . Y
X 613 lbs.
In CD:
Then
MCD == 50,000
l\c == [1000 - 5~~3 ( 613 )J x = 685x
The strain energy was therefore
50
X L)
U = 2"" AE AB
2
( MAC == - -
58.3 613 • Y 526 Y
A8.5
Example Problem D Differentiating under the integral sign

A semicircular pin-ended, uniform ring is
supported and loaded as shown in Fig. A8.3(c). 600
As a first approximation the horizontal floor
tie is to be assumed rigid axially. Find the
au
ax = EI
R
3
X 1
a sin " Q d Q
bending moment distribution in the ring.
3
600
I
Rf p
[ 1 (1 - cos Q) sin Q d Q
a
(c) (d)
Fig. A8.3
+ PR J60cos Q sin Q d Q
AE
Solution:
a
The axial load in the floor was taken to
be the redundant (since the floor was assumed 90 0
rigid, this could have been thought of as a re- " Qd Q+ ~ sin" Q d Q
dundant floor reaction from fixed supports).
The loading is shown in Fig. A8.3(d). J60 0
The bending moment distribution was Evaluating,

M = X R Sin Q - PR (1 - cos Q) 0(Q(60 0
The axial loadings were
s P cos ~ + X sin ~
Therefore
S X sin Q
The strain energy (for only half the 3P

X
structure) was * 2n
U " ~, tE60 0
R sin 0 - PR (I - cos OIJ RdO
Example Problem E
The portal frame
of Fig. A8.3(e) is
I [X 0
three times redundant.
90 Set up the simultaneous
+ 2ir
60°
R sin ~- PR/2 T RdQ
equations in the re-
dundant forces. The
relative bending stiff-
nesses of the segments
are given on the figure.
+ X sin Q J RdQ Solution:
Fig. A8.3e
The redundant forces

selected were the bending-moment, the transverse
shear force and the axial force, all at point A.
The four figures A8.3(f) through A8.3(i) show
the bending moment diagrams of the structure due
to applied loads and due to redundant forces
* Zero strain energy in the rigid floor (AE ~ -).
AB.6 STATICALLY INDETERMINATE STRUCTURES
acting individually (it being easier to com-

pute the loadings in this fashion). The com-
plete loading was obtained by superposition.
50
+ (lOOOs' + M - Vs' + 50 V - 100 T)(50 - s') d '
2.25 s
Jo
50
2!1 = 0 = + O:ooos' + M - Vs' + 50 V - lOar]( 100) d '
aT 2.25 - s
Jo
-+-
1,000# + J~-50,000 sin Q + M + 50 r~;o+ sin Q) -5ar(1 - cos Il)](cos II _ l)bOodQ
S
t .......,
~
Fig. AB.3f 50,000 M Fig. AB.3g After evaluation of the integrals the equations
obtained were
.14-92M + 9.682V 7.4-59T = 2.778 :x' 10"
9.682M + 763.1V 4-84-.OT 288.3 x 10"
{
-7.4-59M - 4-84-.0V + 614-.9T = -111.15 x 10"
AB.3 Redundant Problems by The Methods of Dummy-

Unit Loads.
50V While the Theorem of Least Work may be made
the basis of redundant problem analysis, its
V
.. Fig. AB.3h
1
~T
S
Fig. AB.3i
~
- lOOT
direct application by the calculus, as in Art.
A8-2.1 and A8-2.2, is often impractical. For the
majority of problems the work is facilitated if
carried out by the techniques of the Method of
Dummy-Unit Loads.
The following derivation is for a doubly
redundant truss structure. The extension to a
more general n-times redundant structure, in
The composite bending moments as functions of which other loadings in addition to axial (flex-
s, g and s' were ure, torsion and shear) are present, is indicated
later.
MAE M + Vs Consider the doubly redundant truss of Fig.
A8.4-(a). It may be made statically determinate
~C 50,000 sin g + M + 50 V + 50V sin g by "cutting" two members such as the diagonals
indicated. Application of the external loads to
- 50T (1 - cos g) this determinate ("cut") structure gives a load
distribution, "S", computed by satisfying static
MCD = 1000 s' + M - Vs' + 50 V - 100 T eqUilibrium. At this time discontinuities appear
at the cuts "x" and "y" due to the strains
Then since developed.
= 12 JM"ds au au au
~f
u ~ and aM = av = aT = 0,
~
one has,
50
au =0 _ . (M + Vs) d
aM - 2.50 s S loads Ux loads Uy loads
1 o Fig. AB.4a Fig. AB.4b Fig. AB.4c
+ I o
[n_50 ,ooo sin Q + M + 50V(1 + sin Q) - 5ar(1 - cos II)] 5Od1l
1.50
To compute these and subsequent displacements
the Method of Dummy-Unit Loads may be used (Art.
50
A7-7). For this purpose virtual loads are placed
+ (lOOOs' + M - Vs' + 50 V - 100 T) d ' alternately at the x and y cuts as in Figs. A8.4-(b)
J
o 2.25 s
and (c). From the dummy-unit load equations
50
!l!!:=0= (M+Vs)sds
av 2.25
J
0
A8.7
~L} (3) For continuity these net relative displacements

ox. " , must be zero. Equating the above expressions
each to zero, and rearranging, gives the simul-
o ;: 1: SUyL taneous equations
yO AE
X Z ux"L + Y Z UxuyL
AE AE
" _, SUXL}
AE
ux and uy are the unit-redundant stress dis-
tributions as indicated in Figs. A8-4b, c. The
- - - - - (4)
L u "L Z SuyL
subscript "0" indicates these relative displace- X Z Uyux + Y Z Y =:
AE ---xE AE
ments occur in the determinate ("cut") struc-
ture with the "original" stress distribution. Eqs. (4) are two simultaneous equations in the
It is now desired to close up the dis- two unknowns X and Y. Upon solution for X and Y
continuities by application of redundant loads the true stress distribution may be computed as
X and Y to the x and y cuts, respectively, as
in Fig. A8-4(d). Load X (5)
causes a stress distri-
bution Xu and, likewise,
x
f\Pl\t1 For a structure which is only singly re-
~
Y causes a distribution dundant, eqs. (4) and (5) are applied by setting
YUy • The relative dis- Y=:o giving
placement at cut x due to
redundant load X is given X Z ux"L =: _ Z SUxL
by (0 is read "displace- Fig. A8.4d AE AE
xx
ment at x due to X"). or, simply,
2
6 =: Z Xux • uxL X Z ux L
xx AE AE
X - - - - - - - - - - - - - - - (4a)
and at cut.y by (read 6 as "displacement at
y due to X"). yx
6 Z Xux • UyL and,

yx AE
- - - - - - - - - - - - - (5a)
Similarly the load Y causes displacements at
the cuts y and x given respectively by
A8.4 Example Problems - Trusses With Single Redundancy.
6
yy
=: Z YU y • UyL =: Y Z Uy"L Example Problem #1
AE "AE
19:I~~~~
and
YUy • LlxL =:
Y Z uyuxL
6
xy Z
AE
Now the net relative displacement at each

AE
!a 10i.- " ~ bl QQ9 707 ;
tlOOO# 1000#
cut under the simultaneous action of the three
stress systems S, X·u and Y·u is Fig. A8. 5 Fig. A8.6 Fig. A8.7
x Y
Fig. A8.5 shows a single bay pin connected
6 + 6 + 6 =: Z SUxL + X Z uxuxL truss. The truss is statically determinate With
xo xx xy AE AE respect to external reactions, but staticaily
indeterminate With respect to internal member
loads, since at any joint there are 3 unknowns
With only two equations of statics available for
a concurrent force system. The truss is there-
and fore redundant to the first degree. The general
6 + 6 + 6 Z SuyL + X Z uxuyL procedure for solution is to make the truss
~ yx YY AE AE statically determinate by cutting one of the
membersjon Fig. A8.6, member bc has been selec-
ted as the redundant member, and it is cut as
+ Y Z UyUyL shown. The member stresses S for the truss of
AE
AB B STATICALLY INDETERMINATE STRUCTURES
\1//
Fig. A8.6 are then determined, the results being ABC
~~<r t~
recorded on the members and also entered in
Table A8.1. In Fig. A8.7, a unit 1# tensile /
dummy load has been applied at the cut section
of the redundant member bc, and the loads in all ~l#
the members due to this unit load are calculated o o
l1000
The results are recorded on the figure and also
in Table A8.1 under the head of u stresses. The Fig. AB.9 Fig. AB.I0
solution for the redundant load X in the redund- S loads u loads
ant member bc is given at the bottom of Table
A8.1. The true load in any member equals the S Example Problem # I-A; Deflection Calculation in
stress plUS X times its u stress. a Redundant Truss
T.ab1e A8.1 Calculations of the deflections under load of a redundan
structure are made by application of the methods of Chapter
lIember L A S u SuL u 2L True Stress A-7. Since, however, there are certain pitfalls as regards
-r A = S + Xu symbols and also some important special techniques, the
ab 30 1 0 -.707 0 15 395 following examples are given at this time. The extension of
bd 30 1 -1000 -.707 21210 15 -605 the method to more complex structures is immediate and no
de 30 1 0 -.707 0 15 395
ea 30 1 0 -.707 0 15 395
further work on deflections of redundant structures is given
eb 42.4 2 0 1.0 0 21.2 -559 in this chapter (excepting in the case of the matrix methods of
ad 42.4 1.5 1414 1.0 40000 28.3 855 Arts. AB. 10 et. seq.).
Yo 61210 109.5
X • true load in redundant member be Find the horizontal movement of point "d"
Su1 of Example Problem # 1 under the action of the
_l: A
X =
-2 • -61210
109.5
= -559# load applied there.
u L
Z A
Solution: The equation used to find the
deflection is Eq. (18) of Chapter A-7. Written
Example Problem # 2 for application tv truss deflections it is
(see Example Problem 13, p. A7.11)
Fig. A8.8 shows a singly redundant
3-member frame. Find the member loadings.
- - - -(A)
Member areas are shown on the figure.
T~O ': yC Now for a deflection calculation the symbols

"s" and "u" must be carefully reinterpreted
3'~/.4
from their meanings in the redundant stress
calculation. For a deflection calculation the
symbols of eq. (A), above, mean: "S-loads" are
o the true loads of the redundant structure due
hooo# to application of the real external loading;
Fig. AB. B
"u-loads" are the loads due to a dummy-unit
(Virtual) load applied at the external point
where the deflection is desired and in the di-
Solution: Member OC was selected as the rection of the desired deflection.
redundant and was cut in figuring the S-loads,
as in Fig. A8.9. Fig. A8.10 shows the u-load ThUS, the "s loads" for use in Eq. (A) are
calculation. The table completes the calcu- the true stresses (the solution) of Example
lation. Problem # 1.
SuL u 2L True Load The "u-loads" represent additional informa

Mem. L A S u ~ A =S"Xu tion which WOUld, in general, appear to necessi
tate another redundant stress calculation. As
AO 141. 4 0.2 0 1. 224 0 1059 335.5 will be seen, such is fortunately not the case.
In the present problem the dummy-unit load is
BO 100.0 0.2 1000 -1. 366 -6.B3xl0 5 933 625.6
applied identically as is the 1000# real load
CO 200.0 0.4 0 1. 000 0 500 274.1 and hence the u-loads are simply equal to the
"S-loads" properly scaled down. The following
L: -6. B3x10 5 2492 table completes the calculation.
SuLL:
A
X - - u 2
L - 274.1 lb.
L:-
A
A89
Mem. L A s* u** SuL/A SuL

Mem. L A S* u ~
ab 30 1 395 .395 4,680
AO 141. 4 .2 335.5 V2 3. 355x10 5
bd 30 1 -605 -.605 10,980
BO 100 .2 625.6 -1 -3. 128x10 5
de 30 1 395 .395 4,680
CO 200 .4 274.1 0 0
ea 30 1 395 .395 4,680
0.227x10 5
eb 42.4 2 -559 -.559 6,620
ad 42.4 1.5 855 .855 20,660

:. 6 = 22,700
Z 52,300 E
* true loads from Example Problem #2.

:. 6 = 52,300
E By way of ctemonstration another set of
u-loads, called u ' , were found for this same
* identical with the "true stress" of Table problem, this time by cutting member OA. The
A8.L corresponding calculations follow:
** simply l/lOOOth of the "S-loads" since the Su'L

dummy-unit load is applied exactly as is S u' p;-
the 1000# real load.
335.5 0 0
Example Problem #2-A.
625.6 .578 1. 808x10 5
Find the horizontal deflection of point 0 274.1 -1. 155 -1. 583x10 5
of Example Problem #2 under application of the
vertical 1000# load shown in Fig. A8.8. Z 0.225x10 5
Solution: To compute the deflection use

The results are identical (allowing for
6 = Z SuL
AE
round-off errors).
Proof of Theorem
==
Again the symbols "S" and "un are to be re-
interpreted for a deflection calculation as To prove the theorem above we return to th e
explained above in Example Problem #l-A. The virtual work principle and the argument from
"S-loads" are now the "true loads" computed in which the dummy-unit loads deflection equation,
Example Problem #2, above. The "u-loads" are Eq. (18) of Chapter A-7, was derived (refer to
loads due to placing a dummy-unit load acting p. A7 .10). It will be remembered that the de-
horizontally on the structure at point O. flection was shown to be equal to the work done
Since this load acts on a redundant structure by the internal virtual loads (u-loads) moving
it would appear that another redundant stress through the distortions (n) due to the real
calculation is required. However, this is not loads, Le., 6 = Z un. The internal virtual
necessary. loads are those loads due to a unit load acting
at the point of desired deflection.
Theorem: For the u-loads in a deflection
calculation any set-of stresses~loads) in Now for the statically indeterminate
static equilibrium-with the dummy-unit load structure these internal virtual loads (u-loads
may be used, even from thElsimplest -
of -
"cut" are, in general, indeterminate since the dummy-
structures. - - - - - - - unit load is applied at an external point of t he
structure. However, we recall that,
This theorem says that to get the "u-loads' i-any stress distribution in static
for this deflection calculation we may "cut" equlli bri um with the "applied load" (for the
anyone of the three members and get a satis- moment now we are thinking of the dummy-unit
factory set of u-loads by simple staticsl load as the "applied load") differs from the
Before proving the theorem we complete the correct (true) distribution only by a stress
calculation in tabular form as shown. The distribution having zero external resultant
"u-loads" were obtained by cutting member OC (p. A8.1).
and applying a unit load horizontally at O.
ii - a zero-resultant stress distribution
moving through a set of displacements does zer o
work.
Mathematically expressed these points are;
i . ~RUE =u STATIC + ~=O
where uSTATIC is a u-load distribution obtained

from statics in a simple "cut" structure under
the action of the externally applied dummy-unit
load and ~=O is the zero-resultant u-load Fig. A8.I2 Fig. A8.I3
U
x loads Uy loads
system which must be superposed to give the
true u-load distribution
We note here the rule by which the degree of redundancy
of a planar pinjointed truss can be determined. For a truss of
ii ~ A . uR=O = 0 m members with p joints, the truss is n times redundant where
n = m - (2p - 3). For a spatial truss (3 dimensional truss)
It follows, therefore, that any set of n = m - (3p - 6).
u-loads in static equilibrium with the
externally applied dummy-unit load will do the In the present problem n = 11 - (12 - 3) = 2.
same amount of virtual work when the structure
undergoes its distortion as would a "true" set The calculation is carried out in tabular
of u-loads computed by an indeterminate stress form in Table A8.2. The member dimensions are
calculation. That is, given in the table.
6 = ~ A
~RUE = Table A8. 2
SUxL SuyL U;L u~L uxuyL True Load =

Member L A S Ux Uy -A- S+Xux+Yu
----;;:- ----;;:- A A y
AB 30 .5 750 -.6 o -27,000 0 21. 6 0 0 1680

~ A ~TATIC + ~ A . uR=O =
BC 30 .25 0 0 -.6 0 0 0 43.2 0 560
CD 40.5 -1000 0 -.8 0 64,000 0 51. 2 0 -255

~ A . ~TATIC BD 50 .25 1250 0 1 0 250,000 0 200 0 318
CE 50 .5 0 0 1 0 0 0 100 0 -932
Q,.E.D.
BE 40.7 -3000 -.8 -.8 137,000 137,000 36.6 36.6 36.6 -1015
A8. 5 Trusses With Double Redundancy ED 30 .5 -750 0 -.6 0 27,000 0 21. 6 0 -190
BF 50.8 0 1 0 0 0 62.5 0 0 -1550

Trusses With double redundancy are
AE 50 .5 3750 1 0375,000 0 100 0 0 2200
handled directly by Eqs. (4). By way of
illustration, the structure of Fig. A8.4, from EF 30.7 -3000 -.6 0 77,200 0 15.4 0 0 -2070
which Eqs. (4) were derived, will be solved AF 40.25 0 -.8 0 0 0 105.0 0 0 1240
for a loading P:I. = 2000# and P 2 = 1000#. Z 562,200 478,000 341.1 452.6 36.6
Choices of redundants were made identical with
those of Fig. A8.4a. Figs. A8.11, A8.12 and
A8.13 show the S, u and u load calculations SUbstituting from the table into Eqs. (4)
respectively. x y gives (common factor of E divided out)
X ~ u~L + Y ~ UxuyL = ~ SUxL

A A --P:-
{ X ~ UX~yL + Y ~ uiL = _ ~ S'1L
341.1 X + 36.6 Y 562,000

{ 36.6 X + 452.6 Y 478,000
Solving,
X - - 1550#
Y--932##
Fig. A8.11
Finally (see Table A8.2)
S loads True Load =S + XUx + YUy.
AB.ll
AB.5 Trusses With Double Redundancy, cont'd. SolVing the two equations for X and Y, one ob-
tains X = 521# and Y = 416#. The true load in
Example Problem 3 any member = S + Xu + Yu which gave the values
x y
Fig. A8.14 shows a structure composed of in the last column of the table.
four co-planar members supporting a 2000# load.
With only two equations of statics available for AB. 6 TrusSi!S With Multiple Redundancy.
the concurrent force system the structure, rel- By induction, eqs. (4) may be extended for
ative to loads in the members, is redundant to application to trusses which are three or more
the second degree. times redundant. Thus for a triple redundancy,
Solution:
u 2L UxuyL UXugL
XZ_x_+YZ--+;?,Z
Fig. A8.15 shows the assumed statically AE AE ~
determinate structure; the two members CE and
DE were taken as the redundants and were cut at UxUyL u ~L uyugL SUyL
points x and y as shown. The member stresses XZ + Y Z -y-- + ;?, Z =- Z -(6)
for this structure and loading are recorded on
AE AE AE AE
the members. Figs. A8.16 and A8.17 give the
ux and uy member stresses due to unit (1#) uxUgL uyugL Ug 2L SUgL
XZ + Y Z - - - + ;?, Z - L:
tensile loads applied at the cut faces x and y. Ali: AE AE AE
Table A8.3 gives the complete calculations for
eqs. (4) and (5). The load in the redundant and after solving for X, Y, ;?',
member CE was designated X and that in DE as Y.
True stresses =S + Xu + Yu + ;?,U - - - - -(7)
r;j +t--3,,2',~ T x Y 13
\ ~L~~./Y
\JL:
AB.7 Redundant Structures With Members Subjected to
6' Loadings in Addition to Axial Forces.
O~/Y Eqs. (6) are extended readily to cover
E~ problems in which flexural, torSional, and shear
t 2000# '2000# loadings occur. ThUS, for a three times redundant
Fig. A8.14 Fig. AB.15 structure
Xaxx + Yaxy + ;?,aXg 6
~
"" jil' xo
...;..-:: 1#
~I Xa + Ya +;!,a 6
yx yy YI3 yo - - - - - - - - -(8)
Ux loads u y loads
Xa gX + Ya gy + ;?,a 1313 6
130
Fig. AB.16 Fig. AB.17
Ux 2L uxuyL SUxL Where

X Z -A-+ Y Z -A- == - Z -A-
Substituting values from table a xx = L: U~; L + Jm~~ J +
2446 X + 2350 Y = 2,253,000.
X ~ uxuyL + Y ~ Uy2L _ SUyL

+ II q;: dy
A -A- - - Z -A-
L, L,
a
xy
= a
yx
= Z uxuyL +
AE
Jmxmydx
EI
Substituting:
Jtxty<L'<: + If qxqy dxdy

2350 X + 3039 Y = 2,488,000.
+
GJ Gt
TABLE A8.3
.
Member .'"i! L S
0
Ux uy
.806 1.154
SUx L X 10- 6 S'i L X 10- 6 U 2L
--,;:-
0 0
~
280.5
~
576.5
uXi('L True Load =
402
S+Xux+Yuy
900
a
YY
= L:
2L +
~
U
AE
J 2
my dx +
EI
- - - - - etc •
AE .286.5
BE .172 2000 -1. 564 -1. 729 -2.253 -2.488 1763 2150 1948 465
CE .280.5 0 1. 00 0 0 0 402.5 0 0 521
DE .393.7 0 0 1. 00 0 0 0 312.5 0 416
Total L: -2.253 -2.488 2446 3039 2350

and
Solution:
J J
~
Su L +
_x_ Mm
_x_ dx + Tt
_x_dx
axo AE EI GJ
The shear flow in the sheet panels was
chosen as redundant. Because of symmetry the
problem was only singly redundant. Fig. A8.19
qqxdxdy shows the Ux and qx loadings due to the redund-
Gt ant shear flow X = 1. The real loading in the
determinate structure consisted of a constant
load P in the central stringer alone. The
Byo ~ 2 sUi + J EI
Mmydx + J equation solved was (ref. eqs. (8)).
~ AE
-
qqy dx dy
x (J ux: II ~ "G:Xd}dx +
Gt
-0 SU~EdX + Jj q \~dY)
etc. where
and where S = P = constant, in central stringer

REAL S = 0 in side stringers
S, M, T, q are the real loads in the de- LOADS {
terminate structure; q =0
ux ' mx , t x , qx are the unit (virtual) load- Ux =L - x in side stringers

ings due to a unit load at cut x; VIRTUAL = 2(x-L) in central stringers
LOADS
{ ~ l.0
u y ' my' t y ' qy are for a unit load at cut
y; etc. When evaluated, (note that the double inte-
grals simply reduce to a constant times the
The redundant force(s) need not be an axial panel area)
~ ~; + ~tL) (-~LE
2
force but may be a moment, torque etc.
After solution for the redundants, True X =_ )
Axial Forces =
1bAE )
+ GtL2
True Bending Moments - - - - - - - (9)
Therefore the true stresses were
1
PROOT = P-2LX = P L2 Gt in central stringer
1 + AEb
etc. P 1
= LX = - in side stringers
2 1 + AEb
2
Example Problem 4 L Gt
The symmetric sheet stringer panel of Fig.
A8.18 is to be analyzed for distribution of load Example Problem 5
P between stringers. As a first approximation, The problem of Fig. A8.19 is doubly re-
assume constant shear flow in the sheet panels. dundant as shown. Determine the bending moment
All stringers have the same area. distribution. Both members have equal sectional
L properties.
///////1
t A~
1
AE}
A= x= 0
o L~EI/;}::'bnt '~S'2~
const. L
~~ D .;:;,S~=-~2;:;P::;::~~~:;:;::;:~;i
~LL.--...1L---U1 [1 l1 fJ tJ "Clamped" f-- L ..j.. L i P
/""3" -
"Pinnedj' M~
ry Y ~~;?1- Pt
b-tb
I-- ~
Fig. A8.19 Fig. A8.20
S, M loadings
P ux, iix loadings
x Fig.A8.18
Fig. A8.19
A8.13
Solution: (Lv'2 + L ) (
3
L (1 + v'2!L \
3
X '\ AE 6EI + Y V2AE + 3EI )

The bending moments at point C in member
CBD and at point B in member 4B were selected
as redundants yielding (when cut) the pin- = _ PL (2L,j2__l:_~.\
jointed determinate structure of Fig. A8.20. \" AE 3EIj
The virtual loadings were as shown in Figs.
A8.21 and A8.22. For a specific case it was assumed that
AE
L
= 100 EI
17 giving
A
.3716 X + .1526 Y - .09011 PL

{ .1526 X + .8121 Y = .3616 PL
X=l"~ ux=-I/L
\C B
X - .0645 PL
~
I-- mx-r;-
y
{ Y: .456 PL
Ux, m x loadings uy ' my loadings
Fig. A8. 21 Fig. A8. 22 Then as usual
Note that in Fig. A8.22 the unit redundant True stresses = S + X ux + Y uy
loading was applied as a self-equilibrating set
of unit couples. The real and virtual load- True Moments =M+ Xmx + Ymy
ings were as follows: (member portion BD,
having no virtual loadings, was omitted. It A8.8 Initial stresses.
could not enter the calculation.) In a redundant structure initial stresses
are developed if, upon assembly, certain mem-
Member S M Ux rn.x: uy my bers must be forced into place because of lack
of fit. In some situations intentional misfits
AB 2N2 0 V2/L 0 -l/LV2 ;?'/Lv'2 are employed to obtain more favorable stress
y-L distributions under load ("prestressing").
BC -2P Py - l/L 0 J/L If, in Fig. A8.4(a), the redundant member
L
with the "x cut" was initially oversize (too
long) an amount a (an oversize, corresponding
Xi
The equations which were solved were (Ref. to a distortion in the positive X direction, is
eq. (8)). a positive a )' the modified condition for
Xi
x(~U~:L J\,m} y0U~L JffiX:t)

+ +
continuity at the x cut would be (compare With
the equations just preceding eqs. (4)).
= _(~ s:t + I"':,dX) axO + aXl. +
Similarly if the Y redundant member were too

long
aXX + axy a 0
X (~Ux~t J~:t) y~Ur:L J

+ + ayO + ayi + ayx + ayy o
=- ~S:t J"'l0 +
Then using the previous notations, the appropri-
ate equations for the redundant forces are
< x 2L u L
~
u
After evaluation of the integrals and mUlti- X U + Y
plying through by L2 these become 2JL =
~ AE AE
- -(10)
3
~Lv'2
~
x (( 1 + 2v'2) .!:... + L \ + Y + L3 \ u u L ~u 2L ~ Su L
AE 3Ei) AE 6Ei} X 2..JL+Y .lL-=_ -y--a
AE AE AE yi
= _ PL ((2 + 4V2) L _ ~)
\" AE 6EI The ItS loads" of eq. (10) are present because
of applied external loads. These mayor may not
be zero depending upon the problem.
Example Problem 6 If, as is sometimes the case, the number of

If in example problem 3 member CE was 0.01 misalignments exceeds the number of redundancies,
inches too short before assembly, determine the or if the misalignment does not coincide with
stress distribution after assembly and load the redundant cut chosen but occurs elsewhere,
application. one may use the virtual work principle to com-
pute the effect of these misalignments on the
Solution: redundant cuts proper. Thus, referring to the
"Virtual work" derivation of the Dummy-Unit
Data obtained from the previous problem was load equations, {Chap. A7) one has
substituted into eqs. (10) along with
- -(ll)
::: _.01" (negative because "too short")
°Xi
::: 0 where 0Xi is the initial misalignment in the
°Yi determinate structure at the X redundancy-cut
due to initial imperfections (equivalent to
to give
initial strains) ~i throughout the structure.
{2446 X + 2350 Y 2.253 X 10 6 + .01E u as before, i& the unit loading due to a
:::
x
2350 X + 3039 Y ::: 2.488 X 10 6
virtual load at cut x.Eq. (11) and similar ex-
pressions for the Y, g, etc. cuts may be in-
With E 29 x 10 6 the redundant forces were serted in eqs. (10).
X ::: 985 Ibs. Example Problem 8
Y::: 57 Ibs. Referring back to example problems num-
bers 3 and 6, assume that member BE is .025"
Then, as usual, too long. Determine the initial stresses if
the other membeFs are of proper length an~ no
True stresses S + Xu + Yu • external load is applied.
x y
Solution:
Example Problem 7
Assume that in the structure of example To employ the same equations as those of
problem 5 an angular misalignment occurred be- example problem 3, the initial imperfections
tween members AS and CBD at joint B such that occurring at the same x and y cuts used there
the end of member AB had to be rotated 2.7 0 were computed, in this case due to the initial
clockwise to fit upon assembly. Determine the elongation of BE. Thus
moments developed without external loads
applied.
Solution:
The initial imperfection was 0Yi :::

Then, use of those previously computed coeffi-
- 2.7/57 •3 ::: - .0471 radians cients in eqs. (10) gave,
The sign was determined by noting that the 2446 X + 2350 Y ::: .0391 E
original misalignment was in the negative di-
rection of the redundant couple Y. 2350 X + 3039 Y::: .0432 E
The equations used from the previous
problem were (noting the equations there had With E = 29 X 10 6
psi
been multiplied by L2 x EI/L 3 ::: EI/L).
X ::: 263 Ibs.
.3716 X + •1.526 Y ::: 0 Y ::: 209 Ibs •
{ .1526 X + .8121 Y::: .0471 E

C F'tnally,
Solving, True Initial Stresses::: S + Xu + Yu

x y
X ::: - .0258 EI/L A8.9 Thermal Stresses.
Y::: .0630 EI/L stresses induced in redundant structures by
thermal strains may be computed by application of
True initial stresses and moments were deter- methods presented above. The problem may be ap-
mined as usual. proached from the point of view of computing the
A8.15
relative motions at the cuts of the determinate Example Problem 9

structure caused by the thermal strains and then The end upright of the truss of Fig. A8.23
restoring continuity by applying redundant mem- is heated to the temperature distribution shown.
ber forces to the cuts. Determine the stresses and reactions developed.
Specifically, consider a doubly redundant
truss such as that of Fig. AS.4(a). After
making cuts "x" and "y" to render the structure
A B :TT~
.T~X~;~X
®
determinate, the application of the temperature
distribution is visualized. Relative displace- 30" x
ments occur at the cuts,denoted by 0xT and 0yT'
~ <I
.I e
These displacements may be computed by the
Dummy-Unit Load method as shown in Art. A7.S of
Chap. A7. After this calculation is accomp-
r-- 40" --, (permits
sliding
i
yy
Fig. A8.23 Fig. A8.24

lished, the problem proceeds as for initial vertically)
strains, Art. AS.S. Thus, the continuity con-
dition at the cut gives (compare with the equa- Solution:
tions immediately preceding eqs. (4) and as-
sume for simplicity that the external loads are The structure was made determinate by cuts
absent, making 0xO = DyO = 0) x and y as in Fig. AS.24. The unit loadings are
shown in Fig. A8.25.
0
xT+ 0xx + 0xy = 0 }
ll
0
yT+ 0yx + 0yy = 0
In a truss, the thermal strains produce

relative displacements at the cuts given by the
"virtual work" derivation of the Dummy-Unit
.6
~ 12J
Ux
.8
loading
1#.6
Fig. A8.25
Load equations (ref. Chap. A7, Arts. A7.7 and
A7.S) as
oyT = J uy ::: } - - - - - - - - - - - (12)
where a is the material thermal coefficient of

The thermal coefficient a was assumed
constant. The calculation was set up in tabu-
lar form.
TABLE A8.4
expansion, T is the temperature above the

ambient temperature and ux and u y are the unit
AB
Ie
00
30
'0
20
• .8
• .8 ....
0
0
.853 0
~T
0
-gaT -1.21
.. ....
CD 00 '0 • .8 0 .853 _1.61
load distributions due to virtual loads at the DA '0 20 - 1.0 1.50 •. 80 0.0
x and y cuts, respectively. The sums in eqs. AC .0 O. 1.0 0 1.111 2,01
(12) are written as integrals rather than finite DB '0 O. 1.0 0 1.111 2.01
sums to allow for possible variation in a and T, 5.008 x la-- 1.50 x 10- -.90 J: 10-· - 9 a T
along the members as well as from member to

member. Then the final equations for thermal
stresses in a dOUbly redundant truss become SUbstituting into eqs. (13)
5.008 X - .90 Y = 9 a T x 10 6
~ Ux 2L
X + Y ~ u.:!:....L
u L = u a T dx { _ .90 X + 1.50 Y o
AE AE -J x
- - (13) Solving,
~ u~+y ~ 2L
X
u L
AE
u-y-=
AE -J uy
a T dx X = 2.01 a T x 10 6
Y = 1.21 a T x 10 6
Equations (13) may, of course, be extended
for application to structures other than True stresses are given in Table AS.4.
trusses. The expressions appropriate to other
loadings have been developed previously (eqs. Example Problem 10
(8) et seq. in this chapter and other equations The upper surface of the built-in beam of
in Art. A7 •8 ) • Fig. A8.26 is heated to a uniform temperature T.
Through the depth of the beam the temperature Then the equations corresponding to eqs. (13)
varies linearly to normal (T = 0) at the lower were written (see also eqs. (8)).
surface. Determine the end moments developed,
neglecting axial constraint and influence of
(J ux 2ds + JmxEI2dS) + (J ~ + mxmYdS)
axial forces.
r- L T -I
X AE Y AE J EI
J
~1
X=1"# X=1"# + g uxu g ds + mxmg dS) =
, :---r , 1'-- 1
---J J (J AE EI
(J
T
u ds + J ~m J-y--
EI constant Virtual Loading
m 2dS)
Fig. AB.26a Fig. AB.26b
X
( Ju~
AE
--y-
EI
dS) + Y u 2ds +
-y--
EI EI
Solution:
J myf!Jm dS)
The problem was only singly redundant be-
cause of symmetry and was made deter.minate by
+g
(J u~
u ds
AE
+
EI
=-0
yT
cutting the end bending restraints. Applica- J m dS) +Y (J u yf!J

u ds + J my m dS)
( Ju~+
u ds m g
tion of unit couples (Fig. A8.26b) gave m = 1 = X ~
AE EI AE EI
const. Then (see Example Problem 24, Art. A7.8)
the thermal deflection at the "cut" was
Evaluated, the equations were
1 R aT
The redundant moment equation was (by analogy to Ef X - Ef Y + o·g = + 11
eq. 13)
X
m 2dx
_x__ = - 6xT
-R 1 (1
EI X + 2 AE + EI
2) Y + o· g
R aRT
h
EI J
"Therefore O-X + O·y + ( - 1 + -R 2) g=;O
AE EI
L
X EI - - TaL/h Note that from the last of these equations
g = 0, as it must because of the symmetry of the
X = - TaEI ring. Solving the first two equations
-h-
X = aT:I
The redundant moment compresses the upper fibers
as was to be expected.
Y 0
Example Problem 11
Complete the problem begun in Example A non-zero value of Y would produce a vary-
Problem 24 Art. A7.8, viz, that of computing ing bending moment which cannot be because of
the thermal stresses in a closed ring whose symmetry. Hence this result too, is rational.
inner surface is uniformly heated to a tempera-
ture T above the outside. AB.10 Redundant Problem Stress Calculations by
Matrix Methods.
Solution: In the following section the indeterminate
structural problem is formulated in matrix no-
The ring was made determinate by cutting at tation. The reader is assumed to be f~~iliar
the top as in Fig. A7.30(b). The unit loadings With the matrix applications of Art. A7.9 and
and thermal deflections were determined in the the elements of matriX notation and arithmetic
referenced example. The results of deflection (see Appendix).
calculations made previously were The stress distribution of the structure is
specified by a set of internal generalized forces,
°xT = 2n RaT qi' qj'* (ref. Art. A7.9). Unlike the cas~
h
the determinate structure, these qi' qj cannot be
2
oyT = -2n R aT
11 * In the case of indeterminate structures, wherein some of the
support reactions may also be redundant, these reactions
also are denoted by q's. (see Example Problem 13a).
OgT 0
A8.17
related immediately to the external loads by the SYMBOLS - continued

equations of statics. Thus a certain subset of
the qi' qj are redundants and are denoted by - the temperature (above normal) at
points i, j.
~, qs' (r, s different numerical subscripts). - the member thermal distortions as-
When, finally, the redundant forces, qr' qs' sociated with qi' qj'
are computed (by satisfying continuity) the true
values of all the qi' qj may be found by
In the notation of Arts. A8.3 et seq, the
statics. final true values of the stresses were ex-
pressed as (eq. (5) Art. A8.3)
SYMBOLS - - - - (5)
- internal generalized forces acting

on structural elements and re- In the notation to be employed here, this
actions at support points. equation is restated as
- redundant generalized forces and
redundant reactions (14)
Pm' Pn - applied external loads
gim (:::gjn) - the value of q., (q ) in the de- Here:
terminate strueturejunder appli-
cation of a unit external load
P :: 1 (p =: 1) [g im] ( =[g jn] ) * is the matrix of unit-load
m n
gir (=:gjs) - the value of qi (qj) in the de- stress distributions in the determinate ("cut")
terminate structure due to appli- structure found by the application of unit
cation of a unit redundant force (Virtual) loads at the external loading points.
~=:l(qs:::l)
The product [gim] { Pm} then gives the real loads
Gim (=:G jn ) - the true value of qi (qj) in the
in the determinate structure, corresponding to
redundant structure due to appli- the "s" loads of eq. (5).
cation of load P =: 1 (p =: 1)
m n
Grm (=:G sn ) - the true value of ~ (qs) for a [girJ~[gjsJ) is the matrix of unit-load
unit value of applied load stress distributions found by application of
P =: 1 (p =: 1).
m n unit (Virtual) forces at the redundant cuts in
- member flexibility coefficient: the determinate ("cut") structure. Hence this
deflection at point i for a unit is the matrix of ux ' u y ' etc. loads.
force, qj =: 1 (see Arts. A7.9, 10) The { ~}. of course, correspond to X, Y,
a
mn
- influence coefficient for the de- etc.
terminate structure: displacement Note that the [gim] and [gir] matrices are
at external loading point m for a
unit applied load, P :: 1. load distributions computed and arranged in much
n
(a
mn
::: a )
nm the same fashion as was [G im] of Art. A7.9. The
a (=:a ) - influence coefficient for the de-
rn sm small letter "g" is used to indicate load dis-
terminate structure: displacement tribution in the "cut" structure.
at redundant cut r (s) for a unit By way of illustration, the final result
applied load, P =: 1 (p :: 1). for Example Problem 3, Art. A8.5 is expressed
n m
(a ::: a ) below. FIRST, in the form of eq. (5):
rn nr
a
rs - influence coefficient for the de- True Stresses =: S + X (u ) + Y (u )
terminate ("cut") structure: dis- x Y
placement at redundant cut r for a AE =: o + 521 (.806) + 416 (1.154)
unit redundant force q :: 1.
(a - a ) s BE 2000 + 521 (-1.564) + 416 (-1.729)
rs sr
influence coefficient for complete CE o+ 521 (1.00) + 416 (0)
redundant structure: deflection DE o+
at external loading point m for a 521 (0) + 416 (1.00)
unit applied load, P :: 1.
n * Note that withineach of the sets of subscript symbols (i, j),
(r, s), (m, n) the symbols may be used interchangeably.
AS. IS STATICALLY INDETERMINATE STRUCTURES
SECOND, in the form of eq. (14): of external-point influence coefficients in the

"cut" structure: deflection at point m per unit
load at point n.
(17)
- the matrix
of influence coefficients relating relative dis-

placements at the "cuts" to external loads:
Note that in this case consisted of only displacement at cut r per unit load at point n.
one column, inasmuch as there was only a single (18)
external load.
In Art. A7.9 the strain energy was written - the matrix
of influence coefficients relating relative dis-
2U = Lqi J ~i i] {q j} - - (15) placements at the "cuts" to redundant loads at
the "cuts": displacement at cut r per unit re-
dundant force at cut s.
where [U ij ] is the matrix of member flexibility With the above notation one may write
coefficients (Art. A7.10). If now eq. (14) and
its transpose are used to substitute into (15) 2U = L PmJ ~mnJ {pn } + 2 L ~J ~r;] {pn }
the expression becomes (note the use of (i,j),
(r,s), and (m,n) interchangeably)
2U = (LPmJ [~iJ L~J [~iJ)x

+
Now according to the Theorem of Least Work
~ii] x (~j~ { qs } + ~j~ {pn } )- aU/aqr = 0 for continuity. Then, differentiating

eq. (19):
Multiplying out
(20)
This last result may be verified by writing

eq. (19) out in expanded form~ differentiating
and then recombining in matrix form. Rearranged,
eq. (20) gives
(21)
The reader may satisfy himself that the Eq. (21) is a set of simultaneous equations
"cr0ss product" term in the middle of the above for the redundant internal forces q , q. It
r s
result is correct by observing that, because of may be compared with eq. (6) of Art. A8.6, to
the symmetry of [Uij ], which it corresponds. Eq. (21)" may be solved
directly from the form there displayed or its
solution may be obtained by computing ~rs:J,
L~J [~iJ ~ii] ~j;]{Pn}= the inverse of the matriX of coefficients, giVing
(22)
LPmJ [~iJ ~ii] ~j~ {qs}
The various matrix triple products occurring The matrix product - ~r;~ ~r~ gives the
above are assigned the following symbols, each values of the redundant forces for unit values
having the interpretation given (compare With of the external loads. This may be given the
eq. (24) of Art. A7.9)
symbol ~s;] so that
~m;] == [~iJ ~iil ~j~ - - - - - - - - -(16)
- the matriX
AB.19
If now eq. (22) is subst i tuted int.o eq.

(14) one gets (with exchange of r for s and LZI3 LI2 0 0
m for n)
L/ 2 1 0 0
{qi}::: ~i~ {Pm} - ~i~ ~sr~ ~s;;] {Pm} [ll~ L

EI 0 0 VI3 L/2
0 0 L/ 2 1
(~i;] - ~iJ ~s[] ~s;] ) {Pm} The moments qz and q4 were taken as the redund-
ants. With these set equal to zero, the internal
force distributions due to application of unit
(~i~ + [!i~ ~rmJ ){Pm
} values of P~ and Pz were determined, giving
The matrix set off in parentheses above,

gives the internal force distribution per ~it
value of the external loads. It is given the
symbol
(23)
With the applied loads set equal to zero, unit

values of the redundants were applied yielding
(r :::) (2) (4)
so that
(24)
Eqs. (23), (24) constitute the major re-

sUlt, inasmuch as they present the means for
computing the internal force distribution in a
redundant structure.
Note that redundant load q2 was applied as a
Example Problem 13 self-equilibrating internal couple, acting on
The doubly redundant beam of Fig. A8.27 (a) both beam halfs.
is to be analyzed for the bending moment dis- ---- The following matrix products were formed:
tribution. The beam is loaded by couples over ~r;] = ~r!J ~ljJ ~j ;]
the supports as shown.
EI constant ......P2 -PI

= iEI L1IL
0
1 IlL
0 -IlL
J
1
L"/3
L/2
L/2
1 0
0 0
0
-IlL
1
0
0
~ 1
f-- L
=+== L ----1
(a) 0
0 0
0 L"/3
L/2
L/2
1
IlL
0
-IlL
1
~~
q2( tl
ql L
} (bl = 6EI
q4( tl
q3
Fig. AB.27
~r;] = [~!J ~ljJ ~j;]
Solution:
.~ L 'IL
J~'/3 L/2 0
:J [~L
n
1 'lL
EI 0 0 -IlL 1 L/2 1 0
The choice of internal generalized forces
is shown in Fig. A8.27 (b). The appropriate 0 0 L"/3 L/2 0
member flexibility coefficients were arranged in 0 0 L/2 1 0
matrix form as (ref. Art. A7.10 for coefficient
expressions) .
= 6E1
L
.. ~O ~
1 2
l
AS. 20 STATICALLY INDETERMINATE STRUCTURES
The inverse of Cars] was found (ref. though the matrix of member flexibility coe-
appendix) fficients was expanded to a 6 x 6, the
coefficients for qe and qe were zero. Thus,
~rs
a _;I.
7L C
J = 6EI -1
2
-J 4 o o o 0
1 o o o 0
Next, the unit redundant load distribution
was found (eq. 22). o
o o 1 o 0
o o o o 0
o o o o 0
r:. 286 -. 42 81 With the redundants qe and qs set equal to

L·l43 -.28~ zero, successive applications of unit external
couples P;I. and P sa gave the stress distribution
Finally, the true unit stress distribution
was computed. o o
1 o
o o
1 1
L J
1 0 _l/L 0 o286 -o42 o o
1 0 0 1 0 .143 -.286 o 0
"" L +
0 1 l/L - l/L
and, with P;I. and P sa zero, successive applica-
0 0 0 1 tions of lmit redundant forces qe and qs gave
-1 0
~1 1
1.286 .428
2 L
-1 -1
0
~
1
L 2 - .286L -.428L ~1~ 2L L
3 - .429 .858 1 0
4 .143L -.286L 0 1
(The tabular form of presentation of the Then, multiplying out per eqs. (17 ) and (18)
matrix G , above, is used here only to indicate
im
clearly the functioning of the subscript nota-
tional scheme. In general, it should be unnec-
essary to callout the subscripts in this
fashion excepting for the larger matrices, for
~r~
L2
2EI bJ
the handling of which, the tabular form may
prove helpful.)
Example Problem 13a

~rJ
L3
6EI r :J
1-
6
5
The redundant beam problem of Fig. A8.27 is The inverse was found:
to be re-solved using the redundant reactions as
the unknowns.
6EI -51
Solution: 7L 3 lsJ
The support reactions under the loads P;I.
and P sa were given the symbols qe and qe, respec- Finally,
tively, positive up. These forces did not enter
into the strain energy expression so that, al-
A8.21
1.286 .428
.286L -.428L
1 - .428 .857
-[
.143L -.286L
-1.286 -.428 Loading for column { gi2} of the matrix [gimJ
1.714 -.428 Fig. A8.29
(Compare with solution to Example Problem 13).
Example Problem 14
The continuous truss of Fig. A8.28 is
twice redundant. It is desired to analyze it
for stress distributions under a variety of
loading conditions consisting of concentrated
vertical loads applied at the four external
points indicated.
8 panels at 20" = 160"
Solution:
Fig. A8.28
20
Ix 1
3
0
5
0
20 2 .0 0
The internal generalized forces (qi,qj) 20 3 l.0 0
20 4 .5 .5
employed were the axial loads in the various 20 5 0 l.0
members. These were numbered from one to 20 6 0 .5
thirty-one as shown on the figure. The member 20 7 0 0
flexibility coefficients in this case were of 40 8 0 0
the form a i = L/AE (Ref. Fig. A7.35a). The 40 9 -.25 0
40 10 -.75 0
coefficients are written as a column matrix be- 40 11 -.75 -.25
low. (They were employed as the diagonal ele- 40 12 -.25 -.70
ments of a square matrix in the matrix mUlti- 40 13 0 -.70
plications, but are written here as a column to 1
~i~
ai 40 14 0 -.25
conserve space.) E 40 15 0 0
Member loads q3 and q6 were selected as 22.4 16 0 0
redundants. With q3 and q6 set equal to zero 22.4 17 0 0
("cut"), unit loads were applied successively 22.4 18 .56 0
at external loading points one through four, 22.4 19 -.56 0
the four stress distributions:thus found being 22.4 20 .56 0
arranged in four columns giving the matrix 22.4 21 -.56 0
[!i~ (below). By way of illustration, the 22.4 22 -.56 .56
22.4 23 .56 -.06
loading figure used to obtain the second column 22.4 24 -.56 .56
22.4 25 .56 -.56
of [!i;] is shown in Fig. A8.29. 22.4 26 0 -.56
22.4 27 0 .56
Next, unit forces were applied successively 22.4 28 0 -.56
at the redundant cuts "three" and "five" as shown 22.4 29 0 .56
in Figs. A8.30a and A8.30b. These loads were 22.4 30 0 0
arranged in two columns to give the matrix [!i~ 22.4 31 0 0
~ 1 2 3 4
059 .111 .0065 . 003 51
[ .0035 .0065 .111 .059J
1 1.0
2 .50 -.50
3 The calculation was completed as per eq.
4
5 (23) to give [G ] , the values of the member
im
6 -.50 .50 forces for unit applied external loads.
7 1.0
8 - .50
I~
9 - .7::> .2b
1 2 3 4
10 - .2.5 .25
11
12 1 1.0 0 0 0
13 2 .529 -.445 .003 .001
.25 - .25 .059 .111 .006 .003
14 3
.25 - .75 4
15 .031 .059 .059 .031
- .50 5 .003 .006 .111 .059
16 1.12 -6
17 -1.12 .001 .003 -.445 .529
18 - .56 -.56 7 0 0 0 1.0
19 ..')6 .56 8 - .50 0 0 0
20 - .56 .56 9 - .764 .222 -.001 0
21 .56 -.56 10 - .294 .167 -.003 - .002
22 11 - .045 -.085 -.032 - .017
23 12 - .017 -.032 -.085 - .045
24 13 - .002 -.003 .167 - .294
25 14 0 -.001 .222 - .764
26 -.56 .56 15 0 0 0 - .50
27 16 1.12 0 0 0
.56 - .56 17 0 0 0
28 .56 .56 -1.12
29 -.56 - .56
18 - .527 -.498 .003 .001
30 -1.12
19 .527 .498 -.003 - .001
31 1.12
20 - .527 -.622 .003 .001
21 .527 .622 -.003 - .001
NOTE: VOIDS INDICATE ZEROES 22 - .031 -.058 .058 .031
23 - .031 .058 -.058 - .031
Multiplying out gave 24 .031 -.058 .058 .031
25 - .031 .058 -.058 - .031
26 - .001 -.003 -.622 .527
[136 -8·01 27 .001 .003 .622 - .527
L:8.0 136J 28 - .001 -.003 -.498 .527
29 .001 .003 .498 - .527
30 0 0 0 -1.12
31 0 0 0 1.12
Example Problem 14a

L~'O
-~.J
1 -15.0 0 Fig. A8.31 shows the two bays of a steel
= tUbular tail fuselage truss which is loaded by
E 0 -1.5.0
tail air loads to be resolved into three con-
centrated loads applied as shown. The fuselage
The inverse of ~r;] was found (ref. bulkhead at the attach-points station (A-E-F-K)
appendix) is heavy enough so that it may be assumed to be
rigid in its own plane. Hence, the truss may
be analyzed as if cantilevered from A-E-F-K as
~rs~ = 18,432
E [136 8·01 shown. All members are steel tUbes, their
l2.0 13~ lengths and areas being tabulated below.
Next, the values of the redundant forces, Solution:

for unit values of the applied loads, were
found per eq. (22). The generalized forces were taken to be
the member axial loads, these being numbered as
in the table below. Member flexibility coeffic-
ients, L/A, (E set equal to unity for conven-
ience) were also tabulated.
AS 23
than apply successive unit loads and forces at

points m = 1, 2, 3 and r = 22, 23, 24 and then
carry each loading through the structure, an-
other procedure, often better adapted to large
complex structures, was employed.
In the method used, the equations of static
equilibrium were written for each of the seven
joints.* Summation of forces in three directions
on each joint gave 21 equations in 24 q's (the
unknowns) and the three applied loads P 1 , P s and
P 3 • Some typical equations obtained were:
From ~ F on Joint C.
E .21368 q,. - .2136~ q,. + .18334 q. - .18334 q. = -Po

"'~ P1 q. + q. = 0
A~ -" - - - - /" . . -::---
~!JJ~
~/
A D P2 {
q". + qlO = -1.0236 P ,
I 'B~IY
..... 1'1 .......... ..... c P
3
~II"~.- _----
..... {\I )1'
!JJ
~ IK_ _ _ _ _llll---_::-
_ [,' _- ·-I From ~ F on Joint B.
rt--- H
F ~--- G .081759 q, - -.07617 q,. = .5227 q.
Fig. AS.31
q, - - .91593 qu = 0
{
-.1410 q, + + .4146 qu :::; -.8523 qae
MEMBER NUMBER LENGTH AREA L/A And so forth, for the other joints.
Note that in each case the equations were
AB 1 35.47 .565 62.8 arranged with the applied loads (P ) and the
BC 2 25.43 .499 51.0 n
BD 3 9.20 .165 55.8 redundant q's (qss, qS3, qS4) grouped on the
CD 4 25.43 .499 51.0 right hand side of the equal sign. This ar-
DE 5 35.47 .565 62.8 rangement was observed for all 21 equations,
FG 6 35.47 .565 62.8 af~er which the equations were placed in matrix
GH 7 25.34 .565 44.85 form as
HI 8 5.00 .165 30.3
IJ 9 25.34 .565 44.85
GJ 10 9.20 .165 55.8
JK 11 35.47 .565 62.8
AG 12 40.4 .630 64.1
BG 13 15.0 .165 90.9 i,j = 1,2, ..•••.. 24
BH 14 27.57 .500 55.14
HC 15 11.70 .395 29.62 n = 1,2,3
IC 16 11.70 .395 29.62 s = 22,23,24
DI 17 27.57 .500 55.14
DJ 18 15.0 .165 90.9 (Note that there were 24 equations here,
EJ 19 40.4 .630 64.1 the additional three equations being the ident-
AD 20 37.36 .565 66.1 ities
FJ 21 37.36 .565 66.1
GI 22 26.23 .500 52.46
DG 23 17.60 .165 106.6
BJ 24 17.60 .165 106.6 qS3 = qS3
qS4 = qS4 )
The structure was three times redundant.
In ~ space framework of p joints, 3p-6 inde- On the right hand side of the above matrix
pendent equations of statics may be written equation the matrices are shown "partitioned".
(p. A8.10). Here, however, stress details in The first three columns of [D] are the coeffic-
the plane AEKF are to be sacrificed; six equa- ients of P and the last three are the
tions are lost thereby since only net forces in 11
two directions in this plane can be summed.
3 x 11- 6 - 6 = 21 equations; 24 member unknowns.
Members 22, 23 and 24 were cut. * For structures other than trusses the equilibrium equations
The next step was to compute the unit are written for the various structural elements, equilibrium
stress distributions [gim] and [gir] • Rather of joints alone bemg inappropnate.
coefficients of the q 0 The inverse was found (see appendix)

s
The matrix equation was solved for the q.
by finding the inverse of [C ] (see appendix).
J
_" [5555 - .2880
.277~
~rs-J
ij 10 -.2880 3.04-12 2.301.5
Thus,
.2775 2.3015 3.023
Then (referring to eqs. (22) and (22a) )
where
~-186.4-
.10 513 2289 ~
~s~ =-~rs~J ~rJ
10-" -238 -201.0
- 31.4- 176 121.3
Finally, the complete unit stress distribution

Thus, the unit stress distributions in the
determinate structure were found by a procedure was obtained as
having as its main advantage the reduction of
the work to a routine mathematical operation.
In the conduct of this work appropriate stand-
ardized techniques may be employed, *
The result in this case was
--- R 1 2 3
~ 1
1
.8416
2
2,727
3
6.074
22
-2.3059 0
23 24
0
1
2 -
.84-12
.00017
1.4-75
20208
1.296
,538
2 0 2.727 2.727 .9565 0 0 3 .01523 - .31.53 - .1190
3
4
- .00117 - .2771
0 -2.727
- .2821\.
-2.727
.0991
.9560
0
0
- .5227
0
4- .00017 -2,208 - ,539
5 .8440 -4.059 -7.943 2.786 1.236 -1.236 5 .6529 -3.058 -1.957
6 -1.217 .5456 -5.016 2.725 1.209 -1.209
7 - .8445 0 -3.8607 1.354 0 0 6 -1.4-04- 1.524-7 .832
e .1094 0 .5000 - .1754 0 0 7 - .84-4-3 .7352 - .761
~
0 3.862 -2.3205 0 0
c--=--.8446
0 .00013 - .0007 - .1914 - .5224 0 8 .1094- - .0952 ,0985
11
12
-1.4234 - .5800
.4415 -.6473
5.926
1.371
-3.734
- .4809
-2.417
-1.135
2.417
1.135
9 - .84-50 -1.260 -1.4-50
13
14
- .2623
.9188
.3845
0
- .8145
4.200
.2856
-1.473
0
0
- .8523
0
10 .0973 .0205 - .3338
15 - .5118 0 -2.310 .8;08 0 0
11 -1.04-94- -1.607 -1.84-2
16 - .5118 0 2.340 - .8208 0 0 12 .6638 - .314-3 .733
17
18
.9190
- .2619
0
.4090
-4.202
.7882
1.4737
- .2761
0
- .0327 -
0
.019(2 13 - .2355 .3896 - .264-1
19 .3605 .6884 - .9593 .2445 1.401
-1.302
-1.401 14- .9185 - .7998 .8283
20 .0005~ 1.3745 1.399 .4906 1.302
21 - .06621 - .0003 .3044 .2941 1.302 -1.302 15 - ,5116 .4-4-57 - ,4-61
22
23
0
0
0
0
0
0
1.0
0
0
1.0
0
0
16 - .5120 - .4-4-57 .4-61
24 0 0 0 0 0 1.0 17 .9193 .8002 - 0829
18 - .1061 - .3638 .3205
The member flexibility coefficients L/A 19 .14-34- .24-11 - .8512
were arranged as the diagonal elements of the 20 .2023 1.64-71 .6956
21 - .2680 - .3796 .5580
matrix [Q ij ] Then, multiplying out according 22 .0002 .54-3 2.289
23 - .1864- - .238 - .2010
to eqs. (17) and (18), 24- - .0314- .176 .1213
['77.9 -1137.0
-6430.~ Example Problem 15
The dOUbly symmetric four flange idealized
~r~ 195.0 - 151.7 -2263.0
box beam of Fig. A8.32 is to be analyzed for
-154-.4- 161.6 2273.3 stresses due to load application at the six
points indicated. Flange areas* taper linearly
from root to tip while sheet thicknesses are
-1l1~
1150 constant in each pane10 The beam is mounted
[ 2970
riGidly at the root, prOViding full restraint
~r~ 1150 1221 -1035
-114-8 -1035 1224-
* These are the "effective areas", being the flange area plus
* See references in appendix. adjacent effective cover sheet area plus one-sixth of the
web area. (The factor of one-sixth provides the same
moment of inertia as the distributed web area).
AB.25
against warping of the root cross section due The member flexibility coefficients were
to any torsion loadings. collected in matrix form as shown below. Note
that entries for a 44 , a~~, a 99 and a~o,~o were
collected from two stringers each (as well as
being doubled as discussed above). Coefficients
for these tapered stringers were computed from
the formulas of Art. A7.10.
)z 1 10 11 12 13
1 1220
" 15
2 51 200
3
•
5
6
1220
.. 2830
25.2
20" 7
6
60 000
2630
9 25.2 80.8 15.28
10 160.
11
12 6 , 0
13
I. 15.28 27.6
20" 15 15.28 27.6
\
NOTE: VOIDS INDICATE ZEROES. E = 1, G = .385
A = .12 in 2
(Typ.)
Cover sheet shear flows q2, q7 and q~2 were
selected as redundants. With these set equal to
Fig. AB.32 zero, and with unit loads placed successively at
loading points "one" through "Six", the [gim]
matrix was obtained. When the cover sheets are
Solution: "cut" the two webs act independently as plane-
web beams. The details of the stress calculation
The generalized forces employed are shown for such beams are similar to those of Example
on the exploded view, Fig. A8.33. No forces Problem 21, Art. A7.7 and are not shown here.
were shown for the lower surface, its members
and forces being equal to those of the upper
'{\ 1 2 3 4 5 6
surface because of symmetry. In the [U ij ] 1 .400
2
matrix this fact was accounted for by doubling 3 .400
the member fleXibility coefficients for the 4 4.006
members of the upper surface. 5 4.006
6 .100 .200
"1
8 .100 .200
'"q :l.4 \ q~ 6 9 5.341 2.671
10 5.341 2.671
~\ ~~~"
11 .04445 .08887 .1333
12
13 .04445 .08887 .1333
14 6.231 4.006 2.003
15 6.231 4.006 2.003
'q
NOTE: VOIDS INDICATE ZEROES.
Calculations for [gir] were made by suc-

cessively assigning unit values to the redund-
ql,~~,
ants q 2, q 7 and q a . The calculations are
illustrated by the exploded view of the end bay
in Fig. A8.34 showing the calculation in that
part of the structure for q2 = 1. Note that
q2 = 1 was applied as a self-equilibrating pair
of shear flows acting one on each side of the
q~ "cut". The ribs were considered rigid in their
own planes.
Fig. AB.33
From equilibrium of end rib: (Z M, Z F)

The following matrix products were formed:

Pereq. (18):
3774 .1789
'0520~
.1789 .2678 .07322
[ .1254
.0520 .07322
Per eq. (17):
15 -.03356 .03356 -.009318 .009318 -.001605 .001605

1 x 20 =.75 -.02324 .02324 -.01156 .01156 -.002259 .002259
-.008216 .008216 -.005150 .005150 -.002278 .002278
The inverse of Cars] was found (ref.

appendix) •
q,,= 1
q~ = -11G" .-.-~ q3 = 1 E
t308~ -2.562
- 0113J
~rs-J 10 6
-2.562 6.137 -2.521
- .1137 -2.521 9.499
Fig. A8.34
Then
q~ -1.0
q3 '" 1. 0 ~sJ - ~rs-J ~rJ
~
From equilibrium of 2nd rib:

0699 -.0699 .00598 -.00598 .00016 _'0001~
qa ~ -(.5625 - .25) ~ -.3125
'" .03593 -.03593 .03409 -.03409 .00403 -.00403
q6 .3125 [ .01568 -.01568 .01872 -.01872 .01578 -.01578
q7 0, by hypothesis
Finally, the true stresses were (per eq. 23)
So on, into the next bay. In the idealization
used here, the ribs have z~ro stiffness normal
to their own plane so that the axial flange---
loads are transmitted directly to the flange
ends of the adjacent bay (see q4, q7, qe, q~o
of Fig. A8.33).
fi\ 1 2 3 4 5 6
1 .330 .0699 - .0060 .0060 -.0002 .0002
I'?\
1 - 1.00
2 7 12
2
3
.0699
.0699
4 2.26
5 1.75
- .0699
.330
1.75
2.26
.0060 - .0060 .0002 -.0002
.0060 - .0060 .0002 .0002
- .150
.150
.150 -.0040 .0040
.150 .0040 -.0040
2 l.00 6 .0859 .0141 .168 .0322 -.0040 .0040
3 l.00 7 .0359 - .0359 .0341 - .0341 .0040 -.0040
4 -25.04 8 .0141 .0859 .0322 .168 .0040 -.0040
5 25.04 9 2.83 2.51 1.545 1.126 -.122 .122
6 .3125 - 1.00 10 2.51 2.83 1.126 1.545 .122 -.122
7 l.00 11 .0454 - .0010 .0776 .0112 -.0150 .0150
12 .0157 - .0157 .0187 - .0187 .0158 -.0158
8 - .3125 l.00
13 - .0010 .0454 .0112 .0776 .0150 -.0150
9 -20.87 -29.38 14 3.47 2.76 2.40 1.61 .610 -.610
10 20.87 29.38 15 2.76 3.47 1.61 2.40 -.610 .610
11 .1387 .1955 - 1.00
12 l.00 The reader will observe that the result
13 - .1387 - .1955 1.00 displays the "bending stresses due to torSion";
14 -18.79 -26.44 -31.71 that is, the bUildup of axial flange stresses
15 18.79 26.44 31.71 near the root of a beam under torSion when the
root is restrained against warping. The solu-
NOTE: VOIDS INDICATE ZEROES. tion for application of a torque is readily
A8 27
found by superposing stresses for P~ = 1 and
_1~.J
44 -15.0 o
P 2 = -1. Thus, one finds that under this con-
1 -15.0 38.0 o
dition there is a root flange load of [amn] = ~m!J ~lil ~j!J =E 0 0
38.0
[
q~4 = 3.47 - 2.76 = 0.71 Ibs. o 0 -15.0 144
for a torque of 15 inch Ibs.
472 .888 .0525 .028J
A8. 11 Redundant Problem Deflection Calculations by .888 1.665 .0975 .0525
Matrix Methods.
..0525 .0975 1.665 .888
Deflections (and in particular, the matrix [
.028 .0525 .888 .472
of influence coefficients) are readily computed
from the results of Art. A8.10.
Assume that the redundant forces { qs } have Finally, the sum of the above two matrices gave
been deterTIined from eq. (22). The total de-
~'M
flection of external loading points is then -15.9 - .052 - 002
easily computed as the sum of, ONE, the de-
flection due to external loads acting on the _ 1
- 15.9 36.3 - .098 - .052
"cut" structure, [Qrnn ] {Pn } (ref.

TWO, the deflections due to the redundant
forces acting on this same cut structure,
eq. 16) and,
~m;] - E -
-
.052
.028
-
-
.098
.052
36.3
-15.9
-15.9
144 J
Example Problem 17
Thus, Determine the matrix of influence coefficients
for the box beam of example problem 15.
Solution:
An alternate procedure to that shown by
eq. (25) was followed. The influence coeffic-
SUbstituting from eq. (22) ient matrix was formed as in Chapter A-7, Art.
A7.9, by the product.
{ Om } = ~rnnJ { Pn } - ~m;] ~rsJ ~r~ {Pn }
= ( ~rnnJ - ~m~ ~rsJ ~r;] ){Pn } This product was formed readily, inasmuch as
[G im] was available from example problem 15.
The matrix expression set off in paren-
theses above, giving as it does the deflections The result was
for unit values of the applied loads, is the
matrix of influence coefficients. Let 4363 3500 1522 1243 29.7 -29.7
3500 4363 1243 1522 -29.7 29.7
1522 1243 897.8 594.2 33.7 -33.7
1243 1522 594.2 897.8 -33.7 33.7
~.7 - 29.7 33.7 - 33.7 36.8 -36.8
- 29.7 29.7 - 33.7 33.7 -36.8 36.8
so that
A8. 12 Precision and Accuracy in Redundant stress
Calculations.
Matters of precision are dependent upon the
number of significant figures obtained and re-
Example Problem 16 tained in dealing with the geometry of the
Determine the matrix of influence coeffic- structure and in the care with which arithmetic
ients for the redundant truss of example operations are performed. In the discussion to
problem 14. follow it is assumed that all due caution is
exercised with regard to the precision of the
Solution: work.
Matters of accuracy have to do with the num-
From the previous work the folloWing matrix ber of significant fiGures finally obtained in
products were formed the answer as influenced by the manner of formu
A8 28 STATICALLY INDETERMINATE STRUCTURES
lation of the problem. The accuracy of the re- inverse cannot be found (ref. appendix). Its
sult may be affected by a nQ~ber of factors, two condition is terrible. Example:
of the most important of which are discussed
here.
Two factors influencing accuracy often are
considered together under the heading "choice of
redundants". They are:
I - the number of significant figures

which may be retained in the solution for the In the matrix [ars ] the off-diagonal ele-
inverse of the redundant force coefficient ma-
trix, [a
rs ] ' (eq. 21). ments are measures of the structural cross-
coupling of one redundant force with another.
The strength of this cross-coupling is depend-
II - the magnitude of the redundant forces ent upon the choice of redundants made in
relative to the size of the determinate forces. "cutting" the structure to make it statically
These two factors are concerned respective- determinate.
ly with the left hand side and the right hand FOR EXAMPLE, the doubly redundant beam of
side of eq. 21, viz. -- Fig. A8.35(a) may be made statically determinate
by "cutting" any two constraints.
r- L --l.- L -+- L -,
they are discussed in detail below.

K IS
,q3
Z1
t q4
'" (a) ~(c)
c::=:::::llqI
I - ACCURACY OF INVERSION OF ~r~ ; (b)
ql rc=::=:=l)Q2 ~
Q 2 (c::=:::::l (d)
THE CONDITION OF THE MATRIX.
The characteristic of the matrix [ars ] Fig. A8.35
which determines the accuracy with which its in- Fig. A8.35(b) shows the choice of general-
verse can be computed is its condition. The ized forces. Only two (q~ and q2) are required
condition of the matrix is an indication of the to describe the strain energy, but the central
magnitude of elements off the main diagonal support reactions were also given symbols as it
(upper left to lower right) relative to those was desired to consider them in the discussion.
on. The smaller are the relative sizes of Then
elements off the main diagonal, the better is
the condition of the matrix. A weJl conditioned - -
matrix is more accurately inverted than a poorly 2/3 1/6 0 0
conditioned one. Two extreme cases are now
given for illustration: 1/6 2/3 0 0
a) the diagonal matrix. Its off-diagonal

~ijJ =.1-
EI 0 0 0 0
elements are zero, so that it is ideally con- 0 0 0
ditioned. ThUS, the inverse of '-- -0
FIRST, suppose the beam was made determinate by
selecting the support reactions qa and q4 as
redundants. The "cut structure" in this case
may be visualized as the beam of Fig. A8.35(c)
whose central supports have been removed.
Application of unit redundant forces qa = 1
is easily and accurately obtained as and q .. = 1 gave
I~ 3 4
2 1
1 - 3 L - 3 L
b) a matrix all of whose elements are equal

~i~ = "00 3
2
- ~ L - 3 L
in each row. All the elements off the main 3 1 0
diagonal are equal to those on. --rhe determin-
ant of such a matrix is zero and hence its 4 0 1
A8.29
MUltiplying out, redundants may be chosen simply. In most struc-

tures, however, the additional labor involved in
seeking an orthogonal set of redundants is not
L" warranted.
18EI (2) In choosing a set of redundants which
will yield a well-conditioned redundant matrix,
r, s '" 3, 4 it is best to make such "cuts" as will leave a
statically determinate structure retaining as
The condition of this matrix is poor. many of the characteristics of the original
Physically, a unit load at point "3" causes structure as possible. ThUS, one may consider
almost as much deflection at point "4" as at "3" that the structure of Fig. A8.35(d) retains more
itself. The cros.s-coupling is large. of the features of the original continuous-beam
SECOND, suppose the moments ql and q2 had structure than does that of Fig. A8.35(c).
been chosen as redundants. The "cut structure" (3) The degree to which one redundant in-
in this case being visualized as in Fig. fluences another (extent of cross-coupling) can
A8.35(d). Application of unit redundant forces be visualized by observing how much their indi-
ql '" 1 and q2 '" 1 gave vidual unit-load diagrams "overlap".
To elaborate on this last point, refer once
again to the above illustrative example. The
I~ 1 2 cross-coupling of q" with q4 may be expected to
be large if their unit-load diagrams are drawn as
1 1 0 below in Fig. A8.36(a). This deduction follows
2 0 1 easily if it is recalled that the dummy-unit load
equation for such a cross-coupling term is of
3 2 1
-[ [ the form J m"m4EI dx* .
,ObV1ously large for m",
1 2
4 [ -[ m4• Strictly, the comparison is With the terms
m~ dxI ' Jm:

Edx wh'1C h f orm th e on-diagonal
Multiplying out, JE I
elements of the matrix [Qrs ]
r, s '" 1, 2
The condition of this redundant matrix is

obviously better than that obtained with the
first choice of redundants. There is less
cross-coupling between the redundant forces.
Fig. A8.36
Thus the analyst, by choice of redundants, determines
the condition of the matrix. The choice may be critical in the
case of a highly redundant structure, for it may prove im-
possible to invert a large, ill-conditioned matrix with the study of the unit-load diagrams for the
limited number of significant figures available from the initial redundant choice ql' q2 (Fig. A8.36b) reveals
data. The following statements and rules-of-thumb may be that the cross-coupling should be small here
useful in the treatment of highly redundant problems.
since an integral of t he f orm mlm2 EI dx can
(1) It is always possible to find a set of
J
have a contribution from the center span only.
redundants for which the cross-coupling is at a
Hence, th e mlm2
EI dx 1S. 0b' V10US 1Y considera bl y
minimum-zero, in fact. Theoretically then, by
proper choice of redundants, the matrix [Q rs ]
J
m2 dx m~EIdx J
may be reduced to a diagonal matrix (ideally
smaller than ~ or
J Thus, a
visual inspection of Fig. A.36 reveals that ql,
conditione-d) . q2 is a better choice of redundants than is
(la) The choice of redundants which gives q", q4·
zero cross-coupling ("orthogonal functions") is
not readily found in general. In some special
structures, such as rings and frames, orthogonal * see eq. (8) Art. A8.7.
(4) Combinations of redundants may be em- such a way that the product [gir] [B rp ] does
ployed to yield new unit-redundant stress dis-
tributions which do not "overlap" as extensively exactly what is desired, viz., "take one times
as do those of the individual redundants origin- the q3 column minus one half the q4 column to
ally chosen. give the first column of the new distribution
For example, suppose that in the previous
illustrative problem, the choice q3' q4 for re- [giP] ". Then "take L1inus one half of the q3
dundants had been made originally, leading to
the unit-load dia~rams of Fig. A8.37(a). In- column plus one times the q4 column to give the
spection of the diagrams leads one to anticipate second column of the new distribution [giP] "
a strong degree of cross-coupling and hence a
new set of redundants is sought. Rather than Multiplying out in the above example
return to the structure to choose new "cuts",
combinations of the m3 and m4 diagrams are
looked for which will have less "overlap" and 1
- :2 L 0
hence less cross-coupling.
It is observed by inspection that two new 1
stress distributions-which have the desired 0 - :2 L
property may be formed from the m3, m4 diagrams
by proper combination. Thus, if one half the
~ieJ =
1
1 - :2
m4 diagram is subtracted from the~ diagram-to
form one newstress distribution and one half 1
the ~ diagram ~ subtracted from the m4 dia- - :2 1
gram to form the other new stress distribution,
the results are as shown in Fig. A.37b. There
is· obviously less "overlap" of the diagrams for Now form the matrix of redundant coefficients
these new combinations. for the new unknowns (subscripts p, 0; [gip] =
[gjO] )
m3~3-tm41~_L 1
2 1
m4~ m
4
-"2 m3IL_ l_:L..L...L..J.~...L-Jo.._1
_ _2~
L3
24EI
(a) (b)
Fig. A8.37 The condition of this matrix is greatly
New unit-redundant-force stress distributions (b) obtained by improved over that obtained for q3' q4 alone
(previously computed), viz.,
combining previous distributions (a).
In this way two new unknowns are introducen by ~rJ = L3/18EI (r,s=3,4)
linear combination. In matrix notation, the old
stress distributions [gir] are transformed to a
(That the Capo] matrix obtained here happens
new set [giP] by forming
to be similar to that obtained for q1' q2 in a
previous example, is coincidental.)
Once a transformation has been performed,
leading to a new unit redundant matrix [giP] ,
the problem may be completed in the "p, a
§L ~ L system". The appropriate equations are obtain-
3 3 ed from eqs. (14), (21), (23) and (25) simply
= ~ L § L by replacing all "r, s," by "p, 0". Thus
3 3
1 o (29)
o 1
where the redundants q (= q ) are the solutions

The transformation matrix [B rp ] is written in of p a
A8.3l
(30) aUXiliary rules, test information, or even in-

tuitive guess-work which leads to a distribution
close to the final true distribution. There is
and where no need to set the redundant (cut) member forces
equal to zero in establishing the determinate
distribution. Instead, reasonable approximate
~p~ - ~p~ ~iLl ~j~ - - - - - - - - (30a) values may be employed for them. (The correc-
tions to these values become the unknown-----
redundants:)
Mathematically, the magnitudes of the re-
dundant forces are directly dependent upon the
The final unit load distribution is magnitudes of the elements in the matrices [a rn]
or [a pn] on the right-hand side of eqs. (21) or
(31)
(30). (The right-hand side of such a set of
simultaneous equations is called the non-homo-
and the matrix of influence coefficients is geneous part.) Thus, the relative merits-or-
given by several possible determinate stress distributions
may be jUdged by forming the matrix product earn]
(32)
with each and comparing results.
FOR EXAMPLE, if the dOUbly redundant struc-
•• - THE MAGNITUDE OF THE REDUNDANT ture of Example Problem 3, Art. A8.5 were formu-
FORCES; SIZE OF ELEMENTS IN earn] lated in matrix form the [gir] and [a ij ]
matrices would be (see Fig. A8.38 for numbering
Inspection of eqs. (14) or (29) reveals scheme)
that the redundant forces act in the nature of
corrections to the determinate stress distri- Fig. A8.38
butions originally assumed (the [gim]). It is Generalized force numbering
apparent that if these corrections are large, schem e in illustrative prob-
then any inaccuracies in the redundants (arising lem. q3 and q4 selected as
from the difficulties inherent in accurately redundants.
inverting Cars] or Capo]) will have an im-

portant effect on the accuracy of the final 432 0 0 0
stress distribution. Thus, as a matter of
general practice, it is desirable to keep the 1 0 720 0 0
magnitudes of the redundant forces as small as E 0 0 402 0
possible.
It follows immediately that one should use 0 0 0 312
for the determinate stress distribution one re-
quiring a minimum of correction, i.e., Select
as the determinate stress distribution one which .806 1.154
approximates the true stress distribution as -1.564 -1.729
closely as possible.
The rule given in connection with the 1.0 o
"choice of redundants" (rule 2, above) is an aid o 1.0
in making a good selection for the determinate
stress distribution. If, as suggested, a deter-
minate structure is obtained by making "cuts" Several possible determinate stress distri-
which leave a system haVing properties similar butions [gim] will now be tried. FIRST, the
to the original, the stress distribution obtain-
ed therein by statics should be a fair approxi- stress distribution obtained by statics alone in
mation to the final true stress distribution. the "cut" structure (q" = q4 0)
However, it is even more important to
realize that the determinate stress distribu-
tion only need be in static equilibrium with
the external applied loads and that it may be
determined with the aid of any appropriate
(This is certainly a poor approximation to the Example Problem 18

final stress distribution). Multiplying out It is desired to increase the accuracy of
the calculation in the case of the box beam,
Example Problem 15. It will be asswned for this
-1l26} 1 purpose that the initial data of that problem
~r;] = ~r:U tit] { gjn} = { -124.5 EI were sufficiently precise to warrant an increase
CUT CUT of accuracy.
SECOND, a stress distribution in which loads of Solution:

0.40 lbs were guessed at for q~ and q3' (In a
two times redundant structure any two forces The first step taken was an examination of
may be assigned values arbitrarily while satis- the unit-redundant stress distributions obtained
fying static equilibrium.) Q2' q4 were found with the choice previously made of q2, q7 and Q~2
by statics. Thus, as redundants. The three unit-redundant stress
distributions were represented graphically as in
Fig. 1\8.39.
0040
0.258 }
0.40
0.0672 •
-18.79 18.79 -26.44 26.44 -31.71 31.71

MUltiplying out one finds
~r~ = { 9A9}
GUESS -101.
Note that a reasonable guess at a stress

distribution resulted in non-homogenous terms
only one-tenth as great as those obtained by
use of the "cut" distribution. The magnitudes
of redundants are correspondingly reduced. Fig. A8.39
THIRD, as a matter of interest, the true Unit-redundant-force stress distributions -

stress distribution, obtained in Example Prob-
lem 3, was used. The result: axial flange forces.
.4.50 } Inspection of' the figures showed that the

{ gim}
= { :~~; ; ~r~ ={ .16}...l
-.16 E following combinations of distributions should
.280 TRUE give new distributions likely to have consider-
TRUE
ably less "overlap".
~~:~~{gi7}+ 0 X{gi'U}
The non-homogeneous terms are practically zero,
as they should be*. The redundant forces would (1)---- 1 x{ gi 2}-
be zero also.
(2)---- 0 x{gis}+ 1 X{gi 7 }- ~~:~{gi,a}
(3)---- 0 X{gi 2}+ 0 X{gi 7 }+ 1 X{gi'a}

* The demonstration here suggests a useful check upon the
fina:l result of a redundant stress calculation. After ob-
taining the final true stresses [Gim] (- [Gin]>, one forms In matrix form the transformation was
and compares the result element-by-element with the

matrix previously computed,
~i~ = [!.i~ t l
-.7103
o
o
1
-.8330
[ a rn ] ;: [gri] [ aii ] [gin] .
Using [gir] as previously computed, the mUlti-
The "true-matrix" elements ought to be zero, or
nearly so, if [Gim] is error-free. plication gave
A8.33
3774 .1789
'0520~
~ 1 2 3
.1789
[ .0520
.2678 .07322
1 - 1.00 .07322 .1254
2 l.00
3 l.00
4 -25.04· It remained to select a determinate stress
5 25.04 distribution which would reduce the magnitude of
6 1.023 - 1.00 the redundants. For this purpose, the engin-
7 - .7103 1.00 eering theory of bending was employed to compute
8 - 1.023 1.00 stress distributions satisfying equilibrium for
\:J -29.38 each application of a unit external load. The
10 29.38 result was (refer to Example Problem 30, Art.
11 1.0285 - 1.00 A7.11).
12 - .833~ 1.00
~
13 - 1.0285 1.00 1 2 3 4 5 6
14 - .010 - .026 -31.71
15 .010 .026 31.71 1 .30 .10
2 .10 - .10
NOTE: VOIDS DENOTE ZEROES. 3 .10 .30
4 2.0 2.0
5 2.0 2.0
Fig. A8.39a shows plots of the new distri- 6 .0875 .0125 0.15 .05
butions, these haVing greatly reduced "overlap" 7 .0375 - .0375 .05 - .05
(compare with Fig. A8.39). 8 .0125 .0875 .05 .15
9 2.67 2.67 1.33 1.33
10 2.67 2.67 1.33 1.33
11 .0423 .0023 .0713 .0179 .100 .0333
12 .020 - .020 .0267 - .0267 .0333 -.0333
13 .0023 .0423 .0179 .0713 .0333 .100
-31. 71 31.71 14 3.00 3.00 2.00 2.00 1.0 1.0
15 3.00 3.00 2.00 2.00 1.0 1.0
NOTE: VOIDS DENOTE ZEROES.
Next the matrix [a ] was obtained by

pn
multiplying out per eq. (30b)
Fig. A8.39a
~
495 -3495 -1841 1841 0 0 ~
New unit redundant stress distributions, [gip]
~pg] = ~ 1167 -1167 1554 -1554 -1450 1450
1082 -1082 1445 -1445 1802 -1802
This result compared very favorably with earn]
previously obtained, the elements being from one-

half to one-tenth as large.
The new redundant coefficient matrix [a ] The solution may be carried to completion in
po
·t
was obtained by multiplying out per eq. (30a).
2 584
-.01132
•0000175
-.01132
.2566
-.03124
.000017J
-.03124
.1254
the "p,o system" using the matrices [gip} Capo}
[gim] and [a pn] in eqs. (31) and (32).
A8. 13 Thermal Stress Calculations by Matrix Methods .
The thermal stress problem is conveniently

formulated in matrix notation by an extension of
This matrix is very well conditioned, being a the techniques presented above.
considerable improvement over that of Cars]
found originally in Example Problem 15, ViZ., First, consider eq. (21), written in the
form
may consist of the sum of several contributions

should qi act on more than one member.
Because o.T is desired at each of the re-
In light of the physical interpretations given r -
dundant cuts, eq. (34) is expanded by writing
to Cars] and earn] , the equation is seen to be
a matrix-form statement of the condition for (35)
continuity at the redundant cuts, viz., the dis-
placement at each cut caused by the redundant
forces plus the displacement at each cut caused [gri] is, of course, the transpose of [gir]'
by the external loads, must be equal to zero.
To modify this equation for thermal stresses, the unit redundant force stress distribution in
the appropriate expressions for thermal dis- the determinate structure. SUbstitution of eq.
placements at the cuts must be added. Following (35) into eq. (33) gives
the argument used in Art. A8.9 one writes
Solution of eq. (36) gives the values of the

where arT is the displacement at the r th cut redundants q , after which the problem may be
s
due to thermal straining in the determinate completed in the usual fashion, viz,
structure. The explicit formlfor this term will
be derived below. Rewritten,
(33) It is obvious that the use of combinations

of redundants (the "p,a" system of Art. A8.12)
is possible. One makes a direct substitution of
Eq. (33) is a modified form of eq. (21),
giving as its solution the redundant forces in [giP] for [gir]' [a pn] for earn]' etc. into eq.
an indeterminate structure under the application
(36) •
of both external loads and a temperature distri-
bution.
MEMBER THER1"lAL DISTORTIONS
To derive an explicit expression for arT'
the virtual work concept may be employed to ad- It remains to establish the forms for ~iT.
vantage. Thus, following the argument of the Thermal strains on an infinitestinal ele-
"virtual work" derivation for deflections (Arts. ment of homogeneous material can cause uniform
A7.7, A7.8), the thermal deflection at the r th normal extensions only, so that no shear strains
redundant cut must be equal to the total in- develop. Hence only normal (as opposed to
ternal virtual work done by the rt-h-redundant shear) virtual stresses need be considered in
force virtual stresses (due to a unit r~-re computing the internal virtual work. Note that
dundant force) moving through distortions normal stresses associated with flexure must be
caused by thermal strains. included. It follows that only virtual work in
Since the internal stress distribution is axially loaded bars and in beams in flexure need
expressed in terms of the internal generalized be considered. Hence ~iT is zero for all qi
forces qi' qj' it is convenient to employ these which are shear flows on panels or torques on
q's in writing the virtual work of straining. shafts.
If one lets ~iT be the displacement of internal
BARS
generalized force qi due to thermal straining,
The general expression for the virtual work
then the virtual work done by a single general- done by virtual axial loads u in a bar under
ized force is qi ~iT. The quantity ~iT will be varying temperature T is
called the member thermal distortion. The total
virtual work throughout the structure is ob-
tained by summing, giving the deflection at the
W :: JU a T dx
r th cut as the matrix product where a is the material thermal coefficient of

expansion.
(34)
where gri is the value of the qi due to a unit * It will be convenient later to designate by qST the solution
to eq. (36) when the mechanical loads P n are zero, the
(virtual) load at cut r. Note that the term ~iT stresses in such a case being purely "thermal".
AB.35
8everal specific cases are now treated. BEAMS

A bar under linearly varying load with For a beam the general form of expression
linearly varying temperature is shown in Fig. for the virtual work done during thermal strain-
A8.40. ing by a top-to-bottom-surface temperature dif-
In this case ference OT, varying linearly over the beam depth,
qjn~
qi - qj is (see Art. A7.8, Ex. Prob. 24).
u qj + x
L oT dx
TJ Ti h
Ti - Tj Fig. AB.40
T Tj + x
L where
Then, assuming a constant m = virtual moment (positive for compress-

ion on top fiber)
oT = TBOTTOM - TTOP
w
h = beam depth.
Ti + 2T j Applied to the case of Fig. A8.42 one gets

aL 6 qj
This expression may be put in the form w li J:~qj q, ~ qj x) (BTJ

+ + BT, ~ BT J x~ dx
where
~ (20Ti 6+ OTj) qi + ~L (OTi +6 20T j) qj
OT~OTi
q.~qi
J(_}--x
(
h = constant
Fig. AB.42
Note that variation in the cross sectional
area of the bar does not affect the distortions or
Ll iT , Ll jT •
The alternate choice of generalized forces
for the bar under varying axial load is shown where
in Fig. A8.41. By a
derivation similar to
that above one finds
_ ~ (20Ti: OT~
(OTi +6 20T~
where Fig. AB.41
Special forms of the thermal distortion
expressions for beams of varying depth may be
derived readily as required.
Example Problem 19
Ti + 2Tj The upper surface of the beam of Fig. A8.43
Ll jT = a Lla
6 is subjected to a temperature oT above that of
the lower surface, varying linearly as shown (i.e.,
The simpler cases of uniform load (qi = u
constant) and uniform T (T = T j ) follow im-
i
mediately by specialization of the above forms.
For example, for a bar under constant load
qi = qj' and constant temperature, Ti = T j = T,
one has Ll = a LT.
iT Fig. AB.43
the temperatures are equal at the left end and

differ by aTe at the right end). Determine the 2EI
center reactions assuming a is constant.
Solution:
Solving
In the illustrative example of Art. A8.12
this structure was analyzed by employing as ql}= EI a aTe{.267}
generalized internal forces the bending moments { q2 h .933
q, and q2 over the central supports (see Fig.
A8.35). The two central reactions were denoted The final stress distribution was
by q3 and q~. The matrix of redundant coeffic-
ients, considering q, and q2 as redundants (the
better choice, it will be recalled) was (ref.
Art. A8.12)
::: o+ EI a aTe 1 0
L
'n
6EI 0 1 .267
2 1 .933
The unit redundant load distribution for q, 1
- L L
and q2 = 1 was 1 2
L - L
I~ 1 2
.267 }
1 1 0 EI a aT e .933
2 0 1 h { .399/L
-1.60/L
3 2 1
- L L Thus the reactions were
1 2
4-
L - L q3 = .399
EI a aTe
Lh
Member thermal distortions were computed. (Note EI a aTe (negative indi-
that aT was negative according to the convention Lh cates DOWN)
adopted earlier).
Example Problem 20
1 The symmetric sheet-stringer panel of Fig.
aL 2 x 3 a Te A8.4-4-(a) is to be analyzed for thermal stresses
h (-) 6
developed by heating the two outside stringers
to a uniform temperature T above the center
aL stringer. Assume G = 0.385E.
- 3H aTe
Solution:
The panel was divided for convenience into

three bays. The numbering and placing of gener-
alized forces is shown in Fig. A8.4-4-(b).
2 a.L Transverse members (ribs, not shown on Figure)
= -:3 11 6 To
were considered rigid in their own planes - a
satisfactory assumption for symmetric panels.
Then from eq. (36) Because of symmetry only one half the panel was
handled. All member flexibility coefficients and
aLaTe thermal distortions were doubled where appropri-
-311 ate.
_ 2aL aT The matrix of flexibility coefficients was
3h e set up as (VOIDS DENOTE ZEROES)
o
o
AB 37
177.6 44.4 Then

44.4 177.6 44.4
44.4 88.8
88.8 22.2 r-: .5 0 0 -
22.2 88.8 22.2
22.2 44.4 0 - .5 0
62,400
62,400 0 0 - .5
62,400
1.0 0 0
was
The matrix of member thermal distortions
~i~ =
0 1.0 0
0 0 1.0
40 - .025 0 0
40
.025 - .025 0
20
o ...... 0 .025 - .025
{lI iT } = a T< o :>
o Multiplying out per eq. 18,
o
o
o
~1l'2
- 5.7
o ]
Here, for example lilT = 2 x 2 x 10 = 40 (doubled ~r~ = E
1 - 5.7 211.2 - 5.7
once because q1 acts on two stringer ends and 0 - 5.7 105.6
doubled again to account for the other half of
and
r ,
the panel).
q.~ EJl
5 0 0 1.0 0 0 -.025 -.025 40
~rJ { o~
, I I I II , , II
c~::::m--1I----lI[~::J 1 61T} = a
T -.5 0 0 1.0 0 0 -.025 40
r
[
o -.5 0 0 1.0 0 0 -.025 20
F =..02 i=
i= q.~ 18l
0
T~~
:= tryp
Ell..-----'L-~.L~
t: 1
60"
q.~ E] r
=a T
{
-ZO}
-20
-10
0
0
1--12'+12''-l~A=.15 in
2
(Typ. )
(a)
(b) ll.
Then the redundant equations, per eq. (36),
Fig. AB.44 were
The structure was three times redundant.
Stringer loads q4' q& and qe were selected as
redundants. Fig. A8.45 shows the unit redundant
load sketches for q4 = 1, q& = 1 and qe = 1.
Solving,
[[ ~ ~8l~ fBlf' 0'0974}

{ qrT} = aTE { 0.100
~ 8l ~ ~'8 f
0.100
~ n~ Finally, the complete thermal stress distribu-

tion was
- 48.7
~'18 f ~ 0 ~ ~ 0 ~
- 50.0
- 50.0
97.4
10-
3
( 100.0 >
~li4} loading I'll. f1i5} loading ill. {gi6pOadin~ ll. 100.0
- 2.44
Fig. AB.45 - 0.065
o
The result compares favorably with "exact"

solutions made under the same assumptions (ref. 1
NACA TN 2240) as far as the stringer loads are
concerned. The shear flow result is not very L/2
satisfactory, primarily due to the use of too
few "bays" in this analysis. 1
Example Problem 21 L/ 2 L/ 2
The uniform four-flange box beam of Fig.
A8.46a is to be analyzed for the thermal
stresses developed upon heating one flange to a 1
temperature T, uniform spanwise, above the other
three flanges. L/ 2 L/ 2 L/ 2
L/2 L/ 2 L/ 2 LI2
Multiplying out per eq. (18),
L~~~ + .833 .625 .375 .125
.625 .500 + .583 .375 .125

Uoi k 8
L 3 k lil (b + c)
~r~ = Gt .375 .375 ~+ .333 .125
.125 .125 .125 ~+ .0833
Fig. A8.46
For a specific case let k2L~ 1. The
Solution: ( b) inverse was computed to be
The beam was divided into four equal bays
giving a four times redundant problem. Four 1.070 - .5291 - .2337 - .06587
self-equilibrating (zero-resultant) independent Gt - .5291 1.366 - .3613 - .1018
stress distributions were taken as the unknowns, ~rs::J L (b + c) - .2337 - .3613 1.497 - .1934
these being shown in Fig. A8.46(b). Such zero-
resultant stress distributions are the only ones - .06587 - .1018 - .1934 1.792
possible in a structure having no applied loads*.
The matrix of member flexibility coeffic-
Member thermal distortions (~iT) were
ients was formed by collecting coefficients from
the several members. Unit redundant stress dis- computed for loads q2' q., q6 and qe and were
tributions were prepared, taking q1, q3' q6 and collected from the one heated flange only.
q7 as the redundants and setting these equal to
unity successively.
o (Note that if two ad-
.500 2 jacent flanges are
.667k' .1667k' o heated equally, one must
.500 a L T 2 set the corresponding ~iT
.1667k' .667k' .1667k' 8 o equal to zero; this be-
.500 2
o cause the virtual loads
.1667k' .667k' .1667k'
1
being of opposite sign in
.500
•333k Ii
adjacent flanges, the
.1667k'
virtual work must cancel).
where k 2 ~ Gt / AE (b + c) Multiplying out:
* Sinceexternal loads are not to be applied it follows from
statics that the generalized forces for adjacent webs and
cover sheets are equal, as are the loads in front and rear
spar caps at any given station. Hence an economy of
numbering in the generalized force scheme is possible.
Much labor is saved in the handling of data when the same
symbol can be employed on several members whose loads
are known to be equal. Then the solution to eq. (36) was written as
A8.39
"cut" making it determinate, after which the

temperature distribution is applied producing
thermal deflections
which, when multiplied out gave
4'079 }
Gt a L T 1.938
{ qST } :: - 16 (b + c) . 8562 at the external points (compare with eqs. (35)
{ and (37) ). Simultaneously the redundant cuts
.2413 •
will experience relative displacements.
Finally the -complete set of thermal stresses SECOND, the redundant cuts are restored to
were (still for the case L2 k 2 :: 1) zero displacement by the application of redund-
ant forces (this problem was solved in Art.
A8.13). The qs are given by eq. (36); they pro-
4.079
2.039 L duce additional deflections at points m
1.938
3.008 L
.8562
3.437 L
.2413
3.558 L The total deflection of point m is then
The result compares favorably with an

exact solution (NACA TN 2240) insofar as flange
stresses are concerned, an error of less than But
1 percent being present at the root. Shear
flow values in the sheet are less satisfactory
due to the (relatively) crude assumption of
constant shear in each beam quarter.
Therefore
A8. 14 Thermal Deflections by Matrix Methods.
A. STATICALLY DETERMINATE STRUCTURES

The problem of the thermal deflections of
a statically determinate structure was con- or,
sidered earlier in Art. A7.8 in non-matrix
form. It should be apparent from the deriva- (38 )
tion, that the matrix method presented for the
calculation of redundant-cut thermal deflec-
tions may be applied equally well to the The matrix quantity in parentheses is the
problem of computing the thermal deflections of total strain (thermal plus "mechanical").
external points of a determinate structure. ----For definite reasons the equation for
Thus thermal deflections has been left in the -form of
eq. (38) rather than the more polished forms
(37) which might be obtained by substitution from eq.
(36), FIRST, the qjT' the thermal stresses, will
where OmT is the thermal deflection of external probably have been solved for previously and will
be readily available in explicit form. SECOND,
point m, and [Gmi ] is the transpose of [G im], and far more important from a labor saving stand-
point, the unit load distribution [gim] (whose
the unit-applied load stress distribution
(compare with eq. 35). transpose is used in eq. 38) may be any conven-
ient stress distribution satiSfYing statics in
B• REDUNDANT STRUCTURES the s1riiPl8St of "cut" structures. One need not
In the case of the redundant structure, even use the same~im] distribution (and same
additional strains are present due to the
thermal stresses set up; the effect of these choice of "cuts") as employed in the redundant
strains upon the deflection of external points thermal stress calculation; a more convenient
must be included in the calculation. choice of cuts may be employed: In principle,
The appropriate equation is most easily any stress distribution statically equivalent to
derived by visualizing the action in two the unit applied load(s) may be used for [gimr
stages. FIRST, the redundant structure is in eq. (38). (See p. A8.9). --
Example Problem 22 a L a To L.Q 0 !J

Compute the rotation occurring at the right h
hand end of the beam of Fig. A8.43.
Solution:
aT ;; - .289
a L a To
To compute the rotation a unit couple was h
applied (positive counterclockwise) at the right
hand end. An additional generalized force, It is apparent that the values of the re-
called q5, also was added at that point. Then, dundant moments q~, qg could have been chosen
arbitrarily in {gim}' above, without affecting
t.
5T =
aL (-)(~ 6To·: 26 To)~ -4 a L a To
the result. Here, clearly, {gim}for any "cut"
h 9h
structure visualized will lead to the same re-
and, using some results from Example Prob. 19, sUlt, q5 being equal to unity in all cases.
Example Problem 23
m
Compute the axial movement of the free end
{~iT} ;; _ a
L a
9h
To of the central stringer of the panel of Fig.
A8.44.
1 =1, 2, 5
Solution:
Note that q3 and q4, the intermediate support
reactions, were omitted from consideration. An additional generalized force, q~o, was
They do not enter into any expression for the added axially to the free end of the central
internal virtual work of the structure; or, stringer (ref. Fig. A8.44b). Then
equally, they are not used to describe the
strain energy of the structure. Hence they are 177.6 44.4
not included in writing the total strain. 44.4 177.6 44.4
(Their ~iT are zero.) ---- 44.4 88.8
88.8 22.2 22.2
The member flexibility coefficient matrix
22.2 88.8 22.2
was
22.2 44.4
62,400
62,400
62,400
i ;; 1, 2, 5 22.2 44.4
40
From Example Problem 19, the true thermal 40
stress distribution was 20
o
o
o
i ;; 1, 2, 5 o
o
(q6 was zero by inspection) o
o
To determine [gim]' a unit couple q5 ;; 1
Using the qjT as obtained in Example Problem 20
was applied. Taking as the "cut" structure one
where q~ ;; q~ ;; 0 one has simply (with q~o • 0), the following product was
formed:
- 10.87
- 13.26
6.660
10.87
Then substituting into eq. (38), 13.26 x aT
6.66
1.2 0U(\: aL~To {3}~ ~ 1~ EI a ~ To {:9~f
6~I ~ i ~
26~l) -152.2
{ Omr } = + 4.06
0.00
2.162
A8.41
Then, sUbstituting into eq. (38), a unit load q10 1 gives (writing the transpose
of {gim})
29.13
26.74
13.34 l- ~i~ = l-.50 .50 .50 000 .025 0 0 l~
10.87
13.26
6.66 Then mUltiplying out for the thermal deflection:
-152.2 OmT = 33.0 a T (same answer).
- 4.06 It is apparent from the above result that
0.00 the Simplest determinate ("cut") structure should
2.162 be used to compute [gimJ; it is completely
adequate.
It remained to find the determinate stress dis-
tribution for q10 = 1•. CLOSURE
As the determinate distribution {gim} the
The general thermal stress problem is
stresses due to a unit load q10 = 1 were com- complicated by the fact that the material prop-
puted in the "cut" structure obtained when erties E, G and a vary With temperature. The
q, = qs = q3 = O. (Note that this is a differ- problem created thereby is primarily one of book-
ent choice of redundant cuts from that employed keeping - computing the member flexibility coeffi-
in computing the thermal stresses in Example cients (a ij ) and the member thermal distortions
Problem 20.) Thus (~iT) for a structure whose properties vary from
point to point with the temperature. The vari-
o ations of E, G and a with T Will, of course,
o have to be known from test data.
o Two additional complications, not considered
1 here, are the lowering of the yield point With
1 heating (and the attendant increased likelihood
1 of developing inelastic strains) and the phenom-
o enon of "creep" (the time-dependent development
o of inelastic strains under steady loading).
o Should it prove necessary to analyze for
1 thermal stresses under more than one temperature
distribution, the member thermal distortion
and finally,
matrix {~iT } may be generalized easily into a
OmT = aT Lo 0 0 1 1 100 0 IJ 29.13 rectangular form such as
26,74
13.34
10.87
13.26
6.66
'-152.2 where ~iR is the member thermal distortion as-
- 4.06 sociated with force qi from thermal loading con-
0.00
2.162 dition R. The matriX [C ij ] of member thermal
33.0a T (ANSWER) coefficients consists of the constant coefficients
in the ~iT expressions previously presented, while
As a matter of interest, a different unit
stress distribution was employed with a differ- T jR would be the temperature associated With qj
ent choice of "cuts", If the forces q401 qll and for condition R. (compare with eq, (26b),
qs are set equal to zero ("cut"), application of Art. A7.ll).
A8.42 STATICALLY lND:f;TERMINATE STRUCTURES
AS.15 Problems. (5) For the "King post" truss in Fig. e,

Note: Problems (1) through (9) below may calculate the load in member BD. Members AB,
be worked by either the Least Work or Dummy Unit BC and BD have area of 2 sq. in. each. The con-
Load Methods. The student will be well advised tinuous member ADC has an area of 9.25 sq. in.
to try some problems both ways for comparison. and moment of inertia of 216 in 4 • E is same for
all members.
5000# A B
-~------"T A D E
100"
20"
T
90"
I- A
B
A
B
0'
4000#
1
F~.
100"
a F~.b
~ 60"
+ 60" -lC
(1) Determine the load in all the members Fig. g

of the loaded truss shown in Fig. (a). Values
in ( ) on members represent the cross-sectional 500#
area in sq. in. for that member. All members
of same material. (6) In Fig. f, AB is a steel wire 0.50 sq.
(2) For the structure in Fig. (b), deter- in. area. The steel angle frame CBD has a 4 in.
mine the load in each member for a 700# load at sq, cross section. Determine the load in member
joint B. Areas of members are given by the AB. E = 30,000,000 psi.
values in ( ) on each member. All members made (7) In Fig. g find the loads in the two tie
of same material. rods BD and CEo lac = 72 in.· ; Abd = 0.05 sq.
500 500# in. Ace = 0.15 sq. in. E is same for all
'A
60" -----rB 30" -ic members.
(1) T B
t
1) (1) 30" 60"
~r-------->"::'.i....--_--"",-------,=------"D1 D~C+
~
6000# I
Fig. c
1000# 1000# 180"
E~
(3) For the loaded truss in Fig. c, deter-
mine the axial load i~ all members, Values in 50,000#
parenthesis adjacent to members represent rela-
tive areas. E is constant for all members. Fig. i
Fig. h
C
(8) For the structure in Fig. h, determine

the reactions at points A, B. Members CE and ED
D are steel tie rods with areas of 1 sq. in. each.
Member AB is a wood beam with 12" x 12" cross
section. Esteel = 30,000,000 psi. Ewood =
E
1,300,000 psI.
(9) For the structure in Fig. i, determine
the axial loads, bending moments and shears in
the various members. The structure is continu-
A ous at joint D. Members AB, BC are wires. The
member areas are AB = 1.2, BC = 0.6; CD = 6.0;
(4) For structure in Fig. d, calculate the BDE = 10.0. The moment of inertia for members
axial loads in all the members. Values in paren- CD = 60.0 in.·; for BDE = 140 in.·
thesis adjacent to each member represent relative
areas. E is constant or same for all members. (10) Re-solve Example Problem 10, p. A8.15
5000# using as redundants the two restraints at one end
f ID
(couple and transverse force). Solved in this
way the problem is doubly redundant as no ad-
~3
C vantage is made of the symmetry of the structure.
(11) Add two additional members, diagonals
FB and EC (each with areas 1.0 ina) to the truss
~_i
Fig.e
I-- 120" + of Fig. A7.85, chapter A-7. Find the matrix of
120" ---l
A8.43
influence coefficients. problems of Art. A8.8 by writing

Ans.
21.8 27.2 27.5 18.3

where ~i is the initial imperfection associated
27.2 68.5 65.4 26.4 with force qi' Refer to the argument leading
[limn] = ~ 27.5 65.4 68.4 26.2 to eq. (11) of Art. A8.8.
18.3 26.4 26.2 20.6 (16) Using the equation of problem (15),
above, re-solve Example Problem 6, p. A8.14.
(12) Re-solve the doubly redundant beam (17) Using the equation of problem (15)
of Example Problem B, page A8.3 by matrix above, re-solve Example Problem 7, p. A8.14.
methods. The redundant reactions should be
given "q" symbols. (See Example Problem 13a, (18) Using the matrix methods of Art.
page A8.20). A8.13, re-solve Example Problem 9, p. A8.15.
(13) Re-solve Example Problem 5, page (19) For the dOUbly symmetric four flange
A8.12 by matrix methods. For simplicity, make box beam of Example Problem 15, p. A8.24, de-
your choice of generalized forces including termine the redundant stresses q2, q7 andq12 if
those designated as X and Y in the example so one flange is heated to a temperature T, uniform
that Figs. A8.21 and A8.22 can be used to give spanwise, above the remainder of the structure.
the gir loadings.
Ans.
(14) By matrix methods re-solve Example
Problem 4, p. A8.12 using 3 equal bay divisions
along the panel (3 times redundant). Use the - EaT 2081}
same structural dimensions as in Example Prob- 10" { 1599
lem 20, p. A8.36. Compare the results with 741
those obtained from the formulas developed in
Example Problem 4.
(15) Show that the matrix equation REFERENCES •

eq. (21) is modified to cover the initial stress
See references at the end of Chapter A-7.
Douglas DC-8 airplane. Photograph showing simulated aerodynamic load being applied to
main entrance door of fuselage test section.
PROOF TEST OF DOUGLAS AIRCRAFT MODEL DC-6
DOUGLAS DC-8 AIRPLANE. An outboard engine pylon mounted on a section of wing for static and flutter
tests. The steel box represents the weight and moment of inertia of the engine.
CHAPTER A9
BENDING MOMENTS IN FRAMES AND RINGS
BY
ELASTIC CENTER METHOD
Assumptions
A9.l Introduction
In the derivations which follow the distor-
In observing the inside of an airplane tions due to axial and shear forces are neglect-
fuselage or seaplane hull one sees a large num- ed. In general these distortions are small
ber of structural rings or closed frames. Some compared to frame bending distortions and thus
appear quite light and are essentially used to the error is small.
maintain the shape of the body metal shell and
to provide stabilizing supports for the longi- In computing distortions plane sections are
tUdinal shell stringers. At points where large assumed to remain plane after bending. This is
load concentrations are transferred between body not strictly true ])ocause the curvature of the
and tail, wing power plant, landin~ gear, etc., frame charl";es this linear distribution of bend-
relatively heavy frames will be observed. In ing stresses on a frame cross-section. (Correc-
hull construction, the bottom structural fram- tions for curvature influence are given in
ing transfers the water pressure in landin~ to Chapter 1\13.
the bottom portion of the hull frames which in
turn transfers the load to the hull shell. Furthermore i t is assumed that stress is
proportional to strain. Since the airplane
In general the frames are of such shane stress analyst must calculate the ultimate
and the load distribution of such character strenGth of a frame, this assumotion obviously
that these frames or rings undergo bending does not hold with heavy frames where the rup-
forces in transferring the applied loads to the turing stresses for the frame are above the pro-
other resisting portions of the airplane body. portional limit of the frame material.
These bending forces produce frame stresses in
general which are of major importance in the This chapter will deal only with the the-
strength proportioning of the frame, and thus oretical analysis for bending moments in frames
a reasonable close approximation of such bend- and rings by the elastic center method. Prac-
ing forces is necessary. tical questions of body frame design are covered
in a later chapter.
Such frames are statically indeterminate
relative to internal resisting stress and thus The following photographs of a portion of
consideration must be given to section and the structural framing of the hull of a sea-
physical properties to obtain a solution of the plane illustrate both liGht and heavy frames.
distribution of the internal resisting forces.
General Methods of Analysis:
There are many methods of applying the

principles of continuity to obtain the solution
for the redundant forces in closed rings or
frames and bents. The author prefers the one
which is Generally referred to as the "Ji:IClStic
Center" method and has used it for many years in
routine airplane design. The method was ori~i
nated by MUller-Breslau~. The main difference
in this method as compared to most other methods
of solution is that the redundant forces are as-
sumed actinG at a special point called the elas-
tic center of the frame Which Gives resulting
eq,mtions for the redundants which are ind'-
pendent of each other.
(j)Muller-Breslau, H., Die Neueren Hethden der

Festigkeitslehre und der Statik der Baukon-
struktionen. Leipzig, 1886.
A9.1
A9.2 BENDING MOMENTS IN FRAMES AND RINGS
Consider a small element ds of the curved

beam as shown in Fig. A9.2. Let Ms equal the
bending moment on this small element due to the
given external load system. The total bending
moment on the element ds thus equals,
(Moments which cause tension on the inside

fibers of the frame are regarded as posi-
t i ve moments. )
The following deflection equations for

point (A) must equal zero:-
Q = 0 (angular rotation of (A) = zero)
~x =0 (movement of (A) in x direction = 0)
A9.2 Derivation of Equations. Unsymmetrical Frame. ~y =0: (movement of (A) in y direction = 0)
Fig. A9.1 shows an unsymmetrical curved From Chapter A7, which dealt with deflection
beam fixed at ends (A) and (B) and carrying theory, we have the folloWing equations for the
some external loading P~, P e , etc. This movement of point (A):-
structure is statically indeterminate to the
third degree because the reactions at (A) and Q = Z MInds _ -(2)
(B) have three unknown elements, namely, magni- EI
tUde, direction and line of action, making a
total of six unknowns with only three equations ~x = Z MInds
EI
- -(3)
of static equilibrium available.
~y = Z MInds
EI
- - -(4)
P e
P~
Pa In equation (2) the term m is the bending
t moment on a element ds due to a unit moment
\ applied at point (A) (See Fig. A9.3). The bend-
ing moment is thus
equal one or unity
for all ds elements A~
~it moment
Fig. A9.1
A
of frame.
Fig. A9.3
ds
P Then SUbsti-
tuting in equation (2) and using value of M from
equation (1) we obtain -
Fig. A9.2 whence,
Q = ZMsdS + MAZ~ _ XAZ YdS + Y ZXdS - 0 (5)

In Fig. A9.2 the reaction at (A) has been EI EI EI A EI -
replaced by its 3 components, namely, the forces
XA and YA and the moment MA and the structure In equation (3) the term m represents the
is now treated as a cantilever beam fixed at bending moment on a element ds due to a unit
end (B) and carrying the redundant loads at (A) load applied at point (A) and acting in the x
and the known external loading P~, P e , etc. direction, as illustrated in Fig. A9.4.
Because joint (A) is actually fixed it does not
suffer translation or rotation when structure The applied unit y
is loaded, thus the movement of end (A) under load has a positive ; ds
the loading system of Fig. A9.2 must be zero. sign as it has been :~
Therefore, three equations of fact can be assumed acting toward l~A- A --- J~-x ~ B
the right. The I
written stating that the horizontal, vertical I
and angular deflection of point (A) must equal distance y to the ds y

element is a plus Fig. A9.4
zero.
distance as it is
A9.3
measured upward from aXis x-x through (A). place of XA, YA and MA respectively.
However the bending moment on the ds element
shown is negative (tension in top fibers), thus The axes x and y through the point (0) are
the value of m = - (1) Y = -Yo The minus sign centroidal axes for the values ds/EI of the
is necessary to give the correct bending moment structure. This fact means that the summations-
sign.
and
SUbstituting in Equation (3) and using M
from Equation (1):-
The expressions Z x 2ds/EI, Z y2 ds /EI and
AX
'-'
= _ZMsydS
EI
_ M ~yds + X ~y2dS
"A"" EI A""~ -
Y ~xydS
A""Er=
(6) Z xyds/EI also appear in equations 6 and 7.
0-- These terms will be referred to as elastic
moments of inertia and product of inertia of
In equation (4) the term m represents the the frame about y and x axes through the elastic
bending moment on a element ds due to a unit center of the frame, and for simplicity will be
load at point (A) acting in Y direction as given the following symbols.
illustrated in Fig. A9.5. Hence, m l(x) x = =
Substituting in
equation (4) and Y ds
using M from
equation (1), we x---~ Equations 5, 6 and 7 will now be rewritten
obtain, t- Fig. A9.5
using the redundant forces at point (0).
e = Z MsdS
EI
+ M Z ds -
0 EI - 0
hence,
- - -(7)
Equations 5, 6, 7 can now be used to solve - - - - - - - - - - - (8)

for the redundant forces MA, XA and YA' With
these values known the true bending moment at
any point on structure follows from equation
(1) •
REFERRING REDUNDANTS TO ELASTIC CENTER The term MoZ yds/EI is zero since Z ydS/EI
is zero, thus Mo drops out When substituting in
For the purpose of simplifying equations
5, 6, 7, let it be assumed that end A is Equat i on (6).
attached to a inelastic arm terminating at a
point (0) as illustrated in Fig. A9.6. The - - - -(10)
point (0) coincides with the centroid of the
ds/EI values for the structure. Reference
axes x and y will now be taken with point (0) The term Msds/EI represents the angle
as the origin. The redundant reactions will change between the end faces of the ds element
now be placed at point (0) the end of the When acted upon by a constant static moment Ms'
This angle change which actually is equal in
Y
value to the area of the Ms/EI diagram on the
I element ds will be given the symbol Os' that is,
0s = Msds/EI. With this symbol SUbstitution,
equations 8, 9, 10 can now be rewritten as
follows:-
Z 08 -(11)
Fig. A9.6 Z dS(EI - - - - - -
inelastic bracket, as shown in Fig. A9.6. -Z 0s Y + Xol x - Yol xy = 0 - -(12)

Since point A suffers no movement in the actual
structure, then we can say that point (0) must Z 0s x - XOl XY + Yol y = 0 - - -(13)
undergo no movement since (0) is connected to
point (A) by a rigid arm. Solving equations (12) and (13) for the
redundant forces Xo and Yo we obtain,
Thus equations 5, 6, and 7 can be re-
written using the redundants Xo ' Yo' and Mo in
X = ~¢sY - ~¢Sx (~) - - - - - - -(14)

Table A9. 1
o y= x= Ix Iy
IX (1 - i);) Por- w=ds dist. dist.
Han I fr05. from
wx wy =wy 2+ ix =wx 2+i
Y
x 1_x y-y
- [~0sX - ~0sY (4;-)J AB 3~=10 15

wy2= 2250 wx 2=1440
-12 -120 +150 i = 30 3 =750 . - 30 - 0
x 12x3 lY-12x3 3-
= ys ) - - - - - -(15)
2
Yo I (1 _ I
Y Ix Y wy2=10800 wx 2=0
BC 24=12 30 0 0 360 . _ 24 3
A9.3 Equations for Structure with Symmetry About One Axis 2 iX=~3=
12x2
0 lY-12x2 =576
through Elastic Center.
wy2=2250 wx 2 =1440
CD 30=10 15 12 120 150 .
If the structure is such that either the x 3 lX= 750 iY=~3=0
12x3
or y axis through the elastic center ~s a axis
of symmetry than the product of inertia Sum 32 0 660 16800 3456
~xyds/EI =
I xy = zero. Thus making the term
I xy = 0 in equations 11, 14 and 15 we obtain ,
J1
0
= -~ 0s
~ds/EI
- - -(16)
_ L:0 Y The tenns i x and i y are the elastic moment

Xo - ~s - - - - - - - - - - - - - - (17) of inertia of each portIon of the frame about
its centroidal x and yaxes. Since I is con-
_ -~0sx stant over each portion the centroidal moment
Yo - ry- ------------- -(18) of inertia of each portion is identical to that
of a rectangle about its centroidal axis.
A9.4 Example Problem Solutions. Structures with at least
1- 1- b
One Axis of Symmetry. To explain for member AB:-
--:J f- 1-3-
Example Problem 1
10 lb.
Referring to Fig. a,
11 x-
T
x ds= 30= h
Fig. A9.7 ~" 18" --1 i x = 2 bh 3 = 12 x"3 x 30 3 = 750
12
shows a rectang-
ular frame with
B ,---_L-
I =2
C
1
fixed supports
at points A and
D, and carrying -x
y
I elastic -x
- -7- T i y =~bh3
Referring to Fig. b,
1 1
= 12 x 30 X-w = .09 (negligible)
A
Fig. a
1
c-ent 12
a single load as 30"
shown. The I y
problem is to de- The distance from the
I =3
termine the bend-
ing moment dia-
gra'll under this
'"3 i
two reference axes to the
elastic center can now be
calculated:-
I
I
I
I
T ds = 30 = b
loading. X'-~ I~ __ D _ I
1
' - _ L:wy _ 660 _
y - ~w - 32 -_20.625 in. I
The first Fig. A9.7 I
step in the
- _~wx 0 -j y 1-1I : 13 = h
solution is to x - ~w 32 - 0
Fig. b
find the location of the elastic center of the
frame and the elastic moments of inertia Ix Having the moment of inertia about axis x~x~ we
and I y . can now find its value about the centroidal
Due to symmetry of the structure about the axis xx of the frame, by use of the parallel
Y axis the centroidal Y axis is located midway axis theorem.
between the sides of the frame, and thus the Ix = Ix - ~W(y2) = 16800 - 32 x 20.625 2= 3188
elastic center (0) lies on this axis.
Iy = Iy - ~W(X2) = 3456 - 32(0) = 3456
Table A9.l shows some of the necessary
calculations to determine the location of the The problem now consists in solving equa-
elastic center and the elastic moments of tions (16), (17) and (18) for the redundants at
inertia. The reference axes used are x~-x~ the elastic center, namely
and y-y.
A9.5
_ - ~¢ _ Area of static M/I diagram along the neutral axis of the frame members.
Mo - ~ds/¥ - Total elastic weight of structure
In Fig. A9.10 the area of the Ms/I diagram
Moment of static M/I diagram about equals 0s = 22.5 x 24/2 = 270. The centroid by
= ~¢sy = x axis simple calculations of this triangle would fall
Xo Ix Elastic moment of inertia about x 10 inches from B. Fig. A9.11 now shows the
axis frame with its Ms/I or its 0s load. 0s is
Yo = -~¢sx
Iy
Moment of static M/I diagram about
= =-_:-:- ..,.y~a_:_x-i..,...s-_;_,...____:_-.,__-
Elastic moment of inertia about y
axis
r,z>s= 270 - I "
~ 1--2
B:
x-----l------ x
0:
C
l'
positive since Ms is
positive. The next
375" step is to solve the
equations for the
redundant at the
I 20.625" elastic center. The
Thus to solve these three equations we : I signs of the distances
must assume a static frame condition consistent A :y D ~ x and y from the axes
with the given frame and loading. In general I x and yare conven-
there are a number of static conditions that can Fig. A9.11 tional.
be chosen. For example in this problem we
might select one of the statically determinate ThUS,
conditions illustrated in Fig. A9.8 cases 1 to
5. _ -Z0 s = - (270) = -8.437 in. lb.
Mo - Zds/I 32 (from Table A9.1
_ Z0 Y _ 270(9.375)
Xo - Ixs - 3188 = 0.7939 lb.
Yo
=-Z0sx
I
= -(270)
3456
(-2)
= 0.1562 lb.
y
Fig. A9.8 Fig. A9.12 shows these values of the re-

dundants acting at the elastic center.
To illustrate the use of different static 28. 18

conditions, three solutions will be presented
with each using a different static condition. T
9.375
Solution No.1 J _
In this solution we will use Case 3 as the
static frame condition. The bending moment on
the frame for this static frame condition is
1
20.625"
L. -:A~-c07
given in Fig. A9.9. The equations 9.81
06 D
Bf-.------=' C,di""(1"'/1di,."m
B C Fig. A9.12
6.06
Fig. A9.13
M s is positive, tension The bending moments due to these redundant

on inside of frame. forces will now be calculated.
D
MA =- 8.437 - .1562 x 12 + .7939 x 20.625 =
t 2. 5# Fig. A9.10 6.06 in. lb.
MB =- 8.437 - .7939 x 9.375 - .1562 x 12 =

for the redundants require ¢s the area of the -17.75 in.lb.
Ms/I diagram. Fig. A9.10 shows the Ms/I curve
which is obtained by dividing the values in Me =- 8.437 - .7939 x 9.375 + .1562 x 12 =
Fig. A9.9 by the term 2 which is the moment of -14.00 in.lb.
inertia of member BC as given in the problem.
Since the equations for Xo and Yo require the MD =- 8.437 + .7939 x 20.625 + .1562 x 12 =
moment of the Ms/I diagram as a load about 9.81 in. lb.
axes through the elastic center of frame, the
area of the Ms/I diagram will be concentrated These resulting values are plotted on Fig.
at the centroid of the diagram and along the A9.12 to give the bending moment diagram due to
centerline of the frame, or more accurately the redundant forces at the elastic center.
Adding this bending moment diagram to the static _ -~0sx _ -[-45(-10)-405x6-300(-12)-90OX12]

bending diagram of Fig. A9.9 we obtain the final Yo - Iy - 3456
bending moment diagram of Fig. A9.13.
The final bending moments could also be

= -(-9180) = 2.656 lb.
3456
obtained by substituting directly in equation
(1) using subscript (0) instead of (A). Thus,
The final moments at any point can now be
found by equation (19).
M = Ms + Mo - XoY + Yox - - - - - - -(19)
For example, determine bending moment at Consider point B:-
point B.
Ms = -30 from Fig. A9.14
For point B, x = -12 and y = 9.375, Ms =0 x = -12, Y = 9.375
substituting in (19)
Subt. in (19)
MB =0 + (-8.437) - .7939 x 9.375 +
MB =-30+51.56-.7939x9.375+2.656(-12)
.1562 (-12) =
-17.75 as previously
= -17.75 in.lb. which checks first
found
solution.
AT rOINT D. x =12, Y =-20.625, Ms =O. Consider point D:-
MD =0 + (-8.437) - .7939 (-20.625) +
= -90, = 12, Y = -20.625
.1562 x 12 =
9.81 in.lb. Ms x
Solution No.2 SUbt. in (19)
In this solution we will use Case 4 (See MD = -90+51.56 7939(-20.625) + 2.656 x 12

Fig. A9.8) as the assumed static condition, = 9.80 in. lb.
that is two cantilever beams with half the
external load or 5 lb. acting on each canti- Solution No.3
lever. Fig. A9.14 shows the static bending
In this solution we will use Case 5
moment diagram and Fig. A9.15 the Ms/I diagram. (Fig. A9.8) as the assumed static condition,
Qs =-45 ~ =-405 namely a frame with 3 hinges at points A, D and
-15 -45 E, as illustrated in Fig. A9.16.
9I75 HO~"--I 10# Before the

~16" ~~18"--1
~- -~25:5.6f-
bending moment dia-
E gram can be calcu-
r- 12" lated the reactions
Fig. A9. 14 " L at A and Dare
IDs=-30~ Ps=- 00
fY necessary.
I -x 30"
I To find VD
Fig. A9. 16
!Y take moments about
-30 -90 -10 -30 point A.
Fig. A9.15
D
Fig. A9.15 also shows the results of
.. HD=l mA = lOX6 - 24VD
calculating the 0s value for each portion of +
VD=2.5
=0
the Ms/I diagram and its centroid location. hence, VD = 2.5
SUbstituting in the equations for the re-
dundants we obtain,
To find VA take ~Fy = 0 =-10 + 2.5 + VA
(-45-405-300-900) =0 hence VA =
7.5.
32 = 51.56 in.lb.
To find HD take moments about hinge at E of
all forces on frame to right side of E and
_ ~0sY _ (-45-405)9.375+(-900-300)(-5.625) equate to zero.
Xo - --r;- - 3188
mE = -2.5 x 12 + 30HD = 0, hence HD = 1.
= 0.7939 lb.
Then using ~Fx =0 for entire frame, we obtain
~Fx =-1 + HA =0, hence HA =
1.
A9.7
axes through the frame elastic center are then

calculated as -12 and -0.625 inches respect-
ively.
_ - Z0s _ - (-390)
Mo - Zds!I - 32
=12.19 in. lb.
X = Z¢sy -648 =-0 203

= 3188 Ib
o Ix . •
Yo = -z0s x
Iy
=- 3456
(-540) = 0.1562 lb.
The final moments at any point can now be

found by use of equation (19), namely
M =Ms + ~o - Xoy + Yox - - -(19)
Consider point B:-
x = -12, Y = 9.375, Ms = -30

SUbstituting -
MB = -30+12.19-(-0.203)9.375+0.1562(-12)
- -17.76 in.lb.(checks previous solu-
tions)
The moment diagram is labeled in 6 parts Consider point D:-

1 to 6 as indicated by the values in the small
circles on each portion. Most of the calcu- x = 12, y = -20.625, Ms =0
lations from this point onward can be done
conveniently in table form as illustrated in Subt. in (19)
Table A9.2.
MD = 0+12.19-(-0.203) (-20.625)+0.1562x12
Table A9. 2
= 9.82 in. lb. (checks previous solu-
1 2 3 4 5 6 7 8 tions)
Mom. A= x = dist. y = dist. The final complete bending moment diagram

1 ¢s = from Y from X
Dia- Area for would of course be the same as drawn in Fig.
gram of axis to axis to ¢sx ¢Sy
Beam A C. g. of C. g. of
A9.13 for the results of Solution No. I.
Por- Mom. 1
Section ¢S
tion Dia. ¢s
Example Problem 2.
1 -450 3 -150 -12.00 -0.625 1800 93
Fig. A9.18 shows a rectangular closed
2 - 60 2 - 30 -10.67 9.375 320 -281 frame supported at points A and B and carrying
3 15 2 7.5 - 6.67 9.375 - 50 70 the external loads as shown. The reaction at D
due to rollers is vertical. The frame at point
4 45 2 22. 5 - 4.00 9.375 - 90 211 A is continuous through the joint but the
reaction is applied through a pin at the center
5 -180 2 - 90 8.00 9.375 - 720 -844 of the joint. The problem is to determine the
6
bending moment diagram.
-450 3 -150 12.00 -0.625 -1800 93
Sum -390 - 540 -658 w = lO#/in. .-----(2)---,
B L = 30 C
In order to take moments of the ¢s values 1= 4
in column (4) of the table, the centroid of IY L=24
each portion of the diagram must be determined.
For example, the centroid of the two triangular
r
PI =20# L=24
1=3 __ -i-
1= 3
_-x
bending moment portions marked 1 and 6 is 12' YA=9t 67'~=3~' c.

.667 x 30 from the lower end or 20 inches as 1 J 1=2
shown in Fig. A9.17. Thus the distances x
Fig. A9.18
and y from this 0s location to the y and x
A9 8 BENDING MOMENTS IN FRAMES AND RINGS
The next step is to choose a static frame

condition and determine the static (M s ) bending
B C B c moment diagram. In this solution the frame is
assumed cut on member AB just above point A as
illustrated in Figs. A9.18, 19 and 20. For
(7)
~xl simplicity the static moment curve has been
drawn in three parts, with each part consider-
T
cut - cut PJ ing only one of the three external given loads
IA D lA D 360 on the structure. Figs. A9.18, 19, 20 show
these resulting bending moment curves. The
~-4500 '360
portions of these bending moment diaGrams
Fig. A9.19 Fig. A9.20 nl~bered 1 to 10 are shown in the parenthesis
M s diagram for w loading M s diagram for P2 load. on each portion.
The next step in the solution consists of

.Solution: finding the area of the Ms/I diagrams and the
first moment of these diagrams about the x and
The first step is to find the location of y axes through the elastic center. These simple
the elastic center of the frame. Due to calculations can b est be done in table form as
symmetry of frame about the y axis, the elastic illustrated in Table A9.3.
center will be on a y axis through the middle
Table A9. 3
of the frame. The vertical distance y
measured from a axis through AD equals, Portion Area Os = pc = dist. Y = dist.
of Ms of Ms I to y to X ¢Sy
Z(¥) y __ --.=-....,...",...,..,..._:=:-=-_-
3~X24+234x12X2
372 - 9.67"
Diagram Portion
( )
A
=A
Axis
T
Axis
Y
Z ds 30+(24)2+ 30 - 38.5 - 1 - 1440 3- 480 -15 10.33 7200 - 4958
I 4 3 2
The next step is to determine the elastic

2 - 7200 4- 1800 0 14.33 0- 25794
moment of inertia of the frame about x and y 3 - 1440 3- 480 15 10.33 - 7200 - 4958
axes through the elastic center of the frame.
4 14403 480 15 - 5.67 7206 - 2726
Moment of inertia about x axis = Ix :- 5 36002 1800 5 - 9.67 9000 - 17406
Members AB and DC, 6 - 45000 4 -11250 7.5 14.33 - 84375 -161212
Ix =( i x i X 14.33 3
) 2 = 653.9
7
8
-108000 3 -36000
- 67500 2 -33750
15
5
2.33 -540000 - 83880
- 9.67 -168750 326362

C~ x i x 9.67 3
) 2 = 200.9
9 32403 1080 15 - 3.67 16200 - 3964
Member BC and AD 10 54002 2700 5 - 9.67 13500 - 26109
Ix = (30/4)(14.33") = 1540.0 Sum -77700 -747225 - 4645

(30/2)(9.67") = 1402.7
Solving for the redundants at the elastic
= 3797.5 center,
Moment of inertia about y axis = I y .- Mo =~

Zds/I
= -(-77700)
38.5
= 2018 in.
Ib
•
Members AD and BC,
Xo = Z¢sy
Ix
= -4645
3797
= -1 . 22 Ib
•
I Y -- 12
1 30 3 = 563
1
12 30 3 = 1126 Yo = -Z~~x = -(-~~~~25) = 141.28 lb.
Members AB, CD, Fig. A9.21 shows these redundant forces

acting at the elastic center. Fig. A9.22
3600 shows the bending moment diagram due to these
redundant forces. The calculations with refer-
ence to Fig. A9.21 are -
= 5289
A9.9
1-----15 -----1 Total elastic weight of ring =

z~ =n (36) = 113.1
I 1
~0~.--1-' 2~ The elastic moment of inertia about x and y
114 1. 28 917 axes through center point of ring are the same
for each axis and equal
J,---'----J.,.l Ix = I y = nr = n
3
x 18 3 =18300
Fig. A9.21
The next step in the solution is to assume
a static ring condition and determine the static
MA = 2018 - 141.88 x 15 - 1.22 x 9.67 = -113 (Ms ) diagram. In general it is good practice
MB -- 2018 - 141.28 x 15 + 1.22 x 14.33 = -84 to try and assume a static condition such that
the Ms diagram is symmetrical about one or if
Me = 2018 + 141.28 x 15 + 1.22 x 14.33 = 4153 possible about both x and y axes through the
elastic center, thus making one or both of the
MD = 2018 + 141.28 x 15 1.22 X 9.67 = 4124
redundants Xo and Yo zero and thus reducing
considerably the amount of numerical calcula-
Combining the bending moment diagrams of
Figs. A9, 18, 19, 20 with Fig. A9.22 would tion for the solution of the problem.
give the true or final bending moment diagram.
In order to obtain symmetry of the Ms
Example Problem 3. Circular Ring. diagram and also the Ms/I diagram since I is
constant, the static condition as shown in
Fig. A9.23 shows a circular ring of Fig. A9.24 is assumed, namely, a pin at (e) and
constant cross-section subjected to a sym- rollers at (f). The static bending moment at
points (a), (b), (c) and (d) are the same
metrical loading as shown. The problem is to magnitude ann equal,
determine the bending moment diagram.
50 50# Ms =50(18 - 18 cos 45 0 ) =265 in. lb.
The sign is positive because the bending moment

produces tension on the inside of the ring.
The next step is to determine the 0s ' 0 s x

and 0 s Y values.
0s is the area of the Mg/I diagram, how-

ever since I is unity it is the area of the Ms
diagram. The static Ms diagram of Fig. A9.24
50 50# is divided into similar portions labeled (1)
Fig. A9.23 Fig. A9.24 and (2). Hence
0s(~) =area of portion (1) =

Pr 2 (a -sin a),
where P = 50 lb. and a = 45 0 •
Substituting and multiplying by 4 since there

are four portions labeled (1).,
e 0s(~) = 4[50 x 182 (0.785 - 0.707)J = 5052
The area of portion labeled (2) equals,
Fig. A9.25
Pr 2 e(1 - cos a).
Moments due to redundant
forces. Fig. A9.26 Since there are two areas (2) we Obtain,
Solution. Due to symmetry of the ring 0S (2)= 2[50 X 18 2 x ~ (1-0.707)J = 15000

structure the elastic center falls at the
center of the ring. Since the ring has been Hence, total 0 8 = 15000 + 5052 = 20052
assumed with constant cross-section, a relative
value of one will be used for I. Since the centroid of the Ms diagram due
to symmetry about both x and y axes coincides
with the center or el~stic center of the frame, Solution:

the terms z0s x and Z0s Y will be zero.
The first step is to determine the elastic
Substituting to determine the value of the center of the frame and the elastic moments of
redundants at the elastic center we obtain, inertia. Table A9.4 shows the calcUlations. A
reference axis x'-x' has been selected at the
midpoint of the side AB. Since a static frame
Mo =~~:'I :: - ~~~~~2) :: - 177 in. lb. condition has been selected to make the Ms
diagram symmetrical about y axes through the
X = z0s Y = 20052(0) = 0 elastic center (see Fig. A9.28), it is not
o Ix 18300 necessary to determine I y since the redundant
Yo will be zero due to this s~~etry.
Yo = -Z0s x = -(20052)(0) :: 0 Table A9.4
Iy 18300
Length ds
Fig. A9.25 shows the values acting at the Member ds I r.v=T y wy IX'X' = i x + wy2
elastic center and the bending moment diagram
produced by these forces. Adding the bending BDB' 94.25 1.5 62.8 2r + 30 3080 5400+151200 = 156600
moment diagram of Fig. A9.25 which is a constant n
value over entire frame of -177 to the static = 49.1
moment diagram of Fig. A9.24 gives the final
bending moment diagram as shown in Fig. A9.26. AB 60.0 1.0 60.0 0 0 18000 + 0 = 18000
Example Problem 4. Hull Frame A'B' 60.0 1.0 60.0 0 0 18000 + 0 = 18000
AC 38.4 2.( 19.2 -42 -806 922 + 33850 = 34772

Fig. A9.27 shows a closed frame subjected
to the loads as shown. The problem is to CA' 38.4 2.( 19.2 -42 -806 922 + 33850 = 34772
determine the bending moment di~gram.
Totals 221. 2 1468 262140
bOOO'li-
D
In the last column of Table A9.4 the term
i x is the moment of inertia of a particular
member about its o\Vll centroidal x axis. Thus
for member BDB;
D
static moment )( 1f..~, i x :: .3;" = .3 x 303 /1.5 :: 5400

curve
For members AC and CA ,

T A ~L=38.4
Fig. A9.28
""""~ ~~=~~-~:}Y ix =..1....
12
b L h2
= ;2x~X38.4X242=922
Let y = distance from X'X' ref. axis to

centroidal el~stic axis X-X.
- _ Zwy _ 1468 _
Y - Zw - 221,2 - 6.64 in.
By parallel axis theorem,
Ix = Ix' - 6.642 (Zw)

- 262140 - 6.642 (221.2) = 252400
The next step in the solution is to com-
pute the static moment elastic weights 08 and
Fig. A9.29 their centroid locations. In Fig. A9.28, the
Fig. A9.30 static frame condition assumed is a pin at point
A and rollers at point A' , which gives the
A9.1l
general shape of static moment curve as shown 38.4

in the fit':ure.
Consider member BDB' , x=
1/2
10o {4690 x 100 x")xdx
78400
The bending moment curve will be con- 38.4

sidered in two parts, namely (1) and (2). The
term 08 represents the area of the Ms/I
=1/2 [469~ x" _ 10~ xj = 21. 77"
diagram. o
78400
Thus for portion (1) and (1')
Vertical distance from line AA' to centroid =
0s (1.) + 0 s (1.' ) = 2~" (a - sin a] 21. 77 x 24/38.4 = 13.6". The stat ic moment
weight for A'C' is same as for AC, thus
= 2@00~.~ 30"(0.524 - 0.5J = 172800
0AC + 0A'C' =2 x 784000 = 1568000
The vertical distance from the line BB' to
centroid of Ms curve for portion (1) and (I' ) Fig. A9.29 shows the frame With the moment
is, weights 0 8 located at the centroids, together
with the redundant forces Mo and Xo at the elas-
sin"a tic center. It makes no difference where the
r(l - cos a - -2-) 30(1 - 0.867 - ~) frame is cut to form our residual cantilever,
y = a - sin a 0.524 - 0.5 if one of the cut faces is attached to elastic
center and the other is considered fixed. With
- 10 in. the elastic properties and moment weights known
the redundants can be solved for:
For portion (2) of the Ms diagram the area of (1007000 + 172800 + 1.56800) =
the Ms/I diagram which equals 08 is 221.2
0 s (") = prI"e (1 - cos a)= 6000xl~~"X2.1o.._0.B67) 12430"#
= 1,007,000 Xo = Z0I s Y =
xx
Consider member AC, _1007000 x 48.16 + 172800 x 33.36 + 156800 x -5.24
- 252400
From free body diagr&~ of bottom portion
of frame (Fig. A9.31J the equation for bending = -98.5#
moment 600~ /4690 6000 Yo is zero because of the symmetrical frame
'""tI
A -- A' and loading. The final or true bending moment
at any point equals
'}.I;)I;)~\~'
Fig. A9.31... Thus for point B
on member AC equals: MB =0 + (-12430) - (-98.5 x 23.36) =

-10130"#
Mx =4690 x -100 x"
For point C
Area of Mil curve between A and C when I =2
equals, Ms at point C = 4690 x 38.4 - 20OX(38.4)2/2
38.4 = 32700
1/2
10o
(4690 x - 100 X")dx
Hence Me = 32700 - 12430 - (-98.5 x -60.64)
14300"#
=
_ 14690 x" 100 x31 38.4 _ Fig. A9.30 shows the general shape of the true
- 1/2 L 2 - 3 J - 784000 frame bending moment diagram.
o
Example Problem 5
Distance to centroid of Mil curve along line
AC from A. Fig. A9.31 shows the general details of one
The moment of inertia of the frame

half of a symmetrical hull frame that was used cross sections are given in column
in an actual seaplane. The main external load 2 of Table A9. 6 which have been
on such frames is the water pressure on the hull determined from a consideration
bottom plating. The hull bottom stringers of the actual dimensions of the
transfer the bottom pressure as concentrations frame members.
on the frame bottom as shown. The resistance to
this bottom upward load on the frame is provided
by the hull metal covering which exerts tangen-
tial loads on the frame contour. The question
as to the distribution of these resisting
forces is discussed in later chapters. In this Centroidal
problem the resisting shear flow in the hull -~~-+----..... Elastic XX AXiS
sheet has been assumed constant between the
chine point and the upper heavy longeron. For
Elastic
Center j
Y 8.9"
Longeron
Assumed
analysis purposes the frame has been divided ReT.""Axis I - -
into 20 strips. The centroid of these strips X'X'
located on the neutral axis of the frame sec- 1"'.,1
tions are numbered 1 to 20 in Fig. A9.21. The llJ
tangential skin resisting forces are shown as S
C1l
concentrations on frame strips # 6 to # 16. On
the figure these tangential loads have been UJ
replaced by their horizontal and vertical com- rs

ponents. The sum of the vertical components llJ
10 -0
0-
should equal the vertical component of the 0
C
bottom water pressure. .....
fn 11 .u7
.....
.....
Table A9.5 shows the complete calculations -->
.....188
for determining the bending moment on the frame.
Columns 1 to 7 give the calculations for the
elastic properties of the frame, namely the
elastic weight of the frame; the elastic center
location, and the elastic centroidal moment of
inertia about the horizontal ~entroidal elastic Bottom member of frame
axis. A reference horizontal axis X'X' has
been selected as shown. All distance recorded
in the table have been obtained by scaling from
a large drawing of the frame.
The static condition assumed for computing

the Ms moments is a double symmetrical canti-
lever beam as illustrated in Fig. A9.32. The
frame is cut at the top to form the free end
of the cantilever beams, and the fixed end
has been taken at the centerline bottom frame
section. The static bending moment diagram
will be symmetrical about the y axis through
the elastic center of the frame and thus the
redundant Yo at the elastic center will be
zero since the term Z0s x will be zero. The 7670
calculations for determining the static
moments Ms in Column 8 of Table A9.5 are not
shown. The student should refer to Art. A5.9 Fig. A9.31
of Chapter A5 to refresh his thinking relative
to bending moment calculations on curved beams.
Columns 9 and 10 give the calculations of

the 0s values (area of MsfI diagram) and the
first moment (0s x values). The summations of
columns 3, 9, 10 permits the solution for the
redundants Mo and Xo as shown below the table.
The final bending moment M at any point on the Fig. A9.33
frame is by simple statics equals, Fig. A9.32
A9.13
Table A9. 5
Calculation of Frame Elastic Properties Calculation of Moment Weights and Solution of Redundants
Col. 1 2 3 4 5 6 7 8 9 10 11 12 13
.....0 Elastic Arm w y' w y,2 Arm Static Moment Os y Mo -Xoy *Total
<:: Weight to to Moment Weight Moment
0 0<:5
.,.. bJl
~C'l
<:: .~
.~
~ <:: (j)~ Ref. Axis Os = M
~
C'l U)~ 8 1;; W = ds Axis XX MS Msw

o <::
~
U) ::E ..... I X'X' =Y 1000
ds I = y'
1 7.6 .065 117.00 26. 1 3050 79700 17.2 0 0 0 -9020 16650 7630
2 7.6 .065 117.00 24.4 2850 69500 15.5 0 0 0 " 15020 6000
3 7.6 .065 117.00 21. 4 2497 53300 12.5 0 0 0 " 12100 3080
4 7.6 .065 117.00 17.3 2025 35020 8.4 0 0 0 " 8130 - 890
5 7.6 .065 117.00 12.0 1403 16820 3.1 0 0 0 " 3000 -6020
6 6.4 .25 25.60 5.9 151 892 - 3.0 1220 31. 2 - 93.7 " - 2910 -10710
7 6.4 .25 25.60 0 0 0 - 8.9 2880 71. 6 - 637 " - 8610 -14840
8 6.4 .25 25.60 - 6.3 -161 1017 -15.2 6230 159.3 - 2420 " -14720 -17510
9 6.4 .25 25.60 -12.6 -323 4060 -21. 5 11260 286.5 - 6160 " -20830 -18590
10 6.4 .25 25.60 -19.1 -488 9320 -28.0 18470 473.0 - 13220 " -27100 -17650
11 6.4 .30 21. 30 -25.5 -543 13850 -34.4 29020 618.0 - 21250 " -33300 -13300
12 6.4 .34 18.82 -32.0 -602 19250 -40.9 43180 814.0 - 33100 " -39700 - 5540
13 6.4 .34 18.82 -38.3 -720 27600 -47.2 58980 1110.0 - 52400 " -45600 4360
14 6.4 .34 18.82 -44.6 -840 57500 -53.5 77330 1455.0 - 77800 " - 51700 16610
15 6.4 .34 18.82 -50.8 -959 48750 -59.9 99030 1865.0 -111300 " -57900 32110
Chine -63.0 111700 -61100 41600
16 7.0 12.0 0.58 -55.4 - 32 1780 -64.3 149700 86.7 - 5570 " -62200 78500
17 7.0 12.0 0.58 -58.4 - 34 1980 -67.3 221400 128.3 - 8650 " -65100 147300
18 7.0 18.0 0.39 -61. 4 - 24 1470 -70.3 271200 105.8 - 7440 " -68100 194100
19 7.0 34.0 O. 21 -63.2 - 13 840 -72.1 308600 64.9 - 4680 " -70000 229600
20 7.0 50.0 O. 14 -64.8 - 9 588 -73.7 327900 45.9 - 3385 " -71400 247500
Keel -75.0 332900 -9020 -72600 251280
Totals 811.48 7228 423237 7315.0 -348106
Y= 7228/811. 48 : 8.9" M = -ZOs =_ (7315.7 x 1000) = -9020"#

o Zw 811.5
1XX = 423237 - 811. 48 x 8.9 2 = 358940
Xo=~= [ 34810635894
x 1000J = -968#
1XX
*Total Moment M at any station = Ms + Mo - XOy
240 lb
Columns 11 and 12 record the values of
Mo and -Xoy for each station point. For 6" ---1
example, the value of -Xoy for station (1)
T
1
C
equals - (-968 x 17.2) 16650 and for station =
(20) = - [-968(-73. 7)J -71400. = 6" Y=5.242 ,
Fig. A9.44 shows the shape of the final _l_L_x

E c
1=1 10"
moment curve as the result of the values in 120 I . .
1
column 13.
£ ~y
A9.5 Unsymmetrical Structures. Example Problem Solutions. 9" 5.032
D
Example Problem 1
Fig. A9.34
Fig. A9.34 shows an unsymmetrical frame
carrying the loads as shown. Determine the
bending moments at points A, B, C and D.
Solution:-
The elastic weight of frame = Zds/I =

(15/1) + (12/2) + 10/1 = 31
The distance x from the line AB to the Fig. A9.36 shows the centroid locations of
elastic center is, the 0 8 values along the center line of the
frame. The moment of these 0 s values about the
x = (15)0 + 6 x 6 + 10 x 12
31 = 5.032 in.
x and y axes will now be calculated.
Z0sx = (-2160)(-3.032) + (-21600)(-5.032) +

The distance y from line Be to elastic (-4860)(-5.032) = 139700
center equals,
Z0s Y = (-2160)(5.242) + (-21600)(-2.258) +
y = 15 x 7.5 + 6 x 0 + 10 x 5
31 - 5.842 in.
(-4860)(-6.758) =
70290
The values of the redundants at the elastic

The elastic moments of inertia Ix and I y center can now be calculated using equa~ions
and the product of inertia I xy are required. (11), (14), (15), namely
Ix = (11'25 )" + (15)(2.258)2 + (1 2)(5.242)2 + (10)"

2 12
11
0
=~
Zds/I
=- (-28620)
31
= 923 in. lb.
+ (10)(0.242)2 =606.51 ~
X
o
= Z03 y - Z0s x
----;".,----_v'--
Iy = (1.5) (5.032)2 + (12)" +(12) (0 968)2 +
2x12 2 • Ix (1 - i" fY
x y
(10) (6.968)2 = 942.96
70290 _ 139700 (217.74)
942.96
Ixy = (15) (-5.032) (-2.858) + (1:) (0.968) (5.242) =--------2=1=7:...:.:...:7:...:4'="2--- = 68.46 Ib.
606.51 (1 - (606.51)(942.96))
+ (10)(6.968)(0.242) = 217.74
__ - [Z0s x - Z0s Y (IfxY)J
The next step is to assume some static Yo
frame condition and draw the static bending I
y
(1 _J.xL)
1XIY
moment diagram. Fig. A9.35 shows that the
frame has been assumed cut near point C which
gives two cantilever beams. The bending
moment diagram in three parts for this static =_ r139700 - 70290 (217.74)J
L..= 606.51 = -132.36 lb.
condition is also shown on Fig. A9.35. 217.742
942.96 (1 - (606.51)(942.96)
¢Sl=-2160 The bending moment at any point from eq.
B J C
T (19) equals,
5.242
= 11s
t
11 + 110 - XoY + Yox, for example,
2.258 .C. x
Consider point A.
~:;=- 1600 6. 750
9.758
x = -5.032, Y = -9.758, 11s = -2520
j ¢S3=-4~60 11A = -2520 + 923 - 68.46 (-9.758) +
(-132.36)(-5.032) = -263 in. lb.
A 5.0~
Fig. A9.35 Fig. A9.36 Point B. x = -5.032, Y = 5.842, 11s = -1440
The 03 values which equal the area of
the 11s diagram divided by the I values of the 11B = -1440 + 923 - 68.46 x 5.242 + (-132.36)
particular portion will be calculated. (-5.032) = - 209 in. lb.
0s~ = 3(-1440)/2 = -2160 Point C. x = 6.968, Y = 5.242, 11s =0

0 S2 = -1440 x 15 = -21600 :Me =0 + 923 - 68.46 x 5.242 + (-132.36)
(6.968) = -357 in. lb.
0s" = 4.5(-1080) = -4860 ThUS, the complete bending moment diagram
Z0s =-2160 - 21600 - 4860 = -28620
could be determined by computing several more
values such as point D and the external load
points.
A9.15
Example Problem 2. Fig. A9.37 shows an un- Total Ix = 6451.07

symmetrical closed frame. The bending moment
diagram will be determined under the given Calculation of I y -
frame loading.
Member AB, I y = t~ X 15 2 + 1~ X 30(0.667)3 -
y'
4500.7
PI = 20# IB P2 = 100#
Member BC, I y =1~ x 31367 X 30 =791.7 2
I~ C Member CD, I y = 20
1 X 15 2 +.J:....
12 x 20(1)3 = 4501.7
30" ~ I 1 x 1 x 30 3
1'1emb er AD , I y --12 = 1125 •0
1=1.5 I I 2
1- --- -~E.
x'
Jj
Y
A
:3=50#
10"
C.-----X- 1=1 20"
P4=50#
10"
::lli
L x'
Total I y
Calculation of I xy -
= 10919
30
r:----30" D Member AB, I xy = 1':5 (-15) (3.343) = -1002.9
Iy ' Fig. A9.37
Member BC, 1
= 31 67(13.343)(0) - 12 31.67
Solution: -
I xy
3 x--3-
(30)(10) = -263.92
The elastic weight of frame equals
20
Member CD, I xy =1 (15) (-1.657) = -497.1
2: ~ = 30 + 31.67 + 20 + 30 = 65 58
I 1.5 3 1 2 .
Member AD, I xy = °
The location of the elastic axes will be
determined with reference to assumed axes Total I xy = -1763.9
x'x' and y'y' as shown on Fig. A9.37.
The next step in the solution is to assume
a static frame condition and draw the Ms dia-
___ (;.o~ (0)+ (31 ~67) (15)+ (210) (30)+ (32°) (15) _- gram. Fig. A9.38 shows the assumed static
x
65.58 15" condition, namely pinned at point A and supported
as rollers at point D. The bending moment
diagram is drawn in parts as shown.
(30) (15)+ (31.67) (25 )+(20) (10 )+(30) (0) 750"
- 1.6 3 1 2
Y = ---------=6=5-.5=8=--------- = ~uetoP2
11.657" 600~toPl
60
These distances x
and y locate the x and (1)
y elastic axes as shown in Fig. A9.37.
The elastic moments of inertia and the Due to PI c
elastic product of inertia will now be cal-
culated. Fig. A9.38
Calculation of Ix -
Member AB, Ix-~ x 1\(18.3433 + 11.657 3 )=1723.51 A

L----------D
Member CD, Ix=~ x i (8.343 + 11.657

3 3
) = 721.58
~
-50
1t-_ _
(4) ~
--,'<l'lTr~Due to
-500
P3 and P4
Member BC, Ix = (31.67) (13.343 2

) + _!.J 3 1.67)
3 12 3
(lOS ) =1967.48 The next step is to compute the value of
Member AD, Ix = (32°) (11.657 2
)
0s for each portion of the moment diagram. 0s
is the area of the Ms/I diagram. For reference
the portions of the Ms diagram have been labeled
(0.5) 3 = 2038.50 1 to 4.
x~= 6000 The bending moment at any point equals the

Portion (1) . ¢s = 600
2 1.5 original static Ms plus the moment due to the
redundant forces as shown in Fig. A9.40.
600 31.67
Portion (2) . ¢s = -2- x --3-= 3167
B
31.67 750 -165
Porti:m (3) 0 ¢s =-2- x --= 3958
3
1~43 E
124.8 T
C
Portion (4) . ¢s -- -500(lQ)
2
-500(l:2..)
2 = -5000 ~ 8 343
--- i! ·31. 07-1
-'-~
Z¢S = 13125 - 5000 = 8125 - Al---
4.674
D
11. 657
L
The individual ¢s values are now con- ~1511 15'~
centrated at the e.g. of each of the Ms dia- Fig. A9.40 Fig. A9.41
grams. Fig. A9.39 shows the location of the
0s values with respect to the x and y axes Consider point A: - Ms =0
through the elastic center.
MA = 0 - 124.8 - 4.674 x 15 + 31.07 x 11.657
= 167 in. lb.
~ ~=3167
i I 0 s =3958 Point B. Ms =600
15 13.343
MB = 600 - 124.8 - 4.674 x 15 - 31.07 x 18.343
8 .. 343 1 I = -165 in.lb.
~1+~·-1~ Point C. Ms =0
L 11 657, JI0 ...
I
1
$ __
Me =0 - 124.8 - 31.07 x 8.343 + 4.674 x 15
!
r-15'~
s =5000 = -314 inolb.
Fig. A9.39 Point D. Ms =0
Z¢sx = 6000(-15) + 3167(-5) = -105835 MD =0 - 124.8 + 4.674 x 15 + 31.07 x 110657
Z¢sy = 6000(8.343) + 3167(15) +

= 307 in. lb.
3958(13.343) -5000(-11.657) = 208659 Fig. A9.41 shows the true bending moment
diagram.
The values of the redundants at the elastic
center can now be calculated. A9.6 Analysis of Frame with Pinned Supports.
M
II O
= - Z¢s = -(8125)
Zds!I 65 ..58
= -124 • 8 in .lb.
Fig. A9.42 shows a rectangular frame and
loading. This frame is identical to example
problem 1 of Art. A9.3, except it is pinned at
points A and D instead of fixed.
_ Z¢s y - Z¢sx (~)
Xo 2
The first step will be to determine the
Ix (1 -~) elastic weight of the frame, the elastic center
x y
location and the elastic moments of inertia
208659 - (-10583:5) ( l;g:i:) _ about axes through the elastic center.
= [(-1763.9 2 )
6451 1 - 10919 x 6451
31.07 lb. J The term ds/EI of a beam element of length
ds represents the angle change between its two
end faces when the element is acted upon by a
= - [Z¢sx - Z¢sy (~U unit moment. In this chapter this term has
been called the elastic weight of the element.
Yo 2
I (1 _...!xL) Physically, the elastic weight is the ability
Y IxI y
of the element to cause rotation when acted
-[-105835 208659C1763.9] upon by a unit moment. When a unit moment is
applied to a rigid support, the support suffers
= [(_1763.9624)51J = 4.674 lb. no rotation since the support is rigid, there-
10919 1 10919 x 6451 fore a rigid support has zero elastic weight
and therefore does not figure in the frame
elastic properties.
A9.17
If a support is pinned or hinged it has

no resistance to rotation and thus a unit
moment acting on a hinge would have infinite -14.4~4.46
angle change or rotation and therefore a hinge
or pin possesses infinite elastic weight. ,30.54 ,
10 lb.
~S=270 0.482
6" 18" ,
B 1=2 C
5
10' ~ Fig. A9.45 Fig. A9.46
1=3 Fig. A9.45 shows the bending moment dia-
L=30
MJ ~I
f{o gram due to the redundant Xo ' Adding this
diagram to the original static diagram gives
A D --;d~C.- the final bending moment curve in Fig. A9.46.
y
Fig. A9.42 Fig. A9.44 Fig. A9.43 A9.7 Analysis of Frame with One Pinned and One Fixed
Support.
Due to symmetry of structure about the
centerline y axis the elastic center will lie Fig. A9.47 shows the same frame and loading
on this axis. Since the two hinges at A and as in the previous example but point D is fixed
B have infinite elastic weight, the centroid instead of hinged.
or elastic center of the frame will obviously
The support D
lie midway between A and B. Fig. A9.43 shows
the elastic center 0 connected to the point A has zero elastic
by a rigid bracket. weight and the pin
at A has infinite
~ds/EI for frame is infinite because of
elastic weight,
Fig . A9.47 therefore the elastic
the hinges at A and B.
The elastic moment of inertia about a y

axis through elastic center is infinite since
x-A jY
--x D
center of the frame
lies at point A.
The total elastic
weight of frame is
the hinge supports have infinite elastic weight.
infinite because of
Ix is calculated as follows: - pin at A.
The elastic moments of inertia will be

For Portion AB =~ x ~ X 303 = 3000 calculated about x and y axes through A.
1 1 =
For Portion CD ="3 x 3" x 303
3000 Ix = 16800 (Same as previous example)
Iy = (30 x 24 )+ (1:.. xJ:... x

2
243 = 8064
= 24 x ~
)
For Portion BC X 302 :: 10800 332
Ix =16800 I xy = (330 x 15 x 24) + (2: x 30 x 12) = 7920
Fig. A9.44 shows the static frame con- The static frame condition will be assumed
dition assumed to obtain the Ms values. the same as in the previous example problem,
fr"./J
hence 0 s = 270 and acts 10" from B (Fig. A9.48).
The value of 0s for member BC equals the
area of the Ms curve divided by I for BC, /Os=270
hence 0s 45 x 24 x 1= 270. The centroid = 10
2"" 2" -18.11
of this 0s value is 10 inches from point B.
The redundant forces at the elastic center can
now be solved for 31°" Mo
A Xo
M
o =- ~
~0s
ds
=-:;-:-:--",..2,..,,7:-:;0=
infinity
- 0 Iyo 6.19
Fig. A9.50
EI Fig. A9. 48 Fig. A9.49
Yo = -~0sx
Iy
=-(270)10
infinity
=0 Since the frame is unsymmetrical, the x
and y axes through the elastic center at A are
not principal axes, hence
X = ~0sY = 270 x 30 = 0.482 lb.
o Ix 16800
I Y = - Z¢sx p -(21)
Z¢sy - Z¢sx ( tSr) op I yp
Xo = I 2
IX (1 -~)
XY Where Ixp and I y are the elastic moments of
270 X 30 - 270 X 10 (7920) inertia about th~ principal axes and yp and x p
8064 the perpendicular distance from the ¢s loads
= 7920 2 = 0.6038 lb.
to the principal axes.
16800 (1 - 16800 X 8064)
The author prefers to use the longer
equations (14) and (15) which avoid finding the
- [Z¢sX - Z¢sy (Itxy)J principal axes location, the properties about
Yo = these axes and the perpendicular distances
I y (1 -~)
IxI y with respect to these axes, thus decreasing the
numerical work required.
7920 ~ _
- [ 270 X 10 - 270 X 30(16800)
= 7920 2 .2582 lb. The student Should use equations (19),
8064 (1 - 16800 X 8064) (20) and (21) to solve example problem 1 of
n N
Art. A9.5 in order to compare the two methods
of solution.
Mo = Zds!EI
- Z¢s = -270 0
--:i-n"""""f7in--;-it=-y-= A9.9 Problems. 200
00 00 20#/in.
Fig. A9.49 shows the redundant values at 110" 0'1 ,15' .
point A. The bending moments at the points I: 3 ~ 5" 150
B, C and Dare, L=20" L=20"! 100
1= 1 I. 1 ... 6"
MB = -.6038 X 30 =-18.11 in.lb. Fig.A9.52 Fig.A9.53
1 Fig.A9.54
Me =-18.11 + .2582 X 24 =-11.92 (1) Figs. A9.52, 53 and 54 show the same frame
rith three different loadings. Determine bend-
MD = .2582 X 24 = 6.19 !ing moment diagram for each loading.
100100100 w=8#/in. 20#/in.
Adding the static moment curve to that of 16"6" "6"1
Fig. A9. gives the final moment curve of L-21
1"~
Fig. A9.50.
~:'f ~:': 20
L=2'f 9" 9"
A9.8 Solution of Unsymmetrical Frames Using Principal r-'f .::L -*-
Axes. -16"
150 1 0
6" I-- --.18"
96
8"
96
~ 0
•
6"L
I
For the frame of A9.51 the axes X and y
through the elastic center are not principal Fig. A9. 55 Fig. A9.56 Fig. A9.57
axes because the elastic product of inertia (2) Figs. A9.55, 56 and 57 show same frame With
y 3 different loadings. Determine bending moment
I diagram for each loading.
100
/ yp
R1' 4,,-----J4, R2
Fig. A9.51 Fig. A9.52 bending moment diagram for
I xy is not zero. However, for the principal in Figs. A9.57 and A9.58.
axes x p and yp as shown, the product of inertia
is zero. Thus if the redundant forces at the 100 100
elastic center are referred to the principal
axes as illustrated in Fig. A9.52, then the
same form of equations for the redundants
result as for symmetrical structure, namely
~50
W
lliill
- Z¢s
r8+14''t8',
Mo = Zds!I (Same)- -(19) Fig.A9.60 Fig.A9.61
(4) Determine the bending moment diagram for the

Xo •
p
Z~SYp
xp
- - - - - - - - - - - - -(20) frames and loadings as shown in Figs_ A9.59 and
A9.60.
CHAPTER AlO
STATICALLY INDETERMINATE STRUCTURES
SPECIAL METHOD - THE COLUMN ANALOGY METHOD
A10.l General. The Column Analogy* method is a Now assume we have a frame whose centerline
method that is Widely used by engineers in de- length and shape is identical to that of the
termining the bending moments in a bent or column section in Fig. AIO.l. The width of each
ring type structure. The method considers portion of this frame will be proportional to
only distortions due to bending of the llEI of the member. Fig. AIO.2 shows this
structure. assumed frame. Furthermore, assume that end B
of the frame is fixed and that a rigid bracket
The numerical work in using the column is attached to the end A and terminating at
analogy method is practically identical to point (0) the elastic center of the frame. The
that carried out in applying the elastic center frame is subjected to an external loading,
method of Chapter A9. w~, wa , etc.
A10. 2 General Explanation of Column Analogy Method. Wa
Fig. AIO.l shows a short column loaded in

W~
compression by a load P located at distances
(a) and (b) from the principal axes x and y of
the column cross-section.
Fig. A10.2
Fig. A10.l
This cantilever structure will suffer bend-
ing distortion under the external load system
w~, wa , etc., and point (0) will be displaced.
1I1I )/ltl/"/llf,'I,
Point (0) can be brought back to its original
undeflected position by applying a couple and
two forces at (0), namely, Mo , Xo and Yo as
shown in Fig. AIO.2. Since point (0) is attached
To find the bearing stress between the to frame end A by the rigid bracket these three
supporting base and the lower end of the forces at the elastic center (0) will cause
column, it is convenient to transfer the load point A to remain stationary or in other words
P to the column centroid plus moments about to be fixed. Therefore, for the frame in Fig.
the principal axes. Then if we let cr equal the AIO.2 fixed at A and B, the moment and two forces
bearing stress intensity at some point a acting at the elastic center cause the static-
distance x and y from the yy and xx axes, we ally indeterminate moments Mi when resisting a
given external loading causing static moments
(;iJ,p wr+t.l'l ,II tN·S",· ::-Ttlc ,~JjoInaJ1," c!"ue~oBt1u,nI@:"Iflultren:iJ :1:
= =.......,;;==========_===:::p=+=-{:;=:'p==a:=;)=y=._=+=::;=(P==..b=;J""X=.==========-115n:",,,·-n:e:¥ti'j"~'fdrr=eq'.,;,a.J[j-T' -=====
i cr-A---r-x----Iy-
Where A is the area of the column cross-

section and Pa and Pb the moment of the load From Fig. AIO.2 we can write fo
P about the xx and yy axes respectively. If on the frame such as B that the inde
we let pa = Mx and Pb =
My, the above equation bending moment Mi equals,
can be written,
Mi =Mo +XoY+Yox
- - - (1) In Chapter A9, Art. A9.3, the e
for Mo, Xo and Yo were derived. The
*Method of Analysis due to Prof. Hardy Cross. See "The
Column Analogy" Univ. Ill. Eng. Expt. Sta. Bull. 215.
A10.l
A10.2 THE COLUMN ANALOGY METHOD
- ~ (4) The final or true bending moment M at any

Mo - zds/I ' YO =- Z0sx X - Z0s Y
Iy , 0 - IX point then equals,
The term Z0s represents the area of the - - - - - - - - - - - (5)

static Ms/I curve. (E has been assumed con-
stant and therefore omitted). Let the term Example Problem 1.
z0s be called the elastic load and give it a
new symbol P. The term Zds/I equals the Fig. AIO.3 shows a rectangular frame with
elastic weight of the frame and equals the sum fixed supports at points A and. D. The bending
of the length of each member times its width moments at points A, B, C and D will be de-
which equals III. Let this total frame elastic termined by the column analogy method. This
weight be given a new symbol A. frame and loading is identical to example
problem 1 of Art. A9.4 of Chapter A9 where the
In the expressions for Yo and Xo the solution was made by the elastic center method.
terms Z0s x and Z0s Y represent the moment of
the static Mil curve acting as a load about I 10 lb. 45"#
the y and x axes respectively passing through
the frame elastic center. Therefore let z0s x --j 6" 18"--1
be given a new symbol My and Z0s Y a new B C
I =2
symbol Mx ' With these new symbols, equation L = 24
(2) can now be rewritten as follows: -
L=30" L=30"
=3 1=3
Mi =.£ + MxY + Myx __ - - (3)
A Ix Iy
A D
Comparing equations (1) and (3) we see
7.5
they are similar. In other words, the inde- 2.5
terminate bending moments Mi in a frame are Fig. A10.3
Fig. A10.4
analogous to the column bearing pressures cr,
hence the name "Column Analogy" for the method SOLUTION NO. 1
using equation (1). With this general ex-
planation, the method can now be clearly ex- We first consider the frame centerline
plained by giving several example problem shape as shown in Fig. AIO.3 as the cross-
solutions. section of a short column (see Fig. AIO.5). The
Width of each portion of the column section is
A10.3 Frames with One Axis of Symmetry. equal to l/EI of the member cross-section.
Since E is constant, it will be made unity and
From the previous discussion we can write, the widths will then equal III.
(1) The cross-section of the analogous column 270

consists of an area, the shape of which
is the same as that of the given frame
and the thickness of any part equals
l/EI of that part.
(2) The loading applied to the top of the

analogous column is equal to the Ms/EI
diagram, Where Ms is the statical moment
in any basic determinate structure de- A
rived from the given frame. If Mg
causes bending compression on the inside
face of the frame it is a positive bending
moment and the analogous load P on the Fig. A10.5 Fig. A10.6
column acts downward.
The first step in the calculations is to
(3) The indeterminate bending moment Mi at a compute the area (A) of the column cross-section
given frame point equals the base in Fig. AIO.5 and the moments of inertia of the
pressure at this same point on the column cross-section about x and y centroidal
analogous column. Thus the indeterminate axes.
moment at any point on the frame equals ds 30 30 24_
(from eq. 3), Area A = Z T ="3 + 3"" + 2 - 32
Mi =~+ MxY + Myx (4) The calculation of the location of the

A Ix Iy centroidal axes and the moments of inertia Ix
A10.3
and I would be identical to the complete Mx = 270x9.375 = 2530 (positive because

calculations given in Art. A9.4, and Table A9.1 base pressure is compressive on column portion
where this same problem is solved by the elastic above x axis).
center method. These calculations will not be
repeated here. The results were, Ix = 3188 My = - (270 x 2) =-
540 (negative because
and I y =
3456. base pressure is tension on column portion to
right of yaxis).
Since the frame is statically indetermin-
ate the next step is to assume a static frame Substitution in equation (3).
condition consistent with the given frame and
loading. Fig. AIO.4 shows the condition Frame Point A. x =- 12", Y = - 20.625"
assumed for this solution, namely, pinned at
point A and a pin with rollers at point D. Mi =~ + MxY + MLX
The static Ms diagram is therefore as shown in A Ix y
Fig. AIO.4. we now load the column cross-
section with the Ms/I diagram as a load as Mi = 270 + 2530 (-20.625) + (-540)(-12)
shown in Fig. AIO.5. The static moment sign 32 3188 3456
is positive because the static condition causes
tension on the inside face of the frame. In = 8.44-16.38+1.88 =-6.06 in.lb.
the column analogy method a positive Ms/I The true bending moment from equation (5),
loading is a downward or compressive load on
top of column, and therefore a negative Ms/I
value would be an upward or tension load on
the column. Ms = 0, see Fig. AIO.5
Equation (3) requires the values of the
Mx and My the moment of the Ms/I diagram as a whence, MA = 0- (-6.06) = 6.06 in.lb.
load about the x and y axes. Equation (3)
also requires the total column load P which Frame Point B. x = - 12", Y = 9.325"
equals the area of the Ms/I diagram.
= 270 + 2530 x 9.375 + (-540) (-12)
Mi 32 3188 3456
For this problem the value of P from
Fig. AIO.5 equals, = 8.44 + 7.44 + 1. 88 = 17.77
P = 22.5 x 24/2 = 270 MB = Ms-M i =0-(17.77) =-17.77 in.lb.
The centroid of this triangular loading Frame Point C. x = 12, Y = 9.325
is 2 inches to left of y axis. Fig. AIO.6
shows the column sectioll with this resultant = 270 + 2530 x 9.325 + (-540)12
load P. Mi 32 3188 3456
We now use equation (3) to find the in-

= 8.44 + 7.44 -1.88 = 14.0
determinate moments Mi which are equal to the
base pressures on the column. Equation (3) Me =Ms -Mi = 0 - (14.0) = -14.0 in.lb.
involves bending moments Mx and My and
distances x and y, all of which must have
Frame Point D. x = 12, Y =-20.625
signs. The signs will be determined as follows:
Mi = 270 + 2530(-20.625) + (-540)12
32 3188 3456
When moment Mx produces compression on
base on that portion above x aXiS, then Mx is = 8.44 - 16 .38 - 1. 88 = - 9.82
positive.
MD =Ms -Mi = 0- (-9.82) = 9.82 in.lb.
When moment My produces compression on
base on that portion to right of y aXiS, then Fig. AIO.7 shows the 28.17
My is positive. final bending moment dia- -17.77
gram, which of course -14
A distance y measured upward from x axis checks the solution by the
is posit-ive; measured downward is negative. elastic center method in
Art. A9.4. The stUdent
A distance x measured to right from y should note that the
axis is positive, to left negative. numerical work in the
column analogy method is
From Fig. AIO.6: - practically the same as
in the elastic center
P = 270 method. Fig. A10.7
A10.4 THE COLUMN ANALOGY METHOD
Solution No.2 MA = Ms -Mi =-30- (-36.06) = 6.06 in.lb.

which checks solution 1.
In this solution a different static frame
condition will be assumed as shown in Fig.
AIO.8, namely the frame is cut under the load Frame Point B. x =- 12, Y = 9.375
and one-half the 10 lb. load will be assumed
as going to each cantilever part. Fig. AIO.9 Mi = -1650 + 2531 x 9.325 (-9180) (-12)
32 3188 + 3456
shows the static moment diagram and Fig.
AIO.IO the static Ms/I diagram with centroid = - 12.23
locations of each portion of diagram which
MB = Ms-Mi =-30-(-12.33) =-17.77 in.lb.
Frame Point C. x = 12, Y =9.375
Mi =- 1650 + 2531 x 9.375 (-9180) 12
32 3188 + 3456
=-76.0
-30 -90 Me = Ms -Mi = - 90 - (-76.0) = -14.0 in.lb.
Fig. A10.9
45 Frame Point D. x = 12, Y = -20.625
Mi =-1650 + 2531(-20.625) + (-9180)12

32 3188 3456
(4) = -99.82
(1)
-10 -30 A MD =Ms -Mi = - 90 - (-99.82) = 9.82 in.lb.

Thus solution 2 checks solution 1. The
stUdent should solve this problem using other
Fig. A10. 10 Fig. A10. 11
static conditions.
are numbered (1), (2), (3) and (4). The area A10.4 Unsymmetrical Frames or Rings.
of each of these portions will represent a
In applying the column analogy method to
load P 1 , P s , etc. on the column in Fig. AIO.ll. unsymmetrical frames and rings, the moment of
Since the static moment is negative on each
portion the load on the column section will be the M/EI diagram must be taken about principal
axes and the moments of inertia with respect to
upward. principal axes.
P1 =-10(30) =-300, P" =-45x9 =-405 However, as explained for the elastic cen-
ter method in Chapter A9, the moments and
Ps =- 15 (6) /2 =- 45 , P4 = - 30 x 30 = - 900 section properties with regard to centroidal
axes can be used if Mx , My, Ix and I are modi-
P = L:P =- 300 - 45 - 405 - 900 =- 1650 fied to take care of the dis-Symmetr~ of the
From Fig. AIO.ll structure. In Art. A9.2 it was shown that the
redundant forces at the elastic center to un-
symmetrical frame sections was, (see equations
Mx = - (45 -\- 405) 9.375 + (300 + 900) 5.635 11, 14, 15 of Art. A9.2).
=
2531
Mo --~-p(
L:dsjEI - T same as for symmetrical
My = 300 x 12 + 45 x 10 - 405 x 6 - 900 x 12 frame)
= -9180
L:.0sY - L:.0s x (~ )
Mi = ~ + MxY + Myx X
o
= ----.,_--=I:.ly~- - - - - - - - (a)
A Ix Iy Ix (1 -~)
IxI y
POINT A. x = - 12, Y =- 20.625
Mi = -1650 + 2531(-20.625) (-9180)(-12) Yo = L:.0sx - L:.0s Y ( ~) - - - - - - - (b)
32 3188 + 3456
I y (1 -~)
= -36.06 IxI y
AlO.5
As previously done in Art. AIO.2 let, SOLUTION:

Fig. AIO.13 shows the cross-section of the
analogous column. The length of each member of
Furthermore, let, the column section being the same as in Fig.
AIO.12, and the ",'i0th of each portion being l/I.
The first step in the solution is to find the
My = My (2r3-) and M~ = Mx (~) column section properties.
y x
and k = (1 -~)
x y
The total area (A) of column section equals
(15/1) + (10/1) + (12/2) 31. =
Then sUbstituting in equations (a) and (b) The other properties required are: - Ix'
we obtain, I y and I xy •
The calculations to determine the location
of the centroidal x and y axes and the above
properties about these axes would be identical
From equation (2) we have, to that given in example problem 1 of Art. A9.5
of Chapter A9, where this same problem was
solved by elastic center method. The results
were Ix =
606.51, I y =
942.96 and Ixy 217.74.
The location of the axes were as shown in Fig.
=
Substituting values of Xo and Yo into this
equation, we obtain as the equation for Mi the AIO.13.
indeterminate moment in the column analogy
method, the following - The next step in the solution is to choose
a static frame condition and determine the
Mi =~+ (Mx-My)y + (My-Mx)x (6) static (Ms ) diagram. Fig. AIO.14 shows the
A klx kl y assumed static condition, namely, two cantilevers
because of the frame cut as shown. The figure
The true moment is the same as for the also shows the static bending moment diagram
symmetrical section, namely, made up of three portions labeled (1), (2) and
(3). The area of each portion of the moment
diagram divided by the (I) for that frame por-
tion will give the loads P on the column.
Thus the solution of an unsymmetrical
frame by the column analogy method follows the
same procedure as for a symmetrical section 240
(1)
except that equation (6) is used instead of
equation (3).
~
cut C[
~
AlO. 5 Example Problem - Unsymmetrical Section. (2)
Fig. AIO.12 shows a loaded unsymmetrical - x

frame fixed at points A and D. Required, the 120---"-1
true bending moments at points A, B, C and D.
This problem is identical to example problem
1 of Art. A9.5 where a solution was given by
D
the elastic center method.
240lb
Br--6 " 4c Fig. AlO.14
T
6"
1=2
6"
T
120# 1 =1 10" x
1
D
P 1.
P2
=- 1440 x 3/2 = -
=- 1440 x 15/1 = -21600
2160
A P3 =-1080 x 4.5/1 = - 4860

Fig. AlO.12
Fig. AlO.13 2;P =- 28620
AlO.6 THE COLUMN ANALOGY METHOD
These loads act on the centerline of the Frame Point C. x = 6.968, Y = 5.242
frame members and through the centroid of the
geometrical moment diagram shapes. These Mi = - 923 + 68.37 x 5.242 + 132.36 x 6.968
centroid locations are indicated by the heavy
dots in Fig. AIO.14 and their locations are =357.7
given with respect to the centroidal x and y
axes. The loads P~, P2 and P3 are now placed Me =Ms - Mi = 0 - (357.7) =- 357.7 in .lb.
on the column in AIO.13, acting upward because
they are negative. Frame Point D. x = 6.968, Y =-4.758
We now find the moments Mx and M which Mi =-923+68.37 (-4.758) + 132.36 x6.968
equal the moments of the loads P about the = -326
centroidal axes.
MD =Ms -Mi = 0 - (-326) = 326 in.lb.
Mx = - 2160 x 5.242 + 21600 x 2.258 + 4860
x 6.758 = 70290 The above results check the solution of
this same problem by the elastic center method
My = 2160 x 3.032 + 21600 x 5.032 + 4860 in Art. A9.5 of Chapter A9. The student should
x 5.032 139700= solve this problem by choosing other static
frame conditions.
To solve equation (6) we must have the
terms ~, M; and k. AlO.6 Problems
M' = Mx (~) = 70290 x 217.74/606.51 (1) Determine the bending moment diagram for
'j{ Ix the loaded structures of Figs. A10.15 to
= 25234 AIO.20.
400 400
My =My (~) = 139700 x 217.74/942.96 ~15T15t151 w= 10 lb/in.
I y L= 45"
= 32258 I =3
L=30" L= 30"
I =2
k = (1 _ Ixy 2 ) = 217.74 2
(1 - 606.51 x 942.96)
I =2
IxI y
= .9171 Fig. AlO.15 Fig. AID. 16
SUbstituting values in equation (6) we obtain, 200 300 lb.
-28620 + (70290 - 32258 )y + ~

31 0.9171 x 606.51 r 5"
V'\.() 1"'10
1-
(139700 - 25234)x 100 100
20 L = 20 L = 20
0.9171 x 942.96 I = 5
I = 5
whence, "--+-----;-_-=----.-_--11
8~ f.----24" ~
Mi =-923+68.37Y+132.36x - - - - - (7) 100 100
Fig. AID. I? Fig. AID. 18
For Frame Point A. x = - 5.032, Y =- 9.758
100 100 400
Mi = - 923 + 68.37(-9.758) + 132.36(-5.032)
~ ~ i
=- 923 - 667.15 - 666 •0 =- 2256
=
MA = Ms -Mi -1440 -1080 - (-2256) =
-264 in.lb.
For Frame Point B. x =- 5.032, Y = 5.242

200 lb. Fig. AID. 20
Mi = - 923 + 68.37 x 5.242 + 132.36(-5.032)
=- 923 + 358.39 - 666.0 =-1230.6 Fig. AID. 19
MB = Ms-M~ =-1440- (-1230.6) = (2) Solve problems (2) and (3) at the end of
-209.4 in.lb. Chapter A9, Art. A9.9.
CHAPTER All
CONTINUOUS STRUCTURES - MOMENT
DISTRIBUTION METHOD *
Al!.l Introduction. The moment distribution set up in the two beams at B. The question is
method was originated by Professor Hardy Cross.** how much of this moment is developed by each
The method is simple, rapid and particularly beam. The physical condition which establishes
adapted to the solution of continuous structures the ratio of this distribution to the two beams
of a high degree of redundancy, where it avoids at B is the fact that the B end of both beams
the usual tedious algebraic manipulations of must rotate through the same angle and thus the
numerous equations. Furthermore, it possesses unbalanced moment of 270 will be distributed be-
the merit of giving one a better conception of tween the two beams in proportion to their abili-
the true physical action of the structure in ty of resisting the rotation of their Bend thru
carrying its loads, a fact which is usually a common angle. This physical characteristic of
quite obscure in some methods of solution. a beam is referred to as its stiffness. Thus let
The method of procedure in the Cross method it be considered that the stiffness factors of
is in general the reverse of that used ~n the the beam BA and BC are such that 162 is distribut-
usual methods where the continuous structure is ed to BC and - 108 to BA as shown in Fig. All.4.
first made statically determinate by removing (The question of stiffness factors is discussed
the continuous feature and the value of the rein a following article).
dundant then solved for which will provide the Referring to Fig. All.4 again it is evident
original continuity. In the Cross method each that when the elastic curve rotates over joint B
member of the structure is assumed in a definite that it tends to rotate the far ends of the
restrained state. Continuity of the structure beams at A and C, but since these joints are
is thus maintained but the statics of the struc- fixed, this rotation at A and C is prevented or
ture are unbalanced. The structure is then moments at A and C are produced. These moments
gradually released from its arbitrary assumed produced at A and C due to rotation at B are re-
restrained state according to definite laws of ferred to as carry-over moments. As shown by
continuity and statics until every part of the the obvious curvature of the elastic curves
structure rests in its true state of equilibrium. (Fig. All.4), the carry-over moment is of oppo-
The general principles of the Cross method site sign to the distributed moment at the ro-
can best be explained by reference to a specific tating end. The ratio of the carry-over moment
structure. to the distributed moment, referred to as the
Fig. All.l shows a continuous 2 Span beam. carry-over factor, depends on the physical prop-
Let it be required to determine the pending mom- erties of the beam and the degree of restraint
ent diagram. We first arbitrarily ass~ that of its far end. (Carry-over factors are discus-
each span is completely restrained against rota- sed in a following article. For a beam of con-
tion at its ends. In the example selected ends stant section and fixed at the far end, the car-
A and C are already fixed so no restraint must ry-over factor is -1/2). In figure All.4 a fac-
be added to these points. Joint B is not fixed tor of -1/2 has been assumed which gives carry-
so tnis joint is imagained as locked so it can- over moments of 54 and -81 to A and C respective-
not rotate. The bending moments which exist at ly. To bbtain the final end moments we add the
the ends of eacn member under the assumed condi- original fixed end moments, the distributed bal-
tion are then determined. Fig. All.2 shows the ancing moments and the carry-over moments as
moment curves for this condition. (For calcula- shown in Fig. All.4. With the indeterminate
tion and formulas for fixed end moments see fol- moments thus determined, the question of shear,
lowing articles). Fig. All.3 shows the general reactions and span moments follow as a matter
shape of the elastic curve under this assumed of statics.
condition. It is noticed that continuity of the
structure at B is maintained, however from the All.2 Definitions and Derivations of Terms
moment curves of Fig. All.2 it is found that the
iRternal bending moments in the beams over sup- 1. Fixed-end moments:
By "fixed end moment" is meant the moment
port B are not statically balanced, or specifi- which would exist at the ends of a member if
cally there is an unbalance of 270. The next these ends were fixed against rotation.
step is to statically balance this joint, so
it is unlocked from its imaginary locked state 2. Stiffness Factor:
and obviously joint B will rotate (See Fig. The stiffness factor of a member is a value
All.4) until equilibrium is established, that is, proportional ~o the magnitude of a couple that
until resisting moments equal to 270 have been must be applied at one end of a member to cause
* Also see Chapter C2, Parts F, Land M. unit rotation of that end, both ends of the
** Paper - A. S. C. E. Proceedings, May 1930). member being assumed to have no movement of
All.l
All.2 THE MOMENT DISTRIBUTION METHOD
0.5 when far end is fixed. The letter C

200 will be used to designate carry-over factor.
w = 10#/in.
Fixed~~~~~~TIo=::=;=;;b==~=Fixed
L=~4~'30
r-----'1=
1511
4. Distribution Factor:
If a moment is applied at a joint where two
or more members are rigidly connected the distri-
A
B Fig. All. 1 C
bution factor for each member is the proportional
part of the applied moment that is resisted by
that member. The distribution factor for any
I member which will be given the symbol D equals
K/~K, where K equals stiffness factor of a par-
ticular member and ~K equals the sum of K values
_::-::--_-:-:-=--=- ~ Fixed End for all the Joint members. The sum of the D
Moments values for any joint must equal unity.
5. Sign Convention:
Due to the fact that in many problems where
members come into a joint from all directions as
is commonly found in airplane structure, the
Elastic C;urve for Fixed End Spans customary sign convention for moments may produce
confusion in applying the moment distribution
method. The following sign convention is used in
Fig. All.·3 I this book: a clockwise moment acting on the end
of a member is positive, a counterclockwise one
I is negative. It follows that a moment tending to
Elastic Curve due to Rotation of Joint B rotate a joint clockwise is negative. It should
be understood that when indeterminate continuity
Balancing Joint B moments are determined by the moment distribution
method using the above adopted sign convention,
I ~ -108 162 ~ ~ Carry Over Moments that they should be transferred into the conven-
54 -81 from B to C & A tial signs before proceeding with the design of
the member proper.
426 -588 -588 -831 Final Moments = Values
in Figs. All. 4 plus All. 2 The following sketches illustrate the adapt-
ed sign convention,
Fig. All. 4 Illustrations of Sign Conventions for
End Moments.
Example I 1f--==::::;;::::==::jt Fixed end beam with
J- ~ lateral loads.
831 conventional + ~ tension in bottom fib-

, II moment signs ers is positive bend-
I Final Moment ing moment.
~----~--------=, Curve
Adopted sign moments which tend to

convention rotate end of member
clockwise are positive.
translation. The value of the stiffness factor
will depend in part on the restraint or degree
of fixity of the opposite end of the member
from which the couple is applied. The letter K
is used as a symbol for stiffness factor. Example 2 Translation of supports
of fixed ended beam.
3. Carry-over Factor:
If a beam is simply supported at one end
and restrained to some degree at the other, and
a moment is applied at the simply supported end,
a moment is developed at the restrained end.
conventional 4-
The carry-over factor is the ratio of the mom- moment signs
ent at the restrained end to that at the
simply supported end. For a prismatic beam
without axial load the carry-over factor is Adopted Sign::~ -::;.
convention ~
AI1.3
Example 3
M~ ?P~b~
I L - _ :S;::I M 2
Fig. AII.9
External applied moment
at joint in structure Fig. All. 10
adopted sign Fig. AIl.ll

joint Moments which tend to
rotate joint counter-
clockwise are positive.
moment Fig. All. 12
AIl.,3 Calculation of Fixed End Moments
Since the fixed end moments are statically Since the change in slope of the elastic
indeterminate1additional facts must be obtained curve between ends (1) and (2) is zero, theorem
from the laws of continuity in order to solve (2) as applying to fixed end beams can be re-
for them. In this book the theorem of area stated as follows.
moments will be used to illustrate the calcula- The sum of the areas of the moment diagram
tion of the fixed end moments as well as the must equal zero. And from theorem (1) the stat-
other terms which are used in the moment dis- ic moment of the areas of the moment diagram
tribution method. (Ref. Chapter A7) about any point must equal zero or in equation
The following well known principles or form:
theorem of area moments will be used:- ZM = 0 - - - - - - - - - - - - - -(1)
(1) The deflection of any point "A" on the ZMx = 0 - - - - - - - - - - - - -(2)
elastic curve of a beam away from a tangent to For a beam with variable moment of inertia the
the elastic curve at another point "B" is equal conditions for fixity are:-
to the mom~nt of the area of the M diagram be-
EI /Mdx/EI = 0
tween the points A and B about point "A". /Mxdx/EI = 0
(2) The change in slope between two points Figs. All.ll and All.12 show the static and con-
"A" and "B" on the elastic curve of a beam is tinUity moment areas, the total area of each por-
equal to the area of the M diagram between the tion and its c.g. location.
EI Substituting in Equation (1)
two points "A" and "B". M = Pab M~L M2 L 0
The "area moment" theorems will be illus- 2+2+2= - - - - - -(3)
trated by the applications to the solution of a
and from equation (2)
simple problem. Fig. All.5 shows a simply sup-
ported beam of constant moment of inertia and Mx about left end =Pab L+ a M~L L
modulus of elasticity carrying a single concen- 2 ' -3-+2' 3+
trated load. Figs. All.6 and All.7 show the
static moment curve and the shape of the beam M2 L 2L = 0 __ - -(4)
2 . 3
elastic curve. Now assume that the ends are fix-
ed as shown in Fig. All.8 and let the value of The values of M~ and M2 for any value of a or b
the fixed end moments be required. Fig. All.9 can now be found by solving equations (3) and (4)
shows the shape of the final moment curves made Table All.l gives a summary of beam fixed
up of the static moment curve and the unknown end moments for most of the loadings encountered
trapizoidal moment curve formed by the unknown in routine design and analysis.
end moments. Fig. All.10 shows the shape of the
elastic curve, the slope at the two supports be- W TABLE All. 1 W
ing made zero by fixity at these points. 2
~r f rt llWL2~ ------y 5wL 2
"C i '--+-1
WL r:::WL122
p l2 ~ - - --r 192 I !==y 192
2 2
t------L-------f
Fig. All. 5 r wa
2
12L2 ( 4aL- 3a 2 )
l~ Fig. All. 6
Fig. All. 7
Fig. All. 8
Al1.4 THE MOMENT DISTRIBUTION METHOD
Table All.l - Continued tangent to the elastic curve at point A relative

to a line AB is equal to the shear at A on a
40:- --\V ,r-w simply supported beam AB due to the M curve be-
wL" {fID=r-:-=-r wL30 " 5wL" ~5WL" EI
20 9 6 ~ ~ tween A and B acting as a load.
2 2
Thus
(ML x 1/3) _ 11L =
2EI - - 6EI (Pos1ti ve Shear
wa " 10L
60L 2 ( " - 10aL + 3a ~I ~~
60L" (5L- 3a) Neg. Slope)
11L x 2/3 11L
Gr3 = 2EI 3EI
wa "
30L" (IOL " - l5aL+ 6a "-I
p/ wa"
20L" (5L- 4a) Let flB unity
Then 1 ML or MB = 3EI = stiffness Factor of
3EI L
M Beam BA of Fig. All.13. A moment applied at B
M~(3~-1) ~ ~ J_M~(3~_1) produces no moment at A since end A is freely
L L I-a -+- b --1 L L supported. Thus the carry-over factor for a
beam freely supported at its far end is obvious-
ly zero. Consider the beam of Fig. All.14. Due
t.:f~X(L_X)"dx tl~ ~{LWX'(L_X)dX

to complete fixity at end A, the slope of the
elastic curve at A is zero.
o c:rr 0
(Ref. N. A. C. A.
1/3 + ~
2EI
. 2/3 = 0
General case - Any loading T. N. #534)
or
All.4 stiffness Factor; Carryover Factor: -

Derivation of: Thus the carryover factor for a beam fixed at
its far end is 1/2. Using the conventional mom-
(For definitions of these terms see page ent signs, the carryover moment is of thA oppo-
Al1.2) site sign as shown by the above equation, How-
Consider the beam of Figure All.13. By ever, for our adopted sign convention inspection
Mohr's theorem (see Art. A7.12), the slope of a of the shape of the elastic curve as shown in
Fig. All.14 tells us that the sign of the carry-
EI is constant over moment is of the same sign as the rotating
JMB moment at the near end. That is, the moment act-
ing on each end of the member is in the same di-
---_,.B
L
t rection, and therefore of the same sign.
~-----~
e~~
Elastic Curve
0B Fixed t
A
eA=o
(EI is constant
L Bf
)
MB
----r
,....-=:==-------!----1 ---.1.
::Sl~ -1~
Curve MBL
AL ML 2EI
3 2EI
Fig. All. 13
~
t
~=------+----+---==::=.-+
L
3
Fig. All. 14
flB=~. 2/3+~ .l/3=~+~, but MA=- ~

2EI 2EI 3EI 6EI 2
Then flB = ~ - ~ =~
3EI 12EI 4EI
Let flB = unity, then MB = 4EI stiffness factor
L
of beam BA of Fig. All.14.
A comparison of the stiffness factor of this
beam to that of Fig. All.13 shows that the stiff-
All 5
ness factor of a beam freely supported at its MA = - 2KF - - ~ which is the general ex-
far end is 3/4 as great as one fixed at its far MB K (3 + F) 3 +F
end. Furthermore in one case the carry-over pression for carry-over factor for a degree of
factor is zero and in the other case it is 1/2. fixation F.
It is therefore obvious that the values of these
two terms depends in part upon the restraint AII.5a Example Problems
or degree of fixation of the far end of the
To obtain a definite conception of the true
beam. mechanics of the "Cross" method, the reader is
advised to follow thru the detailed solution of
AII.5 General Expressions for Stiffness and Carry- the following simple problems. In these prob-
over Factor in Terms of Fixation Factor (F) lems, the moment of inertia in any span has been
at Far End of a Beam
taken as constant and all joints have been as-
In the beam of Fig. All.14 (F) fixation sumed to undergo no translation. Problems in-
factor at A was unity since beam had been taken volving variable I and joint translation will be
as completely fixed at A. It was found that:- considered later.
MB = 4EI8B and MA = - 2 EI 8B Example
L L
Then 8B = unity and let EI =K for simplici ty. Problem #1
L
Then MB = 4K
MA = -2K Stiffness Factor Ei = K o1 11 1 0
Likewise the results for the beam of Fig. All.13 Distribution Factor K/1: K o1 .5 .5 10
give Carry Over Factor o .5 .5 .5 .50
MB =3K Fixed End Moments --883768=. 0833wL 2=-768 768=. 0833wL 2 =-768 883
MA =0
1-1s_t_Ba_lan_cl_ng -j0_ll_5~~ ~~_-1_15+0_ _-l
Figs. All.15 and All.16 show these results. Carry Over 1/2 0 ~57. 5 -57.5............... 0
Fig. All.17 shows the general case, the fixa- 2nd Balancing 00 00 00
tion factor at A being F. The difference be-
tween Figs. All.15 and All.17 is that the slope Final Moments -88~ 883 -710.5710.5 -883 883
at end A has changed but 8B the slope at end B Conventional Moment Signs + • • • + +
remains the same. -Overhang Moment = .5 x 51. 5 x 25.75 •. 25 x 51. 5 x 17.17 = 883
6 A=O Example Problem 1 shows a two-span continu-

F =I ~ n _--.1:,B ous beam With over-hanging ends. We first as-
A)~
sume all the joints B, C and D to be locked
MA= -2K MB= 4K against rotation or the beams BC and CD are
Fig. All.15 arbitrarily assumed in a fixed-end condition.
The first line in the solution gives the
stiffness factor of each beam. From Art. All.5
F=O the general expression for stiffness factor is
EI (3 + F)/L, where F = fixation factor of far end
which equals 1, since all joints have been con-
sidered fixed. Thus, stiffness factor K = EI
Fig. All. 16 (3 + l)/L = 4 EI/L. Since EI/L is the same for
each span BC and CD, the stiffness factor has
The change in moment at end A when changing been written as one since it is only the rela-
beam All.15 to that of All.17 = 2K - 2KF = tive values that are necessary. The cantilever
2K(l-F). Since 8B is kept the same value, one- BA has zero stiffness.
half of the moment change at end A appears at The second row gives the member distribution
end B but of opposite sign, or factor D at each joint or D = K/ZK. For example
MB=4K-l/2 2K(1-F) =3K+KF=K (3+F) = at joint B, the distribution factor to BA = 0/(1 +
EI (3 +F) 0) = 0 and to BC = 1/(1 + 0) = 1. At joint C, the
L distribution factor to CB = 1/(1 + 1) = .5 and
6A- F(. 5) likewise to CD.
The third row gives the carry-over factors.
F=F~ From Art. All.5, the carry-over factor C = 2F/(3 +
MA= -2KF MB a K(3+ F) F) = 2 x 1/(3 + 1) = .5 assuming the far end fixed
or F = 1 fo~ each member.
Fig. All. 17
The next step as shown by the 4th horizont
Thus the general expression for the stiffness al row is the calculation of the fixed end mom-
factor of a beam of constant section equals EI ents. The signs or the moments are according to
L our adopted sign convention, that is an inter-
(3 + F). The carry-over factor from B to A = nal resisting moment which tends to rotate the
Al1.6 THE MOMENT DISTRIBUTION METHOD
end of a member clockwise is positive. 4. Calculate the carry-over factor C for each
We now begin the solution proper by first end of all members C = 2F/(3 + F). Thus for
unlocking joint B from its assumed fixed state. beams fixed at far end F = 1 and thus C = 112.
We find a moment of -883 on one side and 768 on For pinned at far end F = 0, hence C = O.
the other side of joint or a static unbalance 5. Calculate the fixed end moments (f) for giv-
of -115. Joint B therefore will rotate until a en transverse member loadings or support deflec-
resisting moment of 115 is set up in the members tions, using equations in summary Table All.l.
BA and BC. End moments which tend to rotate end of beam
The resistance of these members to a rota- clockwise are positive moments. (See Art. All.2)
tion of Joint B is proportional to their stiff- 6. Considering one joint at a time, unlock it
ness. The distribution factor based on the from its assumed fixed state, all other joints
stiffness factors is 0 for BA and 1 for BC. remaining locked. If an unbalanced moment ex-
Thus 1 x 115 = 115 is distributed to BC at Band ists balance it statically by distributing a
o x 115 = 0 to BA. Joint B is now imagined as counter acting moment of opposite sign among the
again locked against rotation and we proceed to connecting members according to their D or dis-
Joint C, which is now released from its assumed tribution factors.
locked state. Since the joint is already stati- 7. These distributed balancing moments produce,
cally balanced, no rotation takes place and the carryover moments at far-end of members equal to
distributing balancing moment to each span is the distributed moment times the carry-over fac-
zero. Next proceed to joint D, and release it. tor C and of the same sign as the distributed
The unbalanced moment is, 115 so the joint is moment. Record these carry-over moments at far
balanced by distributing -115 between DE and DC ends for all distributed moments.
as explained above for joint B. 8. Repeat the proceedure of unlocking each
As pointed out in Art. All.5, when we ro- joint, distributing, and carrying over moments
tate one end of a beam it tends to rotate the until the desired precision is obtained, stopping
far fixed end of the beam by exerting a moment the solution after a distribution. The final
equal to some proportion of the moment causing moment at the end of any member equals the alge-
rotation at the near end. For beams of constant braic sum of the original fixed end moments and
section and fixed at their far ends, the carry all distributed and carryover moments.
over factor is 1/2 as explained before. Thus
the distributing balancing moments in line 4 Example Problem #2
produce the carry-over moments as shown in line .5#/in. ./
l#/in EI is Constant
5 of the table. This completes one cycle of the
moment distribution methOd, which is repeated
-L
-,,- rrff T11 TIl T TTTTTT TTTTTT TTTTTTTr~
until there is nothing to balance or carry-over,

5L5~
B C D E F
}1.5 10------96 72 72 96----J

or in other words until all artificial restraints
have been removed and the structure rests in its Stiffness ~=K 0.0104 :--.!....= .0104 .0139=-k=.0139 .0139 :...!...:. 0139 ,0104 =~= .0104
Factor 72
true state of eqUilibrium. L 96
To continue with the second cycle, go back Distribution D=.1L

Factor L:K
01 .428 .572 .5 .5 .572 .428 10
to joint B and release it again from its assumed Carry-Over C o .5 .5 5 .5.5 .5.5 .50
locked state. There is no unbalance since the Factor
carryover moment from point C was zero, thus Fixed End

Moments
-883 768 =-I2WLZ:: -768 432=f2WL2:: .. 432 432=f2WL2=-432 768 =~WL2:: -768 883
there is nothing to distribute or carry-over.
Proceed to point C, releasing the joint, we find
1st Balancing
Carry-Over
0115?<~ 1 9 2 > < 0 -o
72 57.5
-
o
-
96 -96
X.292
- ~X-1150
0 -57.5 -72
it balanced under the carry-over moments of 57.5
and -57.5. Thus the distributing balancing mom- 2nd Balancing
0-72?X~-32.9 0 O~2.9
~><:: 720
ents are zero. Joint D is likewise in balance Carry-Over -12.3 -36 ;-X~r--
16.5
-
0 36 12.3
since the carry-over moment from C is zero. All 3rd Balancing O12.3:3><~20.6~
-
o
-
0 o XO.6
r----- -
-15.~12.3 0
- - f--
joints can now be released Without any rotation Carry-Over 7.7 6.1 10.3 ·10.3 0 -6.1 -7.7
since all joints are in equilibrium. To obtain 4th Balancing 0-7.7X~

~X~ 0~.5
t--
2.6~7.7
- f-- -
0
~
the final moments we add the original fixed end Carry-Over -1. 3 -3.9 o -1. 8 1. 8 0 3.9 1. 3
moments plus all distributed balancing and car- 5th Balancing o 1. 3 1.72. a 00 -,2.2 -1.7 -1. 30
ry over moments. Final Moments -883 883 -610.4 610.4 -344 344 -610.4 610.4 -883 883
Conventional Signs +. ++ ++ ++ ++
All~6 General Summary of Proceedure
1. All computations should be written on, or Example Problem 2 is similar to problem 1 but two
adjacent to the diagram of the structure. spans have been added. We first assume all joints
2. Determine the stiffness factor K for each locked against rotation. The stiffness factor of
end of each member. K = (3 + F) EIIL, where F each span is proportional to EIIL or llL since EI
is the degree of fixity at far end of member. is constant. The carry-over factor is 1/2 as in
If all members are assumed fixed at far end then previous example. Fixed end moments are calCUlat-
K is proportional to IlL assuming E as constant. ed as shown. Unlock joint B, the unbalanced mom-
3. Determine the distribution factor D for each ent is -115. Balance the joint by distributing
member at each joint of structure. D = K/~K. 1 x 115 to BC and zero to BA. Proceed to joint C,
All 7
and unlock, all other joints remaining fixed necessary. These modifications usually involve
against rotation. The unbalanced moment is rather long expressions for expressing the stiff
(-768 + 432) = - 336. Balance by distributing ness and carryover factors of a member in terms
.428 x 336 = 144 to CB and .572 x 336 = 192 to of the fixation given by adjacent members. It is
CD. Proceed to joint D, and release. The un- felt that it is best to keep the method in its
balanced moment is zero which means that joint simplest form which means that very little is to
D is in equilibrium, thus no distribution is be remembered and then the method can be used in-
necessary. Proceed to joint E and F and balance frequently Without refreshing ones mind as to
in a similar manner. The distributed moments many required formulas or equations.
will be the same as the values for joints Band There are however several qUite simple modi
C due to symmetry of structure and loading, fications which are easily understood and re-
however, the signs will be opposite under our membered and which reduce the amount of arithmet-
adopted sign conventions. The next step is the ic required considerably.
carry-over moments Which are equal to 1/2 the For example in Problem #2, joints Band F
distributed balancing moments. This operation are in reality freely supported, thus it is need-
is shown clearly in the table. Values of all less arithmetic to continue locking and unlocking
moments are given only to first decimal place. a joint which is definitely free to rotate.
The first cycle has now been completed. Cycle Likewise due to symmetry of structure and load-
two is started by again releasing joint B. We ing it is only necessary to solve one half of the
find the joint has been unbalanced by the carry- structure. Due to symmetry joint D does not ro
over moment of 72. Balance the joint by distri- tate and thus can be considered fixed, which
buting - 72 x 1 = - 72 to BC and zero to BA. eliminates the repeated locking and unlocking of
Proceed to joint C. The unbalanced moment is this joint.
57.5. Balance by distributing - 57.5 x .428 = A second solution of Problem #2 is given in
- 24.6 to CB and the remainder - 32.9 to CD. Example Problem #3. As before we assume each
Proceed to joint D. There is no unbalance at joint locked and calculate the fixed end moments.
this joint since the carry-over moments are in Now release or unlock joint B and balance as ex
balance, thus no distribution is necessary. plained in previous example #2. Before proceed
Proceed to joints E and F in a similar manner. ing to joint C, carryover to C from B the car-
The carryover moments equal to 1/2 of the ry over moment equal to 115 x 1/2 = 57.5. Joint
2nd set of balancing distributed moments are now B is now left free to rotate or in its natural
carried over as shown in the table. The second condition. Proceed to joint C and unlock. The
cycle has now been completed. This operation unbalanced moment = (-768+432+57.5) =- 278.5 or
has been repeated five times in the solution 278.5 is necessary for equilibrium. This moment
shown, or until the values of the balancing and is distributed between two beams, CD which is
carryover moments are qUite small or negligible. fixed at its far end D and CB which is freely
The final moments equal the algebraic summation supported at the far end B. The stiffness fac-
of the original fixed end moments plus all distor is equal to (3 + F) EI/L (See Art. All.5).
tributed and carryover moments. One require- Hence for CD stiffness factor = (3 + 1)
ment of the final end moments at any joint is EI/L = 4 EI/L. For CB stiffness factor = (3 + 0)
that the algebraic sum must equal zero. The EI/L = 3 EI/L or in other words the stiffness of
other requirement consistant with the common a beam freely supported at its far end is 3/4
slope to all members at any joint is given by as great as when fixed at its far end. Thus the
equation (5) of Art. All.8. The results at stiffness factor of CB at C is .75 x .0104 =
joint C will be checked using this equation. .0078. The carry-over factor C to B is zero
since B is left free to rotate. (See Art. All.5)
fillcd - .5 tl !'Idc = lS.c..d
fillcb - .5 tl Mbc Kcb Example Problem #3. Simplified Solution of Prob-
lem #2
SUbt. values
(610.4-432)-.5 (344-(-432)]_178.6-44_
[-610.4- (-768)J -.5 (883-768f157.6-57.3- rfT·f 'r'
T (TI#(1 TT TT T TiT f ~ixe~
-r
1.343 j:-51. 5~<-B---96---""C--72---;
E
K"rl .0139
Ratio of stiffness factors = ~ = .0104 = 1.34. Stiffness Factor (k) o .0104=k . 75x. 0104=. 0078 .0139 = 1/72 = .0139 pc
Thus the distribution is according to the K Distribution Factor
o = K/ZK 0 1 .36 .64 o 1
ratios of the adjacent members. Carry-over factor (e) a .5 0 .5 .5
-432
Simplifying Modifications - Example Problem #3 Fixed end moments ·883 768 -768 432
Balance joint B 0 115 ~ _
The solution as given in Problem #2 repre- -=C=ar=ry=_o-'-ve-=r"-I/=2-'o-c:C,-----IF-=--- 5 -----+----
----.c--=-57=-.e:-t
Balance joint C ---===100.1 178. 4 ~-;::--,:t----
_
sents the "Cross" method in its fundamental and -=c"-ar"-ry--o-'-ve-r-'o-C:B-,=-o----If::-o-~ 89.2
most elementary detailed form. Many modifica- o =::----+---------+:--O~
-=B::.:ala:::n.::..::ce-,-io:::in:.:..' 0 342.8
tions of the general method have been presented, Carry-over to C
Final moments -883 883 -610.4 610.4 -342.8 342.8
in the most part for the purpose of eliminating Conventional signs + + ++ ++
part of the arithmetic or the number of cycles
All.S THE MOMENT DISTRIBUTION METHOD
The stiffness factor of the fixed support is in- Solution #1

,-U/in. 50% FiXed)
finity, that is a rigid support has infinite re- Example .5~. rTTTT
sistance to rotation. Pr oblem#4 TTTTTTT TTTTTT
'r V
Joint C is now balanced by distributing 1;-51. 5 96 72---J
278.5 x .36 = 100.1 to CB and 278.5 x .64 = 178.4 B C D E
to CD. Now carryover to joint B, 0 x 100.1 = Stiffness Factor k o .0104 = 1/96 = .0104 .0139=1/72=.0139 .0139
o and 0.5 x 178.4 = 89.2 to D. Proceed to joint Distribution Factor D;ik o1 .428 .572 .5 .5
D and release it from its assumed locked state. Carry-Over Factor 0.5 .5 .5 .5
The unbalanced moment is -432 + 89.2 = - 342.8, Fixed End Moments -883 768 -768 432 -432
we balance by distributing 342.8 between DC 1st Balancing o 115 144 192 216 216
Carry-Over 72 57.5 108 96
which has a stiffness of 0.139 and the support 2nd Balancing 0-72 -70.7 -94.8 -48 -48
E which has an infinite stiffness or zero goes Carry-Over -35.3 -36.0 -24.0 -47.4
to DC and 342.8 goes to rigid support. The car- 3rd Balancing o 35.3 25.7 34.3 23.7 23.7
Carry-Over 12.9 17.6 11. 8 17.2
ry over moment to C from D is zero since 0.5 x 4th Balancin[. 0-12.9 -12.6 -16.8 -8.6 -8.6
o = O. The final moments thus equals the sum- Carry-Over -6.3 -6.4 -4.3 -8.4
mations as shOwn Which of course are equal to th 5th Balancing
Final Moments
06.3
-883 883
4.6 6.1
-644.3 644.3
4.2 4.2
-187.3 187.3
results shown in Problem #2. Conventional Signs ++ ++ ++
Example Problem #4
.5 x 216 = 108 is carried over to C. Joint D is
Problem 4 is similar to Problem #2 and #3, left unlocked or in its true state of restraint.
except that the support at D is assumed as hav- Joint C is now unlocked. The unbalanced moment
ing 50 percent fixity. Thus 50 percent of any
moment at this point produces rotation of the Solution#2
member DC at D. Example .rU/in 50% Fixed
Pr oblem #4 .5:tn . t::=)
r-
T tttttttt 1 1
In continuous wing beams, which fasten to- T
gether by fittings on a support, it is commonly 5
1.5
96 72--->
required that the beam be considered as being B C DE

fully continuous and also that the degree of stiffness Factor K
1
_883 .0104-9'6
7 1
. 75x. 0104=.0078 ,0121-axi2 iZ-·0139 .0139
continuity be taken as 50 percent. Solution 1
Distribution Factor=~ 01 .392 .608 .50 50
of Problem 4 is a detailed solution. The only Carry.over Factor 1/2 o .286 1/2
change that has been made is in the stiffness Fixed End Moments -983 768 -768 432 -432
factor of the support E, which has been taken Balance Joint B
Carry-over to C
0115~
57.5
as equal to the beam DC, thus any unbalanced
~~
Balance Joint D
moment at this point is equally divided between Carry-over to C
Balance Joint C
108
66.8 103.7
the beam and the support. Carry-over to D 29.6
Solution 2 is a modified solution which Balance 0
Final Moments -883 883 -643.7 843.7
29.6
4186.4 186.4
eliminates considerable arithmetic. Thus it is
unnecessary to lock and unlock joint Band D
since we know definitely that one is freely sup- equals (-768 + 432 + 57.5 + 108) = - 170.5, or
ported and the other 50 percent fixed. There- 170.5 is necessary for equilibrium. Joint C is
fore, once we have releasee these joints from balanced by distributing .392 x 170.5 = 66.8 to
their assumed fixed end state, we leave them in CB and the remainder of 103.7 to CD. The carry-
their natural state. The stiffness and carryover moment to B is zero and to D it equals
over factors for beams CB and CD must then be 103.7 x .286 = 29.6. The final moments in solu-
determined for these beams with their modified tion #2 are slightly different than solution #1.
end conditions. If another cycle had been added in solution #1
By reference to the fundamental equations the discrepancy would be considerably smaller.
for stiffness and carryover factors in Art. All.7 Continuous Beams with Yielding or Deflected
All.5, it is readily seen that the stiffness Supports *
factor for CB is 3/4 as much as when fixed at In wing, elevator and rudder beams the sup-
its far end B and the carryover moment is zero. port points usually deflect due to the deforma-
For beam CD the stiffness is 7/8 as much as when tion of the supporting struts or wires in the
fixed at end D, and the carryover factor is case of a wing, or to the deflection of the
2F/3 + F = (2 x .5)/3 + .5 = .286. With these stabilizer or fin in the case of elevator and
modifications the solution is carried thru With rUdder beams. If these beams are continuous this
a relatively small number of steps. Thus in deflection of their support points causes addi-
solution #2, joint B is unlocked. The unbal- tional bending moments in the beams. The moment
anced moment of -115 is balanced statically by distribution method can of course be used to find
distributing 115 to BC. The carry-over moment the additional moments due to this deflection.
of .5 x 115 = 57.5 is carried over to C as shown. Thus Example Problem #5, shows a solution illus-
Joint B is now left unlocked or free to rotate. trating a problem which involves the deflecting
Joint D is unlocked next. The unbalanced moment of the supports of a continuous beam. Due to
is -432. It is balanced by distributing 216 to symmetry of structure and loading, the slope at
DC and 216 to E since support at E is considered D is zero or the beam may be considered fixed at
to give 50% fixity. The carry-over moment of
* Also see Chapter C2, Parts F, M and Q.
All 9
Joint Do Since the moment of inertia is con- Since the fixed end moments are due to both
stant and the spans are constant, the relative lateral loads and support deflection, the values
stiffness factor of the beam is 1. In the so- as listed in the solution table will be explain-
lution shown since beam is freely supported at ed in detail.
B this joint is left free to rotate after re-
leasing and thus the stiffness factor of beam Fixed End 110ment For Lateral Beam Loading
CB is 3/4 x 1 = 3/4, when compared to one llaving The distributed airload is trapezoidal in
fUll fixity at B. shape. The fixed-end moments for a trapezoidal
Since the first step in the solution proper loading from Table All.4 are:
is to assume the joints fixed against rotation,
it is evident that deflecting one support rela- 111_2=L 2 (5u+2v) (See Fig. (a))
tive to an adjacent support will produce mom- 60
ents at the ends which are assumed fixed against 112_1=L 2 (5u + 3v) (See Fig. (a) )
rotation. 60
AII.S Fixed End Moments --L _

Due to Support
Deflections v
Fig. Allo18 shows a T

fixed end beam. The sup- u Fig. (a)
M ~4EI
Err //,
ML port B is deflected a
distance d relative to
the support point C. If
-L
CD
1& L ,®
~
L ~ -E~ the member is of constant
6 ~~
~L
J cross-section the point
of inflection will fall For Span BC:
6 at the beam midpoint and
40 2
Fig. All. 18 11B - tlc. 60 (5x3+1) = 426 in. lb.
By moment area prin-
40 2
cipal deflection 60 (5x3+l.5) 440 in. lb.
d tlL L -tlL 5 L
4EI (3 4EI (3 For Span CD:
or d 40 2
60 (5x3.5+1) = 494 in. lb.
hence 40 2
11 6EId the magnitude for the fixed end 11dc = 60 (5x3.5+l.5) = 507 in. lb.
L"
moment due to a transverse support settlement Fixed End 110ments Due to Support 110vement
of d.
From Art. All. 8 11 = ~EId/L2
Example Problem #5. Continuous beam with de-
For Span BC:
flected supports.
General 1 Deflection of B relative to C = 5/16 inch. Hence,
4 #/in. 3 #/in. lr.049 AI.
11bc =11 c b = 6 x 10,000,000x .03339 x .3125/40 2 = 390
~ I~ I ~~l~:
data:
Oi;::9
in. lb.
----+- -I--
D
40 40 --110~-
Defleciion of SupP!'rt~
For Span CD:
DI I I J 2
11 cd =11dc = 6 x 10,000,000 x .03339 X .1875/40 = 234
- -Fixed due to
~-r---r----,_,--~-,---rilisymmetry
in. lb.
Soluiion: For signs of the moments due to these deflections
A B C D
see Art All.2. Having determined the fixed end
Siiffness Factor K o1 3/4xl=.75 1 1 0<' moments the general distributing and carrying
Distribuiion Factor = K! LK o1 .429 .571 o 1 over process follows as indicated in the solution
Carry-over Factor 1/2 o 1/2 1/2 table. Thus at joint B, the unbalanced moment =
Fixed
End
} Due to
Lateral Load
50 _426 440 -494 507 (50-426 + 390) = 14. Balance by distributing - 14
Moments Due to .support o 390 390 234 234
x 1 = - 14 to BC and zero to BA. Carryover.5 x
deflectlOn -14 = - 7 to C. Considering joint C, the unbal-
Balance joint B o -14 anced moment = (440 + 390 + 234 - 494 - 7) = 563. Bal-
Carry-over to C -7
Balance joint C -241 -322
ance bydistri buting - 563 x .571 = - 322 to CD and
Carry-over to B, D 0 -161 -563x .429 = - 241 to CB. carryover.5 x -322 =
Balance joint D o -580 - 161 to D. At joint D the unbalanced moment =
Final Moments 50 -50 582 -582 580 -580 (507 + 234 - 161) = 580. Thi sis balanced by di s-
Conventional Signs tributing zero to DC and -580 to the fixed support.
All.lO THE MOMENT DISTRIBUTION METHOD
All~8 Check on Final Moaents ~b - .5&Iba = ~ -(5)

To satisfy statics, the algebraic sum of ~c - .5&Ica Kac
the moments in all the members at a joint must If equation (5) is satisfied the accuracy of the
equal zero. This requirement alone will not moment distributions is established.
prove that the final moments are correct, as er-
rors could have been made in the various distri- All.9 End MOIIlents for Continuous l"ra.e1forks Whose
butlng and carryover moments. Continuity tells lleabers are not in a Straight Line. Joint
US that the final slope of the elastic curve at Rotation Only *
anY joint is common to all members meeting at Continuous truss frame works are quite com-
that joint, thus for a complete check it is mon in aircraft construction. Welded steel tub-
necessary to prove that the final moments are ular fuselages are composed of members which
consistent with equal slopes for the various maintain continuity thrU the joints due to the
members. The rotation of the joints from their welding. Landing gears frequently have two
original assumed fixed or locked condition is members which are made continuous at their points
due entirely to the distributed and carryover of connection. The members of such structures
moments. The actual rotation will therefore usuallYl carry high axial stresses which cause
equal the rotation of the end of a simply sup- joint translation which in turn produces bending
ported beam when subjected to end moments equal of the members since the joints are rigid. These
to those produced by the algebraic sum of the moments produce lateral deflection of the memb-
distributed and carryover moments. These end ers which introduces additional secondary moments
moments equal the final moments minus the origi- due to member axial loads times the lateral memb-
nal fixed end moments and will be referred to as er deflections. These influences are treated in
~ moments. later articles. In this article to further fa-
From Chapter A7, the slope at any point on a mlliarize the student with the moment distribu-
simply supported beam equals the shear due to the tion procedure, the effect of joint translation
M/EI diagrams as a loading. and secondary moments will be neglected.
The slope QA from Fig.a' Example Problem #6
equals the beam shear at
I--L/3 ~ MabL A or equals the reaction Fig. All.19 illustrates a simplified landing
r- 2EI at A. gear chassis problem. Let it be required to de-
termine the bending moments in the two members
AMab
~ !- ---= --=::::::..,;B due to the vertical load on the axle. The prob-
Af~· _
lem has been solved using three different de-
L grees of restraint at ends A and B. Joint 0 is
a welded joint and full continuity is assumed
thru this joint. The solutions as given in Fig.
Fig. a' -r ~Mba
All.20 give only the moments due to joint rota-
tion under primary bending moments. The effect
c::::::::=------+----i ---±-.E X of axial deformation and secondary moments due
Al~ B to member deflections is omitted in these solu-
_.- - - - - . l . L L/3 tions. These factors are treated in later arti-
3
~MbaL cles.
~ In a practical problem the degree of re-
straint at points A and B would be determined by
_ ~~L 2/3 _ ~~~L . ~ the type of fitting used and also on the rigidity
of the adjacent fuselage or Wing structure. As
__ (AMab + .5 &Iba) illustrated in later examp~es, the moment distri-
3EKab bution method permits the consideration of the
where rigidity of the adjacent structure without adding
t any difficulty, while such methods as least work
r- -+
Kab = of AB
Since the angle QA must be the same for A
18"
------T-----
~2'~ ~ B
members meeting at A, the general relation be-
tween the moment increments of any two members
such as AB and AC must be, .-t.
.,,~
<.,.
'fl'
--"
... "" '%>
~ 0.
I
I
I
I
i>
~'\; ~
~
"'"v:
~'~
....'O b;:
'!::;~
T
~b + .5 &Iba = Kab ?;;(/ " "'o.~?, I ~
<;:)'0 '\>',' 'i:;~'
~ '~" .?>.
--- ~
AMac + .5 &Ica Kac I "" ....'\."""....c:s
00 ~ ~dl 'l> \"'>'V\
~ ~o ~~
To make this equation consistent with the assum-
ed sign convention, that is, the carryover mom- "'~ ~4
ent has the same sign as the balancing distri-
buted moment, the above equation must be modi- 6'~OO# Fig. All. 19
fied as follows:- * Also see Chapter C2, Part H.
All. 11
Condition I to be freely ,supported, they will be left in

Fixed at A and B B their true state. Thus the carryover moments
from end 0 will be zero. Since A and B are both
pinned, the relation between the relative stiff-
ness factors of members OA and OB remain the
same as in condition I, thus the same K or
stiffness factors that were used in condition I
can be used in distributing moments at joint O.
Joint 0 is balanced in same manner as condition
I but with zero carryover moments to A and B.
Solution for Condition III. A is 50% Fixed and
B is Pinned.
Since each member has a different degree of
fixity at its upper end, the stiffness and carry
over factors will be considered in detail. In
condition I since both members were fixed at
their upper ends the relative stiffness factor
of each member was proportional to I/L for the
Condition ill B member and this ratio was used. The general ex-
~ + A= 50% Fixed pression for stiffness factor is K = EI (3 + F)/L
~~~~:-Pinned ~~":J<:?J carryover factor 2F. For member OA, F
3+F
\P ~ •
r'~ .c'
v equals .5 since A is 50% fixed, and for member
'cP~ ~ \..~, ~ OB, F is zero since B is freely supported. Hence
for member OA
-'
~ 0 <::I)":J"J
18000 K=E{(3+ .5) =3.~ EI=3.5X3~105 E= .0129 E

K = Stiffness Factor
Values in ( ) are carryover factors C.O. Factor from 0 to A = 2 x .5 = .286
3 + .5
Fig. All. 20
For member OB
are not practicable because of the large number
of equations that must be solved to obtain val- K = E{ (3 + 0) = 3 x •~~ x E = .0139E
ues for the many unknowns.
Solution for Condition I, Fixed at Ends A and B C.O. Factor From 0 to B = 2 x 0 = 0
3 + 0
Referring to Fig. All.20, all joints are
assumed locked against rotation or fixed. The Considering joint 0 in Fig. All.20 the external
vertical axle load of 6000# produces a counter- moment of 18000 in.lb. is balanced by distribut-
ing - 18000 between the two members in proportion
clockwise moment of 3 x 6000 = 18000 in. lb.
about joint O. The sign is positive (See Art. to their stiffness factors. Hence - 18000 (.0129/
All.2). Release or unlock joint 0, the unbal- .0129 + .0139) = -8650 in lbs. is resisted by OA
anced moment is 18000 or - 18000 is required and the remainder of -18000 (.0139/ .0129 + .0139)=
for static equilibrium of joint O. Joint 0 is -9350 to OB. The carryover moment from 0 to A =
balanced by distributing - 18000(.464/.464 + .286 x -8650 = -2475 and zero from 0 to B. (See
.369) =-10030 to member OB and the remainder or Fig. A1L20)
or -18000(.369/.464 + .369) = -7970 to OA. Example Problem #7
These distributed balancing moments at 0 produce Fig. All.21 shows a structure composed of 3
carryover moments at A and B.
Thus carryover to B, .5 x - 10030 = - 5015 members. Member AO is subjected to a transverse
load of 120# at its midpoint. Joint A is fixed,
and carry over to A, .5 x - 7970 = - 3985 B is freely supported,
Proceeding to joint A which is a fixed joint, C is 25 percent fixed
the unbalanced moment of -3985 is balanced en- and joint 0 is considered to maintain continuity
tirely by the rigid support, or no rotation between all members at O. The end moments on
the three members due to the transverse loading
takes place when joint is released from its im- on member AO will be determined.
aginary fixed state. Similar action takes place
at joint B. The final end moments are as Solution #1. Fig. All.2l gives a solution using
shown in the Figure. the "Cross" method in its fundamental unmodified
state. The solution is started by assuming all
Solution for Condition II. End A and B Pinned three members as fix-ended. The relative stiff-
For this condition the ends A and Bare ness factor K of each member is therefore pro-
freely supported. Instead of locking and un- portional to I/L of each member. These K values
locking these joints which are definitely known are listed in Fig. All.21. The distribution
to be free
All. 12 THE MOMENT DISTRIBUTION METHOD
factor D for each member at each joint which Member - K- Value C. O. Factor
equal K/ZK is recorded in Don each member 2F
~ (3+ F)
around each joint. Thus any balancing moment 3-F
is distributed between the joint members as per 2 X 1 _1
these distribution factors. The carryover OA=-.!..(3+l) =.4, 3 + 1 -"2
20
factors for all members is 1/2. The fixed end
moments due to external loading are computed OB"~·;(3+0) =.3, 3+0 2xO =0
for the three members. For member AO, the fix-

ed end moments equal PL/8 = 120 x 20/8 = 300 in. OC=i5(3+.25)=~ 2x.25= 3 + • 25 .
154
lb. The other two members having no transverse ZK=.83
loading, the fixed end moments are zero.
In this solution the order of joint con-
sideration has been AOBC and repeat. starting
with joint A the joint is released but since
the member AO is actually held by a fixed sup- Fixed 482
port, no rotation takes place and the balancing -300 f f 300
0~11~
moment of 300 is provided entirely by the sup- ~c,-b,-145
port and zero by the member AO. The carryover -372.5 t f 155
moment C, to 0 is zero. Releasing joint 0, the ...
..... .0 ....
unbalanced moment of 300 is balanced by distri-
buting - 300 between the three members accord- SOLUTION #2 !
ing to their D values, thus - 300 x .416 =
- 125 to OA; - 300 x .168 = - 50 to OC and 000
. . . <ill.... Fig. All. 22
- 125 to OB. To prevent confusion it is recom- by distributing .75 x 25 = 18.75 to CO and the re-
mended that a line be drawn under all distri- mainder of 6.25 to the support, since the fixity of
buted balancing moments, thus any values above the support at C has been assumed as 25 percent. A
these lines need not be given further consid- line is drawn under the 18.75 and the carryover
eration and only values below the lines need moment of 9.37 is taken over to O. One cycle has
be considered in later balancing of the joints.
Immediately after distributing the moments at now been completed. Returning to joint A, we
find -62.5 below the line. This is balanced by
joint 0 the proper carryover moments should be distributing zero to OA and 62.5 to the fixed
taken over to the far end of each member, thus support. A line is drawn under the zero distri-
- 62.5 to A, - 62.5 to Band - 25 to C. Joint
B is next considered. The unbalanced moment is buted moment to AO and the carryover moment of
zero is placed at O. Considering joint 0 for
-62.5 and it is balanced by distributing 62.5
to BO since the pin support has zero stiffness, the second time the unbalanced moment is 9.37 +
31.25 + 0 = 40.62 or the sum of all values below
or no resistance to rotation. A line is drawn the column horizontal lines. The joint is bal-
under the 62.5 and the carryover moment of
31.25 is placed at O. Joint C is considered anced by distributing - 17 to OA and OB and
-6.62 to OC. Lines are drawn under these bal-
next. The unbalanced moment of -25 is balanced ancing moments as shown in Fig. All.2l and the
SOLUTION #1
carryover moments are taken over to the far
ends before proceeding to joint B.
This general process is repeated until joint
A has been balanced 5 times and the other joints
4 times each, as indicated in the figure the dis-
Stiffness Factor K
tributing values have become qUite small and it
I 2 is evident that a high degree of accuracy has
OA=r:= 20 = .10
been obtained. The final end moments at each
1. 5 - 10
OB = 15-' joint equal the algebraic sum of the values in
OC= -h=~ each column. A double line is placed above the
~ K - Joint 0 =. 24 final moments as a distinguishing symbol. In
C. O. Factor = 1/2 the figure the letters band c refer to balancing
For All Memb.
and carryover moments, the subscripts referring
to the member of the balancing or carryover
oneration. Any order of joint consideration can
be used in reaching the same result.
Solution #2 of Problem 7
Fig. All.22 gives a second solution. With
the end conditions known at A, Band C, the modi-
fied stiffness factors of the members can be .
found together With the modified carryover fac-
tors, thus making it necessary to balance joint
Fig. All. 21 o only once and carryover this final far end
All 13
moments of each member. The figure gives the assumed to give 50% fixity to these joints. In
calculation of the modified or actual stiffness Table All.2 a modified stiffness factor is cal-
and carryover factors. With these known the culated for members GI, FI, and FH using a 50
solution is started as before by computing the percent fixity at their far ends. The last
fixed end moments due to transverse loading on column of Table All.2 gives the summation of the
member AO. Joints Band C are released and member stiffness factors for members intersect-
since no fixed end moments eXist, no balancing ing at each joint.
is required and the joints are left in their 120#
true state of restraint instead of locking and
iI2~r-
3200"#
unlocking as in solution #1. Releasing joint 0 r G I r.
I- /'
r --7l\A1l/~ /
from its imaginary fixed state the unbalanced /'
moment is 300 which is balanced by distributing
50% Fixity
- 300 between the 3 connecting members accord- Assumed
ing to the new distribution factors at joint O. '-.. ~"'" 100# I
Thus - 300 x .482 = - 145 to OA; - 300 x .361 =
- 108.2 to OB and - 300 x .157 = - 47 to OC. Engine
Mount
;r"f
2400~
A
-hO"1-
~ .
{4t:0"# F H
The carryover moment to A - - 145 x .5 = - 72.5 \ /
to B = - 108.2 x 0 = 0 and -47 x .154 = - 7.2 to
~ /
C. Fig. All. 23
Example Problem #8
"- \ !Landing
" Gear
Figure All.23 shows the forward portion of
a fuselage side truss. Due to eccentricity of TABLE All. 2
engine mount and landing gear members, external
moments are produced on joints A, Band D as Member Size Tube
Length
L 1
1 x 1000
L
i~tor
L
each joint
shown. Furthermore lateral loads due to equip- Stiffness il-
ment installation are shown acting on members BE It Factor L
Joint Stiffness
and CD. Assuming the fuselage welded joints
IlK 1-1 8 - .049 41.25" .0240 0.582 A 5.80
produce rigid continuity of members thru the KG 1-1 8 - .049 38.75 .0240 0.619 B 4 81
joint, the problem is to find the end moments in Gl
AC
Ul
1-1
8 - .049 34.0 .0240
.049 19.25 .0588
.618*
3.06
C 6.95
6.14
2 - D
all the members due to the eccentric joint mom- D -1 2 - .049 30.0 .0588 1.96 K 2.58
ents and two lateral loads. The effect of joint DF 1-1 2 .049 30.0 .0588 1.96 F 4.42
Fa 1-3 8 - .049 34.0 .0449 1.16* G 3.21
translation and secondary moments due to deflec- AB U3 4 - .049 34.5 .0947 2.74
tions and axial loads is to be neglected in this ED
FG
-
1
4
8
-
-
• 49 35.4
.049 345
.03339 0.944
0240 0.695
example. BC
CD
1 1
1-1
2 - .049 39.5
.035 41.0
.0588
.0178
1.49
o 434
(K is constant
for all aea-
8 -
Solution: DG 1-1 2 - 049 16.0 .0588 1 28 bers)
Table All.2 gives the calculation of the 1'1 1-1 14 _ .049 18.5 .03339 0.602*
( * stiffness factor - 7/8 L because of 50% fiXity)
stiffness factor for each truss member. The L
fuselage truss aft of joints I and H have been
Fig. All. 24
-2981 1725
-282 b -141
=-sa b -176
::-g
-187 b
::-g
- 94
- 18 b b - 17 -16 b - - x .286 - 4
-1 b 2 2b -1 .:-5b - - - - - 1
1b -0 =----2 b -
-0
5 ob -.JL
-684 b_O --.lb -21 - 5
401 -195 b 0
-115
50% Fixed
}---~
50% Fixed
"'~I'"'~I<-~II'"
COMCOCON
U':>~
I I I
co
~
I
.602
.0 .0
'"'1"'''I~~II~
~~~~ll"J~
I I I I
H 50% Fixed
-2400
- 1268 b - _ - 634 - 4451 1 222
318 - - b 635 406 b 203
183 b ~ 3i4 b 629 ~b 314
~ ---HQ ~b -=-45 - 70 --b-140 - 82 b _ x. 286 - 23
60 b 30 44 b 88 88 b ~ =--9b - 3
- 777 - 33 - 21 b -=-iO -7 - - b - 15 - 0b o
--3
:-""50" 209 b __
7 ~b - 90 - 26
1094 647 b 0
206
Fig. All.24 gives the solution of the by distributing 183 to AC and 164 to AB with
problem. The procedure in this solution was as carryover moments of half these values to ends
follows: C and B respectively.
The stiffness factor K for each member as The student should now be able to check the
computed in Table All.2 is recorded in the cir- rest of the solution as given on Fig. All.24.
cles adjacent to each truss member. The carry The solution could be made with any order of
over factors for all members is 1/2 except for joint consideration. If any particular joint ap-
modified members GI, FI and FH for which the pears to be nearly balanced, it is best to skip
carryover factor to the 50% fixed ends is .286. it for the time being and consider those joints
The distribution factor for each member at each which are considerably unbalanced.
joint is recorded at the end of each member, and The final moments at the end of each member
equal s K/L:K. are given below the double lines.
The next step in the solution is to compute
the fixed end moments due to the transverse Example Problem #9
loads on members. Fig. All.25 represents a cross section of a
For member BE welded tubular steel fuselage. The top and bot-
tom members which are web members in the top and
MBE = Pab 2 /L 2 = 120 x 29.25 x 122 /41.25 2 = -al8"
2 2 bottom fuselage trusses are subjected to the
MEB = 120x 29.25 x 12/41.25 = 725"# equipment installation transverse loads as shown.
McD = 100 x 20 2 X 10/302 = - 445 Let it be required to determine the end bending
moments in the rectangular frame due to these
MDC = 100 x 10 2 X 20/30 2 = 222 transverse loads assuming full continuity thru
These moments are placed at the ends of the joints.
members on Fig. All.23 together with the eccen-
tric jOint moments. The process of unlocking Solution:
the joints, distributing and carrying over mom- Fig. All.26 shows the solution. The distri-
ents can now be started. In the solution as bution factors based on the member stiffness fac-
given the order of joint consideration is tors are shown inDat ends of each member. The
ABCDEFG and repeat, and each joint has been bal- first step is to compute the fixed end moments
anced three times. due to transverse loads, on members AB and CD us-
Consider joint A:- ing equations from Table All.l. The magnitudes
Unbalanced moment = 2400. Balance by dis- are 1890"# for AB and 2025"# for CD.
tributing - 2400 as follows:- Joint B is now released from its assumed
To AC 2400 x .527 = -1268. Carryover fixed state. The unbalanced moment of 1890 is
to C = -634 balanced by distributing - 1890 x .247 = - 467 to
To AB 2400 x .473 = -1132. Carryover BA and the remainder of -1423 to BD. The carry
to B = -566 over moment to A = - 465 x .5 = - 233. Due to
symmetry of structure and loading only one half
Proceed to Joint B:- of frame need be considered and hence these car-
Unbalanced moment (-566 + 3200 - 298) = ry over moments to A are not recorded. However,
2336. Joint is balanced by distributing - 2336
in balancing joint A it will throw over to B the
to connecting members as follows:- same magnituae of carryover moments as thrown
To BA = - 2336 x .569 = - 1330. Carryover to over to A from B but of opposite sign since the
A - 665 original fixed end moment at B is minus. Thus
To BC = - 2336 x .310 = - 724. Carryover to
233 comes to B from first balance of A as shown
C - 362 in the figure. The distributing moment to B of
To BE = - 2336 x .121 = - 282. Carryover to -1423 produces a carryover moment of -1423 x .5=
E - 141
The convenient device of drawing a line
under all balancing moments is used to prevent
confusion in later balances of the joint.
-712 at D. ~ 27"
10"
300#
7"
300#
H
1'.... 035 Tube
T
Proceed to Joint C:- K=' 0~~37 = .000458
Unbalanced moment = (-634 - 362 - 445) = -1441
The joint is balanced by distributing 1441 as
follows:- 0>
'<l' '<l'
To CA = 1441 x .44 = 635. Carryover to A = 318 36" o .....
o
CB = 1441 x .214 = 309. Carryover to B = 155 I 0
CE = 1441 x .064 = 92. Carryover to E = 46 co . . . "I..:l
CD = 1441 x .282 = 406. Carryover to D= 203 "'-.
M II
This process is continued for the remainder of 600#
..!. :.::
the truss joints. After all joints has been
balanced once, on returning to joint A we find 1---13.5'+13.5"~
below the lines an unbalanced moment of (318-
665) = -347. The joint is balanced a second time 1-1/8 - . 035 Tube
K =!...:·01782=
L 27 .
00066 Fig. All. 25
All. 15
A B tors will not be considered here but will be

'---=-18=9=Of::--1 treated in a later article.
)( 1 f 1890 (A) Calculation of Fixed End Moments for a
~-467 Given Beam Load
)( 1/3 233
~~
)( 1/3 -27
Fig.All.27 shows a fixed end beam with a
variable moment of inertia and carrying a single
~2L
-15 concentrated load of 100#. The beam moment dia-
b4~ gram for this load is considered in three parts.
1710 D Fig.All.28 shows the static moment curve as-
suming the beam simply supported at A and Band
C -2025f .322 carrying the load of 100#. Fig.All.29 shows
)( 1/3 f 2025 the other two parts, namely the triangular mom-
~-421 ent diagram due to the unknown moments MA and
)( 1/3 210 MB which produce fiXity at the two ends. In
~-93 this figure MA and MB have arbitrarily been taken
)( 1/3 47 Fig. All. 26
~ -30b"
as 100, instead of unity.
~ 15 Fig.All.30 shows the Mil diagram for the 3
b4~ moment diagrams of Fig.All.29. The beam is di-
1743 vided into ten equal strips, and the Mil curves
are obtained by dividing the moment values at the
Joint D is considered next: The unbalanced mom end of each strip or portion by the corresponding
ent of (2025 - 712) =
1313 is balanced by dis- I value from Fig. All.27.
tributing - 1313 to the connecting members. From the conditions of fixity at the beam
To member DB 1313 x .678 =-
892 and =- ends, we know that the slope of the beam elastic
the remainder of -421 to DC. The carry-over curve is zero at each end. Likewise the deflec-
moment to B is - 446. The carry-over from C to tion of one end of the beam away from a tangent
D is one-half the balancing moment B 421 =- at the other end is zero. Stating these facts
but of opposite sign or 210, due to symmetry as in terms of the moment area principles, we ob-
explained before for member AB. tain
Returning to joint B, the unbalanced as
recorded below the single lines is (233 - 446)
=- 213. To balance 160 is distributed to BD
and 53 to BA. Carry-over 80 to D and bring
joL M~. 0 - -(5) (Area of Mil diagrams
equal zero)
over from A to B .5 x-53 =- 27. Continue this
process until joints A and D have been balanced
4 times or 4 cycles have been completed. The
final moments are shown below the double lines.
Fig.All.26a shows the resulting moment diagram
on frame.
IL -(6) (Moment of the Mil dia-
Mx,dx =0
gram as a load about
either end equals zero)
(Note: Since E is usually considered constant
it has been omitted from denominator of the above
equations.)
TABLE All. 3
IIA liB
Static M/I curve Tr1al- curve Trial-curve
I I
Beam 110m. Mom. )10m..
Portion Av. arm about Av. arm Av. Arm 110m.
Fig. All. 26a Ord. left left Ord. left yx Ord. x yx
= y end = x end ... (y) ,~nd x • y
yx
All. 10 Continuous Structures with Members of
1 107 2.67 286 110 2.06 227 6.7 2.67 18
Variable Moment of Inertia* 2 397 6.31 2510 133 6.07 807 24.9 6.27 156
3 770 10.16 7800 143 10.0 1430 48 2 10 16 488
In Arts.All.3, 4 and 5 consideration was 4 1120 14.10 15800 130 13.94 1810 70.0 14.10 990
given to the derivation of expressions for 5 1440 18.07 26000 110 17.96 1976 90.0 18.06 1630
6 1760 22.06 38900 90 21.92 1970 110.0 22.05 2430
fixed end moments, stiffness and carry-over 7 1615 25.88
29.77
41600 67.3 25.88
29.76
1740 123.5 26.04 3480
8 975 29000 40.7 1215 117.0 29.94 3490
factors for beams of uniform cross-section. 9 420 33.65 14180 17.5 33.68 590 91.0 33.88 3080
Many cases occur in routine design where mem- 10 100 37.33 3733 4.2 37.33 157 62.5 37.86 2360
*iy 8504 845.7 743.8
bers have a variable cross section. This ~yx 179809 11922 18122
article will illustrate the calculation of the i=~E· ~ x .11922 .14.1" x .18122 _ 24.4"
i 8504 845.7 743.8
fixed end moments, stiffness and carry-over • 21.15"
factor for a beam With variable moment of in-
ertia. The effect of axial load on these fac- * Actual area is 4 times the results shown. For calculation of
fixed end moments S~c.o. factors only relative values are
necessa~y.
*Also see Chapter C2, Part L.
P identical to Fig. All.30c. For

Fixed l
A
tllOO# B
~ Fixed the conditions of support assumed,
the deflection of A away from a
f-ol-o------- 24" .. , .. l6"----_.j1 tangent to the elastic curve at B
2.0 in.
Fig.
is zero. Thus by the deflection
principle of area moments, the
moment of the Mil diagrams of
I Curve
Figs. All.30 (a and b) about
(10) end A equals zero. Thus,
845.7x14.1+ 7~08 MBx24.4=0
800 960"#
Fig. All. 28 ~~~-~~-~~--r:==_1~

~~
or
MB = - 66"#
Since MA was assumed 100, then
Fig. All. 29 carryover factor from A to B =
-66 = - .66.
Trial MA
= -100 100
To find carryover factor from B
to A, take moments about Band
IM curve equate to zero.
Hence
00
845.7
(a) 743.8x 15.6 + 100 MAx 25.9 = 0
whence
MA = - 53.2"#
(b)
Therefore carryover factor B to
A = - 53.2 = - .532
Fig. All. 30 100
(Note: For the moment sign con-
vention used in this book carry
(c) o over factor would be plus.)
(C) Calculation of Beam Stiff-
ness Factors
24.4 When a beam is freely sup-
ported at one end A and fixed at
743.8 the far end B, the stiffness fac-
tor at the A end is measured by
Table All.3 shows the calculations for computing the moment necessary at 'A to produce unit rota-
the areas and the centroids of the three Mil tion of the elastic curve at A.
diagrams of Fig. All.30. Substituting values In Art. All.4 it was proved for beams of
from Table in equations (5) and (6). uniform section that EGA = MAL/41 or MA = 4EIGAI
L. In a continuous structure at any joint all
845.7 M + 743.8 M + 8504 = 0 _ _ -(7) members have the same G, thus 4EG is constant
100 A 100 B
and the stiffness K of any prismatic beam is
11922 MA + 18122 MB + 179809 = 0- - -(8) proportional to IlL. For beams of variable
100 100 section the stiffness factor K may be written:-
The value 100 in the denominators is due to the K = clolL - - - - - - - - - - - - - - - - (9)
fact that trial values of 100 were assumed for
where 10 is the moment of inertia at a particu-
MA and MB' Solving equations (7) and (8) we lar reference section and c is a constant to be
obtain; found for each non-uniform member. Thus for
MA = - 315"# and MB = - 785"#. non-uniform members
(B) Calculation of Carry Over Factors EGA =MALI4cI 0 - - - - - - - - - - - - (10)
To determine the carryover factor from A to By the moment area slope principle, the
B consider end A as freely supported and B as slope at A when B is fixed equals the area of
fixed. A moment MA = 100 is applied at A which the Mil diagram between A and B.
produces the Mil curve of Fig. All.30b. Due to Taking MA = 100, MB was found to equal
the deflection of the beam under this loading a -66"#. Thus EGA = 4(845.7- .66x743.8) =1419.
restraining moment at B to cause tangent of The value 4 is due to the strip width since
elastic curve to remain horizontal will exist. true area is wanted and only average ordinate
This unknown fixed moment at B has arbi trarlly was used in Table All.3.
been taken as 100, thus producing a Mil curve Equating this result to Equation (10)
All. 17
1419 = 100 x 40
4c xl
, whence c = .707 Example Problem 1.
10 was taken as the value at the A end or 1. Fig. All.33 shows a single bay rectangular
(See Fig. All.27). bent carrying a distributed side load on one
Therefore KAB =
.707 IA/L leg as shown. The K or I/L values for each
member are given on the figure.
Similarly for end B:- Under the given loading it is obVious that
100x40
EGB =
4(743.8 -' .532 x 845.7) 4c xl = the frame will sway to the right or in other
words, joints Band C will undergo considerable
whence horizontal movement. The moment distribution
c = .85 and KBA = .85 IA/L method assumes that only joint rotation takes
place. To make this assumption true for this
All. 11 Frames with Unknown Joint Deflections Due structure we will add an imaginary support at
to Sidesway * joint C which will prevent sidesway of the frame
as illustrated in Fig.All.34. The end moments
In the example problems so far treated the in the frame will then be found by the moment
joints of the particular structure were assumed distribution process. Fig. All.34 shows the
to rotate without translation or With a definite results of this process. To explain, the solu-
amount of translatory movement. Translation of tion begins With computing the fixed end moment
the jOints may however, be produced by shorten- on member AB = WL 2/12 = (300 x 25 2)/12 = 15.63
ing and lengthening of the members due to axial thousands of foot Ibs. This value with the
loads and by lateral sway due to lack of diag- proper sign is written at the head of a column
onal shear members. The problem relative to of figures on member AB as shown in Fig.All.34.
the effect of joint translation due to axial Now considering Joint B, the unbalanced
stresses is treated in a later article. In joint moment of 15.63 is distributed as follows:-
this article only the effect of sidesway of To Member BA 15.63 (40/190) =- 3.27. Carry =-
rectangular frames on the frame bending moments
will be considered. over to Joint A 3.27/2 =-
1.63 =-
Fig. All.31 illustrates cases of frames To Member BC =- 15.63 (150/190) =- 12.36. Car-
where only rotation of joints takes place (neg- ry over to Joint C =- 12.36/2 =- 6.18
lecting axial deformation) whereas Fig.All.32
illustrates conditions in Which sidesway takes
place and the joints suffer translation as well
as rotation. 4 6'1' 70'
'1 6' r--
~T
B 1= 10, 500 C
I
K=r;=150
0
i
"-
0
'<t' g II 25'
'lIo
0
..,
0
;;-"":!:J
:.::
D
rm
Fig. All. 33
Symmetry of Structure Frame prevented
& Loading from sidesway by
extern I restraint
Fig. All.3l
~B--------"'''''8LRC
~: l~j~I~II:
Symmetry
of Loading
but Structure
;:I~I~I~II;
+ I
~
I
Gl
I
t')
1+
~
(O.OO)f
-12.36 b1.------+
+ 2.44
- 1. 93 b ..--.....
f (0. 00)
-6.18
b 1. 4.88
-0.96
~+ + + + +
_ B ~.. ~.,~...
-~~~~
Symmetry of Structure,
Loading Unsymetrlcal
Fig. All ...32
Unsymmetrical
111 1
+ 0.38.
- 0.30 b 3
+ 0.06.
- 0.04 b..
~
b .. +0.76
-0.15
b 3 +0. 12
-0.02
jjj j
There are several methods of determining
-11.75 b.. +0.02
-1. 53
Sit) 0.. . 0l~
OCO..-4oor:-
00000 .
the bending moments due to sidesway. Only one - .... ++++
method will be presented here and it can best D
be explained by the solution of example prob- Fig. All. 34
lems.
* Also see Chapter C2, Part M.

3517
11750
~ B .........
1530 horizontal shear reaction at the upper end of
L
:=...:.-.+ .II =-+ ~i) +--- - -Roc· 3606 lb. the vertical legs of the frame equal to 3606 lb.
89 C
11750~~B=3517 1~0
Bending Moments Due to 8idesway.
T
B HC=8tC Assume the structure sways sideways as
illustrated in Fig. All.36, but with no
rotation of the upper joints. It was proven
25" in Art. All.8 that the end moments produced by
30O#/in
1 r _A
7570
D'J~
....., 770
i
the lateral movement of one end of a beam whose
ends are fixed against rotation are equal to
6EId/L 2 where d is the lateral movement. In
Fig. All.36 the load P causes both vertical
members to suffer the
Fig. All. 35
same horizontal dis-
placement at their
upper ends, hence their
end moments due to this
Now consider Joint C:- displacement are pro- Fig. All. 36
The unbalanced moment is -6.18. To bal- portional to EI/L 2 •
ance we distribute to CB 6.18 x 150/190 = =
4.88. Carryover to B 2.44; to CD = 6.18 x = Since EI and L are
40/190 =
1.30. Carryover to D 0.65 = the same for each vertical
Sidesway Without
The balancing and carryover procedure is now member in our structure, Rotation of Joints
repeated for joints Band C, until the unbal- the fixed end moments will
anced moments become of negligible magnitude. be the same magnitude for
Fig.All.34 shows that 4 cycles have been car- each member. Therefore, for convenience we will
ried through. By keeping the ratio 150/190 set assume fixed end moments of 10,000 ft.lbs. are
produced by the sidesway. We now use the moment
on the slide rule the unbalanced moments at distribution process in permitting the upper
joints Band C are distributed and chased back joints to rotate as illustrated in Fig. All.37.
and forth as rapidly as one can write them The procedure is similar to that in Fig. All.34.
down. Since joints A and D are assumed fixed, For example the solution is started by consider-
they absorb moments but do not give out any.
ing joint B. The unbalanced moment is -10.
This is balanced by distributing 10 x 150/190 =
Fig. All.35 shows a free body diagram of 7.89 to BC and the remainder of 2.11 to BA.
each portion of the frame. The end moments The carryover moments are 7.89/2 3.95 to C =
are taken from the results in Fig. All.34.
Consider member AB as a free body. To find and 2.11/2 =
1.06 to A. Due to symmetry of
the horizontal reaction HB we take moments loading and structure, the distributing and
about point A. carryover moments at joint C will be same as
at joint B, hence it is needless work to show
calculations at these joints. The carryover
m A = 300x 25 x 12.5 + 11750 -17570 - 25HB = 0 moments from C to B will be identical to those
from B to C. Fig. All.37 shows the 5 cycles
hence HB =87930/25 =3517 lb. have been carried out to obtain the final end
moments as shown below the double lines.
Now consider member CD as a free body. + 7.89 b
To find HC take moments about D. + 3.95 -1/2 Due to symmetry
- 3.12 b of loading and
~MD = 1530 + 770 - 25Hc = 0, hence HC =89 lb. ·1.56 -1/2 structure it is not
+ 1. 23 b necessary to con-
We now place these horizontal forces on the .. 0.62 -- sider right hand
- 0.49 b portion as calcu-
top member BC as a free body as shown in the -0.25 -- lations will be
upper portion of Fig. All.35. The unknown ,. 0.20 b identical as shown
imaginary reaction RC at point C that was + 8.47 for left side.
added in Fig. All.34 to prevent sidesway is
= ~ 51~t:I~I~IE B C
also shown. To find RC take ~Fh 0
+~
II~
0
0
ZFh = 3517 + 89 -He = 0, hence He =3606 lb.

I + I +
.... ..c..c..c..c..c
I
0
....
Since the external reaction of 3606 does Fig. All. 37
not eXist, we must eliminate it and find the
bending moments due to the sidesway of the
II~
0
frame. In other words, the frame will sway .... 0
0
sideways until bending of the frame develops a 0 0 " " ' .... 00«1
o co «I co co ""I"" ....
~...:ooooa)
I
+ +
A
I ... I I I D
All. 19
Fig. All.38 shows a free body diagram of 100

the frame vertical members with the end moments
as found in Fig. All.37. The shear reactions ~8" 8" ~
HB and He are each equal to (8470 + 9230)/25 =
708 lb. These shear reactions are then placed
on the free body of member BC in Fig. All.38.
T B 1=16 C
I
12"
For equilibrium the summation of the horizontal 1=10 15"
forces must equal zero. Thus P 708 + 708 = = 100 lb.
1416 lb. and acting to the right as shown in 1=24
D _1
Fig. All.38. Since the reaction He -3606 in = Pin
Fig. All.35 must be liquidated and since the 12"
~
liquidating force P produced by the moments in
A
Fig. All.37 is 1416 lb. it is obvious the \ Pin Fig. All. 41
values in Fig. All.37 must be multiplied by a
factor equal to 3606/1416 2.545. Therefore, = SOLUTION:
the final bending moment values equal those of
Fig. AIl.34 plus 2.545 times those in Fig. Relative Stiffness Factors: -
All.37. Fig. AIl.39 shows the results and
Fig. All.40 the final moment diagram. KBA -- L31 3 x 24 3
24'
= KBC = 41
L
= =4 16x 16 = 4
K ~D
''\.;
=31L = 3 15
x 10 = 2
708 i t
- - -..B.--------------;;-=:-iC- P=1416 lb.
The distribution factors for each member at
708
8470 joints Band C, which equal K/~K are recorded in
~B=708 ---+ 8470 the small D on Fig. Al1.41a.
T
25"
HC=708
Fixed End Moments: -
For Member AB = PL/8 = 100 x 24/8 =300 in .lb.
Fig. All. 38
= PL/8 =100 x 16/8 = 200
1
For Member BC
A D As explained in example problem 1, the
9230 9230
moment distribution is carried out in two steps,
one for joint rotation only and the uther for
effect of sidesway or horizontal translation of
- 11. 75 - 1. 53 joints Band C. Fig. All.41a shows the moment
+ 21. 56 + 21. 56 distribution for no sidesway by placing an
+ 9.81 + 20.03 imaginary reaction He at joint C. The process
is started at joint A and the order of joint
B
",CO", c balancing is ABCBC. As soon as a joint is
Fig. All. 39 t-1t)CX) ll)IOO balanced the carryover moments are immediately
. COil'"
. . . . . Cl)
It)
"'''''1
"';"";0
""IN carried over before proceeding to the next
-+- I I + I , joint. When a joint is balanced a horizontal
line is drawn.
~~II~
F.E.M. - 200 200 F.E.M.
BJ, - 143 - - - 71.5
+ ,
- 42. 8 - - 85. 7 B J,
II II
I
II
D Bs 24.4 - ----r2.2
- 4.1 - - - 8. 1 B s
B.. ~ - ----r:t
-363.2 ---=...Q...1 B"
..
::!l 47.3
I:il
~ ~ ~ ~ ~ ~
~ ~~~ ~~~
~I~I~II~
41. 07 22.73
Fig. A1l.40
Example Problem 2. Bent with Unequal Length j 1j
Legs and Pinned Ends. 00011 0
Fig. All.41 shows a loaded unsymmetrical
frame. The final bending moments at Band C A Fig. All. 41a
will be determined.
With the end moments known the reaction B,,1l9 - 5 9 . 5

He can be calculated by a consideration of the 54. 4 ~108. 8 B"
free body of each member as shown in Fig. B& -3l.0~-15.5
Al1.4lb. -s:2---- 10.3 B s
B" =---.3....Q ----+ =--t.5
-.l...J!. B"
6~ B;;:-------:::.-- .----Rc = 61. 85 144.6
162.6 ... . "
3.15 C III III III
ril ..... "
363. 2 THB=65 HC=3'1~CD~147' 3 rz: III III III
C ~~~I!'>~~1~111~.... ~
<- :1:111:....
::;
.....4OC')lN
""IN
~ B
""IN
I i 111 '
10~
-3.15 'jll[' D ~~Iooor

. ...
::Sill
35------, A
~ ~Iooolio A rzi
Fig. All.41b
.... Fig. All. 41c rz:
::Sill
rzi
To find HB take moments about A, rz: 6.0L __ P=6. 02 ~ 10. 85
100x12+362.3-24HB = 0, HB = 65 lb. B 1tT.85 =16.87 lb.
144.6~ 10'85~I162'6
~.02 C
To find HC take moments about D,
-47.3 + 15Hc = 0, HC = 3. 15
D
Placing these shear reactions on member BC
and writing equilibrium of horizontal forces A Fig. All. 41d
we find that RC =
65 - 3.15 = 61.85 lb. We
now must liquidate this reaction RC and permit To liquidate the reaction RC 61.85 it =
frame to sidesway. will therefore require 61.85/16.87 3.67 times =
the moments in Fig. All.41c. Therefore the
We will assume that the frame is fixed final bending moments equal those in Fig.
ended and that an unknown horizontal force P All.41a plus 3.67 times those in Fig. All.41c.
at C will deflect the frame sideways. The Fig. All.41e shows the results. The bending
fixed end moments for equal horizontal de- moment and shear diagram and the axial loads
flection of Band C will be proportional to now follow as a simple matter of statics.
EI/L& of the vertical member. -363.2 47.3
530.0 ~
I -- 24&
Fo r membe r AB , 1:2 24 -- • 0417 166:8 64IT
B
'"~~II":
F or memb er DC , 1::2 10&-- •
I = 15 OA A t::
'±""hJ
"'0'"
.., ..,I!'> '"....
I I
For convenience we will assume 417 in. lb. as

the fixed end moments on AB and 445 on DC.
The assumed fixed ends will now be eliminated
by the moment distribution process as shown
in Fig. All.41c. The order of joint balance Fig. All. 41e
was ABDCBCBC.
Fig. All.41d shows the shear reaction on Example Problem 3. Bending Moments in Truss
the vertical members at Band C. These forces Involving One Panel Without Diagonal Shear
reversed on the top member show an unbalanced Member.
force of 6.02 + 10.85 =
16.87 lb. Therefore
=
a force P 16.87 was necessary to produce the Frequently, in aircraft structures a truss
bending moments that resulted on the frame is used in which a diagonal member must be left
due to sidesway. out of one or more truss bays. The external
shear load on such bays must be carried by the
truss chord members in bending. Fig. All.42
shows a 3 bay truss with no diagonal member in
the center bay. The bending moments on the
truss members will be determined for the truss
All. 21
loading as shown in the figure. Then the distribution factors are: -

10
te
10 To Member cb =4/7 - .57
td
3 = 1/7 = .14
® ~
T To Member cf
CD CD CD @ 60" To Member cd = 2/7 = .29
@
f
@
e ~ In this problem there are no loads applied
to the members between joints. The external
shear load on the truss to the right of the
1-- 60" -+- 60" --+- 60" ---I center truss panel which equals 10 + 10 20
must be carried through the center panel by the
=
truss chord members in bending. This bending
Fig. All. 42 causes the center truss to sway or deflect
until a resisting shear force equal to 20 is
SOLUTION: developed.
The relative moment of inertia values We will assume the truss center panel is
for each member are given in the circles on fixed at joints b, c, f, and g. The right end
the truss in Fig. All.42. of this assumed fixed ended truss will be given
an upward deflection. This deflection will
The distribution factor to each member cause fixed end moments in members bc and gf
at each joint is then computed and equals which are proportional to IlLS for each member.
K/~K. The values are recorded in thec=J Since L is the same for each member, the fixed
on each member in Fig. All.43. The stiffness end moments will be proportional to I of the
factor K is proportional to IlL for the member. member.
For example for joint c: -
1 7 Ibc =4,
+-=-
60 60
be eb cd de
100 F. E. M. F.E.M. 100
-13.6 -27.2
-28.5
-27.2 . -57
-13.6
-29
5.9 . -14.5
11.8
TO
9.8
---=1.4
4.9
-2.8
4.0
-2.8
~
1.4
4.1
-0.9 . -L 9
0.2
1.0 0.5 0.5 0.9 0.5
-4.4 -0.3 ~ '36.9 -19.4 -0.2
-24.9 45.7 -2.6
ah .10 bg cf de
6
1-:4 '-?~ -7.1 -14 bQ.
2.4
U
~O .@
-3.4
-0.8 fe
~4
2.0 '£?(
0.5
-0.5
~'O
0.4
~'<7
0''/ -3.6 -0.4 -7.0 0.7 <r '~
~'S
~
----0:0
ha
4.4 \9 ,;>
'0
-6.7
-0.4
-0.1
-11. 8
-12.9
1.0
0.2
~
-17.5
&'0'
0''/ ~'S
0'6'
~
3"":lf
h!! -0.9 ~ ,~ ed
1.2 ~
's 0 0.1 0'<7 '~./ 1.5
0.3
'~
~·o -0.1 0.3 '~ M
4.3 O'S -11. 7 -17.2 -0.3
'qy lhQ
0.0
.::...1.
2.0
gh .40 fg .58 .5
hg
.
75 F.E.M.
-19.7 . F.E.M. 75
-39.4 -6.8
ed
-3.4
-1~2 -24.4 -20.6 -10.3 0.5 --.hQ
10.3
-2.6
--.L...l
. 5.1
----.b..l
0.6
2.0
-2.7
0.4
~
-1. 4
0.9
0.6
0.0
0.2
0.3
0.0
--.Q.....l..
28.7 -5.5 -0-
-3.4 -----.lL....Q ~
-22.3 34.0 Fig. All. 43 -2.0
Therefore the fixed end moments on member to axial loads produces translation of the
gf will be 75 percent of those on member bc. joints of the structure. Since the angles be-
tween the members at a joint remain the same
Assume 100 as the fixed end moments on due to continuity, this translation of joints
bc. Then a 75 x 10 =
75 is the accompanying bends the members between joints. (2) The
fixed end moment on gf. bendirg moments on the members due to external
joint or lateral loads or those due to joint
We now remove the imaginary fixed supports rotation produce lateral deflection of the
on the center truss and let them rotate to members between joints. The member axial loads
equilibrium by the moment distribution process times these lateral deflections produce
as shown in detail in Fig. All.43. secondary moments. These secondary moments can
be handled by the general method of moment
The first step is to record the assumed distribution however the stiffness and carry
fixed end moments of 100 at each end of member over factors and the fixed end moments are not
bc and 75 at each end of member gf with due constant but become functions of the axial
regard for sign. The moment distribution pro- loads.
cess will be started at joint C. The unbal-
anced moment is 100 or -100 is necessary for Fig. All.45 shows a prismatic beam simply
static balance. Using the distribution fac- supported at A and fixed at B, with a moment
tors on joint C, we find -57 goes to cb, -29 MA applied at A and carrying an axial compres-
to cd and -14 to cf. Short lines are drawn sive load P. Sub. Figure a, band c show the
below each of these numbers. Fifty percent of 3 parts which make up the moment diagram on the
these balancing moments are carried over to beam. Without the axial load P the portion (c)
the far end of each of these members. This would be omitted.
process is repeated at each joint until the
remaining balancing moments are negli~ible.
p--+(M~__ ~ J__ B
In Fig. All.43 the order of joint balance was
Q Y
.?--~, -) p
cfdebghacfdebghacfdebg. If the student will I
I : MB
follow this order he should be able to check I
I I
the figures in Fig. All.43. I
'''' r
f
.'/>///////7/>,~
....• \a) MA~=:: :=! :: ... - - ...
I
I
I _
!~MB(b) free bodies of the truss

members bc and gf with
,&PY
I
I
!
end moments from Fig.
~~: (c) All.43. The shear re-
action at ends c and f
by statics equals
MOMENT DIAGRAMS Hc + Hf = 36.9 + 45. 1
60
Fig. A'll.45 + 28.7 + 34 =
60
By definition the stiffness factor of beam 1.377 + 1.045 = 2.422 lb.

b. AB is measured by the necessary moment MA to
es produce unit rotation (9:: 1) for elastic 11ne at The external truss shear at 111
point A. By the moment area principles, the Therefore it will take 20/2.42:
slope at A equals the shear at A due to the the final bending moments as f
moment diagrams acting as a load on a simply All.43 to develop a bending sh
supported beam at A and B. For a given unit of 20 lb. Thus the final mome
slope at A it is obvious that the required MA times those in Fig. All.43.
'fect is less When the load P is acting as it pro-
'ness duces a moment load in (Fig. c) of the same sign The solution as given neg
.ins as the moment load due to MA' Thus the stiff- of axial loads upon the value
ness of AB is less when carrying an axial com- and carryover factors. Art.
pressive load compared to that withoMt it. For how to include these effects.
a tension axial load diagram C becomes a nega- All. 12 Effect of Axial Load on Mornen'
)f tive moment diagram and thus MA must become
> a larger to give unit slope at A. Or in other In the previous articles
,xial words a member carrying a tensile axial load axial loads upon the member en
con- has a greater stiffness than one without axial continuous structure was negle
lOrt- loads. It is evident that the moment diagram loads produce bending in the IT
~s due
due to P will also effect the magnitude of any tinuous structure in two ways.
ening or lengthening of the va
All. 23
fixed end moments and alao the carry over fac-

tor.
.. 0
1. 8
All.13 Fixed End Jloaents, Stiffness and Carry
over Factors for Beaa Columns of Constant
Section * 1.
•
In deriving expressions for fixed end mom-
ents, stiffness and carryover factors the beam ··COIlpr. . .loa
columns formulas of Chapter A5, must be used

1.
•
in finding slopes, deflection and moments.
Mr. B. W. James in an excellent thesis at 1.
•
Stanford University and later published by N.A. /
C.A. as Technical Note #534 has derived expres- 1. 0
/
sions for these factors and has provided graphs 1/
ffl r.t1 ;II( 11
of these factors for use in routine design. T.A "'5~1'
Figures All.46 to All.56 inclusive are taken 8
/
from this thesis. The use of these Figures can V
best be illustrated by the solution of several
problems. .. .-----/
V
--....... ---.
All.14 Illustrative Probleas:** .. r---. r--- ~ 'IenaiOD
Example Problem #12 r--- I-- e--
Fig. All.57 shows the same continuous beam •
as used in Example Problem #4 (and solution #2).
For this example it has been assumed that such
axial compression loads exist in spans BC and CD L/l • a
as to make the term L/j = 2.5 and 2.0 for these carry-over factor. C = 2_
spans respectively, where j = vEI/P. Due to the Fig. All. 46
axial loads, new values of stiffness and carry
over moments as well as fixed end moments must
determined as follows.
Span BC .. 0
I 1
K = L = 96 = .0104 1. 8
/ /'
From Fig. All.47 when L/j = 2.5, correction fac- //
tor for stiffness factor = .775 when far end is T_lolL, far n4 tlUll ~'~
/
fixed and .36 when far end is pinned. Hence,
1.
•
/
KBC = .0104 x .775 .00806 1
1.
• // /'
KeB = .0104 x .36 .00374
From Fig. All.46, when L/j = 2.5 the carryover 1. 8
/ ./.. '1'. . 1011.. far eD:l planed.
factors are CBC = .73 and CCB = 0 1f B is con- / V

sidered freely supported.
From Fig. All.48 when L/j = 2.5 fixed end mom-
ents for uniform load = wL 8 /lO.67 = 1 x 96 8 /
1. 0
--- ~
.._/V
/
V
10.67 = 865 in. lb. 8 "'-

--
~
k
Span CD
•
~ '" \
K = I/L = 1/72 = .0139, L/j = 2.0 '\ \ --QOIIP!'•••lOR, tar eDd fixed
From Fig. All.47 correction factor = .86, hence \ \

KeD = KDC = .86 x .0139 = .01198. Carryover • --'
factor from Fig. All.46 = .62. Fixed end moment
Oozapr•••lon,
tar ad plGn~
_....
\
from Fig. All.48 = WL 2 /l1.2 = 1 x 72 2 /11.2 = 462 • \ \
The distribution factor at joint C equals \
(.00374/.00374 + .01198) = .238 to CB and the
\
\
\
remainder .762 to CD. 0 1
• ',.1 • • 8
L/l
The balancing and carryover process is
s1mllar to that in Example Problem 1/4. Stiffness factor coefficient. K c !L
Fig. All.47
* See Chapter C2, Part L for members having a varying EI.
*~ Additional examples in Chapter C2, Parts G, Land M.
1 I I / I I II
-
"C:~~~
_.lb./lu.
"Ad • • ·b.. .
17
-
,. I r----
-----" T
'"
/
v
~-
as
~- c-- 1\
.11.- -lIB- ~a
/ \ 1II.l.~2
15
/
/
-
\
C'''1
•
.
_+c
-~
L
-
--- "--
v'"
~~r-
•
V
~--'I'e»loll
\
•
V
/ 1\
\
a
----I'-----
r--- ao
--- r----
"---
1
0
"- ___ Compression
\
1\
•
6
-
'" I~ -"s
1"\
~-r-- 1\
• •
\ .. \
• ,-r---- 2
\ L/l
•
L/l •
Fixed~end moment coefficient. Fixed-end moment coefficient.
Uniform Load. Uniformly varying load.
Axial -compression.
Fig. All. 48 Fig. All. 49
Q I 1 1 I I I
50
_
'. c:~
~ II •.,lb./u·i
I
~)'. i·.·-~~oi
--l-J[~-2r
~~
I
Jr.... "'
ID ~ ~.--J
I
3r-~~
I
L/2
"t L/2 -+-+
~~+-
I
~
I
,
I
i
V
-t---
f--r---i 'ii af--- j-
i
f-
:'::A '" -liB· ~ !
I I /
TT
~-
io~
--
50 -,"- -~~
T. .
/
I -.-, \,V
/ 1
V V
2
V 0
1/
V 1/
:IS
0. V
V /
V
// 6
----r---- V
0
_v v
7
KCoepr...l0D
./ ~
V ti '\
/' 1\
~, ~ .0. \
---~ "
~ I-- --- \
18
0
til• • •
L/l
Fixed-end moment coefficient. Fixed-end moment coefficient.
Uniformly vary.ing load. Concentrated load at mid-span.
Axial tension.
Fig·. All. 50 Fig. All. 51
3. <
3. 3
f----- I---
.~rn
Vi.
.JI Ib./ln.
I
Fig. All. 55
! I
0
.
~J'
,
TTITT'
.~.
h
· · I +-171--
-fp
-;;; -'V<,·
I----- _,lb./in.
•
Fig. All. 54
0
+-----_
r. t-
I
~a(~la[6-B(~)+3(~)a]0i
L V ~ ~
'i' 0
>0
3. a ,< 1/ ---
I..---- f--" - <••
/ V
<.0
, "'J
a
a.
o
co
....
a
....
N
....
"
....
(])
....
co
r"
o
I I -
1.. / r!G'
.3
Il::I
II ------- . . _ ..tIo"(A)3[<\..) _ .(..)3] 0B / 1/ /~.
tf'
......>
13 L L L
LL/.
1/ V v ~ 3•• J=
L ./ <.0 ,a
V I..---- / /V V C11
~+-+----1
~
-- "" '"
-
./ 3.0
';) c.
Os
V / 'V J.----- t'!l
... g j V
--- <'.5
~. 77/
~
"'1~ ~l>:ol
~
11
~ 1/V
--
1. < V j.---
--V e----- f.- I--- I I 1\ \ 1\ \ \ "", _ '----'N 0'
j ,~ ~~ e------ f.- f.-

a.o
II \ '\.
1:"10'
~
oN
\1:/
_
1. 3 - '.0
~ ~ ..::- r-----
1.0
- -
0-
J,
0'
(])
/
/ Ii \ \ '\. ti'
e------
-- i-----
3 ••
3.0
1.0 ,g ~ ~ ~ l -I -t - ;~ .-t I i I 1\ I\ \ I'",

~
--- -! I \1 \ 1\
L 0
- ~~ I'---- I'----
... (.l / 1/ 1\
-
-... • '" ~ I'---- t----..- i
~
~ .I I 1
• \...-
------- - r-----
g
!
~t--- --r---.. r---..
'------ I'---- ~ ~- I I
II
/
I
I \ ~- I\
\
1\ \
•
.1 .3 .. - •< ..
7
.1 .2
./L
.3 .<
t--
.. ~..... $-o· i!I' 4- I\ \

I
\ 1\'
I\
~ $- I I Ii
./L CD u'
... \ \ \
Fixed-end moment coefficient Fixed-end moment coefficient _
for MB. Uniform load over for MA. Uniform load over
miW-++-!- Jk~ E- 1\
o
part of the span. part of the span.
~
o \
~ ""';~
o
CD
, ~I" ~
"~I
• r.~--.01
.,
(0/L)(0/L)3 .... (./L)3(0/L) .p. CN t-' II
..... VJ If:o ("
~ ('1 l\)
-f--- .0 ,
;, ;.. . -. • . .' . .
- ~ ~ 9 j.. ;.. ;.. ;.. rt 00000 OOC5lp _
Pl I~tg' i -H
rt
CD
• I
I'":;< 1'1
'1:J
0-
f-'
l\-\+- g aW- I I I I I Y I II I· I I
o ..... '"
1
~tmi1lT1I111
~
"'J
4Q' . ooh
......> "
~~
~
."1 17
C11
c.> ~~
~~
-t.
....
~ . ~ § ~
o
~ ~
o
~'r'i1 l> ~ l:
~.1l f'al!)a[ot- 3(!)'
~
~,
~ 2 ~
.l.(B)a
~
r. _.. + 3(B)31
~ g .
o g ...>...
~ 12L II -L LJ
""
C11
Jr---rI --t,--- I !,
1. of---- I-~
~I
+-----c--,
! :
/
I
1.
• /
1/
1. 4
/
1/
1. s
J. .. Coapre••ioa
t. a 1/
/
:" v
I
1. 1
o/,v
Fig. All. 59
1. 0
-----r----
V
~
Fig. All.59 shows the total beam axial loads in
the various portions of the beam by adding the
•
"""'0 .ralWloll
values shown in Fig. All.58. The outer span AB
due to the axial loads induced by drag wires of
0
"" "-...
~
drag truss is subjected to a varying load. As is
customary in design of such wing beams the axial
load in span AB will be taken as equal to the
average load or (-9571-8565-7559)/3=-8565#.
a L/l
Culumn di'Stribution factor.
S
2~
""'"
- a.
The beam bending moments at the support
points will be determined by the moment distribu-
tion method.
Fig. All. 56 Calculation of Factors:-
w= l#/in L/i = 2. 5 1 is constant, L/i =2. 0

Span AB
I = 17.21, L = 100, I/L .172
E = 1300,000, P = 8565
Fig. A11.57
51. 5"----jo---96" ---->1+---72"
j =~=~300,0~~6~ 17.21 51.1
Stiff. Fact. K o .00806 .00374 .01198 .0119800
Distribution Fac. o 1 .238 .762 o1
C.O. Factor o .73 o .62 .62 L/j = 100/51.1 = 1.96
Fixed End Mom. -883 865 -865 462 -462
Ba1anc e Joint B o 18 From Fig. All.47 when L/j = 1.96, correction
C.O. to C 13.2
Bal. Joint C 146.0 243.8 factor = .86 for fixed far.end and .52 for pinn-
Carr -Over o 148 ed far end. Therefore
Balance D
Carr -over o KAB .52 x .172 .0895
Final Mom. -883 883 -705 8 705 8
KBA = .86 x .172 .148

Comparing these results with those of Problem #3 From Fig. All.46, when L/j = 1.96, the carryover
we find moment at C is increased 15.7 percent factors are:-
and that at D is decreased 8.5 percent. For C.O'BA = .62, and zero from A to B since B
larger values of L/j the difference would be is considered in its true state or freely sup-
greater. ported. From Fig. All.48 when L/j = 1.96 equa-
Example Problem #13 tion for fixed end moment for uniform load
Fig. All.58 shows the upper wing of a bI- WL 2
/11.25 = 32 x 100 2 /11.25 = 28420"#.
The overhand moment MBC = 60 x 26.6 x
plane. The wing beams are continuous over 3 (32 + 16)/2 = 38300"#.
spans. The distributed air loads on the front Center Span AA' :-
beam are shown in the figure, also the axial
loads on front beam induced by the lift and drag I = 22, I/L = 22/84 .262, P = - 11758
truss. The bay sections of the spruce are shown j =~,300,000 x 22 49.4
in Fig. All.58. The moment of inertia in each 11758
span will be assumed constant, neglecting the L/j = 84/49.4 = 1.70
influence of tapered filler blocks at strut
points. KAA' = KA'A .262 x .895 .234
C.O. Factor .585
All 27
Fixed end moment /1l.35 = 31 x 84 2 /11.35 = = WL 2

For member BC
P = - 27150, I/L = .0083, I = .332
19300"#
The moment distribution process is given
in Fig. All.60. If the axial loads were neglec-
j ::
9
:: 2.12
, 00~7~~~ x .33~/2 :: 18.82, L/j =1:. 82
0 e
ted, the bending moment, at support A and A'
would be 19480, thus the axial influence in- From Fig. All.47, stiffness factor = .82 x .0083
creases the moment at A approximately 7.5 per- = .00681
cent. From Fig. All.46 C.O •. Factor = .66
w·16 w = 32 w·32 w·16
w =31
Fig.A1l.60
.:..,-r T T T T T T Il--
C' B' A' A B
1.-60" 1<----100" 84" 100" 60"
K Value o .148 .0895 .234 .234 .0895 .148 0

Distri. Factor o1 .276 .724
C.O. Factor o .62 o .585 .585 0 .62 0
F.E.Mom. -3"300 28420 -28420 19300 -19300 28420 -28420 38300
BaI.B&B 0 8"0 _ -9880 0
C.O. -6125 -6125 ......
Bal. AI & A 828 2167 -2167 -828
C.O. 0"- -1265 1265 ....0
Bal. AI & A _ 347 918 -918 -347 _
C.O. 0- -538 538 0
Bal. A' & A -148 390 -390 -148-
C.O. 0..-- -228 228 --'0
Bal. 63 165 -165
Final Mom. -38300 38300 -20909 20909 -20909 20909 -38300 38300
Example Problem #14

Fig. All.6l shows a triangular truss com-
posed of two members fixed at A and B and rigid-
ly joined at C to the axle bar. Let it be re-
quired to determine the end moments on the two {gf/lLB
members considering the effect of axial loads on o A
joint rotation and translation. '"
;;.... ~
C:>
~25"~ ![I~
,,~~1 ~
Ll~ A
'"
. " .L
~""'..g:>
T
I •
~/ Fig.
o '
o ~
~ ~
C:>
C'l ,
Al1.63
: ~~ co ~ "--C
J
~ I'~ ....~,,:::~<f, -72000
~" .... ~
000
arl::0Fig.
L"-
All. 61
Effect on End Moments Due to Translation of
Joint C Due to Axial Loads.
4"r The movement of joint C normal to each
18000 member will be calCUlated by Virtual work.
Solution:- (Reference Chapter A7). Fig. All.64 shows the
Virtual loading of 1# normal to each member at C.
The magnitude of the axial loads in the The Table shows the calculation of the normal de-
members is influenced by the unknown restraining flections at C.
moments at A and B. To obtain a close approxi-
mation of the axial loads, the end moments in Table
the two members will be determined Without con-
SUaL SUb L
sideration of axial loads. Thus the external lIem. L/AE S Ua AE Ub AE
joint moment of 4 x 18000 = 72000 in. lb. at C
is distributed between the two members as shown AC .0000049 10400 1.89 .0962 1.6 .0816
BC .00000244 -27150 -1.60 .1060 -1.89 .1250
in Fig. All.62. With the member end moments .202 .207
known the axial loads and shear reactions at A
and B can be found by statics. The resulting Thus the deflection of joint C normal to
axial loads are BC equals .202" in the direction assumed for
PAC = 7360# and PBC = - 24240# and the the unit load, and the deflection of C normal to
shear reactions, SA = 198# and BB = 2332#. AC = .207".
With the axial loads known the modifted
beam factors can be determined.
All. 15 Secondary Bending Moments in Trusses with

Rigid Joints. *
B A
Often in airplane structural design trusses
with rigid joints are used. Rigid joints are
produced because of welding the members together
at the truss joints or by the use of gusset
plates bolted or riveted to each member at a
truss jOint. When a truss bends under loading
Fig. All. 64 the truss joints undergo different amounts of
Ua = 1# movement or deflection. Since the truss members
C C are held rigidly at the joints, any joint dis-
placement will tend to bend the truss members.
The fixed end moments due to support de- The bending moments produced in the truss mem-
flection equals M = 6Eld/L 2 as derived in Art. bers due to the truss joint deflections are
All.7. However, the secondary moments due to generally referred to as secondary moments and
axial loads times lateral deflection modify this the stresses produced by these moments as
equation. "James" has shown that the modified secondary stresses.
equation is,
Since fatigue strength is becoming more
6Eld I d important in aircraft structural design, the
Mfixed end =L2 (2~-a.)' i f K =Land R L' than
question of secondary stresses becomes of more
_ 6Krn importance than in the past. The moment distri-
M - 2~-a. ' bution method provides a simple and rapid method
for determining these secondary moments in truss
Fig. All.56 shows a plot of 2~-a. against L/j. members due to truss deflection. The general
For Member BC:- procedure would be as follows: -
R = .202/40 = .00505 K = I/L = .0083,
(1) Find the horizontal and vertical displace-
211-0. = 1.08 for L/j = 2.12. ments of each truss joint due to the criti-
Hence, fixed end moments due to translation of cal design condition. (See Chapter A7 for
joint C equals- methods of finding truss deflections).
MBC =MoB = 6 x .0083 x .00505 x 29,000,000 = -6750"#
1.08 (2) From these displacements, the transverse
deflection of one end of a member with
For Member AC respect to the other can be found for each
R = .207/47.2 = .00438 K = .0029, 2~-a. = .92 truss member.
when
L/j =2.42 and member is in tension. (3) Compute the fixed end moments on each member
hence due to these transverse displacements.
M -M_ _ 6x .00438 x .0029 x 29,000,000 -
AC -. \;A - •92 --
(4) Calculate stiffness and carry-over factors
-2400"# for each truss member.
The signs are minus because inspection shows
that the moments tend to rotate ends of members (5) Calculate distribution factors for each
counter clockwise. (Ref. Art. All.2) truss joint.
Fig. All.65 shows the moment distribution
for these moments. The magnitude of the moment (6) Carry out the moment distribution process
at B is 6.7 percent of that in Fig. All.63 and to find the secondary moments at the ends
10.4 percent at Joint C, however, it is reliev- of each member.
ing" in this example.
(7) Calculate the stresses due to these second-
ary moments and combine with the primary
axial stresses in the truss members due to
truss action with pinned truss joints.
All.16 Structures with Curved Members.
The moment distribution method can be ap-
plied to continuous structures which include
curved as well as straight members. The
equations for finding the stiffness factors,
carryover factors and fixed end moments for
straight members cannot be used for curved mem-
Fig. All. 65 bers. The elastic center method as presented
* Also see Chapter C2, Part H.
All. 29
in Chapter A9 provides a rapid and simple diagram. Thus to find the distance from Bx
method for determining these values for curved to this centroid we take moments of the M/EI
members. The use of the elastic center method diagram about B and divide by 0 the total area
in determining the value of stiffness and of the M/EI diagram. Thus
carryover factors will now be explained.
AIl.l7 Structures with Curved Members.
Before considering a curved member a Thus the centroid of the total moment weight 0
which equals one lies at point A or a distance
straight member of constant EI will be con-
sidered. Fig. All.66 shows a beam freely L to left of B.
In using the elastic center method to de-
termine the stiffness and carry-over factor for
Fig. All. 66 ~ a straight beam, we assume that the bending
_ /-_d,= I Rad.
moment curve due to a moment applied at A is of
MAJs " * ' ...rdeflectedbeam such magnitude as to turn the end A through an
~A EI constant' {~ MB angle of 1 radian. As shown above, the moment
L ~ weight 0 for this loading is unity or 1 and its
centroid location is at A. Then by the elastic
Fig. All. 67
=
center method we find the moment required to
__ 4EI~2LEI=MB turn end A back to zero rotation. The value of
MA - --r:- - this moment at A will then equal the stiffness
factor of the beam AB. In order to simplify
the equations for the redundant forces the
Fig. All. 68 M/EI Diagram elastic center method refers them to the elastic
center. From Chapter A9 the equations for the
4
-L redundant forces at the elastic center for a
structure symmetrical about one axis are: -
_ -2: 0s
Mo - 2: dS/EI - - - - - - - - - - - (1)
Yo _-2: 0s X (2)
supported at end A and fixed at end B. A - Iy
moment MA is applied at end A of such magni-
tUde as to turn end A through a unit angle of
one radian as illustrated in Fig. All.66. By 2: 0s Y - - - - - - - - - - - - - (3)
definition, the necessary moment MA to cause Ix
this unit rotation at A is referred to as the
stiffness of the beam AB. In Art. All.4 it Fig. All.69 shows beam of Fig. All.66 re-
was shown that this required moment was equal placed by a beam with the reaction at end A
in magnitude to -4EI/L. It was also proved replaced by a rigid bracket terminating at
that this moment at A produced a moment at point (0) the elastic center of the beam, which
the fixed end B of 2EI/L or a moment of one due to symmetry of the beam lies at the mid-
half the magnitude and of opposite sign to point of the beam. The elastic moment loading
that at A. Fig. All.67 shows the bending is 0s = 1 and its location is at A as shown in
moment diagram which causes one radian rota- Fig. All. 69.
tion of end A. Fig. All.68 shows the M/EI
diagram, which equals the moment diagram
divided by EI which has been assumed constant.
The total moment weight 0 as explained

in Chapter A9 equals the area of the M/EI
diagram. Thus 0 for the M/EI diagram in
Fig. All.68 equals,
Fig. All. 69
-i (~) + ~ (~) = -2 + 1 = -1 Solving for redundants at (0) by equations (1),
(2) and (3).
In other words the total elastic moment
weight 0 equals one or unity. = 2: ds/EI
- 0s = L/EI
-1
Mo =- EI / L
The location of this total moment weight
o will coincide with the centroid of the M/EI
L
- 1(-2) _ 6EI
Yo =- .z I k1s X
= 1 1 L3 - ~
is removed and end A is connected by a rigid
bracket terminating at the elastic center of
y
12 EI the structure. We will now find the required
forces at (0) to cancel the unit rotation at A
Xo = 0s y = 0 because y the vertical
.z which was assumed in Fig. All.71.
Ix '
distance of 0s load to x axis through elastic
center (0) is zero.
J....Mo
X- (~~XQ- - x
Fig. All.70 shows these forces acting at
f
the elastic center. rD.=1 rYo y = 0.6366r
EI/L
s " I B~
A Y
A~EI/L2 Fig. All. 73
Fig. All. 70 The total area of the M/EI curve for the
curve in Fig. All.72 if calculated would equal
one or unity as explained in detail for a
The bending moment at end A equals straight member. The centroid of this M/EI
diagram would if calculated fall at point A.
_ EI _ 6E! (.1::.-) = _ 4EI Thus in Fig. All.73 we apply a unit 0s load at
L L 2 L A and find the redundant force at O. Due to
symmetry of structure about a vertical or y
By definition the stiffness factor is the
moment at A which is required to turn end A axis the elastic center lies on this symmetrical
axis. The vertical distance from base line AB
through an angle of 1 radian. Thus 4EI/L
is the stiffness factor and this result
to elastic center equals y .6366r. (See =
page A3.4 of Chapter A3).
checks the value as previously derived in
Art. Al1.4. The elastic moments of inertia Ix and I y
can be calculated or taken from reference
The bending moment at B in Fig. All.70 sources such as the table on page A3.4.
equals,
Whence,
Ix = .2978r 3
t, but t -- ~
EI
Hence the carry-over factor from A to B
is .5 and the carry-over moment is of opposite .2978r 3
Hence Ix· EI
sign to that of the moment at A.
All. 18 Stiffness and Carry-Over Factors For Curved
Members.
Fig. All.71 shows a curved member, namely, Solving the equations for the redundants
a half circular arc of constant EI cross- at (0), remembering 0 s =
1 and located at point
section. The end B is fixed and the end A is A.
freely supported. A moment MA is applied at
A of such magnitude as to cause a rotation at
Mo - -
.z 0s =.=..i.U.- = - EI
A of 1 radian as illustrated in the Fig. .z ds!EI nr nr
Fig. All.72 shows the general shape of the EI
bending moment curve which is statically in-
determinate. In Fig. All.73 the support at A Y = .z0s x _ - (1)( -r) _ 2EI
o - Iy - nr 3 - nr 2
2EI
--- ....... ~deflected

Xo
=.z0Ixs y = (1)(-.6366r)
0.2978r 3
2.14 ~;
\ member
\
\ EI
r
Fig. All.74 shows these forces acting at
the elastic center. The bending moment at A
equals the stiffness factor for the curved beam
Fig. All. 71 Fig. All. 72 fixed at far end B and freely supported at near
All. 31
end A. The ratio of the bending moment at The method of how to handle this remaining
point B to that redundant force can best be explained by pre-
at point A will senting some example problem solutions.
give the carry-
over factor. -r~.
0
.636,6 r __ .. ' i2EI
2. 4 EI/r
2
All. 19 Example Problems. Continuous Structures
Involving Curved Members.
-'- "'- ~
Bending TJ'r B
Example Problem 1
moment at A, B
Fig. All. 74
Fig. All.76 shows I =1
M - -EI EI 2EI a frame consisting of
A - nr 2.14--
r2
(0.6366r) - (--2)r
nr both straight and
= -0.31SEI 1.36EI _ 0 636
r .
R
r
= 2 . 314- Rr curved members. Al-
r though simplified L=60"
C
relative to shape, I =2
Bending moment at B, this frame is somewhat
representative of a
M _ -0.31SEI 1.36EI + 0.636EI fuselage frame with two L=80" L=80"
"B - 1 . 04-2 Rr 1=1 1=1
r r r cross members, one be-
tween A and C to support
Therefore the stiffness factor for a half- installations above
circular arch of constant EI is 2.314- EI/r. cabin ceiling and the
other between F and D w = 50#/in.
The carry-over factor equals the ratio of to support the cabin
MB to MA or (-1.04-2 EI/r)/(-2.314- EI/r) = floor loads. The frame F f-'-........"""'---'''--''--''-'........."'''-"'--'''---j D
0.452. It should be noticed that the carry- supporting forces are
over moment has the same sign as the applied assumed provided by
moment at A as compared to the opposite sign the fuselage skin as 1= 2
for straight members. In other words, there shown by the arrows
are two points of inflection in the elastic on the side members. E
curve for the curved arch as compared to one Eccentricity of these
for the straight member. skin supporting forces Fig. All. 76
relative to neutral
FIXED END MOMENTS axis of frame member is neglected in this simpli-
fied example problem, since the main purpose of
The fixed end moments on a curved member this example problem is to illustrate the appli-
for any external loading can be determined cation of the moment distribution method to
quite rapidly by the elastic center method as solving continuous structures involving curved
illustrated in Chapter A9 and thus the ex- members.
planation will not be repeated here.
SOLUTION
The student should realize or understand
that when the end moments on a straight member Due to symmetry of structure and loading,
in a continuous structure are found from the no translation of the frame joints takes place
moment distribution process, the remaining due to frame sidesway.
end forces are statically determinate, whereas
for a curved member in a continuous structure, The frame cross members AC and FD prevent
knowing the end moments does not make the horizontal movement of joints A, C, F, and D
curved member statically determinate, since due to bending of the two arches. Any horizontal
we have six unknowns at the two supports as movement of these joints due to axial deforma-
illustrated in Fig. All.75 and only 3 equations tion is usually of minor importance relative to
of static equili- causing bending of frame members. Therefore it
brium. Even when can be assumed that the frame joints suffer
the end moments Fig. All. 75 rotation only and therefore the moment distri-
are determined bution method is directly applicable.
from the moment
distribution Calculation of stiffness (K) values for
process there each member of frame: -
still remains
one unknown, Upper curved member ABC: -
namely the hori-
zontal reaction _ _ EI
at one of the KABC - K CBA - 2.314 r' This value was
beCLll ends. derived in the previous Art. All.lS.
Substituting: - distributed moment.
2.314xl Calculation of Fixed End Moments.

KABC =
KCBA = 30 =
.0770 (E is con-
stant and therefore omitted since only relative The curved members of the frame have no
values are needed for the K values. ) applied external loadings hence the fixed end
moments on the curved members is zero.
The relative moment of inertia of the
cross-section of each frame member is given For member AC, fixed end moment equals
on Fig. All. 76. =
WL 2 /12 10 x 60 2 /12 = 3000 in. lb. and for
member FD = 50 x 60 2/12= 15000 in.lb.
2.314 x 2 - .1540
30 - Since the supporting skin forces on the
side members have been assumed acting along
For all straight members, the stiffness factor
equals 4EI/L. Hence, centerline of frame members, the fixed end
moments on members AF and CD are zero.
K
AC
= KCA = 4x 2/60 0.1333
Moment Distribution Process: -
~D = ~F = 4 x 6/60 = 0.4000 Fig. All.77 shows the calculations in

carrying out the successive cycles of the
K =~A = 4x 1/80 = 0.0500
AF moment distribution process. Due to symmetry
KCD = K
DC
= 4 x 1/80 = 0.0500
DISTRIBUTION FACTORS AT EACH JOINT: -
JOINT A.
ZK = 0,1333 + .0770 + .0500 = 0.2603
Let D equal symbol for distribution factor.

j--------l C
DABC = .0770/0.2603 = 0.296 CA
-3000 3000
D
AC = .1333/0.2603 = 0.512 1536 -1536
-768""-- - -
D = .0500/0.2603 = 0.192 -122
AF ---m:---
-82
JOINT F.
------;u---
~
ZK = .0500 + 0.400 + 0.1540 = 0.6040 12--
--=.Q.
D
FA
= ,0500/.6040 = .082 -2353
D = .4000/.6040 = ,663 .663 D

FD
DFED = .1540/.6040 = .255
FD DF
-15000 15000
9750/-9750
Carry-Over Factors: - -4875
2122
For the straight members the carry-over -106UE
469
factor is 0.5 and the moment sign is the same -234/ Fig. All.77
as the distributed moment when the sign con- 105
vention adopted in this chapter is used. -=52/
~
For a half circular arch of constant I, -8753
the carry-over factor was derived in the pre-
vious article and was found to be 0.452. The
sign of this carry-over moment was t he same of frame and loading, the process need only be
sign as the distributed moment at the other carried out for one-half of frame, thus in Fig-
end of the beam. However, using the sign ure All.7?, only joints A and F are considered
convention as adopted for the moment distri- since the numerical reSUlts at D and C would be
bution in this book, the carry-over factor the same as at A and F respectively. In the
would be minus or of opposite siGD to the figure the distribution factors are shown in
All. 33
the c=J at each joint. The process is started inertia about x and y axes through the elastic
by placing the fixed end moments with due re- center.
gard to sign at the ends of members AC and FD,
namely, -3000 at AC, 3000 at CA, -15000 at FD
and 15000 at DF. We now unlock joint A and
find an unbalanced moment of -3000 which means ~
d=4.6~
a plus 3000 is needed for static balance. • - ------.; I
Joint A is therefore balanced by distributing _ I E.C. \ I 19.1"
.512 x 3000 = 1536 to AC, .296 x 3000 = 880 to Y
-----.:L- ':- ---, I = 2 ----! C ~
L_-.,
ABC, and .192 x 3000 = 576 to AF. Short hori- A r30"------l Fig. All. 78
zontal lines are then drawn under each of these
distributed values to indicate that these are
balancing moments. Carry-over moments are
immediately taken care of by carrying over to
joint F, .5 x 576 = 288. From A to C the
- = 19.1
Y
x n (30/1)
n x 30 60
= 14.49"
carry-over moment would be .5 x 1536 = 768 and 1 + 2
therefore the carry-over from C to A would be
.5(-1536)= -768 which is recorded at A as (NOTE:- 19.1 equals distance from line AC to
shown. For the arch member ABC, the carry-over centroid of arch member ABC.)
moment from A to C would be -0.452 x 880 =
-401 (not shown) and therefore from C to A = Calculation of moment of inertia Ix: -
-0.452 x (-880) = 401 as shown at joint A in
the figure for arch member CBA. Joint C in Member ABC
the figure has been balanced once for the _ .3r" nrd 2
purpose of helping the student understand the Ix - -1- + - 1 -
sign of the carry-over moments which flow to
the left side from the right side of the frame. = .3 x130" x-
+ n
-l 30x 4.612 = 10100
After balancing joint A and taking care 60
of the carry-over moments, we imagine A as Member AC =2"x 14.49 2 = 6310
fixed again and proceed to joint F where we
find an unbalanced moment of -15000 + 288 = x
total I = 16410
-14712, thus plus 14712 is necessary for Total elastic weight of structure equals
balancing. The balancing distribution is .255
x 14712 = 3750 to FED, .663 x 14712 = 9750 to Z dS/I = (n x 30/1) + 60/2 = 124.24.
FD and .082 x 14712 = 1212 to FA. The carry-
The next step in the solution is to draw
over moments are .5 x 1212 = 606 to A, .452 x the bending moment curves on this frame portion
3750 = 1695 from D to F by way of the arch due to the given load on member AC and the end
member and -.5 x 9750 = -4875 from D on member
FD. We now go back to joint A which has been moments as found by the moment distribution
unbalanced by the carry-over moments and repeat process in Fig. All.77. It is composed of
the balancing and carrying-over cycle. In the three parts labeled (1) to f3) in Fig. All.79.
complete solution as given in Figure All.77 Portions (1) and (2) are due to the end moments,
each joint A and F was balanced five times. and portion (3) due to the distributed lateral
The final bending moments at the ends of the load on member AC.
members at each joint are shown below the
double short lines.
The arch member ABC has 3 unknown forces
at each end A and C or a total of 6 unknowns.
With 3 equations of static equilibrium avail-
able plus the known values of the end moments
at A and C as found from the moment distribu-.
tion process, the arch member is still static-
ally indeterminate to one degree. Thus the Fig. All. 79
horizontal reaction at A or C as provided by
the axial load in member AC must be found be-
fore the bending moments on arch ABC can be The next step is to find the 08 (area of
calculated. M/I curve) for each portion and its centroid
location.
The first step in this problem is to find
the elastic center of the frame portion com-
posed of members ABC and AC, as shown in Fig.
All.78, and then find the elastic moments of
0s~ =1094 x n x 30/1 = 103000 46 lb. The member AC also suffers an axial load
due to the shear reactions at the top of members
0 S2 = 60 x 2353/2 = 70590 FA and DC. Fig. All.81 shows free bodies of the
side member FA and DC with the end moments as
0s " = (-.667 x 60 x 4500)/2 .. -90000 found in Fig.All.77.
The shear reaction 3...:.8..:.....1"---- 38:2-
Z0s = 83590 at A and C can be 1259 A C 1259
found by taking (,iA-R
Fig. All.80 shows the 0 s values concen- R- "C
moments about lower A C
trated at their centroid locations and referred ends. Thus for FD, 80"
to the x and yaxes through the elastic center. RA = (1259 + 1791)/80
.. 38.1 lb. Likewise F D
y
RC =
38.1 lb. These ~
1791
~.A
1791
reactions react on
B Fig. All. 81
cross member AC in the
• OSl. = 103000 opposite directions
-----.~-~~oX-O-\---Tt- X thus giving a compression load of 38.1 lb. in
<a = 70590 14.49 member AC, which must be added to the tension
/SL c~ load of 46 lb. from the arch reaction to obtain
A
"n Fig. All. 80 the final load in the cross-member.
~,,= -90000
Bending Moment in Lower Arch Member FED

The frame has been imagined cut at A and the
arch end at A has been connected by a rigid The horizontal reaction on arch FED must
bracket to the elastic center. The redundants be found before the true bending moments on the
at the elastic center which will cancel any arch can be found. The procedure is the same
relative movement of the cut faces at A can as for the upper arch. Figs. All.81, 82, 83
now be computed. show the results for the lower frame portion.
_ - Z 0s -83590_
Mo - Z ds!I = 124.24 - - 673 in. lb.
F Iy D F
~+-+-+-~w,..--'-r'-T"--Ht--,
Xo = Z 0 s Y = [103000 (4.61) + 70590

Ix
(-14.49) - 90000 (-14.49)J /
r ~1JE:C:)~
...Lx
--1-
0__
1
3••34" 19 • 1"
f t
16410 = 46 lb. E
Iy Fig.A1l.81 Fig. All.82
Yo = -Z I 0s x = 0 because x = zero.
y y = 19.1 x n(30/2)
n x -30 + -60
= 900
57.1
= 15.76" Z dIs = 57.1
The bending moment at any point on ABC 2 6
or AC equals that due to Mo and Xo plus the
moments in Fig. All.79. Ix = .3x30" + 15nx3.34 2 = 4578
For example,
2 (60/6)(15.76)2 = 2478
Ix = 7056
At point A on member ABC,
From Fig. All.82,
MA = 1094 - 670 + 46 x 14.49 = 1091 (should =-
0s~ = -8753 x 60/6 87530
be 1094 since moment as found in Fig. All.77
is correct one. Small error due to slide rule 0S2 = .667 x 22500 x 60/6 = 150000
accuracy.
08 " =- 6962n x (30/2) = -328000
At point B on member ABC: -
Z 0s = -265500
MB = 1094-670-46x15.51 -289 in. lb.
Solving for Redundant forces Mo and Xo (See
The horizontal reaction at A will not Fig. Al1.83)
produce any bending on member AC, thus the
values in Fig. All.79 are the true moments. M - - Z 0 8 = - (-265500)
o - Z ds!I 57.1
= 4650 in. lb.
The axial load in member AC by statics

from Fig. All.80 equals Xo or a tension load of
All. 35
x _Z¢SY=-87530X15.76+150000x15.76-328000(-3.34) Example Problem 2. External Loads on Curved

o IX 7056 Members.
Xo =294 lb. Fig. All.85 shows a frame which has ex-
ternal loads applied to the curved member as
well as to the straight members. The frame
-r- supporting forces have been assumed as acting
15.76
-x+ uniformly along the side members AD and CEo
The problem will be to determine the bending
14.24 moments at frame points ABCDE.
-----.L
Fig. All. 83
1000 B 1000
Bending moment at F on member FED, 1 ~ 1. 'j
MF = -6962+4650-294x15.76 = -6962 in.lb.
ME = -6962+4650+ 294x 15.76 = 2340 in.lb.
A C
The axial load in member FD 294 lb. = w = 10 lb/in.
compression plus 38.1 Ibs. tension due to shear
reaction from side members at points F and D 1=1 1=1
or a resultant load of 255.9 lb. compression. Fig. All. 85 L-60 L=60
Fig. All.84 shows the final bending moment
diagram on each member of the frame. 1000 lb.
30'
B
E
1= 3, L = 60
Member ABC
SOLUTION: -
Calculation of stiffness factors K.

1094 A
1094
2147 Member ABC. KABC = 2.314 I/r = 2.314x 1.5/30 =
.1158
Member AC KAC = 4I/L = 4 x 2/60 = .1333
KAD = 4I/L = 4x 1/60 = .0667
2353 KDE = 4I/L = 4x 3/60 .2000 =
C 1259
Calculation of Distribution Factors D.
JOINT A. ZK = .1158 + • 1333 + .0667 = .3158

Member CD
DABC = .1158/.3158 .366 =
DAC = .1333/.3158 = .422
DAD = .0667/.3158 = .212
F 13747
1791
JOINT D. ZK = .0667 + .2000 = .2667
Member FD DDA = .0667/.2667 = .25
DDE = 2000/.2667 = .75
The carry-over factor for ABC is -0.452 as pre-
viously derived for a half circle are, and 0.5
6962
for the straight members.
962
Calculation of Fixed End Moments
For member AC, MA = MB =~

12
wL " = .-l. x 10 x 60 =
12
2
Member FED
E
3000
Figs. All. 84. Final Bending Moment Curves on Each

For DE, MB = ME = PL/8 = 1000 x 60/8 = 7500
Member of Frame.
Curved member ABC. total elastic weight of member and its elastic
moments of inertia about axes through the
The fixed end moments on this curved elastic center.
member due to the external loads will be de-
termined by the elastic center method. The TOTAL Eh~STIC WEIGHT = 2: ds/I -- 1.5
nr _ n x 30
- 1.5
assumed static frame condition will be an arch
pinned at A and supported on rollers at B. = 62.83
(See Fig. All.S6)
Distance y from line AC to elastic center of
Fig. All.S7 arch ABC eQuals 0.6366r = .6366 x 30 = 19.1 in.
shows the general 1000 =P
1= 1.
B 1000 =P Ix =0.2978r 3 /1 =
2978 x 30 3 /1.5 5350 =
shape of the static
moment.curve. For -JlD=120~
the frame portion "
between the re- <><= 300~ /'
/'
lD sa " 167300
A "f" - ~-
actions and the load
points, the bending Fig. All. 86
moment eQuation is
1'1=P(r-r cos a).
(2)
Y=24.8"
\IX
19.1
~
(I' )
For the beam
portion between the (1)
C
two 1000 lb. loads, Fig. All. 87
the bending moment Static Moment Fig. All. 88
is constant and Curve.
eQuals, Calculation of redundant forces at elastic cen-
1'1=P(r -r cos 30 0 ) ter.
Calculation of ¢s Values. 2: ¢, = - (28800 + 167300) = -3122

2: ds I 62.83 in. lb.
The ¢s values eQual the area of the 1'1/1
curves. The moment curve in Fig. All.S7 has X = 2: ¢sY
o
= 167300 x 5.7 + 28800 (-9.1) =
been broken down into three parts labeled (1) Ix 5350 129.3 lb.
(1') and (2).
¢sJ. + ¢S2 = 2~ 2(a - sin aJ Yo = - 2: ¢sx

Iy
=0 (because x = 0)
= 2~0~~;302(.524_ .58 = 28800 The fixed end moments at ends A and C will
eQual the moment due to the redundant forces 1'1 0
The vertical distance y from line AC to and Yo since the static moment assumed was zero
at A and C.
centroid of ¢sJ. and ¢S2 values is,
y = r(l-cos a-~) = 30(1- .867-0.5/2) = 1'1A =I'1c =1'10 + XoY = - 3122 + 129.3 (- 19.1)
= - 652 in. lb.
a-sina .524-.50
10 in.
1'1oment Distribution Process
2
¢8 2 = Pr e (1 - cos a). e = 1200 , a - 30 0 •
I Having determined the fixed end moments,
distribution and carry-over factors, the moment
¢ = 1000 x 30 2 X 2.1 (1 _ .867) = 167300 distribution can now be carried out. Fig.
8 2 1.5 All.89 shows the solution. The first cycle will
be explained. Starting at joint (A) the un-
Vertical distance y from line AC to centroid of balanced moment is -3000 - 652 = - 3652. The
¢S2 eQuals, joint is balanced by distributing .432 x 3652
= 1543 to AC; .364 x 1543 =
1333 to ABC and
- 2r sin e/2 .:. : .2.:. : .x: :. .3: :. .0=--=x--.:0,•.-" :. : .8-,-6-,--7 = 24.8 in. .212 Y 3652 = 776 to AD. The carry-over moment
y = e = 2.1 =
from C to A .5(-1543) =
-772; from C on
member CBA to A =-
.452(-1332) = 601; and from
Fig. All.88 shows the ¢s values and their lo- =
A to D .5 x 776 =
388. Now proceeding to
cation with respect to x and y axes through joint D the unbalanced moment is -7500 + 388 =
the arch elastic center. -7112. The joint is balanced by distributing
.75 x 7112 =
5340 to DE and the remainder 25
The elastic center method reQuires the per cent =
1772 to DA.
All. 37
Static Mom.
Curve
(1) (1)
ON
""'
'"
Co
AC 3000
-3000 F. E. M. 2513
2513
~~~i/
-302/
151
-1~~/
-64 / ' Fig. All. 90 -1/8 wL~ =4500
32/
-~~/ 0s = 2513 X 60/2 = 75390
.25 -=..9.- e
-2513
D .75 I - - - - - - - - - J E
L:0 s = 218290
DE 7500
Fig. All. 89 -7500 Calculation of elastic center location and
5340/
-2670 Ix·
2060/
-1030
802/
-401
Distance y from line AC to elastic center =
~/
-156
19.1xnx30/1.5 = 1202 = 12 97"
nx30 + 60 92.83 •
-lli/
-60 1.5 2
46
-3135
TOTAL Eh~STIC WEIGHT = 92.83.
Ix = .2978x30
1.5
3
+ nx30 (6.13)" +..§Qx 12.97"
1.5 2
The carry-over moments are: -
= 12713
From E to D = .5 x (-5340) = -2670
Fig. All.91 shows the elastic center loca-
To A from D = .5 (1772) =886 tion and the 0 s values together with their
centroid locations.
The first cycle is now completed. Five more
cycles are carried out in Fig. All.89 in order
to obtain reasonable accuracy of results. The
final end moments are listed below the double
short lines. [B ~S2 = 167300
On arch member ABC the end moments of 584 0;0 S 3 =30
are correct. However before the end moments 3---,- i
at any other point on the arch can be found 6.13 M
the horizontal reaction on the arch at A must _~_-'-:-::~---'L.-~ Xo x
be determined. This reaction will be de- =-T---Af~ - - do. = 28800
termined by the elastic center method. I ,oJSi .-L
I
A ~ C
Fig. All.90 shows the bending moment ~s = -90000 + 75390 = - 14610
curves for members ABC and AC as made up of Fig. All. 91
5 parts labeled 1 to 5.
Solving for redundant forces at elastic center.
Calculation of 0 s values which equal area
of each moment curve divided by I of member. No = L: L: ds/I
0s = -218290
92.83
= -2356
20s 1. = 28800
Xo=~= 167300 x 11.83 + 36800 x 6.13 +
03 " = 167300 Ix 12713
28800 (-2.97) - 14610 (-12.97)= 181 lb.
0 S3 = 584 x n x 30/1.5 = 36800 12713
0s 4 =- (.667 x 4500 x 60/2) = -90000
The bending moment at any point on member (2) Find the bending moments at all joints
AC or ABC equals the values in Fig. All.90 and support points of the loaded structures
plus those due to Mo and Xo in Fig. All.91. in Fig~. 7 to 10.
Thus at point A on member ABC, W= 30lb/in.
~
MABC = 584-2356+181x12.97 = 584 in. lb. L- rf ---'
T II
which is the same as found by the moment distri- 20" 400 I =6 ~400
bution process. ! =4 1=4 10"
--l-
At point B on arch: - Fig. 7 Fig. 8
MB from Fig. All.90 = 584 + 4000 = 4584

100
MB from Fig. Al1.91 = -2356 -181 x (30 - 12.97) Pin 0-4" 100
= -5476 1- 5" 1=3 A 1= 1 r- 8"--->j~
Hence MB = 4584 - 5476 = -892 in.lb. 1--16"--1
=2 =3 1=2
STRUCTURES WITH JOINT DISPLACEMENTS Pin
Fig. 10
For unsymmetrical frames or for frames
r- 24"-+- 20"-1 5" ~
Fig. 9
loaded unsymmetrically, the assumption that
only joint rotation occurs may give resulting
moments considerably in error. Joint dis- (3)
placement can be handled in a manner as --;oo!12"k-
previously explained and illustrated for frames I I. 12 lb/in.
composed of straight members. i'l"-.';t-rB't--rt -1,-----,t---r2--.c'· *,. . D. .,. ·t-r--=;·=--=rl"'-T't::.:t,.......,..ta:-,-~,-,-l-,!
The distribution of the skin supporting
forces on the frame boundary are usually taken
~20'~40" -+-32" -r- 40"-+20'~
as following the shear flow distribution for Fig. 11
the shell in bending as explained in a later
chapter. In the loaded beam of Fig. 11, the supports
Band E of the elevator beam deflect 0.1 inch
All.20 Problems.* more than supports C and D. Compute the re-
SUltant bending moments at the supports and find
(1) Determine the bending moment diagram for reactions. E1 =
320,000 lb.in.sq.
the loaded structures shown in Figs. 1
to 6. (4) Fig. 12 illustrates a continuous 3 span
1 1
t
100
f--8"+-16,,=t
t;t=
1100
.. Ik'
~24"---l
w=10lb/in.
k
wing beam, carrying a uniform air lo~d of 20
lb./in. Determine the beam bending moments at
strut points A and B. Take
Fig. 1 Fig. 2
IAB=17in., 1M' = 20 in., and E = 1.3 x lOS in. 2
Ifo . 200
b
W=r'f /
m
. r"~ Neglect effect of support deflection due to
~=10 =i=!i6 --f strut axial deformation.
Fig. 4
ti°O 1-12j
100 100
t12-1
;eo ~
,ft=l/A1
r-- ------+--
14"
1=2
36"
Fig. 5
--+-- 1=2
16" ---l
100 200 w=12lb/in. (5) Figs. 13 to 15 are loaded structures that
•
I ~ ~IILLi suffer joint translation. Determine bending
~ I- 1 ~ 1= 1.5 P, 1= 2 I - 2 moment diagram.
1-6" +-12''---+-12''-----+1-16''-+ 12"---1 100
500 ~5 10'I.j
Fig. 6 l R = l T.. 1= - 0
.... 12 ~ ... 200
24" 1=2 II -.l.... 1=4
1 .... 6" ..!.I ~II
* Also see Chapter C2, Parts F, L, M and Q.
-i.
f.-20'Y
T
Fig. 14
'--20"
I·
---k--
·1·
24"
I I f-
Fig. 13 Fig. 15
CHAPTER Al2
SPECIAL METHODS - SLOPE DEFLECTION METHOD
A12.1 General. The slope-deflection method is Fig. A12.1a

another widely used special method for analyz- Ir<-<- - - L -I
ing all types of statically indeterminate
beams and rigid frames. In this method as in t~ f-I----=:-=:----,--,----------II~
EI Constant I
the moment distribution method all joints are d4 M4
considered rigid from the standpoint that all LC _ - J - - _ I ds
angles between members meeting at a joint are
assumed not to change in magnitude as loads Fig. b " - - . ~
-~
-~- ~~- - -.:=.--
Oft ~il
are applied to the structure. In the slope

deflection method the rotations of the joints
I) M s
are treated as the unknowns. For a member
bounded by two end joints, the end moments can
be expressed in terms of the end rotations. - - --=----1) ~
Furthermore for static equilibrium the sum of Fig. c
the end moments on the members meeting at a
joint must be equal to zero. These equations
of static equilibrium provide the necessary
conditions to handle the unknown joint rota- Fig. d M'i (I----==--
tions and when these unknown joint rotations
are found, the end moments can be computed
from the slope-deflection equations. The ad-
vantages of the slope deflections method will Fig. e
be discussed at the end of the chapter after
the method has been explained and applied to
problem solution.
Figs. A12.1
A12. 2 Derivation of Slope Deflection Equation.
The problem is to determine the relation- distance ~ which causes end moments M'~ and M":, •
ships between the displacements of the end In deriving the slope deflection equations each
supports of a beam and the resulting end of the three beam deflections will be considered
moments on the beam. separately and the results are then added to
give the final equations. Fig. A12.2 shows Fig.
Fig. A12.1a shows a beam restrained at c repeated.
ends 1 and 2. It is assumed unloaded and of
constant cross-section or moment of inertia.
This beam is now displaced as shown in Fig. b, Fig.A12.2 L
namely, that the ends are rotated through the Os= 0
angles 8 4 and 8 s plus a vertical displacement =..---;.----i2
d 4 and d s of its ends from its original posi- t
tion, which produces a relative deflection ~ Fig. A12.3 Mi
of its two ends. The angle 0 representing the
swing of the member equals ~/L.
The problem is to derive equations for

M4 and Ms in terms of the end slopes 8 4 and Fig.A12.4
8 s , the length of the beam and EI. Figs. c,
d and e illustrate how the total beam deflectio -=::::::::::==1
---.J
M's -- -2-
- Mi
in Fig. b is broken down into three separate
deflections. In Fig. c, the end (2) is con-
sidered fixed and a moment Mi is applied to
rotate end (1) through an angle 8 4 , In Fig.d, The applied moment Mi, is positive (tension in
end (1) is considered fixed and a moment M~ is bottom fibers). Since end (2) is assumed fixed
applied at (2) to rotate end (2) through an an unknown moment M's is produced at (2). Fig.
angle 8 s • In Fig. e, both ends are considered A12.3 shows the bending moment diagram made up
fixed and end (2) is displaced downward a of two triangles. M~ being unknown will be
A12.1
A12.2 SPECIAL METHODS - SLOPE DEFLECTION METHOD
assumed also positive, as the algebraic solu- Fig. A12.6 shows the moment diagram for an
tion will determine the true sign of M~. applied bending moment -M~ which rotates end (2)
through an angle 8 2 when other end (1) is fixed
Two moment area theorems will be used in (See Fig. A12.5). In a similar manner as
deriving the slope-deflection equations, described before,
namely (I) that the change in slope of the
beam elastic curve between two points on the e = (-M~L/2)
EI +
(-M~L/2) =- (Wi + M~)L
EI 2EI - (5)
beam is equal in magnitude to the area of the 2
M/EI diagram between the two points, and (II)

the deflection of a point (A) on the elastic Taking moments of M/EI diagram about (2),
curve away from a tangent to the elastic curve
at (B) is equal in magnitude to the first
moment about A of the M/EI diagram between (A)
(M':r.L)!:. + (M'iL) 2L
\2EI 3 2EI 3
=0
and (B) acting as a load.
or, - M~ = 2M1 - - - - - - - - - - (6)
In Fig. A12.2 since 8 2 0, then 8 1 will =
equal the area of the M curves divided by EI. whence,
MilL _ 2EI8 2
8
1
= (MiL/2) + (MkL/2) =
EI EI
- - (1) 82 =2EI and M~ - L - - - - - - - (7)
The deflection of end (1) away from a Then from equation (6)
tangent at (2) equals zero, thus we take
moments of the moment diagram about (1) and Mil _ - 4Ele g - - (8)
equate to zero. 112 - L
Fig. A12.8 shows the third part of the

beam deflection, namely, support (2) is deflected
downward a distance ~ assuming both ends fixed.
whence -Mk =Mi/2 - - - - (2)
Fig. A12.4 shows the true shape of moment

M1\.
r . L 'I
diagram. ,n(~,=o
------ ---r-
Substituting results in equation (2) in Fig. A12.8 =0 ~ Q2
equation (1) we obtain, -------=-.....;f---j-lj M'~'

M~n~
~M'J,'
M"'..h M'n.k
Fig. A12.9 1 2 2 2
Then from equation (2)
Mk =- 2Ele 1
L
_ - - - - - - - - - - __ (4)
Fig. A12.10
The change in slope of the elastic curve
between the ends equal zero, thus by the moment
L
Fig.A12.5 area theorem the area of M/EI diagram equals
zero.
(M'!' + Mm2 )L
Hence, -"-=c=--=::c-"----L:.=-
2EI
=0
whence MT =-M':- - - - - - - - (9)
Fig.A12.6
From the deflection theorem,
M"2 M'~ L (.§.L) + M'~ L (..!:-) = _~ (10)

2EI 3 2EI 3
SUbt. (9) in (10)
Fig.A12.7
Mm _ - 6EI~.
11 1 - --y;-;a-- (ll)
A12.3
M~'
=_+--=.6E=.I::..:6.::..:
L 2 ------------ (12) (18a)
M2-1 = 2K (28 2 + 81 ) + MF 2 - - - - (18b)

The combined effect of deflection 01 , O2
and 6. can now be obtained by adding the separ- A12.3 Hinged End. Slope Deflection Equation.
ate results.
Consider that end (2) of beam 1-2 is free-
ly supported or in other words hinged. This
M _
"1-2 -
M'
111 + Mil
"1 + Mill
111 == L2EI (28 1 + 8 2 - 36./L) means that M2 -:L is zero. Thus equation (18)
EI can be equated to zero.
Let K = Land 0 = 6./L
M2 - 1 = 2K (28 2 + 81. -30) +MF 2 =0
Then, M1 - 2 == 2K (28 1 + 8 2 - 30) - - - - (13)
whence
and M2 -:L = -2K (28 2 + 8 1 -30) (14)
2K8" = -K81 + 3K0 - •5MF 2
These end moments due to the distortion
of the supports must be added to the moments Substituting this value in equation (17)
due to any applied loads on the beam when
considered as fixed ended. Let these fixed M1 - 2 = 3K(8 1 -0) +MF 1.- 2 -O.5MF 2-.1. - - -(19)
end moments be labeled MF 1 and MF. Then A12.4 Example Problems.
equations (13) and (14) can be written includ-
ing these moments.
Problem 1
M1 - 2 = 2K (28 1 +8 2 -30) - MF 1 (15) Fig. A12.11 shows a two span continuous
M2 - - -2K (28 2 + 8 1 - 30) - MF 2 (16) beam fixed at ends (1) and (3) and carrying
1
lateral loads as shown. The bending moments at
MODIFIED SIGN CONVENTION points 1, 2 and 3 will be determined. Relative
values of I are shown for each member.
Equations (15) and (16) have been 100 lb.
developed using the conventional si~ns for
bending moments. In general there ~re ad-
vantages of using a statical sign convention
as was used in the moment distribution method
in Chapter All. Therefore, the following
sign convention will be used in this chapter: - Fig. A12.11
(1) The rotation of a joint or member is posi- SOLUTION:

tive if it turns in a clockwise direction.
Calculation of fixed end moments due to
(2) An end moment is considered positive if it applied loads -
tends to rotate the end of the member
clockwise or the joint counterclockwise. MF 1 _ 2
= -PL/8 = 100x24/8 = -300 in. lb.
This adopted sign convention is the same MF 2 _ 1

=PL/8 =300
as adopted for the moment distribution method
(See Art. All.2). The signs as shown are determined from the
fact that the end moment at (1) tends to rotate
When equations (15) and (16) are revised the end of the member counterclockwise which is
for this new sign convention they become: - a positive moment according to our adopted sign
convention. By similar reasoning the fixed in
M1 - == 2K (28 1 + 8 2 -30) +MF 1 (17) moment at (2) is positive because the moment
2
tends to rotate end of member clockwise. For
(18) span 2-3: -
M2 -:L == 2K (28 2 + 8 1 - 30) + MF 2
Equations (17) and (18) are referred to as the MF 2 _ 3
= wV/12 = 24x24 2
/12 = -1152 in.lb.
slope-deflection equations where K == EI/L and
MF 3 _ 1152
o= 6./L. 2
==
For no yielding or transverse movement of K Values K == EI/L. Since E is constant it

supports, 6. = 0 and equations (17) and (18) will be considered unity.
become FOR BEAM 1-2, K I/L 24/24 = = =1
FOR BEAM 2-3, K = 48/24 2 =
Substituting in slope-deflection equations lSa Example Problem 2

and 19b. (¢ =
0 since supports do not trans-
100# 100
late. )
8"~ w = 12 Ib/in.
Beam 1-2. 81 =0 because of fixity at (1). (I)
L = 20 (4)
I = 20
11 1-" = 2K (28 1 + 8,,) + 11F 1 K=I
111-" = 2xl (0+ 8,,) -300 Fig. A12.12
111-" = 28" - 300 - - - - - - - - - - (a) Fig. A12.12 shows a loaded 3 span beam
fixed at points (1) and (4). This fixity at
11"-1 = 2K (28"+8 1 )+11F,, (1) and (4) causes 8 1 and 8~ to be zero.
11"-1 = 2xI (28" + 0) + 300 Fixed end moments:
11"-1 = 48" + 300 - - - - - - - - - - - (b) I"lF 1_" =- PL/S = 100 x 24/S = - 300,
Beam 2-3. 8a = 0, because of fixity at (3). 11F"_1 =300

11F,,_a = (100xSx16"/24 ) + (100x16xS"/24")
11,,_a = 2x2 (28,,+0)-1152 =-534

- - - - - - (c) I"lF a_" =534
11Fa_~ =wL"/12 = 12 x 20"/12 =-400,
11a_" = 2K (28 a + 8,,) + 11F a
11 a-" = 2x2 (0+8,,)+ 1152

11F~_a = 400
11 a-" = 48" + 1152 - - - - - - - - - - - (d)
SLOPE DEFLECTION EQUATIONS (¢ = 0)
At joint (2) m = 0 for static equilibrium. 111-" = 2xl(0+8,,) -300 = 28" -300 (a)
(b)
Hence
11,,-a = 2x2(28,,+8a )-534 = S8,,+48a
- 534 - - - - - - - - - - - - - - (c)
Substituting from (b) and (c)
11a_" = 2x2(28a +8,,) +534 S8" +48" =
48" + 300 + S8" - 1152 = 0
+ 534 - - - - - - - - - - - - - - (d)
whence 8" = 71 11"-4 = 2 x 1 (28" + 0) - 400 = 48" - 400- (e)
With 8" known the final end moments can be
found by substituting in equations (a), (b), 11~_" = 2 x 1 (0 + 8" ) + 400 = 28" + 400 (f)
(c) and (d) as follows: -
Joint-11oment equilibrium equations: -
111_" = 2 x 71- 300 =- 15S in.lb.
JOINT (2)
11"-1 = 4x 71 + 300 = 5S4 whence
11,,_a = S x 71 - 1152 =- 5S4 48" + 300 + S8" + 48" - 534 = 0 or
11a_" =4 x 71 + 1152 = 1436 128" + 48" - 234 =0 - - - - - (g)
Changing the resulting moment signs to the
conventional sign practice, gives
SO" + 48" + 534 + 48" - 400 =0 or
128" + 48" + 134 =0 - - - - - - - - (h )

A12.5
Solving equations (g) and (h) for 8 2 and 8 3 or free to rotate we use equation (19) in writ-
gives, ing equations for 11,,-1. and 11 3- 4 •
82 = 26.15, 83 =- 19.9
SUbstituting in equations (a) to (f) inclusive MS - 1 = 3xlx8 s +250-0.5 (-250) =
to find final end moments.
38" + 375 - - - - - - - - - - - (a)
MJ.-s = 28 2 - 300 = 2 x 26.15 - 300 =
-247.7 in.lb.
11 3- 4 = 3xlx83 -500-0.5 (500) = 38 3
11 2-1. =4 x 26.15 + 300 = 404.6
= 482 + 300 - 750 - - - - - - - (b)
11"_3 = S8" +463 -534 = Sx26.15+4 (-19.9) Using equations lSa and lSb and substituting,
- 534 =- 404.6
11"-3 = 2K(28" + 83 ) + 11F
11 3- 2 = S63 + 46" + 534 = S (-19.9) + 4 x 26.15 " 3
+ 534 = 479.4 11"-3 = 2xO.5(28 2 +8 3 )+0 = 28,,+ 83 - - (c)
11 3- = 48 -400 = 4 (-19.9) -400 = 11 3-" = 2K(26 3 + 6 2 ) + 11F 3-"

4
- 479.4 11 3-" = 2xO.5(26 3 +6,,) +0 = 26 3 + 6,,- - (d)
11 4 - 3 =283 + 400 = 2 (-19.9) + 400 = 360.2 For statical equilibrium of joints: -
(Note: - Student should convert to conventional JOINT (2) 11,,-1. + 11"-3 =0

moment signs and draw complete bending moment
diagram). Subt. 38" + 325 + 28" + 83 =0
Example Problem 3. Beam with Simple Supports. or - - - - - - - - (e)
Fig. A12.13 shows a loaded 3 span contin- JOINT (3)

uous beam with hinged supports at points (1)
and (4), which means that 111.-" and 11 4 - 3 O. = 28 3 + 8 2 + 38 3 - 750 =0
The moments at supports (2) and (3) will be
determined. or 58 3 + 8" - 750 =0 - - - - (f)
100 lb.
Solving (e) and (f) for 8" and 8 3 gives,
r- 10tl
(1) ~----=---'::=---7ic----;-L-=-----=-:20:;---'7i---;----'----;:;::-----7i\
6" =- 109.4, 83 = 172
I = 10 The end moments at (2) and (3) can now be found
K = 0.5 from equation (a) to (d) inclusive.
Fig. A12.13
11,,-1. =36 2 +375 = 3 (-109.4) +375 = 46.S
in. lb.
SOLUTION: 11"-3 =262 + 63 = 2 (-109.4) 'I- 172 = -46.S
Fixed end moments: 11 3-" = 26 3 +8" = 2x172+ (-109.4) = 234

11F1._" = - PL/S = 100 x 20/S = - 250, 11 3_ 4 = 38 3 - 750 = 3 x 172 - 750 - - 234
11F,,_1. = 250 A12.5 Loaded Continuous Beam with
Yielding Supports.
11F"_3 -- 11F3-" =0 (No load on span 2-3)
The movable control surfaces of an airplane,
11F 3_ 4 = PL/S = 200 x 20/S =- 500, namely the elevator, rudder and aileron are
attached at several points to the stabilizer,
114 - 3 = 500 fin and wing respectively. These supporting
surfaces are usually cantilever structures and
Slope-deflection equations: - thus the supporting points for the movable con-
trol surfaces suffer a displacement thus pro-
Since supports at (1) and (4) are hinged moting a continuous beam on yielding or deflected
supports. The slope deflection equations can 40"

take care of this support displacement as I1F 4 - 3 = -60 (5 x3.5 + 1.5) = 507
illustrated in the following example problem.
The beam has a constant section, hence,
BJcample Problem. K for all spans equals,
K = EI/L = (10,000,000xO.03339) 40 = 8347

Calculation of 0 values.
Span 2-3. The settlement of support (2)
Fig. A12.14 with respect to support (3) = (0.5 - .1875) =
.3125 inches ~ =
o = - ML = - .3125/40 = - 0.007812 rad.
deflected supports.
Fig. A12.15
(Since (2) moves counterclockwise with respect
to (3) the sign of 0 is negative.
Fig. A12.14 is representative of an ele-
vator beam attached to deflecting stabilizer Span 3-4. ~ = .1875 inches
structure at the five reaction points as
indicated in the figure. The elevator beam o =- ML = - .1875/40 =- .004688 rad.
is a round aluminum tube 1-1/4 - .049 in size.
The moment of inertia (I) of this tube equals Since the beam, external loading and support
0.03339 and the modulus of elasticity (E) of settlement is symmetrical about support (4), the
the material equals 10 million psi. The air slope of the beam elastic curve at (4) is hori-
load on the elevator beam is variable as zontal or zero and therefore 6 4 O. Thus only =
indicated in Fig. A12.14. one-half of the structure need be considered in
the solution. Due to the fact that (2) is a
Fig. A12.15 shows the shape of the de- simple support with a cantilever overhang, the
flecting supporting structure, which means moment I1.. _~ is statically determinate and equals
that supports (2) and (3) move downward through 5x3x3.33 = 50 in.lb. Then for static equili-
the distances indicated on the figure. The brium of joint (2), I1..- 3 must equal - 50.
problem will be to determine the bending
moment at the supports under the combined Substitution in slope deflection equations
action of transverse loading and settling of (17) and (18): -
supports.
I1"_ 3 = 2K(26.. + 6 30) + I1F " - 3
3 -
SOLUTION: -
- 50 = 2 x 8347 [26.. + 63 - 3( .007812)J - 426
Calculation of fixed end moments: -
or = 333888.. + 1669463 + 15 = 0 - - - - - - - (1)
For a trapeqoidal loading as shown in
Fig. a, the fixed end moments are, I13 - .. = 2K(26 + 6.. - 30) + I1F 3 - " = 2 x 8347
3
[28 3 +6.. -3(-.007812)J +440
whence I13 _ .. = 333886 3 + 166946.. + 831 - - - - (2)
_ L"
I1.. _~ - 60 (5u + 3v) ~-4 = 2K(26 + 8 -30) +I1F
3 4 (8 3-4 4 = 0)
= 2 x 8347 [28 + 0 - 3(-.004688)J 3 494

SUbstituting in these
equations for the loading Fig. a whence I1 3 - 4= 333888 259 - - - - - - -
3 - - (3)
values as shown in Fig.
A12.4, I1 4 - 3 = 2K(28 4 - 63 - 30) + I1F 4 - 3
40 ..
I1F.._ 3 =60 (5x3+ 1) =-426 in.lb. M..-3 = 2 x 8347 [0 + 6 3 - 3(-.004688)J + 507
whence I1 4 - 3 = 166940 3 + 742 - - - ,... - (4)
For static equilibrium of joint (3),
40 ..
I1F3-4 =-60 (5x3.5+ 1) =.. 494 I1 3 - , . +I1 3 - 4 = 0, hence
3338863 + 166946,. + 831 + 33388~ - 259 =0

A12.7
whence, 667766" + 166946 2 + 572 = 0 - - - - - (5) MF~_2 = (100x8x12 2 )/20 2 + (100x12x8 2)/
Solving equations (1) and (5) for 8 2 and 20 2 =- 480
8" we obtain 8 2 = 0.00439 and 8" =- 0.00967.
MF 2_1. = 480
SUbstituting in equations (2), (3) and (4)
gives the moments at these points. MF ~_" = 100 x 24/8 = 300. 300
= 10 x 20 2/12 = 334. 334

M"_2 = 33388 (-.00967) + 16694x .00439
+ 831 = 582 in. lb. Due to symmetry of loading we know that
8~ = - 8 2 and 8" =- 8 4 , which fact will shorten
M"_4 = 33883 (-.00967) =-582
259 the solution.
M4-" = 16694 (-.00967) + 742 = 580 Slope deflection equations: -

Converting these signs to the conventional M~_2 = 2K (28~ + 8 2) + MF~_2
signs would give M2 = - 50, M" = - 582 and
M4 =-580. M~_2 = 2xl (28~ - 8~) - 480
A12.6 Statically Indeterminate Frames.
Joint Rotation Only.
M~-2 = 28~ - 480 ------ --- - - (a)
The slope deflection equations can be M~_" = 2K (28~ + 8,,) + MF~

"
used in solving all types of framed structures.
M~_" = 2 X1.67 (28~ + 6,,) + 300
In frames as compared to straight continuous
beams, the aXial loads in the members are
usually more important, however the influence M~_" = 6.678~+3.338,,+300 - - - - - (b)
on joint displacement due to axial deformation
of members is relatively small and is there-
M,,_~ = 2K (28" + 8~) + MF,,_~
fore usually neglected in most simple framed
structures. To illustrate the slope-deflection M,,_~ = 2 X1.67 (28" + 8~ ) - 300
method as applied to frames, the structure
shown in Fig. A12.16 will be solved. M,,_~ = 6.678" + 3.338~ - 300 - - - - - - (c)
Example Problem 1. M"-4 = 2K (28" + 8 4 ) + MF"_4

100 100 M"-4 = 2xl (28" - 8,,) + 334
~ 8" -I------ 4"-1-- 8,,----l M"-4 = 28" + 3-'34 - - - - - - - - - - - (d)
1 2
L = 20 Static Joint Equations: -
I = 20
K=l JOINT (1).
L = 24 L = 24 Which gives,
100 - - . I = 40 I = 401--100
K = 1. 67 K = 1. 67 28~-480+6.678~+3.338,,+300 = 0
L = 20 whence 8.678~ + 3.338" -180 =0 - - - - - - (e)

I = 20
K=l JOINT (3).
3tftttttirTittTTf 4
w = 10Ib/in. which gives,
Fig. A12.16
6.676" + 3.338~ - 300 + 28" + 334 = 0
Fig. A12.16 shows a closed rectangular
frame subjected to loadings on its four sides Whence 8.678" + 3.338~ + 34 =0 - - - - - - - (f)
which hold the frame in equilibrium. The Solving equations (e) and (f) for 8" and
bending moments at the frame joints will be 6~ gives,
determined.
SOLUTION: - 8~ =26.3, S" =-13.94
FIXED END MOMENTS: -

Subt. in equation (a) Due to the fixity at joints (1), (3) and
(5), 8~-s = 8 3 - 4 8~-6 = =
O. Therefore, the
1'1,.-s = 2x26.3-480 =-428 in.lb. unknowns are 8s-~, 8 4 - 3 , 8 6 - 6 and ~.
Subt. in equation (d) There are three statical joint equations

of equilibrium available, namely,
M3 - 4 = 2(-13.94) + 334 = 306 in. lb.
Ms-~ + MS - 4 = 0, M4 _ S + M4 - 6 = 0, and
Fig. A12.17 shows the resulting bending
M6- + M6-~ = 0
moment diagram. 4
Another equation is necessary because there are

four unknowns. This additional equation is
obtained by applying the equations of statics
to the free bodies of the vertical members.
Fig. A12.19 shows the free bodies.
Fig.A12.17 233 M 6-11

~
T Tr
L 1- 8
194
A12.7 Frames with Joint Displacements.

1
Fig. A12.19
In the previous example problems, the
conditions were such that only joint rotations Treating each member separately, we take
took place, or if any transverse joint dis- moments about the upper end and equate to zero
placement took place as in the example problem and then solve for the horizontal reaction at
of Art. A12.5, these displacements were known, the bottom end. The reSUlts are,
or in other words the value of 0 in the slope-
deflection was known.
In many practical frames however, the

joints suffer unknown displacements as for
example in the frame of Fig. A12.18. The term
o= ~/L in the slope-deflection equation was
derived on the basis of transverse displace-
ment of the member ends when both ends were
fixed. Thus in Fig. A12.18 under the action
of the external loads, will sway to the right With these horizontal reactions known the
as magnified by the dashed lines. Neglecting static equation ZH =
0 can now be applied to
any joint displacement due to axial deforma- the frame as a whole, which gives,
tion, the upper end of each member will move
through the same displacement~. Then we can (20)
write,
Equation (20) is generally referred to as
the bent or shear equation and is an equation
that supplies the necessary extra condition to
take care of the additional unknown ~.
Ps
A12. 8 Example Problems of Frames with Unknown
Joint Displacement.
Example Problem 1
Fig. A12.20 shows an unsymmetrical frame

with unsymmetrical loading. The bending
1 Fig. A12.18
moment curve will be determined.
SOLUTION: -
The relative moment of inertia of each
A12.9
member is given on 96# M.._.. = 2 x 1.667(26.. + 6.. - 0) - 144

the figure. The
relative K I/L
values are also
=
I 2 I 20
r
= 4
6
-1 M.._.. = 6.6678.. + 3.3338.. -144 - - - - - (c)
indicated adjacent 6" L

K
= 12"
= 1. 667 L : 10"
M.._ .. = 2 x 1.667(28.. + 8.. - 0) + 144
to each member.
Relative values
of 0 due to sidesway
48#
1-9·"
I = 10
L
K
=15"
= .667
I = 10
K =1
3
ITn
M.._ ..
M3 - ..
= 6.6678.. + 3.3336.. + 144 -
- 2x1(0+8.. -3xO.10)+0
- - - - (d)
of frame: -
The angles 0
are proportional to
1~~ Fig. A12.20
M3 _.. = 28.. - 0.60 - - - - - - - - - - (e)
the ~/L for each

member where ~ is M"_3 = 48.. - 0.60 - - - - - - - - - - - (f)
same for members
(1-2) and (3-4) and Equilibrium equations: -
zero for member (2-4).
(See Fig. A12.l0). JOINT (2).
Hence, SUbstituting: -
0 .. _~= l~ = .06670 2.6678.. - 0.26670 + 103.68 + 6 .6678.. + 3.3338..

Fig. A12.10
- 144 0 =
0.100 whence, 9.3338.. -0.26670+3.3338.. -40.32
=0 - - - - - - - - - - - - - - (g)
o JOINT (4).
02-"= 12 = 0
FIXED END MOMENTS: -
Member 1-2. whence, 10.6678.. + 3.3338.. - 0.60 + 144 =0 - (h)
MF
l.-"
= - (48 X 6 2 X 9)/15" =- 69.12 in. lb. Writing the bent equation (See Eq. 20).
MF .. _:I. = (48x9"x6)/15" = 103.68 Hl. + H3 - 48 =0 - - - - - - - - - (i)

Member 2-4. In order to get Hl. and H3 in terms of end
moments see Fig. A12.2l.
MF .._.. =- (96 x 12)/8 = -144
M .. _l. M ..- 3
MF.._ .. = 144
1J
~~H"
TrH..
Member 3-4 has no lateral loads and thus fixed 48 10
end moments are zero.
9
~~H3
The slope-deflection equations are: -
1 LJ'-H:I. M 3 _ ..
Ml.-.. = 2K(26:1. + 6.. - 30) +MF 1. (17) Ml._.. Fig. A12.21
M.._l. = 2K(28.. + 8:1. -30) +MF.. (18) For free body of member 1-2 take moments
about (2) and equate to zero.
Writing above equations for each member
and noting that 81. and 63 are zero because M..-l. + Ml._.. - 6 x 48 + l5H", =0
frame is fixed at support points, gives,
hence
,
H
:I.
= 288 - (M2 -1. +Ml._.. )
15
Ml.-.. = 2 x 0.667(0 + 8.. - 3 x .06670) - 69.12
SUbt. values of M.. _:I. and Ml._.. in the equation,
Ml.-.. = 1.3336.. - 0.26670 - 69.12- - - - (a)
M..-l. = 2 x 0.667(28.. + 0 - 3 x .06670) + 103.68 Hl. =- 0.26678.. + .035550 + 16.88 - - - - (j)
For free body of member 3-4 take moments

M.._", = 2.6678.. -0.26670+ 103.68 - - - (b) about (4) and equate to zero.
Me-.. + M..-3 + lOR 3 =0 ing the symmetrical distributed load as shown.
whence, R3 =- (M
3 to+M . 3
)
w = 12 Ib/in.
1illOllll
Substituting values of M3 _ .. and M.. _3 in above 3
equations, gives, 10"
t
R3 =- 0.68 .. + 0.120 - - - - - - - - - (k) 2 4
SUbstituting values of R ~ and R 3 in equation
(i), gives,
-0.26678 2 - .684 + .15550-31.12 =0- - (1) 2 0" I = 10 Fig. A12.23

I = 10
= 20 = 20"
1
L
Solving equations (g), (h), (1) for the
unknowns 8 2, 8 .. and 0 gives,
5
o= 196.9, 8 2 = 12.17, e.. = - 6.22 40"
The final end moments can now be found by
sUbstituting these values in equations (a) The bending moment diagram under the given load-
to (f) inclusive. ing will be determined.
M~-2 = 1.333x12.17- .2667x196.9 SOLUTION: -

- 69.12 = -105.4 in. lb.
Relative Stiffness (K values)
M2-~ = 2.667 x 12.17 - .2667 x 196.9
+ 103.68 =- 83.6 Members 1-2 and 4-5, K = 10/20 = 0.5
M2-" = 6.667 x 12.17 + 3.333 (-6.22) Members 2-3 and 4-3, K = 20/22.4 =0.892
- 144 = 83.6
Fixed end moments:
M..- 2 = 6.667 (-6.22) + 3.333 x 12.17 + 144
- 142.9 MF 2_ 3 = wL'h/12 = (12x20 2 )/12 =-400 in.lb.
M4 - 3 = 4 (-6.22) - 0.6 x 196.9 = -142.9 MF 3 _ 2 = 400, MF.._ 3 = 400, MF 3 _ .. = - 400
M3 - .. = 2 (-6.22) - 0.6 x 196.9 = -130.45 Relative values of 0: -
Fig. A12.22 shows the bending moment Due to the sloping members, the relative
diagram. The student should also draw the transverse deflections of each member are not
shear diagram, find axial load in members as obvious as when the vertical members are con-
since all these loads are needed when the nected to horizontal members as in the previous
example. In this example, jOints (1) and (5)
are fixed. Because of symmetry of frame and
174.7 loading, joints (2) and (4) will move outward
the same distance ~ as indicated in Fig. A12.24.
Furthermore, due to symmetry, joint (3) will
undergo vertical movement only.
In Fig. A12.24, draw a line from point (2')

parallel to 2-3 and equal in length to 2-3 to
locate the point (3"). Erect a perpendicular
to 2'-3" at 3" and where it intersects a vertical
105.4
Fig. A12.22
through (3) locates the point (3') the final
location of joint (3). The length of 2'-3'
strength of the frame members are computed equals 2-3. In the triangle 3-3'-3", the side
and compared against the stresses caused by 3' -3" =...;r;t;. The relative values of 0 which
the frame loading. are measured by ~/L for each member can now be
calcUlated.
Example Problem 2. 1
0~-2 = - 20 = - .050
Fig. A12.23 shows a gable frame, carry-
A12.11
11 4- 6 = 2 x 0.5(-28 2 + 0 - 3 x .05¢) + 0
3..t3~~ 11 4- 6 =-28 2 -0.15¢ - - - - - - - (g)
3'---_ j~r
4'
/ - - - - (h)
/
f JOINT (2). Equilibrium equation.
1
112-~+112-3 = 0, sUbstituting-
5
282 + 0.15¢ + 3 ..56882 - 0.5352¢ - 400 =0
Fig. A12.24 whence, 5.5688 2 - 0.3852¢ - 400 =0 - - - - (1)
04-6 = .05¢ The joint equilibrium equations at (3) and

(4) will not provide independent equations be-
SUbstitution in Slope Deflection Eguations: - cause in the previous substitutions in the
slope-deflection equations 8~ was made equal to
We know that 8~ 86 = =
0, due to fixity zero and 84 was equal to - 82 •
at joints (1) and (5). Also 8~ = 0 due to
symmetry or only vertical movement of joint Shear Equations: -
(3). Furthermore due to symmetry 84 82 =-
Due to symmetry the horizontal reactions
at points (1) and (5) are equal and opposite
and therefore in static balance. Since we have
11 ~-2 = 2 x 0.5 (0 + 8 2 + .05 x 3¢) + 0 two unknowns 82 and ¢, we need another equation
to use with equation (i). This equation can be
11 ~-2 = 82 + 0.150 - - - - - - - - - - (a) obtained by stating that the horizontal reaction
H2-~ on member 2-1 at end (2) must be equal and
opposite to H2_~' the horizontal reaction at
end (2) of member 2-3.
112-~ = 2xO.5(282 + .05x3¢) +0
Fig. A12.25 and 26 show free body sketches
112-~ = 28 2 + 0.15¢- - - - - - - - - - (b) of members 1-2 and 2-3.
11 2-~ = 2K(28 2 + 8~ - 3¢) +11F 2-~
112-~ = 2xO.892(28 2 +0- .3xO.l¢) -400
112_~ = 3.5688 2 -0.5352¢-400 - - - - (c) T

20"
1 1
"-.:./-H~
Fig. A12.26
M ~-2
11~_2 = 1.78482 - 0.5352¢ + 400 - - (d) Fig. A12.25
11 ~-4 = 2K(28~ + 84 - 30) +11F ~-4

In Fig. A12.25 take moments about (1) and
equate to zero.
11~_4 = 2 x .892(0 - 82 + 3 x O.l¢) - 400
11 ~-2 +11 2-~ - 20H 2-~ =0
11 ~-4 = -1. 7848 2 + 0.5352¢ - 400 (e)
whence, H 2-~
= 11~-2+112-~
20
114-~ =2x .892(-282 + 0 + 3 x O.l¢) + 400 and in Fig. A12.26, taking moments about (3),
114_~ =- 3.5688 2 + 0.5352¢ + 400 - - - (f) 11 2-~ +11 ~-2 + 240 x 20 - 20 x 12 x 10 - 10H 2-~ =0
11 4_ 6 = 2K(284 + 86 -30) +11F = 11 2 ~ +11 ~ 2 + 2400
4-6 whence, H 2-~ 10
Shear equat i on, H.. _~ = H "-3 Fig. A12.27 shows the resulting bending
moment diagram, first drawn in parts for each
whence, member of the frame and then added to form the
composite diagram plotted directly on the
M~_ .. +M .. _~ _ M.._ 3 +M 3 _ .. +2400 frame. The shears and axial loads follow as a
20 - 10 matter of statics. With the moment diagram
known the frame deflected shape is readily
SUbstituting values for the end moments: - calCUlated.
8.. + 0.150 + 28.. + 0.150 _ 3.5688.. - 0.53520 A12.9 Comments on Slope-Deflection Method.
20 - 10
The example problems solutions indicate
- 400 + 1. 7848.. - 0.53520 + 400 + 2400 that the method of slope-deflection has two ad-
10 vantages, namely, (1) it reduces the number of
equations to be solved simultaneously and (2)
whence, it presents equations that are easily and
rapidly formulated.
0.38528.. - 0.1220 + 240 =0 - - - (j)
Thus for structures with a high degree of
Solving equations i and j for 8 .. and 0, gives redundancy, the slope-deflection method should
be considered as possibly the best method of
o = 2805, 8.. = 265.8 solution. The solving of the equations in this
method are readily programmed for solution by
SUbstituting these values in equations (a) to high speed computing machinery.
(h), the end moments are obtained as follows: -
A12. 10 Problems.
M~_ .. = 8.. + O. 150 = 265.8 + 0.15 x 2805
= 686.5 in. lb. (1) Determine the bending moments at support
points A, B, C, D, for the continuous beam
M.. -~ = 28.. + 0.150 = 531.6 + 420.7 = 952 shown in Fig. A12.28.
M"-3 = 3.5688.. - 0.53520 - 400

= 3.568 x 265.8 - 0.5352 x 2805 - 400
=- 952
M3-" = 1.7848.. - 0.53520 + 400
= 1.784 x 265.8 - 0.5352 x 2805 + 400
628 = Fig. A12. 28
M3 - 4 =-1.7848.. +0.53520-400 =-628 (2) Same as problem (1) but consider support A
as freely supported instead of fixed.
M4 - 3 = 952 (3) Determine the bending moment diagram for
M4 _ 6 =- 952 the various loaded structures in Fig.A12.29.
w = 12 lb/in.
M6-4 =- 686.5
,?>f;}
~~...L.,~--"Br- \>~ f;}
\ ~'!)
A
100 100 lb.
300 lb.
A B
1000 1000 lb.
w = 20 lb/in.
4 -L
J~ r
C 6"
V,c. D
= 1 1= 3
20" 1=3
L = 20
Fig. A12.27
B
1 f-- A
20"-------oj
E
A Fig. A12.29
CHAPTER A13
BENDING STRESSES
A13.0 Introduction. applied moment the beam takes the exaggerated

The bar AB in Fig. a is subjected to an deflected shape as shown.
axial compressive load P. If the compressive The applied bending moment vector acts
stresses are such that no buckling of the bar parallel to the Z axis, or in other words the
takes place, then bar sections such as 1-1 and applied bending moment acts in a plane perpend-
2-2 move parallel to each other as the bar icular to the Z axis. Consider a beam segment
shortens under the compressive stress. of length L. Fig. d shows the distortion of
this segment when plane sections remain plan~
11 12 after bending of the beam.
p---+lL... -'-I--'-' ~I--P
A 11 12 B It will be assumed that the beam section is
homogeneous, that is, made of the same material,
Fig. a and that the beam stresses are below the pro-
--t'l# portional limit stress of the material or in
I
I
I
I
other words that Hooke's Law holds.
I I
n-fj'-n
I I
From the geometry of similar triangles,
I I
, I
ShearD~
--hxl<- e - ec or e: L e c (A)
y c c
./ Fig. b Fig. c
. / Moment Dia. .............
We have from Youngs Modulus E, that
In Fig. b the same bar is used as a simply
supported beam with two applied loads P as E : unit stress : -cry
shown. The shear and bending moment diagrams unit strain elL
for the given beam loading are also shown. The
portion of the beam between sections 1-1 and 2-2 where ay is the bending stress under a deforma-
under the given loading are subjected to pure tion e and since it is compression it will be
bending since the shear is zero in this region. given a minus sign.
Experimental evidence for a beam segment Solving for ay,
Ax taken in this beam region under pure bending
shows that plane sections remain plane after a : - E ec y
bending but that the plane sections rotate with y c L
respect to each other as illustrated in Fig. c,
where the dashed line represents the unstressed The most remote fiber is at a distance y : c.
beam segment and the heavy section the shape Hence
after pure bending takes place. Thus the top
fibers are shortened and subjected to compress- E ec
ive stresses and the lower fibers are elongated L
and subjected to tensile stresses. Therefore at whence
some plane n-n on the cross-section, the fibers
suffer no deformation and thus have zero stress. a
y
=- ac y _ - - - _ _ _ _ _ _ _ _ _ _ _ (B)
This location of zero stress under pure bending c
is referred to as the neutral axis.
For equilibrium of the bending stress perpend-
A13.l Location of Neutral Axis. icular to the beam cross-section or in the X
Fig. d shows a cantilever beam subjected to direction, we can write ~ Fx = 0, or
a pure moment at its free end, and under this
~ Fx "" - acc f yda = O.

however in this expression, the term ac is not
zero, hence the term f
yda must equal czero and
Beam thiS can only be true if the neutral axis coin-
Section cides with the centroidal axis of the beam
Fig. d cross-section.
A13.l
A13.2 BEAM BENDING STRESSES
The neutral axis does not pass through the Thus, Yn = (y - x tan 0) cos 0 - - - -
beam section centroid when the beam is nonhomo-
geneous that is, the modulus of elasticity is not
- y cos 0 - x sin ---- - o- (2 )
constant over the beam section and also when Then Eq. (1) becomes
Hooke's Law does not apply or where the stress-
strain relationship is non-linear. These beam o = k (y cos o-
x sin 0) - - - - - - (3 )
conditions are described later in this Chapter.
This equation contains three unknowns, 0,
A13.2 Equations for Bending stress, Homogeneous Beams, k, and 0. For solution, two additional equa-
stresses Below Proportional Limit stress. tions are furnished by conditions of equili-
In the following derivations, it will be brium namely, that the sum of the moments of the
assumed that the plane of the external loads external forces that lie on one side of the
contain the flexural axis of the beam and hence, section ABCD about each of the rectangular axes
the beam is not subjected to torsional forces X-X and Y-Y must be equal and opposite, respect-
which, if present, would produce bending stress- ively, to the sum of the moments of the internal
es if free warping of the beam sections was re- stresses on the section about the same axes.
strained, as occurs at points of support. The Let M represent the bending moment in the
questions of flexural axes and torsional effects plane of the loads; then the moment about axis
are taken up in later chapters. X-X and Y-Y is Mx ~ M cos Q and My = M sin Q.
m The moment of the stresses on the beam section
about axis X-X is /0 da y. Hence, taking
moments about axis X-X, we obtain for equil-
Plane ibrium,
of loads
M cos Q =/ 0 da y
- / k (cos 0 y 2 da - sin 0 xyda)

X --\---:~~---''---+-------1.... X
=k cos 0 / y 2 da - k sin 0 / xyda (4 )
N----- m
In similar manner, taking moments about
the Y-Y axis
Fig. A13.1 m Fig. A13. 2 M sin Q =/ 0 da x
whence
Fig. A13.1 represents a cross-section of a
straight cantilever beam with a constant cross- M sin Q = - k sin ¢ / x 2 da + k cos %/ xyda(4a)
section, subjected to external loads which lie
in a plane making an angle Q with axis Y-Y A13.3 Method 1. stresses for Moments About the
through the centroid O. To simplify the figure, Principal Axes.
the flexural axis has been assumed to coincide In equation (4), the term / y'"da is the
with the centroidal axis, which in general is moment of inertia of the cross-sectional area
not true. about axis X-X, which we will denote by Ix, and
Let NN represent the neutral axis under the the term / xyda represents the product of in-
given loading, and let 0 be the angle between ertia about axes X-X and Y-Y. We know, however,
the neutral axis and the axis X-X. The problem that the product of inertia with respect to the
is to find the direction of the neutral axis and principal axes is zero. Therefore, if we se-
the bending stress 0 at any point on the section. lect XX and IT in such a way as to make them
In the fundamental beam theory, it is as- coincide with the principal axes, we can write
sumed that the unit stress varies directly as equation (4):
the distance from the neutral axis, within the
proportional limit of the material. Thus, Fig. M cos Q = k cos 0 Ix (5)
A13.2 illustrates how the stress varies along a p
line such as mm perpendicular to the neutral
axis N-N. In like manner, from equation (4a)
Let 0 represent unit bending stress at any
point a distance Yn from the neutral axis. Then M sin Q k sin 0 Iy (6 )
p
the stress 0 on da is
To find the unit stress 0 at any point on the
(1) cross-section, we solve equation (5) for cos 0
and equation (6) for sin 0, and then SUbstitut-
where k is a constant. Since the position of ing these values in (3), we obtain the follow-
the neutral axis is unknown, Yn will be expressing expressing, giving 0 the subscript b to
ed for convenience in terms of rectangular co- represent bending stress: -
ordinates with respect to the axes X-X and Y-Y.
A13.3
M cos Q yp + M sin Q x p moment of inertia about the neutral aXis, hence

I xp I yp
ab = [M cos (Q - a)J Yn =MnYn (9)
Let the resolved bending moment M cos Q and
In r;-
M sin Q about the principal axes be given the
symbols Mxp and Myt' Then we can write In can be determined from the relationship ex-
pressed in Chapter A3, namely,
(7)
A13.5 Method 3. Stresses from Moments, Section Prop-
erties and Distances Referred to any Pair of
The minus signs have been placed before each Rectangular Axes through the Centroid of the
term in order to give a negative value for ab Section.
when we have a positive bending moment, or Mxp The fiber stresses can be found without
is the moment of a couple acting about xp-x p , resort to principal axes or to the neutral
positive when it produces compression in the axis.
upper right hand quadrant. Myp is the moment
Equation (4) can be written:
about the Yp-Y p axis, and is also positive when
it produces compression in the upper right hand Mx =k cos 0 Ix - k sin 0 I xy (ll)
quadrant.
where Ix = I y 2 da and I xy =I xyda, and Mx =
BENDING STRESS EQUATION FOR SY11METRICAL M cos Q.
BEAM SECTIONS In like manner,
Since symmetrical axes are principal axes
(term J xyda = 0), the bending stress equation
My =- k sin 0 Iy 0 I XY - - - - - (12)
+ k cos
for bending about the symmetrical XX and YY axes Solving equations (11) and (12) for sin 0 and
is obviously, cos 0 and substituting their values in equation
(3), we obtain the following expression for the
ab =- MxY _ Myx _ - - (7a) fiber stress ab: -
Ix Iy
ab = _ (MyIX - MxIxy ) x _ (MxI y - MyIXY)y _ (13)
A13.4 Method 2. Stresses by use of Neutral Axis for IXI y - I~y IxI y - I~y
Given Plane of Loading.
The direction of the neutral axis NN, mea- For simplification, let
sured from the XX p principal axis is given by
dividing equation (6) by (5). K~ = Ixy/(IxI y - I XY ) 2
Tan a =- Ix
p
tan Q - - - - - (8) K2 = Iy/(IXI y - I\y)
K3 = Ix/(IXI y - I xy ) 2
SUbstituting these values in Equation (13): -
ab =- (K 3 My - K~Mx) x - (K 2 Mx - K~My)Y - (14)

Xp --+--=--¥=-I-----'i'-
In Method 2, equation (8) was used to find
N
the position of the neutral axis for a given
plane of loading. The location of the neutral
axis can also be found relative to any pair of
rectangular centroidal axes X and Y as follows: -
Since the stress at any point on the neutral
The negative sign arises from the fact that axis must be zero, we can write from equation
Q is measured from one principal axis and 0 is (14) that: -
measured in the same direction from the other (K 3 My - K~Mx)x =- (K 2 Mx - K2 My )y for all
principal axis. points located on the neutral axis. From Fig.
Since equation (8) gives us the location of
the neutral axis for a particular plane of load- A13.1 tan 0 = ~ ,
x
ing,the stress at any point can be found by re-
solving the external moment into a plane perpen- Thus tan 0 =_ (K My 3 - K~Mx) - - - - - - - (15)
dicular to the neutral axis N-N and using the (K~x - K~My)
It frequently happens that the plane of the A13.8 illustrative Problems. Example Problem 1.
bending moment coincides with either the X-X or Fig. A13.3 shows
the Y-Y axis, thus making either Mx or My equal a unsymmetrical one
to zero. In this case, equation (15) can be cell box beam with
simplified. For example, if My = 0 four corner flange
members a, b, c and d.
Ixy Let it be required to
tan0=- - - - - (16) determine the bending
I y
axial stress in the
and i f Mx =0 four corner members
due to the loads Px
tan o =~-
I xy
- - - - - - - - - - - - - - - (17) and Py acting 50"
from the section
abed.
A13.6 Advantages and Disadvantages of the Three In this example
Methods. solution the sheet 121t
Method 2 (bending about the neutral axis connecting the corner ~ d ~X'
for a given plane of loading) no doubt gives a members will be con - 8[J1t JI'·I.---16-It-~---e.; ,4[J1t
better picture of the true action of the beam sidered ineffective in . y'
relative to its bending as a whole. The point bending. The stresses Fig. A13. 3
of maximum fiber stress is easily determined by will be determined by
placing a scale perpendicular to the neutral each of the three me-
axis and moving it along the neutral axis to thods as presented in this chapter.
find the point on the beam section farthese away
from the neutral axis. In airplane design, SOLUTION
there are many design conditions, which change
the direction of the plane of loading, thUS, The first step common to all three methods
several neutral axes must be computed for each is the calculation of the moments of inertia
beam section, which is a disadvantage as com- about the centroidal X and Y axes. Table A13.1
pared to the other two methods. gives the detailed calCUlations. The proper-
In determining the shears and moments on ties are first calculated about the reference
airplane structures, it is common practice to axes x'x' and y'y' and then transferred to the
resolve air and landing forces parallel to the parallel centroidal axes.
airplane XYZ axes and these results can be used
directly in method 3, whereas method 1 requires TABLE A13.1
a further resolution With respect to the prin- Area
cipal axes. Methods 1 and 3 are more widely Mem. y' x' Ay' Ax' Ay,2 Ax,2 Ax'y'
"Aft
used than method 2. -192
a 1.0 12 -16 12 -16 144.0 256.0
Since bending moments about one principal b 0.5 8 0 4 00 32.0 0 0
axis produces no bending about the other prin- c 0.8 0 -16 0 -12.8 0 204.8 0
cipal aXiS, the principal axes are convenient d 0.4 0 0 0 0 0 0 0
axes to use when calculating internal shear flow Totals 2.7 16 -28.8 176.0 460.8 -192
distribution.
Calculation of Centroid of Section: -
A13.7 Deflections.
The deflection can be found by using the
beam section properties about the neutral axis x = ZAx'
ZA
:: - 28.8
£.7
~ _ 10.667"
for the given plane of loading and the bending
moment resolved in a plane normal to the neutral - ZAy' 16
Y = LU"I.
"'. = - - = 5.926"
2.7
axis. The deflection can also be found by re-
solving the bending moment into the two prin-
cipal planes and then using the properties about Ix = 176 - 2.7 x ~ = 81.18
the principal axes. The resultant deflection is
the vector sum of the deflections in the direc- Iy =460.8 - 2.7 x 10.667 = 153.58
2
tion of the principal planes.

I xy = - 192 - (2.7 x - 10.667 x 5.926) = - 21.33
A13.5
Solution by Method 1 (bending about principal axes) (Positive moments produce compression in upper
The angle 0 between the x axis and the right hand quadrant).
principal axes is given by the equation,
Calculation of Stresses
2 I XY
2 (- 21. 33 ) - 42.66
tan 2 0 = -=---'~ = 153.58 - 81.18· 72.40 Myp xp
I y - Ix
I yp
=- .589
Stringer a
2 0 = 30° - 30' , o - 15° = 15' xp =- 6.74, yp = 4.45
I xp = Ix cos 2
0 + I y sin
2
0- 2 I xy sin 0 cos 0 hence
268700 x 4.45 - 156200 (- 6.74)
= 81.18 X .9648 2 + 153.58 X .2636 2 - (- 21.33 x 75.38 159.34
2
- .2636 x .9648)2 15900 - 6600 22500#/in.
= 75.56 + 10.67 - 10.85 = 75.38 in. 4 Stringer b
Iy = Ix sins 0 + I y cos 2 0 + 2 I xy sin 0 cos 0 x

P
= 9.75", yp .. 4.82
p
.. 81.18 X .2636 2 + 153.58 X .9647 2 + 10.85 - 268700 x 4.82 - 156200 x 9.75
75.38 159.34
159.34 in. 4
2
IY /Yp
17180 + 9570 7610#/in.
I / Fig. A13.3a Stringer c

6.74'L./
11--9.75"
x p = - 3.58, yp =- 7.12
- 268700 x - 7.12 - 156200 (- 3.58)

75.38 159.34
= 25400 3520 = 21880
and similarly for stringer d, ab = 21900.

Solution by Method 2 (Neutral Axis Method)
Let Q' .. angle between yy axis and plane
of loading.
tan Q' = - 80000 =- .2667

300000
hence
Q' = 14° - 56' and Q = 14° 56 t + 15° - 15' =
Mx = 6000 x 50 : 300,000"i = 30° - 11
My =- 1600 x 50 =- 80,000"# Let a =

angle between neutral axis NN and
the Xp axiS,
These moments will be resolved into bending
moments about the x p and yp principal axes. Ixp tan Q - 75.38 (- 0.5816)
tan a =- -----''------.,,...---
I 159.34 = .275
yp
Mx p = 300,000 x cos 15° - 15' - 80000 x
hence a =
15° - 24' (see Fig. A13.3b).
sin 15° - 15' = 289700 - 21000 = 268700"~ Since the angle between the X axis and the
neutral axis is only 9' , we can say
Myp 300000 x sin 15° - 15' 80000 x IN = Ix = 81.18.
Resolving the external bending moments
normal to neutral aXis, we obtain
cos 15° - 15' =- 79000 - 77200 =- 156200"#
MN = 300000 x .9999 + 80000 x sin 0° - 9' = Kg = Iy/(IxI y - I 2XY ) =

153.58/12016 = .012794
300200"#. The bending stress at any point is
given by, K3 = Ix/(IXI y - I 2xy ) = 81.18/12016 = .00674
Stringer a stringer a
Yn = 6.074 + 5.33 x .0025 = 6.087" x =- 5.333", y = 6.074"

300 200 x 6.087 = _ 22500~/in.2 ab ~ - [.00674 (- 80000) + .00177 x 300000J x
81.18
Stringer b [.012794 x 300000 - .00177 (- 80000)J y
Yn = 2.055" hence ab = (8) x - (3697)y;
- 300 200 x 2.05

81.18
- 7570 hence ab ~ 8 x - 5.33 - 3697 x 6.074 =- 22450
Stringer b
Stringer c
x = 10.667, Y = 2.074
Yn =- 5.92
ab =8 x 10.667 - (3697) 2.074 = 85 -7660 = - 7575
ab
=- 300 200 (- 5.92)
81.18
~ 21850
Stringer c
Stringer d x = - 5.333, y =- 5.926
y =- 5.95 ab ~ 8 x 5.333 - (3697 x - 5.926)
ab
=- 300 200 (- 5.95)
81.18
=22000 42 + 21900 = 21862
Stringer d
y
Yp x = 10.667, Y =- 5.926
9' =14°56'
~ = 15°15'
ab =8 x 10.667 - (3697 x - 5.926) =
9=30°11'
plane of ,loading
= 85 + 21900 = 21985
9' 524' N
X~;;;;;;;::~==-=~:£;F::;=:~ NOTE: The stresses Db by the three methods
-15°15'
were calculated by 10" slide rule, hence the
small discrepancy between the results for the
three methods.
Yp Y Fig. A13.3b
Error in Stresses Due to Assumption that Section
Bends About X and Y Centroidal Axes Due to Mx
Solution by Method 3 (Method Using Properties and My.
About X and Y Axes)
ab =- Mx Y _ My x (Equation 7a)
Mx =300000, My ~ - 80000 Ix Iy
Ix = 81.18, I y = 153.58, Stringer a
I xy =- 21.33 y = 6.074, x =- 5.33

- 21.33 _ 300000 x 6.074 - 80000 x - 5.33
= 81.18 x 153.58 - 21.33" - ab = 81.18 153.58 =- 25190
- 21.33 Stringer b
12016 =- .00177
y = 2.07, x = 10.667
A13.7
stringer b - cont'd. Mx = 14260 x 50 713,000"*
300000 x 2.07 - 80000 x 10.667 M =- 760 x 50 = - 38000"~

ab =- 81.18 153.58 = - 2nO Y
These moments are resolved into bending
In like manner for stringers c and d moments about the principal axes, as follows: -
abc = 19130 and abc = 27440 Mx = 713000 x cos 0 - 38000 x sin 0 = 700,000"#
p
Comparing these results with the previous My

P
=- 71300 x sin 0 - 38000 x cos 0 =-140000"#
results it is noticed that considerable error
eXists. Under these eroneous stresses the in- From equation (7), the general formula for ab is:
ternal resisting moment does not equal the ex-
ternal bending moments about the X and Y aXis.
Example Problem 2.
Fig. A13.4 shows a portion of a cantilever
2-cell stressed skin wing box beam. In this ex- Stress on Stringer 1:
ample, the beam section is considered constant,
and the section is identical to that used in yp = 1.85" and x p 17.86" (scaled from fUll
size drawing).
14260# (700000 x 1. 85 ) - 140000 x (- 17.86)

181.17 437
7150 - 5700 - - 12850Vsq. in.

stress on Stringer 9:
yp = 9.04", x p = 14.24"
(700000 x 9.04) - 140000 x 14.24

ab -- 181.17 437
- - 34,900 + 4560 - - 30340:\Vsq. in.
Stress on Stringer 12:
yp = - 6.80", x p = - 8.22"
700000 x (- 6.80) - 140000 x (- 8.22)

Fig. A13.4 ab - - 181.17 437 =
26200 + 2630 = + 23570i/sq. in.

Fig. A3.13 of Chapter A3 for which the section Solution by Method 2 (Neutral Axis Method)
properties were computed and are as follows: - In Fig. A13.5 let Q' be the angle between
the Y-Y axis and the plane of the resultant
Ix = 186.5 in. I y = 431.7 in. 4 I XY = 36.41 in. 4 bending moment. Resultant bending moment,
o = 8° - 16.25" I xp = 181.2 I yp = 437
M = )713000 2 + 38000 2 = 714600" lb.
The resultant air load on the wing outboard
of section ABC is 14260* acting up in the Y di- tan Q' = - 38000 __ .0533, hence Q' = _ 3° - 3'
713000
rection and 760* acting forward in the X direc-
tion, and the location of these resultant loads Let Q equal the angle between the plane of
is 50 from section ABC (Fig. A13.4). the resultant moment and the Yp axis.
~he bending stress intensity at the cen-
troids or stringers number (1), (9) and (12) Then Q = Q' + 0 = 3° 3' + 8° - 16' = 11° - 19'
will be calculated using all three methods.
From equation (8), the angle between the Xp axis
Solution by Method 1 (Bending about Principal Axes)
and the neutral axis = Q, and
The bending moment at section ABC about the
X and Y axes is, I xp tan Q 181.17 x (- .200)
tan a
I yp 437 =
Y Yp - 36.41
Plane of ~' = 186.46 x 431.7 - 36.41"
Resultant 9
Moment-
-.:,...\-++---------. .00046
\
\ I'
K" = I y/ (I x I y - I "xy ) = 186.46 431. 7 _ 36.41"
x 431.7
-~~-tLxp = 431. 7 = .00545

793000
= 186.46 .002355
793000
stress on Stringer 1:
Y Fig. A13.5
y~ =4.39", x~ =- 17.41"
= .083. Hence, a = 4° - 45'
~ig. A13.5 shows the relative positions of
the neutral axis, principal axes, and plane of
loading. ab = [.002355 x (- 38000) - (- .00046 x 713000)J
~
The component of the external resultant
moment about the neutral axis N equals: - (-17.41 )
MN = 714060 x sin 83° - 26' = 709350 in. lb. - [.00545 x 713000 - (- .00046 x - 38000)J 4.39
IN = I xp cos" a + I yp sin" a (238.5)(- 17.41)- (3868) 4.39

183.37 in.· = 4150 - 17000 = - 12850i/ln. 2
stress on Stringer 1: Stress on Stringer 9:

Yn~ = (yp + x p tan a) cos a = (1.85 + 17.86 x Ye = 6.89", Xe = 15.39"
.083) .9966 = 3.32" abe = - [238.5J 15.39 = [3868J 6.89 =-30320#/in."

ab =- Mn Yn =- 709350 x 3.32 =_ 12850i/sq. in.
~ In 183.37 Stress on Stringer 12:
Stress on Stringer 9: y~" =- 5.55", x~" =- 9.11"
Yn e = (9.04 - 14.24 x .083) .9966 = 7.84" [238.5J (- 9.11) - [3868J (- 5.55) =
abe =- 709350 x 7.84

183.37 =- lI.
30300ft/sq. in. - 23620i/in."
NOTE: In the three solutions, the distance
Stress on Stringer 12: from the axis in question to the stringers 1,
9, and 12 have been taken to the centroid of
= (6.80 + 8.22 x .083) .9966 = 6.10 each stringer unit. ThUS, the stresses ob-
tained are average axial stresses on the
= - 709350 x - 610 = 23600*/sq in. stringers. If the maximum stress is desired,
183.37 • the arm should refer to the most remote part of
the stringer or the skin surface.
Solution by Method 3
Approximate ..Method.
Mx = 713000"#, My = 38000"* It is sometimes erroneously assumed that
the external bending moments Mx and My produce
Ix = 186.46 I y = 431. 7 I xy =- 36.41 bending about the X and Y axes as though they
were neutral axes. To show the error of this
The constants K~, K", and K3 are first deter- assumption, the stresses will be computed for
mined - stringers 1 and 9.
~500
Stringer 1:
y~ = 4 + .39 = 4.39
x~ =- 33.15 + 15.74 =- 17.41
r-- 48 "
=- 713000 x 4.39 (- 38000 x - 17.41) = Fig. A13.8 Fig. A13.9

crb 186.46 431.7 stress intenSity at bottom edge of portion A is
- l8330#/in 2 •
crb (2.25/3) 0.75 = crt.
stringer 9: TA = CA' = Average stress times area
yg = 6.89, xg = 15.39
- 713000 x 6.89 (- 38000 x 15.39)
°b = 186.46 431.7 The distance from the neutral axis to the
centroid of the trapizoidal stress loading on
_ 25000#/in 2. portion A is 2.64 inches. In like manner,
Example Problem 3.
The previous example problems were solved
TB = CB' = (0 +2 crt) 3 x 0.25 = 0.375crb
by sUbstituting in the bending stress equations.
The student should solve bending stress prob- The arm to TB is 0.667 x 3 = 2 inches.
lems by equating the internal resisting moment
at a beam section to the external bending mom- For equilibrium of the beam free body, take
ent at the same section. To illustrate Fig. moments about point (0) and equate to zero.
A13.6 shows a simply supported loaded beam. The
shear and bending moment diagram for the given ZM o =- 500 x 48 + 2.64 x TA + 2.64 CA' + 2 TB +
beam loading is also shown. Fig. A13.7 shows
the beam section which is constant along the 2 CBl =0
span.
500# 1000 300 Substituting values as found for TA, TB etc,
~ .60 1 •
A ;AI ! - 500 x 48 + 2.64 x 1.15 crb + 2.64 x 1.15 crb
T
!
~8" ..-r-- l20"-----r36'~

1110 690 + 2 x 0.375 crb + 2 x 0.375 crb =0
610# _ 6"
~ 1
390
p hence crb = 24000
7.58 = 3170 ps i .
100
~OO"# Solution using Bending Stress Formula.
_ Mc
Fig. A13.7
crb - -1- where M =- 24000
c = 3"
I =Moment of inertia about
Neutral axis.
Calculation of I.
For Portions A and A'
For Web Portions b, b",
I = l~ x 0.25 X 6 3 = 4.50 in 4 •
ITotal =22.76 in
4
•
hence ternal force action from this method of solu-

tion. In solving nonhomogeneous beams and
°b =- 24000 x 3 = _ 3170
22.76
i
ps ., beams stressed above the elastic limit stress,
this method of solution often proves necessary
which checks first solution. or advantageous because no simple beam bending
stress formula can be derived.
Example Problem 4.
In Fig. A13.12, let 0b be the intensity on
Fig. A13.10 shows a loaded beam and Fig. the most remote fiber, or 3.09" above the neu-
A13.11 shows the cross-section of the beam at tral axis. Table A shows the calculations of
section a-a'. Determine the magnitude of the the total stress on each of the portions of the
maximum bending stress at section a-a' under the cross-section and their moment about the neutral
given beam loading. The beam section is unsym- axis, all in terms of the unknown stress, 0b'
metrical about the horizontal centroidal axis. TABLE A
Simple calculations locate the neutral axis as
shown in Fig. A13.11. 1 2 3 4 5 6
_
12#/i~.
Portion Area Average Total y = arm Resisting
~5#/in.
A stress Load to N.A. moment
See in = (2) (4) x (5)
Fig. terms x (3)
~ 100"-------+ 30"~ 100" ~ A13.13 of 0b
1 1 _~tf-- 2" - ..d-f-
Fig. A13. 10 RA RB r 1 1.045 -.3370b -. 353 Ob 1. 39 ... 490 Ob
1
1 1/2 2.35 -1. 780
2 1.00 -.757 -.757
I d7.2 in. i -- - 3.09" 3 - 3' 1.00 -.837 -.837 2.62 -2.190
31- 1 4 1.09 .350 .387 -1.44 -0.557
LN- N1'" 5 - 5' 0.56 .578 .323 -1. 93 -0.622
il
I
6 1. 50 • 821O b 1. 2370b -2.57 -3.180 0b
Fig. AI3.11 Totals 6.19 O.OOOOb -8.82 0b

:: 2
~ r.-2"--j Explanation of Table A.

Fig. A13.12 shows a free body of the por- Column 3 gives the average stress on each
tion of the beam to the right of section a-a. of the 6 portions. For example, the stress on
The bending stress intensity diagram is shown by portion (1) varies from zero at the neutral
the horizontal arrows acting on the beam face at 2.09 0b
ax i s t 0 3.09 .674 0b a t the upper edge of
=
section a-a. Fig. A13.13 shows the cross-sec-
tion of the beam at section a-a. (1). ThUS, the average stress _ •674ob + 0 =
The general procedure will be to determine 2
the total bending stress load on each portion, .337 0b' On part 3 or 3' , the stress on the
1 to 6, of the cross-section and then the mom- lower edge = .674 0b and 0b on the upper edge.
(1 + .674)
ent of each of these loads about the neutral Then, the average stress = 2 0b =
axis, the summation of which must equal the ex- .837 0b' Column 5 is the distance from the
ternal bending moment.
neutral axis to the centroid of the load on
each of the blocks. For portion (1) the stress
pattern is triangular, and y = 2/3 x 2.09 =
1.39". On portions (2), (3), (5) and (6) the
bending stress distribution is trapezoidal, and
the arm is to the centroid of this trapezoid.
The total internal resisting moment of
- 8.82 0b from Table A equals the external bend-
ing moment of 36670"#.
100" ~ ThUS,
Fig.AI3.13 a - 36670
Fig. A13.12 0b =
8.82 =-
4160#/in. 2 •
This method of solution involves more cal- For equilibrium, the total compressive
culations than that required in sUbstituting in stresses on the cross-section of the beam must
the bending stress formula, however, the stUdent be equal to the total tensile stresses, or m
should obtain a better understanding of the in- must equal zero. Column 4 of Table A gives the
A13.11
total load on each portion of the cross-section

and the total of this column is zero.
The bending stress on the lower fiber of
the cross-section is directly proportional to
ower =
the distance from the neutral axis or, abl 3.33"
2.91 x 41 60 -- 3930 jj:/ in. 2
3.09 6"
-N.A.
a
b
hence
SOLUTION USING BENDING STRESS FORffiJLA.
=-l1L
IN.A. I N•A. = 27.2 in. 4
J 2.73"
1~
2
I, 22.3x1.5=33.4"
1
16
-\
ab (upper fiber) = - 36670 x 3.09 ;:: - 4160 psi.

Fig. A13.14 Fig.A13.15 Fig. A13.16
27.2 Transformed Section
into Equivalent Spruce
which checks the above solution.
ent materials is: -
A13.9 Bending Stresses in Beams with Non-Homogeneous
Sections, Stresses within the Elastic Ranges. Espruce = 1,300,000 psi.
In general, beams are usually made of one = 1,600,000 psi.
material, but special cases often arise where Emaple
two different materials are used to form a beam = 29,000,000 psi.
section. For example, a commerical airplane in Esteel
its lifetime often undergoes a number of changes
or modifications such as the addition of ad- SOLUTION:
ditional installations, fixed equipment, larger
engines, etc. This increase in airplane weight The first step in the solution is to
increases the structural stresses and it often transform the reinforced beam section of Fig.
becomes necessary to strengthen various struct- A13.15 into an eqUivalent beam section composed
ural members at the critical stress points. of the same material throughout. This is possi-
Since space limitations are usually critical, it ble, because the modulus of elasticity of each
often necessitates that the reinforcing material material gives us the measure of stiffness for
be stiffer than the original material. Years that material. In this solution the reinforced
ago, when spruce wood was a common material for beam will be transformed into a spruce beam
wing beams, it was normal practice to use rein- section as illustrated in Fig. AI3.16.
forcements at critical stress points of stiffer
wood material such as maple in order to cut down Emaple 1,600,000
the size of the reinforcements. In aluminum Espruce
= 1,300,000 =1.23
alloy beams, the use of heat-treated alloy steel
reinforcements are often used because steel is 29,000,000
Esteel
29000,000/10500,000 or 2.76 times stiffer than = 1,300,000
= 22.3
aluminum alloy. Espruce
Thus to transform the maple reinforcing strips

SOLUTION BY MEANS OF TRANSFORMED in Fig. A13.15 into spruce we increase the
BEA1'1 SECTION. width of each strip to 1.23 x 0.5 = 0.615 inches
as shown in Fig. AI3.16. Likewise, to trans-
To illustrate how the stresses in a com- form the steel reinforcing strip into spruce, we
posite beam can be determined, two example prob- make the width equal to 22.3 x 1.5 = 33.4 inches
lems will be presented. as shown in Fig. A13.16.
This transformed equivalent section is now
Example Problem 5. handled like any homogeneous beam section Which
Fig. A13.14 shows the original beam section is stressed within the elastic limit of the
which is entirely spruce wood. Fig. A13.15 materials. The usual calculation would locate
shows the reinforced or modified beam section. the neutral axis as shown and the moment of in-
Two ears of maple wood have been added to the ertia of the transformed section about the
upper part as shown and a steel strap has been neutral axis would give a value of 51.30 in .•
added to the bottom face of the beam. The prob-
lem is to find the stress at the top and bottom Bending stress at upper edge of beam section: -
points on the beam section when the beam section
is SUbjected to an external bending moment of - 60000 x 3.33
--=-=--==-- = - 3900 psi.
60000 in. Ibs. about a horizontal axis which 51.30
causes compression in the upper beam portion.
The modulus of elasticity (E) for the 3 differ- Referring to Fig. A13.15, this stress would be
the stress in the spruce section. Since the re- alloy. The original area of 0.1 sq. in. each
inforcing strips are maple, the stress at the have been multiplied by these stiffness ratio
top edge of these maple strips would be 1.23 values.
times (- 3900) =-
4880 psi. 0.446 0.1615
The bending stress at the lower edge of the ·a-------:cb...
transformed beam section of Fig. A13.16 would be: T
°bt
=_ 60000 x (- 2.73) = 3200
51. 30
thus equals 22.3 times 3200 71500 psi. =

i
ps .
The stress in the steel reinforcing strap

t-
'.65 16t
'=2.3~ ~.:35 ~
I
Fig. A13. 18
Transformed Beam
Section into Magnesium
Since all these stresses are below the y = 5.35"
Alloy.
elastic limit stress for the 3 materials the
beam bending stress formula as used is
applicable. ~ c
0.10
d
e-=-----=e 0.431
Example Problem 6. The solution for the beam section of Fig.

Fig. A13.17 shows an unsymmetrical beam A13.18 is the same as for any other unsym-
section composed of four stringers, a,b,c and metrical homogeneous beam section.
d of equal area each and connected by a thin The first step is to locate the centroid
web. The web will be neglected in this example of this section and determine the moments of
problem. Each of the stringers is made from inertia of this section about centroidal X and
different material as indicated on Fig. A13.17. Y axes.
The beam section is subjected to the bending
moment Mx and My as indicated. Let it be re- -y _
- "YA
~ Ay = .446 x 10 + 0.1615 x 10 =5.35"
quired to determine the stress and total load on 1.1385
each stringer in resisting these applied extern-
al bending moments.
= 0.1615 x 6 + .431 x 4
1.1385 = 2.365"
------j
\ - - 6"
.. ..
r
10"
Steel ~10 Sta~nl~::
Steel
x-~
Y
tI My = 5000"#
Mx = 10000"#
Ix = 0.6075 x 4.65 + 0.531 x 5.35
Iy = 0.546
..
x 2.365 + 0.165 x 3.635 + 0.431 x
1.635
..
= 6.34 in.·.
..
= 28.27 in.·
11 Magnesium
c .1 Alloy _.10
4" -----it!
Alum. Alloy
Area of Each
Stringer =.1 sq. in.
I xy =0.446 (- 2.365)(4.65)
0.1615 x 3.635 x 4.65
0.10 (- 2.365)(- 5.35)
0.431 x 1.635 (- 5.35)
- - 4.90
= 2.73
= 1.27
- - 3.77
TOTAL - - 4.67 in.· = I xy
Fig. A13.17
SOLUTION; The bending stresses will be calculated
by using method 3 of Art. A13.5.
Since the 4 stringers are made of different
materials we will transform all the materials From Equation (14) Art. A13.5 --
into an equivalent beam section with all 4
stringers being magnesium alloy.
Esteel 29 K = I xy = - 4.67 =
Emag. = 6,500,000
Emag.
= 6.5 = 4.46 1 2
I x I Y - 1xy 28.27 x 6.34 - 4.67 2
Esteel = 29,000,000 28
ES . S . - 4.67 = _ 0 0296
Estain.steel= 28,000,000 Emag. = 6.5 = 4.31 157.4 •
Ealum.alloy = 10,500,000 = IxI/: I xy = l~~~ = 0.0403

Ealum •
Emag. -
10.5
= 6.5 = 1.615
K..
Using the ratio of stiffness values as indicated

28.27
157.4
= 0 •1797
above gives the transformed beam section of Fig.
A13.18 where all material is now magnesium 10,000 in. lb. My • 5000 in. lb.
A13.13
Consider Stringer (c): - to see if they equal zero.

x =- 2.365, y =- 5.35 ~ P =6 + 1167 - 162 - 1010 =1 check.
Substituting in 0b equation Forthermore the moments of the stringer loads
about the X and Y axes must equal the applied
0c =- [0.1797 x 5000 - (- 0.0296)(- 100 00 )Jx external bending moments.
Take moments about a Y axis through
stringers (a) and (c).
- [(0.0403)(- 100,00 ) - (- .0296 x 5000 )J y
~ My =- 6 x 162 - 4 x 1010 + 5000
(898 - 296 ) x - [(- 403 ) + 148 JY =- 12 in. lb.
- 602 x + 255· Y - - - - - - - - (18) Take moments about X axis through stringers

(c) and (d),
For stringer (c) x = - 2.365, y = - 5.35 ~ Mx = 1162 x 10 - 162 x 10 - 10000
hence
=0 in. lb.
0c =- 60Z (- 2.365) + 255 (- 5.35)
(The calculations in this example being
= 1423 1360 = 63 psi. done on a slide rule can not provide exact
checks) •
The total load Pc in stringer (c) thus equals
0c (Area) = 63 x 0.1 = 6 Ibs. tension. A13.10 Bending stresses of Homogeneous Beams
stressed above 'the Elastic Limit stress Range.
Stringer (a): In structural airplane design, the applied
loads on the airplane must be taken by the
x =- 2.365, Y = 4.65 structure Without suffering permanent strain
which means the stresses should fall within the
from equation (18) elastic range. The airplane structural design
loads which in general equal the applied loads
0a - - 602 (- 2.365) + 255 x 4.65 = 2613 psi. times a factor of safety of 1.5 must be taken
by the structure without collapse or rupture
Pa = 2613 x 0.446 = 1167_ lb. tension. With no restriction on permanent strain. Many
airplane structural beams will not fail until
the stresses are considerably above the elastic
Since the true area of this stringer is 0.1 stress range for the beam material.
square the stress in this steel stringer equals Since the stress-strain relationship in
1167 /0.1 = 11670 psi. tension. the inelastic range is not linear _and also
since the stress-strain curve for a material in
Stringer (b): the inelastic range is not the same under ten-
sile and compressive stresses (See Fig. A13.19),
x = 3.635, Y = 4.65 the beam bending stress formulas as preViously
derived do not apply since they were based on a
0b - 602 x 3.635 + 255 x 4.65 =- 1006 psi. linear variation of stress to strain. Experi-
mental tests however, have shown that even when
Pb = - 1006 x 0.1615 = - 162 Ibs. stressed in the inelastic range, that plane
sections before bending remain vplane after
True stress in stainless steel stringer after bending, thus strain deformation is still
=- 162 /0.1 =- 1628 psi. linear which fact simplifies the problem since
the stress corresponding to a given strain can
Stringer (d): be found from a stress-strain curve for the
beam material.
x = 1.635, Y =- 5.35 A general method of approach to solving
beams that are stressed above the elastic range
ad =- 602 x 1.635 + 255 (- 5.35) =- 2345 psi. can best be explained by the solution of a
problem.
Pd =- 2345 x 0.431 1010 Ibs.
Example Problem 7.
Portion (a) of Fig. A13.20 shows a solid
True stress =- 0.1
1010 =- 10100 psi. round bar made from 24ST aluminum alloy ma-
terial. Fig. A13.19 shows a stress-strain
To check the results, check total stringer loads curve for this material. Let it be assumed
that the maximum failing compressive stress A13.19 shows that for a given strain in the in-
occurs at a strain of 0.01 in. per inch. The elastic range the resulting tensile stress is
problem is to determine the ultimate resisting higher than the resulting compressive stress,
moment developed by this round bar and then and furthermore from internal equilibrium the
compare the result with that obtained by using total tensile stress on the cross-section must
the beam bending stress formula based on linear equal the total compressive stress.
variation of stress to strain. The problem as stated assumed that a
compressive unit strain of 0.01 caused failure.
Fig. b thus shows the strain picture on the
beam just before failure since plane sections
remain plane after bending in the inelastic
range. Table A13.2 gives the detailed calcu-
lations for determining the internal resisting
moment developed under the given strain con-
dition.
TABLE A13.2
1 2 3 4 5 6 7
Strip Strip Unit Res.

No. Area y E Stress F = crA Moment
"A" cr M = Fr
1 .058 0.935 .00867 53000 3075 2760

2 .102 0.840 .00773 52500 5350 4300
3 .135 0.75 .00685 52100 7025 5040
4 .153 0.65 .00591 51500 7870 4820
5 .165 0.55 .00494 51000 8410 4310
6 .180 0.45 .00398 43000 7740 3200
7 .185 0.35 .00302 33200 6140 1920
8 .195 0.25 .00205 22800 4450 945
9 .197 0.15 .00108 12500 2460 280
10 .200 0.05 .00012 3200 640 10
11 .200 -0.05 -.00084 - 7250 -1450 130
12 .197 -0.15 -.00181 -17800 -3510 660
13 .195 -0.25 -.00276 -29500 -5750 1650
14 .185 -0.35 -.00374 -35500 -6560 2540
15 .180 -0.45 -.00470 -40000 -7200 3510
(a) Fig. A13.20 16 .165 -0.55 -.00566 -43000 -7100 4170
TENSION TeotSlLf 17 .153 -0.65 -.00663 -44800 -6850 4710
STRESS
18 .135 -0.75 -.00759 -46000 -6210 4880
19 .102 -0.84 -.00846 -47200 -4810 4210
20 .058 -0.935 -.00937 -48000 -2780 2690
Total 3.140 740 56735
Col. 1 Rod divided into 20 strips . 1" thick.

Y = distance from centerline to strip c. g.
1/
IZ Col. 3
• Col. 4 E = strain at midpoint = (y - .0375)/103.75
Col. 5 Unit stress for 6 strain from Fig. A13.19
Col. 6 Total stress on strip.
Col. 7 Moment about neutral axis. M = (y - .0375)F.
.010· COIWlPIilE~I"E
COMPR £6SION UNIT
:iTll.AII\l STltli5S
The summation of column (6) shculd be
SOLUTION: zero. Since the discrepancy is 740 Ibs., it
means that the assumed position for the neutral
Since the stress-strain diagram in tension axis is a little too high, however the dis-
is different from that in compression (See Fig. crepancy is negligible. The total internal
A13.19) the neutral axis will not coincide with resisting moment is 56735 in. Ibs. (Col. 7).
the centroidal axis of the round bar regardless If we take a maximum unit compressive
of the fact that it is a symmetrical shape. strain of 0.01 we find the corresponding stress
Thus the method of solution is a trial and error from Fig.A13.19 to be 48500 psi. If this
one since the location of the neutral axis can stress is used as the failing stress in the
not be solved for directly. In Fig. b of Fig. beam formula M = ~ we obtain,
A13.20 the neutral axis has been assumed 0.0375 c
inches above the centerline axis of the bar. It
was assumed toward the tension side because ob- M = 48500 x 0.785 = 38000
in. Ibs.
servation of the stress-strain curve of Fig. (0.785 = ~ of round bar = nr 3 /4)
A13.l5
Thus using the same failing stress at the given in most books on advanced engineering
far extreme fiber the beam formula based on mechanics and will not be repeated here.
linear stress-strain relationship gives an ul- It is convenient however to express the
timate bending strength of 38000 as compared to influence of the beam curvature in the form of
a true strength of 56735 or only 67 percent as a correction factor K by which the stresses ob-
much. tained by the beam formula for straight beams
Fig. c of Fig. A13.10 shows the true stress can be multiplied to obtain the true stresses
distribution on the cross-section, which ex- for the curved beams. Thus for a curved beam
plains why the resisting moment is higher than the maximum stress can be calculated from the
when a triangular distribution is used. equation
The problem of the ultimate bending
strength of structural shapes is discussed in cr =K ~
I
- - - - - (19)
detail in Volume II.
A13.11 Curved Beams. Stresses Within the Elastic Table A13.3 gives the value of K for
Range. various beam section shapes and beam radius of
The equations derived in the previous art- curvatures. The table shows that for only
icles of this chapter were for beams that were rather sharp curvatures is the correction ap-
straight. Thus in Fig. A13.21, the element of preciable. In general for airplane fuselage
length (L) used in the derivation was constant rings on frames the curvature influence can be
over the depth of the beam. The strain (6L/L) neglected. However there are often fittings
was therefore directly proportional to 6L which and mechanical structural units in airplane
had a linear variation. construction whose parts involve enough curva-
In a curved beam, the assumption that plane ture to make the influence on the stress of
sections remain plane after bending still primary importance. The concentration of
applies, however the beam segment of a curved stress on the insirle edge of a curved unit in
beam cannot have equal width over the depth of bending may influence the fatigue strength of
the beam because of the curvature as illustrated unit considerably, thus a consideration of the
in Fig. A13.22, or in other words the length of possible influence of curvature should be a
the segment is greater on the outside edge (L ) regular part of design procedure.
than on the inside edge (Li). Thus in calcu- In the inelastic or plastic stress range,
the influence of beam curvature should be
<1L-1 ~ considerably less since the stiffness of a
material in the inelastic range is much less
MIl
rl
Straight Beam
1M
1 Fig.A13.2l
than in the elastic stress range and changes
rather slowly as the stress increases.
A13.l2 Problems.
(1) Fig. A13.23 shows the cross-section of a

single cell beam with 12 stringers.
Assume the walls and webs are ineffective
in bending. Calculate load in each
stringer by use of beam formula. Also
calculate stringer loads by equating in-
ternal resisting moment to external bend-
Stress ing moment. Each stringer area is same
Fig. A13. 22 Distribution and equals 0.1 sq. in. applied bending
moment Mx =- 100,000 in. lb.
lating the strain distribution over the beam

depth the change in length 6L at a point must be
divided by the segment width at that point.
Thus even though plane sections remain plane
the strain distribution over the section will
not be linear. The width of the segment at any
point is directly proportional to the radius of
k
x-- .-=-----e.. j ~,,--_ _=h__ t=x
curvature of the segment and thus the strain at Fig. A13. 23
a point on the segment is inversely to the
radius of curvature. This gives a hyperbolic (2) Same as problem (1) but change external
type of strain distribution as illustrated in
Fig. A13.22, and if the strains are within the
bending moment to My =
200,000 in. lb.
elastic limit the stress distribution will be (3) Fig. A13.24 shows a beam section with 4
similar. The development of a beam formula stringers. Assume web and walls in-
based on a hyperbolic stress distribution is effective in bending. Stringer areas
TABLE A13.3 TABLE A13.3 (continued)

VALUES OF K FOR USE IN THE VALUES OF K FOR USE IN THE
BEAM FORMULA a = K ~C BEAM FORMULA a = K ~c
R FACTOR K e* FACTOR K
SECTION INSIDE OUTSIDE SECTION R
INSIDE OUTSIDE e*
c R C R
FillER FillER FillER FIBER
~R~ 1.2 3.41 0.54 0.224 1.2 3.55 0.67 0.409

1.4 2.40 0.60 0.151 1.4 2.48 0.72 0.292
C)'
I~j I
1.6
1.8
2.0
3.0
1. 96
1. 75
1. 62
1. 33
0.65
0.68
0.71
0.79
0.108
0.084
0.069
0.030
1.6
1.8
2.0
3.0
2.07
1. 83
1. 69
1. 38
0.76
0.78
0.80
0.86
0.224
0.178
0.144
0.867
~I
4.0 1. 23 0.84 0.016 4.0 1. 26 0.89 0.038
6.0 1.14 0.89 0.0070 6.0 1.15 0.92 0.018
8.0 L10 0.91 0.0039 8.0 1.10 0.94 0.010
10.0 1. 08 0.93 0.0025 10.0 1. 08 0.95 o 0065
1.2 2.89 0.57 0.305 1.2 2.52 0.67 0.408
dJl
1.4 2.13 0.63 0.204 1.4 1. 90 0.71 0.285
1.6
1.8
1. 79
1.63
0.67
0.70
0.149
0.112
r-
-It 1-4t -it I 1.6
, 1.8
1. 63
1. 50
0.75
0.77
0.208
0.160
: I 2.0
3.0
1. 52
1. 30
0.73
0.81
0.090
0.041 i rY±:l I
2.0
3.0
1. 41
1. 23
0.79
0.86
0.127
0.058
f.-- R-j 4.0
6.0
1. 20
1.12
0.85
0.90
0.021
0.0093
UT ~1-1 4.0
6.0
1.16
1.10
0.89
0.92
0.030
0.013
8.0 1.09 0.92 0.0052 8.0 1.07 0.94 0.0076
10.0 1.07 0.94 0.0033 10.0 1. 05 0.95 0.0048
1.2 3.01 0.54 0.336 1.2 2.37 0.73 0.453
rb~ 1.4 2.18 0.60 0.229 1.4 1. 79 0.77 0.319
1.6 1. 87 0.65 0.168 1.6 1. 56 0.79 0.236
T 1 I 1.8 1. 69 0.68 0.128 1.8 1. 44 0.81 0.183
b I 2b I 2.0 1. 58
1. 33
0.71
0.80
0.102
0.046
2.0
3.0
1. 36
1.19
0.83
0.88
0.147
0.067
..1.. 3.0
I
c~-J
1I 4.0
6.0
8.0
1. 23
1.13
1.10
0.84
0.88
0.91
0.024
0.011
0.0060
4.0
6.0
8.0
1.13
1. 08
1. 06
0.91
0.94
0.95
0.036
0.016
0.0089
10.0 1.08 0.93 0.0039 10.0 1. 05 0.96 o 005'1
1.2 3.09 0.56 0.336 1.2 3.28 0.58 0.269
r- --I , 1.4 2.25 0.62 0.229 1.4 2.31 0.64 0.182
-lCJT
3b 1.6 1. 91 0.66 0.168 1.6 1. 89 0.68 0.134
1.8 1. 73 0.70 0.128 1.8 1. 70 0.71 0.104
I
2.0 1. 61 0.73 0.102 2.0 1. 57 0.73 0.083
4- Ii' 3.0 1. 37 0.81 0.046 3.0 1. 31 0.81 0.038

4.0 1. 26 0.86 0.024 4.0 1. 21 0.85 0.020
r- R --.j
..j c to- 6.0 1.17 0.91 0.001 6.0 1.13 0.90 0.0087
8.0 1.13 0.94 0.0060 8.0 1.10 0.92 0.0049
10.0 1.11 0.95 0.0039 10.0 1. 07 0.93 o 0031
1.2 3.14 0.52 0.352
r-- I 1.2 3.55 0.67 0.409
i8:Jr.
1.4 2.29 0.54 0.243 4t -j 1.4 2.48 0.72 0.292
1.6
1.8
2.0
1. 93
1. 74
1. 61
0.62
0.65
0.68
0.179
0.138
0.110
1tQ}t2
4t t t l 2.0
I
1.6
1.8
2.07
1. 83
1. 69
0.76
0.78
0.80
0.224
0.178
0.144
~11
c r-
! 3.0
4.0
1. 34
1. 24
0.76
0.82
0.050
0.028 1 t/2 I
3.0
4.0
1. 38
1. 26
0.86
0.89
0.067
0.038
6.0 1.15 0.87 0.012 I-c-...j I 8.0
6.0 1.15 0.92 0.018
R~ 8.0
10.0
1.12
1.12
0.91
0.93
0.0060
0.0039 J-- ----J
R 10.0
1.10
1. 08
0.94
0.95
0.010
0.0065
1.2 3.63 0.58 0.418
1 3t
i 42Y2
r- I
1.4
1.6
2.54
2.14
0.63
0.67
0.299
0.229
* e equals distance from centroidal axis to neutral axis.
~~TI
4t I
-l I
1.8
2.0
1. 89
1. 73
0.70
0.72
0.183
0.149
References: Wilson and Quereau. "A Simple Method of
Determining Stresses in Curved Flexural
3.0 1. 41 0.79 0.069 Members, "
-Jcf- I 4.0 1. 29 0.83 0.040 "Advanced Strength of Materials", by Seely.
f.--
R--I 6.0
8.0
1.18
1.13
0.88
0.91
0.018
0.010
10.0 1.10 0.92 0.0065
A13.17
shown in () on figure. External applied (7) Fig. A13.27 shows 3 different beam sections.
bending moments are; They are made of aluminum alloy Whose
stress-strain diagram is the same as that
Mx= -500,000 in. lb. and My 200,000 in. lb.= plotted in Fig. A13.19. Determine the
Find stress on all four stringers by all ultimate internal resisting moment if the
three methods which were explained in this maximum compressive strain is limited to
chapter. 0.01 in./in. Consider that upper portion
is in compression. Compare the results
a
1.2)
obtained with formula M = 0bI/c' where 0b =
compressive stress when unit strain is
(0.5)
1:5 ---r- 0.01.
6"
(1.0) (0.4)-.L
j--G"~12" --!j Fig.A13.24
r- 1" -l ~1/21
10 01
f4
(4) The Zee section shown in Fig. A13.25 is
subjected to bending moments of Mx = 500 1 11/4"
in. lb. and My =
2000 in. lb. Find bending I-- 4
stresses at points a,b,c,d.
1 Fig.A13.27
1"-1
k4
~~4~
_.1.
3/4
L x ~L16 c r3/4 Fig. A13. 25 500# 500#

d -L
2"
--11/:j c
(5) Fig. A13.26 shows a beam section composed

of three different materials. Find the Fig. A13 .. 28
stress at top
and bottom
points on
beam due to a Alum. 0.1" 0.1"
bending moment Alloy ~ !-O.4'-!..j I-
= r:1"~
~
Mx BO,OOO in.lb. T
EMag.= 6.5xl0 6
ESteel= 29 x 10
6
@Sec. 2
0.1"
Sec. 3
L
0.6"
EAlum. = 10.5 x 10 6 Fig. A13. 26
(6 ) Fig. A13.26 shows (B) Fig. A13.2B shows a curved beam, carrying
a cross-section of a two 500 lb. loads. Find bending stresses
wood beam composed at points C and Cn , when beam cross-section
of 3 kinds of wood is made 3 different ways is indicated by
labeled A, Band C, sections 1, 2 and 3. Use Table A13.3.
glued together to
form a composite x- - - x 6"
beam. If the beam A
~1
is sUbjected to a
bending moment
Mx = 75000 in.lb.,
find intensity of 3/4
l--J'-----"-_'-----' -*'---
bending stress at Fig. A13. 26
top edge of beam.
Also find total end load on portions B
and C.
=
A Spruce. E =1,300,000 psi
=
B Maple. E = 1,500,000 psi
C = Fir. E = 1,600,000 psi
DOUGLAS DC-8 AIRPLANE. Over-all view of the test wing section representing center wing section of DC-8.
A close-up view of wing test section showing details of wing ribs, stringers, etc.
CHAPTER A 14
BENDING SHEAR STRESSES· SOLID AND OPEN SECTIONS
SHEAR CENTER
A14.1 Introduction.
In Chapter A6, the shear stresses in a Fig. A14.2
Load
I
--Idxl-
member subjected to pure torsional forces were • D' F'
considered in detail. In Chapter A13, the sub-
ject of bending stresses in a beam subjected to _ I yN'A~xr=_=_=='=c-~l ±:_~~=l_x =-:
pure bending was considered in considerable de-

---l!' I I
dA~ C ~:C-)
tail. In practical structures however, it I I
seldom happens that pure bending forces (coup- t

les) are the loading forces on the beam. The
Beam Section
Fig. A14.1
D F t
usual case is that bending moments on a beam ac ldxr-- a
n'
I
are due to a transfer of external shear forces. ~
~o.~A:~r~·:J
Thus bending of a beam usually involves both
bending (longitudinal tension and compression
stresses) and shear stresses.
The same assumptions that were made in
Chapter A13 in deriving the bending stress
equations are likewise used in deriving the
N.A.
" D F " ,
atf
c Yo L[ RJ J
'y
! D
o. F\.
C Yo
equations for flexural shear stresses. With at at at at

flexural shear stresses existing, the assumption Fig. A14.4
Fig. A14.3
that plane sections remain plane after bending
is not completely true, since the shearing is as indicated. The stress at 1s greater than
strains cause the beam sections to slightly at because the bending moment due to the given
warp out of their plane when the beam bends. beam loading is greater at beam section DD' than
This warping action is usually referred to by at FF'. Now consider that this beam portion dx
the term "shear lag". However, except in cases is further cut as indicated by the notch DCEF in
of beams with wide thin flanges, the error in- Fig. A14.1, and this segment is shown in Fig.
troduced by neglecting shearing strains is quite A14.4 as a free body with the forces as indicated.
small ffild therefore neglected in deriving the Let at = maximum tensile stress at a distance
basic flexural shear stress formula. The prob- c from the neutral aXis,
lem of shear lag influence is considered in Then the stress at a distance y from neutral
other C'hapters. axis is a y = at y/c.
The total load on an element of area dA of
A14.2 Shear Center. the beam cross-section (see Fig. A14.2) thus
When a beam bends without tWisting, due to
some external load system, shearing stresses are equals at J y dA .
c
set up on the cross sections of the beam. The Now, referring to Fig. A14.4, the total
centroid of this internal shear force system is tensile load on each face of this segment will be
often referred to as the shear center for the calculated.
particular section. The resultant external
shear load at this section must pass through
the shear center of the section if twist of the Total load on face CD = act IC y dA (1)
section is to be prevented. Thus, if the shear Yo
center is known, it is possible to represent the
external load influences by two systems, one
that causes flexure and the other which causes
only twist.
Total load on face FE = at
c
,f Yo
c y dA (2)
AB.3 Derivation of Formula for Flexural Shear Stress. From Chapter A13, the equation for flexural
Fig. A14.1 shows a loaded simply supported stress a was derived, namely at = Mc/I. Let M
beam. When the beam bends downward due to the equal the bending moment at beam section DD' and
given loading, the beam portion above the neu- M' that at beam section FF' and let I and l'the
tral axis is placed in compression and that be- moment of the inertia of the cross-sectional area
low the neutral axis in tension. Consider a about the neutral axis at these same beam sections
short portion dx of the beam at points DF on the respectively. Then SUbstituting value of at in
beam and treat it as a free body as shown in equations (1) and (2) we can write,
Fig. A14.3. The variation of tensile and com-
pressive stress on each face of the beam portion Total load on face CD =~JCy dA - - - - - (3)
Yo
A14.1
A14 2 BENDING SHEAR STRESSES SOUND AND OPEN SECTIONS. SHEAR CENTER.
Total load on face FE =~: JC y dA - - - - (4)

Yo
Now let ~ b dx equal the shearing force on
face CE of the segment in Fig. A14.4 where ~
equals the shearing stress and b dx the shearing
area. For equilibrium of the segment, the total
forces parallel to x-x axis must be zero. If
the beam is of uniform cross-section, which is
the case in our problem, then I = I' and c and
Yo are the same in both equations (3) and (4).
Then the resultant horizontal force on the
sides of the segment equals the difference of
the values in (3) and (4), or Fig. A14.5
Resultant horizontal load =M

t
-I M Jc ydA. neutral axis are given on the figure.
Yo SOLUTION: -
For equilibrium of segment in x-x direction, Shear stress at neutral axis x-x
2: Fx =- -M-M'
1-- JC ydA + ~ b dx =0 V J3.09
yg ~x-x = -.:.L y dA
Ix b
hence ~ = MI b- M'
dx
Jyc dA - - - - - - - - - - (5)
o
Table A shows the calculation of the term

o
M-M' dM 3.09
However ~ = dx = V = the external shear on y dA.
the beam section.
1o
hence ~ V JC ydA
=r-b (6) TABLE A. REFER TO FIG. a.
Yo
PORTION Area d A y ydA
It is important to note that equation (6) 1 2.09 x 0.5 = 1.045 1. 045 1. 096
applies only to beams of uniform section (con- 2 3 x 0.5 = 1. 50 2.34 3.510
stant moment of inertia). In airplane wing 3 0.5 x 0.5 = 0.25 2.84 0.710
structures the common case is for beams to vary 3' 0.5 x 0.5 = 0.25 2.84 0.710
in cross-section or moment of inertia, and if SUM 6.026
this variation is considerable, equation (6)
should not be used and resort should be made to
~
~~___j:~~~( 2)
b -~
equations (3) and (4). This fact is illustrated (2)
2
in example problem 2. This matter of variable -I -j i, I
I
cross-sections is discussed later in this I I
chapter. Fig. a : Fig. b I I ,III Fig. c
'
x- ~x x-_U i[JY : : -.1- -

X-_L - 1-
Y
~
A14.4 Example Problems. Symmetrical Sections.
External Shear Loads Act Thru Shear Center. hence ~ =850 x 6.026 = 377
27.2 x 0.5
i
ps •
Example Problem 1.
Fig. A14.5 shows the cross section of a Calculation of Shear stress at point 1-1, : -
beam symmetrical about the Y-Y axis. Assume
that a beam with this cross-section is SUbjected
to a loading which produces a shear load in the ~l-l = IxYb
V J3.09Y dA •
Y direction = to Vy = 850 lb. Its location is
through the shear center of the section which 2.09
lies on the centroidal y axis of the beam sec- Fig. b shows the effective areas in this
tion due to the symmetry of the section about intregration, thus in Table A, we leave out
this axis. Let it be required to determine the portion 1, hence I y dA = 4.93.
shearing stress at the neutral axis x-x and at
points 1-1 and 2-2 of the cross-section as 850 x 4.93
shown in Fig. A14.5. The neutral axis location SUbstituting, ~l-l = 27.2 x 0.5 = 308 psi.
and moment of inertia of the section about the
A14.3
This shearing stress is at a section just SOLUTION: -

adjacent to portion (2). If it was taken just
adjacent to portion (1), then the width b in the Bending Moments: -
equation would be 3 inches instead of 0.5 inch M132 = -600 x 132 = - 79,200"#
and the shearing stress would be 0.5/3 times that
shown above, or in other words the shearing M175 = -600 x 175 = -105,000"#
stress changes abruptly when the shear area M218 = -600 x 218 = -130,800"#
changes abruptly.
Table A14.1 shows the results of calculating the
Shearing stress at point 2-2 on cross-section: - bending stresses at 3 points on each side of the
Fig. c shows effective area, thus in Table neutral axis for the 3 stations. For example
A, the portions (1) and (2) are omitted and for station 132
3.09
± C9;~~6~ 4)
y dA equals 1.42.
J,
2.59 0MAX. ± (M/) =
± 8180 psi. = =
850 x 1.42 (tension at top edge and compression at lower
SUbstituting, ~2-2 = 27.2 x 1.0 = 44.4 psi. edges) .
(The width b = 0.5 + 0.5 for the two portions For a point 1 inch from either edge of the
3, 3' ). beam
The shearing stresses as calculated act in the
plane of the beam cross-section in the Y di-
rection and also with the same intensity parallel
to the Z axis which is normal to the beam TAl>LE A14.1
section.
1 2 3 4 5 6 7 8
Example Problem 2. VARIABLE MOMENT OF INERTIA. Bend. Stress 0b = My/1 Total Bending Stress
Load on: -
Fig. A14.6 shows a cantilever beam loaded On Top Point Point
with a single load of 600 lb. at the end and 801 § Bending or 1" From 2" From
acting through the centroid of the beam cross- .~
"'~
Moment Bottom Top or Top or Portion Portion Portion
"A" "B" "C"
section. The beam section is constant between o:l~ M F'iber Bottom Bottom
Y = 4" Y = 3" y = 2"
stations 0 and 132, then it tapers uniformly to
132 79,200"# +8180 +6135 +4090 7157# 5112# 1023#
the sections shown for stations 175 and 218. .6135 ,4090 7157 10224 1023
175 105,000"# .8180
The shear stress distribution on the beam cross- 218 130,800"# .8180 .6135 ,4090 7157 15336 1023
section at stations 175 and 218 will be deter-
mined.
600#
~43"-----.l...-43"'-'+1~-- 132"------+1
From the results in Table A14.1, it should be
Fixed 1- -----r-- noticed that the change in moment of inertia be-
tween the three stations is directly porportional
to the change in bending moment, hence the same
Sta. '218 Sta: 175 sci. 132 Sta.O value for the bending stresses for all three sta-
tions. Columns 6, 7 and 8 give the total bending
1r2~~
stress load on portions A, Band C of the three
cross-sections (see Fig. A14.6). These values
___ L equal the average stress on the portions times
1" the area of the portion; for example, for station
132, the load on portion A:
c
------ ----
load = 8180 + 6135 x 1 x 1 7157 #, and for
2
portion C:
Sta. 218 Sta. 175 Sta. 132 load = 4090 + 0 x 2 x .25 - 1023#
1= 64.0 in. 4 I = 51. 33 in. 4 I= 38.67 in. 4 2
Fig. A14.7 shows the tension and compressive
Fig. A14-6 stresses acting on a portion of the beam between
A14.4 BENDING SHEAR STRESSES. SOUND AND OPEN SECTIONS. SHEAR CENTER.
stations 175 and 218. Fig. A14.8 shows the re- shear stresses between stations 132 and 175
sulting horizontal shear stress pattern result- would be the same, since the change in bending
ing from the loads in Fig. A14.7. For example, moment and moment of inertia have been made the
i f we take a section along the beam 1" from the same as between stations 175 and 218.
top or bottom edge of the beam and treat this Figs. A14.9 and A14.l0 show the shear
V ~~
portion as a free body as shown in Fig. A14.l2
applying ZH = 0, stress patterns if the formula 'L = dA be
used for each station. The discrepancy is con-
ZH =- 7157 + 7157 + 'L X 43 x 1.0 = 0, hence 'L = 0 siderable as the equation does not apply to
beams of varying section.
To illustrate the calculation by the shear
stress formula, the shear stress will be calcu-
lated at the neutral axis for the beam section
~1023# at station 175.
'L"'I\J>dA
where f~ y dA = 1 x 1 x 3.5 + 2 x 1 x 2.5 + 2 x

Fig. A14.7 0.25 x 1 ~ 9.0
600 x 9.0
hence, 'L = 51.33 x 0.25 = 420 psi. as compared
Fig. A14. 8 Fig. A14.10 Fig. A14.9
to the true shear stress of 477 in Fig. A14.8.
TABLE A14.2
MAXIMUM SHEAR STRESS FOR SIMPLE SECTIONS

477#/0"
Location of
C ross Section Max. Shear Stress Max. Shear Stress
3V e :: 0
True Sta 218 Sta 175 N-B-N 'L :: 2A
Shear Shear Stresses
Pattern By Equation (6)
--43"
1+--1' "
6 ---1r-1/ 4"
7157
15336
7157
10224
F s x 43 x. 25
N - 8 -N 'L
4V
= 3A e =0
Fig. A14.13
N-~-N 'L = 4V(

3A 1
Dd)
+ D'" + dlil e =0
Fig. A14.12 d
3V
Similarly, treating the portion between the 'L = 2A e = 6"
tN-8-t-N
edge of the beam and a point 2" from the edge as
a free body diagram as shown in Fig. A14.l3,
ZH =- 7157 + 7157 - 15336 + 10224 + 'L x 43 x eL()~

e T----
_.--!
-N
9V
'L :: 8A e ::
d
11
.25 = 0, therefore 'L =477 psi.
Obviously, the shear stress on portion C is con-
stant, since the end load on this portion at
both stations is the same, or 1023#.
Fig. A14.8 shows the general shape of the
shear stress distribution on the beam section at
N(j:" 'L
4V
= 3A e =0
any point between stations 175 and 218. The

A14.5
A14.5 Maximum Shear Stresses for Simple Cross-Sections. dy + (C y - CY) = 0, hence

Table A14.2 gives the value of the maximum
shear stress on a few simple sections and where
it occurs on the cross section, (e is distance q =- ( C'y - CY)
from neutral axis to point of maximum shear Yb dy
stress). V equals the shear load normal to the
neutral axis and it acts along the centerline
axis, thus no tWisting on the section. A is the But (C y - Cy) =Mx,Ix- Mx Jba z dA (See Art. AI4.3)
total cross-sectional area. The maximum shear
stress is given in terms of the average shear
stress which equals VIA. hence qy
b
=_dM!-Jb
dy Ix
ZdA
a
A14.6 Derivation of Flexural Shear Flow Equation.
Symmetrical Beam Section.
To emphasize further the fundamental re- =
however ~ Vz ' the external shear in the Z
lationships, a second derivation of the equation direction.
for shear stress distribution will be presented.
Fig. A14.14 shows a portion of a cantilever beam hence qy
carrying a load P at the free end as shown. b
VZJb Z
Ix a
dA - - - - - - - - - - (7)
This load is so located as to cause the beam to

bend in the YZ plane without twist about a Y Equation (7) gives the change in shear flow force
axis. The problem is to derive relationships qy between points (a) and (b) and since in figure
which will give the magnitude and sense of the
shear flow distribution on the cross-section of A14.15 the value of qy at (a) is zero because of
the beam. a free surface, the value of qYb in equation (7)
Fig. A14.15 shows a free body of a small is the true shear flow force in lbs. per inch at
portion of the beam cut out from the upper point (b). The student should realize that
flange of the beam at points (a b) in Fig. equation (7) gives the shear flow q in the Y di-
AI4.14. Under the given external load P it is rection. The minus sign in equation (7) means
obvious that the upper half of the beam is sub- that the positive sense as assumed by the arrow-
jected to compressive stresses. In Fig. AI4.15, head on qYb in Fig. A14.15 is incorrect or should
y
C is larger than Cy since the cantilever bend-
be reversed.
ing moment is greater at station Y' .
The initial problem was to determine the
shear flow force system in the plane of the beam
cross-section or the ZX plane. From elementary
engineering mechanics, we know that if a shearing
stress occurs on one plane at a point in a body,
a shearing stress of the same intensity exists on
planes at risht angles t a the first plane, or in
I.,L;-I~~ dy general at a point,
~::::;~~33j:jj::'=;1f-Sta.y
L:: sta.y'L qy = qx = qz - - - - - - - - - - - (8)
Sta.y' f-hl-l:---=-r;/
Before the shear flow in other planes is
x- completely defined its sense (positive or nega-
Free tive) must be known. Equation (7) gives the mag-
surface nitude and sense for qy at any desired point on
the cross-section. The question of the sense of
the associated qx and qz is easily determined from
an observation involving eqUilibrium of moments.
Fig. A14.14 Fig. A14.16 Fig.A14.15 This fact will be explained by referring to a
number of free body diagrams.
The free edge of beam flange forms the Fiv. A14.17 shows a free body of a small
right side face of the element in Fig. A14.15 element"cut from the beam in Fig. A14.14 at point
and thus the shear flow force on this face is (a) on the cross-section. The forces on this
zero as indicated. The shear flow forces on the free body are the compressive forces c P and C on
internal or cut faces in lbs. per inch are qy the front and rear faces and the shear forces on
and qx as indicated. Since the sense of these the various faces as indicated. The right side
shear forces is unknown, they will be assumed as face of the element is a free surface and thus q
acting in the positive direction. (See Fig. on this face is zero.
A14.16 for positive sense of forces acting par- The shear flow qy on the left side face is
allel to each of the coordinate axes XYZ. calculated from equation (7) namely
Now consider the equilibrium of forces in V
the Y direction for the element in Fig. A14.15 qy =- ~ L: Z A •
L:F y = 0, or Ix
A14.6 BENDING SHEAR STRESSES. SOUND AND OPEN SECTIONS. SHEAR CENTEH.
back side and through a point on the line of

action of C' and C, the shear flow qz on the
front face must act downward in order to balance
the moment due to qy.
Fig. A14.22 shows the results for a free
body of the lower flange plus a small web por-
tion. qy is positive from equation and thus qz
must be downward for moment equilibrium.
From the results obtained for these 6 dif-
ferent locations, a simple rule can be stated
relative to sense of shear flows in the plane of
the cross-section, namely: -
If the calculated shear flow is directed toward the
boundary line between the two intersecting planes of the par-
ticular free body, then the shear flow on the other plane is
also directed toward the common boundary line, and con-
versely directed away if the calculated shear flow is directed
away.
Fig.A14.22
T' Fig.A14.19 T' Fig. A14.l4 shows the sense of the shear
Fig.A14.20 flow pattern on the beam section as determined
for the given external loading.
A14.7 Shear Stresses and Shear Center for Beam Sections
For the given beam loading in Fig. A14.l4 the with One Axis of Symmetry.
load P is up, therefore Vz has a positive sign. Example Problem. CHANNEL SECTION.
For any portion of area (A) of the cross-section Fig. A14.23 shows a cantilever beam With
the distance z in the above equation is there- channel shaped cross-section carrying a 100 lb.
fore positive. Therefore in sUbstituting in the downward load as shown. The problem is to de-
above equation qy comes out negative for any termine the lateral position of this load so that
point above neutral axis, and likewise for any the beam will bend without tWist. This position
point on beam section below the neutral axis the will coincide with the lateral pOSition of the
distance z would have a negative sign and qy centroid of the shear flow system on the beam
would come out positive. Therefore the sense of cross-section'which holds the external load in
the calculated shear flow qy on the left side of equilibrium without tWisting of the beam section.
The cantilever beam has been cut at a section
the element in Fig. A14.l7 is negative or as abc d (Fig. A14.23) Which is far enough from the
indicated by the arrow on the force vector. Now fixed end of the beam (not shown) so that the
to find the sense of the shear flow qx on the effects of beam end restraint against section
front face take moments about a Z axis acting in warping can be neglected. In Fig. A14.23, the
the plane of the rear face and through point (0) internal forces holding the beam in eqUilibrium
which is on the line of action of the forces C' are sketched in. They consist of a longitudinal
and C. Only two forces have moments, namely the stress system of tension and compression and
side shear force qy dy and the front face shear variable shear flow system in the plane of the
force (qx/2)dX. It is obvious by observation cross-section. In this problem we are only con-
that qx on front face must act to the left as sidered with the internal resisting shear flow
shown if the moment is to equal zero. The total system.
shear force on front face is (qx/2)dx because For solution of this problem the moment of
shear flow at right edge is zero and it varies inertia Ix must be known. If calculated it
linearly to qx at left edge of front face. would be Ix = 0.2667 in 4 •
Figs. A14.l8, 19 and 20 show free bodies of
elements taken at the other three corners of the SOLUTION: - From equation (7)
beam section which are labeled c, d, e in Fig.
A14.l4. For all free surfaces q is zero. The
sense of qy as before is given by the equation qy = - ~~ ~ z A - - - - - - - - - - - - - - - (8)
as explained before, hence qy is negative in
Fig. A14.l8 and pOSitive in figures 19 and 20. we know that the shear stress is zero at a free
A simple consideration of moment eqUilibrium as edge, thus the solution of equation (8) is
explained for Fig. A14.l7 gives the sense of the started at either points (a) or (d).
shear flow Qx as shown in the 3 figures. The shear Vz is - 100 lb. Thus equation
Fig. A14.2l shows a free body of the entire (8) for our problem reduces to,
upper beam flange and a short portion of the
beam web. qy from the equation is negative and - 100
this acts as shown in the figure. Now if we 0.2667 ~ z A =375 ~ z A - - - - - - - (9)
take moments about a X axis in the plane of the
A14.7
100 lb. boundary line as shown in Fig. A14.24.

Common sense tells us that the resisting
Free end shear flow qz on the channel web must be directed
upward because it is the only force system that
can balance the 100 lb. load as far as ~Fz = 0
is concerned.
In general the shear flow is continuous
around the section and only reverses when it
passes through zero which only happens in closed
tubular sections. In general, it is possible in
most cases by observation only, to determine the
Shear Flow Diagram sense of the shear flow at some one point on the
Fig. A14. 25 beam cross-section. The shear flow being like a
flow of liquid will continue in the same general
direction along the center line of the parts that
make up the beam section.
The small arrows on the beam section of
Fig. A14.23 show the sense of the shear flow
pattern over the beam section. In Fig. A14.25,
Fig. A14.23 the shear flow values as calculated at the var-
ious points are plotted to form a shear flow
diagram for the beam section. Between points a
We will start at point (a) in solving equa- and b or d and c, the arm z in equation (9) is
tion (9) and proceed around the section. constant and thus q varies linearly as plotted.
Point (a) q
Ya
= 0 (free surface) Between b and 0 or a and c the arm z changes and
the area is also a function of z, thus q varies
Point (b) z = - 1, A = area between (a) and (b). as z" or parabolic as plotted.
The initial problem was to locate the
qy = qy + 375 ~ b z A centroid of this final shear flow system which
b a a
is generally referred to as the shear center.
q = 0 + 375 (- 1)(1 x 0.1) 37.51b/in. In Fig. A14.25, Qab, QbC and Qcd represent the
Yb resultant of the shear flow force system on
these three portions of the beam cross-section.
Point (0) on X axis. Each force is equal in magnitude to the area of
the shear flow diagram for the particular beam
qy = - 37.5 + 375 ~~ z A section portion. Hence,
a
Qab =1 x 37.5/2 = 18.75 lb.
= - 37.5 + 375 (- 0.5)(1 x 0.1) = - 56.3 lb./in.
Qbc =2 x 37.5+ (56.3-37.5) x2 x 2/3 = 100 lb.
Point (c).
Qcd = 1 x 37.5/2 = 18.75 lb.
qy = -56.3+375 (0.5)(1 x 0.1) =-37.51b./in.
c The resultant R of these three shear forces
will now be determined.
Point (d).
~ Fz = 100 lb., ~ Fx = 18.75 - 18.75 =a
qy =-37.5+375 (1) (1 x 0.1) = 0 (free surface)
d Hence
We know that the intensity of shear flow in the R = (~ Fz " + h Fx "J1/2 = (lOa" + 0)1/" = 100 lb.
ZX plane at any point equals that in the Y di-
rection at the same point. Thus qx or qz equal
The moment of the resultant about any point such
the qy values above. The sense of the qx and qz as (b) in Fig. A14.17 must equal the moment of
shear flows must be known before they are com- the shear flow force system about point (b).
pletely defined or known. Fig. A14.24 shows a Let e be distance from point b to line of action
free body of a small element at point (a) on the of the resultant R.
end of the lower beam flange. For any point be-
low the X centroidal axis equation (9) will give Hence R e =h M b (of shear flow system) or
a minus sign for qy. Thus in Fig. A14.24 qy
acts as shown, or directed toward the boundary 100 e = 18.75 x 2 or e 0.375 inches.
line between the side face and the front face.
Then by the simple rule as given in the previous Thus the centroid of the internal shear re-
article qx is also directed toward this common sisting force system lies on a vertical line
0.375 inches to left of point b as shown in Fig. Method 3. The k Method.

A14.23. For bending about the centroidal Z axis
without tWist the resultant of the internal
shear flow system would obviously, due to sym-
metry of section about X centroidal axis, lie on ~ z A (14)
the X axis, hence shear center for the given
channel section is at point 0 in Fig. A14.23. EXAMPLE PROBLEM USING THE THREE DIFFERENT
The external load of 100 lb. would have to be METHODS.
located 0.375 inches to the left of the center-
line of the channel web if bending of the Fig. A14.26 shows a Zee Section subjected
channel without twist is to occur. to a 10,000 lb. shear load acting through the
shear center of the section and in the direction
A14.8 Shear Stresses for Unsymmetrical Beam Sections. as shown. The problem will be to calculate the
In chapter A13, which dealt with bending shear flow qy at two points on the beam section,
stresses in beams, three methods were presented namely points b and c as indicated on the figure.
for determining the bending stresses in beams The shear flow at these two places will be cal-
with unsymmetrical beam sections. The bending CUlated by all 3 methods.
stress equations for these three methods will be Since all 3 methods require the use of
repeated here: - beam section properties and since the direction
of either the principal axes or the neutral
Method 1. The Principal Axis Method. axis are unknown the z
first step in the 1--1"--1-1
Mxp zp Mzb x p solution regardless of -L1 II
(J =- I xp I zp
- - - - - - - - - - (9) which method is used TIl b~
1
is to calculate the .1" I
section properties
Method 2. The Neutral Axis Method. about centroidal X and
Z axes. Table A14.3
2
x- - -c -x 3"
I
gives the calculations.
- - - - - - - - - - (10) The section has been
divided into 4 portions
labeled 1, 2, 3 and 4.
Method 3. The Method using section properties ,,'-.r....._ L -
about centroidal Z and X axes. For .1"
brevity this method will be called
the k method. Fig.AI4.26
TAbLE A14.3
Area Arm Arm Az· Ax'

Po rtion (Al z x Axz
where,
1 0.10 1. 45 -0.45 -.06525 .21025 .02025 .00838 .000017
2 0.14 0.70 0 0 .06860 0 .000117 .02287
3 0.14 -0.70 0 0 .06860 0 .000117 .02287
4 0.10 -1. 45 0.45 -.06525 .21025 .02025 .00838 .000017
~ -.1305 .55770 .04050 .01690 .04577
k Iz
2 =Ix I z - liz
IX = ~ A Z2 + ~ ix = .6035 in 4 •
In referring back to the derivations of Iz =~ A x 2 + ~ i z = .0574

equations (5), (6) and (7) the above equations
(9), (10) and (11) can be written in terms of
I xz = ~ A xz = - .1305
beam external shears instead of external bending (Note: In Table A14.3 i x and i z are the mom-
moments as follows: - ents of inertia of each portion about
Method l. The Principal Axis Method. its own centroidal axis) •
Vz p Vx SOLUTION BY PRINCIPAL AXES METHOD. (Method 1)
P
qY = ~ zp A - - - ~ x p A - - - - - - (12)
I xp I zp Let ~ be angle between Principal axes and
the X and Z axes. From chapter A13,
Method 2. The Neutral Axis Method.
tan 2 ~ = 2 I xz
Vn I z - Ix
qy = ~ zn A - - - - - - - - (13)
In = 2 (- 0.1305) _
0.0574 _ 0.6035 - 0.4778
A14.9
hence 2 ¢ = 25° - 32.2' or ~ = 12° - 46.1' Now SUbstituting in equation (15)

sin ¢ = 0.2210 and cos ~ = 0.97527 . 7348
0.63316 (1 x 0.1 x 1.5136)
The moments of inertia about the principal axes
can now be calculated.
Ix = Ix cos s ¢
+ I z sins ~ - 2 I xz sin cos ~ ¢ 6792
0.02782 (1 x 0.1) (- 0.1184)
=- 1756 + 2890
p
S S
= 0.6035 x .97527 + .0574 x .2210 2 (-0.1305) = 1134 Ib/in.
x .97527 x .2210 = .63316 in 4 •
Calculation of shear flow at point (c).
s ~ + 2 I
Iz = Ix sins ¢ + I z cos xz sin ~ cos ~ For portion (2) area A = 1.4 x 0.1 = 0.14
p
S S
zp = .70 x .97527 = 0.6825 in.
= 0.6035 x 0.2210 + .0574 x .97527 + 2 (-0.1305) xp .70 x .2210 = 0.1547
x .97527 x 0.2210 = 0.02782 in 4 •
The shear flow at point (c) equals the
The equation for shear flow q is, shear flow at point (b) plus the effect of the
portion (2) between points (b) and (c)., hence
Vz Vx
q = - -.-R .z zp - -p-.z x p - - - - - - - (15) 7348
Ix p I zp qc = 1134 - 0.63316 (0.14 x 0.6825) -
In Fig. A14.26 the external shear load is

10000 Ibs. acting in a direction as shown. Re-
0~62;282 (0.14 x 0.1547) = 1134 - 1109 - 5285 =
solving this shear load into z and x components,
we obtain, - 5260 Ib./in.
Vz = 10000 x cos 30° 8667 lb. The shear stresses at these two points (b)
and (c) would equal q/t = 1134/0.1 and - 5260/0.1
Vx = 10000 x sin 30° 5000 lb. or 11340 psi and - 52600 psi respectively,
Resolving these z and x components further SOLUTION BY NEUTRAL AXIS METHOD. (Method 2)
into components along the principal axes we
obtain, In this solution
it is necessary to find
nrf
Vz p = 8667 x .97527 - 5000 x .2210 = 7348 lb. the neutral axis for
the given external load-
Vx = 8667 x .2210 + 5000 x .97527 = 6792 lb. ing. In Fig. A14.28,
p the angle Q is the angle
between the plane of
Calculation of shear flow at point (b). (See loading and the zp prin-
Fig. A14.26). cipal axis, and this
Fig. A14.27 shows the position of the angle Q equals 30° + 12° Fig. A14.28
principal axes as calculated. The shear flow - 46" = 42° - 46". 150-29' z n
at the free edge of the upper portion (1) is Let a equal angle between x p p
zero. For the shear flow at point (b), the principal axis and neutral axis n-n.
area to be used in the summations Z zpA and From chapter A13, we find,
.z xpA is the area of element (1). The arms zp
I xp tan Q
and x p can be calculated by simple trigonometry. tan a =-
Fig. A14.27 shows the value of these distances,
namely x p =-
0.1184 and zp 1.5136. =
0.63316 x 0.9245
0.02782
= - 21.052
portion ( ) whence, a - - 87° - 17' (See Fig. A14.28 for

location of neutral axis.
sin a = 0.9989 , cos a = 0.04742

In = Ix p cos s a + Iz sinS a, substituting,
p
Fig. A14.27
2 2
In =0.63316 x .04742 + .02782 x:9989 = .02919 q = - [34.25 x 5000 - (- 7.406 x 8667)J Z xA -
The component of the given external shear load [3.257 x 8667 - (- 7.406 x 5000)J Z z A, whence
normal to the neutral axis n-n equals
Vn = 10000 x sin 45° - 29 t = 7130 lb. q =- 235438 Z x A - 65258 Z z A

Shear flow at point (b)
From equation (13)
For portion (1), x =- 0.45 in., z = 1.45 in.
q :::: - Vn Z zn A
In A = 1 x 0.1 = .1
Substituting,
Shear flow at point (b): -
The distance from the neutral axis to the qb =- 235438 (- 0.45) 0.1 - 65258 x 1.45 x 0.1
centroid of portion (1) equal zn - - 0.0466 in.
hence = 10597 - 9462 = 1135 Ib./in.
Point (c)
qb = - .~~~~9 (1 x 0.1) (- 0.0466) = 1135 Ib/in.
For portion (2), x = 0, z = 0.7, A = 1.4 x .1
= .14
Shear flow at point (c): - Zn = 0.1868 in.
qc = 1135 - 235438 (0) .14 - 65258 x 0.7 x 0.1
7130
qc = 1135 .02919 (1.4 x 0.1) .1868 = 1135 - 0 - 6395 = - 5260 Ib./in.
= 1135 - 6390 =- 5255 Ib./in. General Comments.

The author prefers solution method number 3
SOLUTION BY METHOD 3 - (The k Method). since it avoids the calculation of additional
angles and section properties as required in
In this method, only the section properties methods 1 and 2. Furthermore, in calculating the
about the centroidal X and Z axes are needed. shears and moments on the airplane wing, fuse-
These properties as previously calculated in lage and other major sturctural units it is con-
Table A14.3 are, venient to refer these shears and moments to the
conventional X Y Z axes, and thus these values
Ix = 0.6035, Iz = 0.0574, Ixz =- 0.1305 can be used in method 3 without further resolu-
tion. From an investigation of many airplane
The shear flow equation from eq. (14) is, stress analysis reports, it appears that the en-
gineers of most airplane companies prefer to use
q =- (k 3 Vx - k 1 Vz ) Z xA - (k 2 Vz - k 1 Vx ) method 3.
Z z A - - - - - - - - - --- - - - - - (16) A14.9 Beams with Constant Shear Flow Webs.
= 0.6035 - 0.1305
2
x .0574 - (- 0.1305) =
Fig. A14.29 shows a
beam composed of heavy
flange members and a curved ~r""e
Z
kfj111'r
thin web. For bending
= - 0.1305 = _ 7.406 about the X-X aXiS, the
.01762 web on the compressive, x-- t--~h
side of the beam absorbs \ I~
k = Ix Iz = 0.0574 = 3.257 very little compressive ~ I
2 I z - Ix~ .01762 stress, since buckling of ~
the web will take place 'Flange
Ix 0.6035 under low stresses, par- Z Fig. A14.29
k3 = -I-x-I"':'z=---I=-x-=-z = 0.01762 = 34.25 ticularly when the curvature of the web is
small. On the tension side, the web will be
Resolving the external shear load of more effective, but if the flange areas are
10,000 into x and z components, we obtain, relatively large, the proportion of the total
bending tensile stress carried by the web is
small as compared to that carried by the tension
Vx = 10000 sin 30° = 5000 lb.
flange. Thus for beams composed of individual
flange members connected by thin webs it is
Vz = 10000 cos 30° = 8667 lb.
often assumed that the flanges develop the en-
tire longitudinal bending resistance which
SUbstituting values of Vx ' Vz and k values in therefore means that the shear flow is constant
equation (16) we obtain -
AI4.11
over a particular web. In other words in the EXAMPLE PROBLEM - RESULTANr OF A CONSTANr FLOW
FORCE SYSTEM.
shear flow equation q = Vz I z d A, if the area
Ix Fig. A14.30 shows a constant flow force
of the web is neglected then q is constant be- system thru points ABC D E With q = 10 lb.
tween flange members. per inch. The resultant of this force system
is required.
RESULTANr OF CONSTANr SHEAR FLOW FORCE SYSTEMS
Fig. A14.29 shows a beam assumed to be
carrying a downward shear load (not shown) and R= 20(}#
to cause bending about axis x-x Without tWist.
Assuming the two flanges develop the entire
bending reSistance, the shear flow q is con- e
" 1
= 18.93TB
~10" •
il
stant on the web and acts upward along the web q =10#/in.
to balance the assumed external downward load. fA I - - -
The resultant of this resisting shear flow 5" r-- h =20':-
force system will give the lateral position of
the shear center for this beam section. The
t------
o Fig.AI4.30
problem then is to find the resultant of the SOLUTION: -
shear flow system.
Let q = load per inch along web (constant). Draw closing line between the beginning
and end points of force system. (line AE). The
Let R = resultant of the q force system. length h of this closing line is 20 inches.
From elementary mechancis, From eq. (17) R = q h = 10 x 20 200 lb. =
The direction of the resultant is parallel
R = J z q~ + Z q~ ,where qx and qy are the x to line AE or horizontal in this problem. To
find the location of R take moments about any
and y components of the q forces along the web. point such as (0). Draw lines from point (0) to
Since q is constant, Z qx is zero, hence, points A and E. The enclosed area (A) equals
(17) 5xlO+5xlO+ .511X5 2 +lOx 5 = 189.3 sq. in.
Equation (17) states that the magnitUde of From eq. (19)

the resultant of a constant flow force system 2 A 2 x 189.3
is equal to the shear flow q times the straight e = 11 20 = 18.93 in.
line distance between the two ends of the shear
flow system. Fig. A14.30 shows the resultant of 200 lb.
Since Z qx is zero, the direction of the acting at a distance e from (0) and parallel to
resultant is parallel to the straight line line AE.
joining the ends of the web.
The location of the resultant force is A14.10 Example Problems for Beams with Constant Shear
found by using the principle of moments, namely, Flows Between Flange Members.
that the moment of the resultant about any
point must equal the moment of the original EXAMPLE PROBLEM 1. Beam Section Symmetrical
force system about the same point. In Fig. About One Axis.
A14.29 assume point (0) as a moment center. Fig. A14.31 shows an open beam section com-
Then R e = q L r posed of 8 flange members connected by thin
but R = q h z
hence e = ----q1l
q L r _ L r
- tl (18)
In equation (18) the term L r is equal to

twice the area (A), where area (A) is the en-
closed area formed by drawing straight lines
from moment center (0) to the ends of the shear x
flow force system. Thus
2 A
e =~ -------------- (19)
The shear center thus lies at a distance e

J,
to the left of point (0), and the external shear
load would have to act through this point if Fig. A14.31
tWisting were to be eliminated.
A14.12 BENDlliG SHEAR STRESSES. SOUND AND OPEN SECTIONS. SHEAR CENTER.
sheet to form the webs and walls. The flange in order to indicate at what point the shear
members are numbered a to h and the areas of flow q is being calculated.
each are given on the figure. It will be as-
sumed that the webs and walls develop no bending %b = 2.5 ~a z A =- 2.5 x 5 x 0.1 =
resistance and thus the shear flow between ad-
jacent flange members will be constant. The = 1.25 Ib./1n.
problem is to determine the shear center for the
beam section. The first letter of the subscript refers
to the flange member where the shear flow q is
SOLUTION: - being calculated and the second letter indicates
on which adjacent side of the particular flange
Since the beam section is symmetrical member. Hence qab means the shear flow at
about the X axis, the centroidal X and Z axes flange (a) but on the side toward (b).
are also principal axes, since the product of
inertia I xz is zero. qab = qba = - 1.25 (Since no additional flange
The vertical position of the beam section area is added, and thus
centroid due to symmetry is midway between the shear flow is constant on
upper and lower flanges. sheet abo
To find the horizontal position of the
centroid, take moments of the flange areas qbc = qba - 2.5 Zb z A
about the left end or line bc: - - 1.25 - 2.5 x 5 x 0.4 = - 6.25 Ib./in.
= 0.4 x 15 + 0.2 x 10 + 0.2 x 5 6.25
1.6 qcb - 2.5 Zc z A
= 5.625 in. =- 6.25 - 2.5 x (- 5 x 0.4) =- 1.25
qdc = qcd =- 1.25
The moments of inertia for the section about
the centroidal x and y axes are: - qde = qdc 2.5 Zd z A =- 1.25 - 2.5 (- 5 x 0.1)
=0
qed = qde = 0
2 2 qef = 0 - 2.5 Ze z A =0 - 2.5 (- 5 x 0.1) = 1.25
Iz = 0.8 x 5.625 + 0.2 x 0.625 + 0.2 x
2 qfe = qef = 1.25
4.375 + 0.4 x 9.375 8 = 64.4 in 4 •
qfg = 1.25 - 2.5 Zf z A = 1.25 - 2.5
HORIZONTAL POOr-PION OF SHEAR CENTER: - (- 5 x 0.2) = 3.75
The horizontal position of the shear center qgf = qfg = 3.75
will coincide with the centroid of the shear
flow system due to bending about axis xx with- qgh = 3.75 - 2.5 Zg z A = 3.75 - 2.5 x 5 x 0.2 =
out twist. For simplicity, to eliminate large 1.25
decimal values for shear flow values an ex- qhg = qgh = 1.25
ternal upward shear load Vz = 100 lb. will be
assumed and the internal resisting shear flow qha = 1.25 - 2.5 Zh z A = 1.25 - 2.5 x 5 x 0.1 =
system will be calculated for this external 0 (checks free edge at h).
loading.
From equation (8) The sign or sense of each shear flow is for
Vz the shear flow in the y direction as explained
q - - -- Z z A , sUbstituting values of Vz and in the derivations of the shear flow equations.
y Ix Ix The procedure now is to determine the sense of
the shear flow in the plane of the cross-section
100 or in the xz plane. It is only necessary to
qy = - 40 Z z A =- 2.5 Z z A determine this sense at the beginning point, that
We could start the solution at either of is in sheet panel abo The surest way to deter-
two points (a) or h since these points are free mine this sense is to draw a simple free body
edges and thus qy is zero. In this solution, sketch of flange member (a) as illustrated in
Fig. A14.31. The shear flow on the cut face is
we will start at the free edge at point (a) and qy(ab) =-
1.25 and
go counterclockwise around the beam section.
The area of each flange member has been concen- this value is shown
q
trated at a point coinciding with the centroid on the free body. By ""''''--~ --f- (ab)
simnle rule given at /j':::- \.~
of each flange area. In solving for the q val- b O'-:"'- - - "
ues the subscript y will be omitted, and sub- the end of Art. A14.6, qX(;;P
scripts using the flange letters will be used the shear flow in the a
plane of the cross- Fig. A14.31
A14.13
section is also directed toward the common We will again start at the free edge ad-
boundary line and thus qx(ab) has a sense as jacent to flange (a) where qy = O.
shown in Fig. A14.31. The sense of the shear
flow on the cross-section will now continue in
qab =- 1.55 Za x A =- 1.55 (- 0.625 x 0.1) =
this direction until the sign changes in the 0.0971 Ib./in.
origional calculation, which means therefore the qba = qab = 0.0971
shear flow sense will reverse. Fig. A14.32
shows a plot of the shear flow pattern with the qbc = 0.0971 - 1.55 Zb x A = 0.0971 - 1.55
sense indicated by the arrow heads.
(- 5.625 x 0.4) = 3.592
qcb = qbc = 3.592
qcd = 3.592 - 1.55 Zc x A = 3.592 - 1.55 x
5.625 x 0.4) = 7.087
-6.25 qdc = 7.087
qde = 7.087 1.55 (- 0.625 x 0.1) = 7.184
qed = 7.184
qef = 7.184 - 1.55 x 4.375 x 0.1 = 6.504
qfe 6.504
Fig. A14.32
qfg = 6.504 - 1.55 x 9.375 x 0.2 = 3.589
The results will be checked to see if
static equilibrium eXists relative to Z Fx and qgf = 3.589
Z Fz O. qgh 3.589 - 1.55 x 9.375 x 0.2 = 0.674
Z Fz = 100 (ext. load) - 10 x 6.25 - 10 x 3.75 +
Fig. A14.33 shows the plotted shear flow
1.25 x 0.5 x 4 - 1.25 x 0.5 x 4 = 0 (check). results. The signs of the calculated shear
flows are for shear flows in the y direction.
Z Fx =- 5 x 1.25 + 5 x 1.25 - 5 x 1.25 + 5 x Simple consideration of a free body of flange
1.25 =0 (check). member (a) will give the sign or sense of the
shear flow in the plane of the beam section.
The shear flow force system in Fig. A14.32 Thus in Fig. A14.34 qx must act as shown when
causes the section to bend about axis xx with- qy is positive.
out tWist. The resultant of this system is 100 .674
.0971
lb. acting down in the z direction. The posi-
tion of this resultant will thus locate the
lateral position of the shear center.
Equating the moments of the shear flow
system about some point such as (c) to the mo- 3.592
ment of the resultant about the same point we
obtain: -
100 e = 10 x 3.75 x 15 - 1.25 x 0.5 x 2 x 5 -

1.25 x 0.5 x 2 x 10 + 1.25 x 0.5 x 2 x 15 ~~~
7.087 7.184 6.504
hence e = 562.5/100 = 5.625 inches. -R
V x = 100#
Fig. A14.33
Thus the shear center lies on a vertical line
5.625 inches to right of line bc.
CALCULATION OF VERTICAL POSITION OF SHEAR I

I I I
~
CENTER.
-----For convenience as before, we will assume I
- 1" I
a shear load Vx = 100 lb. and compute the re- qx I- 5.625 -j
sisting shear flow system to resist this load
in bending about axiS zz without twist. The I
a t
resultant of this shear flow system will give
the vertical location of the shear center. The
shear flow equation is, Fig. A14.34 1
S.C.
6.43
+
qy = - Vx Z x A = - ~ Z x A = - 1.55 Z x A Fig. A14.35

Iz 64.4
Checking to see if Z Fz =0 and Z Fx = 0: - Calculation of Ix, I z and I xz --

Z Fz = 3.592 x 10 - 3.589 x 10 + .5 x .0971 - .5 Ix = Z A Z2 = .5 (5.333 2 + 2.667 2 ) + 1
x .0971 + 5 x .674 - .5 x .674 + .5 x .0971 - .5 (5.333 2
+ 6.667 2
) = 90.667
x .674 - .5 x 6.504 + .5 x 7.087 =0 (check) Iz =Z A x 2
= 1 (10.667 2
) + 2 (5.333 2 ) = 170.667
Z Fx = 100 + 5 x .0971 + 5 x .674 - 5 (7.087 +

I xz = Z A xz = 1 x 6.667 (- 5.333) + 1 (- 5.333)
(- 5.333) + 0.5 x 10.667 x 2.667 + 0.5 x 10.667
7.184 + 6.504) = 0 (check)
(- 5.333) = - 21.333
The resultant R of the internal shear flow
system is a horizontal force of 100 lb. acting The constants k~, k s and k" are now determined -
toward the left. To find the location of the
resultant take moments about a point 0.5 inch I xz - 21.333
below point (c).
k = = 90.667 x 170.667 - 21.333 2 =
~
Ix I z - I xz
Re = Z M - 21.333
- - .00142
15019
100 e =5 (.0971 + .674) 11 + 10 x 3.589 x 15 +
.5 x .0971 x 5 - (.5 x .674 x 10) + .5 x .674 x Iz 170.667
ks = = = 0.01136
15 + .5 x 6.504 x 15 + .5 x .674 x 10 - (.5 x Ix I z - Ix~ 15019
.0971 x 5) Ix 90.667
k" = I = = 0.006037
100 e = 643 I
x z - Ix~ 15019
e = 643/100 = 6.43 inches ResolVing the given shear load of 141.14 into z
and x components, we obtain,
Fig. A14.35 shows the resulting shear center
location for the given beam section. Vz 141.14 x sin 45° = 100 lb.
ElCAMPLE PROBLEM 2. Unsymmetrical Beam Section. Vx 141.14 x cos 45° 100 lb.
Fig. A14.36 shows a four flange beam sec-
tion. The areas of each flange are shown ad- From equation (14) -
jacent to flange. The external shear load V
equals 141.14 lb. and acts in a direction as
shown. The problem is to find the line of
qy =- (k" Vx - k~ Vz ) Z x A - (k s Vz - k~ Vx )
Z z A
action of V so that section will bend without SUbstituting
twisting.
qy =- [.006037 x 100) - (- .00142 x 100J Z x A -
! Load = 141.14
a 1.0 I Z ft45° [.01136 x 100 - (- .00142 x 100)J Z z A
"i
x- ~_- +r:- ~5
I d
--r whence
Hlz
';:b,.......,
•
0.5
..--~----=-=ec
i
qy = - 0.7457 Z x A - 1.278 Z z A
Fig. A14.36
H16"---l
We will start at flange member (a) where qy is
SOLUTION: - zero on the free edge side of the member.
To solve this problem, method (3) will be Qab =- 0.7457 x 1 (- 5.333) - 1.278 x 1 x 6.667
used.
To locate centroidal x and z axes: -
qba = Qab = - 4.544
z -
_ ZAx
~
=1 x 12 +
3
0.5 x 8 =5 •333 i
n. qbc =- 4.544 - 0.7457 x 1 (- 5.333) - 1.278 x
1 (- 5.333) = 6.249 Ib./in.
x-~
-: Z A z = (0.5 + 0.5) 16 =5.333 in. qcb = qbc = 6.249
3
A14.15
qcd = 6.249 - 0.7457 x 0.5 x 10.667 - 1.278 x A14.11 Shear Center Location By Using Neutral Axis
Method.
0.5 (- 5.333) = 5.680
In a beam subjected to bending there is a
~c = qcd = 5.680 definite neutral axis position for each differ-
qda = 5.680 - 0.7457 X .5 X 10.667 - 1.278 X 0.5 ent external plane of loading on the beam. The
shear flow equation with respect to the neutral
X 2.667 = 5.680
- 3.977 - 1.704 = axis is,
o (checks free edge at d where qy
must be zero.) - in Z ZnA - - - (20)
n
Fig. A14.37 shows the resulting shear flow re- where, Vn = Shear resolved normal to neutral axis
sisting pattern. The sense of the shear flow in In =Moment of inertia about neutral axis
~~~ Zn =Distance to neutral axis
U
,,~.
a / v= 141.14
In finding the shear center location of an
Fig.A14.37a unsymmetrical section, it is convenient to as-
jfR= 141.14
d ....- a
qx
q= 5.680
sume that the Z and X axes are neutral axis and
find the shear flow system for bending about
each axis by equation (20). The resultant of
b q= 6.249 c each of these shear flow force systems will pass
Fig. A14.37 through the shear center, thus the intersection
of these two resultant forces will locate the
shear center.
the plane of the cross-section is determined in
web at flange member (a) by the simple free Example Problem
body diagram of stringer (a) in Fig. A14.37a
Check Z Fx and Z Fz to see if each equals The same beam section as used in the
100. previous article (see Fig. A14.36) will be used
to illustrate the neutral axis method.
6.249 X 16 = 99.99 (checks Vx = 100)
Fig. A14.38 shows the section with the
12 X 4.544 - 8 X 5.68 =-99.94 (checks centroidal axis drawn in. The X axis will now
Vz = 100).
The resultant of the internal resisting

shear flow system equals V 100 2 + 100 2
lb.
= 141.14 d
-x ~T
~:.47
To locate this resultant we use the prin- c
ciple of moments. Taking point (b) as a moment b c
center,
Fig. A14.38 Fig. A14.39
141.14 e = 8 x 5.68 x 16 - 4.544 x 36 n be assumed as the neutral axis for an external
212 plane of loading as yet unknown. We will further
hence e = 141.14
= 1.50 inch. assume that when this unknown external loading
is resolved normal to the X neutral axis, that
Therefore external load must act at a distance it will give a value of 100 lb., or Vz
e = 1.50" from (b) as shown in Fig. A14.37. The From the previous article Ix
100. =
90.667. =
load so located will pass thru shear center of
section. To obtain the shear center location, Since the X axis has been assumed as the
another loading on the beam can be assumed, and neutral axis, equation (20) can be written
where the line of action of the resultant of the
resisting shear flow system intersects the re-
sUltant as found above would locate the shear
qy = -
x
-?
Z ZA, hence,
center as a single point. If the shear center
location is desired it is convenient to assume Qab - - 9~~~67 x 6.667 x 1.0 -7.35 lb./in. =
a unit Vz and Vx acting separately and find the
horizontal and vertical locations of the shear qbc - - 7.35 - 9~~~67 x (-5.333)1 -1.47 =
center from the 2 separate shear flow force
systems. 100
qcd = -1.47 - 90.667 (-5.333) 0.5 = 1.47
Fig. A14.39 shows the resulting shear flow

values.
~Fx = 1.47 x 16 = 23.52 lb. Take moments about (b) and let (e) equal
distance to reSUltant R.
~Fz = 7.35x12+8x1.47 = 99.96 lb. (check)
100.8e = 3.125xnx6 2
+3.13x8x16
R = V 100 2
+ 23.52 2 = 103 lb. e = 753/100.8 = 7.47 in.
tan Q = 23.52/100 =.2352 Fig. A14.40 shows the position of this re-
hence Q= 13 0 - 16' lOTIR'to> sultant force. Where it intersects the previous
resultant force gives the shear center location.
Let e =
distance from 23.52
A14. 12 Problems
resultant R to point b.
(1) Fig. A14.42
shows the cross-
Equating moments of section of a wood -,
resultant about (b) to that of shear flow
I I
t
beam glued to-
system about (b), gether on lines a a
a-a and a-b. The
103e = -7.35x6n 2
+1.47x8x16 beam is subjected 4"
644 to a vertical
=- =-6.25
e 103
Fig. A14.40 shows the location of the

resultant. We know the shear center lies on
in. shear V 2400
lb. Determine
shearing stress
on sections a-a
=
I
,
~
J
I- .1"
i-
the line of action of this resultant. Thus we and a-b. Find
maximum shearing
Fig. A14.42
stress on beam
section.
,?>.
\.~~ Fig. A14.43
6.25 3.13 (2) Fig. A14.43
shows a Zee sec- ~
r-1"~ a
tion loaded by a 1/4" a---~
1000 lb. load T 1/16 I
Fig. A14.41
acting through
b- -b
Fig. A14.40 the shear center S. C. 2"
~4l
as shown. Find
must obtain another resultant force which the shear stress
passes through the shear center before we can at sections a-a l00Olb. <So
definitely locate the shear center. Therefore and b-b by three 1-1,,-1
we will now assume that the Z centroidal axis different methods.
is a neutral axis and that a resolution of the ...L .2°"
external load system gives a shear Vx 100 lb. = f T...--.--1--·..L....,·~10-"~.10"
r
=I-=-=-=O-1 3/8"
I T 1" 1\' .11"
qY Vx ~ XA
=- Iz 170.667 = II .05" T} .040" 10" Neglect Web in
_~,14.44~"
Iz
2' :11_F_ig_. . -.L I Resisting Bending
17~~~67 x (-5.333)1 = 3.125 lb./in. ~=:::::JJ5/16L .2°" .1 0 " .1°"
%b - -
1" 1"+-1" T r- 6" + 6 " - 1
qbc = 3.125 -1~g~66 (-5.333)1 = 6.25 lb./i Fig. A14.45 Fig. A14.46
(3) Determine the shear flow diagram and the
shear center location for bending about hori-
qcd = 6.25 -1~g~66 (0.5) (10.667) = 3.13 zontal centroidal axis for the beam sections
as given in Figs. A14.44 to A14.46.
Fig. A14.41 shows the shear flow results.
LO .4 d
~Fx =-6.25 x 16 = -100 lb. Fig. A14.47
~Fz =-8x3.13+12x3.125 = 12.5 lb. .15 0 ' I

8" tt "4" 8"
R = V 100 2
+ 12.5 2
= 100.8 R=100.8 l ---±- bLO .4c
.1°".:19" .<B' .mp"
12.5~100
Fig. A14.48
t an Q -- 12.5
100 -- .125 ~
(4) Determine the shear center location for the
beam sections in Figs. A14.47 and A14.48. As-
A14.17
sume flange members develop entire bending

stress resistance. q=10#/in.
a b
T
(
l~o4/1
EJ 10"
I
1
r-20"-l~'
t-t
+
I
10"
10"
l£50"~
Fig.A14.51
Fig.A14.52
(7) In Fig. A14.5l, the shell structure is sub-

jected to a torsional moment M 50,000 in. lb. =
The shell skin shown dashed is cut out, thus the
torsional moment is resisted by the constant
(5) Determine the shear center for the beam shear flow on the two curved sheet elements ac,
section of Fig. A14.49. Assume only the 8 and bd. Determine the value of the shear flow.
stringers as being effective in bending. Area
of stringers (a) and (b) =2 sq. in. each.
(8) Determine the moment of the constant flow
force system in Fig. A14.52 about point (0).
All other stringers 1 sq. in. each. Also find the resultant of this force system.
1 sq. in.
a r
,
e i P
c
r- -+ t
1.5°"
12 8 10
1. 0 0.8 1.2
'4j
T
.5 sq. in.
15" 10" 20"
=5 b ...,::::...------:::.a.-"
1.2 0.7°" 0.8 1.0 1
r +-
I
.8 sq. in. 11--- 20"

Fig. A14. 53
·1 7 15 "-f 8 1
Fig. A14. 54
~--19'L----" . 5 sq. in. (9) In Fig. A14.53, the four stringers a, b, c
Fig.A14.50 and d have the same area. Assume the webs in-
effective in resisting bending stresses. De-
termine the distance (e) to product bending
about the horizontal axis without twist.
(6) Determine the shear center for the un-
symmetrical beam section of Fig. A14.50. As- (10) For the wing cell beam section in Fig.
sume sheet connecting the four stringers as A14.54, determine the location of the shear
ineffective. Areas of stringers shown on Fig. center. Assume webs and walls ineffective in
bending.
STRUCTURAL TESTING IS AN IMPORTANT PHASE OF STRUCTURAL DESIGN.

CHAPTER A 15
SHEAR FLOW IN CLOSED THIN - WALLED SECTIONS
A1S. 1 Introduction. The wing, fuselage and em- open sections we could start the summation at a
pennage structure of modern aircraft is essen- free surface where q would be zero, thus the
tially a single or multiple cellular beam with summation to any other point would give the true
thin webs and walls. The design of such shear flow qy. In a closed cell there is no
structures involves the consideration of the free end, therefore the value of qy is unknown
distribution of the internal resisting shear for any point.
stresses. This chapter introduces the student
to the general problems of shear flow distri- Equation (1) gives the shear distribution
bution. Chapter A14 should be covered before for bending about the X axis without tWist.
taking up this chapter. The general procedure is to assume a value of
the shear flow qy at some point and then find
A1S.2 Single Cell Beam. Symmetrical About One Axis. the shear flow pattern for bending without
All Material Effective in Resisting Bending tWist under the given external load. The cen-
Stresses.
troid of this internal shear flow system will
Fig. A15.1 shows a single cell rectangular be the location where the external shear load
beam carrying the· load of 100 lb. as shown. should act for bending without twist. Since
The problem is to find the internal resisting the given external shear would have a moment
shear flow pattern at section abed. about this centroid, this unbalanced moment
must be made zero by adding a constant shear
100# flow system to the cell.
To illustrate we will assume qy to be zero
qo = O. The term
a
*
at point 0 on the web ad.
= 100/62.5 = 1.6
%0 = -1.6 2: ZA = -1.6x2.5x5xO.l = -2 Ib/in.
o
qba = -2 -1.6 2:~zA = -2 -1.6 x 5 x 20 x .05 = -10
Fig. A1S-l
x+ z
qo'b = -10-1.6 2:~'zA=-10-1.6X2.5X5X.05=
qco' = -11 - 1.6 2:~, zA = -11-1.6 x (-2.5)5 x
d
.05 = -10
qed = -10-1.6 Zc zA = -10-1.6(-5)20 x .05

-11
= -2
Solution 1
qod = -z-1.6 2:~ zA =-2-1.6(-2.5)5xO.l=0
Due to symmetry of material the X cen-
troidal axis lies at the mid-height of the
beam. The shear flow equation requires the Fig. A15.2 shows a plot of the shear flow
value of Ix, the moment of inertia of the reSUlts. On the vertical web the increase in
section about the X axis. shear is parabolic since the area varies
directly with distance z.
Ix = 1~X.15XIO"+2~OX.05X5lilJ = 62.5 in. 4
The intensity of qx and qz in the plane of
From Chapter A14, the equation for shear the cross-section is equal to the values of q
flow is, found above which are in the y direction. The
sense of qx and qz is determined as explained
qy = - i~ 2: zA - - - - - (1) in detail in Art. A14.6 of Chapter A14.
This equation gives the change in shear
flow between the limits of the summation. In
A1S.l
A15.2 SHEAR FLOW IN CLOSED THIN-WALLED SECTIONS. SHEAR CENTER.
Adding this constant shear flow to that of

10 it/in. Fig. A15.2 we obtain the final shear flow
2 it/in.
pattern of A15.3.
1. 66 it/in.
1I1-;-':-':-~ q = 11 it/in.
-8.34 it/in.
2.66 it/in.
2 it/in.
4.+y:.,.Lfl---
Fig. A15-2 -6.34 it/in.
Fig. A15.2 also shows the resultant shear Fig. A15-3

flow force on each of the four walls of the If this constant shear flow of -8.34 was
cell. The resultant shear force on each not added then the external load of 100 lb.
equals the area of the shear flow diagram on would have to be re-located so that it passed
each position. through the centroid of the shear flow pattern
in A15.2. To find this centroid location, we
For example, can equate the moment of the internal shear
flow pattern about some point to the moment of
Qab =( 2 ; 10 )20 =120 lb. the resultant of the system about the same
point.
1
Qda =3 x 2 x 10 = 6.67 lb. Resultant R = Y ~Fx "+ ~Fz 2= V 100" + 0
= 100 acting down.
Qbc =10 x 10 + 0.667 x 1 x 10 =106.7
Take moments about point d. Let x- equal
The internal shear flow force system as distance to resultant.
given in Fig. A15.2 will now be checked forw
equilibrium with given external shear loading Hi = m d (internal system)
of 100 lb. as shown in Fig. A15.1.
100 x = 120 x 10 + 106.67 x 20
~Fz = 100 (external) + 6.67 - 106.67 hence x = 33.33 in. Thus the externalload
- 0 (check) would have to be moved 33.33 inches to right if
the pattern of A15.2 would hold it in equilibrium.
~Fx =120 - 120 =0 (check) Since we assume q =
zero at point 0, this means
an open cell with the free end at 0 would bend
without tWist if the external load was moved the
distance -x.
Equilibrium of moments must also be
satisfied. Take moments of all forces about
point d. The external load has no moment The student should work this same problem
about d. by assuming q at some other point is zero in-
stead of point 0 as assumed in the above solu-
~~ =120 x 10 + 106.67 x 20 tion.
= 3344 in. lb. clockwise
Solution No.2. Shear Center Method.
Thus we have an unbalanced moment which
must be made zero if we are to have equil- In this solution, we determine the centroid
ibrium. of the internal shear flow system for bending of
the closed section about axis X without twist.
The unbalanced moment of 3334 in. lb. This point is called the shear center. The
can be balanced by adding a constant shear external shear load can then be resolved into a
in a counter-clock#ise direction around the shear force acting through the shear center plus
cell. The value of this balancing shear flow a torsional moment about the shear center.
would equal,
We start the solution, exactly as in solu-
q -- - 2A
M -- 2 -3334
x 200 -- - 8 • 34 Ib/'In.
tion 1, by assuming the shear flow q 0 at =
point O. In a sense we are cutting the cell at
o and making it an open section. The resulting
(A equals area of cell = 10 x 20) shear flow is as given in A15.2. This open
A15.3
section will bend without tWist if the ex-

ternal shear load acts through the shear
center of the open section.
The closed section will be assumed to

bend without tWist, and the resulting shear
flow pattern will be determined. 6.36 JIlin.
.64 JIlin.
The equation for angular twist g (See
Chapter A6) per inch length of beam is,
_ 1 qL
g - 2AG Z t ' where L equals the length
of a web or wall.
or 2Ag =~ i
Z L • The right hand 3ide of
Fig. A15-4
The location of the resultant of the shear

this equation represents the total shearing flow force system of Fig. A15.4 will locate the
strain around the cell which must be zero for horizontal position of the shear center. Due
no twist of cell. Since G is constant, we to symmetry of the section about the X axis,
can assume it as unity as only relative the vertical position of the shear center will
values of strain are needed in the solution. be on the X axis, because for bending about the
Thus the total shearing strain 6 around cell Z axis, the shear flow would be symmetrical and
is proportional to, thUS, the resultant would coincide with the X
6=ZiL----
axis.
- - - - - - (2)
Fig. A15.4 also shows the resultant shear
Using the values of q from Fig. A15.2 load Q on each portion of the cell wall which
and substituting in (2), equals the area of the shear flow diagram for
various portions as shown.
6 =Z qLt =1..:9
f] x 5 +
x .1
(10 + 2) 20 + 10 x 5
2 . 05 .05
ZF z =-
56.9 - 43.1 =-
100, which balances
the external load of 100 lb.
+ (11 - 10)(.667 x 5)1 2 = 7000 ZFx =
0 by observation. Take moments about
.05 J point 0, the intersection of axis XX and side ad.
Since the section and shear flow pattern _ ZM o _ 43.1 x 20 + (-23.8 + 16.58)10 _
is symmetrical about the X axis, the substi- x - ZF z - 100 -
tution above is written for one-half of the
cell and the results multiplied by 2. When 7.90 in.
the shear flow q varies over a portion the
average shear flow is used in the above Hence the shear center lies on the X axis, 7.9"
substitution. from side ad.
If the cell is not to twist the relative The moment of the external load of 100 lb.
tWist of 7000 must be cancelled by adding a about shear center equals 100 x 7.9 =
790 in.
constant shear flow q around the cell to give lb. clockwise. The moment of the internal shear
a total shear strain of -7000. flow of Fig. A15.4 is zero thus we have an un-
balanced moment of 790. Therefore for equil-
The shearing strain for a constant q ibrium of moments we must add a constant shear
equals, flow q around cell to develop -790 in. lb. or
= Z ~ = q 5~1 .~~ .~~2 = -7000

_ 11 _ - 790
6 + + q - 2A - 2 x 200 1.98 Ib/in.
7000 _ Adding this constant shear flow to that of

hence q = "llO6 - -6.36 Ib/in. Fig. A15.4, we obtain the final shear flow
pattern which will be identical to that in
Fig. A15.4 shows the resultant shear Fig. A15.3 or the results of solution 1.
flow pattern if the constant shear flow of
-6.36 is added to the shear flow pattern of A15.3 Single Cell - 2 Flange Beam. Constant Shear Flow Webs.
A15.2.
As discussed in Art. A14.9, the common
aircraft cellular beam is made up of thin sheet
Al5.4 SHEAR FLOW IN CLOSED THIN-WALLED SECTIONS. SHEAR CENT ER.
walls and webs which are stiffened by members hence q1 =16.21 Ib/in. (sign comes out pOSitive,
usually referred to as flange members. A hence assumed sense of q1 is correct).
simplifying common assumption is to assume
that the flange members alone develop the Fig. A15.7 shows a plot of the internal
resistance to the bending moment. This shear flow resisting system.
assumption therefore means that the shear flow
is constant between flange members. A single cell beam having only two flanges
can resist only external loads which are parallel
Fig. A15.5 shows a single cell beam with to the line AB, and thus a two flange box beam
two flanges at A and B. Find the internal is not used very often in aircraft structure.
shear flow force system when the beam carries When the bending moment in a plane at right
the external load of 100 lb. as shown. angles to line AB is small, the resistance of
the curved panel to compressive bending stresses
may be SUfficient to resist such external bend-
1-10,J ing moments and thus be satisfactory.
-.---'1-' V=100#
1 100 6.21
I A15.4 Shear Center of Single Cell - Two Flange Beam.
I
1
,~ 1/2" Let it be required to find the shear center
......--.--o;:r=---L of the beam as given in Fig. A15.5. In other
z
i
~10"~
words where would the external load have to be
placed so that the beam would bend without
twist.
A~
6.21
Fig. A15-6 Fig. A15-7
~.=w.I--C1/2"
2.05 7.95
Fig. A15-5
It is assumed that the two flanges

develop the entire bending stress resistance.
This means that shear flow is constant on each
web. Let q1 and qs be the constant shear floW
as shown in Fig. A15.6. The sense or direction
of these shear flows will be assumed as indi- Fig. A15-8 Fig. A15-9 Fig. A15-10
cated by the arrow heads in the figure.
The force system in the plane of the

cross-section has 2 unknowns, namely q1 and
qs and thus we can solve for q1 and qs by
simple statics.
Fig. A15.8 shows the resulting shear flow
system in resisting the 100 lb. exterr-al load
To find qs take moments about point A. acting as shown in Fig. A15.5. This shear flow
system will cause the cell to twist. Therefore
ZMA = 100 x 10 - qs (2 x 80.52) =0 we add a constant shear flow q to the cell to
produce zero twist (Fig. A15.9). The centroid
hence qs = 1000/lb 161.04 = 6.21 Ib/in. of the combined shear flow system will then
locate the lateral location of the shear center.
The sign of qs comes out plUS, thus the
assumed sense of qs as shown in Fig. A15.6 is To find q we must write an expression
correct. which measures the twist when SUbjected to the
shear flows of Figs. A15.8 and 9 and equate the
In the above equation of moments about result to zero, and then solve for the one un-
point A, the moment of the constant flow
known q.
system qs about A, equals qs times double the
enclosed area formed by lines running from A
to the ends of the web which carries the shear o =~ ~L =O. (Clockwise q is positive)
flow qs. In this case the area is the area
of the cell, or 80.52 sq. in. Substituting, using the shear flows of Figs.
A15.8 and A15.9,
To find the remaining unknown q1 we use
the equilibrium equation,
~Fz =0 = - 100 - 10 x 6.21 + 10Q1 =0

A15.5
6 - - 16.21.04x 10 6.21 x 24.28 + 10q

.025 .04
of the cell, the web and wall thickness do
influence the amount of twist for a given
torsional load. In the shear center solution,
24.28q it is known what portion of the shear flow is
+
.025 =0 due to torque or pure tWist, and also that due
to bending without tWist, which fact is some-
hence q =8.26 lb. in. times of importance.
Adding this constant shear flow to that TORSIONAL DEFLECTION OF CELL
of Fig. A15.8, we obtain the shear flow of
Fig. A15.10. The angUlar twist as given by the final
shear flow pattern of Fig. A15.8 equals
The lateral position of the shear center 2QAG = ZqL/t, whence
is given by the location of the resultant of
the shear flow system of Fig. A15.10. 2QAG = -6.21 x 24.28 _16.21 x 10 = -10082 __ (3)
.025 .04
After finding the shear center location,

we found that the external load had a moment
ZF x =0 of 1330 in. lb. about shear center, which was
ZF z =10 x 7.95 + 2.05 x 10 = 100 lb. resisted by a constant shear flow of - 8.26
Ib/in. The angUlar twist under this pure
Therefore R = 100 lb. torque shear flow should therefore give the
same result as equation (3) above.
Equate moments of R about point A to
moment of shear flow system about A. 2QAG =- 8.26 x 24.28 _ 8.26 x 10 = _ 10082
.025 .04
Re = ZMA which checks the result of equation (3).
100e = 2.05 (2 x 80.52) A15.5 Single Cell-Three Flange Beam. Constant Shear
Flow Webs.
e =3.30 in.
100
Thus shear center lies 3.30 inches to 100 lb.
left of line AB. (See Fig. A15.10).
1 'II ~5" I
0::'b T'
A "
Thus if the given external load of 100 .____-----.'=- 25 lb.
C~B
lb. acts through the shear center, it will
produce the shear flow system of Fig. A15.10. y
However it acts 13.3 to right of shear center,
hence it produces a clockwise moment of 100 Fig. A15-12
x 13.3 =
1330 in. lb. on the cell. For
equilibrium, this moment must be balanced by
a constant resisting shear flow around cell
which will produce a moment of -1330.
- 1330
The required q = 2 x 80.52 =-
8.26 lb. in.
Area Cell
C-...;f-----+--__4
~2.639
which if added to the shear flow system of • 128. 54 sq. in. Fig. A15.13
A15.10 will give the true shear flow system Fig. A15.11
of A15.8.
Fig. A15.11 shows a single cell beam with
Thus having the shear center location, three flange members, A, Band C, carrying the
the external load system can be broken down external load as shown. A three flange box if
into a load through the shear center plus a the flanges are not located in a straight line
moment about the shear center. The shear can take bending in any direction and therefore
flow due to each is then added to give the is often used in design because of its simpli-
true resisting shear flow. city.
It should be noticed that the web or skin For such a structure, there are six un-
thickness does not influence the magnitude of knowns, namely, the axial load in each stringer
the shear flow system in a single cell beam. and the shear flow q in each of the three sheet
A change in thickness, however, effects the panels that make up the cell. For a space
unit shearing stress and therefore the shear- structure, we have six static equations of
ing strain and thus in computing angular tWist equilibrium, thus a three flange single cell
A15.6 SHEAR FLOW IN CLOSED THIN. WALLED SECTIONS. SHEAR CENT ER.
beam can be solved by statics if we assume

that the three flange members develop all the 100
bending stress resistance, thus producing ~5" 1---
constant shear flow webs. ............ _! A
V <lac +
Fig. A15.12 shows the cross section ABC. 'I qbc qab l
+--+ ~ ~,
+ =
The three unknown resisting shear flows have C B
been assumed with a positive sign. (Clockwise
flow is positive shear flow). These three Fig. A15-l4 Fig. A15-l5 Fig. A15-l6
unknown shear flows can be determined by
statics.
To find qca take moments about point B The three unknown resisting shear flows
and equate to zero. will be assumed with the sense as indicated by
the arrow heads.
2MB = 100 x 5 - 25 x 10 + qca(128.54 x 2) = 0
To find Qac take moments about B
250 -- - 0 • 972 Ib/in.
h ence qca - - 257.08
2MB =100 x 5 - qac (128.54 x 2) =0
To find qab take ZF z =0
Qac =1.945 Ib/in.
ZF z =100 - 10 x 0.972 - l0Qab =0
ZFx =- 15 x 1.945 + 15qbc = 0
Qab = 9.13 Ib/in.
qbc = 1.945
To find qbc take ZF x =0
ZF z =100 - 10 x 1.945 - l0Qab =0
ZF x - - 15 x 0.972 - 25 - 15qbc =0
Qab =8.055 Ib/in.
hence qbc =- 2.639 Ib/in.
The algebraic signs of the unknown q value
The signs of qca and qbc came out negative, all come out positive, thus the assumed direction
hence the sense of the shear flow on these cell of shear flows in Fig. A15.14 is correct.
wall portions is opposite to that assumed in
Fig. A15.12. The resulting shear flow pattern To make the cell twist zero, we must add a
is plotted in Fig. A15.13. constant shear flow q to the cell (see Fig.
A15.15). The relative tWist under the shear
The student should realize the thickness flow of Figs. 14 and 15 will be equated to zero.
of the wall elements does not influence the
shear flow distribution if we assume the three Z~ =_ 1.945 x 20.71 1.945 x 15 + 8.055 x 10
flanges develop the entire resistance to the t .03 .025 .04
bending moment. + 20. 7lq + 10q + 15q =0
.03 .04 .025
A15.6 Shear Center of Single Cell-Three Flange Beam.
Constant Shear Flow Webs. Whence, q = 0.322 Ib/in. with sense as
assumed in Fig. A15.15. Adding this constant
Let it be required to determine the shear shear flow to that of Fig. A15.14 we obtain the
center location for the beam in Fig. A15.11. shear flow system of Fig. A15.16. The resultant
The shear center is a point on the beam cross- R of this shear flow system is obviously - 100
section through which the resultant external lb., since the external load was 100 lb. The
shear must act if the cell is to bend without location of this resultant R will therefore
twist. locate the horizontal position of the shear
center. Equate moment of resultant R about
The shear center location will be deter- point B to the moment of the shear flow system
mined in two steps, first its horizontal about B, whence,
location and then its vertical location.
100e = 1.623 (128.54 x 2)
Calculation of horizontal location:-
or e = 417/100 = 4.17 in. from line AB.
We will assume any vertical shear load, (Fig. A15.16)
as the example, the same vertical shear as used
in the problem of Art. A15.4, namely, a 100 Calculation of Vertical Position of Shear Center
lb. load acting five inches from A, as
illustrated in the following Fig. A15.14. A convenient horizontal shear load will be
A15.7
assumed acting on the cell. Since we used a A15.7 Single Cell-Multiple Flange-One Axis of
25 lb. load in the example problem of Art. 15.5, Symmetry.
we will assume the same load in this solution.
Fig. A15.17 shows the loading and the assumed Fig. A15.20 shows a single cell beam with
directions of the three unknown shear flows. 8 flange members, carrying a 100 lb. shear load.
The resisting shear flow system will be calcu-
lated •
• ()l\:...~~~--,
R=2.Q. T
e z
~ .40" . 10" . 10" . 20"
J[
B
g
Fig. A15-17 Fig. A15-18 Fig. A15-19
x
Solving for the three unknown shear flows
in Fig. A15.17. d e
----.- """!'---~-----!"--~-
2MB =- 25 x 10 + qca (128.54 x 2) = 0 • 5"
qca =0.972 Ib/in.

Fig. A15-20
ZF x =- 25 + .972 x 15 + 15qcb =0
qcb =0.695 The beam section which is symmetrical about
the X axis is identical to the beam section
ZF z =10 x 0.972 - l0Qab = 0 relative to flange material which was used in
example problem 1 of Art. A14.10.
Qab = 0.972
SOLUTION:-
A constant shear flow q is now added to
cell to make tWist zero (Fig. A15.18). Assuming the 8 flanges develop all the
bending stress resistance, the shear flow will
Writing ZqL/t for both loadings and therefore be constant between flanges. Since
equating to zero:- the beam section is a closed one the value of
the shear flow q at any point is unknown. Thus
ZqL/t = 0.972x20.71 _ 0.695x15 + 0.972xlO we will imagine the top cover cut between flange
.03 .025 .04 members a and h, thus making q zero in this
+ 20. 7lq + 10q + 15q =0 panel due to the free end at the cut. We now
find the internal resisting shear flow system
.03 .04 .025
for bending of this open section about axis
whence, q =- 0.324 Ib/in. x-x under a external shear load Vz 100 lb. =
Adding this constant shear flow to that of The calculations would be exactly like
Fig. A15.17 we obtain the values in Fig. those in example problem 1 of Art. A14.10 and
A15.19. will not be repeated here. Fig. A15.21 shows
the plotted results as recopied from Fig. A14.32.
R (the resultant) = 25 lb.
I-- t 100 lb.
Equating moment of resultant about B to I 7.5",
moment of shear flow system about B,
Re = 2MB
25e =0.648 (128.54 x 2)
Therefore e = 6.65 inches.
Thus shear center lies 6.65 inches above B,

and 4.17 inches to left of B as preViously
found.
Fig. A15-21
A15.8 SHEAR FLOW IN CLOSED THIN-WALLED SECTIONS. SHEAR CENT ER.
If ~Fx and ~Fz are considered for equil-

ibrium of external and internal loads, they
will be found to equal zero.
To check ZMy =
zero take moments about
some point such as C.
100 x 7.5 + 10 x 3.75 x 15 - 1.25 x 5

- 1.25 x 10 + 1.25 x 15 =- 187.5 in.lb.
Therefore to make ZMc =

0, a constant
shear flow q equal to M/2A =
(187.5/2 x 11 x 15)
= 0.57 Ib./in. is required. Adding this con-
stant shear flow to that in Fig. A15.21, we
obtain the final shear flow pattern of Fig.
A15.22. This final pattern is not much Fig. A15-23
different from that of Fig. A15.21, the
reason being that the location of the imaginary
cut to make q equal zero, was not far from the Solution 1. Using Section Properties and Ex-
true fact, since the final q in this panel was ternal Shears with Reference to
only 0.57. If we had started the solution by Centroidal Axes (The K Method).
assuming the web bc cut or qbc =
0, then the
correction constant flow that would be needed In Art. A13.8 the calculations for this
to satisfy ZM =
0 would have come out q = beam section gave -
-5.88, since this is the final q in web bc.
Since the shear flow which is a load on the Ix =81.18, Iz = 153.58, Ixz =- 21.33
cell wall influences the required thickness of
sheet required, it is good practice to try to Fig. A15.23a shows the location of x and z cen-
place the imaginary cut at a point where the troidal axes.
shear is near zero, so that preliminary z
estimates in routine design relative to shell
thickness required will be based on shear flow 1.0 I
~.074 I
values that are near the final values.
a b .50 L
---+-x 2.074
c I d I926
.80
--.1 I
1-110.667
~I' 40
r
.57 1.82 5.333 I
Fig. A15-23a
In Art. A14.8 of Chapter A14, the method

of solution was referred to as the K method.
The Shear flow equation (see Eq. 14 of Art.
A14.8) is,
qy = -(k;,VX-kl.Vz) ~xA -(k"VZ-kl.Vx) L:zA

Fig. A15-22
k 1. -- I XZ /1 XI z- I xz" = 81.18 x -21. 33 =
A15.8 Single Cell - Unsymmetrical - Multiple Flange.
153.58 - 21.33"
-21.33
Example Problem 1 =- .001775
12016
= Iz/IxI z - I xz " =153.58/12016 = .01279

Fig. A15.23 shows a 4 flange unsym-
metrical single cell beam carrying two external kg
loads as shown. This beam is identical to the k" = Ix/lxI z - I xz " = 81.18/12016 = .00674
one used in ex~~ple problem 1 of Art. A13.8
which dealt with bending stresses. For the given beam loading the external
shear loads at section abcd are,
Vz = 6000 lb., Vx =- 1600 lb.

,Al5.9
Substituting in the equation for qy as Adding this value of q to those in Fig.

given above, A15.24 we obtain the final shear flow resisting
pattern in Fig. A15.25.
qy = - [.00674 (-1600) - (-.00177 x 6000)]ZxA
Solution 2. Principal Axes Method
- [.01279 x 6000 - (-.00177) (-1600DzzA
The shear flow system can of course be
found by referring section properties and ex-
whence ternal shear loads to the principal axes of the
beam section. The equation for shear flow is
qy = 0.16 ZxA - 73.91 ZzA - - - - - - (4) (see Eq. 15 of Chapter A14),
In using equation (4) to compute the

shear flow pattern we will imagine top panel
qy =- r zp Z zpA -
xp
~Z xpA - - - - - (5)
zp
ab cut, thus making the shear flow qy = zero
in this panel. Subt. in (4) - - - - (The subscript p refers to principal axes.)
The section properties about the principal

Qac =0 + .16 x 1 (-5.333) - 73.91 x 1 x axes were computed for this same beam section
6.074 =- 449.78 Ib/in. on page A13.5 of Chapter A13. The values are:-
qcd - - 449.78 + .16 x .8 (-5.333) - 73.91 I xp =75.38, I zp = 159.34

x .8 x (-5.926) =- 100.05 Fig. A15.26 which was also taken from page
A13.5 shows the location of the principal axes
qdb - - 100.05 + .16 x .4 x 10.667 - 73.91 and the distances from the four flange members
x .4 (-5.926) = 75.80 to the principal axes.
qba =75.80 + .16 x .5 x 10.667 - 73.91

x .5 x 2.074 =0
Fig. A15.24 shows the plotted shear flow
reSUlts. This pattern satisfies ZF z 0 and =
ZFx = O. To check equilibrium of moments
1600=
- Vx
q=75.8
Fig. A15-26
Fig. A15-24 Fig. A15-25

Before substitution in equation (5) can be
made, the given shear loads Vz =
6000, and Vx =
-1600 must be resolved normal to the principal
about a y axis. Assume a y axis going through axes.
point d.
Vzp = 6000 cos 15 0 -15'-1600 sin 15 0 -15'
ZMct = 6000 x 8 - 1600 x 8 - 449.78 x 12 = 5367.9 lb.
x 16 = - 51160 in. lb. Vx p = -6000 sin 15 0 -15'-1600 cos 150 -15'

= -3121.8 lb.
Thus for equilibrium a moment of plus
51160 in. lb. is required. This is produced Vz Vx
by adding a constant shear flow q around the Hence ~
xp
= 71.21 and --p
I zp
= -19.59
cell walls, where
Subt. in equation (5),
q 2A = 251160
=~ x 160 = 159 • 88 lb. in.
qy = -71.21 Z zpA + 19.59 Z xpA - - - -(6)
(A = area of cell = 160)
Assume qy =0 in top panel abo

Solution. The K method of solution will
~c =-71.21 x 4.45 x 1 + 19.59 x (-6.74) be used.
x 1 = -448.92 = I xz = 36.41
k:L IxIz-Ixz" 186.5x431.7-36.41'"
qcd =-448.92 - 71.21 (-7.12).8 + 19.59
(-3.58).8 =- 99.41 ::: ~~i~~ = .0004598
qdb - - 99.41 - 71.21 (-2.90).4 + 19.59 .005452
x 11.80 x .4 ::: 75.65 lb/in.
qba =75.65 - 71.21 (4.82).5 + 19.59 k"

::: Ix
- IxIz-I xz '" = 186.5
79185 =
.002355
x 9075 x .5 =0 (check)
These shear flows are practically the qy = (k"Vx-k:LVz)~xA - (k ..Vz-k:LVx)~zA
same as obtained in solution no. 1 as recorded SUbstituting in above equation,
in Fig. A15.24. Discrepancies are due to
slide rule accuracy. qy -:: - ( .002355 x 400 - .0004598 x 1000) ~xA
For equilibrium of moments, take moments -( .005452 x 1000 - .0004598 x 400)~zA
about (b).
qy = -0.4822 ~xA -5.268 ~zA - - - - - (7)
2Mb =6000 x 8 - 1600 x 8 - 448.92 x 12
Since the value of the shear flow is un-
x 16 = -50993 in. lb. known at any point on the cell walls, it will
be assumed that the cell wall is cut between
A constant shear flow q around cell must flange members 1 and 10, thus making q zero on
be added to produce 50993 in. lb. for equil- the sheet panel numbered (1-10). Then using
ibrium. This balancing shear flow is, equation (7) the shear flow is calculated in
going clockwise around the panel. Columns 1 to
q =J:!..
2A = 250993
x 160 ::: 159 . 35 lb/in • 10 of Table A15.1 show the calculations in
solVing equation (7). For explanation on how
to determine sense of shear flows qy in Columns
which is the same as in solution no. 1. 9, 10, and 11, review Art. A14.6 of Chapter
A14.
Example Problem 2
Fig. A15.27 illustrates a typical single

cell wing beam with multiple flange members. The shear flow values in column 11 would
The external shear load on this beam section be the results if the external loads as given
is Vz =
1000 and Vx =
400 located as shown. were so located as to act through the centroid
of this shear flow force system. Since they
The internal shear flow resisting pattern will
be calculated. do not we will solve for the unbalanced moment
on the beam section about point (0) the cen-
This beam section is the same as that troid of the beam cross-section. The moment of
used in example problem 5 of Chapter A3, where the shear flow force on a sheet panel between
the calculations of the section properties any two adjacent flange members is equal in
were made. magnitude to q times double the enclosed area
formed by drawing lines from the moment center
The results were: (0) and the ends of the particular sheet panel.
Fig. A15.28 illustrates this explanation. The
I xz = 36.41 value m in column 12 of the table lists the
double areas of these various triangular areas.
9
Taking moments of all forces both external
7 8 and internal about point (0),
Vx=400# 5
r~ . x
ZM o = 1000x2+400x3+17123 = 20323 in.lb.
10
(17123 equals summation of column 13)
14 15 16
Thus for eqUilibrium a negative moment of -20323
z Fig. A15-27 is needed. This moment is provided by adding a
AI5.11
Table A15.1
1 2 3 4 5 6 7 8 9 10 11 12 13 14
Area Arm Arm qy = -0.4822 qy = -5.268 qxz = m Final

Member A Z ZA X XA L:ZA L:XA (L: xA) (L: zA) (col 9 + sq. in. qxz m q=
col 10) ~z -20.6
0 0 0 0 0 0 -20.6
1 0.14 4.396 0.615 -17.41 -2.437 0.615 -2.437 1.175 - 3.240 2.065
2 0.14 6.446 0.902 -13.54 -1. 896 1. 517 55.2 114 -18.53
2.089 - 7.991 5.902 44.2 261
3 0.38 7.396 2.810 - 9.11 -3.462 4.327 -4.333 3.758 -22.794 19.036 609
-14.70
4 0.17 7.766 1. 320 - 5.44 -0.925 5.647 -7.795 4.205 -29.748 25.543
32.0
38.2 976
- 1.56
4.94
5 0.17 7.946 1. 351 - 0.86 -0.146 6.988 -8.720 4.2'15 32.590 1075
6 0.17 7.896 1. 342 3.14 0.534 -8.866 -36.865 33.0 11. 99
7 0.17 7.696 1. 308 7.14 1. 214 8.340 -8.332 4.017 -43.935 39.908
47.393
33.2 1325 19.31
8 0.17 7.296 1. 240 11.74 9.648 -7.118
1.996 10.888 3.432 -50.825 40.2 1905 26.79
9 0.29 6.896 2.000 15.39 4.463 12.888 -5.722 2.470 -57.358 54.888 32.6 1789 34.29
0.318 67.576
16 0.35 -8.411 -2.964 15.39 5.386 9.924 -0.659
4.727
-67.894
-52.280 54.559
251. 8
46
1702
2509
46.98
33. 16
15 0.31 -8.224 -2.587 9.64 2.988 -2.279
14 0.31 -7.734 -2.395 3.32 1. 629 7.327 7.715
4.982 8.744
-3.720 -38.862 42.582
30.44
48.2
46.4
2052
1413
21.98
9.84
13 0.31 -7.004 -2. I'll -0.917 2.813 7.827 -4.216 -26.245
- 2.96 -3.774 -14.819 18.59 47.6 885
12 0.28 -5.554 -1. 554 - 9.11 -2.551 1. 259 5.276 - 2.01
11 0.17 -4.504 -2.544 - 6.632 9.176 36.6 336 -11. 43
-0.766 -13.54 -2.302 0.493 2.974 4.031 42.6 172
10 0.17 -2.904 -0.493 -17.51 -2.976 -1. 434 - 2.597 -16.57
0 0 0 0 0 -20.6
1
L: 17123
constant negative shear flow around cell where A15.9 Two Cell-Multiple Flange Beam. Symmetrical
magnitude equals About One Axis.
_ Ii _ 20323 _ Fig. A15,30 shows a two cell cantilever

q - 2A - 2 x 493 - -20.6 lb./in. beam with 10 flange stringers. The cross-sec-
tion is constant. Let it be required to deter-
(493 = area of cell) mine the internal shear flow in resisting the
1000 lb. load acting as shown. For simplifi-
Adding this constant shear flow to that in cation, the top and bottom sheet covering and
COllnru1 11, we obtain the final shear flow in the three vertical webs will be considered in-
column 14. Fig. A15.29 shows true shear flow effective in taking bending flexural loads.
pattern. Since the beam section is symmetrical about the
X axis, the beam will bend about this axis in
resisting the given external load. The moment
of inertia of the section about the X axis
equals 250 in. 4
.11 .12 ,13 ,14 ,15
Fig. A15-28
46.98
Fig. A15-30
Solution 1 (Without use of shear center)
The internal shear flow is statically in-

Fig. A15-29 determinate to the second degree, since the
shear flow at any point in each cell is unknown. zero. Thus the shear flow in the vertical web
Therefore, to make the flexural shear flow cc ' is determined by the stringer c alone,
statically determinate, a value for the shear namely
flow q in each cell will be assumed at some
point, and the flexural shear flow for each qcc' = -4Z~' zA = - 4x 1 x 5 = -20 Ib./in.
cell will then be calculated, consistent with
the assumed conditions. These resulting static We can now continue around cell (2) starting
shear flow systems will, in general, produce a with stringer c' where we were previously
different total shearing strain around the stopped.
perimeter of each cell, or in other words,
produce a different cell tWist. Since full
continuity exists between cells, this condition
qc'd' = qb'c' + qcc' 4Z~:zA
cannot eXist, and therefore an unknown constant =0 20 - 4 x (-5) 1 =0
shear flow of q in cell (1) and q 2 in cell (2)
must be added t6 make the tWist or both cells qd'e' =0 4 x (-5) 0.5 =10 Ib./in.
identical. This fact gives us the basis for
one equation and the other equation necessary qe'e =10 - 4 x (-5) 1 =30 Ib./in.
for the solution of the two unknowns q1 and q2
is given by the requirement of equilibrium, qed = 30 - 4 x 5 x 1 = 10 Ib./in.
namely, that the moment of the external and
the internal shear forces about any point in qdc =
10 - 4 x 5 x 0.5 = 0, which
checks
the plane of the cross section must equal zero. the assumed value of q = 0 in panel cd.
In Fig. A15.31 the flexural shear flow has
been assumed as zero just to the left of The shear flows in cell (2) could of
stringer c in cell (1) and just right of course been found by starting in panel cd
stringer c in cell (2). The balance of the where the shear has been assumed zero and
flexural shear system consistent with this proceeding clockwise around cell as for example
assumption is calculated as follows:
%e =0 - 4 x 5 x .5 =-10 Ib./in.
qee ' = -10 -4x5xl = -30 Ib./in.
qe'd' =-30 4x (-5) 1 = -10 Ib./in.

50#/in. - qd'c' = -10 4x (-5) 0.5 =0 Ib.lin.
The magnitude of the results are the same
as previously calculated but the signs are
L 10#/in. opposite. As emphasized previously the shear
flow calculated together with its sign is in
Fig. A15-31 the y direction or qy. The direction of the
The general shear flow equation is, shear flow along the cell walls in the xz
plane can be detennined by the simple rule as
qy =_ Vz ZzA =_ 1000 ZzA = -4ZzA given in Art. A14.6 of Chapter A14. To illus-
trate, refer to the free bodies in Figs. A, B,
Ix 250 and C. The shear flow q in panel ba was
Cell (1). Starting in panel cb where the calculated to be -10. ThiS value is shown in
shear has been assumed zero and proceding Fig. A with its calculated sense and acts
counter-clockwise around cell toward the intersection point (0). Then by
the rule of Art. A14.6, qba must also act
qcb =zero (assumed)
x
toward point (0), thus giving the sense as
shown in Fig. A.
%a = -4ZzA = 0-4x5xO.5 = -10 Ib./in.
To obtain the shear flow in web panel cc',
qaa' = -10-4x5x2 = -50 Ib.lin. the calculations started at point c and went
toward point c I . q on this panel was calcu-
Qalb' = -50-4x (-5) x2 = -10 Ib.lin. lated to be -20. The free body in Fig. B
shows this value of qy = -20, acting with its
qb'c' = -10-4x (-5) xO.5 = Olb.lin. true sense or acting toward point (0'). Then
qcc must also act toward point (0'), as
We cannot proceed beyond stringer c' be- z
cause there are two connecting webs with un- shown in Fig. B. The free body in Fig. C
known shear flows. We can get around this shows how the sense of qde was obtained.
x
difficulty by going back to stringer c, where
the shear flow on each side of c was assumed
A15.13
Fig. (A) Since there is continuity between cells, Q~ =

Ql1l' Also since area of each cell is the same,
Fig. (B)
A~ =
Al1l • Equating (1) and (2),
1533q - 1583q - 10840

~ "
=0 - (3)
One other equation is necessary to solve for

unknowns q~ and q" , and it is given by the
moments of the external and internal shear
forces about any point in the plane of the cross
section, which must equal zero for equilibrium.
Fig. (C) <:>
Take moments about pOint (b) of the shear
~!f7r7;~[ flow system of Figs. A15.31 and A15.32 and also
the external shear load of 1000 lb., which in
--
o.t>/L}LJ' ~
q=O -
d <Ix=IO
this case has no moment about our assumed
moment center.
mb = -50 x 10 x 5 + 20 x 10 x 5 + 10 x 30 x 15 + 200q~
Fig. A15.32 illustrates the unknown con-
stant shear flow systems q~ and q" which must + 200q" = 0
act on cells (1) and (2) respectively to pro-
duce the same cell tWist when added to the hence, 200q~ + 200q " + 3000 = 0 - - - - - - - (4)
snear flow system of Fig. A15.31. The sense
Solving equations (3) and (4) for q~ and q", we
of q~ and q" has been assumed clockwise or obtain
positive in each cell.
q~ =-4.07 Ib./in. q" = -10.80 Ib./in.
The final or true internal shear flow sys-
t---"7i
ql ~ t~--t·
tem then equals that of Fig. A15.31 plus that of
(I) ~ ~ r(2)
L-___ ~
Fig. A15.32 when q~ = -4.07 and q" = -10.80 Ib./
ft___ ~
in., which gives the shear flow diagram of Fig.
~ A15.33.
Fig. A15-32
The equation for the angular twist Q per

unit length of beam is,
L
2AGQ ~ ~qt
in Ib./in.)
using the shear flow values in Figs. A15.31
and 32, the angular tWist of each cell will Fig. A15-33
be calculated by substituting in the above
equation. Solution 2 (By use of shear center)
For cell (1) In this solution, we find the flexural
shear flow for bending about axis X-X without
2A GQ = - 10 x 5 _ 50 x 10 _ 10 x 5 + 20 x 10 + ~ + tWist. The centroid of this internal shear
~ ~ .03 .05 .03 .03 :DO system locates the shear center. The moment of
the external shear load about the shear center
3 :O~Og], - ~~'" Hence, 2A~GQ~ = 1200q~ - 333q" produces pure torsion on the 2 cell beam. Thus,
adding the shear due to this pure torsion to
- 6670 - - - - - - - - - - - - - - - (1) that of pure bending, we obtain the final re-
sisting internal shear flow.
For cell (2)
In bending about axis X-X Without tWist,
2A"GQ,,= 10 .~~ x 2 + 3~MIO _ 2~0310 + 3 :&008 +~ the shearing strain for each cell as given by
equations (1) and (2) must equal zero. Hence:
- ~g~~. Hence, 2A"GQl1l = -333q ] + 1250q ,,+ 1200 q~ -333q" - 6670 =0 (5)
4170 - - - - - - - - - (2 ) -333q~ + 1250ql1l + 4170 =0 - (6 )

Solving equations (5) and (6) for q:l. and q'a> the resisting internal torque must therefore
we obtain q2 =-
2.0 lb./in. q:l. =
5.00 lb./in. equal -3600. Therefore,
Therefore, taking these values of q:l. and qe
in Fig. A15.32 and adding the results to that qt(l) = .00254(-3600) = -9.17 lb./in.
of Fig. A15.3l, we obtain the shear flow
pattern of Fig. A15.34 which is the shear flow Solving for qt(2)
system for bending without twist about X axis.
The centroid of this shear system locates the lfa.o:l.Ae + a:l. 2AJ
shear center. qt (2) =2"[ 31190000 T
In Fig. A15.34,
1:.Q.. + 10 + J:.Q) 100 + J:Q. x 20~
ZV = 0 = -10 x 45 - 10 x 28 - 10 x 27 = - 1000
lb., which checks the external shear of 1000
=1
2 Q (.03 .03 .05)
31190000
.03 T= 00245T
.
lb. ZH =
0 by observation of Fig. A15.34.
hence: qt(2) = .00245x-3600 = -8.85 lb/in.
V= 1000#
Therefore, if we add to the shear flow sys-
tem of Fig. A15.34, a constant shear flow of
-9.17 lb./in. to cell (1) and -8.85 lb./in. to
cell (2), we will obtain the true internal re-
-45
sisting shear flow of Fig. A15.35, which checks
solution 1, any discrepancy being due to slide
rule accuracy.
Fig. A15-34
To find the horizontal position of the centroid

of the shear flow in Fig. A15.34 take moments
about point "a:
Z:t1a = 10 x 27 x 10 +10 x 28 x 20
+ 5 x 8 x 10 - 5 x 2 x 10 = 8600 in .lb.
hence, x = ~ggg = 8.6" to the right of web aa' •
Fig. A15-35
Torsional Deflections
The external shear load of 1000 lb. acts 5" to
the right of aa' , and therefore causes a moment The angular twist of each cell is the same.
The value of the angular twist Q per unit length
about the shear center equal to (8.6 - 5.0)
1000 = 3600 in.lb. To resist this torsional of the beam can be found using the shear flow
pattern of Fig. A15.35 which is the true result-
moment, a constant torsional shear flow qt(l) ant shear flow, or the pure torsional shear flows
and qt(2) must act on cells (1) and (2)
respectively. of qt(l) = -9.17 lb./in. and qt(2) -8.85 lb./in. =
may be used if desired.
The values of qt(l) and qt(2) can be found The results will, of course, be the same.
by using equations (29) and (30) of Art. A6.ll For example:
of Chapter A5. Thus
For cell (1) due to qt(l) = -9.17 lb./in.
and qt(2) =
-8.85 lb./in.
200A - ZqL - (9.l7xlO) 3+9.17XlO_8.85xlO

:I. - -t- - ( .03) .05 .03
= 8000
Cell (2)
200A = ZqL = (8.85 x 10) 3 + 8.85 x 10 9.17 x 10
2 t (.03) .04 .03
= 8000
Since the external torque equals 3600 in.lb.,
AI5.I5
Cell (1) Final stresses - Fig. A15.35 cell the same, since if any twisting takes
place, all cells must suffer the same amount.
2GGA _2:qL=(14.17 x 5)2 + (4.17 x 5)2 + 54.17 x 10 Furthermore, for equilibrium, the moment of
J. t ( •03 ) ( .03) .05
the internal shear flow system plus the moment
_26.68 x 10 = 8000 of the external shear load must equal zero •
•03
For bending about axis XX, the flexural
Cell (2) Final stresses shear flow will be assumed as zero at a point
just to the left of stringer a in cell (3) and
200A _ZqL_(l0.85X5)2+ (0.85x5)2+26.68 x 10 just to the left and right of-stringer c in
2 t ( .03 ) ( .03) .03
cells (1) and (2) respectively. One might con-
-19.15 x 10
.04
= 8000 sider the cells as cut at these three points.
Fig. A15.37 shows the flexural shear flow under
these assumptions. Since the leading edge cell
AI5.IO Three Cell - Multiple Flange Beam. Symmetrical (3) has no stringers and the covering is con-
About One Axis.
sidered ineffective in bending, the shear flow
Fig. A15.36 shows a 3-cell box beam sub- will be zero on the leading edge portion since
jected to an external shear load of 1000 Ibs. the shear flow was assumed zero just to the left
as shown. The section is symmetrical about of stringer a. The resulting flexural shear
axis XX. The area of each stringer is shown flow for the 3 cell section will therefore be
in parenthesis at each stringer point. The identical to Fig. A15.31 and the calculations
internal shear flow system which resists the for the flexural shear flow will be identical to
external load of 1000 Ibs. will be calculated those in Art. A15.7.
assuming that the webs and walls take no
bending loads, or, the stringers are the only
-
IO#/in.
effective material in bending. The moment of
inertia about the XX axis of effective material
equals 250 in ~ (Note: this beam section is
identical to the two cell beam of Fig. A15.30
plus the leading edge cell (3).
50#/in. 20#/in.
IO#/in.
-
.025 Fig. AI5-37
.05 .03 .04
-x Fig. A15.38 shows· the unknown constant
shear flows qJ,' qs' and q3 which must be added
(3) (1) 10" (2) to the flexural shear flow of Fig. A15.37 to
t make the tWist Q of each cell the same. The
L 5"__
I
~2.0)
a'
'
1(·5)
5"--1- 5 " _
b
.03 c'
1~1.0)
d'
1(·5)
5"(-J-5"~
.03 e'
1(1.0)
sense of each has been assumed po~itive in each
cell.
- t- --r
qi -.t tf ~---"""J----
....
Fig. AI5-36
1 1t
q3
q2
•
Solution No.1 (Without use of shear center)
tt •t •
(3) (1)
~t
(2)
•
The system is statically indeterminate, to H t •,
the third degree, since the value of the shear
flow q at any point in each cell is unknown.
The value of the shear flow will be as-

~t
- f......-....-.- .......... ~
~
-- - ---
......... ..- f
Fig. A15-38
sumed at a point in each cell and the flexural
shear flow for bending about the XX axis will The angular tWist Q for each cell equals Q =
be determined consistent with this assumption. 1
A constant unknown shear flow q J,' q 2' and q" 2AG 2: q Lit
for cells (1) and (2) and (3) respectively
will be added to the static flexural shear Using the values of q in Figs. A15.37 and
flow so as to make the angular twist Q of each A15.38, the value GG will be computed for each
cell.
Cell (3)
2QA G
3
= 2":qL
t
2Qx39.4G = 10x50
•05
+ 15.71 q + 10~ _lOqJ.
--:c525 3 •05 •05 22.91 #/in.!
or 00 = 10.5q21-2.55qJ. + 127 - - - - - - - - (1) (2)
-24.97 #/in.
Cell (1)
2QA G
J.
= 2":qL
t
-12.12#/in. -2.12#/in. -7.09#/in.
2QxlOOG = -50xlO_2(10x5) +10x20+3(10qJ.)
.05 .03 .03 .03 Fig. A15-39
10qJ.
+---:05-~-~
10q3 10q" h ence 00 -- 6qJ.-q3- 1 •67 q"
A15.11 Shear Flow in Beam with Multiple Cells. Method of
Successive Approximations.
- 33.34 - - - - - - - - - - - - - - - - - - (2)
The general trend in airplane structural
Cell (2) design appears to be to the use of a relatively
large number of cells. There are various
2QA G
"
=2":qL
t
reasons for this trend some of which are:
(1) using multiple interior webs, the detri-
2Q x 100 x G =_ 20 x 10 + 2 (10 x 5) + 10 x 30 + mental effect of shear deformation on bending
stress distribution is decreased; (2) the fail
.03 .03 .040
safe characteristic of the wing is increased
3xlO xq"+lOq,, 10q. hence 00 = 6.25q,,- because the wing is made statically indeter-
.03 .04 -.03
minate to a high degree and thus failure of
1. 67qJ. + 20.83 - - - - - - - - - - - - - - - (3) individual units due to fatigue or shell fire
can take place Without greatly decreasing the
Taking moments of the internal shear flow over-all ultimate strength of the wing;
systems of Fig. A15.37 and A15.38 and the ex- (3) the ultimate compressive strength of wing
ternal load of 1000 lbs. about stringer a and flange units is usually increased because
equating to zero:- - column action is prevented by the mUltiple webs
which attach to flange members.
ZMa = 10 x 20 x 10 + 10 x 30 x 20 - 5 x 1000
In Chapter A6, Art. A6.l3, the method of
+ 78.6q3 + 200qJ. + 200q" = 0 successive approximation was presented by de-
= 3000 + 78.6q3 + 200qJ. + 200q,,=0 - - (4) termining the resisting shear flow system when
a mUltiple cell beam was subjected to a pure
Solving equations (1) (2) (3) and (4) for the torsional moment. This method of approach has
unknown qJ.' q", q21 and 00 we obtain: now been extended to determine the resisting
shear flow when the beam is subject to flexural
qJ. - - 2.12 lb./in. bending without tWist*. Using these two methods
the shear flow in a beam with a relatively large
q" - - 7.09 lb./in. number of cells can be determined rather rapidly
q3 = -14.5 lb./in. as compared to the usual method of solving a
number of equations.
00 = - 19.9
Adding these constant shear flows to the flex- PHYSICAL EXPLANATION OF THE METHOD
ural shear flow of Fig. A15.37, we obtain the Fig. A15.40 shows a 3-cell beam carrying
true internal resisting shear flow as shown in and external shear load V acting through the
Fig. A15.39. shear center of the beam section but as yet un-
known in location. In other words, the beam
bends about the symmetrical axis X-X without
twist. The problem is to determine the internal
resisting shear flow system for bending without
* "The Analysis of Shear Distribution for Multi-Cell Beams

in Flexure by Means of Successive Numerical Approxi-
mations." By D. R. SAMSON. Journal of the Royal
Aeronautical Society, Feb. 1954.
A15.17
tWist. In this example, it is assumed that the Now consider each cell as a separate cell.
bending moment is resisted entirely by the The static shear flow qs acting on each cell
flange members as represented by the small will cause each cell to twist. Since zero
circles on the figure, which means that the twist is necessary a constant shear flow q~ to
shear flow will be constant between the flange cell (1), q~ to cell (2), and q~ to cell (3)
members. must be added as shown in Fig. A15.42, and the
magnitudes of such value as to make the twist
The first step in the solution is to make of each cell zero. However, the cells are
the structure statically determinate relative actually not separate but have a common web be-
to shear flow stresses for bending without tween adjacent cells, thus the shear flow q~
tWist. In Fig. A15.41 imagine each cell cut acts on web 2-1 which is part of cell (1), and
at points a, band c as shown. For the given thus causes cell (1) to twist. Likewise cell
shear load V, the static shear flow qs can be (3) is tWisted by q~ and cell (2) by both q~
calculated, assuming the modified section bends and q~. Therefore to cancel this additional
about axis X with no tWist. Fig. A15.41 shows cell tWist, we must add additional constant
the general shape of this static shear flow shear flows q~, q~ and q: as shown in Fig.
pattern. A15.43, and considering each cell separate
again. However, since the cells are not separ-
ate these additional shear flows effect the
twist of adjacent cells through the common web.
ell Cell Cell As before this disturbance in cell twist is
-x (1)- - (2) - -(3) - -x Fig. A15-40 again cancelled or made zero by adding further
closing shear flows q~, q:, q: as shown in
Fig. A15.44. This procedure is repeated until
the closing shear flows become negligible. In
general the converging of this system is qUite
rapid and only a few cycles are necessary to
give the desired accuracy of results.
Fig.AI5-41
The total closing shear flows qJ.' q2 and
q" are then equal to -
qJ. = q~ + q~ + q~ + ----
qs _
O-t
~I
----------t
~--
qs
----t
~--qr-
qs
Fig. A15-42
q2 =
q" =
q~ +
q~
q: + q: +
+ q: + q: +
'\ ~1lt tqi
t__ -J
t (2) }
~ ________tt qs
(3)
The final shear flow on any panel then
"<:::...........
'<--'
1
--- qs --- qs
equals, (See Fig. A15.45)
q = qs + qJ. + q2 + q" - - - - - - - - - (1 )
The centroid of this final shear flow
Fig.AI5-43 system locates the shear center of the section,
relativA to bending about the X axis.
DERIVATION OF EQUATIONS FOR USE IN SUCCESSIVE

APPROXIMATION METHOD
Fig.AI5-44 Fig. A15.46 shows cell (2) of the 3-cell

beam shown in Fig. A15.45. qs is the static
shear flow and q" ~
and Q" are the re-
qs -"'"
~ -~- -~-
4-- _ _ '-t-....-..-~
dundant or unknown
H q2 !~ q3 !l Fig. A15-45 shear flows. Since
cell (2) does not twist
i ql H H Hqs under these shear flows
~,--U f t..----- U we can write in general,
qs
~ --.........- ..........
qs qs 2:~ =0 - - - - - - (1) --- ~-
Substituting the vari- Fig. A15-46

ous shear flows on cell
(2) in Fig. A15.46 into eq. (1),
A15.18
L L L L and thus equation (7) becomes,

~
2
qs t- q a L;2 -t+ q 1 ~ -t+ q a 3-2
~-2
~ -t =0 - - - - (2)
The subscript (a) on the summation symbol q a =::::: q'a + C1_a (q ~ + q ~) + Ca-a (q ~ + q ~ ) (8 )
implies summation completely around cell (2)
whereas the subscripts 1-a and a-a implies Thus by repeating the above procedure, a
summation only along webs 1-2 or 3-2 respect- power series of the carryover influence factor
ively. L is the length of a sheet panel and t is obtained. In general the convergency is
its thickness. rapid and only a relatively few cycles or oper-
ations are needed for sufficient accuracy for
Solving equation (2) for q final shear flows. A solution of a problem
will now be given to show how the necessary
(~L)
operations form a very simple routine.
a T
L;qsL 1-a tL)
(L; a-at
~~ ~ L q1 +
+ L; L qa - - - (3) A15.12 Example Problem Solution. Problem No. 1.
a tat a t
Fig. A15.47 shows a cellular beam with
The first term in equation (3) represents five cells. The flange areas and the web and
wall thicknesses are labeled on the figure.
the proportion of the static shear flow qs
which must act as a constant shear flow around The problem will be to determine the internal
shear flow pattern when resisting an external
cell (2) to cancel the twist due to qs. The
resulting value of this first term will be
shear load of Vz =
1000 lbs. without tWist of
given the term q~. the beam. Having determined this shear flow
system the Shear center location follows as a
The second and third terms in (3) repre- simple matter.
sent the constant closing shear flows required
Fig. A15.48 shows the assumed static
in cell (2) to cancel the twist of cell (2)
condition for determining the shear flow system
due to the influence of q1 and qa in the ad-
jacent cells acting on the common webs between in carrying a Vz load of 1000 lb. without twist.
The static condition is that all webs except
the cells. The ratio in equation (3) before
the right end web have been imagined cut as
q1 will be referred to as the carryover
indicated thus making the shear flow q at
influence factor from cell (1) on cell (2) and these points zero. s
will be given the symbol C1- a , and the ratio
before qa in equation (3), the carryover
In this example problem it will be assumed
influence factor from cell (3) to cell (2) and
that the flange members develop all the bending
it will be given the symbol Ca_ a • Thus
stress reSistance, which assumption makes the
equation (3) can now be written as,
shear flow constant between adjacent flange
- - - - (4) members.
As explained above, q~ is the value of the The total top flange area equals 5.5 in. a ,
necessary closing shear flow for zero twist and also the total bottom flange area. Due to
when the adjacent cell shear flows are zero. symmetry the centroidal X axis lies at the
Hence first approximations to the final shear mid-depth point.
flows in each cell can be taken as neglecting
the effect of adjacent cells, or in other words Hence, Ix = (5.5x5 a )2 = 275 in. 4
each cell is considered separate. Hence the

first approximations are, qs = _.:!.:£ ~zA = - 1000 ~zA = -3. 636 ~zA
Ix 275
starting at the lower left hand corner,

(5 ) the static shear flow qs will be computed going
counter-clockwise around beam.
By SUbstituting (5) in (4) a second ap- Qab = - 3.636(-5)2 = 36.36 lb./in.

proximation for qa is obtained, namely, qbc = 36.36 -3.636(-5)1 = 54.55 lb./in.
(6) qcd = 54.55 - 3.636(-5)0.5 = 63.64
Continuing in like manner around the beam, the
(7)
values of qs as shown in Fig. A15.48 would be
where q~ is the correction added to the first obtained.
approximation. In a similar manner corrections
q~ and q~ are made to the approximations for
The solution from this point onward is
made in table form as shown in Table A15.2
q1 and qa • Therefore as a third estimate for
which should be located below a drawing of the
qa' these further corrections should be added
A15.19
cellular beam as illustrated, and the numbers

in the Table should be lined up with respect
to the cells as indicated.
C
,,-~
J~J..-~
~ L
= 200
856
= .2336
~ t
The solution as presented in Table A15.2
is carried out in 17 simple steps. The first We are now ready to start the solution
step as given in row 1 of the Table is to proper by successive approximations. In row 5
compute for each cell the value for ~ qs 1. , of the Table, the first approximation is to
t assume a value q' added to each cell which will
where qs is the static shear flow on each cancel the twist due to the static shear flow
sheet panel of a cell; L the length of the when the cell webs are not cut, but each cell
panel and t its thickness. Values for qs are is considered separate or independent of the
taken from Fig. A15.48. other. This constant closing shear flow q'
equals,
For example, for CellI L
~qSt
q' =----L--. The minus sign is necessary
~ qs ~ = 2(36.36 x 10) .~ = 18180 ~t
The sign is positive because qs is posi- because the tWist under the static shear flow
tive. (Clockwise shear flow on a cell is must be canceled. The values for q' are re-
positive.) Row 1 in the Table shows the values corded in row 5 of the Table.
as calculated for the 5 cells.
For example for cell (1),
The second step as indicated in row 2 of q' = _ 18180 = _ 21.238
the Table is to calculate the value of the 856
expression ~ Lit for each cell.
For cell (2),
For example, for CellI,
27275 -- - 28. 71
q I = - 95()
Steps 6 to 13 as recorded in rows 6 to 13

For cell 2, of the Table are identical in operation, namely,
the carryover influence from one cell to the
adjacent cell because of the common web between
~....!:...= ~+ 2(10) +.l2.. = 950 the cells. As a closing shear flow is added to
"t .05 .04 .04
each cell to make the cell tWist zero when they
The third step as indicated in row 3 is to are considered separate, this result is contin-
calculate the value for the Lit of the common ually disturbed because of the common web.
web between two adjacent cells. Gradually these corrective shear flows become
smaller and smaller until the cells reach their
For example, for web bb ' between cells true state and possess zero tWist. In the
(1) and (2), Table, arrows have been used for two cycles
to help clarify the operations.
[L)
\ t ~-2
=..l2....
.05
= 200 For example in row 6, the carryover shear
flow from cell (1) to cell (2) is,
The fourth step is to determine the
carryover factor from one cell to the adjacent - 6.700 x .2105 =- 1.414
cell. The results are recorded in row 4 of
the Table. From cell (2) to cell (1), the carryover value
is
Referring to equation (3) for general
explanation, the carryover factor from cell (-4.480 - 8.330) 0.2336 =- 3.000
(1) to cell (2) is,
From cell (2) to cell (3),
= (t) +_" 200 (-4.480 - 8.330) 0.250 =- 3.216

= .2105
~ L 950
" t
and the carryover factor from cell (2) to

cell (1) equals,
1f4------ i -------.t 1000 lb. =Vz

2 a', 1 b' •5 c , . 5 d' . 5 e' 1.f,
Fig. A15-47
Flange and
Web Data
---r 10" X .064
.04
Cell
(1) .05
.04
(2) .04
.04
(3) .04
.03
(4) .03
.03
(5) .03 X
~ 2a
.04
1b
.04
.5 c
.04
.5 d
.03
.5 e
.03
If
I 5 Cells at 10" = 50" ---------------11
Fig. A15-48 a' 36.36 b' 54.55 c' 63.64 d' 72.73 e' 81. 82 f
Assumed Static
Condition !or cut Cell Lcut • cut Lcut Lcut 100
Shear Flow qs 0 (1) 0 (2) 0 (3) 0 (4) 0 (5)
a 36.36 b 54.55 c 63.64 d 72 73 e 81. 82

Table A15. 2
Row ,OPERATION
1. ~ qs Lit for each cell 18180 I 27275 I 31820 I 48487 I 87880

2. ~ Lit for each cell 856 I 950 I 1000 I 1250 I 1333
3. Lit of cell web 200 250 250 333
4. Carry Over Factor (C) . 2105 I . 2336 .250 I .2633 .200 I .250 .250 I .266
5. 1st Approx. q' = -~qs L/t/~ Lit -21. 238....... _-28.71......... ......-31. 82 ....... ......-38.79--.... /-65.910
6. q" = Cq' (Carryover) - 6.700~-4.480 -8.38~7.170 -9.70otS6.364 -17.560~- 9.700
7. q'" = Cq" (Carryover) - 3.000"'- "'--I. 414 -4. 43()¥"-3. 216 -5.975""'-3.372 -2.580K' ~- 5.975
8. ETC (Carryover) - 1. 362 -0.631 -2.518 -1. 460 -1. 489 -1. 837 -1. 592 - 1. 489
9. Carryover - 0.736 -0.287 -0.777 -0.787 -0.872 -0.590 -0.396 - 0.872
10. Carryover - 0.248 -0.155 -0.436 -0.266 -0.246 -0.332 -0.232 - 0.246
11. Carryover - 0.138 -0.052 -0.135 -0.148 -0.141 -0.102 -0.065 - 0.141
12. Carryover - 0.044 -0.029 -0.076 -0.047 -0.042 -0.058 -0.037 - 0.045
13. Carryover - 0.003 -0.009 -0.003 -0.003 -0.003 -0.002 -0.012 - 0.003
14. q = q' + q" + q'" + - - -) -33.47---. _-52.52 ...... ___ -63. 39 ...... ___ -73.92 ---. ..... -84.38
15. 1st Reiteration -12.27 ----... -7.05 -16.6gx:13.13 -18. 4B><:12. 68 -22.44~-18.48
16. q' from row 5 -21. 238 -28.71 -31. 82 -38.79 -65.91
17. q -33.51-. _-52.45 ....... ..... -63.43, .......... -73.91_ ......-84.39
18. 2nd Reiteration 12.23"'-- -......-7.05 -16.7~13.10 -18.481><-12.69 -22.44~-18.48
19. q' from row 5 -21. 238 -28.71 -31. 82 -38.79 -65.91
20. q -33.47 -52.46 -63.40 -73.92 -84.39
Fig. A15-49
Closing Shear
t----~t r - - -- t~- -- --t~-- -- ---tt- - - - t
Flows to Make
Twist of Each
t~ q =33. 51
U
t~ q =52. 45
U
U q =63. 43
U
H q =73. 91
H
U q =84. 39
t
~
Cell Equal Zero
~----.. q tt ~q---.... u ~_~ __tt
2.85 2.10 0.27 1.18 2.57

Fig. A15-50
Final Shear Flows. 33.51 18.94 10.98 10.48 10.48 15.61
(Fig. A15. 48 plus Fig. A15.49)
2.85 2.10 0.27 1. 18 2.57

CALCULATION OF SHEAR CENTER LOCATION
In Fig. A15.47 let x : distance from left end of beam to shear center. Taking moments about
upper left corner of the shear flow forces in Figs. A15.48 and A15.49 and equating to 1000 X.
1000x = 10(36.36+54.55+63.64+72.73+81.82)10 + 100 x 10 x 50 - 2 x 100(33.51 + 52.45 + 63.43 + 73.91
+ 84.39), hence 1000x =
19472 or X 19.47 inches. =
A15.21
In row 13 of the Table, the carryover condition where the static shear flows in the
values are so small that the process is termin- webs should be much closer to the final values
ated. The final constant shear flow that must and thus hasten the convergency in the succes-
be added to each cell to cancel the tWist due sive approximation procedure.
to the static shear flow equals the algebraic
sum of the values from the beginning row 5 to Thus in Fig. A15.51, we have assumed the
row 14. The results are shown in row 14 of the top panel in each cell as cut to give the static
Table. condition. The static shear flow is now con-
fined to the vertical webs and zero values for
The results in row 14 are obtained after a top and bottom sheet. Table A15.3 shows the
considerable number of multiplications and calculations for carrying out the successive
additions of numbers, thus it is easy to make a approximation~ and needs no further explanation.
numerical mistake. To check whether any It should be noticed that after the first ap-
appreciable mistakes have been made, we take proximation was made in row 5, only three carry
the values in row 14 and consider these values over cycles were needed in rows 6, 7 and 8 to
of constant shear flow in each cell as that obtain the same degree of accuracy as required
causing zero twist if cells are separate. Then in 8 cycles in Table A15.2 for solution 1. Fig.
bringing the cells together, through the common A15.52 shows the final shear flows which equal
webs causes a disturbance in twist and this is the constant shear flows in each cell from row
made zero by the carryover values. This step 9 of Table added to the static shear flows in
in the Table is referred to as a reiteration Fig. A15.51. These values check the reSUlts of
and is indicated in row 15. Then adding the solution 1 as given in Fig. A15.50, within slide
values in row 15 to the initial approximation rule accuracy. In Table A15.3 no reiteration
q' in row 5, which value is repeated in row 16, steps were given. The student should make it a
we obtain the final value of q in row 17. The practice to use such checks.
values in row 17 are practically the same
magnitude as in row 14, thus no appreciable A15.13 Example Problem 2.
mistakes have been made. If the difference All Material Effective in Bending Resistance.
was appreciable, then a second, and if needed,
even a greater number of reiterations should The general trend in supersonic Wing struc-
be carried out. In the Table a second reiter- tural design is toward a large number of cells
ation is shown in rows 18, 19, 20 and the and relatively thick skins, thus in general, all
results in row 20 are practically the same as cross-sectional material of the Wing is effective
in row 17. in resisting bending stresses and thus the shear
flow varies in intensity along the walls and
It will be assumed that the solution was webs of the beam cells. Fig. A15.53 shows a ten
stopped after first iteration, and thus the cell beam with web and wall thicknesses as
values in row 17 are the constant shear flows shown. It will be assumed that all beam mater-
that must be added to the static shear flows ial is effective in bending. The shear flow
to produce bending Without tWist. Fig. A15.49 resisting system for bending about the horizontal
shows these final closing constant shear flows. axis Without tWist will be determined. The Cen-
Adding these values to those in Fig. A15.48 we troid of this system will then locate the shear
obtain the final shear flows in Fig. A15.49. center.
The lateral location of the shear center Fig. A15.54 shows the static condition that
for this given 5 cell beam coincides with the has been assumed, namely, that the upper sheet
centroid of the shear flow force system in panel in each cell has been imagined cut at its
Fig. A15.50. The calculations for locating midpoint, thus making the static shear flow zero
the shear center are given below Fig. A15.50. at these points. The static shear flow values
qs are shown on Fig. A15.54. To explain how
Solution 2 of Problem 1 they were calculated, a sample calculation will
be given.
In solution 1, the assumed static condi- The moment of inertia of the entire cross-
tion involved cutting all vertical webs except section about the horizontal centroidal axis is,
the right end web. Thus the static beam
section became an open channel section and the Ix for top and bottom skins,
resulting static shear flows must obviously be = (50 x 0.125 x 2.5 3 ,)2 = 78.0
far different than. the final true shear flow Ix of all webs = (0.912 x 5 3 )/12 = 9.5
values, since the webs always carry the greater Total Ix = 87.5 in 4
shear flows in bending without tWist. This
fact is indicated by the relatively large num- For convenience an external shear load
ber of steps required in Table A15.2 to reach Vz = 8750 lb. will be assumed acting on this
a state where successive corrections were small beam section.
enough to give a desired accuracy of final re- Hence, q = -:!z..."L,zA = - 8750 "L,zA =- 100 "L,zA
sult. Thus it is logical to assume a static Ix 87.5
Now consider Fig. A15.55 which shows a Starting at point (a) in cell (1) where the
sketch of cell (1) plus half of cell (2). As shear flow is zero and going counter-clockwise
previously explained the upper cell panels were around the cell, the static shear flows are
assumed cut at their midpoints (a) and (m). as follows: - (Cont. on PI. A 15.23 ----_J--I)
Solution II cut cut cut cut cut
0 0 0 0 0 0 0 0 0 0
Fig. A15-51 ,
Assumed Static
Condition for 36.36 (1) 18.18 (2) 9.09 (3) 9.09 (4) 9.09 (5) 18.18
Shear Flow qs and
Resulting qs Values.
0 0 0 0 0
Table A15. 3
Row OPERATION
1. L: qs Lit for each cell -2045 I 0

-1362 758 I I I 3031
2. L: Lit for each cell 856 I 1 950
1000 I 1250 1 1333
3. Lit of cell web 200 250 250 333
4. Carrv over factor C . 2105 1. 2336 .250 1.2633 .2001 .250 .2501 .266
5. 1st approx. q' =-L: qg L/t/2; Lit 2.389 1. 434 0 -0.606 -2.274
6. q" =Cq' (carryover) 0.334 0.503 0 0.358 -.152 0 -.605 -0.152
7. q'" =Cq" (carryover) 0.118 0.071 .055 0.126 -.152 0.041 -.041 -0.152
8. ETC (carrv over) 0.029 0.025 -.007 0.031 0 -0.005 -.040 0
9. a =a' + a" + am + - - -) 2.870 2.081 0.211 -1. 256 -2.578
2.87 2.081 0.211 1. 256 2.578

Fig. A15-52
Final Shear Flows
(Row 9 Plus Fig. AI5-51) 10.96 10.55 10.41 15.60
33.49 18.96
Compare Results with
Solution I (See Fig. AI5-50)
2.87 2.081 0.211 1. 256 2.578
Example Problem 2 Ten Cell Beam - All Material Effective in Bending.

Top Skin =. 125 Inches Thick
Fig. A15-53
00#'
ec (1)
00#'
(2)
00#'
(3)
00#'
(4)
00#'
(5)
00#'
(6) ~
... (7)
...
<Xl (8)
...
t- (9)
... T(10) ~ 5"
o '"o '"
0 '"
0 ~ '"
0 0 0
t-
0
Bottom Skin =. 125 Inches Thick 1--5"-1

1
-78 0 -
78 78 0 78 78 0 78 78 0 78 78 0 78 -78 0 78 -78 0 78 -78 0 78 -78 0 78 -78- - 78
-
0
-78 ~ -156:- -156 -156 -156 -156 -156 - -156 -156 -156 - ~78
Fig. A15-54 -98 185.4 185.4 185.4 185.4 185.4 181.3 181.3 178.5 178.5 98
-78 ~o-- - o + - - 0 - ~ o - - 0 - - 0 - -
-78 -78 78 -78 78 -78 78 -78 78 -78 78 -78 78
-0- -78 78
~o-
-78 78
-0- -0- -78 78 -78
Table A15.4
L:qs Lit 2200 I 0 I 0 I 0 I 0 -4760 I 1350 I 0 I 1190 I 0 I

L:L/t cell 211. 4 I 186.4 I 186.4 1 186.4 1 186.4 I 1 227.7 195 I 203.6 1 211.3 I 219
Lit web 53.2 53.2 53.2 53.2 53.2 61. 8 61. 8 69.5 69.5
C .2851.252 .2851.285 .2851.285 .2851.285 .2851.285 .3041.316 .2921.304 .3171.329 .3061.317
q' =-L: qg L/t/L: Lit -10.42 0 0 0 0 -6.92 0 -5.63 0 20.91
C.O. 0 -2.980 00 00 0-1.97 00 -2.10 1.71 00 -1.7 6.63 0
C.O. -0.75 00 -0.860 0-0.57 00 -0.56 -1.20 00 -1.11 1.60 0 1.48
C.O. 0 -0.21 -0.25 0-0.16 -0.250 -0.16 -0.50 00 -0.53 0.15 00 0.1 0.47 0
C.O. -Q 11 0-0.00 -0.13 -0.07 -0.05 -0.19 -0.070 -0.19 0.11 00 -&.11 0.20 0 0.19
C.O. -0.00 -0.03 -0.06 -0.00 -0.07 -0.06 -0.02 -0.07 -0.08 -0.020 -0.09 0.03 00 0.0 0.06 0
q -11. 28 -3.53 -1. 29 -1.14 -2.85 -9.00 -4.25 -5.05 5.56 22.58
Reiterate -0.89 -3.211-0.36 -1.011-0.32 -0.361-0.81 -0.321-2.56 -0.81-1.34 -2.741-1.53 -1.2411.83 -1.6017.15 1.70
q' -10.42 0 0 0 0 -6.92 0 -5.63 0 20.91
q -11.31 -3.57 -1. 33 -1.17 -2.88 -9.07 -4.27 -5.04 5.55 22.61
-89.31 _._
66.69 -81.57 74.43 -79.33 76.67 -79.17 76.83 -80.88 _ -82.27
75.12 -87.07 68.93 - __
73.73.83.04 72.96 -72.45 - 4 - _ ~
83.55 _55.39 100.61
109.31 177.66 183.16 185.24 187.11 191.59 176.50 182.07 167.91 161.44 120.61
Fig. AI5-57.
-89.31
Final Shear Flow Values.
66.69 -81.57 - 74.4' -79.33 76.67 -79.17 76.83·80.88 -
75.12 -87.07 68.93 -82.27 73.73 -83.04 72.96·72.45 83.55 -55.39 100.61
(Note: Shear Flow at Ends of Webs Equal Sum of Shear Flows in Adjacent Skin Panels.)
A15.23
c---5"-----+·+<I<:--2.5'L-..+i
10+1
Proceeding to (g)
.cut ,cut
TXT =qhg + 100ZgZA

b a h m
qg i= 156 + 100 (1.25 x 2.5 x
h
5"
.094) = 156 + 29.4 = 185.4 lb./in.
li·
c - - (1) - - - - g l - - ( 2 ) - -
qfg = 185.4 + 100ZfZA = 185.4 + 100 (-1.25)
g
5
(2.5x .094) = 156 lb./in.
d
'-=-_ _---..:e::-. ...:f..L...- k
qfk = qfg - qfe = - 156 + 78 = - 78
Fig.AI5-55
=0 k
Qa qkf =-78-100Z ZA =-78-100 (-2.5)(2.5x
f
qb =_lOm;bZA =-100x2.5x2.5xO.125
a 0.125) =-78+78 =0
= -78 lb./in.
Fig. A15.56 shows a plot of these calcu-
qc = qb - 100ZcZA = - 78 - 100 (1.25 x 2.5 lated values. The arrows give the sense of the
b shear flows.
x 0.064) =-78-20 =-98 lb./in. iM78~mm[II:Iii~~~~~mrn~
qd = - 98 _ZdZA = - 98 -100 (-1.25)(2.5 x
c
0.064) = -98 + 20 = -78 Fig.
98 (2) A15-56
qe =_78_Z e ZA =-78-100 (-2.5)(2.5 x
d
0.125) = -78 + 78 = 0
qfe. O_ZfZA = 0-100 (-2.5)(2.5 x 0.125)

e
78 = Fig. A15.54 shows the calculated static
At point (f) there are two other connecting shear flow values for the entire 10 cells. The
sheet panels so we cannot proceed past this values are recorded at the ends of each sheet
joint in calculating the shear flow in the two panel and at the midpoints of each sheet panel.
connecting sheets at (f). Clockwise shear flow in a cell is positive shear
flow. Since an interior web is part of two ad-
Thus we go back to point (a) and go jacent cells, the sign of the shear flow on
clockwise, vertical webs is referred to the left hand cell
in order to determine whether sense is pas it i ve
Qa = 0
or negative.
qha = 0 + 100ZhZA
a
=0 + 100 (2.5 x 2.5 x 0.125) Having determined the static shear flows
which will be referred to as qs' we can now
= 78 lb./in. start the operations Table A15.4. The first
horizontal row gives the calCUlations of the
With two other webs intersecting at joint twist of each cell under the static shear flOWS,
(f) the shear flow summation cannot continue which is relatively measured by the term
L
past (f), hence we go to point (m) in cell (2) ZqsTfor each cell.
where shear flow is zero due to the assumed
cut at point (m).
With all material effective in bending the
shear flow varies along each sheet. Fig. A15.56
Clm =0 shows this variation on the sheet panels of cell
qhm = O-lOOZhZA =-100(2.5x2.5xO.125) (1). The term Zqs L/t is nothing more than the
m area of the shear flow diagram on each sheet
=-78 divided by the sheet thickness. To illustrate,
consider cell (1) in Fig. A15.56.
Now at joint (h) we have the shear flow of 78
magnitude on each top panel, thus the shear Upper sheet panel: -
flow in the vertical web at (h) equals the sum
of these two shear flows or 156. Zq ..h = -(0 + 78) ~ +(0 + 78) 2.5 =0
s t 2 0.125 2 0.125
For lower sheet panel: - listed in last row in Table A15.4: -
M = +(-11.31 - 3.57 -1.33 -1.17 - 2.88 - 9.07

l:qs ~ = 0 (Same figure as for upper sheet) - 4.27 - 5.04 + 5.55 + 22.61) 2 x 25 = - 520
For left hand vertical web: - Total Moment =

175000 + 22520 - 520 = 197000
Treating the shear flow diagram as a hence x= 197000/8750 =
22.5 in., which equals
the distance from left end of beam to shear
rectangle with height 78 and a parabola with center location.
height 98 -78 = 20,
Zq .1=..=-78x5 _(20X5xO.667) =-7142 Referring to the final shear flow values
s t 0.064 0.064 in Table A15.57, it will be noticed that the
final results are not much different from the
For right hand web of cell (1), assumed static shear flows with the possible
exception of the two end webs. If we had
The shear flow diagram is likewise made assumed all the webs cut except one to form the
up of a rectangle and a parabola. static condition, then Table A15.4 would have
required several times as many carryover cycles
L 156 x 5 0 667 to obtain the same accuracy of final results.
ZqSt= 0.094 + (186.5-156)5X O: 094 =9342
A15.14 Use of Successive Approximation Method for Multiple
Therefore for entire cell (1) Cell Beams when Subjected to Combined Bending and
Torsional Loads.
Zqs ~ = -7142 + 9342 = 2200, Which is the
The internal shear flow resisting force
value in row (1) of Table A15.4 under cell (1). system for a beam subjected to bending and tWist-
The results for the other nine cells as calcu- ing loads at the same time is carried out in two
lated in a similar manner are recorded in row
1 of the Table. The procedure as followed in distinct steps and the results are added to
given the true final shear flow system. First,
the remaining rows of Table 4 is the same as the shear flow resisting system is found for
explained in detail for example problem 1. In being without twist as was explained in this
Table A15.4 only one reiteration is carried chapter. The results of this first step locates
through as the values in the bottom or last
row are practically the same as arrived at after the shear center. The external load system is
the fifth carryover cycle. Adding the con- then transferred to the sheat center, which
normally would produce a torsional moment about
stant shear flow values in the last row in the
Table to the static shear flow values in Fig. the shear center. The internal resisting shear
flow system to balance this torsional moment is
A15.54 we obtain the final shear flow values
then handled by the successive approximatIon
of Fig. A15.57. The resultant of this shear method as explained and illustrated in detail
flow pattern is a force of 8750 acting down
in the Z direction. Its location would be in Art. A6.13 of Chapter A6.
through the centroid of the shear flow force A15.15 Shear Flow in Cellular Beams with Variable Moment
system. Let X equal distance from upper left of Inertia.
hand corner of beam to line of action of shear The previous part of this chapter dealt
flow resultant force.
entirely with beams of constant moment of iner-
tia along the flange direction. In airplane
Taking moments of shear flow force system wing and fuselage structures, the common case
of Fig. A15.54 plus the constant shear closing
values in each cell as given in the last row is a beam of non-uniform section in the flange
of Table A15.4 and equating to 8750 x, we d1rection. In cases where the change of the
cross-sectional areas is fairly well distributed
obtain;
between the various flange members which make up
Due to uniform static shear flow on each the beam cross section, the shear flow reSUlts
as given by the solution for beams of constant
web: - moment of inertia are not much in error. For
M = (156 x 5)( 5 + 10 + 15 + 20 + 25 + 30 + 35 beams where this is not the case, the shear flow
+40 + 45) + 78 x 5 x 50 = 175000 results may be considerably different from the
actual shear flows. This fact will be illus-
trated later by the solutIon of a few example
Due to parabolic static shear flow in eachweb:-
problems.
M = (29.4x5xO.667)(5+10+15+20+25) A15. 16 The Determination of the Flexural Shear Flow
+ 25.3 x 5 x 0.667 (30 + 35) + 22.5 x 5 x 0.667 Distribution by Considering the Changes in Flange
(40 + 45) + 20 x 5 x 0.667 x 50 = 22520 Loads. (The 1l.P Method. )
Due to constant closing shear flows as Fig. A15.58 shows a single cell distributed
A15.25
flange beam. Consider the beam acts as a canti- flange members, see Fig. A15.60. Again ZF y = 0,
lever beam with the bending moment existing at
section (A) being greater than that eXisting at whence ~Pa + ~P4 + qyd =0
section (B) and that the bending moment produces
(~Pa + ~P 4)
compression on the upper surface. By the use of or qy =- d
the flexural stress equations, the bending
stress on each stringer can be found, which if
Therefore starting at any place where the
multiplied by the stringer area gives the value of qy is known, the change in the average
stringer axial load. Thus at beam section (B), shear flow to some other section equals
let P~, P 2 , Pa' etc. represent the axial loads
due to a bending moment M. The external bending
qy =- Z ~ - - - - - - - - - -(1)
If the summation is started where qy is zero

then equation (1) will give the true average
shear flow qyo
Fig. A15.61 shows sheet panel (3,3' ,4,4' )

y isolated as a free body. Taking moments about
corner 4' and equating to zero for equilibrium,
Lx ~
~!4'
-- d(~P3)b
whence, qx
d- - qxbd -- 0
= ~Pa /d - - -(2)
Thus for rectangular sheet panels between
flange members the shear flow qx or qz equals
the average shear qy.
Fig.AI5-58 The same rules as previously presented to

determine the sense of qx or qz after having qy
can be used and will not be repeated here.
To show that equation (1) reduces to the

shear flow equation preViously derived and used,
consider a beam with constant cross-section and
take a beam length d = 1 inch. Then,
~ =Vzd = Vz(l) =Vz

~P =~ ZA =.Y.z... ZA (where A = area of
Fig.AI5-59 Fig.AI5-60 Fig.AI5-61
Ix Ix stringer)
moment at section (A) is M + ~, hence the From equation (1) qy =- Z~P

stringer axial loads at section (A) will equal
SUbstituting value of ~P found above,
P, + ~P" P 2 + ~P2' Pa + ~Pa' etc. These
stringer axial loads are shown on Fig. A15.58. V
qy = -~ZZA - - - - - (3) which is the
Ix
Imagine the upper sheet panel 2, 2' , 3, shear flow equation previously derived for beams
3' is cut along line (a-a). Furthermore con-
with constant moment of inertia.
sider stringer number (3) cut out and shown as
a free body in Fig. A15.59. Let qy be the
average shear flow per inch over the distance A15.17 Example Problem to Compare Results in Using
Equations (1) and (3).
d on the sheet edge bb. It has been assumed
positive relative to sense along y axis. Fig. A15.62 shows a square single cell beam
with six flange stringers. Between points Band
For equilibrium of this free body, C, the beam has a constant flange section which
is shown in Section B-B. The numerals beside
ZF y = 0, hence ~Pa + qyd =0 each stringer represent the area of the stringer.
Between points Band D, the flange material
whence qy =- ~Pa/d tapers uniformly with the flange material at
point A as indicated in Section A-A. It should
For a free body including two stringers or be noticed that the increase in flange area is
A1S.26 SHEAR FLOW IN CLOSED THIN-WALLED SECTIONS. SHEAR CENTER.
with the shear flow system of Fig. A15.63, which
(l :: lO~rD
therefore is the final shear flow system for
this method of sOlution.
I-- 20" -+- 30" -----1

o
U'>
II
C'
~ ~q= iO~
/~~~J
~
o
U'>
II
C'
f5"
-.l
2 1 2
-r
x 10"
Da
1 1
b
1
T10" Fig. A1S-63 Fig. A1S-64
c' a' b' J. c' 'b' -L Solution No.2, Considering ~P Loads in Flange
h"+s"~ 1rs,,-+s,,~1 Stringers. (Equation 1)
SECTION A-A
I x = SxS 2 x2 =2S0in4
SECTION B-B TO C-C
I x =3 xS 2 x2= lS0in'l:
Bending moment at section AA = 1000 x 50 =50000
in. lb.
Fig. A1S-62 Bending moment at section BB = 1000 x 30 = 30000
obtained by increasing only corner stringers b in.lb.
and c. The shear flow on section A-A will be
computed using equation (3) which applies only Considering Section B-B:
to beams with constant section and also by Bending stress intensity at midpoint of
equation (1) which applies to beams with vary- stringers by the flexural formula:
ing moment of inertia.
a -~ - 30000 x 5 - 1000 psi.
b - Ix - 150 -
Solution 1. Using Shear Flow Equation for
Beams of Constant Cross-Section. Axial load in each of the stringers a, b, and
(Equation 3) c = 1000 x 1 = 1000 lb.
Since q at any point on the cell is un- Considering Section A-A:
known, it will be assumed that the upper sur-
face on Section A-A is cut through the midpoint - Mz _ 50000 x 5 - 1000 i
ab - IX - 250 - ps •
of flange stringer (a), thus making the shear
flow qy equal to zero on this free surface. Axial load in stringer (a) = 1000 x 1 = 1000 lb.
One-half of stringer (a) thus acts with each
side of the top surface. In this solution the Axial load in stringer (b) or (c) = 1000 x 2 =
webs and walls will be assumed ineffective in 2000 lb.
resisting bending stresses, thus the shear flow
is constant between adjacent stringers. These resulting axial loads are shown act-
ing on the portion between points A and B in Fig.
Starting at midpoint of stringer (a) and A15.65, which equals the results as shown in Fig.
going counter-clockwise around cell, A15.66.
qa = 0 (assumed cut)
V 1000
Qac = --?-'L.ZA
x
= - 250 x5xO.5=-10 lb/in
QcA = -10
1000
qcc' = -10- 250 x5x2=- 50 lb/in.
Proceeding around the cell the balance of the
shear flow c·ould be calculated, but due to sym-
metry enough values have been found for the
shear flow to draw the complete shear flow pic-
ture for bending about the X axis when it is Having found the ~P flange loads over a
assumed that one-half of the area of stringer length (d) of 20", the shear flow can be com-
(a) acts with each adjacent web. Fig. A15.63 puted by equation (1).
shows the resulting shear flow diagram. The
resultant of this shear flow pattern is a It will be assumed that one-half of the ~P
1000 lb. force in the Z direction and its load in stringer (a) will flow to each adjacent
location through the midpoint of the box since wet. However, there is no ~P load in stringer
the flow is symmetrical. The external load of (a) hence Qab =
Qac =
O. Then from equation (1),
1000 lb. also acts through the midpoint of the ~?,... 1000
cell hence the external load is in equilibrium qbb' =
O-'L. T = O--mr-= - 50 lb./in.
A15.27
6Pb' -1000
qb I a I = -50 + Z-d- = -50 - 2 0 = 0
Due to symmetry the left side of cell would

give the same results. The results are plotted
in Fig. A15.64. Since the increase in section
moment of inertia between beam points B and A is
increasing at the same rate as the external
bending moment, the average shear of 50 Ib./in.
is constant between the two beam section A and
B. Comparing Figs. 63 and 64, we find the
first method gives a shear flow of 10 Ib./in. Fig. A15-70
in the top and bottom webs, whereas actually it _ Vd .. V
is zero. This seems reasonable since the entire CIa - bId b1
increase of flange area was placed in stringers
band c.
Example Problem No.2
The same beam as in Problem 1 will be used

except that the cross-sections at beam points B Fig.AI5-71
and A are as shown in Fig. A15.67. The increase
in flange area between beam points B and A has whence q~ Vb"
b~
=
been placed entirely in stringer (a) which
changes from 1 sq. in. at B to 3 sq. in. at A.
But V = q"b", whence
1 3 1
c b
1
c
1
a
1
Il q~ =( ~: )" q" - - - - - - - - - - - - - (4)
Section Ix = Ix = Section Fig.AI5-67
A-A 250in 150 in 4 B-B From Fig. A15.71
c' a' b' c' a' b'
I 1 % =~ =~ hence q = b~Qa
b~ b~' "b"
The results of using equations (3) and SUbstituting value of q in (4)
(1) relative to the shear flow pattern are
given in Figs. A15.68 and A15.69 respectively. q~=(~:)Qa--- - - - - - - (5)
(The stUdent should check these results.) It
should be noticed that the true shear flow is Thus having the shear flow on the stringer
greater in the top and bottom skin than that edge of the sheet panel, the shear flow on the
given by equation (1) which applies only for large end of the tapered panel can be found by
beams of constant cross-section. equation (5).
A15.19 Example Problem of Shear Flow in Tapered
Multiple Flange Single Cell Beam.
o
Il>
~q=3~ 00
Il>Il>
o
Il>
"0' +-i.=.~ ~
1111
0'0' "0'
Fig. A15.72 shows a tapered single cell
beam with 6 spanwise stringers or flange members.
The beam is loaded by a 1000 lb. load located as
Fig.AI5-68 Fig.AI5-69 shown. Assuming the webs ineffective in bending
A15. 18 Shear Flow in Tapered Sheet Panel.
the internal resisting shear flow pattern will
be determined.
Major aircraft structural units such as
the wing, fuselage, etc., are tapered in both In this solution the shear flow at Station
plan form and depth and therefore the sheet 120 will be determined by considering the 6P
panels between flange members usually are flange loads over a length of 30" or between
tapered in width. Fig. A15.70 shows a canti- Stations 90 and 120.
lever beam tapered in depth and carrying a
load V at its end. The flange reactions at Consider section at Station 120:-
the left end have been found by statics. A
free body diagram of the web is shown in Fig. Bending stress ab --1-
- Mc _ 1000 x 120 x 5 = 1333.33
- 450
A15.71. Take moments about point (0) and psi.
equate to zero.
The horizontal component of the axial load
mo =(~~) b" - q~b~d = 0 =
in a stringer equals aoA (where A area of the
stringer) •
Since panel ab is tapered in Width from 6"

Fig. A15-72 at station 120 to 5.5" at station 90, the shear
flow qX(ab) at station 120 can be found from
T 3 . 2i a I'l" c 1.75
equation (5).
8.5"
L-
Ix=271
a' b' c'
Sec. Sta. 90 qX(ab) = 24.87x5.5/6 =- 23.2 lb./in.
3.25 2.5 .75
qy(bC) -- 24.87 - 471/30 = - 40.57
Sta.90
1000# f----8" ----l -- 40.57x5.5/6 = - 37.2
t qX(bC)
a' b' c' I

~=---
~a 0
11
T Ta' ~1
1
1
f 11
1
qy(cc' ) =
=
40.57 -196.6/30
47.12x8.5/10 =
=- 47.12
40.0 lb./in.
---.L 90" Sec. Sta. 0 qZ(cc' )
4 k"~6"~ 2 12 O
Sta'1--_ _ _ 120'
Since the ~P loads are the same on the
1
Sec. Sta. 120 lower stringers but tension the shear flow cal-
hence, the horizontal components, of stringer cUlations if continued would give the same
loads are, values as found on the top surface. Fig. A15.74
shows the shear flow pattern on station 120.
Pa = 4x 1333.33 =5333.33 lb.
IN PLANE FORCES PRODUCED BY INCLINATION
Pb =3 x 1333.33 = 4000 lb. OF FLANGE MEMBERS
Pc =2 x 1333.33 = 2666.7 lb.
Since the box tapers in depth and Width, the
Consider beam section at Station 90:- flange stringers are not normal to section 120,
thus X and Z force components are produced on
1000 x 90 x 4.25
crb = 271 = 1411.5 psi. section by the stringer loads.
hence stringer loads are, These in plane force components are:-
Pa = 3.25 x 1411.5 = 4587.4 For stringers a and a' ,

Pb =2.50 x 1411.5 = 3528.7 Px = 5333.33 x 2/120 = 88.9 lb.
Pc = 1.75 x 1411.5 = 2470.1 Pz =
5333.33 x 3/120 = 133.3 lb.
The change in axial load ~P in the For stringers band b' ,
stringers between stations 90 and 120 equals
the difference between the above loads, whence
Px = 4000 x 0/120 =0
~Pa = 746, ~Pb = 471, ~Pc = 196.6 Pz = 4000 x 3/120 = 100 lb.
Fig. A15.73 shows these ~P loads acting For stringers c and c' ,
on wing portion. Since shear flow is unknown
at any point, assume q equals zero in web aa' .
The average shear flow in each sheet panel Px = 2666.7 x 2/120 = 44.4 lb.
=
over a length d 30 inches.can now be calcu- Pz = 2666.7 x 3/120 = 66.7 lb.
lated by using equation (1).
Fig. A15.74 shows these in plane force
:Z~P
qy =Qa' a - - d - qa'a = 0 (assumed) components due to the flange axial loads.
1000#
- 24.87 lb./in.
133.3 t 100 66.7
88.9_
!.-3"-+1! ~
-444
a -23.2 - -G
- -b ~37.2 .
1
10"
t
~
40.0~ Fig. A15-74
j 88.9
I~'
~
23.2 b' 37.2 : ,
- - - - - 44.4
6" 1 6" l
133.3 1(~0 66.7
Fig. A15-73
/
471
A15.29
The forces in Fig. A15.74 will be checked Taking moments about point a of the forces
for equilibrium. in Fig. A15.74 but replacing the shear flows in
the top and bottom panels by the values found
ZF z = 1000 - 266.6 - 200 -133.4 -10 x 40 =0 above, we would obtain an unbalanced moment of
7970 in. lb. instead of 8670 previously found.
ZFx =0 by observation. The correcting shear flow would then be q =
-7970/240 =
-33.2 instead of -36.1 as previously
Take moments about stringer (a), found. The final shear flow pattern would be
as shown in Fig. A15.76, which values should be
ZMa =- 3 x 1000 + 200 x 6 + 133.4 x 12 + 10 x 40 compared to those in Fig. A15.75.
x 12+(+88.9 - 44.4)10 + 23.2 x 6 x 10
+ 37.2 x 6 x 10 = 8670 in. lb. 15.4 2.1
For equilibrium a 33.2 6.8 Fig. A15-76

moment of -8670 is 15.4 _ 2.1
needed, which is
provided by a constant 12.9 1.1
shear flow q =
M/2A =
-8670/240 =
-36.1 Ib./in. 36.1 3.9
A15.20 Problems
Adding this constant
shear flow to that of 12.9 1.1
Fig. A15.74 gives the 1000 r"--t 2000 lb.
final shear flow as
shown in Fig. A15.75.
Fig. A15-75
f-4"1_ p j-12~'-j-12:'-1
T =.===03===;ij
rr=1
10,7 +-.05 -.<172 12'-r .... •072

I
uti4 .. 0.4
• __ -.04
Solution No. 2
L [.03 _ L }
This same beam and loading will now be f-----.!- 20'~ t.03
solved using the shear flow equation derived Fig. A15-77 Fig. A15-78
for beams of constant cross-section.
(1) Determine the resisting shear flow pattern
Since the stringers are not perpendicular for the loaded single cell beam as shown in
to the beam cross-section they have a z force Fig. A15.77. Assume load P = zero in this
component which thus assists in carrying the problem. Assume all material effective in bend-
external shear load in the z direction. These ing. Make two solutions, one of them involving
Pz components at station 120 have been calcu- the use of the shear center.
lated in the other· solution.
(2) Same as problem (1) but add load P = 1000 lb.
ZPz =- (2x133.3+2xlOO+2x66.7)=-600 Ib
(3) Fig. A15.78 shows an unsymmetrical single
Total Vz (external) =1000 lb. cell beam loaded as shown. Assume all material
effective in bending. Determine resisting shear
Let VZ(net) be shear load to be taken by flow diagram.
cell walls, 400 lb.
5 L."
4
tD'~
VZ(net) = Vz -ZPz = 1000 - 600 = 400 lb.
Calculation of static shear flow assuming Fig.
.04 .03
q in sheet panel aa" is zero. 10" A15-80
qy(ab) = VZi~et) ZZA = 400 x5x4 =17.8 ~ B Fig.A15-79
450 lb. in.

r.4 sq. in.
This corresponds to value of 23.2 in (4) Figs. A15.79 and A15.80 show two loaded
previous solution (see Fig. A15.74). single cell - 2 flange beams. Assume the flanges
develop all the bending resistance. Determine
qy(bc) = 17.8+~gx5x3 = 31.11b./in. the shear flow resisting pattern by two solu-
tions, namely, without and with use of shear
As compared to 37.2 in Fig. A15.74. center.
qy(cc') = 31.1+~gx5x2 = 40.0 Ib.in. (5) Fig. A15.81 shows a single cell - 3 flange
beam subjected to loads as shown. Assume the 3
Which is the same as in Fig. A15.74.
flanges develop the entire bending resistance. (11) Fig. A15.84 shows a multiple flange-circu-
Determine the internal shear flow resisting lar beam section. Find the resisting shear
system. flow pattern when carrying the external shear
OO#
¥
300 lb. load of 5000 lb. as located in figure. Assume
Q~'A
0
100 lb. L1 cell skin ineffective in resisting beam bending
45 stresses •
•04 .02
fL\'02 10" t1000 lb.
l-_~
•8--,"" 1~2
L1---'1B
.03.4r}
- l
c---2"
r- ~500 lb. 0.5
f V= 3000 lb.
r--10'~ .1
r 6 " ---j
.2
Fig. A15-81 Fig. A15-82 -,..
03 .03 6"
"'--it-----i.-._---i..----4t--__ ----L
.3 .1 .1 .1 .2
(6) Find shear center location for beam in
Fig. A15.81 if the 3 flanges provide the entire
bending resistance.
r-t 0',10''l
I- 4 panels a 5" = 20" 'I
Fig. A15-85 Fig. A15-86
(7) Find the internal resisting shear flow (12) Determine the shear flow resisting system
pattern for the 3 flange-single cell beam of for the beam section of Fig. A15.85. The 6
Fig. A15.82. Assume webs or walls ineffective flanges have areas of 0.2 sq. in. each. Skin
in bending. is .032 thickness. Assume skin ineffective in
bending.
(8) Determine shear center location for beam
of Fig. A15.82. Webs and walls are ineffective (13) Find the shear flow resisting system for
in bending. the unsymmetrical beam section in Fig. A15.86.
t
2000#
Flange areas and skin thicknesses are given on
figure. Assume skin ineffective in bending •
.3 .4
~~041
Top Skin .03 (14) Determine shear center location for beam
section in Fig. A15.86.
Fig. A15-83 .035 10"
Bottom Skin .03 I r "-i 1000 f 500 lb.
$02
5
#
.3 .4
~ 6 panels at 5" = 30"
---1 T8" .06"
.05
.03
[:303
.04
.-l .05 .03
(9) Fig. A15.83 shows a multiple flange-single
cell beam section. Find resisting shear flow
1--12"---+-8" --1
system if webs and skin are ineffective in Fig. A15-87 Fig. A15-88
resisting bending stresses. All skin flange
members have area of 0.1 sq. in. each. (15) Fig. A15.87 shows a 2 cell beam section.
Consider all material effective in resisting
(10) Find shear center location for beam bending stresses. For the given beam loading
section in Fig. A15.83, if webs and skin are determine the internal resisting shear flow
ineffective in bending. system.
(16) Find shear center location for beam section

18"15000 lb.
in Fig. A15.87. All material effective in bend-
ing.
-. +3"
5" (17) For the 2 flange-2 cell beam in Fig. A15.88,
.---+ All flange areas determine the resisting shear flow pattern when
.---+
5" = 0.1 sq. in. beam section is loaded as shown. Webs are in-
5" each. Skin
x- .---+ thickness = .035 effective in bending. •
5"
7" (18) Fig. A15.89 shows a 4 cell beam section
with 6 flange members. Assume walls and webs
6" ineffective in bending.
(a) As a first problem assume that the left

Fig. A15-84 and right curved sheet panels are removed,
A15.31
1 8" 11000 lb.

flow equation for
I 10 "
.03
.5 0 "
.03
.50 " beams of constant
section. Compare
the results. All six
04 Cell Cell Cell Cell 03
12" . (1) .05 (2) .04 (3) .04 (4) . flange members have
~ .03 .03
Fig. A15-92 0.2 sq. in. area
each at section A-A
10" .50" .50"
and tapers uniformly
1-12"--t- 8"--'j to 0.1 sq. in. at
end C-C.
Fig. A15-89
leaving a 2 cell beam section. Find the

resisting shear flow system under the
given loading.
(b) Now add the left curved sheet to form

cell (1) thus giving a 3 cell beam. Find
shear flow system.
(22) Add two interior webs to the beam of Fig.
(c) Now add the right curved sheet to A15.92, connecting flange members a-a" and b-b",
form cell (4), thus giving a 4 cell beam thus making it a 3 cell beam. Find the shear
section. Find the shear flow system. flow resisting pattern at section A-A by the
~p method.
~ i .t V = 2000 lb.
I' Top Skin .04
Fig. A15-90
.04 .04 '~"
(19) Fig. A15.90 shows a 10 cell multiple Fig. A15-93

flange beam section. Area of each of the 22
flange members equal 0.3 sq. in. Assume webs
and skin ineffective in resisting bending (23) Fig. A15.93 shows a circular single cell
stresses. Find internal resisting shear flow beam with 8 flange members. The area of each
system for 2000 lb. shear load acting through flange member is 0.1 sq. in. throughout the beam
shear center of beam section. Find location length. For the given 400 lb. external loading
of shear center. For solution use method of determine the resisting shear flow pattern at
successive approximation. section A-A using the ~p method over a distance
of 25 inches between sections A-A and B-B.
Assume cell wall ineffective in resisting bend-
ing stresses.
(21) Fig. A15.92 shows a single cell-6 flange

tapered beam carrying a 1000 lb. load as shown.
Calculate resisting shear flow pattern at
section A-A by two methods. (1) By ~p method
over a distance of 20 inches between sections
A-A and B-B, and (2) By using general shear
A15.32 SHEAR FLOW IN CLOSED THIN-WALLED SECTIONS. SHEAR CENTER
DOUGLAS DC-B JET AIRLINER. BASIC FUSELAGE CONSTRUCTION

CHAPTER A-16
MEMBRANE STRESSES IN PRESSURE VESSELS

ALFRED F. SCHMITT
A16.1 Introduction. of adjacent meridian curves and a pair of ad-

The structural designer is often called jacent parallels.
upon to develop a vessel which is to contain a
fluid under pressure. Occasionally the design
of such a vessel is not critical from either a
weight or shape standpoint and almost any suit-
ably strong sealed vessel will suffice. More
meridian
often, the strength, weight and form of such a curve
unit are closely prescribed and rigidly con-
trolled. ThUS, the pressurized cabin of a
modern aircraft is a sealed pressure vessel con-
taining an atmosphere at near sea level pres-
sures and Whose functional requirements include:
i-the transmittal of heavy loads from
the tail surfaces and from internal dead loads,
ii - the necessity for nonstructural cut-
outs for doors and windows,
iii - an efficient shape from both the aero- Fig. A16.1
dynamic and space utilization points of View,
iv - a minimum of weight. The radii Rm and Rt shown on the figure
are found by erecting local normals to the sur-
Structurally, the most efficient form of face of the element at its corners. Rm is the
pressure vessel is one in which the lateral radius of curvature of the meridian curve: it
pressures are supported by tensile stresses may be either positive (inward pointing), nega-
alone in the curved walls of the vessel. Ex- tive (outward pointing as in Fig. A16-1), or in-
amples of such shapes are those assumed by finite (at inflection points or straight-line
pressurized rUbber balloons and canvas fire meridian segments). Rt is the radius Df curva-
hoses and by the free surface of a drop of ture of the section normal to the meridian
water (in which the surface tension forces pro- curve. For simple forms of pressure vessel Rt
vide the support). The walls of these vessels is always positive; all radii Rt point inward
have zero bending stiffnesses and hence have the and intersect the axis 0-0, although not gen-
properties of a membrane. The stresses develop- erally normal to 0-0 (see radius Rt erected from
ed, lying wholly in tangential directions at point C of Fig. AI6-1).
each point, are called membrane stresses. Fig. A16-2 is a detail of the surface ele-
In shells of technical importance, the ment. The forces per unit length* in the merid-
walls do, of course, have some bending stiffness ional and tangential directions are denoted by
and hence may carry some transverse loadings by lim and Nt, respectively. Shear stresses are
flexural stresses. Indeed, the boundary con-
ditions imposed on the shell may be sucll as to
necessitate some localized bending near edges
and seams. An efficient pressure vessel design
is one in which the configuration minimizes these
departures from a true membrane stress system,
i.e., minimizes the degree of local bending
N 1o -
--.-.
::p
stresses induced.
A16.2 Membrane Equations of Equilibrium: Shells of
Revolution Under Rotationally Symmetric Pressure
Loadings. Nt - - - tUm", - Nt
P Fig. A16.2
Consider the equilibrium of a differential
element cut from the shell of revolution of Fig.
A16-1. (The figure is drawn to resemble a
familiar folding paper Christmas bell, since R
t
such an object may aid in visualization.) The * Hereafter referred to as "stresses" although their
element is cut out by the intersection of a pair units are pounds per inch rather than psi.
Al6.l
A16.2 MEMBRANE STRESSES IN PRESSURE VESSELS
absent due to the symmetry of the problem. The its support, is neglected.) In the lower hem-
included angles between the pairs of meridional
and tangential forces are dsm/Rm and dStIRt,
respectively.
Summing forces normal to the differential
element, one has
o o
or,

p - - - - - - - - - - - - - - - - (1)
Here p is the internal pressure, positive out- ispherical portion meridional stresses are re-
ward. Note that the shell wall thickness does quired as shown in Fig. A16-4b. Hence, in this
not enter eq. (1). The pressure p may vary in class of problems it is best to derive the nec-
essary second equilibrium equation (correspond-
the meridional direction but is constant in the ing to eq. 2) by considering the individual
tangential direction by hypothesis (rotational characteristics of the structure.
symmetry assumed).
Eq. (1) is one equation containing two un- A16.3 Applications to Simple Pressure Vessels.
knowns. Another equation may be obtained by the
condition of equilibrium of a portion of the Example Problem 1
shell above or below a parallel circle. Thus in Determine the membrane stresses in a cylin-
Fig. A16-3, the pressures acting downward on the drical pressure vessel of circular cross section
lower portion of the shell are equilibrated by (radius Ro ), having hemispherical ends, if the
the upward vertical components of the meridional internal gas pressure is p. Also find the
stresses, Nm. greatest combined normal stresE.
Fig. A16.5
Fig. A16.3 SOLUTION:
Summing forces vertically In the hemispherical ends Nm = Nt by

symmetry and, of course, Rm = Rt = Ro •
Hence eq. (1) is sufficient to determine the
stresses in this portion of the structure. One
Solving, has
Nm Nt
2-=2-=p
Nm = p Rt (2 ) Ro Ro
2
P Ro
Eqs. (1) and (2) determine completely the Nm = Nt = -2-
membrane stress state in the rotationally sym-
metric shell problem: the problem is thus seen Nm Nt P Ro
to be statically determinate. Stress =t = t = 2 t
We note that eq. (2) should not be used in
cases of hydrostatic pressure loadings. The In the cylindrical portion the radii are
basic concept is that of shell equilibrium, and Rm = 00 (the curve of Fig. A16-5 is the meridian
consequently for this class of problems the curve and this is a straight line for
manner of shell support must be considered. the cylindrical portion).
ThUS, in the tank of Fig. A16-4a, the upper cy-
lindrical portion requires no meridional stress-
es since the load is reacted at the supporting
ring 0-0. (In these analyses the structural Eq. (1) becomes
weight, which always requires some stresses for Nm + Nt =p
00 Ro
Al6.3
Nt = P Ro From eq. (1), with Nm = Nt = Nb'
__ Nt __ P Ro 2.7 R o p
stress Nb = 2
t t
Therefore
The meridional stress in the cylindrical portion
is found from eq. (2):
P Ro
= 1.25 Ro P
2
Finally, the compressive ring load, F, is (ref.
stress =P R o
Fig. A16-6c)
2t
Since shear stresses are absent everywhere 2F = J:l.25 Ro P sin g Ro d g = 2.5 R~ P

in the meridional and tangential directions,
these are the directions of principal stresses.
Hence the greatest combined normal stress is In computing reinforcing ring stresses
identically equal to the greatest meridional or from this result it is necessary to include
tangential stress as just computed. It is seen some effective skin from the adjacent shell
to be walls when the ring cross sectional area is
figured.
smax = p Ro/t •
Example Problem 3.
Example Problem 2. Another form of bulkhead used to close a
An important problem in pressurized cabin circular pressure cylinder is elliptical in
design concerns the shape of the end bulkheads. section as shown in Fig. A16-7. Such a bulk-
While hemispherical bulkheads (such as used in head shape provides tangential meridional forces
Example Problem 1) are highly desirable from a at the seam (requiring no reinforcing ring as in
stress standpoint, such forms are uneconomical the last example) and yet is reasonably efficient
as regards space utilization. On the other ex- as regards space utilization. Problem: de-
treme, a flat bulkhead, while providing far more termine the membrane stresses in such a bulk-
useful volume, cannot resist the pressure load- head.
ing by membrane stresses and hence is structur-
ally inefficient. A compromise configuration is
that shown in Fig. A16-6, in which the bulkhead
is a spherical surface of low curvature,
(ftdished head") supporting the pressure loading
by membrane stresses. A reinforcing ring,
placed at the seam, resists the radial component
of these stresses. Problem: find the com-
pressive load acting in the reinforcing ring. Fig. Al6.7
Solution: In cartesian coordinates the equation

of the bulkhead meridian is
F
(0)
From the calculus, the radius of curvature of
this meridian curve is
Fig. Al6.6
Solution: as shown in the exploded view, Fig.

A16-6b, the radial components of the bulkhead =
stresses are resisted by the reinforcing ring,
the cylinder wall being presumed to offer no
resistance to concentrated transverse forces. The radius Rt is found most readily by ob-
For this case 9= arc sin 2~7 = 21.7°. serving that it is normal to a tangent to the
meridian curve (see figure). After finding the
slope of the tangent (angle ~), one computes SOLUTION:

Rt = x/sin~. The results are
From eq. (1) the "hoop" loadings in the
bllx upper and lower cylindrical lobes are
sin ~ =""(~a--=4-y-;;Il~+~b""4-x-;;1l7)'],'7nll
(a 4y ll + b 4x ll )],/1l
Rt = bll
The meridional stress is found from eq. (2). Summing forces horizontally at the floor joint:
Thus,
Nf =p (RU cos a], + RL cos all) •
Assume all stresses are equal and are given by
s.
SUbstituting the expressions for Rm, Rt and Nm
into eq. (1) one finds Su =s = p RU/tU
sL = s = P RL/tL
Sf = s = t:-
(RU cos a], + RL cos all)
f
Of particular interest are the stresses at Letting the weight denSity of the material
the seam. Here y = 0 and x = a. One finds be w, the weight per unit length (axially) along
the cabin is (w times the developed length of
walls and floor).
Nm = P2a
w = W\ 2 tuRU (n - a],) + 2 tLR L (n - all)
Nt = pa (1 - 2~:) + 2 tf RU sin a], ~ •
This last result is important since it in- Solving for the various thicknesses from above
dicates that compressive tangential stresses are and substituting, one finds (to obtain a result
possible i f a> b...[2. As will be seen below, symmetric in appearance use was made of the
such a situation is undersirable because of the
large resultant difference in radial expansions
fact that RU sin a], =
RL sin all)
between the cylinder and bulkhead (the bulkhead
actually contracts radially if Nt is negative) W =~ lRU (n - a], + ~ sin 2 a],) +
producing high secondary bending stresses in the
vessel walls. RL (n - all + ~ sin 2 all) ~
Example Problem 4. Since one may show that the cross sectional area of this
Determine the weight per unit length of the fuselage is
double-cylinder fuselage cross section as a
function of the internal pressure, allowable A = Ru(n - a], + -} sin 2 a],) t Ri. (n - Gil + -} sin 2 as}
stress and the geometric parameters of Fig.
an important consequence of this calculation is that the ratio
A16-S. For structural efficiency it is desired of shell weight to shell volume is
to maintain equal membrane stresses in the skin
and floor. w _ 2wp
A --s-
and is therefore independent of the combinations of RU, RL,

a], and a II used. The designer is thus free to choose these
shape parameters so as to satisfy other requirements.
Example Problem 5.
Determine the membrane stresses in a
conical vessel of height h and half apex angle
Fig. A16.8 a. The cone forms the bottom of a large vessel
filled With a liquid of specific weight wand
having a head of liqUid H above the cone. The
complete-unit is supported from above.
A16.5
these secondary stress peaks, argUing that local

yielding of the material will level them out.
However, such stress peaks may prove to be of
great importance in cases of repeated loadings
wherein fatigue failure ~ considered likely.
To point up the major weakness in the mem-
brane analysis one need only compute the radial
displacements in the two different elements that
make up the pressure vessel of Fig. A16-5, viz;,
Fig. A16.9 the cylinder and the hemispherical bulkhead. By
Hooke's law, the tangential strain in the cyl-
inder is
SOLUTION: (fl. =
Poisson's
ratio, =
.3 for
The meridian curve of the cone has a aluminum)
=
radius of curvature Rm 00 and, at any point a
distance y down from the top of the cone, the
radius Rt is
R - (h - y) tan a By integration the radial displacement of any

t - cos a point on the cylinder is seen to be (ref. Fig.
A16-10) .
Then from eq. 1, at any level y
N - R - w (H + y) (h - y) tan a
t - P t - cos a
To find the meridional stresses Nm in the

R
cone, the equilibrium of a segment of height Fig. A16.10 .85 ~
h - Y is considered (Fig. A16-9b). Summing Et
forces vertically,
Adjacent to the seam, the tangential strain
Nm • 2 n (h - y) tan a cos a = in the hemispherical bulkhead is
n (h - y)2 tan ~ a • w (H + y) + P Ro
= .35 p Ro
E:
tBLKHD -- 2Et (1 - rII) Et
+ W 3 (h - y) 2 tan 2 a • (h - y)
Hence, by integration
Solving
P R~
= .35 E"t
1\1
""ID.
= w-2 -
tan-a
cos a (h - y) [ H + Y + 1
- (h - y) ]
3 Thus the cylinder tends to expand more than
the bulkhead - a situation prevented by the seam
A16.4 Displacements, Boundary Conditions and Local
between these elements. It follows then that
Bending in Thin Walled Shells. the seam experiences a transverse shearing action
It is appropriate at this point to examine as indicated in Fig. A16-11. These shear forces
some of the foregoing illustrative cases to de- in turn produce bending moments in the shell
termine whether or not the membrane stresses wall as shown on the figure.
computed gave satisfactorily accurate measures
of the shell stresses. Anticipating the answer,
we state that, while the membrane analysis will °blkhd
~ive the primary stress system in a shell-like
pressure vessel, a careful (and often lengthy)
analysis of induced bending caused by boundary
effects will reveal localized secondary stress
peaks. In static strength analyses of properly
designed* vessels it is the practice to neglect
* various codes and standards give proportions of common Fig. A16.1l

vessels which will correctly limit secondary stresses.
See .for example reference (1).
While it is not our purpose here to take up shell bending 13% above the membrane stress (as against only
in detail, some indication of the character and magnitude of 3% above for the hemispherical bulkhead) (ref-
these bending stresses should be available to place them in
proper perspective. The most striking thing about these wall
erence 2, p. 410). For this type of bulkhead,
moments is that they are quickly damped out, becoming neg- boiler codes sometimes permit a ratio of alb as
ligibly small (down to 1'10 of their maximum value) at a distance high as 2.6, however.
of about 4VRc;t from the seam. Thus, for an instance, in a
circular cylindrical shell of 40" radius and. 065" wall thick- A16.4 Special Problems in Pressurized Cabin Stress
ness, these moments are so damped at 6.5" from the seam. Analysis.
The next important consideration is an appreciation of Because of functional requirements over
the magnitude of these secondary bending stresses. For the and above those of a simple pressure vessel,
case of the pressure vessel of Fig. A16-5, the meridional the pressurized cabin shell of an airplane has
stresses are increased about 3<Y,0 at the point of maximum a number of stress analysis problems peculiar
moment, while the tangential stresses are increased only to its configuration. Several of the more
about 30/0, Fortunately, in this class of vessel, the tangential general of these will be considered here.
stresses are the ones designed by (they are twice as great as
the meridional stresses) and hence the secondary stresses
have little importance for this case (see Chap. 11, pp. 389-<;22 DISTRIBurION OF STRESSES BETWEEN SHELL
of reference 2). In other configurations one is not always so
fortunate, and detailed analysis may be required. (see refec-
AND STRINGERS.
ences 3, 4, 5 and 6). To stabilize the shell wall in transmitt-
The situation at the seam of the above ing heavy tail loads through the fuselage,
vessel is typical of many seams or boundaries longitudinal stringers are added. These
where elements are joined which would experience same stringers will also help to carry the
different expansions if loaded separately. meridional pressure loads. The skin and
Among such seams and boundaries are those: stringers must, of course, have equal strains
in the longituditudinal_ directions but, because
i-where the meridional curvature changes the skin is in a two-dimensional state of stress,
abruptly. It changes from Rm Ro = they cannot have equal longitudinal stresses:
to Rm = QO at the seam in Fig. A16-5. hence the following analysis.
ii - where a SUdden change in direction of Let the meridional (longitudinal) stresses
the meridian curve occurs. In Example in the skin and stringers be 3m and sL, respect-
Problem 5, above, considerable shell ively. St will be the tangential (hoop) stress
wall bending would be induced near the in the skin. From eq. (1) we again have
seam. In fact, a reinforcing ring
would probably have to be added at the
seam as was done in Example Problem 2, St = P-t-
Ro
above.*
iii - at which structural members of differ- If N is the total number of stringers, each of
ent stiffnesses and different loadings cross sectional area AL' then eqUilibrium longi-
join. In Example Problem 2, the cyl- tudinally reqUires
inder tends to expand the most, the
bulkhead quite a bit less and the re-
inforcing ring, being loaded in com-
pression, tends to contract. Other
seams and/or boundaries of this type The condition of equal longitudinal strai~ in
are those where an abrupt change in the skin and stringers yields
shell wall thickness occurs (addition
of a doubler) or where a shell is fast-
ened into a foundation.
where ~ is Poisson's ratio (= .3 for aluminum).
Good design tends to minimize the magnitude
of the secondary bending stresses by avoiding SolVing these three equations one finds
combinations of elements which would have highly
incompatible distortions. ThUS, the analysis of
Example Problem 3 shows that i f one closes a
circular pressure cylinder with an elliptical
bulkhead in Which a = 2 b, compressive tangen- = P Ro (1 + 2 ~ a) (1 + .6 a)
tial stresses would develop in the bulkhead. In 2t (1 + a) (1 + a)
such a case the bulkhead would tend to contract
radially while the main cylinder would tend to P R o (1 - 2 j.l) _ p R o .4
expand as always. Thus, the shear and induced sL = 2 t (1 + a) - 2t (1 + al
moment at the seam would be aggravated, prodUC-
ing (as it happens) a tangential maXimum stress where a =
NAL/2n Rot is the ratio of total
stringer area to skin area. A little study will
show that t(l + a) is a sort of "effective shell
* certain details of the design of such reinforcing rings are wall thickness": it is the result of taking all
given in the codes and standards. the cross sectional area (skin plus stringers)
A16.7
and distributing it uniformly around the peri- ations of these effects in a specific case are
meter. On this basis, the results are a little to be found in references 2 (PP. 395-406), 6
disappointing: the stringers are carrying only and 8.
40% of the stress one might expect if the net One proposed solution to the ring-shell
longitudinal load (p n R~) were distributed interaction problem is the floating skin.
evenly over the entire cross sectional area Basically, the idea is to reduce the radial
(2 n Rot (1 + a)). stiffness of the connection between the shell
The meridional skin stresses are reduced by and the rings so as to allow the shell to ex-
the factor (1 + .6 a)/(l + a) from what they pand freely under the pressure loading. The
would be without the stringers. For structures connection between ring and skin must still re-
of usual proportions this decrease may amount to tain its shear stiffness so that ring loads may
20 to 30% but clearly can never exceed 40%. In- be transferred to the shell wall by tangential
asmuch as the bending stresses due to tail loads shear flows. Fig. A16-13 shows the basic idea
will be superposed on these pressure membrane of the radially flexible connection. Many
stresses, the reduction is certainly beneficial*.
INTERACTION BETWEEN RINGS AND SHELL.
Because of the necessity for transmitting
various concentrated loads from within the cabin Fig. A16.13 Ring skin cross
and from the wings and tail to the main shell section showing the action of
and because it is also necessary to provide some a radially flexible connection.
lateral restraint which will stabilize the
stringers and skin against an overall instabilit
failure, the pressurized fuselage of an airplane
contains a considerable number of rings and variations of this type of "mount" suggests
frames distributed along the length of the shell. themselves, some of which may have merit for
These rings are seldom, if ever, spaced closely other reasons. For instance, the transmittal
enough such that they can be considered effect- of wind and other vibration noise into the
ive in carrying a part of the hoop stresses (in cabin of a high speed transport is a problem
the way the stringers were effective in carrying which might be treated simultaneously by the
part of the meridional stress). Rather, they proper choice of connection between the ring
act more like widely spaced restraining bands and the shell.
having the effect shown exaggerated in Fig. DOORS AND WINDOWS.
AI6-12.
The various cutouts in the shell of a
pressurized cabin reqUire special considera-
tion if an excessive weight penalty is to be
avoided.
Consider the panel removed from the pres-
surized cylinder of Fig. A16-14a. Following a
Fig. A16.12 Restraining rings along a press-

urized tank. The action is representative of a
fuselage with widely spaced rings inside.
It is obVious that the rings in this case

will produce secondary bending stresses in the
skin and hence may have a detrimental effect on
the simple membrane stress system. Equally
harmful are the tensile loadings developed in
the rivets joining the skin and rings. Detailed
I
analyses which will permit quantitative evalu- I
Fig. A16.14 (b) ~
* If one looks at the problem from the point of view of a /
stiffened shell, loaded primarily by bending and shear loads
from the tail, on which the pressure membrane loadings are common practice in dealing with cutouts, we
to be superposed, an interesting effect appears. Because determine what forces the panel-to-be-removed
the internal pressure tends to stabilize the curved skin applies to the main structure around the border
panels on the compression side, the effective width of skin of the cutout, and then superpose a set of
acting with the stringers is increased. The section proper- equal but opposite, self-equilibrating stresses
ties of the cross section may then change in such a way as
to produce little or no variation in the maximum fiber
to cancel these. The cutout border is then un-
stresses. Indeed, the maximum tensile stresses may actu- stressed and the panel may be removed without
ally be reduced by the addition of the internal pressure load- disturbing the new stress system in the main
ing! (see reference 7). structure.
A16. B MEMBRANE STRESSES IN PRESSURE VESSELS
In examining the figure to determine what Fig. A16-16 shows the typical cutout
sort of canceling stress system must be supplied, structural arrangement. While analytical
we see that the tangential hoop stresses border- approaches have been tried, it is probably
ing the cutout cannot be canceled by a self- safe to say that the true elastic stress dis-
equilibrating set since they have a radial com- tribution in such a configuration cannot be
ponent. However, the radial component of these computed. The necessity for avoiding high
stresses will actually be supplied by the door intensity stress concentrations (With their
or window pressing outward against its frame. attendant fatigue likelihood) makes empirical
Hence, it is only the component of the hoop information most useful in such cases. On
stresses along a chord which need to be canceled the other hand, a simple rational analysis,
(Fig. A16-14b)*. based on principles outlined above, will very
The immediate problem becomes one of de- likely suffice for a static strength check and
signing a structure to effectively support a set for most design purposes.
of uniformly distributed self-equilibrating The above discussion has concentrated
stresses acting in the plane of the chord con- attention on the problems of carrying the hoop
necting the upper and lower edges of the opening stresses around a cutout. The longitudina-l---
(Fig. A16-15a). pressure stresses, while being smaller them-
selves, are intensified by bending stresses
from the tail loads. Hence, the longitudinal
stresses across the cutout may make this con-
dition (or the combination) most severe.
(a)
Fig. A16.15
All that appears necessary to support the

stress system is to provide horizontal headers
at the top and bottom of the cutout, Which, as
beams, will carry the loads across to the sides
of the frames Where the loads cancel (Fig. A16-
15b). For cutouts of usual sizes in pressurized
fuselages, the stress system to be supported in
this manner is qUite large and it proves un- Fig. A16. 16 Structural arrangement
economical to design a single horizontal frame around a cutout. Most or all of the
member of sufficient bending stiffness to resist shaded skin area would probably be
them. Instead, the shell wall itself is em- doubled.
ployed to help carry these loads across. The
skin is used to form a beam of considerable
depth, the skin being the web of this beam, with
the horizontal frame member and one or more LARGE DEFLECTIONS OF PLANE PANELS; "QUILTING".
longitudinals forming the beam flanges (Fig. A16- The use of flat skin panels in a pressur-
15C) • ized fuselage cannot always be avoided. Since
Because of the heavy shear flows and direct the thin skin has little bending stiffness, it
stresses developed, the skin is usually doubled cannot support the lateral pressure as a beam
in this region. Additional stringers may also ("plate", more correctly) and hence must deflect
be added to relieve the stresses. The rings to develop some tensile membrane stresses which
bordering the cutout (and forming part of the will then carry the loading. The resultant
frame) are extended some distance above and be- bulges of the rectangular skin panels between
low the cutout proper (unless they coincide with their bordering stiffeners give a "quilted" ap-
a regular ring location, in Which case they pearance to the surface.
carryall the way around). Even in the case of curved skin panels
quilting will occur: if the internal stiffening
* Clearly one of the design requirements will be to make the framework (transverse rings and frames and
frame sufficiently stiff in bending against radial forces so longitudinal stringers) is relatively rigid and
that the door or window can bear up evenly against the is everywhere tightly fastened to the Skin,
frame. then each skin panel is restrained along its
A16.9
four sides (borders) against the radial expan- REFERENCES

sion normally associated with the shell membrane
stresses. The result is a sort of three-dimen- (1) API-ASME* Unified Pressure Vessel
--- --
Code
sional-case of the behavior depicted in Fig. 1951 Edition, et. seq.
A16-12* •
From a structural viewpoint, the unfortun- (2 ) Timoshenko, S. "Theory of Plates and
ate aspects of quilting lie in the high concen- Shells"
tration of stresses occurring near the panel McGraw-Hill, N.Y., 1940
edges and in the tensile loadings on the rivets (3 ) Watts, G. and Lang, H., Stresses in a
which join shell to stiffeners. The aerodynam- Pressure Vessel With a Conical Head,-
ic characteristics of a quilted surface are Trans. ASME, vol. 74,-1952, pp. 315-324.
highly undesirable in a high performance air-
plane; hence again qUilting is to be avoided. (4 ) - - - - - - - - - - - , Stresses in a
Computations of stresses in quilted panels, Pressure Vessel With a Flat Head Closure,
inasmuch as they involve large, nonlinear de- Trans. ASME, vol. 74,-1952, pp. 1083-1090.
flections, are difficult. An additional (and (5 ) - - - - - - - - - - - , Stresses in a
qUite necessary) complication is that of having Pressure Vessel With a Hemispherical-Head,
to introduce the stiffness properties of the Trans. ASME, vol. 75,-1953, pp. 83-89-.---
bordering members. The reader is referred to
Chapter A.17 for a further discussion of the (6 ) Roark, R. J. "Formulas for Stress and
problem. A simplified approach, indicative of Strain", MCGraw-Hill, N. Y. 3rd Edition,
trends, is given there along with further ref- 1954.
erences to the literature. (7 ) Howland W. and Beed, C. Tests of Pres-
surized Cabin Structures, Jour~ Aero.
A16.5 Shells of Revolution Under Unsymmetrical Loadings. Sci. vol~Nov. 1940.
Problems in which the shell of revolution
experiences unsymmetrical loadings are not un-
common in aircraft structural analysis. The
nose of a fuselage, the external fuel tank and
the protruding radome are shells of revolution Designs by Krafft
which may be loaded unsymmetrically by external Ehricke of Convair
aerodynamic pressures. Again, the same external for Space Travel
fuel tank shell receives an unsymmetric internal
hydrostatic pressure load from th~ weight of
fuel directed normal to the shell aXis.
Because of the unsymmetry of the problem,
membrane shear stresses are now present and so
the analyst must solve not two, but three equa-
tions in three unknowns (Nm, Nt and Ns )' More-
over, these become differential rather than
algebraic equations.
Because the derivation of the differential
equations of equilibrium is rather lengthy, and
because their general solution cannot be written
(rather, only specific solutions for certain
cases may be be found), no details are repro-
duced here. The reader is referred to pp. 373-
379 of reference 2 for the derivation of the
equations and for an example problem.
Outer Space Vehicles will Present Many New Problems to

* One design which reduces quilting in thecurved skin, the Aeronautical Structures Engineer.
fastens rings and frames to the inner surface of "hat" section
stringers only. Thus the ring is not directly fastened to the
skin which is therefore not continuously restrained around
each ring circumference. The result is a modified floating * American Petroleum Institute - American Soeiety of
skin. Mechanical Engineers.
Douglas DC-8 Under Construction and Assembly.

t~.. ~
Fuselage is Pressurized to Permit High Altitude Flight.
Section of DC-8 Fuselage being Lowered into Hydrostatic Test Tank.

CHAPTER A -17
BENDING OF PLATES
ALFRED F. SCHMITT
A17. I Introduction.
It was seenin the last chapter that thin with another such, we write the equations down
curved shells can resist lateral loadings by by a direct appeal to past experience and
means of tensile-compressive membrane stresses. intuition.
As will be seen later, thin flat sheets, by de- Fig. A17-1 shows the differential element
flecting enough to provide both the necessary of a thin, initially flat plate, acted upon by
curvature and stretch, may also develop mem- bending moments (per unit length) Mx and My about
brane stresses to support lateral loads. In the axes parallel to the y and x directions respect-
analysis of these situations no bending strength ively. Sets of twisting couples Mxy (= - Myx)
is presumed in the sheet (membrane theory). also act on the element.
In contrast to the membrane, the plate is
y
a two-dimensional counterpart of the beam, in z
which transverse loads are resisted by flexural
and shear stresses, with no direct stresses in
?' yx
its middle plane (neutral surface).

The skin may also be classified as either a -M'
plate or a membrane depending upon the magnitude xy
of transverse deflections under loads. Trans- (Twisting
l40::!:...,e::.~~=t::::::;=---_--.1'---
couples shown
verse deflections of plates are small in compar- by right hand
Y
ison with the plates' thicknesses - on the order vector rule. )
of a tenth of the thickness. On the other hand, Fig. A17.I
the transverse deflections of a membrane will be
on the order of ten times its thickness.* As in the case of a beam, the curvature in
Unfortunately for the engineers' attempt at an orderly the x,z plane, a 2 w/ax 2 , is proportional to the
cataloging of problems, most aircraft skins fall between the moment Mx applied. The constant of proportion-
above two extremes and hence behave as plates having some ality is l/EI, the reciprocal of the bending
membrane stresses. stiffness. For a unit Width of beam I = t 3 /12.
Plate bending investigations have for a In the case of a plate, due to the Poisson
longtime been important in aircraft structural effect, the moment My also produces a (negative)
analyses in their relation to sheet buckling curvature in the x, z plane. ThUS, altogether,
problems. Recently they have assumed new im- with both moments acting, one has
portance with the introduction of thick skinned
construction and still more recently with the
use of very thin low aspect ratio wings and
control surfaces which behave much like large
plates, or even are plates in some cases. where ~ is Poisson's ratio (about .3 for alumin-
It is the purpose of this chapter to pre- um). Likewise, the curvature in the y, z plane
sent briefly the classic plate formulas and is
some applications. Appropriate references are
cited in lieu of an exhaustive treatise, which
could hardly be presented in one chapter (or
even one volume) as Witness the voluminous
literature on the SUbject. These two equations are usually rearranged to
give the moments in terms of the curvature.
A17.2 Plate Bending Equations **. They are written
Technical literature in this field abounds
with many excellent and elegant derivations of
the plate bending equations (references I and 2,
Mx =D(aa
2
x2w+ ~ R
a w) -
2
(1)
for instance). Rather than labor the subject
(2)
* As will be seen later, the presence or absence of mem-
brane stresses is not wholly dependent upon the magnitude
of deflections, but is also determined by the form of de-
flection surface assumed by the sheet (in turn dependent upon where
the shape of boundary and loading).
The twist of the element, a 2 w/axay
** the assumptions implicit in the following analysis are (= a~/ayax) is the change in x-direction-
spelled out in detail in Art. Al 7. 5, below. slope per unit distance in the y-direction (and
A17.1
A17.2 BENDING OF PLATES
(and visa versa)***. It is proportional to the In a similar manner, a moment summation about
tWisting couple Mxy • A careful analysis (oee the x axis yields
references 1 and 2) gives the relation as
a2 w (5 )
My
.JC
= D (1 - II)
r-
-axay (3 )
(Equations (4) and (5) correspond to V = dM!dx

Equations (1), (2) and (3) relate the applied in beam theory).
bending and twisting couples to the distortion One final equation is obtained by summing
of the plate in much the same way as does forces in the z direction on the element:
M = EI d 2 Y/dx 2 for a beam.
While a few highly instructive problems may
be solved with these equations (see reference 1, (6)
pp. 45-49 and reference 2, pp. 111-113), they
are of little technical importance. Hence we Equations (4), (5) and (6) provide three
move on to consider bending due to lateral loads. additional equations in the three additional
Fig. A17-2 shows the same plate element as quantities Qx , Qy and q. The plate problem is
in Fig. A17-1, but with the addition of internal thus completely defined.
shear forces Qx and Qy (corresponding to the nV n To summarize, we tabulate below the quan-
of beam theory) and a distributed transverse tities and equations obtained above. For com-
pressure load q (psi). With the presence of parison, the corresponding items from the
these shears, the bending and twisting moments engineering theory of beams are also listed.
now vary along the plate as indicated in Fig.
A17-2a. (For clarity, the several systems of
forces on the plate element were separated into CLASS ITEM PLATE THEORY
BEAM
the two figures of Fig. A17-2. They do, of THEORY
course, all act simultaneously on the single Coordinates x y x
element) • Geometry Deflections w y
M +dM
yx yx
a"w a"w
, a"w ~
/YM.6 Distortions ax" , ay" axdy dx"
Structural Bending Et"
D = 12 (1 _ 1-'") EI
Characteristic Stiffness
C,ouples Mx , My, MXY M
Loadings Shears Qx, Qy V
(a) Lateral q q or w
Fig. A17. 2. The differentials are increments which "Hooke's Moment- (a"w a"w)
Mx = D ax" + I-' ay"
should be written more precisely as, for instance, Law" Distortion
d~y ~ (a~y/aY)dy. Relation ( a"w a"w)
My = D ay" + I-' ax" M = EI ~
a"w
The next relations are obtained by summing Mxy = D (1 - 1-') aXdy
moments in turn about the x and y axes. For ex-
ample, we visualize the two loading sets of Fig. aMx aMxy
'i.x=-+-
A17-2 acting simultaneously on the single ele- ax ay
Equilibrium Moments v= ~
ment, and sum moments about the yaxis. ~ aMxy d x
Qy = +
ay ax
Mx dy + (Myx + d Myx) dx + (Qx + d Qx ) dx dy = _ aQX + aQy
Forces q - ax ay q= ~
d x
(Mx + d Mx ) dy + Myx dx
Dividing by dx dy and discarding the term of Finally, one very important equation is
higher order gives obtained by eliminating all internal forces
(Mx , My, Mxy , Qx, Qy) between the above six
Q =aMx_~
equations. The result (which the student
x ax a y should obtain by himself as an exercise) is a
relation between the lateral loading q and the
or,
deflections w*:
Qx = a Mx + a Mxy _ (4 )
ax ay (7 )
*** If w,
the deflection function, is a continuous function of
x and y (as it must be, of course, in any technically im-
portant plate problem) then at each point a 2 w/axdy = * the corresponding equation for a simple beam is
a~;aydx, as is proven in the calculus. q/EI = d.y/dx •.
A17.3
The plate bending problem is thus reduced

to an integration of eq. (7). For a given
lateral loading q (x, y), a deflection function
w (x, y) is sought Which satisfies both eq. (7)
and the specified boundary conditions. Once Hence the required deflection surface (and the
found, w (x, y) can be entered into eqs. (1) to solution to the problem**) has the equation
(5) to determine the internal forces and stress-
es. 1
qmn( ~+~
w=f)=4 2 2)2 si nmnxsinnn
-a- -b-
y
- (10)
A17. 3 An illustrative Plate Bending Analysis. n a 2 b2
Assume a lateral loading applied to a rec-

tangular plate having all edges simply supported The maximum deflection is seen to occur
(hinged). The coordinates are chosen as in Fig. where the trigonometric functions have values
A17-3. With foreknowledge of the general use- of unity and q is also a maximum.
fulness of the result, we assume a sinusoidal If eq. (10) is substituted into eqs. (1),
loading of the form (2) and (3) one obtains
q .. qmn Sin m an x sin -

nbn -
y- (8)
m, n = 1, 2, 3,
M _ _
..x n 2(illa ~mn
2 + nb 22) 2 (illa 22 + I-l n
b 22) S In
. ill- aT1 -X S In. n- bT1 -y
qmn (1 - I-l) mn ill T1 X n T1 y

Mxy = n (~:
2 + ~) 2 ab cos a cos -b-
Fig. A17. 3 Sinusoidal loading on a In a similar manner the transverse shears

rectangular plate. Sections through may be found from eqs. (4) and (5).
the loading shown for m=3, n=2. With such results as these the plates'
stresses may be determined as desired. For
example, the maXimum direct bending stresses
To find the resulting deflected shape of are seen to occur where the shear stresses (due
the plate we try a solution of the form to Mxy ) are zero. Thus
w = Amn sin m n x sin n n y

a b
(9) Sx (= MC)I = Mxt
21
Where Amn is the unknown deflection amplitUde. and hence

This trial deflection function is known to sat-
isfy the boundary conditions on the plate since
at x = 0, a and at y.= 0, b we have - - - - - - - - (11)
w= ° (zero deflection at the supported
edges)
The reader haVing a familiarity with Fourier series
methods will recognize immediately that the above analysis
(zero moment at the hinged provides the key to the solution of the problem of any general
° edges: see eqs. 1 and 2) loading q (x, y) on the same plate. Such an application is
made by determining the proper combination of sinusoidal
It remains only to find the value of Amn which pressure terms (each of the form of eq. 8) such that their sum
will closely represent the desired loading. The sum of the
will satisfy eq. (7). SUbstituting (8) and (9) corresponding deflection functions (each of the form of eq. 10)
into (7) one obtains gives the desired solution. Details of this type of analysis are
to be found in reference 1 on pp. 113-176 and 199-256.
In common With all problems which are
formulated in terms of a partial differential
* ** theuniqueness of solutions to the differential equation of

or the form of eq. (7) is a classical proof appearing in Ilum-
erous advanced texts on mathematics and mathematical
physics. Since the equation is known to have a unique
* the common factor sin m: x sin f has been divided out.
solution, then any solution found for it is the one and
only correct solution.
A17.4 BENDING OF STRESSES
equation, the solution of the plate bending prob- TABLE A17.1

lem depends strongly upon the boundary conditions
(both the shape of the boundary and the types of Stress and Deflection Coefficients for a Uniformly
support provided there). The above example may Loaded Rectangular Plate Having Various Edge
Conditions
be said to have been deceptively easy because of
both the simple shape of the boundary and the Long Sides Short Sides
type of support. Plate problems wherein the All Sides Pinned, Pinned, All Sides
plate planform is not a simple geometric figure Pinned Short Sides Long Sides Clamped
Clamped Clamped
must be solved by numerical means. As to the
type of support, a fUll discussion of boundary b/a a 13 a 13 a f3 a ~
conditions for plates is to be found in refer-
1.0 .0443 .2874 .0209 .420 .0209 .420 .0138 .3078
ence 1, pp. 89-95.
1.2 .0616 .3756 .0340 .522 .0243 .462 .0188 .3834
A17. 4 Compilations of Results for Plate Bending Problem s. 1.4 .0770 .4518 .0502 .600 .0262 .486 .0226 .4356
Fortunately for the practicing engineer, it 1.6 .0906 .5172 .0658 .654 .0273 .500 .0251 .4680
is not necessary to perform analytic computations 1.8 .1017 .5688 .0799 .690 .0279 .502 .0267 .4872
as discussed above for the great majority of 2.0 .1106 .6102 .0987 .714 .0284 .504 .0277 .4974
practical plate problems. Problems of the type 3.0 .1336 .7134 .128 .750 - - - -
illustrated above, plus the myriad variations 4.0 .1400 .7410 - - - - - -
possible, became very fashionable exercises 5.0 .1416 .7476 - - - - - -
amongst mathematicians following the discovery 00 .1422 .7500 .1422 .750 .0284 .498 .0284 .498
by LaGrange of eq. (7) in the year 1811. The Rectangular Plates Under Various Loadings
results of many researchers' labors have been
compiled in various forms for handy reference. • S. Timoshenko, "Theory of Plates and Shells",
A com~on and important case is that of a pp. 113-176, 199-256.
uniformly loaded rectangular plate (Fig. A17-4).
The major engineering results are the values of • J. P. Den Hartog, "Advanced Strength of
Materials", pp. 132-134.
the maximum deflections and the maximum stresses
developed. These may be put in the form (a is • R. J. Roark, "Formulas for Stress and
the length of the short side): Strain", pp. 202-207.
= q a4 Circular Plates Under Various Loadings

a E t3 - - - - - - - - - - - - - - - (12)
(same three references, in order)
- - (13) • PP. 55-84, 257-287.
• PP. 128-132.
where the coefficients a and ~ are given in
Table A17.1 for the four most Common edge
• PP. 194-201, 209-211.
conditions. A17.5 Deflection Limitations in Plate Analyses.
In the introductory remarks of this chap-
ter it was stated that a plate may be distin-
gUished from a membrane by the small order of
its deflections (on the order of a few tenths
of its thickness). We will re-examine this
statement here to show that this is not so
much a definition as it is an accuracy limita-
tion imposed by one of the assumptions made in
the plate analysis.
Fig. A17.4 There are several familiar assumptions
from beam theory which, of course, carryover
here, inasmuch as the plate analysis resembles
Similar presentations may be made for many the beam analysis rather closely. These "beam
dozens of other cases. With the readyavaila- theory assumptions" are:
bility of comprehensive catalogings of these
problems in references devoted to the purpose, i-elastic stresses only are presumed,
there appears to be little Virtue in duplication ii - small slopes (so that a 2w/ax 2 and
here. Hence the following list of selected a~/ay2 are good approximations to
references is presented. Additional references the curvatures),
are to be found in turn within these works. We iii - at least one transverse dimension
note that, because of the linearity of the plate (length or width) be large compared to
bending problem, superposition of solutions is
the thickness so that shear deflections
possible to extend even further the usefulness may be neglected.
of these extensive listings.
AI7.5
However, the beam theory assumptions do not

place a very severe-restriction on the-magnitude
of defleCt1ons-permitted. Deflections of sev-
eral times the plate thickness would be per-
missible if these were the only restricting
assumptions.
In deriving the plate bending equations it
was assumed that no stresses acted in the middle
(neutral) plane of the plate (no membrane
stresses). Thus, in summing forces to derive
eq. (6), no membrane stresses were present to
help support the lateral load. Now in the sol- Fig. A17-5b shows the desired one-dimen-
utions to the great majority of all plate bend- sional problem which now resembles a loaded
ing problems (solved as in Art. A17.3), the de- cable. The differential equation of equilib-
flection surface solution found is a non-de- rium is obtained by summing vertical forces on
velopable surface, i.e., a surface which cannot the element of Fig. A17-5c (draw with all
be formed from a flat sheet without some strech- quantities; loads, deflections, slopes and
ing of the sheets' middle surface*. But, if curvatures shown positive). One obtains
appreciable middle surface strains must occur,
then large middle surface stresses will result,
invalidating the assumption upon which eq. (6) +qdx=O
was derived.
Thus, practically all loaded plates deform
into surfaces which induce some middle surface or
stresses. It is the necessity for holding down
the magnitude of these very powerful middle
surface stretching forces that results in the - - - - - - (14)
more severe rule-of-thumb restriction that plate
bending formulae apply accurately only to prob- where s is the membrane stress in psi.
lems in which deflections are a few tenths of Eq. (14) is the differential equation of
the plates' thickness. a parabola. Its solution is
AI7.6 Membrane Action in Very Thin Plates.
There is still another source of middle W-2~\ (a-x) - - - - - (15)
surface strains in plates: this is the re-
straint against in-plane movements offered by The (as yet) unknown stress in eq. (15)
the edge supports. While not important in prob- can be found by computing the change in length
lems wherein deflections are limited in accord- of the strip as it deflects. This "stretch" is
ance with the restriction of the last article, given by the difference between the curved arc
such restraint does assume great importance in length and the original straight length (a).
the case of large deflections of very thin Thus:
plates which support a major share of the load
by membrane action. It is, in fact, useful to
consider the limiting case of the flat membrane *
which cannot support any of the lateral load by
bending stresses and hence has to deflect and
stretch to develop both the necessary curvatures
and membrane stresses.
The two-dimensional membrane problem is a
nonlinear one whose solution has proven to be
very difficult. Rather than attempt to treat
the complete problem, we can study a simplified
version whose solution retains the desired
general features. The one-dimensional analysis
of a narrow (unit width) strip will be treated.
This strip is cut from an originally flat mem- Since the slope dW/dx is small compared with
brane whose extent in the y-direction is very unity, we use the binomial theorem to write
great (Fig. A17-5a).
( 1+ ( -dx
dW))~/2,....;
-1+1 (dW)2
2 -dx
-
* The cone and cylinder are examples of developable sur-

faces, the sphere is a nondevelopable one. It is a familiar * here"ds" is the differential arc length of the calculus
experience that the skin of an orange cannot be developed in- and has no kinship with the s which denotes the membrane
to a flat sheet without tearing. stress throughout the remainder of the analysis.
Hence at the middle of the long side of the panel.
o• J} ~ (~~ 'J
+ dx - a
We note that the limiting case, alb = 0, cor-
responds to the one-dimensional case analyzed
earlier. Unfortunately, an extrapolation of
these two-dimensional results to that limit
does not show agreement with the one-dimen-
=~ J1:/ dx •
siDnal result. Presumably the discrepancy may
be traced to the excessive influence of inac-
curacies in the assumed deflection shape of the
membrane as used in the approximate two-dimen-
Substituting through the use of eq. (15) and in- sional solutions.
tegrating we find Experimental results reported in reference
4 show good agreement with the theory for
= q" a" square panels in the elastic range.
6 24s"'t"
TABLE A17.2
Now by elementary considerations
Membrane Stress and Deflection Coefficients
alb 1.0 1.5 2.0 2.5 3.0 4.0 5.0
n1 .318 .228 .16 .125 .10 .068 .052

Equating these last and solving we find
n2 .356 .37 .336 .304 .272 .23 .205
s = .347 [E (qta)"J~~"- - - - - - - - - - - (16)

A17.7 Large Deflections in Plates **.
In the previous articles of this chapter
If eq. (16) is substituted into eq. (15) the results of analyses were outlined for the
two extreme cases of sheet panels under lateral
one gets for the maximum deflection (x =~ ) loads. At one extreme, sheets whose bending
stiffness is great relative to the loads applied
mx = .360 a
W (i ~r/" - - (17)
(and which therefore deflect only slightly) may
be analyzed satisfactorily by the plate bending
solutions. At the other extreme, very thin
Equations (16) and (17) display the essen- Sheets, under lateral loads great enough to
tial nonlinearity of the problem, the stress and cause large deflections, may be treated as mem
the deflection both varying as fractional ex- branes whose bending stiffness is ignored.
ponents of the lateral pressure q. As it happens, the most efficient plating
Solutions of the complete two-dimensional designs generally fall between these two ex-
nonlinear membrane problem have been carried tremes. On the one hand, if the designer is
out*, the results being expressed in forms iden- to take advantage of the presence of the in-
tical with those obtained above for the one- terior stiffening structure (rings, bulkheads,
dimensional problem, viz., stringers, etc.), Which is usually present for
other reasons anyway, then it is not necessary
Wmx =n~ a ( E~)~/"
t - - - - - - - - - (18) to make the skin so heavy as to behave like a
"pure" plate. On the other hand, if the skin
is made so thin as to necessitate supporting
~/" all pressure loads by stretching and developing
- - - - - - - - - - (19) membrane stresses, then permanent deformation
results, producing "quilting" or "washboarding'
Here "a" is the length of the long side of the The exact analysis of the two-dimensional
rectangular membrane and n~ and n" are given in plate which undergoes large deflections and
Table A17.2 as functions of the panel aspect thereby supports the lateral loading partly by
ratio alb. its bending resistance and partly by membrane
The maximum membrane stress (sMAX) occurs action is very involved. A one-dimensional
** The discussion to follow will be concerned primarily with

* The work of Henky and Foppl is summarized in reference 3, problems dealing with the support of a uniform pressure
pp. 258-290 and in reference 4. The partial differential load on a flat skin panel. n may, therefore, help the
equation solved is given in reference 1 on p. 344 (eq. 202) reader to fix his ideas if he visualizes the discussion as
and the approximate method of solution usually employed is applied to the problems of analysis of a single rectangular
sketched out on pp. 345, 346 of thls same reference. The skin panel taken between the stringers and bulkheads of a
reader who would compare presentations amongst these ref- seaplane hull bottom. Equally useful is the picture of the
erences should note the differences in the definitions of the very nearly flat panel between rings and stringers in the
plate dimensioning symbols "a" artd "b". slightly curved side of a large pressurized fuselage.
A17.7
analysis, parallel to that of Art. A17.6, is to The approximate large-deflection method

be found in reference 1, pp. 4-10. A more outlined above has serious shortcomings insofar
elaborate two-dimensional analysis is shown on as the prediction of stresses is concerned.
pp. 347-350 of this same reference. For simply supported edges the maximum combined
An approximate solution of the large de- stresses are known to occur at the panel mid-
flection plate problem can be obtained by adding point. Fig. A17-7 shows plots of these stresses
together the flat plate and membrane solutions for a square panel as predicted by the approxi-
in the following way: mate method (substituting q' and q" into eqs.
Solve eq. (12), the plate bending relation, (13) and (14) respectively and cross plotting
for q; call it q' , with the aid of Fig. A17-6*). Also shown are
the maximum stresses computed by the exact
q' = wMAX E t" large-deflection theory (reference 5).
a a6 ~o I-------.e-x-a-c-t--.------.-----.----;
Now solve eq. (18), the membrane relation, for approx.

q; call it q",
20/----I----I-----+------,=+--""--'-=--I
sa
Et
2
2
---
10 I------i-=-'£----:;oo"""l'-----+----+-----i
The sum of these two pressures gives the total
lateral pressure, called simply, q.
O~ _ _-'- .L-__:_--'---~.J.---':":'__
q = q' + q" o 50 100 4 150 200 250
q =-1a Et" 1 Et
- - b - WMAx + ~ - 6 WMAJC
"
- - - - - - (20)
7
Fig. A17. 7 Large-deflection theories' mid-
a6 (_) 4 n1 a panel stresses; simply supported square panel.
a
Because of the obVious desirability of
Eq. (20), we see, is based upon summing the in- using the results of the more exact theory, some
dividual stiffnesses of the two extreme be- of these are presented in Table A17.3. The
havior mechanisms by which a flat sheet can treatment of additional cases (other types of
support a lateral load. No interaction between edge support) may be found in reference 6, pp.
stress systems is assumed and, since the system 221, 222.
is nonlinear, the result can be an approximation TABLE A17. 3
only. Large Deflection Rectangular Plate Coefficients
Eq. (20) is best rewritten as (Uniform Pressure Load, q. Simply Supported Edges)
Deflection
Pressure Coefficient, qa 4/Et 4
=; (~)(%y +n~" (~j" - - - - - (21) alb or Stress

Coefficient
wit
12.5 25 50 75 100 150 200
0.4055 0.6083 0.846 1. 0031 1.1240 1. 3104 1. 4557 1. 5770

250
1 sb :YEt Ii! 3.191 5.158 7.770 9.720 11.339 14.055 16.364 18.424
Fig. A17-6 shows eq. (21) plotted for a
5MbJil/Et~ 0.609 1.377 2.683 3.792
square plate using values of a and n 1 as taken 4.785 6.542 8.103 9.533
from Tables A17.1 and A17.2. Also plotted are wit 0.6958 0.9463 1.241 1.440 1. 596 1. 840 2.033 2.196
the results of an exact analysis (reference 5).
2 tooo sb 2 lEt 2 4.871 7.164 10.26 12.62 14.61 18.01 20.93 23.55
As may be seen, eq. (21) is somewhat conserva- 2 2
tive inasmuch as it gives a deflection which is sMb /Et 1.293 2.403 4.155 5.612 6.911 9.220 11. 28 13.16
too large for a given pressure. Notes: 1. s = total stress = sB + 8M

2. sB = "bending stress" component of stress
3. sM = "membrane stress" component of stress
A17.8 Considerations in the Applications of Large-De-

flection Plate and Membrane Analyses.
Before concluding this chapter it is
pertinent to note several serious omissions in
the developments outlined above with regard to
their application to flat pressure-panel
analyses within a ship hull or fuselage. The
o l..-~:::::::::L':""'-_--l..._ _. J_ _~ L
o 0.5 wit 1.0 1.~ 2.0 * but using n2= .260 in eq. (19), This value gives the
stresses at the center of a square panel whereas n2 =
Fig. A17. 6 Deflections at the midpoint of a simply .356 in Table A17. 2 is for stresses at the panel edge.
supported square panel by two large-deflection
theories.
large-deflection plate and the membrClne analyses REFERENCES

were developed for applications where the plate
bending analysis appeared inadequate. However, 1. Timoshenko, S. "Theory of Plates and Shells",
these analyses themselves presumed conditions MCGraw-Hill, N. Y., 1940.
seldom encountered in practice.
FIRST, the analyses assume unyielding sup- 2. Den Hartog, J. P. "Advanced Strength of
ports on the boundaries of the sheet panel. In Materials", McGraw-Hill, N. Y., 1952.
practice, the skin is stretched across an elas- 3. Sechler, E. and Dunn, L. "Airplane Struc-
tic framework of stringers aml bulkheads. It tural Analysis and Design", John Wiley,
follows, therefore, that the heavy membrane N. Y., 1942.
tensile forces developed during large deflec-
tions will cause the supports to deflect towards 4. Heubert, M. and Sommer, A., Rectangular Shell
each other thereby increasing the plate de- Plating Under Uniformly Distributed Hydro-
flection and relieving some of the stresses. static Pressure, NACA TM 965.
A simple one-dimensional analysis for a 5. (selected large-deflection plate references).
membrane strip having elastic edge supports
(parallel to the analysis of Art. AI7.6), shows a) Moness, E. Flat Plates Under Pressure,
errors on the order of 25 per cent are likely if Journ. Aero. Sc1., §, Sept. 1938.
the framework elasticity is neglected (reference b) Ramberg, W., McPherson, A. and Levy, S.,
7). At this writing no two-dimensional treat- Normal Pressure Tests of Rectangular
ment of this problem is known to the writer. Plates, NACA TR 748, 1942.
SECOND, it is seldom that the analyst has
to check a panel for lateral pressure loads c) Levy, S. Square Plate With Clamped Edges
alone. Most often, the entire "field" of panels Under Normal Pressure Producing Large
on the framework of stringers and bulkheads must Deflec~ NACA TR 740, 1942. ----
simultaneously transmit in-plane loadings from d) Levy, S. Bending of Rectangular Plates
the tail load bending stresses and the cabin With Large Deflections, NACA TR 737, 1942.
pressurization stresses.
Inasmuch as the large-deflection plate and e) Chi-Teh Wang, Nonlinear Large Deflection
membrane analyses are nonlinear, it follows that Boundary-value Problems of Rectangular
correct stresses cannot be found by a straight Plates, NACA TN 1425, 1948.
superposition. The magnitude of the error in- f) Chi-Teh Wang, Bending of Rectangular
troduced by such a procedure is difficult to Plates With Large Deflections, NACA TN
estimate in the absence of an exact analysis. A 1462, 1948. - - -
one-dimensional analysis, parallel to that of
Art. A17.6, but with elastic supports and axial 6. Roark, R. J. "Formulas for Stress and Strain~
load, is given in reference 7. These reSUlts, McGraw-Hill, N. Y. 3 rd Edition, 1954.
which indicate the effect of the axial load to 7. Flugge, W. Stress Problems in Pressurized
be qUite importClnt, may be used as a guide in Cabins, NACA TN 2612, 1952.
lieu of more complete two-dimensional studies.
The interested reader is referred to the orig- 8. Engineering Research Bulletins of Oklahoma
inal work for details. State University.
a) No. 15. Influence Coefficients for
Circular Plates.
b) No. 17. Analysis of Thin Plates Sup-
ported on their Corners.
c) No. 118. Analysis of Flat Plates by
Algebraic Carryover Method.
CHAPTER Al8
THEORY OF THE INSTABILITY OF COLUMNS AND THIN SHEETS
(BY DR. GEORGE LIANIS)
PART I
ELASTIC AND INELASTIC INSTABILITY OF COLUMNS
A18. I Introduction. Thus the total bending moment at section z
will be:-
Part 1 of this chapter will be confined
to the theoretical treatment of the instability Qaz on upper portion - - - - ( d)
11 = Pu + --1-
of a perfect elastic column and an imperfect
elastic column. The column is the simplest of
the various types of structural elements that 11 = PU + Q(l-a) (l-z) on lower portion (e)
1
are sUbject to the phenomenon of instability.
The theory as deYeloped for columns forms the From mechanics of simple bending, we have
basis for the study of the instability of thin the deflection equation,
plates, which subject is treated in Part 2.
d 2u 11
A18.2 Combined Bending and Compression of Columns.
dz 2 = -
Ely --- - - - - - - - (1)
Consider a column with one end simply
Thus the deflection u(z) of the column
supported and the other end hinged (Fig. A18.1)
is,
under the simultaneous action of a compressive
load P and a transverse load Q. Without the d 2u Qaz
load P the bending moment due to Q would be:- EI dz2 = - PU 1
, (0 :: z :: 1 - a)
p
d2 u (l-a) (l-z)
EI dz 2 = - Pu - Q 1
/
(1 - a ::= z ::=. 1) - - - (2)
a /
1
1- j---++--Q If we introduce the notation,
I - (3)
I
\ The general solution of eq. (2) is:
\
\ Qaz
u = C], cos Kz + C2 sin Kz - Pl '
-4,U (O:::zl-a) - - - - (4a)
u =C 3 cos Kz + C4 sin Kz - Q (l-a) (l-Z)

1 '
(l-a::z::l) - - - - (4b)
where C]" C2 , C3 , and C4 are constants of

integration to be determined from boundary
On the lower portion of the column conditions.
11 = Qaz - _ - - - - _ - (a) For eqs. (4), since u =0 for z = 0 and
], 1
z ; 1, it follows that:
On the upper portion
11], = Q(l-a)(l-z)
C], = 0 and C =- C tan Kl - -
3 4 - - - (f)
], 1 - - - - - - - - - (b)
At z = (1 - a) the two portions
of the
Due to the deflection u(z), the axial deflection curve given by (4a) and (4b)
load P contributes to the bending moment by respectively must have the same deflection and
the amount:- slope. From these two conditions we determine
C2 and C4 •
(c)
Al . I
A18.2 THEORY OF THE INSTABILITY OF COLUMNS AND THIN SHEETS
Q sin Ka _ Q sin K(l-a) the body would cause only infinitesimal changes
C,,= PK sin Kl ' C.. - - PK tan Kl - - - - (g) in the displacements and the body recovers if
the added loads are removed. When the dis-
SUbstituting (f) and (g) into eq. (2), placements are continuously increased with
we obtain: little or no further increment of loads, the
system is unstable. If the body will remain
u = Q sin Ka i Kz Qa in the displaced position after the removal
PK sin Kl s n - Pl Z of the disturbance, the body is said to be in
(0 5 Z 5 1 - a) - - - - - - - - - - -(5a) neutral equilibrium. Having these definitions,
we will not investigate the behavior of the
column before and after the critical load is
u = QPKsinsinK(l-a)
Kl
i K(l-)
s n Z
Q(l-a) (l-z)
reached.
Pl
(1 - a 5 Z 5 1) - - - - -(5b) p A B P
~ ~
~~ ~/> z
If load Q is applied at the middle of the s~~ _
column the maximum deflection is:-
Q tan Kl Fig. A18.2
2 Ql
Umax = 2 PK - 4P - - - - - - (6 ) Assu~e, as shown in Fig. A18.2, that a
simply sunported column loaded by an axial load
Kl
It is obvious that for:2 =2n ' P is bent by a small disturbance. If the
deflection at a distance s along the deflected
KI - axis is u, the bending moment due to P is Pu.
tan """"2 00. Thus the maximum deflection of From engineering theory of bending, we
know that,
column becomes infinite for KI = n and from
EI
eq. (3). If = -M , whence
= n"EI EI
Pcr - - - - - - - - - - - - (7) If =-Pu - - - - - - - - - - - - -(8a)
The exact expression for the curvatuve of
Equation (7) is an important equation the neutral axis is:-
derived first by Euler. It gives the critical 1 _ dG
compressive load which causes infinite deflect- R- ds ' where s is the arc length of the
ion in a column and it specifies the ultimate
strength of a column in compression. deformed aXis, and G the angle between the
tangent to the curve and the z axis. Thus,
It is obvious-from eq. (7) that Euler's
critical load is independent of the magnitude EI dG + Pu =0
of the transverse load Q. It seems, therefore, ds - -(8b)
that even in the absence of the transverse
load Q, the maximum deflection becomes infinite Differentiating (8b) with respect to s
under the action of only a compressive load as
given by eq. (7). and since ~~ = sin G, we obtain:
d"G
A18. 3 Elastic Stability of a Column. EI -dB" + P sin G =0 - - - - - - - -(8c)
The above conclusion as to the critical Multiplying (8c) by dG and noting that:-
load was based on purely mathematical reasoning.
We have found a critical value of a compressive
load which causes infinite deflection. ~:~ dG = ~~ d (~~) , and integrating
Far more important, however, is an in- Elf dG d (dG) + pfsin GdG = C, or

vestigation of the stability of a column which ds ds
should be based on physical arguments. The
question arises as to what happens before the
load P reaches its critical value as given -("sd)
by eq. (7) and also how the column behaves if
this critical value is exceeded.
Since at end A, G=-a and M = EI ~~ = 0,
An elastic system is called stable under we find that C =-
P cos a.
given loads when infinitesimal loads added to
AIB.3
4K 2
Now let k = ~
EI - - - - - - - - - (8e)
P PlIO
Pcr = n 2 EI =7 (8n)
Then eq. (8d) becomes, 5 =_"",2P,"== - - - - - - - - - - - - - - (80)

1 dQ
1
n j-/-cr-
V2 ds = k V cos Q- cos a , or
Let us now write the bending moment M =
Po at the middle point in non-dimensional form:
ds = dQ (8f)
2
I 2k (cos Q- cos a) m=--'-=-2-
P
-
5 4KP
(8p)
P cr 1 n
The total length of the column in the
deflected shape is given by:- Since p = sin "2a is a function of a so is the
P 5
elliptic integral K and the ratios Pcr and 1
1 = J. B ds =/ a dQ _ - - (8h)
calculated from equations (8a) and (80).
A -u /2 k J cos Q- cos a
or 1 =/ a dQ _ _ _ _ (8i) Thus P~r is a function of ~ calculated by

-u 2k/sin ~ -sin 2 Q/2
2
means of tables giving elliptic integrals.
Thus m can be plotted against 5/1 as shown in
a
Denoting sin 2 by p and introducing a new Fig. A18.3.
variable 0 :-
5/1
Q
pJ. = Per
sin "2 = p sin 0
Equation (8i) then becomes,

n
2
1=5:...1 d:in20=2kK ------(8j)
k 0 Il_p2
n
where K = ! 2 dQ is called the Fig. AIB.3
o J I_p2 sin 2 0
IL... m
complete elliptic integral of the first kind,
~
and it can be found in tables. If a and

therefore p is very small, then p 2 sin 2 0 can Let us now examine the stability of
be neglected in equation (8j) and then various equilibrium configurations. Assume
that a load pJ. is acting on the column and
n
_2"2 n lEI the column has a certain maximum deflection 5
1-- 1 d0=-=nv~- where pJ. does not correspond to 5.
k 0 k P
The non-dimensional maXimum bending
whence P - n'"EI
- 2- - - - - - - - - - - - - (8k) moment is:-
1
J. pJ. 5
The deflection at the midpoint of column
m = -(-)
P 1 - - (8g)
cr
is:- Q = 0, du = ds sin Q, and from Eq. (8f)
The mJ. versus 5/1 curves are straight lines.
u(z= .1.)=5=L! a sin QdQ (81) The column is in eqUilibrium if m = mJ., or in
2 2kOji2U
sn 2-sn"2i 2Q other words, if the mJ.(5/1) curve intersects
the m(5/1) curve.
or in terms of 0:- We see from Fig. A18.3 that these curves
n intersect at the origin only i f P J. <: Pcr . The
2P 2 2P column, therefore, has only one possible
5 =- ! sin 0d0 = -k
k 0
- - (8m)
eqUilibrium form, for example, that for which
5/1 - 0, Which is the straight form. When
From equation (8j) and equation (8k), we obtain pJ. > Pcr, there are two points of intersection,
one when 0/1 = 0 and the other (point A) for (b) At P/Pcr = 1, or at point (A), a bifurcation
which 0/1 ~ O. The column thus has two of equilibrium occurs and the column starts
possible equilibrium forms, one straight and to acquire two possible neighbor positions
one bent. of equilibrium, the straight and the bent.
Let us now assume that at 0 = 0, the (c) Above P/Pcr = 1, the column has two possible
column is displaced by a small disturbance equilibrium positions 0/1 = 0 and 0:1./1 ~ O.
and acquires a deflection 0:1.' For p:I. -< Pcr ,
we see from Fig. 3 that m> m:l.. Thus p:I. is Thus as far as initiation of instability is
not sufficient to maintain the column in concerned, the Euler load as given by Eq. 7 can
equilibrium in the bent form and it will spring be considered as the critical load. The question
back to its straight form. Thus for p:I. c: Pcr , arises whether this load has a practical use
the straight form is stable. for design purposes. A logical design criterion
is obviously the maximum load which a column can
If p:I. > Pcr , then m:l. > m. Thus m:l. will sustain. We observe from Fig. A18.4 that the
bend the column still further. This means load P increases for increasing displacement o.
that if p:I. > Pcr , the straight form of This behavior is due to the development of large
equilibrium is unstable. The column will deflections due to bending. However, over a
continue to bend until m:l. becomes equal to m considerable range of deflections 0, the P -+ 0
(point A in Fig. 3). If the column is dis- curve is practically horizontal (for instance,
placed further from A, the deflection becomes between points A and B the ratio 0/1 varies
larger than O2 and m ~m:l. at the new position. from zero to ~ 0.4). For such large deflections
The column will spring back to point A. Point for which the column load does not change
A is therefore stable. practically, it is obvious that the column
ceases to function properly. Therefore, from
At P = Pcr , the m:l. versus 0/1 line is this point of view, the Euler load can be con-
tangent to the m curve at the origin. There- sidered that which characterizes the maximum
fore, for an infinitesimal disturbance, the strength of the column.
column will remain in equilibrium at the
displaced position since for such small The rising part of the curve BD holds as
disturbances m:l. remains equal to m. The column long as the material behaves elastically. At
is therefore in neutral equilibrium. some point B, however, inside the almost flat
portion of the curve, the inner fibers of the
A18.4 The Failure of Columns by Compression. column acquire maximum stress equal to the
yield stress. If we carry out an elastic-
In discussing the stability of a column plastic analysis of the subsequent behavior,
in the previous section, it was shown that we observe that the curve drops almost
below the critical Euler load (Eq. 7), the immediately. Again this maximum load PB is
straight form is stable, above the Pcr the very near the Euler load. For design purposes,
bent form is stable and at Pcr the equilibrium therefore, the Euler load, which is a buckling
is neutral. By plotting the curve p/pcr versus load, is a very good approximation to the
0/1 as shown in Fig. A18.4, we observe the Ultimate load which the column can sustain.
following behavior.
Another argument will confirm the above
(a) Below P/P cr = 1, there is only one conclusion. In discussing the buckling of
equilibrium position, 0/1 = O. columns in the previous paragraphs, we have
assumed that the column is initially straight,
centrally loaded and made of
piper homogeneous material. Actual
columns, however, are imperfect
due to initial crookedness (for
instance, due to unavoidable
tolerances in their manufacture),
due to slight load eccentricities
t B and due to lack of complete
10
• A- - '-'I .... ..... homogenity. Therefore, a certain
I ......... amount of bending is always
I u present even for small loads.
I
I Let us now examine the
I
J behavior of such initially im-
1~/l perfect columns by assuming a
o'1r----::-::--=-=--=-=--...l..::---:-~0 11
O. I 0.2 O. 3 0.4 certain initial deflection 11"
Fig. of the column axis (see Fig.
Fig. A18.4 A18.5 A18.5). For small deflections,
A18.5
the change of curvature due to sUbsequent approaches the horizontal line PIPer 1 =
bending (after loading) is:- asymptotically. This curve, however, is valid
for small deflections for which the approxi-
° (l)R =~
dz
_ d l};o
2
dZ
mation:-
is valid.
In the differential equation of deflection
one can prove that llR is the change of
curvature which for an initial straight column By a treatment similar to that in the
coincides with the curvature itself. Thus in previous paragraph, we will find that for
the present case, where the bending moment is large deflections the load deflection curve
Pu, the equation of deflection becomes:- raises after the point I (curve FIH). Due to
the onset of plasticity, the actual curve
drops at the point 11. (curve FII1.H1.). The
failing load at 11. can be either greater or
smaller than Per' but it is usually very near
to it.
Let us express Uo in Fournier series:-
In the above discussions we have shown
- - - - -(10) that for all practical purposes the Euler
buckling load can be considered as the ulti-
mate load which a real or practical column
Substituting (10) in (9), we find the can sustain. Besides its closeness to the
solution which satisfies the boundary actual ultimate load, the critical load can
conditions (u = 0 for z 0, Z = 1) is:-
be easily calculated from equation (7) with-
00
out the necessity of carrying out a lengthy
u =Z on sin n~ - - (lla) calculation which will include the initial
n=l imperfections and plastic effects.
6n
"un '" _---':..::....,-_ It should be noted, however, that the
were
h - - - - - - - - - (llb) buckling load given by equation (7) is valid
1 - PIPn
when the uniform stress due to a compressive
load (0 = PIA, where A is cross-sectional
area) is below yield stress. If 0 is above
the yield stress, the theory of plasticity
predicts another value for the buckling load.
The deflection of the column at the center Referring now to equations (11) we find:-
is:-
Pn =n 2
Pcr (p cr from equation 7)
- -(12)
6n
If we plot the deflection versus the load on = ---=-~~P~=
1
we obtain the curve (Fig. A18.6), which n 2 Pcr
piPer Thus as P approaches Pcr ' we see that
~ _ 9/8 etc.
c
i\
H
1.0 A
"H1. In a buckling test we measure Omax °=

- 6, where 6 is the initial deflection at the
middle point. Thus:-
°;- Omax - 61. = =-~61."--_

P cr
and,
1
II P
Fig. A18.6 - - - - - - - - - - (13)

_.... 0 THEORY OF THE INSTABILITY OF COLUMNS AND THIN SHEETS
If in a buckling test we plot ojp versus 0 1 0 1

0, we can obtain the critical load experi-
'mentally without knowing the initial deflection 0 1 0 0
0
61 , It is simply the inverse of the slope of sin kl cos kl 1 1
this curve. - sin kl - cos kl 0 0
AlB. 5 Buckling Loads of Columns with Various End

Conditions. or sin kl = 0 or kl = nn (n = 1, 2, 3 - - -)
n" n"EI
From the conclusions reached in the
previous discussion, we can consider the
whence Pn = 1 " (16a)
buckling problem as an instability problem of

an initially straight column. Thus we assume Thus for Pn equal to the right hand side
a certain deflected position near the straight of equation (16a), we have a possible form of
equilibrium of the bent form. The smallest
configuration as another possible equilibrium
form and seek the loads under which the non- value of Pn occurs at n = 1, and this is the
buckling load:
straight form is possible. Furthermore, only
a small deflection analysis is necessary. n"EI
The general differential equation of
P cr =-r - - (16b)
bending-buckling is:-
The buckling load for other end conditions
4 can be derived in similar manner.
d u + k" d"u = 0 - - - - - - (14)
dz" dZ"
INELASTIC COLUMN STRENGTH
and the general solution is:-
AlB. 6 Inelastic Buckling. Introduction.
Euler's theory of buckling is valid as

The coefficients C1 , C", C3 and C4 long as the stress in the column nowhere ex-
depend on the conditions of the end supports. ceeds the elastic limit of the column material.
The various end conditions are:- We have seen that the analysis for perfect and
imperfect elastic columns leads to the same
d"u d3 u result, namely, equation (7).
Free end:- dz" = 0 , dz 3 = 0
The case of the inelastic buckling, that
is, instability under axial load exceeding the
Pin end: - u =0 elastic limit stress, presents some difficulties.
As We will see, the perfect column analysis
Fixed end:- u = 0 du leads to a different expression for the critical
dz = 0 load than for the perfect column. This is due
to the fact that in the plastic stress range,
Thus we have 4 end conditions. These the material behaves differently under loading
give systems of four linear homogenous and unloading, as illustrated in Fig. A1S.7.
equations. A trivial solution of these is Let us now examine the two cases:
the zero solution. For the buckling state,
however, C~, C", C3 , C4 are not all zero.
The condition of non-zero solution of the
above system is that the determinant of the
/,
//~IOading
coefficients of C1 , C", C3 and C4 is equal to
zero. From this equation, we calculate the
buckling load.
/
/
For example, in a simply sUDported beam, /
d"u
(u = 0, dz" = 0, at both ends), the end /
/
conditions furnish give Deformation E
Fig. AIB.7
C" =0
AIB.7 Perfect Column. Reduced Modulus Theory.
C1 sin kl + C" cos kl + C3 1 + C4 =0
Let us assume that the perfect column is
For buckling we must have:- first compressed uniformly up to the stress o.
To study the critical value, ocr, of for which °
Al8.7
the column becomes unstable we assume:- - (19)

(1) That the displacements are small so that The distribution of the compressive (-)
the relation between the radius of stress and the tensile (+) stress due to
curvature R and the deflection u of the bending is shown in Fig. A18.Sa. The stress
elastic axis is, becomes zero on line (a a~), which is at a
distance e from the centroid c. For
equilibrium of stresses on the cross-section
- - - - - - - - - - - (17a) we have,
(2) Plane sections remain plane, therefore - La,.
a oS ~ a oS ,. dA
dA + La,. =0 - - - - (20)
the change of strain due to bending at a
distance h on the plane of bending is, and for equilibrium for moments,
h d"u
OE =R = h dz" - (17b) - - (21)
(3) The stress-strain relation follows the Due to the linear distribution of stress,
simple tension curve for the material. we have:-
(4) The plane of bending is a plane of

symmetry of the cross-section.
00,.
- - - - - - - - - - - (22)
Assume now a column with the cross-section
as shown in Fig. A18.8a be compressed in the oS,. = -a:-,. h,.
Introducing now (17b), (lS), (19) and
(22) in equation (20), we obtain,
- - - - (23)
00,.
where,
Q~ = L0 a~ h~ dA , Q,. = La,.
0 h,.dA - - - (24)
are the moments of the cross-sectional areas
~~~~~~E to the right and left of line a a~.
I Fig. Al8.8b From eq. (21) we obtain,
la~ r
I I
----·~IE-- a~ ----+1 -I d"u
E dz" = Pu - - - - - - - - - - (25)
Fig. Al8.8a
where,
plastic stress range and that the compressive
stress prior to instability be a. To consider
the condition of buckling, let the column be E= - (26)
slightly deflected transversely. The stress
on one side of the column will then increase E is the so-called reduced modulus, and
due to the bending following the stress-strain I~ and I,. being the moment of inertia of the
curve, while on the other side the stress will two sides.
decrease and will therefore follow the un-
loading elastic line shown in Fig. A18.8b. We observe that the position of the
neutral axis in terms of the axial stress is
For small changes of the stress, on the given by eq. (24), while the buckling eq. (25)
first side, the variation of stress is related is similar to the elastic buckling eq. (14).
to the variation of strain by:- However, the value of K here is not given by
eq. (3), but by,
- - - - - (18)
where Et (a) is the slope of the stress-strain
curve at stress a. On the second side the
K=Af - - (27)
changes will follow the elastic relation, Therefore all the results of the previous
that is, analysis will be valid for the case of in-
elastic buckling. For instance, for a simply
Al8.8 THEORY OF THE INSTABILITY OF COLUMNS AND THIN SHEETS
supported column according to eq. (7), we will

have,
(28)
d
Since E is a function of ocr given by eq. (28) b

and the value of Et at the unknown ocr, the
calculation of the critical stress requires a
trial and error simultaneous solution of
equations (23), (26) and (28).
Al8.8 Imperfect Column. Tangent-Modulus Theory.
The tangent-modulus theory, originally

proposed by Engesser (Ref. 1), is based on the Fig. Al8.9
assumption that at the critical state, no
stress reversal takes place, and the critical ing the effect of a load P above the critical
stress, therefore, is determined only by the load Pt corresponding to tangent-modulus,
tangent modulus Et. This theory was abandoned Shanley proves the relation,
early since according to the previous dis-
cussion with the classical definition of
instability (perfect column, bifurcation of P = Pt 1 +
~+
1
1 +'C ) - - - - - - (32)
equilibrium) strain reversal does take place. (
2d l-'C
In recent years, however, this tangent modulus
theory has been proved useful. where
Under the assumption of no strain reversal
both sides of the cross-section in Fig. A18.8a,
will be characterized by the same linear stress It must be emphasized that the buckled con-
distribution, corresponding to the tangent- figuration is a stable one similar to that
modulus Et. Thus the buckling equation will be, considered in the refined Euler's theory.
Shanley has recognized the fact that such a
stable configuration may exist after exceeding
- - - - - - - (29) the tangent modulus load.
and the critical stress for simply supported =
If R P/Pt , Shanley found that the
end conditions becomes, relation betwe.en the variation of stress due
to bending and the compressive strain et
corresponding to at is;-
- - - - - - - - - - - - (30)
Since 1 1 + 12 > I and E > Et, it follows Concave side;

from (26)
- - - - (33)
E> Et and or> at - - - - - - - - - (31) 2 (R-l)
The critical stress at, therefore pre- Convex side:
dicted by the tangent-modulus theory is Et
smaller than or from the reduced modulus
theory. Although for perfect columns, the In Fig. A18.10, OE1/Et and OE2/Et are
assumption of no strain reversal is in plotted against R for 'C = 0.75. We observe
contradiction to the material behavior in the that while the strain on the concave side
plastic range, most experiments have given increases very rapidly and reaches an infinite
results more closely to the results by the value at the reduced-modulus load, the strain
tangent-modulus theory. on the convex side decreases initially very
slowly. Due to this picture we can conceive
To resolve this controversy, Shanley that in a real column, which has initial im-
(Ref. 2), proposed the following explanation. perfections, the compressive strain will
For simplicity, let a two-flange buckled increase more rapidly. Furthermore, the
column be formed by two rigid legs (see Fig. rapid increase of OE 1 will cause a fast re-
A18.9) joined in the middle by a plastic duction of Et. The column, therefore, loses
hinge. Assume that this column starts to its usefulness after the tangent-modulus has
buckle as soon as at is reached. By consider- been slightly increased. Thus the tangent-
Al8.9
000
I ...-CoN"o{RE-'ENGESSER (~." RQ
70
\ I
.".....IJ
FLAT END TESTS

I
000
~ 0000
\
,~
\\.
'..;\ ~EULER
k
\ 000
ENGESSER- i"..A
(TA-r MDIJ)
t--~_
,
~~
\
\ 000
~
i'- "'""'"-
\ J
0
20 ~ ~ ~ m lro I~ I~ ~
r-..
~
--
2~ 240
\ SLENDERNESS RATIO (I,.t)
"''''~
Coillmn lest TUults on 17S-T solid round rod (0.282 in. diameter).
,.. ,,10m' nl Mea Iils LtJI1rI
(> Fig. Al8.11
0 '-or ~l~
.,.. '- p.m,c6c1 ModtJlfII Load 00 0
70000
o FlAT END TESTS
.ROUt«) END TESTS
000
Fig. Al8.l0
ooo~ ~= ~
7f~ It:
modulus, even though it does not actually ~ ~
define an unstable configuration, it repre- .......
sents the lower limit of a spectrum of possible ~ ~ ~ r.::::......
buckled configurations, the upper limit of
which is the reduced modulus load which
..~
~
.~
~
corresponds to infinite deflections.
Thus to summarize, sufficient experi-
mental results are available to show that the
'0
0
~ ~ ~
.......
~
~
m
-- f0&-
lro ~
...........
~
r--
~
- ~
r---
~ m
failing stress of a column in the inelastic SLENDERNESS RATIO, (LIf)
('"h",,,, ,,.111 rt'Imllttmt 17:':'-Ta"glt' (O,{jll.I'!ll~ in. iy ~ ;',.).
range can be found by replacing E by the
tangent modulus Et in Euler's equation, or, Fig. Al8.l2
virr-
References:
n 2 Et
= r = A - - - - - - - (34) Ref. 1. Engesser F., Schweezeriche Bauer
(L/r)2 '
Zeitung. Vol. 26, p. 24, 1895.
Figs. A18.11 and 12 show how experimental Ref. 2. Shanley, F.R., Inelastic Column
results check the strength as given by the Theory, Jour. Aeronautical Sciences,
Euler equation using the tangent-modulus Et. 1947, p. 261.
PART 2
THEORY OF THE ELASTIC INSTABILITY OF THIN SHEETS
A18.9 Introduction. curvature of the middle surface can be found

aw aw
Thin sheets represent a very common and apprOXimately by omitting powers of ax ' ay ,
important structural element in aerospace as compared to unity, as it has been done for
structures since the major ·units of such the curvature of beams. Thus the curvature
structures are covered with thin sheet panels. of the deflected middle surface in planes
Since compressive stresses cannot be eliminated parallel to xz and yz planes respectively are:
in aerospace structures, it is important to
know what stress intensities will cause thin 1
- (la)
sheet panels to buckle. Equations for the Ry
buckling of thin sheet panels under various
load systems and boundary conditions have Another quantity used in the problem of plates
been derived many years ago and are readily is the so-called twist of the middle surface
available to design engineers. Part C of this given by:
book takes up the use of the many buckling
equations in the practical design of thin (lb)
sheet structural elements. The purpose of
this chapter is to introduce the student to
the theory of thin plate instability or how The strains can now be expressed by means of
these buckling equations so widely in use by curvatures and twist of the middle surface.
design engineers were derived. For a broad In the case of pure bending of prismatic bar
comprehensive treatment of the sUbject of a rigorous solution was obtained by assuming
instability of structural elements, the student that cross-sections of the bars remain plane
should refer to some of the references as after bending and rotate so as to remain
listed at the end of this chapter. perpendicular to the deflected neutral axis.
Combination of such bending in two perpen-
A18.10 Pure Bending of Thin Plates. dicular directions brings us to pure bending
of plates.
To derive the theory of instability of
thin plates, we must first derive the theory .--+-----,£--+--,--_ x
of the pure bending of thin plates.
In the following the analysis will be

confined to small deformations. Let x, y be
the middle plane of the plate before bending
occurs and Z be the axis normal to that plane.
Points of the x,
~------~x
y plane undergo y z
small displace- Fig. 2a
ments, w in the
z-direction,
which will be
referred as the
y
deflection of
the plate. The
z slope of the
middle-surface
in the x- and
r-------~x
y-directions
after bending are
y
Fig. 2b
For small de- Let Fig. (2a) represent a thin rectangular
Fig. I flections, the plate loaded by uniformly distributed bending
A18.1l
Mx , My per unit length at its edges. These a2w __ 2z

moments are considered positive when they are E xy = 2z axay Rxy
- (3)
directed as shown in the figure, i.e. when
they produce compression in the upper surface Since we treat the problem of plates as a plane
of the plate and tension in the lower. Let stress problem, we find by means of Hook's law
also (Fig. 2b) be a rectangular element cut
out of the plate with sides dx, dy, t. The
thickness t is considered very small compared
with the other dimensions. Obviously the
stress conditions at the edges of all such
- - - - - - - (4a)
elements will be identical to that of Fig. 2a.
Assume now that the lateral sides of the
element remain plane during bending and rotate = Ez (1 lI)=
about the axes so as to remain normal to the Oz (1_lI 2 ) Ry + Rx
deflected middle surface. Due to symmetry
the middle surface does not undergo any
extension and it is therefore the neutral - - - - - - - (4b)
surface.
These normal stresses are linearly distributed
From the geometry of the above described over the plate thickness. Their reSUltants
form of deformation the displacements in the must be equal to Mx and My respectively:
x, y, z directions can be found as follows:
J h/2 Ox z dydz =Mxdy
A point B on the middle surface has been -h/2
displaced to B~ by W in the z-direction. An
element of surface dzdy has rotated by an J h/2 Oy z dxdz =Mydx
angle equal to the slope of the deflected
middle surface in the direction so as to -h/2
remain normal to the middle surface. See
Fig. 3. SUbstituting from (4) we find:
~-----:;:----,>x
Mx = -D 2
(aaxew + V a
2
w)
aye (5a)
z
Fig. 3
This angle for small displacements is obviously

aw
equal to ax' Thus the horizontal displacement From eq. (3) we obtain:
Ux in the x-direction of a point at distance z
from the middle surface is: 2Gz aSw
Oxy = ayx = Rxy = 2Gz axay - - - - - (4c)
aw
Ux =- z ax (The sign - indicates
negative displacement for positive z). Mxydx = J h/2 OxyZ dxdz , MyzdY =
-h/2
In a similar manner we find the displacement
h/2
in the y-direction. The complete displacement J oyxz dydz , Mxy =Myx =
system is: -h/2
aw
Ux = - z ax ' uy
aw
= - z ay , Uz = W (x, y ) - - (2) D(l-lI) ~
axay
- - - - (5c)
The corresponding strains are: Equations (5) give the moments per unit length
for pure bending and tWisting of a plate.
a
2
w z _ z a2
w -_-L
EX = - z ax 2 = Rx ' Ey - - ay 2 Ry '
A18. 11 The Differential Equation of the Deflection Surface. aMxy aMy

ax dxdy - ay dxdy + Qy dxdy =0 or
To develop the theory of small deflections
of thin plates we make one more assumption.
At the boundary of the plate we assume that aMxy aMy
its edges are free to move in the plane of the ax -ay-+ Qy =0 - - - (6b)
plate.
From the moment equilibrium in the y-direction
we find:
aMxy aMx
--+---Qx=O
ay ax --------(6c)
By eliminating Qx, Qy from 6 a,b,c, we find

the equilibrium relation between the moments:
(7 )
To represent this equation in terms of the

deflections w of the plate, we make the
assumption that the expression (5) derived for
pure bending holds approximately also in the
case of laterally loaded plates. This
Fig. 4 assumption is equivalent to neglecting the
effect on bending of the shearing forces and
Thus the reactive forces at the edges due to the compressive stress Oz. This is an ex-
the supports are normal to the plate. With tension of the engineering theory of bending
these assumptions we can neglect any strain of beams. As in the case of beams it gives
in the middle plane during bending. good approximation for bending of plates under
transverse loads.
Let us consider, Fig.4, an element dxdy
of the middle plane. Along its edge the Introducing equation (4) into (7) we find:
moments Mx , My, Mxy are distributed. These
are the resultants of the bending and twisting a4 w a4 w
stresses distributed linearly along the - -4+ - - + - (8)
ax ay 4
thickness of the plate (see eqs. 4 and 5).
If the plate is loaded by external forces
normal to the middle plane in addition to the The problem of bending of plates is thus
above moments there are vertical shearing reduced to integrating eq. (8) for w. The
forces Qx, Qy, acting on the sides of the corresponding shearing forces in terms of the
element of Fig. 4. displacements are found from eqs. 4 and 6b
and c:
"x
~ = Jh/2 Oxzdr, Qy = Jh/2 oyzdz -- (6 )
aMxy
Q =--+-=
aMx a
D-(-
a2 w
+-) - - -
a2 w
(9a)
-h/2 -h/2 x ay ax - ax2 a-x ay2
Let q be the transverse load per unit area _ aMy aMxy a a2 w a2 w

Qy - ay - ax -- D ay (ax 2 + ar) - (9b)
acting normally to the upper face of the
plate. Considering the force equilibrium in
the z-direction of the element of Fig. 4 we The above analysis is sufficient to seek
find: solutions of specific problems. The general
procedure is to find approximate solution of
aQ,x aQy the fourth order differential equation (8)
- - dxdy + - - dydx + qdxdy = 0
ax ay
or = which satisfies the given boundary displacement
and force conditions.
aQx aQy
-- + -
ax ay
+ q =0 - - - - - (6a)
A18. 12 Strain Energy in Pure Bending of Plates.
In evaluating the strain energy of a thin

Taking the equilibrium of the moments acting plate we shall ignore the contribution of the
in the X-direction we obtain: shear strains which are generally small for
small deflections.
A18.13
dVx + dVy + dVxy + dVyx

Vo = dxdy =
aw+a"w dx
ax a x"
-(10)
z SUbstituting Mx ' My, M~y in terms of the dis-

Fig. 5a placement from eqs. (5) we find:
1 a"w a"w"
V0 .. "2 D (ax" + ay") -
ow a"w
...-ay+a xOy dx
a"w a"w
2(1-V) [ ax" • ay" - (axay)
a"w "J - - - - -(11)
2 ~ This expression will be modified later when we

Mxy will consider the superposition of compressive
loads in the plane of the plate which are
Fig. 5b related to the problem of plate buckling.
The strain energy stored in a plate A18. 13 Bending of Rectangular Plates.

element is obtained by calculating the work
done by the moments Mxdy, and Mydx on the The general differential equation for
element during bending. Since the sides of bending plates was given in section C2.3
the element remain plane during bending the (eq. 8). Two very useful methods of solution
work done by Mxdy is obtained by taking half have been widely used, namely, the Fourier
the product of Mxdy and the relative angle Series Method and the Energy Method. Both
of rotation of the two sides of the element. methods will be developed in the following
Since the curvature in the x-direction is for rectangular plates and various edge-
a"w supporting conditions. The edge support
- ax" , the relative angle of rotation of the conditions are classified as follows:
sides 1 and 2 of distance dx will be - a"w
ax" dx.
a) BUilt-in edge or Fixed: The deflection
Thus the work due to MxdY is: along the bUilt-in side is zero and the tangent
1 a"w plane to the deflected middle surface is
dVx = - 2" Mx ax" dxdy - - - - (a) horizontal. Thus if for instance the x-axis
coincides with the bUilt-in edge these con-
Similarly the work due to Mydx is: ditions are:
dVy = - 2"1 My a"w

ax" dxdy - (b) (w)y = 0 =0 = 0 - - - -(12a)
b) Simply supported edge: The deflection

The tWisting moment MxydY also does work along the simply-supported side is zero and the
against rotation of the element about the bending moment parallel to this side is also
x-axis. The relative angle of rotation of the zero. Thus if the plate is simply supported
a"w along the x-axis we have:
two sections 1, 2 is obviously axay dx. 'rhUS
the work done by MxydY is:
(w) _ =0
y-o
1 a"w
dVxy = "2 Mxy axay dxdy - - (c)
c) Free edge: The bending moment,
tWisting moment and shear force along the free
and the work due to Myxdx =Mxydx is: side is zero. Thus if the free side coincides
With the straight line x =aL we have:
a"w
dVyx =2"1 Mxy axay dxdy - - - (d)
(Q,x) =0
x=a
(It is noted that the twist does not affect
the work produced by the bending moments, However, as was proved by Ki rchoff two bourldary
neither the bending affect the work produced conditions are only necessary to find a unique
by the torsional moments).
solution of the bending problem. He has shown
that the two last equations of the above
Thus the total strain energy per unit conditions can be replaced by one condition.
volume of the plate is:
( a"w (2 V) a"w) = These boundary conditions are satisfied if we

ax" + - axay" x=a 0 take:
00 00
Expressing the condition (Mx)x=a = 0 in terms W = Z Z Cmn sin m~x sin n~y - - - (14a)
of w we find the final form of the boundary m=l n=l
conditions along the free edge:
By sUbstituting in eq. (18) with q given by
eq. (13) we find:
00 00
(12c) Z Z Cum
m=l n=l
In the following solutions for various 1 00 00
edge conditions will be developed. -D Z Z amn sin mroc sin nny - - - - - (14b)
m=l n=l a b
1. Simply supported rectangular plates
This relation is identity if:
Let a plate with sides a and b and axes
x, y, as shown in Fig. 6, be simply supported
around the whole periphery and loaded by a
distributed load q = f (x,y).
Two methods of solutions will be and thus:

developed:
1 00 00
W = -4 - Z Z amn sin mroc sin nny
a) Navier solution by means of double n D m=l n=l (m" + n"" a· b
Fourier Series: as l)2 )
- - - (14c)
We can always express f (x,y) in the form
of a double trigonometric (Fourier) series: In the case of a load qo uniformly distributed
over the whole surface we have:
00 00
q = f(x,y) = Z Z amn sin m~ sin n~y -(13a)

f (x,y) = qo = const.
m=l m=l
_ 4qo jajb i mroc nny =
where: 4 j 0 a j 0 b f (x,y ) sina:-sinTdxdy
a mn = ab mroc nny amn - a:i.) 0 0 s n -a- • sin -b- dxdy
- - -(13b) n"mn - - - - - (1 5a )
169 0
where m, n are both odd integers. If either

r---r------.--+ X or both are even amn = 0 and SUbstituting in
a (14c) :
b 169 0 00 00
sinmm sin nny
W=--
neD Z L: a b
m=1,3,5 n=1,3,5 " n"
mn (~+ - ) "
a" b"
q
y
- (15b)
with maximum deflection at the center,
Fig. 6
m+n_
DO l
Wmkx - nS D ~ h (-1) 2
The boundary conditions are:
m=1,3,5 n=1,3,5
W = 0 Mx = 0 at x = 0, x = a
W = 0 My = 0 at y = 0, y = b or:
- - - - - (15c)
(1) W = 0 (2) a"w = 0 at x = 0, x = a
ax" This is a rapidly converging series and a
a"w _ satisfactory approximation is obtained by
(3) ~ = 0 (4) ay" - 0 at y = 0, y = b taking only the first term. For a square
A18.15
plate this approximation becomes:

T
b/2~a~
which is by 2-1/2% is error with the exact

t
b/2
I------+----i~ x
Fig. 7
solution. 1
The expressions for bending and tWisting
moments are not so qUickly convergent. To y
improve the solution another series solution
can be developed as follows: where for symmetry m = 1,3,5 •
b) Levy alternate single series solution: This equation can be satisfied for all values
of x if:
The method will be developed for uniform
load qo = const. Levy suggested a solution in
the form:
00
W = Z Ym (y) sin mme

a
- - - - - - - (16) The general solution of (19b) is:
m=l
4
qa [Am cos h mny + Em mny sin h mny
Ym (y) = ---D
where Ym is a function of y only. Each term a a a
of the series satisfies the boundary conditions
a2 w + em s i n h
mny + T\..._
---a- mny
~Ill -a- cos
h mnYJ -a
w = 0, ax 2 = Oat x = a. It remains to
determine Ym so as to satisfy the remaining
-(20)
two boundary,conditions w = 0, ~;: = ° at y = b.
Since the deflection is s~~etric with
A further simplification can be made if respect to the x-axis it follows that em =
we take the solution in the form Dm = 0. Thus:
w = w~ + W2 - - - - - (17a)
where
(Am cos h mny + Em mny sin h mny ) sin mme
w - - (17b) a a a a
is the deflection of a very long strip with or

the long side in the x-direction loaded by a
qa (4 A h mny Em mny
uniform load q 0 supported at the short sides W =---n-
4
Z n5m6 + m cos -a- + a
x = 0, x =
a, and free at the two long sides. m
Since (17b) satisfies the differential nny mme
equation and the boundary conditions at x = 0, sin h --) sin--
x = a, the problem is solved if we find the a a
solution of:
where m 1,3,5 = SUbstituting this
expression into the boundary conditions:
(18 )
w = 0, for g = +- 2"b we find:
with w2 in the form of (16) and satisfying
together with w~ of eq. (17b) the boundary
a-22 Am cos h urn + am Em sin h urn + n 6m6
4
=°
conditions w = 0 , -ay
w
°
= at y = -+ -2b (see
Fig. 7). (An + 2Em) cos h am + urn Bm
sin h Om = °- -(21a)
Substituting (16) into (18) we obtain: where um = mnb/2a

00
Z
From these equations we find:
m=l
- -(19a)
2(am tan h am + 2)
Am =-------- la 00
m-l
rr 6 me cos ham (M)
x max
= qa8 + (l-V)qa"n la Z (-1) 2
2
m=1,3,5
Bm=----- - - - - - - - - (21b )
Thus:
00
m-l
W
4qa 4
=-- Z
2-
[1 _ (am tan h am + 2 ) (-1) 2
neD m=1,3,5 me 2 cos hOm
cos h 2amY + am • 2y sin h 2amY ] m (Am + l~V

la
Bm) - - (2lf)
b 2 cos h am b b
c) Solution by means of the principle
sin mnx - - -(21c) of virtual work
a
The maximum deflection occurs at the middle From the discussion of the method (a) we
can represent w in double Fourier Series:
- a
x-Z·y= 0:
00
m-l
00
mnx nny
00 W= Z Z Cmn sin -a- sin -b- - - - ('22a)
'-1) 2 ~_"m tanho.,+2\ m-l n=l
m=1,3,5 me 2 cos ham -)
The coefficients Cmn may be considered as the
co-ordinate defining the deflection surface.
- - - - - - - -(21d) A virtual displacement will have the form:
The summation of the first series of terms mnx nny
corresponds to the solution of the middle of 6w = 6Cmn • sin ---a- . sin t ) - - - (22b)
a uniformly loaded strip, eq. (17b). Thus:
00 m-l The strain energy Vi can be found by substi-
_ 5 qa 4 4qa 4 Z (-1 )---z-
Wmax - 384 D - -3 - tuting (22a) into eq. (6). After a few
n D m=1,3,5 me algebraic manipulations we find:
Om tan h Om + 2 00 00
~a fa b Vodxdy = n :b D
4
- - - - - -(21e)
2 cos ham Vi = Z Z emn la
m=l n=l
This series converges very rapidly. Taking a la
square plate, alb = 1, we find from (21a): (m + .!f-)la_ (23)
a" b"
Let us now examine the deflection of the plate

of Fig. 6 with a concentrated load P at the
_ 5 qa 4 4qa 4 point with co-ordinates x = { , y n. The =
Wmax - 384 D - - - (0.68562 - 0.00025 + •••• ) increment of strain energy due to the incre-
n6D ment of the deflection by:
~ qa 4 mnx nny
= 0.00406 D
6w = 6Cmn sin-
a - sin -b- - - - - - - (24a)
We observe that only the first term of the is found from (23)
series in (21e) need to be taken into con-
sideration.
The bending moments are found by substi-

tuting (21c) into eqs. (5). The maximum The increment of the work of the load Pis:
bending moments at x =~ , y = 0 are:
6W = P6Cmn sin mf sin n~?7 - - - - - - (24c)
From 6Vi - 6W = 0 we obtain:

A18.17
P sin m;:t . sin n~YJJ Let us consider the case of a uniform load q.
we write the deflection in the form:
6Cmn = ° -(26a)
Since 6C mn is arbitrary its coefficient must where Wi is the deflection of a simply
be zero. Thus: supported strip of length, a, which for the
system of axes of Fig. 8 can be written (see
sin ~
Levy's method in previous section):
4-P a . s i n nnn
b
Cmn = ----"'--,;2---;;-2-"---
n 4abD rl!!- + ..!1.-) 2
'a 2 b2
00 sin mng sin nn!] and w2 is represented by the series:

~
00
W 4-P ~ a' b
= n 4abD m=l n=l m2 n2 2
00
(""8:"2 + b 2 w2 = ~ Ym (y) sin mnx - - - -(26c)

)
a
m=1,3,5
mnx nny
sin -a- • sin -b- - - - - - -(24d)
where w2 being a solution of
This series converges rapidly. For a square , a4 W2 a4 w2

plate (alb = 1): +- -
a y4 + 2 ax 2 ay 2 =0
00
1
~ ~ - - - -(24-e) is found as in the previous section:
m=l n=l
_ qa4 mny +
Ym - -D (Am cos h mny
a + R mny
a sin h--
a
By taking only the first four terms we find '-'!Il
Cm sinh mnY+n mny cosh mny) - -(26d)

Wmax =: 0.01121 p~2 a '-1Il a a
It is obvious that the two first boundary

which is 3-1/2% less than the correct value.
conditions are identically satisfied by w =
(2) Rectangular plate with two opposite Wi + w2 • The coefficients Am, Em, Cm, Urn
edges simply supported, the third edge free must be determined so as to satisfy the last
and the fourth edge built-in or simply four boundary conditions. Using the conditions
supported. (25b) we obtain:
H-a-7
r-.,..--------.-_x
b
By the conditions (25c) we find:
t Bm = _4_.
nllm ll
(3+J)) (l-V) cos h" Bm + 2V cos hBm -J)(l-VIBm sin h Bm-(l-J)")
(3+V) (I-V) COS h ll Em + (l_V)"Bm ll + (l+V)2
- - - -(268)
em = _4_. ~(3::.::+JJ:..c)-C.:(1:.-:-V,!-):,,:S h Bm~C:.:::OS:...:h:.:Bm~+J):.:-(c:.-
n ::-
ic:: 1 +..:...J))c::.S:.::inc:::h~Bm,,--=-:V(-=-l-..:...V:.::IBm=CO_S..:...h-=.!Bm,,---,--(l--=-V--,-)"~Bm
11 II mil (3+11) (l-V)cos h iii Em + (l_V)liI Elms + (1+V)liI
Fig. 8
y
Assume that the edges x and x a are =° = SUbstituting Am, Em, Cm and Dm in eq. (26d) we
simply supported, the edge y = b free and the find the deflection. The maximum deflection
°
edge y = bUilt-in (Fig. 8). In such a case
the boundary conditions are:
occurs at the middle of the free edge.
A18. 14 Combined Bending and Tension or Compression of
Thin Plates.
w = 0, ~:'::" = 0, for x = 0, x = a (a)
In developing the differential equations
w = °, aw = 0, for y = (b) of eqUilibrium in previous pages, it was
2
ay
a 2 w = 0,
° - - (25)
assumed that the plate is bent by transverse
loads normal to the plate and the deflections
(a w+
ax2 a y 2) were so small that the stretching of the middle
plane can be neglected. If we consider now the
t~~~ + 2(1-V) = ° for
3
a w] y =b (c) case where only edge loads are active coplanar
ax2 ay
A18.l8 THEORY OF THE INSTABILITY OF COLUMNS AND THIN SHEETS
with the middle surface (Fig. 9) we have a r----------?~ x

plane stress problem. If we ass~me that the
stresses are uniformly distributed over the Nx
dx
thickness and denote by Nx ' Ny, NXY ' Nyx the
resultant force of these stresses per unit
length of linear element in the x and
y-directions (Fig. 9) it is obvious th8t:
Nx = hO x , Ny = hOy
- - - - - - - - (27) y
Nxy = Nyx = hoxy = hoyx Fig. lOa
Consider an element of the middle surface
dX, dy (Fig. lOa). The conditions of force
eqUilibrium in the x, y-directions are given
byeq. (28). Consider now the projection of
the stresses Nx , Ny, Nxy in the z-direction:
(a) Projection of Nx : From Fig. lOa it

follows that the resultant projection is:
y
and neglecting terms of higher order:
Fig. 9
(N a2w+~ aw)
The equations of equilibrium in the absence x ax 2 ax· ax - - - - - - - - (a)
of body forces can be written now in terms of
these generalized stresses Nx , Ny, Nxy by (b) Projection of Ny: By similar argument
SUbstituting from eq. (2 7 ) to the equations we find that this projection is equal to:
of eqUilibrium
(N a2~ + aNy~) - - - - - - - - - - - (b)
aNxy y ay ay ay
ax - 0 - - - - - (28)
On the other hand i f the plate is loaded by

transverse loads the stresses give rise to dx ~
,,-====-----;:c;-_~z
pure bending and tWisting moments only. The ::::::::--. Nxy :'
equations of equilibrium for the latter have
been given in before (see eqs. 6, 7, 8). If
both transverse loads and coplanar edge loadS
are acting simultaneously, then for small
vertical deflections the state of stress is
the superposition of the stresses due to
Nx ' Ny, Nxy and Mx ' My, MXY . For large
vertical deflection of the plate, however,
there is interaction of the coplanar stresses
and the deflections. These stresses give rise
to additional bending moments due to the non-
zero lever arm of the edge loads from the Fig. lOb
deflected middle surface, as in the case of (c) Projection of Nxy and Nyx: From Fig. lOb
beams. When the edge loads are compressive we find:
this additional moments might cause instability
and failure of the plate due to excessive aw ( aNxy ) (aw + a 2W dx)
vertical deflections. - Nxy ay + Nxy + ax dx ay axay
In this chapter the problem of instability and neglecting terms of higher order:
of plates will be examined.
(N ~+ aNxy aw) dxd - - (c)
When the edge loads are compressive and xy axay ax· ay y
give rise to additional bending moments eq. (8)
must be modified. Simil~rly we find the projection of Nyx = Nxy :
A18.19
(Nxy ~+
axay
aNxy . aw) dxd
ay ax y - - - - - - - (d) ViN =~ / / (NXE~ + NyE; + Nxyfxy) dxdy =
Thus in eq. (6a) the terms given by (a), (b), 2~JJ (N~+N;-2IJNxNy+2(1+V)Nxy") (a)
(c) and (d) should be added (divided of course
by dxdy): During bending due to transverse loads
or/and due to buckling we assume that the edge
aQ.x + aQ.y + + 1\1 a"w N a"w + 2N a"w loads and consequently Nx , Ny, Nxy , remain
ax ay q "x ax" + Yay" xy axay +
constant. Its variation is thus zero and we
(aNX + aNxy) aw + (~+ aNxy)aw = 0 (e) do not consider it in the following. Let us
ax ay ax ay ax ay apply now the transverse load that produces
bending. (We can also consider bending due to
But due to the equation of equilibrium (28) the other transverse disturbance, which is the
two terms inside the parentheses in (e) are case of buckling). If u, v, are the displace-
zero. Thus: ments of the middle surface due to the coplanar
loads (which are assumed constant across the
aQ.x + ~ + q + N a"w + thickness) and w the bending deflection of the
ax ay x ax" plate it can be shown that the strains are:
a"w a"w
Ny ayr + 2Nxy axay =0 - - - - - -(29)
EX
= ~+l(aw)"
ax 2 ax
Eq. (29) replaces eq. (6a) when edge loads are
present. Eqs. (6b) and (6c) are, however, still Ey
=~ay
+.l (aw)"
2 ay
valid since they express moment equilibrium
of the element dxdy in which the contribution '0
xy
= au
ay
+ av + aw
ax ax·
aw
ay - - - - - - (b)
of Nx , Ny, Nxy is zero. Thus eliminating Qx ,
Q.y between (6b), (6c) and (29) we find: Let us apply now bending with constant coplanar
stresses. Due to stretching of the middle
surface the energy is:
/ / (NxEX + NyEy + Nxy 0xy) dxdy - - (c)
Introducing (b) into (c) and adding the strain

energy due to bending, eq. (11), we find the
Eq. (30) replaces eq. (128) when edge loads total change of strain energy due to bending
are present and the deflections are large so
that instability might occur. which is:
The distribution of the coplanar stresses

Nx , Ny, Nxy can be found from eqs. (28) by
solving the plane stress problem. In the
following the above theory will be applied to
rectangular plates.
A18.15 Strain Energy of Plates Due to Edge
Compression and Bending.
The energy expression for pure bending,

eq. (11), must be complemented to include the Here u, v are the additional coplanar displace-
contribution of the edge coplanar loads. ments after bending has started. It can be
Assume that first the edge loads are applied. shown by integrating by parts that the first
Obviously the strains due to the stresses integral is the work done during bending by the
Nx , Ny, Nxy are: edge loads. For instance taking a rect~ngular
plate this integral becomes:
iai b [N aU+ N ~+N (au+av)J dxdy =
o 0 x ax y oy xy ay ax
=NhGxy - - - - - - - - - - -(31)
fa b (I Nxul ~ + INxyvl ~ (I Nyvl ~ + INxyul ~ )dx-
0
If xy ) dy + fa a
The strain energy is:
faa£ b{U(O:xX + o~~Y) + v (0:: + a~~y)} dxdy - - (e)
Obviously the first two integrals represent buckled configuration (deflections different
the work done by the edge loads while the from zero) is possible. It was found in the
second integral is zero due to the equilibrium case of column that this latter solution
equations (28). Thus the work of the edge (Euler's solution) approaches asymptotically
loads is: the first at the limit where the deflections
become extremely large but for even small
deflections the edge load acquires a value
very near the Euler's critical value. The
latter technique is mathematically more
We assume now that for small deflections convenient and it gives for plates also a very
the stretching of the middle surface of the good estimate of their compressive strength.
plate is negligible. (This is the so-called In the following we shall use this latter
inextensional theory of plates). In this case approach by assuming a plate with edge loads
by zeroing the strains in eq. (b) and substi- and no transverse load. Eq. (30) becomes in
tuting in (f) we find: this case:
(32a) (33)
In the strain energy expression, eq. (a) By solving eq. (33) we will find that the
the first two terms cancel each other and the assumed buckling made is possible (w 1 0) for
strain energy is due only to bending: certain definite values of the edge loads, the
smallest of which determines the critical load.
Vi
2
= 21 D!! l,(aax2w2 + aay2
w)2 _ The energy method can also be used in investi-
gating buckling problems. In this method we
2 assume that the plate is initially under the
a w ay"
2(1-21) [ ax2. a~ "]} dxdy - - (32b)
a2w - (axay) plane stress conditions due to the edge loads
and the stress distribution is assumed as known.
We then consider the buckled state as a possible
In the absence of transverse loads the work configuration of equilibrium. The change of the
of external forces is simply due to the edge
loads: work is given by eq. (32a). We interpret here
w as a virtual displacement though we do not use
the variation symbol o. Thus the increment of
(32c) work oW is given by (32a) and the increment of
Expressions (32b) and (c) will be used in strain energy oVi is given by eq. (32b). If
solving the buckling problem by means of the OW<oVi for every possible shape of buckling the
principle of virtual work. flat equilibrium is stable. If OW>oVi for a
certain shape of buckling then the flat con-
A18. 16 Buckling of Rectangular Plates with Various Edge
figuration is unstable and the plate will buckle
Loads and Support Conditions. under any load above the critical load. If
oW = oV i , the equilibrium is neutral and from
General discussion this equation we find the critical load. The
critical load therefore is found from the
equation:
In calculating critical values of edge
loads for which the flat form of equilibrium
becomes unstable and the plate begins to
1 ! ! [ Nx (ax)
- 2: aw" + Ny (ay)
aw " + 2Nxy aw aw] dxdy
ax . ay =
buckle, the same methods and corresponding 2
reasonings as for compressed bars will be ~2 !!{(a"w + a"w)" _ 2(1-))) [a w . a"w _ ( a"w )2l}dXd
ax" ay2 ax" ay2 axay J y
employed.
- - (34)
The critical values can be obtained by
assuming that the plate has a slight initial Here w is a certain assumed deflection which
curvature or a small transverse load. These satisfies the boundary conditions (Virtual
values of the edge loads for which the lateral deflection) .
deflection w becomes infinite are the critical
values (see Part 1 for similar treatment in A18.l7 Buckling of Simply Supported Rectangular Plates
columns) . Uniformly Compressed in One Direction.
Another way of investigating such Let a plate of sides a and b (Fig. 11)
instability is to assume that the plate buckles simply supported around its periphery be com-
due to a certain external disturbance and then pressed by load Nx uniformly distributed
to calculate the edge loads for which such a along the sides x =
0 and x =
a. From the
A1B.2l
__ ....,...-------.-.-~ X
a + c + e +
- - - - (b)
b + d + f +
This fraction has some intermediate value

Nx
between the maximum and the minimu~ of the
fractions (a). It follows that if we wish to
make the fraction (b), which is similar to the
fraction of eq. 35d, a minimum, we must take
Fig. 11
y only one term in the numerator and the corres-
ponding term in the denominator. Thus to make
obvious solution of the corresponding plane the fraction of eq. (35d) minimum, we must put
stress problem we find that the state of all the parameters Cu , CJ.2, C 2 J. • • • • except
stress is everywhere a simple compression one, zero. This is equivalent to assuming that
equal to Nx (the load at the periphery). the buckling configuration is of simple
sinusoidal form in both directions, i.e.
The deflection surface of a simply mnx nny
supported plate when bending takes place have Wmn = Cmn sin -a- • sin -b-' The minimum
been found previously (see eq. 14a). Its expression for Cmn obtained by dropping all the
general expression can be written in a double terms except Crnn becomes:
series form:
00 00 - - (3.5e)
W= Z Z Cmn sin mnx sin n~y (35a)
m=l n=l
a
It is obvious that the smallest value of Nx is
The increment of strain energy found by obtained by taking n = 1. This means that the
plate buckles always in such a way that there
substituting (35a) in the right-hand side of
can be several half-waves in the direction of
(34) is:
compression but only are half-wave in the
perpendicular direction. Thus for n = 1, eq.
7 (' ~rnn
2 (ma 2 + ~)2
2
b2
(35b) (35e) becomes:
n=l
The increment of work done by the external

forces is found by sUbstituting (35a) into the
left-hand side of eq. (34) and Nx = const, The value of m (in other words the number of
Nx = Nxy = O. Thus: half-waves) which makes this critical value
the smallest possible depends on the ratio
OW=l.N
2. x 10 a10 b (aw)'"dxd =
ax y
alb and can be found as follows:
n2 b 00 00 Let us express (36a) in the form:

-sa Nx z Z m2 Crnn 2 - - - - - - (35c)
m=l n=l
- (36b)
From the equality oW = OVi, solving for Nx , we
obtain: a
where k is a numerical factor depending on (t)'
From (36a) and (36b) we have:
00 00 2 2
n2 a2 D Z Z Crnn 2
(1!L + ..!L)
a2 2 2
b
Nx = m=l n=l
- - - (35d)
k =~
2
a
(m + 1:..
m b 2
~)2 (36c)
00 00
z Z m2 Crnn 2
If
a
we plot k against b for various values of the
m=l n=l
integer ill = 1,2,3, ..... we obtain the curves
Here Cmn is arbitrary. We are interested, of Fig. (12). From these curves the critical
however, to find that values of Cmn which make load factor k and the corresponding number of
Nx minimum. To that effect we use the follow- half-waves can readily be determined. It is
ing mathematical reasoning: only necessary to take the corresponding point
a
Imagine a series of fractions: (j)) as the axis of abcissas and to choose that
curve which gives the smallest k. In Fig. 12
ace the portion of the various curves which give
b ' d ' f ' · .... · .. - - - - - - - (a)
the critical values of k are shown by fUll
lines. The transition from m to (m + 1) half-
If we add the numerators and the denominators waves occurs at the intersection of the two
we obtain the fraction: corresponding lines. From eq. (36c) we find:
A18.22 THEORY OF THE INSTABIl.ITY OF COl.UMNS AND THIN SHEETS
k \
7
6
\ '",3\
\ m=2 \
\
1\ m=4,/ "
/ \ / /
,/
\
" ,,"
,/,/
5
~ '\. /"/
4
~ ¥;:,<.../
I
3
I
I I ,I
I
2
r
I ,
I
I
I I
,
,,
I
1 I ,
o ! 0
, Y Fig. 13
o 1j2 2y16 3/12 4 5 alb
22 22 22 22
m n + N B....B.-
Nx~ = D(.!!!..lL. +B....B.-) 2 _ (37a )
Fig. 12 y bB a2 b2
mb a (m+l)b a a~d by introducing the parameter:

-+-= + --- or:
a mb a (m+l)b
- - - - - - - - - - - - (37b)
a =
b
Jm (m+l) - - - (36d)
We obtain:
Thus the transition from one to two half-waves
occur for:
..§:.
b
=J 1 (1+1) = ,f2 Taking any integer m and n the corresponding
deflection surface of the buckled plate is given
by:
from two to three for
mnx . nny
Wmn = Cmn sin -a-· SIn -b- (38)
a =
b
J2 (2+1) = ~
and the corresponding ox, 0y are given by (37c)
and so on. which is a straight line in the diagram 0y' Ox
(Fig. 14). By plotting such lines for various
The number of half-waves increases with the pairs of m and n we find the region of stability
ratio alb and for very long plates m is veFy and the critical combination of ox, 0y which is
large. on the periphery of the polygon formed by the
full lines of Fig. 14.
This means that a very long plate buckles in

half-waves the lengths of which approach the
width of the plate. The buckled plate is
subdivided into squares.
The critical value of the compression

stress is:
2
(Nx)cr = kn E - - - (36e)
°xv =
h 12(1-))2)
(t = thickness) Fig. 14
A18.18 Buckling of Simply Supported Rectangular Plate A18. 19 Buckling of Simply Supported Rectangular Plate
Compressed in Two Perpendicular Directions. Under Combined Bending and Compression.
Let (Fig. 13) Nx , Ny the uniformly distri- Let us consider a simply supported
buted edge compressions. Using the same as rectangular plate (Fig. 15). Along the sides
before expression for the deflections (eq. 35a) x = 0, x = a there are linearly distributed
and applying the energy equation (33) with edge loads given by the equation:
Nx , Ny = constants (which is the solution of
the corresponding plane stress problem) we find:
Nx = No (1 - A ~) - - - - - - - - - (39a)
A18.23
which is a combination of pure bending and We examine for each value of m the solutions of
pure compression. Let us take the deflection the system (38f). Starting from m = 1 and
again in the form: denoting:
00 00
mnx nny = (No)cr
w Z Cmn sin a-' sin -b- - - - (39b) - - - - - - - - - - (39g)
n=l h
we obtain from (39f):
These are homogeneous equations in a 11 , a 12 '"

y etc. The system possesses a non-zero solution
Fig. 15 (which indicates the possibility of buckling of
the plate) if the determinant of eq. (39h) is
zero. From this condition an equation is
SUbstituting in the right-hand side of obtained for ocr' ObViously the system (39h)
eq. (34) we find the variation of strain is of infinite number of equations (n = 1, 2,
energy: 3 ... ). A sufficient approximation is
obtained by taking a large but finite number
- - - - - (39c) of terms and find the solution of the finite
determinant (USing for example digital
computers). Thus a curve of ocr versus alb is
while the increment of work is: obtained for m = 1 like that of Fig. 12.
Repeating the same calculation for m = 2, 3
etc., we find similar curves of two, three, etc.
half-wave lengths. The regions of the curves
N {ab 0000 m2n 2 )a 0<:> m2n 2 with minimum ordinates define the region of
.....Q
- Z Z Cmn2 -2- - _1\- Z stability as in Fig. (12).
4 2 a b m=l ~
A18.20 Inelastic Buckling of Thin Sheets
-
b
Z C
00
---Z 2
8b 2 0<:> ~ Cmn Cmi ni]} (39d)
[ 4 n=l mn n'2 n=l
L> - -
i=l (n 2_i 2 )2 The problem of the inelastic buckling of

thin sheets has been extensively studied by
various authors. The main difficulty in such
Equating (39c) and (39d) and solving for No studies is in reference to the stress-strain
we find: relations of plasticity under complex states
of stress. Many controversial discussions
have appeared in literature without resolving
the theoretical difficulties. For this reason
we will not develop the theory of inelastic
buckling in this chapter. Some of the better
references on this sUbject are listed below.
- - - - (3ge)
Chapter C4 presents the plasticity
The coefficients now Cmn are so adjusted that correction factors to use in calculating the
(No)cr becomes minimum. By taking the inelastic buckling strength of thin sheets.
derivative of expression (38e) with respect
to each coefficient Cmn and equating these to
A18.21 References.
zero we find:
2 2 2 2 (1) Bleich, F: Buckling Strength of Metal
DC mn n4 (2:2 n 2= (No) cr {
m + 1)2) m
Cmn n A. m2n 2
~ - 2" ---a" Structures. Book by McGraw-Hill.
(2) Stowell, E.Z., A Unified Theory of Plastic
Buckling of Columns and Plates. NACA
c 16 7Cmi ni ]} (39f) Report 898, 1948.
[ mn - ; ; i (n2_i2)2 (3) Gerard and Becker: Handbook of Structural
Stability. NACA T.N. 3781, 1957.
(4) Gerard: Introduction to Structural
Stability Theory. Book by McGraw-Hill
Co., 1962.
AIB.24
PART 3
STRENGTH OF COLUMNS WITH STABLE CROSS-SECTIONS
(BY PROFESSOR E. F. BRUHN)
A18.22 Methods of Column Failure. Column Equations. The range AB in Fig, 16 is for a range
of Lip values of below 20 to 25, and repre-
In Part 1, the theory of the elastic and sents a range where failure is due to plastic
inelastic instability of the column was pre- crushing of the column, In other words, the
sented, The equations from Part 1 for a pin column is too short to buckle or bow under
end support condition are:- end load but crushes under the high stresses.
This column range of stresses is usually
For elastic primary failure, referred to as the block compression strength.
n"E A column, however, may fail by local

Fc = --- - - - - - - - - - (1) buckling or crippling due to distortion of
(Lip )" the column cross-section in its own plane.
The horizontal dashed line in Fig. 16
For inelastic primary failure, represents the condition where the primary
column strength is limited by the local weak-
n"Et
Fc = - - - - - - - - (2) ness. This line moves up or down according
(Lip )" to the value of the local weakness. The
determination of the column strength when
Where Fc =
compressive unit stress at failure is due to local weakness is covered
failure =
piA stress, in volume 2.
E = Young's modulus
A18.23 Column End Restraint. Fixity Coefficients,
Et = tangent modulus Column Effective Length.
L = column length The column strength is influenced by the

p = radius of gyration of cross- end support restraint against rotation and by
section any lateral supports between the column ends.
The letter c is commonly used to indicate the
Fig, 16 shows a typical plot of Fc versus end fixity coefficient, and C = 1.0 for zero
Lip, If the column dimensions are such as to end restraint against rotation, which can be
cause it to fail in range CD in Fig, 16, the produced mechanically by a pin or ball and
primary failure is due to elastic instability socket end support fitting. Thus including
and equation (1) holds. This range of Lip the end restraint effect equations (1) and (2)
values is often referred to by engineers as the can be written,
long column range,
A B C D
Fc =
cn"E
, Fc = cn"Et - - - - - (3)
(Lip )" (Lip)"
I I I I
Local Crippling Limit Let.L' = effective length of the column
which equals the length between inflection
points of the deflected column under load.
Then L' = Live (4 )
Thus equation (3) can be written as,

o L'/p
Fig. 16 Fc = n"E
Fc = n"Et --- (5)
(L'/p)" (L'/p) "
The range BC represents the range of L~ If we let P = failing or critical load,
values where failure is due to inelastic in- equation (5) can be written as equation (6)
stability of the column as a whole and equation
(2) applies, This range BC is often referred by realizing that P = FcA and p = vIlA.
to as the short column range, n"EI n"EtI
P = (L')" P =(L!)2 - - - - - - (6)
A18.25
A1B.26 STRENGTH OF COLUMNS WITH STABLE CROSS-SECTIONS
Fig. 17 shows the deflected column curve TABLE A1B.1

under the load P for various end and lateral
supDort conditions. The effective lengths L' 17-7 PH(TH1050) Stainless Steel Sheet
and the end fixity coefficients are also listed. Ftu = 180,000 E = 29 X 10 6
F cy = 162,000
Et* L' /p =11."; Et/Fc
{T - { t' f { ( 20,000 29 x 10 6
119.56
- ;-T ---'I - -r 30,000

40,000
29
29
x 10
x 10
6
6
97.62
84.54
I L L L'
12"=L' r I l
IL=~ I
.70L=L'
1-
50,000
60,000
70, 000
80,000
29
29
29
29
x 10
x 10
x 10
x 10
6
6
6
6
75.62
69.03
63.91
59.78
I 2 I
J
89, 600 29 x 10 6
56.49
96,000 27.55x10 6
53. 19
L/4 102,000 26.10 x 10 6
50.22
--"f~ -'f~ 107,400 24.65 x 10 6

47.57
112,200 23.20 x 10 6
45.15
117,000 21. 75 x 10 6 42.81
C=4 C=1/4 C=2.05
121,500 20.30 x 10 6 40. 58
125,800 18.85 x 10 6 38.43
Fig. 17 130,000 17.40x10 6
36.32
134,000 15.95 x 10 6 34.25
A18.24 Design Column Curves for Various Materials. 137,700 14.50 x 10 6 32.22
147,000 11.60 x 10 6 27.89
For routine design purposes it is con- 158, 100 8.70 x 10 6 23.29
venient to have column curves of allowable 167,600 5.80 x 10 6 18.47
failing column stress Fc versus the effective 173,800 2.90 x 10 6 12. 82
slenderness ratio L' /p. In equation (5) we
will assume values of Fc , then find the *Values of Et calculated from stress-strain
tangent modulus Et corresponding to this diagram for this material.
stress and then solve for the term L' /p.
Table A18.1 shows the calculations for
17.7 PH (THI050) stainless steel sheet at
180
room temperature. The results are then plotted
in Fig. 18 to give the column strength curve. 170
Similar data was calculated for the material
under certain exposure time to different 160
elevated temperatures and the results are also 150

plotted in Fig. 18. Figs. 19 to 22 give
column curves for other materials under various 140
temperature conditions.
130
A18.25 Selection of Materials for Elevated Temperature 120

Conditions.
110
Light weight is an important requirement 100

in aerospace structural design. For columns
that fail in the inelastic range of stresses, .~
90
a comparison of the FCY/W ratio of materials 80

gives a fairly good picture of the efficiency
of compression members when sUbjected to 70
elevated temperature conditions. In this
60
ratio Fcy is the yield stress at the particular
temperature and w is the weight per cu. inch 50
of the material. Fig. 23 shows a plot of
Fc /w for temperature ranges up to 600 0 F. 40
Wi~h 1/2 hour time exposure for several im- 30
portant aerospace materials.
20
A18.26 Example Problem.
10 20 30 40 50 60 70 80 90 100
PROBLEM L'/f'
Fig. 24 shows a forged (I) section mem- Fig. 18

ber 30 inches long, which is to be used as
A1B.27
170
7079-T6 Alum. Alloy Hand Forging
160 t" 6.0 In.
17 -4 PH Stainless ~~f!1 Temp. Exp. • 1/2 HR.
(Bars & Forgings) Ftu = 67000
Temp. Exp. up to 1/2 HR. Fey = 59000
Ftu = 180000 50
140 Fey = 165000
40
120
110
30
100
90
<.>
20
'"'
70
10
60
50
20 30 40 50 60 70 80 90
40
L'/P
Fig. 20
10 20 30 40 50 60 70 80 90 100
L'/f
Fig. 19
75 R. T. 150 R.T.
7075 -T6 Alum. Alloy Extrusions
t ~ 0.25 In. steel A1SI 4130, 4140, 4340
140 Heat Treated To F tu = 150000
Temp. Exp. = 2 HR.
F tu =75000 Exp. = 1/2 HR.
130 .145000
Feu = 70000
60
120
110
50
100
<.>
90
'"'
40
80
<.>
70
'"'
30 60
50
20
40
30
10 600 0 F 20
10
10 20 30 40 50 60 70 80 90 100 10 20 30 40 50 60 70 80 90 100
L'/P L'/f'
Fig. 21 Fig. 22
Values of Fe in KSI. The horizontal dashed line is the compressive yield stress. Values
above these cut-off lines should be substantiated by tests.
A18.28 STRENGTH OF COLUMNS WITH STABLE CROSS-SECTIONS
Fig. 23 Solution:
(1) AISI Steel, F tu = 180,000
Since the column may fail by bending about
(2) 17-7PH. Stainless Steel, F tu =210,000 either the X or Y axes, the column strength for
(3) 7075-76 Alum. Alloy bending about each of these axes will be calcu-
(4) A~31B Magnesium Alloy lated. Since the column strength is a function
of the radius of gyration of the cross-section,
(5) 6AL-4V Titanium Alloy
the first step in the solution will be the cal-
culation of Ix and I y , from which p x and Py can
be found.
800
~ (2) Calculating Ix: In Fig. 25 the section will
700~~ be first considered a solid rectangle 2.5 x
2.75 and then the properties of portions (1)
600 ~ and (2) will be sUbtracted.
g + +~
S 500 (3) (1) Ix (rectangle) = l~ x 2.5 X 2.75 3 = 4.32
:><
::: Portions (1) and (2)
';;.,
<.J
400 - -___
~ Ix = -112 x 1.5 x 1.25 3 - 4( .625 x .25 x 1.292 2 ) =
300 -1.29
200 (1 0 of (2) about its x centroidal axis is
negligible)
100
Ix = 4.32 - 1.29 = 3.03 in. 4
100 200 300 400

TEMP. of
500 600
PX = I!i:.. , Area A = 2.5x2.75-2x .75x1.25-
A
4x .25x .625) = 4.375 sq.
in.
a compression member. Find the ultimate
strength of the member if made from the
following materials and subjected to the given Px = V 3.03/4.375 = .83 in.
temperature and time conditions. Calculation of I y :
Case 1. Material 7079-T6 Alum. Alloy hand
1
forging and room temperature. I y (solid) = 12 x 2.75 x 2.5 = 3.58 3
Case 2. Same as Case 1, but sUbjected 1/2 Portion (1) = -(1.25x .75x .875 )2 -(1.25x
2
hour to a temperature of 300 oF. .75 3 /12)2 =- 1.52
Case 3. Same as Case 2, but for 600 oF. Portion (2) = -(.25x .625x .833 2
)4-4(.25X
3
1.25 /36) =- .488
Case 4. Material 17-4 PH stainless steel,
hand forging at room temperature. Iy = 3.58 - 1.52 - .488 =1.58 in. 4
Fig. 24 }- ----->=n
------
Py = J 1.58/4.375 = .60"
Column strength is considerably influenced
by the end restraint conditions. For failure
by bending about the x-x axis, the end restraint
against rotation is zero as the single fitting
30 bolt has an axis parallel to the x-x axis and
thus c the fiXity coefficient is 1. For
failure by bending about the y-y axis we have
end restraint which will depend on the rigidity
of the bolt and the adjacent fitting and
structure. For this example problem, this
restraint will be such as to make the end
fixity coefficient c = 1.5.
SEc.A-A
Fig. 25
AIB.29
For failure about x-x axis, these end conditions depend on the type of end
fitting or connection to the adjacent structure.
L' = L/ve= 30/VT= 30, L'/px = 30/.83 Furthermore, to save weight columns are often
tapered in cross-section, that is, a variable
=36 moment of inertia along the column length,
For failure about y-y axis, these equations (1) and (2) which assure con-
stant moment of inertia do not apply. Since
L' = 30/~= 24.6, L'/P y = 24.6/.60 elevated temperature conditions effect the
=41 strength of a material, the column strength of
a structural member is likewise influenced.
Therefore failure is critical for bending (See Fig&. 26, 27, 28).
about y-y axis, with L' /p = 41.
Since this Volume 1 is concerned primarily
Case 1. The material is 7079-T6 Alum. Alloy with stress and strain analysis of structure,
hand forging. Fig. 20 gives the failing the subject of column strength is only slightly
stress Fc for this material plotted against considered or presented. However, Volume 2,
the L'/p ratio. Thus using L'/p = 41 and the which deals primarily with materials and
room temperature curve, we read Fc = 50500 psi. strength analys i sand d es ign of structure,
Thus the failing load if P = FcA = 50500 x presents a very complete and broad coverage of
4.375 = 220,000 Ibs. column design, involving over 180 pages of
theory, design charts and many sample design
Case 2. Using the 3000F curve in Fig. 20 problems.
for the same L'/p value, we read Fc = 40,400,
and thus P = 40,400 x 4.375 = 177,000. AIB.2B Some Mechanical and Physical Properties of a
Few Typical Aircraft Materials.
Case 3. Using the 600 0 F curve, Fc reads 6100
and thus P = 6100 x 4.375 = 26700 Ibs. Thus In order to determine the true deflections
subjecting this member to a temperature of or distortions df structures under a given load
600 0 F for 1/2 hour reduces its strength from system, the values of the stiffness properties
220,000 to 26,700 Ibs., which means that of the material used in the structure must be
Alum. Alloy is a poor material for carrying known. Chapters A7 and A8 involve the deflec-
loads under such temperatures since the tions of structures, thus the value of (E),
reduction in strength is qUite large. modulus of elasticity and (G), modulus of ri-
gidity must be known if true deflections are
Case 4. Material 17-4 PH stainless steel to be wanted. Table A18.2 gives these stiff-
forging. Fig. 19 gives the column curves ness values for a few common aircraft structur-
for this material. For L' /p = 41 and using al materials and also several other strength
the room temperature curve we read Fc = properties.
=
135,200 and thus P 135,200 x 4.375 =
591,000 Ibs. When a material is SUbjected to elevated
temperatures for various periods of time, the
AIB.27 Comprehensive Treatment of Column Strength. strength and ~tiffness properties are con-
siderably influenced as illustrated in Figs.
Columns with stable cross-sections fail as 26 to 28, which have been taken from Volume 2.
described in Art. A18.22. However, compression
members having open sections such as angle, For a comprehensive treatment of the be-
channel and zee shapes, often fail by tWisting havior of aircraft materials and their mechani-
or by local buckling or crippling of one or cal and physical properties, the reader is re-
more of the elements that make up the shape of ferred to Chapter Bl and B2 of Volume 2, where
the cross-section. Furthermore, the end fiXity over 50 pages are devoted to this important
condition influences the column strength and subject.
A18.30 STRENGTH OF COLUMNS WITH STABLE CROSS-SECTIONS
o.... Strength at temperature

Q)
Q. Exposure up to 1/2 hr.
E
Q)
to- 120
E
o
o
0:::
- 100
c
<.)
W
"0
c
0 E
w
80
::J
LL
>.
u:-
- 60
0
c=
Q)
....Q)
ll.
40
20
0'
-400 o 400 800 1200
Temperature, F
Fig. 26. Effect of Temperature on F tu ' F ty

and E of AISI alloy steel
TABLE A18. 2
W
Material Symbol Form Ftu Fty E Ec G
psi/lO psi/lO 6 psi/lO lbs./
KSI KSI 6 6
Cu. in.
AISI Alloy Steel 4140 Sheet-Tube 95 75 29.0 29.0 11.0 0.283

AISI Steel - Heat Tr. 4130 Sheet-Tube 180 163 29.0 29.0 11. 0 0.283
Stainless Steel 17-7PH Sheet 180 150 29.0 30.0 11. 0 0.276
Aluminum Alloy 2024 Sheet-Tube 62 40 10.9 10.7 4.0 0.100
Aluminum Alloy 7075 Extrusion 78 70 10.3 10.5 4.9 0.101
Magnesium Alloy AZ31B Sheet 32 15 6. 5 6. 5 2.4 0.064
Titanium Alloy 6Al-4V Sheet 130 120 16.0 16.4 6.2 O. 160
Inconel Steel X Sheet 155 100 31. 0 31. 0 -- 0.304
F tu = Ultimate Tensil Str. Fty = Yield Tensile Str.

E = Modulus of Elasticity. G = Modulus of Rigidity.
A18.3l
EFFECT OF TEMPERATURE ON 7075 ALUMINUM ALLOYS
100
S
E
!l BO
E
!!!
E
&
0 40
...
~
20
IOU 100
~ !
~ ~
I
!!!
80 100 200 300 400 500
Temperature, F
600 700 BOO !l
E
!!!
80
E 60 E 60
& Effect of exposure at elevated temper- ~
0 0
40
atures on the room-temperature tensile 40
~
yield strength (Fty) of 7075-T6 alumi- ...;.
~ 20
num alloy (all products). Iu 20
~ ~
100 200 300 400 500 600 roo 800 100 200 300 400 500 60D 700 800
Temperature, F 100
Temperature, F
'Effect of temperature on the ultimate i!l 80 Effect of temperature on the bearing

E
tensile strength (Ftu) of 7075-T6 !!! yield strength (Fbry) of 7075-T6
aluminum alloy (all products). j 60
aluminum alloy (all products).
0 40
...u
100 ~ 20
~ &
~ 80 100 200 300 400 500 600 700 800
~ Temperature, F
!!!
E 60
& Effect of temperature on the compres-
0 w" 60
40 sive yield strength (Fcy) of 7075-T6
~ aluminum alloy (all products). o
W 40
~
~
100 200 300 4CO 500 600 700 800 -400 -200 200 400 600 BOO 1000
Temperature, F 100 Temperature, F

~
~
~
BO
Effect of temperature on the tensile Effect of temperature on the tensile
yield strength (Fty) of 7075-T6 alumi- !!! and compressive modulus (E and Eel
E 60
num alloy (all products). 8 of 7075-T6 aluminum alloy.
'"0 40
...i!.
~
100 20
!
~ 80
~
!l 100 200 300 400 500 600 700 800
E
!!! Temperature. F
E 60
& Effect of temperature on the ultimate
0 40
shear strength (F su ) of 7075-T6
... aluminum alloy (all products).
~ 20
~
100 200 300 400 500 600 700 800 100 200 300 400 500 600 700 800
Temperature, F 100 Temperature. F
Effect of exposure at elevated temper-

atures on the room-temperature
t
!!!
80
Effect of exposure at elevated temper-
atures on the elongation of 7075-T6
ultimate tensile strength (Ftu) of 7075- j 60
aluminum alloy (all products except
T6 aluminum alloy (all products). 0 40
thick extrusions).
J.
I
u
of
100 200 300 400 500 600 700 800
Temperature, F
Fig. 28
Effect of temperature on the ultimate
bearing strength (Fbru) of 7075-T6
aluminum alloy (all products).
CHAPTER A19
INTRODUCTION TO WING STRESS ANALYSIS
A19.1 Typical Wing Structural Arrangement relatively thin. In general the wing structural
flange arrangement can be classified into two
For aerodynamic reasons, the wing cross-
types; (1) the concentrated flange type where
section must have a streamlined shape commonly
flange material is connected directly to in-
referred to as an airfoil section. The aero-
ternal webs and (2) the distributed flange type
dynamic forces in flight change in magnitude, where stringers are attached to skin between
direction and location. Likewise in the various
landing operations the loads change in magni- internal webs.
tUde, direction and location, thus the required
Fig. A9.3 shows several structural arran(~e
structure must be one that can efficiently
ments for wing cross-sections for supersonic
resist loads causing combined tension, com-
aircraft. Supersonic airfoil shapes are
pression, bending and torsion. To provide
relatively thin compared to subsonic aircraft.
torsional resistance, a portion of the airfoil
surface can be covered with a metal skin and
then adding one or more internal metal webs to Distributed Flange Type of Wing Beam.
produce a single closed cell or a multiple cell
Fig. d ~-----
~--------]
wing cross-section. The external skin surface
which is relatively thin for subsonic aircraft
is efficient for resisting torsional shear
stresses and tension, but qUite inefficient in
resisting compressive stresses due to bending
of wing. To provide strength efficiency, span-
wise stiffening units commonly referred to as Fig. e
flange stringers are attached to the inside of
the surface skin. To hold the skin surface to
airfoil shape and to provide a medium for
transferring surface air pressures to the
cellular beam structure, chordwise formers and
ribs are added. To transfer large concentrated
loads into the cellular beam structure, heavy
ribs, commonly referred to as bulkheads, are
used.
Figs. A19.1 and A18.2 illustrate typical

structural arrangements of wing cross-sections
for subsonic aircraft. The surface skin is
Fig. a
~--------\
~l _
Fig. b Fig.hc=J L
: 1,:n
Fig. c
cqA19) I~__ J Fig. iC[ :1: : OJ

Concentrated FlanGe Type of Wing Beam.
Dashed line represents secondary structure. In Fig. A19.2 Common Types of Wing Beam Flange
many cases this portion is fabric covered. Arrangement.
A19.1
A19.2 ANALYSIS OF WING STRUCTURES
To withstand the high surface pressures and to ment (Fig. A19.2), a portion of the covering on
obtain sufficient strength much thicker wing the compression side is made effective since it
skins are usually necessary. Modern milling is attached to closely spaced stringers or corru-
machines permit tapering of skin thicknesses. gations. However, the flange stringers between
To obtain more flan~e material integral flange cell webs are supported only at rib or bulkhead
units are machined on the thick skin as illus- points and thus suffer column action normal to
trated in Fig. k. the cell covering. This factor reduces the
flange allowable compressive stresses since it
Fig. j is not practical to space wing ribs less than
12 to 18 times the flange stringer depth. Thus,
if there were no other controlling factors, one
could easily make calculations to determine
which of the above would prove lightest. In
general, if the torsional forces on the wing
Fig. k are small, thus requiring only a thin covering,
the concentrated flange type of wing structure
should prove the lightest.
In general, the flange material should be

placed to give the largest moment of inertia in
"~
Fig. the Z direction, which means in general that the
flange material should be placed between the 15
and 50 per cent of wing chord from the leading
edge.
Fig.A19.3 Wing Sections - Supersonic Aircraft The secondary or distributing structure aft
of the structural box beam should be made as
In a cantilever wing, the wing bending light as possible and thus in general the farther
mOments decrease rapidly spanwise from the forward the rear closing web of the box beam, the
maximum values at the fusela~e support points.
Thus thick skin construction must be rapidly lighter the wing structure as a whole.
tapered to thin skin for weight efficiency, but
In the layout of the main spanwise flange
thinner skin decreases allowable compressive members bends or changes in direction should be
stresses. To promote better efficiency sand-
wich construction can be used in outer portion avoided as added weight is required in splicing
of wing (Fig. 1). A light weight sandwich core or in transverse stiffeners which are necessary
is glued to thin skin and thus the thin skin is to change the direction of the load in the flange
capable of resisting hi~h compressive stresses members. If flange members must be spliced, care
since the core prevents sheet from buckling. should be taken not to splice them in the region
of a maximum cross-section. Furthermore, in
A19.2 Some Factors Which Influence Wing Structural general, the smaller the number of fittings, the
Arrangements lighter the structure.
(1) Light Weight: (2) Wing-Fuselage Attachment:
The structural designer always strives for If the airplane is of the low wing or the
the minimum weight which is practical from.a high wing type, the entire wing structure can
production and cost standpoint. The hi~her the continue in the way of the airplane body. How-
ultimate allowable stresses, the lighter the ever, in the mid-wing type or semi-low wing
structures. The concentrated flange type of type, limitations may prevent extending the
wing structures as illustrated Fig. (a, b and c) entire wing through the fuselage, and some of
of Fig. A19.1 permits high allowable compressive the shear webs as well as the wing covering
flange stresses since the flange members are must be terminated at the side of the fuselage.
stabilized by both web and covering sheet, thus If a distributed flange type of cell structure
eliminating column action, which permits design were used, the axial load in the flange string-
stresses approaching the crippling stress of the ers would have to be transferred to the members
flange members. Since the flange members are extending through the fuselage. To provide for
few in number, the size or thickness required is this transfer of large loads requires structural
relatively large, thus giving a high crippling weight and thus a concentrated flange type of
stress. On the other hand, this type of design box structure might prove the best type of
does not develop the effectiveness of the metal structure.
covering on the compressive side, which must be
balanced against the saving in the weight of the (3) Cut-outs in Wing Surface:
flange members.
The ideal arrangement where continuity of
In the distributed type of flange arrange- structure is maintained over the entire surface
A19.3
of the structural box is seldom obtained in ac- the lower side of the Wing. They are usually
tual airplane design due to cut-outs in the fastened to two spanwise stringers with screws
wing surface for such items as retractable and the removable panels are effective in re-
landing gears, mail compartments and bomb and sisting bending and shear load. (See Fig.A19.5)
gun bays. If the distributed flange type of
box be&~ is used, they are interrupted at each
cut-out, which requires that means must be
provided for drifting the flange loads around Removable panel for
the opening, an arrangement which adds weight Fig. A19. 5 assembly and inspection
because conservative overlapping assumptions purposes.
are usually made in the stress analysis. The
additional structure and riveting to provide
for the transfer of flange load around large
openings adds considerably to the production
cost. Cutouts in the wing structural box destroys
the continuity of the torsional resistance of
For landing gears as well as many other the cell and thus special consideration must be
installations, the wing cut-outs are confined to given to carrying torsional forces around the
the lower surface, thus a structural arrangement cut-out. This special problem is discussed
as illustrated in Fig. A19.4 is qUite common. later.
The upper surface is of the distributed flange
type whereas the lower flange material is con- (4) Folding-Wings:
centrated at the two lower corners of the box. For certain airplanes, particularly Carrier
In the normal flying conditions, the lower sur-
face is in tension and thus cell sheet covering based Naval airplanes, it is necessary that pro-
between the cut-outs is equally effective in vision be made to fold the outer wing panels up-
bending if shear lag influence is discounted. ward. This dictates definite hinge points be-
tween the outer and center wing panels. If a
For negative accelerated flying conditions, the distributed flange type of structure is used,
lower surface is in compression thus sheet cov- the flange forces must be gathered and trans-
ering between corner flanges would be ineffec- ferred to the fitting points, thus a compromise
tive in bending. However, since the load fac- solution consisting of a small number of span-
tors in these flight conditions are approximate-
ly one half the normal flight load factors, this wise members is common practice.
ineffectiveness of the lower sheet in bending is (5) Wing Flutter Prevention:
usually not critical.
With the high speeds now obtained by modern
Cut-outs likewise destroy the continuity of airplanes, careful attention to wing flutter
intermediate interior shear webs of such prevention must be given in the structural lay-
sections as illustrated in Figs. (c and i), and out and design of the wing. In general, the
the shear load in these interrupted webs must critical flutter speed depends to a great ex-
be transferred around the opening by special tent on the torsional rigidity of the Wing.
bUlkheads on each side of the cut-out, which When the mass center of gravity moves aft of the
means extra weight. 25 per cent of chord paint, the critical flutter
speed decreases, thus it is important to keep
weight of the wing forward. At high speeds
where "compressibility" effects become important,
the torsional forces on the wing are increased,
Fig. A19.4 which necessitates extra skin thickness or a
larger cell. Designing for flutter prevention
is a highly specialized problem.
(6) Ease and Cost of Production:

In many cases, cut-outs in the leading edge
are necessary due to power plant installations, The airplane industry is a mass production
landing gear wells, etc. Furthermore, in many industry and therefore the structural layout of
airplanes, it is desirable to make the leading the wing must take into account production
edge portion removable for inspection of the methods. The general tendency at this time is
many installations which occupy this space in to design the wing and body structure, so that
the portion of the wing near the fuselage. If sub-assemblies of the various parts can be made,
such is the case, then an interior web should which are finally brought together to form the
be located near the front of the wing section. final assembly of the wing panel. To make this
process efficient requires careful consideration
Inspection doors for the central portion of in the details and layout of the wing structure.
the box beam structure are usually loce.tee-I on Photograph A19.6 illustrates the sUb-assembly
_ ________
_~'""'-"2S1T~R~U~C~T~UU!R~E~S'--~~~9~;,,;---------
...;A~N~A~L~!1!YS~I~S OF WING
:t;
Photograph A19 .6
Designin g T 0 Facilitate Pr o duction.
__ Photographs bY courtesy of North A mencan

. Av'la t ion Co.
A19.5
break-down of the structural parts of an air- is necessary. Another important distinction is

plane of a leading airplane company. Fabri- that a statically determinate wing structure
cation and assembly of these units permits the has just enough members to produce stability
installation of much equipment before assembly and if one member is removed or fails, the
of the units to the final assembly. entire structure will usually fail, whereas a
statically indeterminate structure has one or
A19.3 Wing Strength Requirements more additional members than are necessary for
static stability and thus some members could
Two major strength requirements must be fail without causing the entire structure to
satisfied in the structural design of a wing.
They are: - (1) Under the applied or limit collapse. In other words, the structure has
loads, no part of the structure must ~ a fail safe characteristic in that a redistri-
stressed beyond the yield stress of the mater- bution of internal stress can take place if
some members are over loaded. In general,
ial. In general, the yield stress is that statically indeterminate structures can be
stress which causes a permanent strain of
0.002 inches per inch. The terms applied or designed lighter and with smaller overall de-
limit refer to the same loads, which are the flections.
maximum loads that the airplane should en-
METHODS OF STRESS ANALYSIS FOR STATICALLY
counter during its lifetime of operations. INDETERMINATE WING STRUCTURES
(2) The structure shall also carry Design Loads
without rupture or collapse or in other words
failure. The magnitude of the Design Loads Two general methods are commonly used,
equals the Applied Loads times a factor of namely,
safety (F.S.). In general, the factor of (1) Flexural beam theory with simplifying
safety for aircraft is 1.5, thus the structure assumptions.
must withstand 1.5 times the applied loads
without failure. In missiles, since no human (2) Solving for redundant forces and stresses
passengers are involved, the factor of safety
by applying the principles of the elastic
is less and appears at this time to range theory by various methods such as virtual
between 1.15 to 1.25.
work, strain energy, etc.
Aircraft factors of safety are rather low The second method is no doubt more accu-
compared to other fields of structural design,
rate since less assumptions are necessary. A
chiefly because weight saving is so important wing structure composed of several cells and
in obtaining a useful transportation vehicle many spanwise stringers is a many degree re-
relative to useful load and performance.
Since safety is the paramount design require- dundant structure. Before the development of
ment, the correctness of the theoretical design high speed computing machinery, so-called
must be checked by extensive static and dynamic rigorous methods were not usable because the
computing requirements were impossible or
tests to verify whether the structure will
carry the design loads without failure. entirely impractical. However, present day
computer facilities have changed the situation
A19.4 Wing Stress Analysis Methods
and rigorous methods are now being more and
more used in aircraft structural analysis. Art.
In many of the previous chapters of this A7.9 and A8.10 in Chapters A7 and A8 present
book, internal forces were calculated for both matrix methods for finding deflections and
statically determinate and statically inde- stresses to be used with computer facilities.
terminate structures. The internal loads in a
A19.5 Example Problem 1. 3-Flange ~ Single Cell
statically determinate structure can be found Wing.
by the use of the static equilibrium equations
alone. The over-all structural arrangement of Fig. A19.7 shows a portion of a cantilever
members is necessary, but the size or shape of wing. To provide torsional strength a single
no individual member is required. In other closed cell (1) is formed by the interior web
words, design consists of finding internal AB and the metal skin cover forward of this
loads and then supplying a member to carry this web. Thin sheet is relatively weak in resist-
load safely and efficiently. In a statically ing compressive stresses thus 3 flange stringers
indeterminate structure, additional equations A, Band C are added to develop efficient bend-
beyond the static equilibrium equations are ing resistance. The structure to the rear of
necessary to find all the internal stresses. spar AB is referred to as secondary structure
The additional equations are supplied from a and consists of thin metal or fabric covering
consideration of structural distortions, which attached to chordwise wing ribs. The air load
means that the size and shape and kind of In this portion is therefore carried forward
material for members of the structure must be by the ribs to the single cell beam.
known before internal stresses can be deter-
mined. This fact means a trial and error method A wing is subjected to many flight condi-
A19.6 ANALYSIS OF WING STR UCT ORES
tions. The engineers who calculate the applied Az

I-- 22" ---J !--1"
loads on the wing usually refer the resulting 3/8"
shears and moments to a set of convenient x, y
and z axes. Fig. A19.7 shows the location of
these reference axes. The job of the stress Vz =1l500
engineer is to provide structure to resist 6" ~<:>(;)(;)
b,.~
these loads safely and efficiently. The general --~~;UL~-----.-", Vx= a ~-<- M =
procedure is to find the stresses or loads in - x
all parts of the cell cross-section at several 700# ~ 1, 100,000
stations along the spanwise direction and from
these loads or stresses proportion the required Fig. A19. 8
~ 10,,-.jMz =80000
areas, thicknesses and shapes.
SOLUTION. sta.240
In this example, the internal loads will
be calculated for only one section, namely, ASSUMPTIONS: - It will be assumed that the
that at Station 240. It will be assumed that 3 flange stringers A, Band C develop the entire
the design critical loads from the critical resistance to the bending moments about the Z
flight condition are as follows. and X axes. For skin under compression this
assumption is nearly correct since the skin will
tlx = 1,100,000 in. lb. Vz = 11,500 lb. buckle under relatively low stress. Since sheet
can take tensile stresses, this assumption is
tlz = 80,000 in.lb. Vx =700 lb. conservative. However, since the thin sheet
tly = 460,500 in. lb. cover must resist the shear stresses we will
make this conservative assumption. The main
Fig. A19.8 shows these shears and moments advantage of this approximate assumption is
referred to the reference axes with origin at that it makes the structure statically deter-
point (0). tloments are represented by vectors minate.
with double arrow head. The sense of the
moment follows the right hand rule. Fig. A19.8 shows the wing cut at Station
240. The unknown forces are the three axial
loads in the stringers A, Band C and the three
shear flows Qab' Qac and qbc on the three sheet
panels, making a total of 6 unknowns and since
there are 6 static equilibrium equations avail-
STA.O able for a space structure, the structure is
statically determinate.
Since the size and shape of flange members

A, Band C are unknown, we guess their centroid
locations as indicated by the dots in Fig.A19.8.
The axial load in each of the 3 stringers has
been replaced by its x, y and z components as
shown on the figure. The external applied loads
...; are given at the reference origin (0) as shown
~
CIl in Fig.A19.8. Y axis is normal to cross section.
0
~
'gl
"" 0
..:l
We now apply the equations of equilibrium
... to find the 6 unknowns •
....
0
'" To find Cy take moments about Z axis through

~ points A and B. Cy is normal to cross section.
:><
I ZMz(ab) =- 22 Cy + 80000 =0, whence
Cy =3636 lb.
The result comes out with a plus sign thus
STA.240
indicating that the assumed sense of tension
was correct.
To find By take moments about an X axis

~ -=~ (1) thru (A). By is normal to section.
ZMx(A) =- 11 By + 1,100,000 + 3636 x

0.125 = 0
Fig. A19.7
A19 7
+9065
whence, By = 100,043 lb. tension as whence qbc =~ = + 13.66 Ib./in. and
assumed. having the sense as assu~ed.
To find Ay take ZF y =0 In the above moment equation the moment of

the shear flow qbc about point (a') equals qbc
ZF y =- 100,043 - 3636 + Ay 0, = times twice the area of the cell or 665.
whence, Ay =
103679 lb. and com-
To find Qac take ZFx =0
pression as assumed.
ZFx = - 2679 + 2584 + 316 + 22 qac - 22 x
Since the direction of the 3 stringers is 13.66 - 700 0 =
known, we can find the X and Z components of
the stringer loads by simple geometry. whence Qac =35.45 Ib./in. with sense as
assumed.
The y, x, and z length components of the
three stringers from the dimensions given in To find Qab take ZF z =0
Fig. A19.7 are found to be,
~Fz =- 1296 - 1250 + 45 + 35.45 x 0.5 -
Member y x z
11.5 x 13.66 + 11500 - 11 Qab 0 =
A 240 6.2 3
whence, Qab = 805 lb. /in.
B 240 6.2 3 The loads on the stringers and sheet panels

have now been determined. The axial load in the
C 240 20.93 3 stringers is practically equal to their y com-
ponent since axial load equals their y force
component divided by the cosine of a small
The force components are therefore: - angle. The stress engineer would find similar
stresses at a number of stations along the span
Az =103679 x 3/240 = 1296 lb. These 6 stresses are generally referred to as
primary stresses. Usually in most structures
Ax =103679 x 6.2/240 = 2679 lb. there are secondary stress effects which must
Bz = 100043 x 3/240 = 1250 lb. be considered before final member sizes can be
determined. For example, internal webs of a
Bx = 100043 x 6.2/240 = 2584 lb. box type beam are designed usually as semi-
Cz = 3636 x 3/240 = 45 lb. tension field beams. Tension field beam theory
shows that the flange members are SUbjected to
Cx = 3636 x 20.93/240 = 316 lb. additional stresses besides the primary stresses
as found above. The SUbject of secondary
Fig. A19.9 shows these forces applied in stresses and the strength design of members
the plane of the cross-section at Station 240, and their connections to carry given stress
together with the unknown shear flows and the
external forces acting in the plane of the loads is taken up in detail in later chapters.
cross-section. A19.6 Example Problem 2. Metal Covered Wing With Single
External Brace Strut.
1296
2"li+r:=:==
1 l.
2679-1- A
22"
<lac
:~3:1.8".-.~1:·- .,1 ~I-. . . .
~o C
1" In Chapter A2, the stress analysis of an
externally braced fabric covered monoplane wing
was considered. To provide sufficient torsion-
316 6 t
I C1ab 45 / Vz=11500 al strength and rigidity, two external brace
2584-+LB_~_+___q_bc
struts were necessary. However, if a wing is
,._:-Vx=look metal covered, a single external brace strut
iJ t My =460,500 can be used, since the closed cell or cells
2 1250 ~10''-1 formed by the metal sheet covering and the in-
Fig. A19.9
ternal webs prOVide the torsional resistance
and the Wing can be designed as a s imply sup-
ported beam with cantilever overhang. An
To find qbc take moments about point (at) excellent example of this type of wing struc-
ture is the Cessna aircraft Model 180 as shown
ZMa , =+ 2679 x .5 - 2584 x 11.5 - 316 in the photograph. An excellent airplane
x 0.375 - 45 x 22 + 700 x 12 relative to performance, ease of manufacture
- 11500 x 39 + 460,500 - qbc and maintenance.
(665) 0 =
To introduce the student to the general
approach of stress analYZing such a wing struc-
ture, a limited discussion with a few calcula- Fig. A19.10 and A19.11 shows the wing
tions will be presented. dimensions and general structural layout of a
monoplane wing with one external brace strut.
The wing panel is attached to fuselage by
single pin fittings at points A and B with pin
axes parallel to x axis. The mating lugs of
the fittings at point A are made snug fit but
those at B with some gap, thus drag reaction
of wing loads on fuselage is resisted entirely
at fitting A. Since the fittings at A and B
cannot resist moments about x axis, it is
necessary to add an external brace strut DC to
make structure stable. The panel structure
consists of a main spar ACE and a rear spar BF.
The entire panel is covered with metal skin
forward of the rear spar.
Cessna Aircraft Model 180
Metal Covered Wing with A simplified air load has been assumed as
One External Strut shown in Fig. A19.12, namely, a lmiform load
=
w 30.27 lb./in. of span acting at the 30 per
cent of chord point. When this resultant load
Rear Web
B-- ----- is resolved into z and x components the results
IA. _ ---.!'ront Spar _ _
are Wz =
30 lb./in. and Wx =
4 lb./in. as shown
l in Fig. A19.12.
C
The general physical action of the wing
structure in carrying these air loads can be
considered as 3 rather distinct actions,
namely, (1) The front spar ACE resists the
bending moments and shears due to load wz .,
(2) The skin and webs of the two-cell tube
resists all moments about y axis or broadly
speaking torsional moments, (3) the entire panel
cross-section resists the bending moment and
Fig. A19.l0 flexural shear due to drag load wx , with the
top and bottom skin acting as webs and the two
spars as the flanges of this box beam.
General Calculations: -
z The unknown external reactions (see Fig.
A19.11) are Ay , Az, Ax, By, Bz and DC., or a
total of 6. Since 6 static equations of
equilibrium are available, the reactions are
statically determinate. Reaction DC is also
the load in brace strut DC.
To find reaction DC take moments about x
axis through points A, B
ZMx(AB) = (- 170 x 30 x 170/2) + DC(80/99.4)60

=0
Fig. A19.l2 whence, DC =
8979 lb. (The sign comes out
plus so the sense assumed in Fig. A19.11 was
correct.) The load in the strut is therefore
8979 lb. tension.
To find B , take moments about a z axis

through point 1A).
ZMZ(A) =- (4 x 170 x 170/2) + 27 By =0

Fig. A19.l!
whence, By =2141 lb. acting with sense
assumed.
A19.9
To find Ay take ZF y =0 points A and B. This couple force equals the

moment divided by distance between spars or
ZFy =2141 - 8979 (80/99.4) + Ay =0 15300/27 =
566.7 lb.
whence Ay = 5085 lb. Front spar (ACE) loads due to Wz = 301b/in:

To find Bz take moments about y axis w = 30 lb. lin.
through (A). t t t f t t t t t f tt
ZMy(A) =+ (30 x 170 x 3) - 27 Bz =0 Ay_ E
IAz
whence Bz =567 lb. acting as assumed.
60"
I
I L y
To find Az take ZF z =0
-L
Fig. A19.14
ZFz =170 x 30 - 567 - (8979 x 59/99.4)

+ Az =0
BO"-+- 90" ---j
whence Az =796 lb. acting as assumed. Fig. A19.14 shows free body of front spar
ACE. To find strut load DC take moments about
The fittings at A and B should be designed (A) •
to take the reactions at these points as found
above. The external strut DC and its end ZMA =(- 30 x 170 x 170/2) + 60 (DC x 80/99.4)
fittings must carry the tension load of 8979 lb. =0
The next step is to find the stresses and whence DC 8979 lb. = Tension which checks
loads on the structural parts of the wing panel. value previously found.
Since the spar ACE must resist the bending The y and z components of the strut re-
moments about x axis the airloads in Fig. action at C will then be,
A19.12 are moved to the spar centerline as
shown in Fig. A19.13. Cy = 8979 (80/99.4) = 7226 lb.
30#/in. Cz = 8979 (59/99.4) = 5329 lb.
+_M =3x30=90"#
<:' :::...-=L)4#/in~ These values are indicated on Fig. A19.14.
Fig. A19.13 To find Ay take ZF y =0

ZF y =- 7226 + Ay = 0, hence Ay = 7226 lb.
The torsional moment of 90 in. lb. per
inch of span is resisted by the cellular tube In finding the reaction Ay previously, the value
made up of two cells (1) and (2). In many was 5085 lb. The difference is due to the drag
designs the leading edge cell is neglected in bending moment Which tends to put a tension load
resisting the torsional moments due to many on front spar and compression on the rear spar.
cutouts, etc., thus cell one could be assumed wx=4#/in.
to provide the entire torsional shear resist-
ance and the shear flow for this case would be BY-B
tllltlltttttt 'FT
=
q Mv/2A Where My equals the torsional moment Rear Spar
27"
at a given panel section and A the enclosed l.<:!A'---_~~;z::.:'---_-.I L
A
area of cell (1). If both cells were con-
sidered effective then the sheet thickness is y-~x y--jE
necessary before solution for shear flow can Fig. A19.15
be computed. Refer to Chapter A6 for computing
torsional shear flow in mUltiple cell tubes. Fig. A19.15 shows the air drag load of 4
Ib./in. The bending moment on panel at a
The maximum torsional moment would be at distance y from the wing tip equals Wx (y)(y/2)
the fuselage end of the wing panel and its =
= 4y"/2 2Y". The axial load Py in either
magnitude would be 170 x 90 =
15300 in. lb. spar at any distance y from tip thus equals
Since the top and bottom skin is not attached bending moment divided by arm of 27" or Py =
to fuselage, this torsional moment myst be 2y"/27 =
.074y". The axial load at points A
thrown off on a rib at the end of the panel and B thus equal .074 x 170" = 2141 lb. (tension
and this rib in turn transfers this moment in in front spar and compression in rear spar).
terms of a couple reaction on the spars at Thus each spar is subjected to an axial load
increasing from zero at tip to 2141 lb. at would have to be replaced by two fittings, one
fuselage attachment points and varying as y2. on the upper flange and the other on the lower
At point C on from spar the axial tension load flange in order to be able to resist a Mx
would be .074 (90 2 ) 600 lb. = moment. Fig. A19.16 shows this modification.
The fitting at B could remain as before, a single
The design of the front spar between pin fitting.
points C and E would be nothing more than a
cantilever beam subjected to a bending forces
plus an axial tensile load plus a torsional
shear flow. The design of the spar between
points C and A is far more complicated since
we have appreciable secondary bending moments
to determine, which must be added to the pri-
mary bending moments. Fig. A19.15a shows a
free body of spar portion AC.
w = 30
t t t t t t t t t J,'i Mc=30X90X45 Fig. A19.16
5085 ---<:- -> -> - -> -> ..... ~
1 Fig. A19. 15a ~ The stress analysis of this wing would consist
of the spar AE resisting all the Mx moments and
The lateral load of 30 Ib./in. bends the the Vz shears and acting as a cantilever beam.
beam upward, thus the axial loads at A and C The torsional moment about a y axis coinciding
will have a moment arm due to beam deflection with spar AE would be resisted by shear stresses
Which moments are referred to as secondary in the cellular tubes formed by the skin and the
moments. To find deflections the beam moment spar webs. The drag bending and shear forces
of inertia must be known, thus the design of would be resisted by the beam whose flanges are
this beam portion would fall in the trial and the front and rear spars and the web being the
error procedure. Articles A5.23 to 28 of top and bottom skin.
Chapter A5 explain and illustrate solution of
problems involving beam-column action and such A19.8 Stress Analysis of Thin Skin - Multiple Stringer
a procedure would have to be used in actually Cantilever Wing. Introduction and Assumptions.
designing this b6am portion.
The most common type of wing construction
The rear spar BF receives two load systems, is the multiple stringer type as illustrated by
namely a varying axial load of zero at F to 2141 the six illustrative cross-sections in Fig.
lb. at B and the web of this spar receives a A19.2. A structure with many stringers and
shear load from the torsional moment. The sheet panels is statically indeterminate to
rear spar is not subjected to bending moments. many degrees With respect to internal stresses.
Fortunately, structural tests of complete wing
In Fig. A9.10 the secondary structure structures show that the simple beam theory
consisting of chordwise ribs and spanwise light gives stresses which check fairly well with
stringers riveted to skin are not shown. This measured stresses if the wing span is several
secondary structure is neceosary to hold wing times the wing chord, that sweep back is minor
contour shape and transfer air pressures to and wing is free of major cutouts and discon-
the box structure. This secondary structure tinuities. Thus it is common procedure to
is discussed in Chapter A21. The broad subject analyze and design a wing overall by the beam
of designing a member or structure to Withstand theory and then investigate those portions of
stresses safely and efficiently is considered the wing where the beam theory may be in error
in detail in later chapters. by using more rigorous analysis methods such as
those explained and illustrated in Art. A8.10
A19.7 Single Spar - Cantilever Wing - of Chapter A8.
Metal Covered
ASSUT1PTIONS - BEAM THEORY
A single spar cantilever wing with metal
covering is often used particularly in light In this chapter the wing bending and shear
commercial or private pilot aircraft. stresses will be calculated using the un-
symmetrical beam theory. The two main assump-
Suppose in the single spar externally tions in this theory are: -
braced wing of Fig. A19.11, that the external
brace strut DC was removed. Obviously the wing (1) Transverse sections of the beam originally
would be unstable as it would rotate about hinge plane before bending remain plane after bending
fittings at points A and B. To make the of beam. This assumption means that longitudi-
structure stable the single pin fitting at (A) nal strain varies directly as the distance from
AI9.11
the neutral axis or strain variation is linear. corner flange members (a) and (b) are stabilized
in two directions by the skin and webs and
(2) The longitudinal stress distribution is usually fail by local crippling.
directly proportional to strain and therefore
from assumption (1) is also linear. This Now continuing the loading of the wing
assumption actually means that each longitudinal after the skin has buckled, the stringers and
element acts as if it were separate from every corner members will continue to take additional
other element and that Hook's law holds, compressive stress. Since the ultimate strength
namely, that the stress-strain curve is linear. of the stringers is less than that of the cor-
ner members, the stringers (c) will start to
Assumption (1) neglects strain due to shear bend elastically or inelastically and will take
stresses in skin, which influence is commonly practically no further stress as additional
referred to as shear lag effect. Shear lag strain takes place. The corner members still
effects are usually not important except near have considerable additional strength and thus
major cut-outs or other major discontinuities additional external loading can be applied until
and also locations where large concentrated finally the ultimate strength of the corner
external forces are applied. members is reached and then complete failure of
the top portion of the beam section takes place.
Assumption (2) is usually not correct if Therefore, the true stress - strain relationship
elastic and inelastic buckling of skin and does not follow Hook's Law when such a structure
stringers occur before failure of wing. In is loaded to failure.
applying the beam theory to practical wings,
the error of this assumption is corrected by In the above discussion the stringers (c)
use of a so-called effective section which is were considered to hold their ultimate buckling
discussed later. load during considerable additional axial
strain. This can be verified experimentally
A19.9 Physical Action of Wing Section in Resisting External by testing practical columns.. Practical col-
Bending Forces from Zero to Failing Load. umns are not perfect relative to straightness,
Fig. A19.17 shows a common type of wing uniformity of material, etc. Fig. A19.18
cross-section structural arrangement generally shows the load versus lateral deflection of
referred to as the distributed flange type. column midpoint as a column is loaded to failure
and fails by elastic bending. Fig. A19.19 shows
similar results when the failure is inelastic
bending.
a b
' - - - -.......v ,.-----'
web c(stringers) web Fig. A19.18 Fig. A19.19
Fig. A19.17
The corner members (a) and (b) are considerably
larger in area than the stringers (c). The
skin is relatively thin. Now assume the wing {j = Central Deflection
{j = Central Deflection
is subjected to gradually increasing bending
forces which place the upper portion of this
wing section in compression and the bottom The test results show that a compression
portion in tension. Under small loading the member which fails in bending, normally con-
compressive stresses in the top surface will tinues to carry approximately the maximum load
be small and the stress will be directly pro- under considerable additional axial deformation.
portional to strain and the beam formula Thus in the beam section of Fig. A19.17 when
0c = MxZ/I x will apply and the moment of the stringers (c) reach their ultimate load,
failure of the beam does not follow since cor-
inertia Ix will include all of the cross-
section material. As the external load is inner members (a) and (b) still have remaining
creased the compressive stresses on the thin strength.
sheets starts to buckle the sheet panels and
A19.10 Ultimate Strength Design Requirement
further resistance decreases rapidly as further
strain continues, or in other words, stress is The strength design requirements are: -
not directly proportional to strain when sheet
buckles. Buckling of the skin panels however (1) Under the applied or limit loads, no
does not cause the beam to fail as the stringers structural member shall be stressed above
and corner members are lowly stressed compared the material yield point, or in other
to their failing stress. The stringers (c) words, there must be no permanent de-
are only supported transversely at wing rib formation or deflection of any part of the
points and thus the stringers act as columus and structure.
fail by elastic or inelastic bending. The
(2) Under the design loads which equal the

applied loads times a factor of safety,
no failure of the structure shall occur.
The usual factor of safety for conventional
aircraft is 1.5, or the structure must
carry loads 50 percent greater than the
applied loads without failure. The higher
the stress at failure of a member the less
material required and therefore the less
structural weight. The stress engineer
thus tries to design members which fail in
the inelastic zone. Fig. A19.21
stress equal to the stringer stress. In Fig.
The bending stress beam formula 0 C Mc/I = A19.21 2w is the effective sheet width to go
does not take care of this non-linear stress- with each stringer. The total stress on the
strain action and thus some modification of the effective widths carrying a uniform stress equal
moment of inertia of the beam cross-section is to the stringer stress equals the total load on
necessary if the ultimate strength of a wing the sheet panels carrying the actual varying
section is to be computed fairly accurately. stress distribution. The equation for effective
The stress engineer usually solves this problem width is usually written in the form,
by using a modified cross-section, usually ~
referred to as the effective cross-section. * 2w = kt (E!ast)2"
AI9.11 Effective Section at Failing Load. a widely used value for K = 1.9, whence
In order to use the beam formula which
assumes linear stress-strain relation, 2w = 1.90 t (E/0st)~ - - - - - - - - - (1)
corrections to take care of skin buckling and
stringer buckling must be formulated. The 0st =stress in stringer
effectiveness of the skin panels will be con-
sidered first. Therefore if we know the stress in the
stringer we can find the width of sheet to use
When a compressive load is applied to a with the stringer to obtain an effective section
sheet stringer combination, the thin sheet to take care of the sheet buckling influence.
buckles at rather low stress. For further
loading the compressive stress various over Effective Factor for Buckled Stringers.
the panel width as illustrated in Fig. A19.20.
The stress in the sheet at the stringer attach- Consider the beam section in Fig. A19.1?
ment line is the same as in the stringer since If we take a stringer (c) and attach a piece of
it cannot buckle and therefore suffers the same sheet to it equal to 2w, the effective width
strain as the stringer. Between the stringers and test it in compression and brace it in a
the sheet stress decreases as shown by the plane parallel to the sheet, the resulting test
dashed line in Fig. A19.20. This variable stress versus strain shortening curve (c) of
stress condition is difficult to handle so the Fig. A19.22 will result. The length of the test
stress engineer makes a convenient substitution specimen would equal the rib spacing in the
by replacing the actual sheet with its variable wing. The corner members (a) or (b) in Fig.
stress by a width of sheet carrying a uniform A19.1? being stabilized in two directions will
fail by local crippling, thus if a short piece
Stress Distribution on
Stiffener
if}i.'(7~,~'1J,':~I}'['
f Stress Distribution on
Sheet
VI)' r of this member is tested to failure in com-
pression the test curve (A) in Fig. A19.22 is
II, ~:\ obtained. Curve (C) shows that the stringer
I I I' \
/ I I'
I I,
I I '
I / I ' I
[I ' \ holds approximately its maximum load for a
! , II " /
I I I
\ '----.t.. 0st considerable strain period. Curve (A) shows
+ /
I
I I \
I I I
\
\~I-I
I
I I
\ / II
I I
II
r '-foc -.l that for the same unit strain member (a) can
take considerable higher stress. If we take
b
2 a unit strain of .006, the point at which the
maximum stress of 47000 is obtained in member
(a) (see point (3) on Curve A) then the stress
at the same strain for member (C) will be 38000
(see point (2) on Curve C).
Fig. A19.20
* The buckling of sheets is taken up in detail in

Chapter C5.
A19.13
sheet to use with each rivet line depends on

the stress in the stringer to which it is
attached. The failing stress for the stringer
will be used, which means that the failing
stress of the stringers and corner flanges must
be known before the effective width can be
found. For this example the zee stringers have
been selected of such a size as to give an
ultimate column failing stress of 38000 psi,
and the corner flange members have been made of
such size and shape as to give a failing stress
of 47000 psi. These failing stresses can be
computed by theory and methods as given in
Chapter C7. The sizes have purposely been
selected to given strengths represented by the
Fig. A19.22 test curves of Fig. A19.22.
The effective width with each rivet line
from equation (1) would be,
For stress analysis procedure using the
beam formula, we assume a linear stress vari- For zee stringer 2W =1.90 x .04 (10.5 x
ation from zero to 47000 psi. Since stringers
(c) can only take 38000 psi at the same strain, 10 /38000)2
8
J.
= 1.25 inches
or in other words, stringer (c) is less effect-
ive than members (a) and (b). The effective- Thus the area of effective skin 1.25 x =
ness factor for stringer (c) equals its ulti- .04 =.05 in. 2 • The area of the zee stringer
mate strength divided by the ultimate strength is 0.135 sq. in. which added to the effective
of member (a) or 38000/47000 = .808. skin area gives 0.185 sq. in. which value is
entered in Column (2) of Table I opposite zee
A19.12 Example Problem
stringers numbered 2, 3 and 4 in Fig. A19.23.
The same procedure is done for the corner mem-
The wing section in Fig. A19.23 is sub- bers 1 and 5 with the resulting effective areas
jected to a design bending moment about the x as given in Table I. On the bottom side which
axis of 500,000 in.lb., acting in a direction is in tension all material is effective. The
to put the upper portion in compression. The skin width equal to one-half the distance to
problem is to determine the margin of safety adjacent stringers is assumed to act with each
for this design bending moment. The material stringer. Taking the area of the angle section
is 2024 aluminum alloy. as 0.11 and adding skin area equal to 6 x .035
= .21 or a total of 0.32 sq. in. which value is
I-- 6" -r- 6" ~
.04
6" -t- 6" ---1 shown in Column 2 of Table I opposite stringers
7, 8 and 9. The areas of the lower corner
members plus bottom skin and web skin would
T
10"
come out as recorded in the Table.
The next step is to correct for stringer
effectiveness in compression. The failing
.04
1 stress for the zee stringer was given as 38000

and for the corner members 47000 psi. The
effectiveness factor for the zee stringer thus
equals 38000/47000 = .808. This factor is re-
corded in Column 3 of Table I. For the corner
Fig. A19.23
members 1 and 5 and all the tension members the
SOLUTION factor is of course unity. The balance of
Table I gives the calculation of the effective
Tne beam formula for bending stress at any moment of inertia ~AZ2 about the x neutral axis.
=
point is crb MxZ/I x • To solve this equation
we must have the effective moment of inertia The compressive stress intensity at the
of the beam cross-section. The bottom surface centroid of the zee stringers thus equals,
being in tension under the given design bending
moment is entirely effective, however the top crb ~ MxZ/I x =500,000 x 5.57/59.80 =46600 psi.
surface has a variable effectiveness since the
skin, stringers and corner flange members have The true stress equals the effectiveness
different ultimate failing stresses. factor times this stress =
.808 x 46600 =
37400 psi. The failing stress equals 38000
From equation (1) the effective width of hence margin of safety = (38000/37400) - 1 =
.01 or one percent.
TABLE A19.1 The stress analysis of this wing would

1 2 3 4 5 6 7 8
show the resulting bending and shear stresses
for a number of spanwise stations for the criti-
Stringer cal design load conditions. In this example
Stringer !Effect-
Stringer
Area
Effect- ive 2
solution the bending longitudinal stresses will
Plus z' Az' z=z'-z Az be determined on cross-sections at two stations,
Number iveness Area
Effective namely, stations 20 and 47.5, and the shear
Skin Factor (A)
stresses will be determined for station 20. In
1 0.37 1.0 0.370 4.50 1.664 5.47 11.05 this example problem, the leading edge cell will
2 0.185 0.808 O. 149 4.60 0.685 5. 57 4.62 be considered ineffective as well as any struc-
0.185 0.808 O. 149 4.60 0.685 5.57 4.62
ture to rear of rear beam, hence structure is a
3
one cell beam with multiple stringers. A second
4 0.185 0.808 O. 149 4.60 0.685 5.57 4.62
solution including the leading edge cell to form
5 0.370 1.0 0.370 4.50 1.664 5.47 11.05 a two cell beam will also be presented.
6 0.417 1.0 0.417 -4.60 -1.920 -3.63 5.50
7 0.320 1.0 0.320 -4.63 -1.480 -3.66 4.28 ANALYSIS FOR BENDING LONGITUDINAL STRESSES
8 0.320 1.0 0.320 -4.63 -1.480 -3.66 4.28
9 0.320 1.0 0.320 -4.63 -1.480 -3.66 4.28 Longitudinal stresses (tension or com-
pression) are produced by external forces normal
10 0.417 1.0 0.417 -4.63 -1.920 -3.63 5.50 to the cross-section and by bending moments
L; 2.981 -2.897 I x = 59.80 about x and z axes. The stress equations are:
<L
z' = distance from x axis to centroid of stringer area
On =PIA - (2)
Z = L; Az' /L;A = -2.897/2.981 = -.97 in.
where on = longitudinal stress
On corner members (1) and (5) the com- P = external load acting through cen-
pressive stress = 500,000 x 5.47/59.80 = troid of effective wing cross-section
45000 psi. With the failing stress being
47000 the margin of safety is 2 percent. A = effective area of cross-section.
Now suppose we would have omitted con- For any given flange member With area (a)
sideration of the stringer effectiveness factor the load Pn on the member would be,
and omitted column (3) of Table I. Carrying
through the calculations of Table I, the value Pn =ona - - - - - - (3)
of Z would be -0.76" and Ix would be 63.08.
The stress intensity on the zee stringers The stresses due to bending moments are
would then be 500000 x 5.36/63.08 42500 psi. = from Chapter A13, Art. A13.5: -
Since the failing stress for stringer is 38000
the margin of safety would be (38000/42500) - 1
= a negative 10.5 percent. The previous result
was a plUS 1 percent, thus a difference of 11.5 where 0b = bending stress with plus being ten-
percent in the results. Sion
The purpose of this simple example problem
was to emphasize to the stUdent that failure of
real aircraft stiffened skin structures occurs K2 = Iz/(Ix Iz
under non-linear stress-strain conditions and
the elastic theory must be modified if fairly
accurate estimates of the failing strength of
K" = I x/ (Ix Iz
a composite structure is to be obtained. The normal component of the axial load in
A19.13 Bending and Shear stress Analysis of Tapered - a flange member equals 0b a) where (a) is the
Multiple stringer Cantilever Wing. Unsymmetrical area of the flange member. Since the angle be-
Beam Method. tween the normal to the beam section and the
In general cantilever wings are tapered in centroidal axis of a stringer is generally qUite
both depth and planform. Fig. A19.24 illus- small, the difference between the cosine of a
trates a typical structural layout of the outer small angle and unity is negligible and thus
wing panel for a small airplane. The structure the normal component can be considered as the
consists of a front and rear beam (spar) with axial load in the stringer.
spanwise stringers between the two beams.
Tapering in cross-sectional material is obtained Before equations 2, 3 and 4 can be solved,
by decreasing size of members by cutting off the effective cross-section area must be known
portions of the spanwise stringers and corner as well as the moments of inertia and product
flanges and decreasing the skin and web thick- of inertia about x and z centroidal axes.
ness.
A19.15
WING BODY SECTION PLAN VIEW LOOKING INBOARD FROM TIP

:t 3 4
6 7
o4OSpolHT
+------ - - - - o
/4 /3 12 /I 10 q B
UPPER SURFACE STRINGER AND SKIN ARRANGEMENT

160 10 2.20
~~~I~~~~~~~~~~~~t~H~40% is normal to air-

of chord
plane ZX-Plane
line
_
_ _ _ _ .032. - - - - l - - - - .02.S ----+-t Skin thickness
between beams
LOWER SURFACE STRINGER AND SKIN ARRANGEMENT

110 , 0 ,.,0
.------.oai'---1---.020 - - - _.. 1Skin thickness

between beams
Fig. A19. 24 Structural Layout of Tapered Cantilever Wing

r--=2 1/4~
~Ot=1.2'~ -+- 1.2-.j
~2
1/4 :::=--I
4
TUpper Front Flange I
.050
1 area of angles = . 508

area of skin =. 090
Upper Stringer.
No. 2 & 6 (t =.064) T Upper Rear Flange
area stringer = . 257 1 area of angles = . 508

area of web =. 050
.648
area skin = . 096
.353 1 area of skin =. 090
area of web =.032
.630
No.3, 4 & 5 (t = .050)
area stringer = . 205
area skin = .096
.301
c"'\.O~
0. 0'1. 9:>e (X' reference axis
----+--C. - ----I ~
X Centroidal Axis
I
Lower Surface Stringer
area of bulb angle = . 11
area skin = .16
:27
I- ~I
Lower Rear Flange

2
I..
1
area of angles = . 508
area of skin =.120
area of web =.064
.692 Lower Front Flange
area of angles = . 508
r
2
i
area of skin =. 160
area of web =. 100
.768
Fig. A19. 25 WING SECTION AT STATION 20. EFFECTIVE AREAS.
f t
--rir=.----~_W_=uT-----,__w_=O~f.F---,----.,,,-tr-----'. 032 upper skin
~ W ~-_~ __
en ~ 7
~~ 9:>ec~
: ~~
x...
u
~
!",_,O_'__ fX' reference axis_
_+-- -=====_~ ========-=-======....::4r~. ==--=--~====::.........:=--=t==X'X
~
OW X Centroidal Axis
~ ~
~ ·~w
£ .~
14 C;;<
~
U
N 8
Z'
Fig. A19.26 WING SECTION STATION 47.5
A19.17
Fig. A19.25 shows the cross section at Substituting in equation (4)

Station 20 divided into 14 longitudinal units
numbered 1 to 14. Since the external load con- ab =- [.00098 x - 285000 - (-.0002125 x
dition to be used places the top surface in
compression, the skin will buckle and thus we 1300000)J x - [.004378 x 1,300,000 -
use the effective width procedure to obtain the (-.0002125 x - 285000)J z
skin portion to act with each stringer. Fig.
A19.25 shows the effective skin which is used whence, ab = 3.3 x - 5639 z - - - - (5)
with each flange member to give the total area
of numbers (1) to (7). The skin on the bottom Column 12 of Table A19.2 gives the results
surface being in tension is all effective and of this equation for values of x and z in
Fig. A19.25 shows the skin area used with each columns 10 and 11. Multiplying these bending
bottom flange member. stresses by the stringer areas, the stringer
loads are given in column 13. The sum of the
The next factor to decide is the stringer loads in thiS column should be zero since total
effectiveness as discussed and explained in the tension must equal total compression on a sec-
previous example problem. For the structure of tion in bending.
Fig. A19.25 we will assume that the compressive
failing stress of the stringers is the same as Stresses at station 47.5:
that for the corner members, thus we will have
no correction factor to take care of the Ix =157.4, Iz = 700, I xz = -35.4 (Table
situation of flange members having different A19.3)
Ultimate strengths.
K4 = 35.4/(157.4 x 700 - 35.4 2 ) =
Table A19.2, columns 1 to 11, and the -35.4/108950 =-.000324
calculations below the table give the calcula-
tions for determining the section properties K2 = 700/108950 = .00643
at Station 20, namely A, Ix, I z and I xz .
Table A19.3 gives the same for wing section at K3 = 157.4/108750 = .001447
Station 47.5. The areas in column (2) are less
since sizes have changed between Stations 20 ab =- [.001447 x -215000 - (-.000324 x
and 47.5.
1,000,000)J x - [(.00643 x 1,000,000
Calculation of longitudinal stress due to (-.00324 x -215000)J z
Mx and My bending moments: -
whence
The design bending moments will be assumed
and are as follows: - ab =-14.5 x -6360 z
Station 20 Mx = 1,300,000 in. lb. Column 12 of Table A19.3 gives the results
of this equation and column 13, the total
Mz =- 285,000 in.lb.
stringer loads at station 47.5.
Station 47.5 Mx = 1,000,000 in. lb.
The stresses in column 12 of each table
Mz = - 215,000 in. lb. would be compared to the failing stress of the
flange members to obtain the margin of safety.
The moments about the y axes are not
needed in the bending stress analysis but are ANALYSIS FOR SHEAR STRESSES IN WEBS AND SKIN
needed in the shear analysis which will be made
later. The shear flow distribution will be cal-
culated by using the change in axial load in
To solve equation (4) the constants K4 , the stringers between stations 20 and 47.5, a
K2 and K" must be known. method commonly referred to as the 6P method.
For explanation of this method, refer to Art.
For Station 20 from Table A19.2, Ix = A15.16 of Chapter A15.
230.3, I z =
1030 and I xz =-
50, whence -
The shear flow in the y direction at a
K4 = Ixz/O x I z - I xz 2 ) = -50/(230.3 x 10.30 point n of the cell wall equals,
- 50 2 = -50/235500 = -.0002125
)
n 6P
= qo - Z- - - - - - - - - - - - (6)
K2 = I z/235500 = 1030/235500 = .004378 o d
K3 = I x/235500 = 230.3/235500 = .00098 where qo is a known value of shear flow at some

point 0 and the second term is the change in
shear flow between points 0 and n.
TABLE A19. 2
SECTION PROPERTIES ABOUT CENTROIDAL X AND Z AXES
Wing Section at Stntion 20 (Compression on upper surface)
1 " 3 4 5 6 7 8 9 10 11 12 13
Flange Area
No. A Z' AZ' AZ,2 X' AX' AX';': AX'~I Z = Z' - Z X
= x' - x CYb P
= ""b A
1 . 648 5.50 3.56 19.61 -0.10 -0.06 0 -0.36 6.68 -17.4 -37660 -24390
2 ,,353 5.90 2.08 12.30 5.65 I. 99 11.30 11.78 7.08 -1 I. 65 -39940 -14070
3 .300 5.85 I. 76 10.28 11.20 3.36 37.60 19.67 7.03 - 6.10 -39620 -11890
4 .300 5.55 I. 66 9.25 16.85 5.05 85.00 28.10 6.73 - 0.45 -37950 -11370
5 .300 5.05 I. 52 7.65 22.40 6.72 150.50 34.00 6.23 S.IO -35115 -10510
6 .353 4.40 I. 55 6.83 28.00 9.89 277.00 43.50 5.58 10.70 -31450 -11090
7 .630 3.55 2.24 7.95 35.72 23.15 826.00 80.00 4.73 18.42 -26560 -16700
8 .692 -8.40 -5.81 48.90 35.72 24.70 881. 00 -208.00 -7.22 18.42 40750 28230
9 .270 -8.50 -2.30 19.55 27.80 7.50 208.00 - 64.00 -7.32 10.50 41280 11180
10 .270 -8.50 -2.30 19.55 22.60 6.10 138.00 - 52.00 -7.32 5.30 41250 11170
l1 .160 -8.30 -I. 33 II. 05 16.70 2.67 44.60 - 22.20 -7.12 - 0.60 40100 6430
12 .27 -8.00 -2.16 17.30 10.90 2.94 32.00 - 23.50 -6.82 - 6.40 38400 10390
13 .27 -7.50 -2.02 15.20 5.60 I. 51 8.45 - II. 35 -6.32 -11.7 35570 9600
14 .768 -6.50 -5.00 32.50 - 0.10 .08 0 0.50 -5.32 -17.40 29950 23020
f ~.584 -6.56 238.0 96.44 2700 -163.8 0.00
Z = -6.56 = -I. 176 General Notes:

5.584
See Fig. A19.25 for Section at Station 20.
X = 96.44 _ 17.3 Reference axes XIX' and Z'Z' are assuITled as shown.
5.584 -
Properties are calculated with respect to these
axes and transferred to the centroidal X and
Ix = 238 - 5.584 x I. 176:l = :l30.3 in. 4
Z axes.
Ie: = :noo - 5.584 x 17.3 2 ~ 1030 in. 4
I xz = -163.8 - 5.584 x-I.176 x 17.3 = -50 CYb = 3.3 X - 5639

"
TABLE A19. 3
SECTION PROPERTIES ABOUT CENTROIDAL X AND Z AXES
Wing Section at Stn tion 47.5 (CoJ:1pression on upper Surface)
1 ~ 3 4 5 6 7 6 9 10 11 12 13
Flan~;'c Area
No. A Z' AZ' AZ! :~ X' AX' AX· 2 AX'Z' L.
= Z' -Z X = X' - X CYb P = cb A
1 .476 5.CO 2.66 14.95 - 0.10 - 0.05 0 - 0.27 6.21 -15.60 -39275 -18700
" .318 5.83 I. 86 10.80 4.40 1.40 6.16 8.16 6.44 -1 I. 10 -40790 -12990
3 • ~~66 5. "/5 I. 53 8.80 10.00 2.6G 26.60 15.30 6.36 - 5.50 -4025 -10710
4 .266 5.4::::' I. 44 7.81 15.60 4.15 64.75 22.40 6.03 0.10 -38400 -10200
5 .266 4.90 I. 30 6.40 21. 20 5.64 120.00 27.60 5.51 5.70 -35075 - 9320
6 .318 4.30 I. 37 5.88 26.70 8.50 2:'::6. S,Q 36.60 4.91 II. 20 -31360 -10000
7 .476 3.60 1.71 6.17 31.70 15.10 478.00 54.40 4.21 16.20 -27030 -12880
8 .353 -7.10 -3.93 27.90 31.70 17.51 555.00 -124.50 -6.49 16.20 41050 22770
9 • ;J,3~ -7.40 -1.74 12.90 26.50 6.23 165.00 - 46.10 -6.79 11.00 43090 10140
10 .~3~ -7.40 1.74 12.90 21. 30 5.00 107.00 - 37.18 -6.79 5.80 43180 10160
11 • 1;.~5 -7.20 -0.90 6.48 15.10 I. 89 28.50 - 13.60 -6.59 - 0.40 41900 5240
1" · ;,3" -6. 90 -1, 62 11, 20 9.60 ~.26 21. 70 - 15.60 -6.29 -55.90 40085 9410
13 .235 -6. "'5 -I. 47 9.20 4.50 I. 06 4.75 - 6.61 -5.84 -II. 00 36060 8470
14 .60r -5.40 -3.27 17.70 - 0.10 - 0.06 0 0.33 -4.79 -15.60 30725 18600
i 4.61 -2.80 159.1 71. 3 1804 - 79.0 0
Z = -:;.80/4.61 ~ -0.61"
General Notes:
X = 71.3/4.61 a 15.50" See Fig, A19.26 for section at Station 47.5.
Reference axes X 1 X 1 and Z·Z' are assuTI1ed as shown.
1)< = 159.1 - 4.61 x.61~
= 157.4
2 Cfb = -14.5 X - 6360 Z
I z = 1304 - 4.61 x 15.5 = 700
I~z = -79 - 4.61 x -.61 x 15.5 - 35.4

A19.19
~p equals the change in stringer axial load areas are referred as m values (See Fig. A19.27).
over a distance d in the y direction. Column 6 of Table A19.4 records these double
areas which were obtained by use of a planimeter.
Since the cell in our problem is closed the Column (7) gives the moment of each shear flow
value qo at any point is unknown. We assume it about the c.g. and the total of this column
zero on web 1-14 by imagining that the web is gives the moment about the c.g. of the complete
cut as shown in Fig. A19.27. Equation (6) thus shear flow system of Fig. A19.27 or a value of
reduces to, 256060 in. lb.
The double areas (m) can be found approxi-

- - - - - - - (7) mately as follows:
Columns 1 to 5 of Table A19.4 show the solution The moment of

of equation (7). The shear flow values of qy the shear flow q on
in column 5 are plotted on the cell wall in the web (2-3) about
Fig. A19.27, remembering that qy = qx = qz. For point 0 equals q
rules giving direction of qx and qz refer to times twice the area
Chapter A14, Art. A14.6. (A 1 + A2 ) . In most
cases, the area A2
2BB.B 1 374.6
j2:!45~.~9_~~~,-"'-"-A.L!L-~':"'="---r---1414.2 can be neglected.
206.B
4 553. By simple
geometry, the area
z A1 = 1/2 (X 3 Z2 - X 2Z 3 ) . The moment of the
shear flow q on web (2-3) thus equals q (X 3 Z2
- X2 Z3)' Since values of X and Z for all
flange points with reference to section (c.g.)
are given in the Table AI9.2, it is unnecessary
to use the planimeter except for regions of
sharp curvature.
2BO.0 316.7 354.5
Fig. A19.27 MOMENT OF EXTERNAL LOADS
ABOUT c.g. OF STATION 20
TABLE A19.4
As stated before the engineers in the
Calculation of Flexural Shear Flow Assuming q - zero in applied loads calculation group supply the shears
Web Between Flange Members (1) and (14). and moments at various spanwise stations. We
(Shear values are average between slations 20 and 47. 5) will assume that these loads are: Vz 12000 =
1 2 3 4 5 6 7 8 9
=
lb., Vx +·2700 lb., My =-
390,000 in.lb. The
location of the reference y axis used by the
Flange
P P
LIP
q: m q :
r loads group will be assumed as located at point
No.
at Station at Station
20 47.5
-27:5 LIP
Z 27.5
sq.
in.
mq q1
q + q1
o in Fig. A19.28 relative to cross-section at
Station 20.
1 -24390# -18700 206.8
-14070 -12990 39.1 206.8 48.4 10000 -328 -121. 2
2
-11890 -10710 42.9 245.9 43.8 10780 -328 - 82.1
3
-11370 -10200 42.5 288.8 42.0 12120 -328 - 39.2
4
-10510 43.3 331. 3 42.6 14100 -328 3.3
5 - 9320
-11090 -10000 374.6 43.6 16340 -328 46.6
6 39.6
-16700 -12880 139.0 414.2 42.0 17500 -328 86.2
7
28230 22770 -198.7 553.2 197.0 108900 -328 225.2
8
11180 10140 - 37.8 354.5 47.4 16800 -328 26.5
9
11170 10160 - 36.7 316.7 37.2 11800 -328 - 11.3
10
6430 5240 - 43.2 280.0 40.8 11420 -328 - 48.0
11
10390 9410 - 35.6 236.8 42.2 10000 -328 - 91. 2
12
9600 8470 - 41. 1 201. 2 38.0 7650 -328 -126.8
13
160.1 54.0 8650 -328 -167.9
14 23020 18600 -160.3
0 204.0 0 -328 -328
1
Therefore moment of external loads about
Total mq : 256060
c.g. is,
For equilibrium of all the forces in the Z'Mc.g. =12000 x 33.3 - 2700 x 11.8 - 390000
plane of the cross-section ZMy must equal zero. = 41800 in. lb.
For convenience we will select a moment y axis
through the c.g. of the cross-section. The Moments Produced by Inclination of Flange Loads With
moment of the shear flow q on any sheet element Beam Section.
equal q times double the area of the triangle
formed by joining the c.g. with lines going to Since the flange members in general are not
each end of the sheet element. These double normal to the beam sections, the flange loads
A19 20 ANALYSIS OF WING STRUCTURES
have components in the Z and X directions. Due to external loads = 41800 in.lb.
Columns (4) and (7) of the Table A19.5 give the
values of these in plane components. The slopes Then the total unbalanced moment = 4936 +
dx/dy between stations 20 and 47.5 are found by 256060 + 41800 =
302796 in.lb.
scaling from Fig. A19.24. Fig. A19.29 shows
these induced in plane forces as found in Table For equilibrium, this must be balanced by
A19.5. a constant shear flow q~.
TABLE A19. 5 hence

In Plane Moments About Section c. g. Produced by in Plane 302796 328 Ib./in.
Components of Flange Loads 2 x 461.5
1 2 3
Station 20
4 5 6 7 8
(Note: 461.5 = total area of cell)
Px = M c . g. Pz = Me. g. = The shear stresses q~ are listed in Column

Flange
P
dx dz (8) of Table A19.4.
No. dY p~ =Px z dy dz
P<iY
-(P z X)
dy
The final or resultant shear flow qr at any
1120 7480 .025 -610 -10600
1
2
-24390
-14070
-.046
0 0 0 .022 -310 - 3610
point therefore equals
3 -11890 0 0 0 .022 -262 - 1600
4
5
-11370
-10510
0
0
0
0
0
0
.022
.022
-250
-231
- 112
1180
qr =q + q~
6 -11090 0 0 0 .025 -278 2970

7 -16700 .021 - 350 -1660 .025 -417 7770 The resulting values are given in Column 9
.021 594 -4300 .018 -508 9350
8
9
28230
11180 0 0 0 .017 -190 2000
of Table A19.4. Fig. A19.30 illustrates the
10 11170 0 0 0 .015 -167 885 results graphically.
11 6430 0 0 0 .015 - 96 - 57
12 10390 0 0 0 .017 -177 - 1130
13 9600 0 0 0 .018 -173 - 2020
14 23020 -.046 -1060 5640 .018 -416 - 7250
Z 7160 - 2224
NOTES:
Column (2); P from Table A19. 2
Column (5) and (8): Values of Z and X are found in
Columns 10, 11 of Table A19. 2
610 417 Fig. A19.30 Final shear flow diagram. For val-
r
~ ues see Column 9 of Table A19.4.
OI ~3 J4 ~5
310
'
262 250
+c.g.
231 2~:
'['50 Havins determined the shear flOWS, the
shear stress on any panel would be q/t. In
checking the sheet for strength in shear and
combined shear and tension, interaction rela-
1060 14 ,13 1.2 l,1 19 ~ 8 594 tionships are necessary. The strength design
~ 1~0 508t
of sheet panels under combined stresses is
416 173
t
1~7 96
t IJ7 covered in considerable detail in Chapter C5.
A19. 14 Bending and Shear Stress Analysis of 2-Cell
Fig. A19.29 In plane forces produced by flange Multiple Stringer Tapered Cantilever
axial loads. Wing.
The moments of these in plane components about A two-cell beam is also quite common in
the section e.g. are given in Columns (5) and wing structural design. A two-cell structure
(8) of Table A19.5. In general, these moments in bending and torsion is statically indeter-
are not large. minate to the second degree since the shear
flow at anyone point in each cell is unknown.
Total Moments of All Forces About Section e.g. However, due to continuity between cells the
at Station 20: angular twist of each cell must be the same,
which gives the additional equation necessary
Due to flanges =7160 - 2224 =
4936 in. lb. for solving a two-cell beam as compared to the
(Ref. Table A19.5) single cell analysis.
Due to assumed static shear flow 256060 = Example Problem

in. lb. (Ref. Table A19.4)
To avoid repetition of similar type calcu-
A19.21
lations as was used in the previous single cell The first 7 columns of this table are the same
problem, the bending and shear stresses will be as in Table A19.4, since no stringers have been
determined for the same structure as in the added to cell (1), and the shear q is assumed
previous example except that the leading edge zero in cell (1).
cell is considered effective, thus making a
2-cell structure. Since there are no spanwise To make the twist of each cell the same and
stringers in the leading edge, very little skin also to make the summation of all torsional
on the compressive side will be effective. On forces zero will reqUire two unknown constant
the tension side, the leading edge skin would shear flows, q~ in cell (1) and q2 in cell (2).
be effective in resisting bending axial loads Thus two equations will be written, namely:
and thus the moment of inertia would be slightly
different from that found in example problem 1. (8)
Since this problem is only for the purpose of
illustrating the use of the equations, the
leading edge skin will be neglected in computing
m c •g . 0 (9) =
the bending flexural stresses. With this The twist 9 per unit length of a cell
assumption, the bending stresses and flange equals
loads at stations 20 and 47.5 are the same as
for the previous problem. (See values in column 1
12 and 13 of Tables A19.2 and A19.3.) 9 = 2AG t;qL/t (10)
,Shear Flow Calculations: - The modulus of rigidity G will be assumed con-

stant and thus will be omitted.
To compute the static shear flow, each cell
is assumed cut at one point as shown in Fig. Consider cell (1): (Refer to Table A19.6,
A19.31, and thus the shear flow is zero at Columns 10 and 11)
points (a) and (b).
Area of cell (1) = 83.5 sq. in.
t
Fig. A19.31
2
.
3 4
.
5
. g~ = _1
2A~
G q Lit + Z q~ Lit + t; q2 L/~
I
6
~ q~
a C,II (l) b Cell (2) 1
Area = 83.5 Area = 461. 5 sq. in. g~ =2 X 83.5 + 1278 - 230 qJ, whenc e
14 1.3 12 1.1 10 .
9
g~ = 7.65 q~ - 1.378 q2- - - - - - - - (a)
For cell (2):

Table A19.6 (Column 5) gives the value of
the static shear flow under these assumptions. Area of cell (2) = 461.5 sq. in.
TABLE A19. 6
SHEAR FLOW - 2 CELL - MULTIPLE FLANGE TAPERED BEAM
(A veragc Shear Flow Between Stations 20 and 47.5)
1 ~ 3 4 5 6 7 8 9 10 11 12 13 14
Flange
No.
P
at Sta.
P
at Sta. _6P
q .:
i6.P m mq L t
L/t
Cell Cell
a (L/t)
Cell Cell ql q2
qr =
~O 47.5 27.5 27.5 (in. ) (1) (2) (1) (2) q+qrtq 2
1 - 24390 -18700 206.8 206.8 48.4 144 29800 0 -317 -110.2
10000 5.75 .04
2 -14070 -12990 39.1
3 -11890 -10710 42.9
245.9 43.8 10780 5.50 .04 137 33700 0 -317 - 71.1
4 -11370 -10200 42.5
288.8 42.8
331.3 42.6
12120 5.50 .04 137
137
39600
45400
0
0
-317
-317
- 28.2
14.3
14100 5.50 .04
5 -10510 - 9320 43.3 374.6 43.6 16340 5.50 .04 137 51400 0 -317 57.6
6 -11090 -10000 39.6
414.2 42.0 137 56700 0 -317 97.2
17500 5.50 .04
7 -16700 -12880 139.0 553.2 197.0 382 211500 0 -317 236.2
22770 -198.7 108900 12.25 .032
8 28~30
354.5 47.4 .032 172 61000 0 -317 37.5
9 11180 10140 16800 5.5
- 37.8 316.7 37.2 11800 5.0 .032 156 49400 0 -317 0
10 11170 10160
11 6430 5240
- 36.7
_ 43. 2 280.0 40.8 11420 5.8 .032 181 50650 0 -317 - 37
12 10390 9410 - 35. 6
236.8 42.2
201.2 38.0
10000 5.8 .032
.032
181
166
42900
33400
0
0
-317
-317
- 80.2
-115.8
13 9600 8470 .41.1 7650 5.3
160.1 54.0 8650 5.5 .032 172 27600 0 -317 -156.9
14 ~30~0 18600 160.3
0 204 0 11.75 .051 230 230 0 0 -62.8 -317 -254.2
1
Lead
edge 0 371 33.5 .032 1048 0 -62.8 -62.8
web
Sum 256060 1278 2469 0 733050
NOTES: L length of web sheet between flange members. Col. 4 (Col. 2 + Col. 3)/27.5
t :z web thickness.
q~
A19.15 Bending Strength of Thick Skin - Wing Section
Q" =2 x 1
461.5 [733050 - 230 + 2469 qJ
Figs. i and k of Fig. A19.3 illustrate
approximate shapes for airfoils of supersonic
Q" = 794 - .2495 q~ + 2.678 q" - (b) aircraft. Such airfoils have relatively low
thickness ratios and since supersonic military
For continuity Q~ must equal Q", hence
aircraft have comparatively high wing loadings,
equating (a) and (b):
it is necessary to go to thick skin in order to
resist the wing bending moments efficiently.
7.899 q~ - 4.056 q" - 794 = 0 - (c)
The ultimate compressive stress of such struc-
tures can be made rather uniform and occurring
For equilibrium the summation of all mom- at stresses considerably above the yield point
ents in the plane of the cross-section about the
of the material. Since structures must carry
section (e.g.) must be equal to zero, or
the design ,loads without failure, it is neces-
sary to be able to calculate the ultimate bend-
Z Mc • g • =0 ing resistance of such a wing section if the
margin of safety is to be given for various
The moment of the external loads about the
load conditions.
section c.g. is the same as in previous problem.
The question of the ultimate bending
M external forces = 41800 in. lb.
resistance of beam sections that fail at stresses
beyond the elastic stress range is treated in
The induced moment due to the in plane
Article A13.10 and example problem 7 of Chapter
components of the flange axial loads is likewise
A13 and should be studied again before pro-
the same as in previous problem (see Table ceeding with the following example problem.
A19.5) •
A19.16 Example Problem
Mdue to flange 1oa ds = 4936 in.lb.
To illustrate the procedure of Art. A13.10,
The torsional moment due to the static a portion of a thick skin wing section as
shear flow from Column (7) of Table A19.6 equals illustrated in Fig. A19.34 will be considered.
256060 in. lb. The torque due to the unknown
constant shear flows of q~ and q" is equal to
twice the enclosed area of each cell times the
shear flow in that cell, whence
M( due to q~ and q" )=2X83.5q~ +2x461.6 q"

=
167 q 1 + 923 q a
Therefore ZM c • g • = 167 q~ + 923 q" +

302796 = 0 - - (d)
Solving equations (c) and (d), we obtain,

Fig. A19.34
q~ = -62.8 lb./in., q" = -317 lb./in.
For simplicity the section has been drawn
These values are listed in columns (12) and (13) symmetrical about the x-x axis. The material
of Table A19.6. The final or resultant shear is aluminum alloy. In this problem the
flow qr on any sheet panel equals the sum of material stress-strain curves will be assumed
q ~ q, + q". The results are shown in column the same in both tension and compression. The
(14) of Table A19.6. Fig. A19.32 shows the problem is to determine the margin of safety
potted shear flow pattern. Comparing this for this beam section when subjected to a
figure with Fig. A19.30 shows the effect of design bending moment Mx = 1,850,000 in. lb.
adding the leading edge cell to the single cell
of the previous problem.
SOLUTION:
Since it is desirable to use the beam

formula ab = Mxz/I x , it is necessary to obtain
a modified beam section to correct for the non-
linear stress-strain relationship since the
give structure will fail under stresses in the
inelastic zone. The maximum compressive stress
at surface of beam will be assumed at 50000 psi.
This value could be calculated from a consider-
A19.23
ation of crippling and column strength of the The values in column (3) of Table A19.7
stiffened skin, a subject treated later. represents the true compressive stress at the
midpoint of a strip area when the beam is
Curve (A) of Fig. A19.35a is a portion of resisting its maximum or failing bending moment.
the compressive stress-strain diagram of the The values in column (4) represent the com-
aluminum alloy material from zero to 50,000 psi. pressive stress at the midpoint of the strip
areas if the bending stress is linear and vary-
Due to symmetry about the x-x axis, we ing from zero at neutral axis to 50000 psi at
need only to consider one half of the beam edge of beam section (Curve B of Fig. A19.35a).
section. We divide the upper half of the beam
section into horizontal strips, each 3/8 inch To illustrate, consider strip area number
thick. Each beam portion along these horizontal (2) in Fig. b of A19.35. Project a horizontal
strips can be placed together to form the areas' dashed line from midpoint of this strip until it
labeled (1) to (8) in Fig. b of Fig. A19.35. intersects curves A and B at points (a) and (b)
Since plane sections remain plane after bending respectively. From these intersection points
in both elastic and inelastic stress zones, project downward to read values of 48000 and
Fig. c shows the beam section strain picture. 40600 psi respectively.
Fig. A19.35 In using the linear beam formula, the
Fig. c stress intensity on strip (2) would be 40600
Fig. b
' ' 'f

Fig. a but actually it is 48000. The ratio between
.010 the two is given the symbol K. Thus to modify
I::
.s the linear stress to make it equal to the
~ .008 nonlinear stress we increase the true strip
1-0
areas by the factor K, giving the results of
~ .006 column (6).
o 3"
....
§ .004 The modified moment of inertia (column 8)
II
w.002
equals Ix =
115.52.
The design bending moment was 1,850,000
10 20 30 40 50 in. lb.
a c x 1000 (psi) Consider point at midpoint of strip (1)
Table A19.7 shows the calculations for

Z = 2.8125 inches
obtaining the modified moment of inertia of the
cross-section to use with the linear beam
0b = Mx Z/I x = (1,850,000 x 2.8125)/115.52
formula. = 45100 psi.
TABLE A19. 7
This stress is based on the modified strip
1 2 3 4 5 6 7 8 areas. The true stress Ob(t) on strip 1 thus
Area Arm
equals KOb = 1.049 x 45100 = 47400. The allow-
K=
Portion A l l
Z Ae Z
2 able stress at failure equals 49200 from column
Oc Oc Ocloc A e = KA
(3) of Table A19.7. Hence margin of safety =
1 4.500 49200 46900 1. 049 4.71 2.8125 37.30 (49200/47400) - 1 =
.04 or 4 percent.
2 2.437 48000 40600 1.182 2.88 2.4375 17.14
3 0.211 46000 34400 1. 340 0.282 2.0625 1. 20
The margin of safety for other points on the
beam section will likewise be 4 percent. For ex-
4 0.211 43200 28100 1. 539 0.324 1. 6875 0.92
ample at midpoint of strip (3), Z = 2.0625. K=1.34
5 0.211 39200 21840 1. 790 0.378 1. 3125 0.65
0.211 33200 15620 2.121 0.449 0.9375 0.38
whence OCt =[(1,850,000 x 2.0625)/115.52Jl.34
6
7 0.211 20500 9360 2.190 0.461 0.5625 0.15 = 44400.
8 0.211 7000 3120 2.240 0.473 0.1875 0.02
~
whence margin of safety = (46000/44400)
57.76
2
- 1 = .04
Ix = 115.52
Col. (2) = Area of Strip The moment of inertia Without modifying
Col. (3) Stress at midpoint of strip as read from Curve (A) the strip areas would come out to be Ix 104.42 =
Col. (4) Stress at midpoint of strip as read from Curve (B) hence the stress at midpoint of strip (1) would
Col. (5) Nonlinear Correction Factor K = a cia c I calculate to be 0b =
(1,850,000 x 2.8125)/104.42
Col. (6) Modified Area = A e = KA
Col. (7) Arm from Neutral Axis to Midpoint of Strip
= 49900. The allowable strvss for linear stress
Col. (8) Modified section moment of inertia variation would be 46900 from column (4) of
Table. Hence margin of safety would be
(46900/49900) - 1 = -.06. The elastic theory A19.18 Shear Lag Influences

thus gives a margin of safety 10 percent less In the beam theory, the assumption is made
than the strength given when true stress-strain
or non-linear relationship is used. that plane sections remain plane after bending.
In a beam involving sheet and stringer panels,
this assumption means that the sheet panels
If the Slli~e comparison was made for bending have infinite shearing rigidity, which of course
about Z axis of this same beam section the is not true as shearing stresses produce shear-
difference would be considerably more than 10
percent as more beam area is acting in the ing strains. The effect of sheet panel shear
region of greater discrepancy between curves A strains is to cause some stringers to resist
less axial load than those calculated by beam
and B. theory. This decreased effectiveness of
A19. 17 Application to Practical Wing Section stringers is referred to as "shear lag" effect,
since some stringers tend to lag back from the
A practical wing section involves these position they would take if plane sections re-
facts: - (1) The section is unsymmetrical; main plane after bending.
(2) external load planes change their direction
under different flight conditions; (3) the In general, the shear lag effect in sheet-
material stress-strain curves are different in stringer structures is not appreciable except
tension and compression in the inelastic range. for the following situations: -
Since the stress analyst must determine (1) Cutouts which cause one or more stringers
critical margins of safety for many conditions, to be discontinued.
it would be convenient to have an interaction
curve involving Mx and Mz bending moments which (2) Large abrupt changes in external load
would cause failure of the wing section. This applications.
interaction curve could be obtained as follows:-
(3) Abrupt changes in stringer areas.
(1) Choose a neutral axis direction and its
location. In Chapters A7 and AS, strains due to
shearing stresses were considered in solving
(2) Assuming that plane sections remain plane, for distortions and stresses in structures in-
and taking the maximum strain as that volving sheet-stringer construction. Even in
causing failure of the compressive flange, these so-called rigorous methods, simplifying
use the stress-strain curve to determine assl®ptions must be made as for example, shear
the longitudinal stress and then the stress is constant over a particular sheet
internal load on each element of the cross- panel and estimates of the modulus of rigidity
section. A check on the location of the for sheet panels under a varying state of
assumed neutral axis is that the total buckling must be made. The number of stringers
compression on cross-section must equal and sheet panels in a normal wing is large,
total tension. Since the location was thus the structure is statically indeterminate
assumed or guessed, the neutral axis must to many degrees and solutions necessitate the
be moved parallel to itself to another use of high speed computors. Before such
location and repeated until the above analyses can be made, the size and thickness of
check is obtained. each structural part must be known, thus rapid
approximate methods of stress analysis are
(3) Find the internal resisting moment about desirable in obtaining accurate preliminary
the neutral axis and an axis normal to the sizes to use in the more rigorous elastic
neutral axes. Resolve these moments into analysis.
moments about x and z axes or Mx and Mz .
These resulting values of Mx and Mz are To illustrate the shear lag problem in its
bending moments which acting together will simplest state, consider the three stringer-
cause failure of the wing in bending. sheet panel unit of Fig. A19.36. The three
stringers are supported rigidly at B and equal
(4) Repeat steps 1, 2 and 3 for several other loads P are applied to the two edge stringers
directions for a neutral axis which results labeled (1) at point (A). The center stringer
will give additional combinations of Mx (2) has zero axial load at (A), but as end B is
and Mz moments to cause wing failure. Thus approached, the sheet panels transfer some of
an interaction curve involving values of the load P to the center stringer by shear
Mx and Mz which cause failure of wing in stresses in the sheet. At the support points B
bending is obtained and thus the margin of the transfer of load from side stringers to
safety for any design condition is readily center stringer is such as to make the load in
obtainable. all three stringers approximately equal or
equal to 2P/3.
A19.25
B .(1) A equals 3b. At beam section 2-2 stringers (5)

P and (7) have become effective since they inter-
I
2P/3 - -
Sheet J sect triangle at points (a) on section 2-2. At
section 3-3 point c, stringer (6) becomes fUlly
2P/3
B '(2) J!- effective.
Sheet b
2P/3 ~ ~P
B "'(1) A To handle shear lag effect in a practical
wing problem another column would be inserted
Fig. A19.36 in Table A19.1 between columns (3) and (4) to
take care of the shear lag effect. The shear
lag effectiveness factor which we will call R
The theoretical load in center stringer would equal the effectiveness obtained from a
can be calculated by methods of Chapter A7 and triangle such as illustrated in Fig. A19.38.
AS, and the results would give the solid curve
of Fig. A19.37. To simplify the solution, it For example, the shear lag factor R at
,is common practice to aSSlli~e the load distri- beam section 1-1 in Fig. A19.38 would be zero
bution in the center stringer to vary according for stringers (5), (6) and (7) and one for all
to the dashed curve in Fig. A19.37 which other stringers. At beam section 2-2 stringers
indicates that in a distance 3b, the load 2P (5) and (7) have a factor R = 1.0 since they
is equalized between the three stringers. are fUlly effective at points (a). Stringer
(6) is only 50 percent effective since section
2-2 is halfway from section 1-1 to point (c),
2P/31-- - - - - - - " - - Theory =
thus R 0.5 for stringer (6). At beam section
~
I" ---. Approximate 3-3, stringer (6) becomes fully effective and
Load in "~ =
thus R 1.0 for all stringers. The final
Stringer
(2)
I" ~, modified stringer area (A) in column (4) of
Table A19.1 would then equal the true stringer
o area plus its effective skin times the factors
B IE 3b --jA KR. The procedure from this point would be the
same as discussed before. Thus shear lag ap-
Fig. A19.37 proximations can be handled qUite easily by
modifying the stringer areas. Using these
A19.19 Application of Shear Lag Approximation to modified stringer areas, the true total loads
Wing with Cut-Out. in the stringers are obtained. The true stresses
equal these loads divided by the true stringer
Fig. A19.38 shows the top of a multiple area, not the modified area.
stringer wing which includes a cut-out in the
surface. The stringers (5), (6) and (7) must A19.20 Approximate Shear Lag Effect in Beam Regions
be discontinued through the cut-out region. where Large Concentrated Loads are Applied.
Wing and fuselage structures are often re-

qUired to resist large concentrated forces as
"I-H-~:tt-------;-------r---------l for example power plant reactions, landing gear
reactions, etc. To illustrate, Fig. A19.39
" represents a landing condition, with vertical
load. The wing is a box beam with 7 stringers
Top Surface A
,~ Effective-
3b ----'1 ..., ~ ness Tri- v
~I ~ ane:le
t -5, /1
Ib ;;1 ./ I
~ ./
I At-- 3b --I /
1
~ t t t f f .j-
Fig. A19.38
*~
~
Dead wi.
inertia forces
It is assumed that the effectiveness of
these 3 interrupted stringers is given by the
triangles in the figure. At beam section 1-1 0t Ben'!lng Mom.
these stringers have zero end load. The
stringer load is then assumed to increase
linearly to full effectiveness when it inter- ~Fi.a
sects the sides of this triangle whose height Fig. A19.39
and flange members. Fig. (a) shows the bending .i. A

moment diagram due to the landing gear reaction 6"
T
c=r-:::----....
C B
alone. The internal resistance to this bending
TTr2~'~
moment cannot be uniform on a beam section
adjacent to section A-A because of the shear
strain in the sheet panels or what is called
shear lag effect. To approximate this stringer
effectiveness, a shear lag triangle of length
3b is assumed, and the same procedure as
discussed in the previous article on cut-outs 150" II I
9= I ~
is used in finding the longitudinal stresses.
It should be understood that the bending moments
1-rgien I
due to the distributed forces on the wing such
as air loads and dead weight inertia loads are j I Y axis at
not included in the shear lag considerations, ,I 33. 3% of chord
only the forces that are applied at concentrated
points on the structure and must be distributed
into the beam. A side load on the gear or -49"1-21" ~ 60" .,
power plant would produce a localized couple
plus an axial force besides a shear force as in tZ r~ A~
Fig. A19.39. The resistance to this couple and 4 x ~lrl) B~
axial force would likewise be based on the 3" l..C
effectiveness triangle in Fig. A19.39. Fig. A19.41
A19.21 Approximation of Shear Lag Effect for Sudden intensity acting upward in the z direction
Change in Stringer Area and 0.25 lb./in.s average intensity acting
stringers of one size are often spliced to rearward in the x direction. The center of
stringers of smaller size thus creating a dis- pressure for z forces is on the 25 percent
continuity because of the sudden change in of chord line measured from the leading edge
stringer area. edge and at mid-height of spar AB for the
x air forces. Assume the 3 stringers A,
Fig. A19.40 shows the stringer arrangement B, C develop the entire resistance to ex-
in a typical sheet-stringer Wing. Stringer B ternal bending moments. Find axial loads
is spliced at point indicated. The stringer in stringers A, B, C and the shear flaw in
area As is decreased suddenly by splicing into the 3 sheet panels of cell (1) at wing
a stringer with less area A1 • stations located 50", 100" and 150" from
wing tip. Consider structure to rear of
Top Surface of Wing
cell (1) as only carrying airloads forward
to cell (1) and not resisting wing torsion
;; Snlk~ Point or bending.
1-- --I
/
Fig. A19.42
/~ ~ J ~
/ s h 1
I I I T b Rear spar
/ I
Single , -- --
/
I
-3b +3b-.j- Pin !
Fittings .-,..,
a
_ _ -
-=-Front §par
C
Fig. A19.40
To approximate the shear lag effect,
assume the area of stringer B at splice point
to be the average area of the two sides or
(A 1 + As )/2. This average area is then assumed
to taper to As and A1 at a distance 3b from the
splice point. The shear lag effectiveness
factor R will therefore be greater than 1.0 on
the side toward the smaller stringer A, and
less than one on the side toward the stringer
with the greater area As, since the average (2) Fig. A19.42 shows a monoplane wing with
area was used for the splice point. one external brace strut. The Wing is
fastened to fuselage by single pins at
A19.22 Problems points (a) and (b). The fitting at (b) is
designed to take off drag reaction. The
(1) Fig. A19.4l shows a cantilever, 3 stringer,
single cell wing. It is subjected to a airloads are Wz =
40 Ib./in. of wing span,
distributed airload of 2 Ib./in. s average with center of pressure at 30 percent of
chord from leading edge and acting upward,
A19.27
and Wx =
5 lb./in., acting to rear and Table A gives the stringer areas at sta-
located at mid-depth of wing. Find re- tions 0 and 150. Assume stringers have
actions at points (a), (b) and (d). Find linear variation in area between these two
axial loads on front and rear spars. Find stations. Use 30t as effective skin with
primary bending moments on front spar. compression stringers.
Find shear flow on webs and walls. Neglect
structure forward of front spar and rear- Find axial loads in stringers at stations
ward of rear spar. 150 and 130 and determine shear flow
system at station 150.
(3) Fig. A19.43 shows a portion of a single
cell - mUltiple stringer cantilever wing. (4) Same as problem (3) but add an internal
The external air loads are: web of .04 thickness connecting stringers
(3) and (8).
Wz =
100 lb./in. acting upward and whose
center of pressure is along a y axis coin- (5) Same as problem (4) but add a leading edge
ciding with stringer (8). cell with radius equal to one-half the
front spar depth. Take skin thickness as
Wx =
6 lb./in. aC4ing to rear and located .04 inches.
at mid-depth of wing.
fDt
-L
6 5
..i
No.
1
2
TABLE A
Flange n,-;=-,.A~r:,=eTa..,-,:::-n-j
Sta.150 Sta 0
1.0
.25
.20
.10
~3..jo.3+-3+ 3j 3 .25 .10
t--- 12 4 .25 .10
5 .80 .10
6 .70 .10
7 .25 .10
8 .25 .10
9 .25 .10
10 .85 .20
II \
/ I
- STA.O
I
I
I~
I
!
I
\\
.032
T
12"
.050
.032
1"1~
1 .025
t:5~ 5 -+ 5
20
+5 j ,.. 150" _I
STA.150 STA.O
Fig. A19.43
North American Aviation FJ3 "Fury Jet". View Shows Bottom Wing Skin.
Note Integral Construction of Skin and Stringers.
North American Aviation F-IOO "Super Sabre" Wing. View Shows End
of Outer Wing Panel. Note Thick Skin.
Douglas DC-8 Wing. View Showing Lower Surface of Outer Wing Panel
Between Center and Rear Spar.
CHAPTER Azo
INTRODUCTION TO FUSELAGE STRESS ANALYSIS
A20.1 General. In general the purpose of an air- to Chapter A5, where exanple calculations of
plane is to transport a commercial payload or a fuselage shears and moments are presented.
military useful load. The commercial payload
of a modern airliner may be 100 or more pass- The basic fuselage structure is essentiall=
engers and their baggage. These passengers a single cell thin walled tube with many trans-
must be transported safely and comfortably. verse franes or rings and longitudinal stringerr
For example, an airliner flies at high altitudes to provide a combined structure which can absorl
where temperatures may be far below zero and and transmit the many concentrated and distri-
where the air density is such as not to sustain buted applied forces safely and efficiently.
human life. These facts mean that the body The fuselage is essentially a bean structure
which carries the passengers must be heated, subjected to bending, torsional and axial
ventilated and pressurized to provide the forces. The ideal fuselage structure would be
necessary safety. Air travel must be acceptable one free of cut-outs or discontinuities, how-
to the passengers, thus the airplane body must ever a practical fuselage must have many cut-
shield the passengers from excessive noise and outs. Fig. (a) shows the basic interior fuse-
vibration, and furthermore efficient, restful lage structure of a small airplane with skin
and attractive furnishings must be provided to removed. It consists of transverse frames and
make travel enroute comfortable and enjoyable. longitudinal stringers. Photographs 1, 2 and 3
The portion of the airplane which houses the illustrate fuselage construction of late model
passengers on payload is referred to as the large aircraft.
fuselage. Fuselages vary greatly in size and
configuration. For exam)le, the fuselage of a
supersonic military airplane may house only one
passenger, the pilot, the remainder of the
fuselage interior space being used to house the
power plant, to provide retracting space for
landing gear, and to house the many mechanical
and electronic installations which are necessary
to fly the airplane and carry out the various
operations for which the airplane was designed
to accomplish. Many groups of engineers with
various backgrounds of training and experience
are therefore concerned with the design of the
fuselage. The structures engineer plays a very
important part because he is responsible for
the strength, rigidity and light weight of the
fuselage structure. Fig. (a)
A20.2 Loads. Basic Structure.
The wing, being the lifting body is sUb-

jected to large distributed surface air forces,
whereas the fuselage is subjected to relatively
small surface air forces. The fuselage is sub-
jected to large concentrated forces such as the
wing reactions, landing gear reactions, empen-
nage reactions, etc. In addition the fuselage
houses many items of various sizes and weights
which therefore SUbject the fuselage to large
inertia forces. In addition, because of high
altitude flight, the fuselage must withstand
internal pressures, and to handle these internal
pressures efficiently requires a circular cross-
section or a combination of circular elements.
The student should refer to Chapter A4 for PHOTO. NO.1

further discussion of loads on aircraft and also Fuselage Construction of Fairchild F-27 Transport
A20.1
FUSELAGE
PHOTO. NO.2
View Looking Inside of Rear Portion of Fuselage
of Beechcraft Twin-Bonanza Airplane.
PHOTO. NO.3
Fuselage Construction of Boeing
707 Jet Airliner.
(FOR GENERAL DETAILS OF DOUGLAS DC-8 FUSELAGE CONSTRUCTION SEE PAGE A15. 32)
A20.3
A20.3 Stress Analysis Methods. Effective Cross- (1) A small width of sheet w. on each side of
Section. the attachment line of skin to stringer is
It is common practice to use the simplified considered as carrying the same c6mpressive
stress as the stringer, as was discussed in
beam theory in calculating the stresses in the
skin and stringers of a fuselage structure. If Chapter A19. These effective sheet Widths
the fuselage is pressurized, the stresses in the w. are shown as the blackened elements ad-
jacent to the stringers on the compressive
skin due to this internal pressure must be added
to the stresses which resist the flight loads. side in Fig. A20.l.
In wings the skin in the middle region of the
airfoil is relatively flat and thus the skin is (2) The remainder of the curved sheet between
usually considered as made up of flat sheet stringers, namely, b- (w. + we) carries a
panels. In fuselages, however, the skin is maximum compressive stress ocr = .3 E t/r.
curved and curved sheet panels have a higher This value for ocr is conservative. E is
critical compressive buckling stress than flat the modulus of elasticity of the skin
panels of the same size and thickness. In material, t the skin thickness, and r the
small airplanes, the radius of curvature of the radius of curvature of the skin. These
fuselage skin is relatively small and thus the curved sheet elements are shown by the
additional buckling strength due to this curva- hatched skin lengths in Fig. A20.1.
ture may be appreciable. A simple procedure of
approximately including the effect of sheet Since the thin curved skin between the
curvature will now be explained. stringers normally buckles under a compressive
stress far below the buckling strength of the
Fig. A20.l illustrates a distributed stringers, the curved sheet is treated as an
stringer type of fuselage section. Assume that element With varying effective thickness which
external loads are applied which produce bending depends on the ratio of the curved sheet buck-
of the beam about the Y axis With compression on ling stress ocr to the bending stress 0b exist-
the upper portion of the cell. ing at that point for bending of the fuselage
section. Hence the effective sheet thickness
for the curved sheet panels can be written,
- - - - (1)
or an effective area can be written
- - - - (2)
where b' is the width of curved sheet between

the effective sheet widths w., we' etc. (See
y y Fig. A20.l).
---f-N.A.
To illustrate this approach in obtaining
the effective cross-section of a fuselage
section, an example problem will be presented.
The example problem will be broadened to some
extent for the purpose of introducing the stu-
dent to design procedure.
A20.4 Example Problem.
Fig. A20.1 Let it be required to determine the stringer
arrangement for the approximate elliptical
Up to the point of buckling of the curved sheet shaped fuselage section shown in Fig. A20.2.
between the skin stringers, all the material in
the beam section can be considered fUlly effec- The following data will be assumed: -
tive and the bending stresses can be computed
by the general flexure formula crt =
Mz/I y , Design bending moment about y axis = 1,600,000
where I y is the centroidal moment of inertia of in. lb. (producing compression on upper portion).
the entire section. When a bending compression Zee stringers, one inch deep and with an area
stress is reached which causes the curved sheet equal to 0.12 sq. in. shall be used.
between stringers to buckle, are-arrangement The ultimate compressive strength of the zee
takes place in the stress distribution on the stringer plus its effective skin and a length
section as a whole. Theory as well as experi- equal to fuselage frame spacing is assumed to
mental results indicate that the Ultimate com- be 32000 psi. The skin thickness is .032 and
pressive strength of a curved she~t With edge all material is (2024) aluminum alloy with E =
stringers can be approximated by the following 10,300,000 psi. The fuselage stringers are to
two assumptions. be symmetrical about section center lines.
A20.4 FUSELAGE STRESS ANALYSIS
w = 1.7 x .032-/10,300,000/32000 = .975 in.
~2.75' which equals a width of .975/.032 = 30.5 sheet

".,.-~' 3\
/\ 2 1
thicknesses. Since the bending stress decreases
to zero as the neutral axis is approached, and
,,;' 4 -1.---r-~ since the curved sheet between the Z stringers
can carry loads up to its buckling strength, a
~ 5 =
preliminary value of effective width w 40t
will be assumed acting with each stringer. Thus
total area of stringer plus effective skin
equals 0.12 + 40 x .032 2 = 0.16 sq. in. The
number of stringers required is therefore 2.00/
0.16 = 12.
Fig. A20,2 shows how the stringers were

placed to give 12 stringers on the top half.
y' ..L---lH!-+=--.L.----+-----_4-_ Since the skin on the lower half is in tension
and therefore fUlly effective, the neutral axis
will fall below the center line and thus the
two stringers on the center line will be con-
sidered as part of the required 12 stringers.
15 A fuselage cross-section has now been obtained.
The desired final result is that the maximum
16
Fig. A20.2 compressive stress will be near but not over
17
32000 psi. The procedure from this point is
18 still a trial and error process since the
19 Symmetrical about
the Z axis
effective sheet on the compressive side depends
20 on the ma~nitude of the compressive bending
21 2223 stress which in turn is influenced by the amount
Location of stringers of effective sheet and the buckling load carried
below Y' - Y' same as by the curved sheet.
above Y' - Y' .
Using the preliminary stringer arrangement
Solution: of Fig. A20,2, Tables A20.1 and A20.2 give the
calculation of the effective moment of inertia
The first thing to do is to determine ap- of the section about the horizontal neutral
proximately how many Z stringers will be re- axis, Table A20.1 deals with the stringers and
qUired so that a section can be obtained to the effective sheet elements and Table A20.2
work with. Since the internal resisting mom- deals with the curved buckled sheet elements.
ent must equal the external bending moment, one
can guess at the internal resisting couple in In the trial No.1, the following assump-
terms of total compressive flange stress and an tions are made:
effective internal couple arm.
(1) A width of 30 thicknesses of skin act
For elliptical and circular sections with with each stringer on the upper or compressive
distributed flange material, the approximate side.
effective resisting arm of the internal couple
can be taken as 0.75 times the height h, and (2) The area of the curved sheet between
the average tension or compressive stress as the effective sheet widths as found in (1) is
2/3 the maximum stress. Thus equating the modified to give an effective area by multi-
external bending moment to the internal resist- plying by a K factor of Ocr/ob' where ocr is
ing moment an approximate total area Ac for the the buckling compressive stress and 0b is the
compressive side of the fuselage section can be bending stress at the center of the curved sheet
obtained. element assuming 32000 at the extreme upper
fiber of the beam section and zero at the hori-
My = Ac (.67 0b)(' 75 h), whence, zontal center line, with linear variation in
between these points.
1,600,000
A
c = .75x50x.67x32000 = 2.0 sq
. in
• NOTE: Since the entire skin on the lower half
is effective, a more logical assumption would
Part of this total area is provided by the be to guess at the location of the neutral axis
effective skin area. The effective width to and use a variation of 0b between the neutral
use with each rivet line equals w = ct )E/Ost. axis and the extreme fiber. This approach will
=
We will take C 1.7 which is a commonly used not be used in this example.
value.
A20.5
TABLE A20.1
TRIAL NO. 1 TRIAL NO. 2
1 2 3 4 5 6 7 8 9 10 11 12
Explanatory Notes for Table A20.l
"
0
II:
1"1
1 ,... 0
"I
~~
.,......
".
.... iO .... TRIAL NO. 1
...."., .....
bll .. Area - ::l a>
., " "'''
.l<l Col. 1 For numbering of stiff-
.." " sq.a in.
.... El
Arm
z'
az' az t2 "I
1
0
.... .......... .....
..." "> ......
az az 2 eners-ind sheet elements, see
Fig. A20.2.
""".... 1>1
" ~
"I
I
"
1>11
iO ..
1>1 ...
0::1>1 - - - 30 x .032 x .032 .12.15+
Col. 2 Stiffener area
2 .15 24.2 3.63 87.8 27.58 -30020 1.0 .032 .152 4.19 115.4 sq. in. For stiffeners
.... 46 .15 22.0 3.30 72.6 25.38 -27700 1.03 .033 .153 3.88 98.5 2, 4, 6, 8, 10. Below
"
.l<l
m 8
.15 18.2 2.73 49.7 21.58 -23550 1.12 .036 .156 3.36 72.4 the centerline each
.15 13.3 1.99 26.4 16.68 -18200 1. 28 .041 .161 2.69 44.7 stiffener is considered
" 10 .15 6.9 1.04 7.2 10.28 -11200 1.66 .053 .173 1. 78 18.3 acting separately. The
.~ 12 .12 0 0 0 3.38 - 3680 2.88 .092 .212 0.72 2.4 entire skin between
t 13 .224 -3.2 -0.72 2.3 -0.18 .224 - .04 0.7 stiffeners is consider-
" 14
~ 15
.120
.224
-6.9
-10.1
-0.83
-2.28
5.7
23.0
-3.52
-6.72 "
....0 ...."
0
.120
.224
-0.42
-1. 52
1.5
10.2
ed as a unit.
Col. 3 All arms z' are measur-
1>1 16 .12 -13.3 -1.60 21.3 -9.92
III ., .12 -1.19 11.8 ~ ed to horizontal cen-
417
., 18
.160
.12
-15.8
-18.2
-2.53
-2.19
40.0
39.8
-12.42
-14.82
""
foo ""
foo
.160
.12
-1. 99
-1. 78
24.7
26.4
terline axis.
"'19
g',,20
;; 21
.16
.12
.16
-20.3
-22.0
-23.7
-3.24
-2.64
-3.80
65.8
58.0
90.0
-16.92
-18.62
-20.32
.."
....
"
"
....
...."
.16
.12
.16
-2.70
-2.24
-3.25
45.6
41. 6
66.0
TRIAL NO. 2
Col. 6 z = distance
axis as found from re-
to neutral
t 22 .12 -24.2 -2.90 70.2 -20.82 ..::" .l<l .12 -2.50 52.1 suIts of Trial No. 1.
m 23 .088 24.9 -2.19 54.5 -21.52 m m .088 -1. 89 40.7 Col. 7 Gb =
1600000 z/1470
SUM _ 2.49 Col. 8 _ effective width based
-12.23 714.3 2.627 -2.86 673 on stress in Col. 7
TABLE A20. 2
TRIAL NO 1 TRIAL NO 2
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18
0 - ...1>1 .0.
'+-l
." ..
II:
,...
. .,
• .0
.. b
,,' ".0
.....
1"1
1
"I
.....;1>1 ,... '"
0"
..
.." "
..::"
mEl "
b' t r
11
... 1>1 ....
'""§
.,
.0
~ bb "<:.: z' az' az t2 -N ill}
• 6'
11
0
a>
0
....I ...
>- (,)
" .. 0::",
.....
• b" ."
.,..::
>",. >< •
..
'".......
".,,,:.: az az 2
...."
1>1 b"'". Po
"b
<
...
... 11
1>1:':
...
... 11
1>1" N 11 0::"":':
II
QJ-I"'4,.c
"
0:: ..
.,"
..::
1 2.27 .032 11 -9000 -31900 .282 .020 24.9 0.50 12.4 28.28 -30800 .292 2.25 .021 .59 16.7
3 4.04 .032 11 -9000 -30300 .296 .038 23.7 0.90 21. 3 27.08 -29500 .305 3.99 .039 1. 06 28.7
'""
.... 5
7
4.04
4.04
.032
.032
24
38
-4510
-2600
-26000
-20200
.173
.128
.022
.016
20.3
15.8
0.45 9.1
0.25 3.9
23.68
19.18
-25800
-20900
.174
.124
3.93
3.80
• 022
.015
0.52 12.3
0.29 5.5
.l<l
";:l
III
9
11
6.04
6.52
.032
.032
38
38
-2600
-2600
-12900
- 4100
.201
.63
.039
.131
10.1
3.2
0.40 4.0
0.42 1.3
13.48
6.58
-14700
- 7200
.177
.362
5.53
4.73
.031
.055
0.42 5.7
0.36 2.4
TOTALS .266 2.92 52.0 TOTALS .183 3.24 71.3
Explanatory Notes for Table A20.2 Explanatory Notes for Table A20.2
Trial No. 1 Trial No. 2
Col. 1, 2, 3, 4 (see Fig. A20.1 for meaning of terms) Col. 12 z - distance to neutral axis
Col. 5 E = 10,300,000 for aluminum alloys as found in results of Trial
Col. 6 0b varies as a straight line from 32000 at No. 1
top of cell to zero at centerline Col. 13 Gb = 1600000 z/1470
Col. 9 z' = distance from centroid of element to Col. 14 based on stress J b of Col. 13
centerline axis of cell
Results of Trial No.1 Results of Trial No.2

Considering results of both Tables and multiplying by 2 since Total effective area _ (2.627 + .183)2 5.62
only one half of cell was considered:
~az • (-2.86 + 3.24)2 • 0.76
i a total effective area = (2.49 + .266) 2
_ 5.51 sq. in.
z= .76/5.62 = .135" above N.A. of trial 1.
2
INA = 2(673 + 71.3) - 5.62 x .135 = 1489
faz' • (-12.23 + 2.92)2 = -18.62
z • Zaz'/Za = -18.62/5.51 _ -3.38"
2
INA = 2(714.3 + 52) - 5.51 x 3.38 • 1470 in.
Due to the s~~etry of the section, Tables V

A20.1 and A20.2 give calculations for only one- qo + I z Zaz
y
half of the material, thus the results are
multiplied by two. General explanatory notes =0 +
are given below each table.
(Vz ) 1 9. 67
1489 = .0132 Vz lb. / in.
The results of trial No. 1 give a neutral The shear stress ~ = q/t = .0132 Vz /.032 = .413
axis 3.38" below the center line and a moment of Vz
inertia of 1470 in. 4 • In Trial No.2, the ef-
fective sheet widths are based on the moment of The average shear stress on the section would
inertia of 1470. The results of trial No.2 be ~av' = Vz /2ht = Vz /2x50x.032 = .312 Vz ·
give a moment of inertia of 1489 in. 4 with a
neutral axis .135" above the first location. If Thus for this shape of cross-section and
a third trial were used, making use of the 1489 stringer arrangement the maximum shear stress
moment of inertia, the change would be qUite is .413/.312 times the average shear stress or
small since the effect of a small change in approximately 4/3 times as large.
stress on the effective sheet width is negligi-
ble. The procedure as given above is qUite con-
servative relative to the true or actual margin
The compressive stress on stringer No. 2 of safety, because a linear variation of stress
using the resulting moment of inertia and with strain has been assumed and failure of the
neutral axis location, therefore becomes section is assumed to occur when the most remote
stringer reaches its Ultimate compressive stress.
ab = MyZ/I y = 1,600,000x27.45/1489 =29,500 Actually in a static test of a fuselage to
psi destruction, the fuselage section as a whole
will not collapse when one stringer buckles,
The allowable stress was 32000, hence the but will continue to take increasing load until
margin of safety is (32000/29500) - 1 = .08 or other stringers have reached their ultimate
eight percent. If a smaller margin of safety strength. Furthermore, in a typical fuselage
was desired some material would be eliminated structure, stringers of various sizes, shapes
and the calculations of Tables A20.1 and A20.2 and therefore different compressive strengths
would be repeated. are used, and thus to obtain a better measure
of the ultimate strength of a fuselage section,
Calculation of Shear Stress in Skin at Neutral modifications in stress procedures are made to
Axis measure stringer effectiveness. This SUbject
was discussed in some detail in Arts. 11 and 12
The equation for the shear flow q at some of Chapter A19. To illustrate stringer effect-
point on the skin is, iveness in fuselage bending stress analysis, a
simple example problem will be presented.
q qo - Vz Zaz - -' - - - - - - - - - - (3) A20. 5 Ultimate Bending Strength of Fuselage Section.
Iy
Example Calculation.
Due to symmetry of cross-section about the Z Fig. A20.3 shows the cross-section of a
axis the shear flow qo is zero at a point on
the center line Z axis. The summation of the circular fuselage. The Z stringers are arranged
symmetrically with respect to the center line
term az between a point on the Z axis and the Z and X axes.
neutral axis is given in Table A20.3. The
values of areas (a) and arms (z) are taken Three sizes of Z stringers are used as
from Tables A20.1 and A20.2. illustrated in Fig. A20.4 and are labeled S~,
TABLE A20.3
S2 and Sa' These symbols are used on Fig.
A20.3 to indicate where each type of stringer
Element No. az is used. The stringers on each side of the
section are numbered 1 to 13 as shown on Fig •
(1)
(2)
(3)
.021(28.28
.152(27.58
.039(27.08
-
-
-
• 13)
.13)
.13)
··= 0.59
4.17
1,05
A20.3. Fig. A20.5 shows a plot of the stress-
strain curve for the three stringer types loaded
(4) .153(25.38 - .13) = 3.86
in compression and with a column length equal to
·==
(5) .022(23.68 - .13) 0.52
(6) .156(21.58 - .13) 3.34 the fuselage frame spacing. Fig. A20.5 also
(7) .015 (19.18 - .13) 0.28
(8) .161(16.68 ~ .13)
= 2.66 shows a tension stress-strain diagram for the
(9) .031(13.48 - .13) 0.41
(10) .173(10.28 - .13) 1.75 material which is aluminum alloy (2024). The
(11) .055(6.58 - .13) 0.35 ultimate bending strength will be calculated
(12) .212(3.38 -
Total az
.13) 0.69
19.67 for bending which places the upper portion in
compression.
Substituting in equation (3)
A20.7
1 Since the location of the neutral axis is

unknown, a location will be assumed, namely, 7
inches below the center line axis as shown in
Fig. A20.3. The entire calculations for deter-
mining the effective moment of inertia can best
be done in Table form, as shown in Table A20.4.
Due to symmetry about the Z axis only one-half
of the structure need be considered since the
CL Axis. results can be multiplied by two.
x Column (1) lists the stringer numbers

Assumed"Y relative to location and Column (2) according
Neutral Axis
to types S1.' 8 2 and S". Column (3) gives the
stringer area. On the tension or lower side of
10 "
the section, the skin is all effective and that
;(4 11 s" ~ S
area of skin halfway to each adjacent stringer
is assumed to act with stringers numbered 9 to
'O~ Y 12
13 . 0036 13 and this skin area is recorded in Column (5) •
Fig. a On the compressive side the skin is only
Strain partially effective. The effective width w for
Fig. A20.3 Diagram each stringer rivet line depends on the stringer
stress. We will take the effective width w =
1.9t VE!OST. The effective area Ae will then
13 equal wt. These effective skin areas are re-
.!.! ~ 16 -of 9 corded in Column (5). In solving this equation
~r1iH+.05"8
~~h~~
.06 11. 1;11. ll\~'
the stringer stress 08T has been taken as -36500
1" [.!
L L ~ L
8 L:5 psi on stringer number (1), and then varying
linearly to zero at the neutral axis as indi-
cated in Column (4). This assumption is not
Stringer S1. Stringer S2 Stringer S" true but accurate enough to obtain effective
Area; .135 Area; .18 Area; .08 skin areas. To illustrate, consider stringer
Fig. A20.4
number (1). The effective area Ae equals wt 2 =
1.9 x .032 2 (10,300,000/36500)i = .032 sq. in.
Column (6) gives the sum of the stringer and
effective skin areas or As + Ae .
In this example problem, the effectiveness
of the curved sheet panels between the sheet
effective widths will be neglected since its in-
fluence is small. It could be inclUded as il-
lustrated in the previous example problem.
Column (7) lists the distances from the assumed
neutral axis to the centroid of each stringer-
skin unit. We now assume that plane sections
remain plane or a linear strain variation.
Referring to Fig. A20.5, it is noticed that when
a unit strain of .006 is obtained in stringer
81. type the compressive stress is 36500, which
represents its ultimate stress. Stringer (1)
is of 81. type and is located farthest from
the neutral axis. Sub. Fig. (a) of Fig. A20.3
shows the strain diagram with .006 at stringer
(1) and varying as a straight line to zero at
tt, ; t
: C , [ the neutral axis. Column (8) of Table A20.4
l~ :,
records the unit strain at each stringer cen-
;I~
troid. The true stress at each stringer point

,I
due to these strain values is read from the
l-t curves on Fig. A20.5 and recorded in column (9)
l th Hi of the Table. It should be noted that stringer
;~
(3) although closer to the neutral axis than
r j H : ,1
~
" stringer (1) carries a higher stress than

stringer (1). This is possible because when
stringer (1) reaches its maximum-stress, it
bends but continues to hold the same stress with
TABLE A20.4
1 2 3 4 5 6 7 8 9 10 11 12 13
Stringer Stringer Linear Effect- Total Arm Un.it Stress

ive Skin Stringer ZJ. KAZJ. KAZ 2
Area Stress Area Area Strain ATe. K = a/aJ. KA (Z = zJ.)
No. Type As aJ. (in) e a
b Ae A
1 SJ. 0.135 -36500 .032 .167 35.7 -.00600 -36500 1. 00 .167 5.96 213
2 SJ. 0.135 -34700 .034 .169 33.8 -.0057 -36500 1. 05 .178 6.00 203
3 S2 0.180 -31000 .036 .216 30.3 -.0051 -39100 1. 26 .272 8.25 250
4 SJ. 0.135 -26600 .039 .174 26.0 -.0048 -36000 1. 35 .235 6. 10 159
5 S" 0.080 -20500 · 044 .124 20. 1 -.0034 -31500 1. 54 .191 3.85 77
6 S" 0.080 -13400 · 054 · 134 13.7 -.0023 -24000 1. 79 · 240 3.30 45
7 S" 0.080 -7150 · 074 .154 7.0 -.0012 -12500 1. 75 · 270 1. 89 13
8 S" 0.080 0 0 · 080 10.0 0 0 1.0 .080 0 0
9 S" 0.080 6130 · 216 · 296 - 6.0 · 0010 10000 1. 63 · 483 -2.88 17
10 S" 0.080 12280 · 216 · 296 -12.0 · 0020 20500 1. 67 .494 -5.91 71
11 S" 0.080 16800 .216 · 296 -16.5 .0028 30000 1. 78 .526 -8.69 144
12 S" 0.080 20400 · 216 · 296 -20.0 · 0034 35000 1. 71 · 506 -10. 10 202
13 S" 0.080 21700 · 216 · 296 -21. 2 · 0036 38000 1. 75 · 517 -10.93 232
Z 4.159 -3.16 1626
increasing strain, but stringer (3 ) which has to Z. The effective moment of inertia is there
not reached its maximum strength of 39000 fore twice the sum of Column (13) or 3252.
continues to take increasing load.
Calculation of Ultimate Resisting Moment.
Since we wish to use the beam formula ab =
MyZ/I x in computing stresses, we must modify the The maximum stress at the most remote
stringer areas to give a linear stress variation stringer which is number (1) is 36500. From
since the formula is based on a linear stress the beam formula,
variation. The stringer modification factor K
equals the ratio of the true stress in column Mx = ObIx/Z
(9) of Table to linear stress value in column
(4) or K = a/a J.. The results are recorded in = (36500 x 3252)/35.7 + 0.7
column (10). The modified stringer areas are = 3,260,000 in. lb.
then equal to KA and are recorded in column
(ll) . Column (12) gives the first moment of This bending strength when compared to any desi gn
the modified areas about the assumed neutral bending moment about the X axis would give the
axis, giving a total value of -3.16. margin of safety relative to bending strength.
The distance Z from the assumed neutral If the moment of inertia had been computed
axis to the true neutral axis is thus, without regard to non-linear stress variation,
or in other words, using K equal 1 for all
Z = Z KAZ'/ZKA stringers the neutral axis would have come out
_ -3.16 _ 4.9 inches below the centerline axis and the
- 4.159 - - 0.76" moment of inertia would have calculated to be
2382 in. 4 The resisting moment developed
The true N.A. would fall about .70 inches would then be (36500 x 2382)/33.6 = 2,600,000
below assumed position. The effect on total in.lbs. Thus the true strength is 25 percent
greater than the strength for linear stress
sum of column (13) would be negligible, thus
variation. This result explains why such
Table A20.4 will not be revised.
structures test overstrength if designed on
linear stress variation basis.
Column (13) gives the calculation of the
effective moment of inertia with Z 1 being equal After stringer stresses are obtained
A20.9
using the modified areas of Table A20.4, the I y = (15 2 x .1 x 2) + (13.86 2 + 10.61 2 + 5.74 2 ).1
true stringer areas must be used to find the
true stringer loads, which must be used in the x 4 = 180 in. 4
shear flow analysis.
Table A20.5 gives the necessary calculations
A20.6 Shear Flow Analysis for Fuselage Structures for determining the flange bending stresses and
the net total shear load to be taken by the cell
The shear flow analysis can be made once skin. Since the cell is tapered, the stringers
the effective cross-sections of the fuselage have a z component, thus the stringer axial
are obtained. The procedure is the same as was loads help resist the external shear load. The
illustrated for wing structures in Chapter A19. summation of column (8) of Table A20.5 gives
To illustrate, two example problems will be -333.4 lb. for a summation for half the fuselage
presented. section.
Example Problem 1. Symmetrical Tapered Section. Hence, net web shear at station 0 equals:
Fig. A20.6 shows a portion of a tapered VWeb = Vext • + Vnange = 2000 + (2 x -333.4)
circular shaped fuselage structure that might
= 1333.2 lb.
be representative of the rear portion of a
fuselage for a small airplane. Since this
example is only for the purpose of illustrating The results in this particular problem show
shear flow analysis, it will be assumed that that at station 0 the flange stringer system re-
the 16 stringers are the only effective mater- sists one third of the external shear load. At
ial. In an actual stress analysis, the effect- station 150 the web system will resist the en-
ive cross-section would have to be used as tire external shear load of 2000 lb. since the
illustrated in previous articles A20.3 to A20.5. load in the stringers is zero.
The problem will be to determine the In actual design the net web shear should
stringer stresses and the skin shear flow stress be used since in many cases it will decrease
system at Station (0) under a given load system the sheet thickness required one or more gauges.
at Station (150) as shown in Fig. A20.6. Calculation of Flexural Shear Flow.
Solution No. 1 - Solution by Considering Beam Properties
at Only One Section. vZ(web) :6 a z = qo - li~~·2 :6 az
Iy
If the change in longitudinal stringer or
flange material is fairly uniform this method = qo - 7.40 :6 az (A)
can be used with little error in the resulting
shear flow stresses. Due to symmetry of the section about the Z
axis, the flexural shear flow in the web at the
Moment of inertia of section at station (0) center line is zero. Therefore, qo will be taken
about centroidal Y axis: as zero and the summation in equation (A) will
Area of all Stringers = . 1" 2000#

1
Frame spacing = 15"
15 t
II
q
5 II
13
II
I
11 7
~
10 8
9 Section at Station
Section at Station 0 14 150" 150
STA.O Length not to scale

STA.150
Fig. A20.6
A20.10 FUSE LAGE STRESS ANALYSIS
TABLE A20. 5 q34 :: -7.06 + 23.66 :: 16.60

1 2 3 4 5 6 7 8 9 q45:: -7.06 + 27.91 :: 20.85 etc.
Stiff- Arm
!Area Db = Px=Db
a Pz = Py =
ener
dz ~ Fig. A20.7 shows the results in graphical
z a -1667z (3)(4) dX dx _Px dz -Px~
No. (psi) lbs. dx dx form, on the left side of the section the shears
are of the same sign and therefore add together.
1 15 *.05 -25000 -1250 -.0333 0 -41. 6 0
2 13.86 .10 -23100 -2310 -.0308 -.0127 -71. 3 -29.4
3 10.61 .10 -17700 -1770 -.0236 -.0236 -41. 7 -41. 7
4 5.74 .10 - 9550 - 955 -.0127 -.0308 -12.1 -29.5
5 0 .10 0 0 0 -.0333 0 0
6 - 5.74 .10 9550 955 .0127 -.0308 -12.1 29.5
7 -10.61 .10 17700 1770 .0236 -.0236 -41. 7 41. 7
8 -13.86 .10 23100 2310 .0308 -.0127 -71. 3 29.4
9 -15 * .05 25000 1250 .0333 0 -41.6 0
Shear taken by stringers = -333.4 0
*1/2 of stringer one is assumed acting with each half of

cell.
NOTES:
Col. 4 Db = -Mz/I y = -2000 x 150 x z/180 = -1667z
Col. 5 Total x component of load in stringer member.
For practical purposes, it equals axial load in
stringers since cosine of a small angle is
practically one.
Col. 6 The slope of the stringers in the z and y Fig. A20.7

directions can be calculated from the dimen-
sions of the two end sections and the length of Solution No.2. Shear Flow by Change in Stringer Loads
the cell. (see Fig. A20.6) Between Adjacent Stations. 6P Method.
Col. 8 The in plane components of the stringer axial The shear flow will be calculated by con-
&9 loads at station O. sidering the change in the axial load in the
longitUdinal stringers between fuselage sec-
tions at stations (0) and (30).
start with stringer (1)
Fig. A20.8 shows the beam section at sta-
q:L .. :: 0 - 7.40 x .05 x 15 :: - 5.55 lb ./in. tion 30, the stringer areas being the same as
at station 0, but the section as a whole is
q .. 3 :: - 5. 55 - 7.40 x . 1 x 13. 86 .. - 15. 80 smaller due to the taper of the cell.
q34 :: - 15.80 - 7.40 x .1 x 10.61 =- 23.66
area of each
q45 :: - 23.66 - 7.40 x .1 x 5.74 =- 27.91 stringer = .10
The torsional moment T about the centroid y

of the section at station (0) equals 5 x 2000 ::
10000 in. lb. (clockwise when looking toward
station 150). Due to the symmetry of the sec-
tion at station 0, the in-plane components of 8
the stringer loads produce zero moment about Fig. A20.8 9
the section centroid.
Table A20.6 gives the calculations for the
For eqUilibrium a constant internal shear flexural shear system. The procedure is the
flow q:L is necessary to make Z Mx :: 0 same as illustrated for wing structures in
Chapter A19.
T
ql.:: 2A -10000
::--,---='::"'::"'::"":::"'--:: -7.06 lb./in.
2 x n x 15" Comparing the results of column 13 With the
flexural shear flow as found by solution No.1,
Adding the torsional shear flow q:L to the we find the second solution gives a maximum
flexural shear flow q, the following results shear flow of 28.95 lb./in. against a value of
are obtained: 27.91 for the first solution. The first method
deals with the properties at only one section
q:L .. = -7.06 + 5.55 :: -1.51 lb./in. and this cannot include the effects of change
in moment of inertia on the shear flow. The
q .. 3 = -7.06 + 15.80 = 8.74
A20.11
TABLE A20.6
1 2 3 4 5 6 7 8 9 10 11 12 13
Sta. 0 Sta. 30 Sta. 0 Sta. 30 Stringer Load liP Panel Flexural Shear
Stringer Area Area Sta. 0 Sta. 30 -°b=
- °b-= Sta. 0 Sta. 30 30- Taper llPK Flow
No. a a Arm Arm -1667Z -1522Z P x = P x = - (Col. 8 - Col. 9) Corr. 30 q = L: llPK
sq. in. sq. in. z z (psi. ) (psi. ) 0b a °b a 30 Factor 30
K lb./in.
1 .05 .05 15.00 14.00 -25000 -21300 -1250 -1065 6.17 .935 5.76
5.76
2 .10 .10 13.86 12.93 -23100 -19700 -2310 -1970 11. 33 .935 10.60
16.36
3 .10 .10 10.61 9.90 -17700 -15050 -1770 -1505 8.83 .935 8.25
24.61
4 .10 .10 5.74 5.36 - 9550 - 8155 - 955 - 815 4.65 .935 4.34
28.95
5 .10 .10 0 0 0 0 0 0 0 0 0
28.95
6 .10 .10 - 5.74 - 5.36 9550 8155 955 815 - 4.65 .935 - 4.34
24.61
7 .10 .10 -10.61 - 9.90 17700 15050 1770 1505 - 8.83 .935 - 8.25
16.36
8 .10 .10 -13.86 -12.93 23100 19700 2310 1970 -11. 33 .935 -10.60
5.76
9 .05 .05 -15.00 -14.0 25000 21300 1250 1065 - 6.17 .935 - 5.76
NOTES:
Col. 6 0b = -2000 x 150 z/180 = -1667 z
Col. 7 0b = -2000 x 120 z/157. 2 = -1522 z
Col. 10 Change in axial load in each stringer between stations 0 and 30 divided by distance between Stations. This
result represents the average shear flow induced by the loading up of each stringer between stations 0 and 30.
Col. 11 The width of a skin panel at Station 0 is 5.88 inches and 5.5 inches at Station 30. The shear flow on the edge
of the panels at Station 0 equals (5.5/5.88) llP/30. (See Art. A15. 18 of Chapter A15 for explanation). This
refinement is usually neglected and the average values as given in Col. 10 are used which are conservative.
Col. 13 Due to symmetry of structure, the shear flow is zero on z axis. Thus shear flow at any station equals the
progressive summation of the shear flow values in Col. 12.
second method is recommended for practical will assume the stringers are the only effective
analysis procedure. material. In actual design practice the effect-
iveness of the skin and each stringer would have
Since the section is symmetrical, there are to be considered as explained in Articles A20.4
no moments induced by the in-plane components of and 5.
the stringer forces at station O.
The problem will be to determine the
The torsional shear flow forces are the stringer stresses and the skin shear flow val-
same as in solution method No. 1 and these are ues at station (0) due to the given external
added to the values of column 13 of Table A20.6 loads of Pz = 4000 lb., Py = 1000 lb. and Px =
and give a pattern similar to Fig. A20.7. 1500 acting at station (150) as shown in Fig.
A20.9.
A20.7 Example Problem. Tapered Circular Fuselage
with Unsymmetrical Stringer Areas. SOLUTION:
Fuselage cross-sections are seldom all sym- Since we choose to use the liP method in
metrical relative to stringer and skin areas finding the shear flow system at station (0),
because the practical fuselage has cut-outs such we will find the stringer loads at two stations,
as doors, etc. To illustrate the unsymmetrical namely, station (0) and station (30). The first
case a simplified case will be presented. step is to find the moment of inertia of each
fuselage section about centroidal z and y axes
Fig. A20.9 shows a portion of a tapered and the product of inertia about these axes.
fuselage. The stringer areas are such as to
Table A20.7 (Columns 1 to 11) gives the calcu-
make the cross-sections unsymmetrical relative lations of the section properties for station
to bending material. Again for simplicity, we
FUSELAGE STRESS ANALYSIS
Fig A20 9
=--
d
~
(.1)
--.0::
e
-
a
10.5
Y'
<{,+----''------t-----.---tY'
Y
(.3) (.8)
......-:
-=
Sta. 150
Z
(.2) (.2) g - L- x
(.2)
-
foo
. ~---------150"
o.
l ----------_.I Section at Sta. 150
Z' Z Sta. 0
The skin stringers are located symmetrically with respect to the
Section at Station 0 centerline axes, however the stringer areas as given in ( ) on
the figure are not symmetrical with these axes. It is assumed in this problem that
the stringers taper uniformly between the values as given for station 0 and 150.
The cell would of course have interior transverse frames which are not shown on the
figure.
TABLE A20. 7
Section Properties at Sta. 0 Total Stringer Loads at Sta. 0
1 2 3 4 5 6 7 8 9 10 11 12 13 14
z = y= °c=F/Za P s = a(Ob + oel

Stringer Area Arm Arm az' az,2 ay' ay,2 az'y' °b
No. a z' y' z' - Z y' - y = -1500/3.40 = (2) (12 + 13)
a .60 10.5 -12.00 6.30 66.20 -7.20 86.50 - 75.50 11.36 -14.46 -15080 -441 -9312
b .10 18.98 - 8.48 1. 90 36.00 -0.85 7.20 - 16.10 19.84 -10.94 -21960 -441 -2239
c .10 22.50 0 2.25 50.80 -0 0 0 23.36 - 2.46 -22600 -441 -2304
d .10 18.98 8.48 1. 90 36.00 0.85 7.20 16.10 19.84 6.02 -16742 -441 -1719
e .80 10.5 12.00 8.40 88.10 9.60 115.10 100.80 11. 36 9.54 - 7692 -441 -6506
f .80 -10.5 12.00 -8.40 88.10 9.60 115.10 -100.80 - 9.65 9.54 11958 -441 9216
II: .20 -18.98 8.48 -3.80 72.00 1. 70 14.40 - 32.30 -18.13 6.02 18798 -441 3673
h .20 -22.50 0 -4.50 101. 60 0 0 0 -21. 65 - 2.46 19485 -441 3809
i .20 -18.98 - 8.48 -3.80 72.00 -1. 70 14.40 32.30 -18.13 -10.94 13610 -441 2634
j .30 -10.50 -12.00 -3.15 33.10 -3.60 43.25 37.80 - 9.65 -14.46 4592 -441 1246
Sum 3.40 -2.90 643.9 8.40 403.2 - 37.70 -1500
Reference Axes Z' and Y' are taken as the centerline axes. General Notes:
(see Fig. A20.9)
Col. 12 0b = 307.0y - 936.1z
Location of centroid and transfer of properties to
centroidal axes. Col. 14 Since the total tensile stresses equal to
total compressive stresses in bending,
Z = -2.90/3.40 = -.855" the sum of Col. 14 should equal the ex-
ternal applied normal load.
y = 8.40/3.40 = 2.46"
Iy = 643.9 - 3.40 x 855 2 = 641. 4
I z = 403. 2 - 3. 40 x 2.46 2 = 382. 6
I zy = -37.7 - 3.40 x 2.46 x -.855 = -30.55

A20.13
TABLE A20.8
Section Properties at sta. 30 Total stringer Loads at sta. 30
1 2 3 4 5 6 7 8 9 10 11 12 13 14
z = y= 0c = Fa a Ps =
Stringer Area Arm Arm az' az,2 ay' ay,2 az'y' °b
No. a z' y' z' - Z y' - y = -1500/2.98 a(Ob + 0c)
a .50 9.80 -11. 2 4.90 48.1 -5.60 62.8 -54.9 10.90 -13.31 -14800 -503 -7651
b .10 17.72 - 7.92 1.77 31. 4 -0.79 6.3 -14.0 18.82 -10.03 -21090 -503 -2159
c .10 21. 00 0 2.10 44.1 0 0 0 22. 10 - 2.11 -21447 -503 -2195
d .10 17.72 7.92 1.77 31. 4 0.79 6.3 14.0 18.82 5.81 -15647 -503 -1615
e .66 9.8 11. 20 6.47 63.2 7.40 83.0 72.5 10.90 9.09 - 7088 -503 -5011
f .66 - 9.8 11. 20 -6.47 63.2 7.40 83.0 -72.5 - 8.7 9.09 -11282 -503 7100
g .20 -17.72 7.92 -3.54 62.8 1. 58 12.6 -28.0 -16.60 5.81 -17583 -503 3416
h .20 -21. 00 0 -4.20 88.2 0 0 0 -19.90 - 2.11 17943 -503 3489
i .20 -17.72 - 7.92 -3.54 62.8 -1. 58 12.6 28.0 -16.60 -10.02 12135 -503 2327
j .26 - 9.8 -11.20 -2.55 25.0 -2.92 32.6 28.6 - 8.70 -13.31 3570 -503 797
Sum 2.98 -3.29 520.2 6.28 299.2 -26.3 -1500
NOTES:
- -- N Z'
(j)
Reference Z' and Y' axes are taken as the centerline axes. r0-
b d
Z = -3.29/2.98 = -1. 10" .10 .10
7f"'Y.2"
Y= 6.28/2.98 = a
-Iz e Section at
It'·
2.11 .50 _.66 Station 30
I y = 520.2 - 2.98 x 1. 10 2 =516.6 lL.~
~Y
, -1.1
c-= b-- -~
I z = 299.2 - 2.98 x 2.11 2 = 286.0 1t Y=
9.8 2.11
I zy = -26.3 - 2.98 x 2.11 x -1. 10 = -19.4 j .26 .66 f
\ I 1 . "
FIg. A20.10
(0) and the similar columns of Table A20.8 gives My = 4000x120+1500x8.10 = 492150 in.lb.
the calculations for station (30). Mz = -1000 x 120 + 1500 x 2.11 = 116820 in .lb •
Before the bending and shear stresses can Pn = -1500 lb., Vz =4000 lb., Vy = -1000
be calculated, the external bending moments, lb.
shears and normal forces at stations (0) and (30)
must be known. Calculation of Bending Stresses.
Stat i on (0):
At station (0): -
The bending moment about y neutral axis at
station (0) equals, where
My = Pz (150) Px (7.85)
+
K1 = Iyz/(IyI z - I yz 2)
= 4000 x 150 + 1500 x 7.85 = 611800 in. lb. ~ = Iz/(IyI z I yz 2)
Mz = Py (150) - Px (2.46) K" = Iy/(IyI z I yz 2)
= -1000 x 150 + 1500 x 2.46 = -146310 in. lb. Substituting values from Tables A20.7 and
A20.8:
The shears at station (0) are Vz = Pz =
4000 lb. and Vy =
Py = -1000 lb. K1 = -30.55/(641.4 x 382.6 - 30.55 2 ) =
The normal load Pn at station (0) referred = -30.55/244670 = -.0001248
to centroid of section equals ZPx -1500 lb. =
K2 = 382.6/244670 = .00156
In a similar manner, the values at station
(30) are, (see Fig. A20.10) K" = 641.4/244670 = .00262
A20 14 FUSELAGE STRESS ANALYSIS
SUbstituting K values in equation for ab: moments in the plane, of all internal and ex-
ternal forces must be zero. Column (7) of Table
ab = - [00262 x -146310 - (-.0001248 x A20.9 gives the moment of the flexural shear
about this point. (See notes and Fig. below
611800)J y - [.00156 x 611800 - Table for explanation.)
(-.0001248 x -146310)J z
TABLE A20.9
whence SHEAR FLOW CALCULATIONS
ab = 307.0 y -936.1 z (plus ab is tension) 1 2 3 4 5 6 7 8 9
Station (30): q~ ~ ~
Stringer P x at Px at l\p x Z l\p x m mq ql

K:I, = -19.4/(516.6 x 286 - 19.4 2
)
No. Sta. 0 Sta. 30 30 3D" sq. in. q + ql
= -19.4/147620 = -.0001315 a
b
-9312
-2239
-7651
-2159
55.37
2.67 55.37
150.04 8300 -52 3.37
-2195 58.04 202.46 11820 -52 6.04
c -2304 3.63
K 2 = 286/147620 = .001936 d -1719 -1615 3.47 61.67 202.46 12500 -52 9.67
-52
e -6506 -5011 49.83 65.14 150.04 9780 13.14
K 3
= 516.6/147620 = .0035 f 9216 +7100 -70.53 114.97 252.0
150.04
29000
6660
-52
-52
62.97
g 3673 3416 - 8.57 44.44 - 7.56
35.87 202.46 7270 -52 -16.13
ab = - [:0035 c -116800 - (-.0001315 x h 3809 3489 -10.67
-52
i 2634 2327 -10.23 25.20 202.46 5100 -26.8
492150] y - [001936 x 492150- 1246 797 -14.97 14.97 150.04 2240 -52 -37.03
j
-52 -52.0
(-.0001315 x 116820 II z
a 0 252.0
Sum
0
92670
whence NOTES:
ab = 344.3y- 937.7z Col. (2) and (3) from tables

A20. 7 and A20. 8
Col. (4) l\ P x ~ - [!XSta. 0
Column (12) in Tables A20.7 and A20.8 gives the
- P XSta . 3~
results of solving the equations for abo
Since an external load of 1500 lb. is act- Col. (6) m ~ double areas
ing normal to the sections and through the (see Fig. a).
Col. (7) mq ~ moment of
section centroids, an axial compressive stress shear flow q on
a c is produced on the sections. (See Columns each web element
13). The total load Ps in each stringer equals about 0' (Fig. a)
the area of the stringer times the combined
bending and axial stresses. (See column 14 of
each table).
Calculation of Flexural Shear Flow q. Moments Due to In Plane Components of Stringer Loads.
Table A20.9 gives the necessary calcula-
tions to determine the shear flow at station (0) Since the stringers are not normal to the
based on the change in stringer loads between section at station (0), the stringers have in-
stations (0) and (30). The correction of the plane 'components which may produce a moment
average shear due to the taper in the skin about the intersection of the symmetrical axes
panels as was done in example problem (1), which has been selected as a moment center.
Table A20.6, column (11), is omitted in this Table A20.10 gives the calculations for the in-
solution since it tends toward the conservative plane components and their moments about point
side. Since the effective cross section is un- 0' .
symmetrical, the value of the flexural shear Moment of External Load System About Point (0').
flow q at any point is unknown thus a value for
q at some point is assumed. In Table A20.9 The 1000 lb. load at station (150) acting
the shear flow q in the web aj is assumed zero. in the Y direction has a moment arm of 7" about
Column (5) gives the results at other points the point 0' of station (0).
under this assumption.
Hence external moment = 1000 x 7 = 7000
Moment of Shear Flow about Intersection of in. lb.
Centerline Axes
Therefore the total moment about the
For equilibrium in the plane of the cross assumed moment center 0' =
section at station (0), the summation of the
A20.15
;1;'
TABLE A20.10
MOMENTS DUE TO IN PLANE COMPONENTS

OF STRINGER LOADS
1 2 3 4 5 6 7 8
Py = Mo = P z = Mo =
Stringer P x ~ P~ dz Px~
No. (lbs. ) dx xdx P y z' dx dx P z y'
a -9312 .0026 248 2600 -.0233 217 -2600

b -2239 .0188 42 798 -.0421 94 - 796
c -2304 0 0 0 -.050 115 0
d -1719 -.0188 32 - 608 -.0421 72 610
e -6506 -.0266 173 -1820 -.0233 152 1820
f 9216 -.0266 245 -2570 .0233 214 2560
g 3673 -.0188 69 -1310 .0421 154 1307
h 3809 0 0 0 .050 190 0
i 2634 .0188 49 930 .0421 111 - 941
j 1246 .0266 33 346 .0233 29 - 348
Sum -1634 1612
NOTES: 115
94 72
Col. (2) from Table A20. 7 c
+r IT
Col. (3) equals the slope 217 2,3 Fig. A20.11
of stringers
in y and z ~48
Ft 1?l
18.98"
directions. Shear flow distribution.
(see Fig. A20. 9)
H2'-'-l10.5" ~
Col. (5) Values of z'
y' ~-~-- where large concentrated loads are applied can
and (8) and y' from o Fig.b W be determined by the procedure given in Articles
Table A20. 7. 33~ j : I M 18 to 20 of Chapter A19. A more rigorous
Fig. b shows ~-I f ~45
29 analysis can be made by the application of the
the P y and Pz .00 I g 214 basic theory as ~iven in Chapter A8.
components from 49 1 h 69
Cols. (4) and (7). 154
111190 The problem of shell stresses due to in-
Total moment about 0' = Looking Toward Sta.150 ternal pressures is presented in Chapter A16.
-1634 + 1612 = -22"# The strength design of the fUselage skin in-
volves a question of combined stresses. The
broad problem of the strength design of struc-
92670 due to shear flow q tural elements and their connections under all
-22 due to in plane components of types of stress conditions is covered in Volume
stringers. II.
7000 due to the external loads.
A20.9 Problems.
Total= 99648 in. lb.
(1)
Therefore for equilibrium a moment of
-99646 is required which can be provided by a
constant shear flow q ~ around the cell, hence
I
_ T _ -99648 = -52 Ib./in. (957 =
20"
q ~ - 2A - 2 x 957
-~"-----
Skin .035
enclosed area of cell)
This value of q is entered in column 8 of

Table A20.9. The resulting shear flow in any
web portion qr equals the algebraic sum of q
and q~. (See Col. 9, Table A20.9). Fig. I
A20.11 shows the results in graphical form.
Fig. A20.12
A20.8 Discontinuities - Shear Lag - Pressurization
Stresses - Combined Stresses.
Fig. A20.12 shows the cross-section of a
A practical fuselage has many cut-outs. circular fuselage. All stringers have same
The approximate effect of these discontinuities area, namely 0.12 sq. in. Skin thickness is
as well as the shear lag effect at sections .035 inches. Stringers are 1 inch in depth.
A20.l6 FUSELAGE STRESS ANALYSIS
All material is aluminum alloy. E 10,500,000 =

psi. The ultimate compressive strength of
stringer plus its effective skin is 35000 pSi.~
For effective sheet width use w 1.9t (E/OST)s, =
For buckling strength of curved panels use
ocr = .3 Et/r. Determine the Ultimate bending
moment that the fuselage section will develop
for bending about horizontal neutral axis. Use
linear stress distribution. Follow procedure
as given in example problem in Art. A20.4.
(2)
Fig. A20.13 shows

4 spaces @7"=28" the cross-section of a
S2
I- S~ S~ S~
-I S rectangular fuselage.
2 The dots represent
·~.032 skin stringer locations.
S" S" Three types of string-
S" $v..\~ S"
ers are used, namely,
j..-.()~ .04'::- S~, S 2 and 8". Fig.
A20.14 shows the
S" S" Ultimate compressive
S:. s" stress-strain curve
l032 for each of the three
S2 S2 stringer types and
s... S~ S~ also the tension
Fig.A20.l3 stress-strain curve
of the material.
Determine the Ultimate bending resistance (4) Same as Problem (3) but change area of
of the fuselage section about the horizontal stringer no. (2) to 0.3 sq. in., thus making an
neutral axis if the maximum unit compressive unsymmetrical section.
strain is limited to .008. Refer to Art. A20.5
for method of solution. A20. 10 Secondary Stresses in Fuselage Stringers and Rings.
ADDITIONAL DATA. Area stringer 8~ = .12 sq.in.; The stresses that are found in the
82 =.25 sq. in.; S" .08 sq.in. E 10,500,000 = = stringers or longerons of a typical fuselage by
psi •. use of the modified beam theory or by the more
rigorous theory of Chapter AS, are referred to
(3) Fig. A20.15 shows a tapered circular as primary stresses. Because of the necessity
fuselage with 8 stringers. The area of each of weight saving, most fuselage structures are
stringer is 0.1 sq.in. Assume stringers develop designed to permit skin buckling, which means
entire bending resistance. Find the axial load that shear loads in the skin are carried by
in stringers at station (110) due to Pz and Px diagonal semi-tension field action. This
loads at station (0). Also find shear flow diagonal tension in the skin panels produces
system at station 110 using ~P method. Use additional stresses in the stringers and also
properties at station (90) IN OBTAINING AVERAGE in the fuselage rings. These resulting stresses
SHEAR FLOWS. are referred to as secondary stresses and must
be properly added to the primary stresses in
P z = 2300# Pz
1
r============:if
I-----------i 7
9 fl
n"~"
Px=500#
the strength design of the individual stringer
or ring. Chapter Cll covers the SUbject of
these secondary stresses due to diagonal semi-
tension field action in skin panels. It is
suggested to the stUdent that after studying
l:==~======::::::::::~
L Sta.O 6 5 ~ Chapters A19 and A20, that Chapters C10 and Cll
5 ta.110 110,, .1 be referred to in order to obtain a complete
stress picture for skin covered structures.
Fig. A20.l5
CHAPTER A21
LOADS AND STRESSES ON RIBS AND FRAMES
A21.1 Introduction. For aerodynamic reasons the loads into the fuselage shell such as those
wing cantou, in the chord direction must be from landing gear reactions, wing reactions,
maintained without appreciable rJistortion. tail reactions, pONer Dlant reactions, etc.
Unless the wing skin is quite thick, spanwise The dead wei";ht of all the payload and fixed
stringers must be attached to the skin in order eqUipment inside the fuselan'e must be carried
to increase the bending efficiency of the wing. to frames by other structure such as the
Therefore to hold the skin-stringer wing surface fuselage floor system and then transmitted to
to contour shape and also to limit the length the fuselage shell structure. Since the dead
of strinc\ers to an efficient coltunn compressive weight must be multiplied by the design accel-
stren,a;th, internal support or brace units are eration factors, these internal loads become
reqUired. These structural units are referred qUite lanse in mo!:nitude.
to as wing ribs. The ribs also have another
major purpose, nam81y, to act as a transfer or Another important purpose or action of ribs
distribution unit. All the loads apnlied to and frames is to redistribute the shear at dis-
th8 wina; are react8d at the wing supporting continuities and Dractical Wings and fuselages
points, thus these applied loads must be trans- contain many cut-outs and openings and thus
ferred into the wing cellular c'tructure com- discontinuities in the basic structural layout.
posed of skin, stringers, snars, etc., anrJ then
reacted at the win': support Doints. The 3TJTJlied A21. 2 Types of Wing Rib Construction.
loads may be only the rJistributed surface air-
loads which requi re relativ!31y light int8rnal FiGS. A21.1 to 6 illustrate the COffinon
ribs to pr'ovids this carry through or transfer types of winG construction. FiG' 1 illustrates
reqUirement, to rather rugged or heavy ribs a sheet metal channel for a leading edge 3
which must absorb and transmit large concen-
trated applied loads such as those from landing
gear reactions, power GIant reactions and fuse-
lage reactions. In between these two extremes +
of applied load mao;nitudes are such loads as
O~O
reactions at supporting points for ailerons,
flaps, leading edge high lift units and the
many internal dead wei,:ht loads such as fuel
and military armament and other insta1l8tions.
Thus ribs can vary from a very light structure Web
which serves primarily as a former to a h3avy
structure which must receive and transfer loads Fig. A2L 1
involving thousands of pounds.
Since the airplane control surfaces (verti-

cal and horizontal stabilizer, etc.) are nothing
more than small size Wings, internal ribs are
likewise needed in these structures.
The skin-stringer construction which forms

the shell of the fusela~e likewise needs in-
ternal forming units to hold the fuselage
cross-section to contour shape, to limit the
coluJnn length of the stringers and to act as Fig. A21. 2
transfer agents of internal and externally
applied loads. Since a fuselage must usually
have clear internal space to house the payload
such as passengers in a commercial transport,
these internal fuselage units which are usually
referred to as frames are of the open or ring
type. Fuselage frames vary in size and strength
from very light former type to rugged heavy
types which must transfer large concentrated
Fig. A21. 3
A21.1
A21. 2 LOADS AND STRESSES ON RIBS AND FRAMES
Wing Skin
Ribs in 3 Spar Wing.
Fig. A21. 4 Fig. A21. 5
Fig. A21. 6
stringer, single spar, single cell wing were used to transmit external loads into the
structure. The rib is riveted, or spot-welded, wing cellular beam structure. Concentrated
or glued to the skin along it boundary. Fig. external loads must be distributed to the rib
2 shows the same leading edge cell but with before the rib can transfer the load to the
spanwise corruga~ions on the top skin and wing beam structure. In other words, a con-
stringers on the bottom. On the top the rib centrated load applied directly to the edge of
flange rests below the corrugations, whereas a thin sheet would cause sheet to buckle or
the stringers on the bottom pass through cut- cripple under the localized stress. Thus a
outs in the rib. Fig. 3 illustrates the gen- structural element usually called a web stiff-
eral type of sheet metal rib that can be ener or a web flan~e is fastened to the web and
qUickly made by use of large presses and rubber the concentrated load goes into the stiffener
dies. Figs. 4 and 5 illustrate rib types for which in turn transfers the load to the web.
middle portion of wing section. The rib To ~et the load into the stiffener usually re-
flanges may rest below stringers or be notched quires an end fitting. In general the distri-
for allowing stringers to pass through. Ribs buted air loads on the wing surface are usually
that are sUbjected to considerable torsional of such magnitude that the loads can be distri-
forces in the plane of the rib should have some buted to rib web by direct bearing of flange
shear ties to the skin. For ribs that rest normal to edge of rib web without causing local
below~tringers this shear tie can be made by buckling, thus stiffeners are usually not
a few sheet metal angle clips as illustrated in needed to transfer air pressures to wing ribs.
Fig. 5. Fig. A21.7 shows an artist's drawing
of the wing structure of the Beechcraft Bonanza EXAMPLE PROBLEM ILLUSTRATING TRANSFER OF CON-
commercial airplane. It should be noticed that CENTRATED LOAD TO SHEET PANEL.
various types and shapes of ribs and formers
are required in airplane design. Photographs Fig. A21.8 shows a cantilever beam com-
A21.1 to 3 illustrate typical rib construction posed of 2 flanges and a web. A concentrated
in various type aircraft, both large and small. load of 1000 lb. is applied at point (A) in the
Since ribs compose an appreciable part of the direction shown. Another concentrated load of
wing structural weight, it is important that 1000 lb. is applied at point (E) as shown.
they be made as light as safety permits and
also be efficient relative to cost of fabrica- To distribute the load of 1000 lb. at (A),
tion and assembly. Rib development and design a horizontal stiffener (AB) and a vertical
involves considerable static testing to verify stiffener (CAD) are added as shown. A fitting
and assist the theoretical analysis and design. would be required at (A) which would be attached
to both stiffeners. The horizontal component
A21. 3 Distribution of Concentrated Loads to Thin of the 1000 lb. load which equals 800 lb. is
Sheet Panels. taken by the stiffener (AB) and the vertical
component which equals 600 lb. is taken by the
In Art. A21.1 it was brought out that ribs
A21. 3
/L~~7}.I._. __././-----.. ----

I If ~L:_._-~--~-~~ .
c.. . //~---~'
Fig. A21. 7 General Structural Details of Wing for Beechcraft "Bonanza" Commercial Airplane.
vertical stiffener CD. The vertical load at E Figs. A21.9 and A21.10 show free bOdies of that
would be transferred to stiffener EF through portion inclUding web panels (1) and (2) and
fitting at E. The problem is to find the shear stiffeners CAD and AB and the external load at
flows in the web panels, the stiffener loads (A) . In Fig. A2l. 9 the shear flows q ~ and q 2
and the beam flange loads.
~ 10"--t-- 10"
G
-+--
flange
10" ----t- 10"-----1
F I
r q~
tr---"'=-~"'=-nD
I~- -D
q~
/t=====;~==:::::::::=;;===t===;;;:::====;] D 9" q t (1) q~ t(1)

I ~t t
9" (1) -+- B ~800 l
B
A
Web 3" qn ~ (2) q2! (2) 600 lb.
(3)
B - .-800 lb.
<[ T """E---- C ~x E-- - - C
~1~~-J
3" (2) 3 q2
C ( 1000 lb.
E 600 lb. Fig. A21. 10
flange
Fig. A21. 9
Fig. A21. 8 1000 lb.
on the top and bottom edges respectively have

SOLUTION: It will be assumed that the beam been assumed with the sense as shown. Taking
flanges develop the entire resistance to beam moments about point E,
bending moments, thus shear flow is constant
on a web panel. mE = 800x3-12xIOq~ = 0, whence q~ =
20 Ib./in.
The shear flows on web panels (1) and (2)
will be computed treating each comDonent of L:Fx = 800 - 20 x 10 - 10q2 = 0, whence q2 =
the 1000 lb. load as acting separately and the 60 Ib./in.
results added to give the final shear flow.
A21.4 LOADS AND STRESSES ON RIBS AND FRAMES
PHOTO. A21. 1 Type of Wing Ribs Used in Cessna 180 Model Airplane, a 4 Place Commercial Airplane .
• • • ,6" • t
\
1 '
PHOTO. A21. 2 Rib Type Used in Outer Panel-Fuel Tank Section- of Douglas DC-8 Commercial Jet Airliner.
PHOTO. A21. 3 Rib Construction and Arrangement in High Speed, Swept Wing, Fighter Type of Aircraft.
North American Aviation - Navy Fury - Jet Airplane.
A21. 5
Referring to Fig. A21.10, Solving equations (1) and (2) gives,
L:I1E = 600 x 10 - 12 x 10q 1. = 0, whence q 1. = q1. =

70 Ib./in., q2 = 10 Ib.lin., which
50 lb. in. checks first solution.
L:F): = -50 x 10 + 10q2 = 0, whence q 2 = The shear flow q3 in web panel (3) is ob-
50 Ib./in. tained by considering stiffener EBF as a free
body, see Fig. A21.16.
Combining the two shear flows for the two loads,
ZFy = 12q3 + 10 x 3 - 9 x 70
q:l. = 20 + 50 = 70 Ib./in.
- 1000 = 0
50 = 10 Ib./in.
whence, q3 = 133.33 Ib./in.
Fig. A21.11 shows the results. Fig.
A21.12 shows stiffener AB as a free body, and The shear flow q3
Fig. A21.13 the axial load diav,ram on stiffener could also be found by
AB, which comes directly from Fig. A21.11 by treating entire beam to
starting at one end and adding the shear flows. right of section through
panel (3). For this free
body,
r--q-1.-=~
770~~D
800#(tension) Fig. A21.16 = -600 - 1000 + 12q 3 = 0
L:F y
,e) B~ Fig. A21. 13
whence, q3 = 133.33
q, =70
A 800 B' =-___.800
- -~A Fig. A21.17 shows diagram of axial load in
600 q2 = 10 stiffener EF as determined from Fig. A21.16 by
C starting at one end and adding up the forces to
Fig. A21.11 Fig. A21. 12 any section.
Fig. A21.14 shows a free body of the After the web shear
F flows have been determined
vertical stiffener CAD, and FiG' A21.15 the ~ Fig. A21. 17
axial load diagram for the stiffener. the axial loads in the
beam flanges follow as
B~lb. the algebraic sum of the
shear flows. Fig. A21.18
E 1000 lb. shows the shear flows
(tension)
along each beam flange
as previously found. The
upper and lower beam flange loads are indicated
by the diagrams adjacent to each flange.
Fig. A21. 15
4700 r-----
The shear flows q1. and q2 could of course

be determined using both components of forces
at (A) acting simultaneously. For example,
./
/'
Tension
- -
-
700 lb.
.-
13i33- 133.33 133-:33- 70
consider free body in Fig. A21.15a. /
/
./ -800
f- 10"-1 / 133.33
F_..h.-_ D ..!31:~ ~ 1~3.:Jt. - -- -10- - 600
1000
I 100# tenslOn
I 9" ........----===
~;g.A21.18
qq
Fig. A21. 15a
~y
~~I
I~B===~IFA.:...-800#
600 lb.
-+
3" 3900# compression
E- - C -L
q2
In this example problem the applied extern-
ZF x = 800 - 10q, (1) al load at point (A) was acting in the plane of
the beam web, thus two stiffeners were suffi-
L:F y = -600 + 9q:l. 3q" =0 - (2) cient to take care of its two components. Often
loads ~re applied which have three rectangular

components. In this case, the structure should
be arranged so that line of action of applied
force acts at intersection of two webs as
illustrated in Fig. A21.19 where a load P is
applied at point (0) and its components Pz ' Py
Fig. A21. 20
z
~y
x
With q known the shear and bending moment

Fig. A21. 19 at various sections along the rib can be deter-
mined. For example, consider the section at B-B
in Fig. A21.20. Fig. A21.21 shows a free body
and Px are distributed to the web panels by of the portion forward of this section.
using three stiffeners S~, S2 and S3 inter-
secting at (0). The bending moment at section B-B equals:
In cases where a load must be applied MB =

2qA~ where A~ is the area of the
normal to the web panel, the stiffener must be shaded portion.
designed strong enough or transfer the load in
bending to adjacent webs. Let Fx equal the horizontal component of
the flange load at this section.
In this chapter, the webs ar~ assumed to
resist pure shear along their boundaries. In
most practical thin web structures, the webs
will buckle under the compressive stresses due
to shear stresses and thus produce tensile
f~eld stresses in addition to the shear
stresses. The sUbject of tension field beams Fig. A21. 21
is discussed in detail in Volume II. In gen-
eral the additional stresses due to tension
field action can be superimposed on those
found for the non-buckling case as explained
in this chapter.
STRESSES IN WING RIBS

A21. 4 Rib for Single Cell 2 Flange Beam.
Fig. A21.20 illustrates a rib in a 2-

flange single cell leading edge type of beam. The true upper flange load Fu = Fx/cOS Q~
Assume that the air-load on the trailing edge and the lower flange load equals FL Fx/cos Q2. =
portion (not shown in the figure) produces a
couple reaction P and a shear reaction R as The vertical shear on the rib web at B-B
shown. These loads are distributed to the cell equals the vertical component of the shear flow
walls by the rib which is fastened continuously q minus the vertical components of the flange
to the cell walls. Let q = shear flow per loads. Hence
inch on rib perimeter which is necessary to
hold rib in equilibrium under the given loads V
web =q a - Fx tan
P and R.
2 q A~
=q a - a
Taking moments about some point such as (1)
of all forces in the plane of the rib: Illustrative Problem
ZM~ = -Ph + 2Aq =0 The rib in the leading edge portion of the
wing as illustrated in Fig. A21.22 will be
hence analyzed.
q = Ph/2A. (A = enclosed area of cell) A distributed external load as shown will

be assumed.
A21. 7
Solution:
The total air load aft of beam = 8 x 40/2 =

160 lb. The arm to its c.g. location from the
beam equals 40/3 =
13.33". Hence the reactions
at the beam flange points Bue to the loads on Area of Shaded
the trailing edge portion equals: Portion = 15.4 sq. in.
P = 160 x 13.33/10 = 213.2 lb. (See Fig.

Fig. A21. 24
A21.23)
Shear reaction Vr = 160 lb.
A21. 5 Stresses in Rib for 3 Stringer Single

w = 8 lb. /in. Cell Beam.
--- ---- Fig. A21.25 shows a rib that fits into a

- -/-
----
--
single cell beam with 3 stringers labeled (a),
----....,
(b) and (c). An external load is applied at
---- ---- ---- ---- point (a) whose components are 5000 and 3000
----- -----
- --/ Ibs. as shown. Additional reactions from a
------
/' trailinG edGe rib are shown at points (b) and
10" ~
(c). A vertical stiffener ad is necessary to
~=====~L _- --- -------- distribute the load of 5000 lb. at (a). The
-+--- 40"
following values will be determined: -
(1) Rib web shear loads on each side of stiff-

Fig. A21. 22
ener ad.
I~ !J-IJlIJJ III ~ll

w = 8 lb. /in.
(2) Rib flange load at section ad.
p. 213 2lb (3) Rib flange and web load at section just
to left of line bc.
~\~
~
Area
A = 139.3 sq. in.
I
,
l~r 10" Fig. A21. 25
! 160
~===_====.=::::::::::.==-(
1) P = 213. 2 lb.
Fig. A21. 23
Let q be the constant flow reaction of the

cell skin on the rib perimeter which is neces-
sary to hold the rib in equilibrium under the
applied air loads.
Take moments about some point such as the

lower flange (1).
SOLUTION. It will be assumed that the 3 string-
ZM = -213.2xlO+8x15x7.5+2x139.3 q ers develop the entire wing beam bending resist-
=0 ance, thus the wing shear flow is constant be-
tween the stringers. The wing rib is riveted
whence, q = 1232/278.6 = 4.42 Ib./in. to the wing skin and thus the edge forces on
the rib boundary will be assumed to be the same
With the applied forces on the rib known, as the shear flow distribution. In other words,
the shears and bending moments at various the three shear flows Gadc' qba and qcb hold the
sections as desired can be calculated. For external loads in equilibrium. The sense of
example, consider a section B-B, 2.5" from the these 3 unknown shear flows will be assumed as
leading edge. Fig. A21.24. shown in Fig. A21.25.
Bending moment at section B-B = To find Gadc' take moments about point (b)
8x2.5x1.25+4.42x2x15.4 = 161 in.lb.
2Mb = -2(A~ + As) Qadc + 5000 x 15 - 500 To find web shear

x 11.5 0 = d T'
~d take ZF z 0 =
= -2(60 + 160) Qadc + 75000 - 5750 =0 2158~b-t ZF z = 5000-9.3x74.6
= 157.3
74.611,
.I~d
9.3"
- 9.3 q~d =0
whence, Qadc
assumed.
lb./in. with sense as
2034-4 1_
a t-
ell
3000 Whence, q~d = 463 lb./
in.
To find qcb take ZF z =0 5000
Fig. A21. 27
ZF z =5000 + 200 - 157.3 x 11.5 - 11.5
To find flange
load C' take ZFx 0, =
qcb =0 considering joint (a) as a free body,
Whence, qcb = 295 lb./in. ZFx = 2034 + 3000 - C' = 0, whence, C' 5034 lb. =
At joint (d) T' obviously equals 2158 lb. The
To find qba take ZFx =0 stiffener ad carries a compressive load of 5000
lb. at its (a) end and decreases uniformly by
ZF x = -500 + 3000 + 500 - 157.3 x 15 the amount equal to the two shear flows or
-15 qba =0 463 + 74.6 =
537.6 lb./in.
The results obtained by considering Fig.

whence, qba = 42.7 lb./in. A21.27 could also be obtained by treating the
With these supporting skin forces on the entire rib portion to the left of a section
rib boundary, the rib is now in equilibrium just to right of stiffener ad, as shown in Fig.
and thus the web shears and flange loads can A21.28.
be determined. Consider as a free body that
portion of the rib just to the left of the To find rib flange load T' take moments
about point (a).
stiffener ad centerline as shown in Fig. A21.26.
ZMa, = (16/17) T' x9.3-157.3x60x2 = 0
T
Whence, T' = 2158 lb.
~T'
(1-
t I 9.3"
Fig. A21. 26 t~d I
~::::::;:==~ ----l...- e '
a./<3000
To find flange load T take moments about 15000
point (a), Fig. A21. 28
ZMa = (16/17)T x 9.3 - 157.3 x 60 x 2 =0

To find flange load C' take ZFx =0
whence, T =2158 lb.
ZFx = -C' +3000+2158(16/17) =0
To find flange load C take ZFx 0 =
ZFx = =
2158 (16/17) - C 0, whence C 2034 lb. = Whence, C' = 5034 lb.
To find web shear Qad take ZF z = 0 To find q~d take ZF z =0
ZFz = 2158 (6/17) - 157.3 x 9.3 + 9.3 Qad =0 ZF z = 2158(6/17) - 157.3 x 9.3 + 5000 - 9.3
whence, qad =
74.6.1b./in. q~d =0
To find the shear in the web just to right whence, ~d = 463 lb./in.
of stiffener ad, consider the free body formed
by cutting through the rib on each side of the The above values are the same as previously
stiffener attachment line as shown in Fig. obtained.
A21.27. The forces as found above are shown
on this free body. The rib flange loads and web shear will be
calculated for a section just to left of line
A21 9
cb. Fig. A21.29 shows the free body for the

rib to left of this section. 6000
Fig. A21. 30
Fig.A21. 29
SOLUTION:
To find flange load C take moments about The total shear load on the wing in the Z
point (b). direction equals Vz = -6000 - 5000 + 2000 =
-9000 lb. and Vx = -8500 + 7500 - 4000 + 4500 =
mb = -157.3x2 (160+60) + 5000 x 15 -500 lb.
-11.5 C = 0 The boundary forces on the rib will be
whence, C =500 lb. equal to the shear flow force system on the
cell walls due to the given external force
To find flange load T take ~Fx =0 system.
From Chapter A14, page A14.8, equation (14)

~F x = 3000 - 157.3 x 15 - 15 x 42.7 - 500 + T =0 the expression for shear flow is,
whence, T = 500 lb.
qy = -(K3VX-K~Vz) ~xA - (K 2 VZ -K~Vx)
To find q;b take ~Fz =0 ~zA - - - - - - - - - - - - - - - (1)
~Fz = 5000 -157.3 x 11.5 - 11.5 q~b =0 The constants K depend on the section
properties of the wing cross-section. Table
whence, qbb = 278 Ib./in. A21.1 gives the calculation of the moment of
inertia and product of inertia about centroidal
The above results could have been obtained Z and X axes. In this example the 4 stringers
with less numerical work by considering the a, b, c and d have been considered as the entire
forces to right of section cb in Fig. A21.29. effective material in resisting wing bending
stresses.
A21. 6 Stress Analysis of Rib for Single Cell
Multiple stringer Wing. TABLE A21. 1
1 2 3 4 5 6 7 8 9 10 11
When there are more than three spanwise
stringers in a wing, there are four or more Flange Area
No. A Z' X' AZ' AZ,2 AX' AX,2 AX' z' Z=Z'- Z X=Xl_X
panels in the cell walls, thus the reactions of
the cell walls upon the rib boundary cannot be a 2.00 5.50 0 11.00 60.5 0 0 0 6.36 -11.8
found by statics as was possible in the 3 b 1. 25 3.55 30 4.43 15.72 37.50 1125 132.9 4.41 18.2
stringer case of the previous example problem.
c 1. 15 -8.40 30 - 9.66 82.20 34.50 1035 -289.8 -7.54 18.2
Fig. A21.30 illustrates a wing section con- d 1. 70 -6.50 o -11. 04 71. 90 0 0 0 -5.64 -11.8
sisting of four spanwise flange members. The ~ 6.10 - 5.27 230.3 72.00 2160 -156.9
concentrated loads acting at the four corners
of the box might be representative of reactions
from the engine mount or nacelle structure and z =~Z'/~ =-5.27/6.10 = -.865"
x- = ~'/~ = 72.0/6.10 = n.8"
the reactions from a rib which supports the
wing flap. These loads must be distributed
into the walls of the wing box beam which neces-
sitates a rib. Before the rib can be designed, Centroidal x and z moments of inertia:
the bending and shear forces on the rib must be
determined. The calculations which follow
Ix =230.3 - 6.10 x .865 = 225.8 2
illustrate a method of procedure. I z =2160 - 6.10 x 11.8 = 1310 2
I xz = -156.9 - 610 x -.865 x 11.8 =-94.7

With the wing section properties known, ZM c .g. = -llOOO x ll.8 - 8500 x 6.36 - 7500 x
the constants K can be calculated. 5.65 - 4000 x 4.41- 4500 x 7.54 - 2000 x 18.2 - 444 x
K1 = 1xz/(1x1 z - 1xz 2
) 2x90-748x2xl08.7-461x2x99 =-648400 in.lb.
= -94.7/(225.8x1310-94.7 = 2
)
F0r equilibrium ZMc • g • must equal zero,
-94.7/286700 = -.00033 therefore a constant flow shear ql acting around
the rib perimeter is necessary which will pro-
K2 = 1z/286700 = 1310/286700 = .00456 duce a moment of 648400 in. lb.
K3 = 1x/286700 = 225.8/286700 = .000786 M
ql = 2A = 2648400
x 368.5 = 880 lb./in.
SUbstituting in equation (1),
(Note: 368.5 = total area of cell)
qy = - G000786 (-500) - (-.00033) (-90008
ZxA - [00456 (-9000) - (-.00033)(-5001 Adding this shear flow to that of Fig.
A21.31, the resulting force system of Fig.
ZzA A21.32 is obtained. The reactions of the beam
cell walls on the rib have now been determined
whence, and the bending moments and shears on the rib
can now be calculated.
qy = 3.363 ZxA + 41.205 ZzA - (2)
Since the shear flow at any point on the 6000

cell walls is unknown, it will be assumed zero 8500 ~ _ _q = 436
on web ad, or imagine the web is cut as shown -t a - ~ - --±'OOO
in Fig. A21.31. The static shear flows can
now be found. gt
ex>
b I1_.
..
M
Gab = 3.363 (-ll.8) (2.0) + 41.205 x 6.36 ~f Fig. A21. 32 I:

7500 td I 0'
x 2 = 444 lb./in. -t~-=-~~-==-==-:=-=--==~Cl
~---- -~~~t_ 4500
q = 419
qbc = 444 + 3.363 x 18.2 x 1.25 + 41.205 5000 2000
x 4.41 x 1.25 = 748 lb ./in.
Fig. A21. 32
qcd = 748 + 3.363 x 18.2 x 1.15 + 41.205 To illustrate, consider the rib section
(-7.54)(1.15) = 461 lb./in. B-B which passes through the c.g. of the beam
section. Fig. A21.33 shows a free body of the
These shear flows are plotted on Fig. bulkhead portion to the left of section B-B.
A21.31. Refer to Chapter A14 regarding sense
of shear flows.
B
rooo Fig. A21. 31 - ' - - Fu
850() a ~= 444
~ ~ , I -;-; , - ; - - - b 00
d6 = -'~~~rea=900" ~/I' ~ 4 41
~. ~rea -~.~~ t~
I - 70.8 1 / ~rea = 108. 7 lex> r
5.64 - /IL- 7,/// ~ 754
- ~Area = 99 / til'
7500 d-~_ I I 1\ ~
~ -t=-
--q :46lib7in.-
C
4liOO -t
5000 2000 Fig. A21. 33 Fig.A21. 34
l--- 11.8 18.2 -----l
I Total Cell
Area = 368. 5°" Moments at section B-B will be referred to
the point (0):
The moments of the forces in the plane of ZMO = -llOOO x 11. 8 - 8500 x 6.36 - 7500 x 5.64 + 436
the rib will now be calculated: x 2 x 36 + 880 x 2 x 70.8 + 419 x 2 x 38.3 =
Taking moments about the c.g. of the beam -38200 in. lb.
cross section (See Fig. A21.31):
A21. 11
The resultant external shear force along the boundary forces on a rib located adjacent
the section B-B equals the summation of the z to a cut-out is to find the applied shear flows
components of all the forces. in the wing on two sections, one on each side
of the rib. Then the algebraic sum of these
v = ZFz = -11000 + 12 x 880 - 436 x 0.36 + 419 two shear flows will give the rib boundary
forces. With the boundary forces known the rib
x 0.96 = -195 lb.
web and flange stresses can be found as pre-
The resultant load normal to the section viously illustrated. The procedure can best
B-B equals the summation of the force compon- be illustrated by example problems.
ents in the x direction. A21. 8 Example Problem. Wing with Cut-Out
Subjected to Torsion.
H = ZF x = -8500 + 7500 + (436 - 419) 11.8
Fig. A21.35 shows a rectangular single
= -800 lb. cell wing beam with four stringers or flanges
located at the four corners. The upper surface
Fig. A21.34 shows these resultant forces
skin is discontinued in the center bay (2).
referred to point (0) of the cross-section. If
The wing is SUbjected to a torsional moment of
we assume that the rib flanges develoD the
entire resistance to normal stresses·, we can
find flange loads by simple statics. STA.O
cS"
~ ~
T: jRlb I' 1 •
To find upper flange load Fu take moments ~'rt <Ii 'I
I
1 _
about lower flange point. 'I II
STA. 2~ 1~1j /,
= 12.6 =0
~~f:/1
ZM Fu -38200-800x6.6
Rib
whence, Fu = 3443 lb. tension /""
"'~ .~';"V 1-------:;;7';t--f-::!L Upper skin
<no ":) ~ surface re-
~ ~ moved in
To find FL use ZF x = 0
STA.5¥ 4,,,0. /, 4000 lb. Bay (2)
ZF x = 3443 - 800 - FL = 0, whence FL = B ~ ! I( Rib (A) I, I "
2643 lb. compression. ~
:. <Ii If II
1 ;-- - - - / , -I" 4000 lb.
STA~O I I; I, II
The shear flow on web equals V/12.6 = .
8 1/ Rib
~
\ 80000 in. lb. /
j
Fig. A21. 35
195/12.6 = 15.5 Ib./in. This result neglects
effect of flanges not being normal to section
B-B, which inclination is negligible in this
case.
80000 in. lb. at Station (70) and a couple force
If the entire cross-section of rib is at Station (50) as shown in Fig. A21.35. The
effective in bending, then the web thickness problem will be to determine the applied forces
and flange sizes of the rib would be needed on rib (A).
to obtain the section momGnt of inertIa which
is necessary in the beam equation for bending SOLUTION:
stresses. The forces at (0) would then be
referred to neutral axis of section before The applied shear flow on the cell walls
bending and shear stresses on the rib section will be found for two cross-sections of the
could be calculated. Wing, one on each side of rib (A).
To obtain a complete picture of the web In bay (1) the torsional moment M is 80000
and flange forces, several sections along the in.lb. The applied shear flow on a cross-
rib span should be analyzed as illustrated for section of the wing in bay (1) thus equals,
section B-B.
M
q -_ 2A = 80000 = 100 Ib./in.
A21. 7 Rib Loads Due to Discontinuities in Wing 2 x 10 x 40
Skin Covering.
This shear flow system is shown on Fig.
As referred to before, ribs in addition to A21.36 which is a free body of rib (A). In
transmitting external loads to wing cell bay (2), since the top skin is removed, the
structure are also a means of re-distributing torsional moment must be taken by the front and
the shear forces at a discontinuity, the most
rear vertical webs, since any shear flow in the
common discontinuity being a cut-out in one or
bottom skin could not be balanced.
more of the webs or walls of the wing beam
cross section. The usual procedure in finding The torsional moment in bay (2) is,
Bay (2)
Jf..:..:.=-----""-----7'
Rib A
t__ 4000# Bay (1) --1--- - - - - L = 30"----+1+_ Bay (3)
t
= 100 t
,+- - --;
'/- - ..9. -
100
-
- -, :
-llJOll
t q'=300 lO" P~i 3000
h =T Fig. a
q t/-.~ ..-+.-+ - ---. 4000# "1 P-'?====-- r:::-P
Deflected Front Bea;;Cmm======d~
u. _V=3000
P
q = 100
Fig. A21. 36 VL~~ Fig.b

M=2 ~ Beam Bending Moment Diagram
11 = 80000 + 4000 x 10 = 120000 in.lb. V=3000"---- _

I_ _--=B=--e--=a=--m---=S:..:.h_e.:..:.ar---=D=--i--=a-".gr--=a--=m _ Fig. c
The total shear load on eaco vertical web
thus equals 120000/40 = 3000 lb., which gives q = 300
a shear flow q' = 3000/10 = 300 Ib./in. on each
web. This applied shear is shown on the free Fig. d
body of rib (A) in Fig. A21.36. On the left q = 300
end of the rib a shear flow of 100 is acting Fig. A21. 38
Fig. A21. 37
--='!'-=="'-q----='="--l~~""O~-""__=_""_'-=_ _ _ 4000
flange loads at the beam ends are P = 45000/10
= 4500 lb o (See Figo a).
q = 200 11 _ _ _ _ _ _ _ Ii q = 200
-4000
The deflection of the rear beam would be
q = 100 the reverse of Fig. a, and thus all forces
would also be reversed.
up and on the other side a shear flow of 300 Fig. A21.39 shows bay (1) of the wing as a
is acting down, thus the rib web must take the free body acted upon by the flange loads due to
difference or 200 acting down. On the right end bending of the beams in bay (2). These internal
of the rib the load on the rib web is 200 Ib./ flange forces from bay (2) must be held in
in. up. The loads on the top and bottom equilibrium by the internal stresses in the ad-
flanges of the rib is obviously 100 Ib./in. jacent wing structure of bay (1).
Fig. A21.37 shows the loads applied to the rib
boundary when the torsion in bay (1) and the r§ <§'
external couple force is transferred to the ::J ::J
r§
cross-section of bay (2).
STA.50 - - , -
I 1/
I
I /~
,cRIb (A)
,~
I
/
::J
ADDITIONAL EFFECTS DUE TO DIFFERENTIAL BENDING y" II~ 1/
OF BEAMS IN BAY (2). $ J] / ;./~ /7
t:I:; I I ,I 11/ h
The torsion in bay (1) and the external ST A. 70 - L ,: /
I
.f! /, / Bay (1)
couple force are thrown off as couple force on II h
the front and rear beams of middle bay (2), ~ ~t
with the total shear load on each beam being

Fig. A21. 39
3000 lb. as previously calculated. These beam
shear loads must be transmitted to bay (3) and
thus cause bending of the beams in bay (2). According to the well known principle of
Since each beam is attached to relative rigid mechanics formulated by Saint Vennat, the
box structures at each end, namely bays (1) stresses resulting from such an internal force
and (3), the beams tend to bend with no rotation system will be negligible at a distance from
of their ends. If we neglect the deflections the forces. This distance in case of a cut-out
of these end box structures, we can assume that is usually assumed as approximately equal to
the beams bend with no rotation of their ends the width of the cut-out, or in general to the
or each beam is fixed ended. Fig. A21.38a width of the adjacent wing bay. Thus in Fig.
illustrates the deflection of the front beam in A21.39 the flange loads of 4500 Ibs. each are
bay (2) under the assumption of no end rotation. assumed to be dissipated at a uniform rate for
The beam elastic curve has a point of inflection a distance of 20 inches. Thus the shear flow
at the span midpoint. Figs. 38b, c show the created by each stringer load which equals the
beams bending moment and shear diagrams. change in axial load per inch in the stringer
in bay (1) equals 4500/20 225 lb. =
The end moments are 11 VL/2 3000 x 30/2 = =
= 45000 in.lb. Assuming the beam flanges Fig. A21.40 shows a segment 1 inch wide
develop the entire bending resistance the beam cut from wing bay (1) with the 6P load in each
A21. 13
flange member. To find the slE~ar flow on the bay (1) as shown, The areas of corner stringers
cross-section the front web is first assumed a, b, c and d are shown in ( ) adjacent to each
cut, and thus the static shear flow qs = Z6P stringer.
from cut face where qs is zero. Fig. A2l.40
shows this static shear flow.
8800 lb.
No Skin on
Bottom of
Bay (2).
r--15"~
Fig. A21. 40 '-.:
~.'
~ ~:I Rib
j:2;'00
t<:;'" ~ I
-l __ /-- T
1"
For eqUilibrium of the cross-section, the iV /$/: /- -- - --

moment of the forces in the plane of the cross- ~
'-.:: . '"
cf; t<$: I/ Rib A
section must equal zero. TakinG moments about t<:; .... I
1 1/
lower left hand corner of the qs force system, 4,~_ / - - - - - --
I 1------
=
M 225x40xlO = 90000 in.lb. For eqUili- I /
brium a moment of -90000 is necessary •. There- &., Rib
fore a constant shear flow system q must be ~ ~
t<:;'" }------
added to develop a moment of -90000. Thus
q = M/2A = (-90000/2xlOx40) = -112.5 lb./in. I (1) / (0.5)
Adding this shear flow to that for qs in Fig.
A21.40 gives the final values in Fig. A21.41.
fa / b t
12" / Rib 8" Fig. A21. 43
This shear flow system represents the stress
-h)~/d~-----c----:: -1
(0.5)
I~ 30" --_>1
112.5
Fig. A21. 41 The mii'dle bay (2) has no skin on the

112.5 botto:n surface, or in other words, the middle
bqy hos a channel cross-section, which fact
often hapnens in practical wing design as for
system caused on cross-section of bay (1) due example a space or well for a retractable land-
to the differential bending of the beams in ing gear. The problem will be to find the
bay (2). This shear flow system must therefore shear flow in bays (1) and (2) and the boundary
be resisted by rib (A) as it must terminate at loads on rib (A) between bays (1) and (2).
end of bay (1). Therefore the shear flows in
Fig. A21,41 are applied boundary loads to Solution No.1
rib (A) and these must be added to the rib
loads in Fig, A21.37 to give the final rib loads This method of solution will make use of
of Fig, A21,42. With the final rib loads the shear center location for bay (2) in order
to obtain the true torsional moment on bay (2).
With this torsional moment known, the procedure
312.5
tl
t
- - -If
212.5
-4000#
1312.5
is similar to the previous example involving
wing torsion only.
'c-.-=::-:;;=--=:-::;;:::-::;=-:;=- t __ 4000# We will first calculate the shear flow in
212.5 wing bay (1). Fig. A21,44 shows the cross-
Fig. A21. 42
section.
8800#
known, the rib flange and web stresses can be
found as previously explained. b-15"
1
-,- (0.5)
A21. 9 Example Problem. Wing with Cut-Out Subjected 6"
to Bending and Torsional Loads. x---t-- b
x
Fig. A21.43 shows a portion of a 4 strlnf!:er

. 2400# -L:....-
3"
6"
. ~d~_---1~_ _----~~r
-f'
single cell cantilever beam composed of 3 bays T~) _ I
formed by the four ribs. The loads on the jX ---
structure consist of loads applied to end of rIO" Iz 30" Fig. A21. 44
The section moments of inertia are needed Therefore we consider bay (2) in its true
in calculating shear flows. condition with bottom skin removed. Fig.
A21.47 shows the cross-section of bay (2).
2
in~
2
Ix = (lx6 x2) + (0.5X4 X2) = 88
I .8800
X = 'ZAx/'ZA = (1 x 30)/3 = 10 in. I- 15,,--J
I z = (2xI0 2 ) + (lx20 2 ) = 600

1t-"--~ __ b
in~
2400 12"
3~li
I qad %b t
%cjL
r;"
Vz 8800 lb., Vx = 2400 lb. .t- 'd
c
_ V" "zA _ Vx "xA Fig. A21. 47
-"""'- L. L. SU b S t·t
1 U t·lng
Ix Iz '
The three shear flows can be determined by
qy =- 100 'ZzA - 4 'ZxA (a)
statics.
Since the shear flow is unknown at any
point on cell, we will assume front web (ad) as 'ZFx = 2400 - 30 Qab = 0, whence Qab = 80
cut or carrying zero shear.
'ZMd = 2400 x 3 - 8800 x 1.5 - 80 x 30 x 12 + qbc
-100 (-6)(1) -4 (-10)(1) 640 Ib ./in (8 x30) = 0, whence qbc = 640
qcb = 640 -100 (-4) (0.5) - 4 (20) 0.5 = 800 'ZFz = 8800 - 8 x 640 + 2 x 80 - 12 Clad = 0
whence Qad = 320
qba = 800-100 (4)(0.5) -4 (20) 0.5 = 560
Fig. A21.48 shows the results. This shear
Fig. A21.45 shows these static shear flows. flow system is the final or true shear on bay
(2) •
L .4- 8800
I[a-
,- 15'~
80 lb. /in.
a --- _.~ .!!::Ji~
cut
bl
r 320 I blfi, 640
2~ A=Cell Area = 300 c t 800 Id c
~ l:d~--,=-==--:;::=-=;:::::::---::;;=~
-~-640-- Fig. A21. 48
Fig. A21. 45
Since we have a channel or ODen wing cross-
section in bay (2), any torsional moment on
To this shear flow, a constant shear flow this bay must be transmitted by differential
must be added to make ZM = O. Take moments bendin~ of the front and rear beams. To obtain
about Doint (d). the torsional moment on bay (2), the shear
center location must be knoNll.
'ZMd = -8800 x 15 + 2400 x 3 + 560 x 30 x 12 + 800
x 8 x 30 = 268800 in. lb., or -266800 Horizontal location of shear center: -
Assume the section bends about centroidal X axis
is required for eqUilibrium, hence the required without twist under a Vz load of 8800 lb.
constant shear flow q = -M/2A -268800/2 x 300=
= -448. Adding this shear flow to that of Fig.
q =- ~~ 'ZzA, or q =- 100 'ZzA
A21.45, we obtain tha shear flow of Fir;. A21.46.
qcb =- 100 (-4)(0.5) =200

qba = 200 100 (4)(0.5) =0
Clad =0 - 100 (6)(1) = -600
192 Fig. A21.49 shows the shear flow results
Fig. A21. 46 for bending about x-x without twist. The line
of action of the resultant of this shear flow
force system locates the horizontal position of
This shear flow system would be the shear flow the shear center.
system for all 3 bays if the bottom skin in bay
(2) was not re~oved. Removing the bottom skin X = (200x8x30)/8800 = 5.45 in.
in bay (2) will modify these shear flows of
Fi~. A21.46.
A21.15
ential bending of beams in bay (2), thus equal

600 1~'4~ the beam shear times half the span of bay (2)
divided by the beam depth.
+d R=8800
For front beam P = 3841 x 12.5/12 = 4000 lb.
Fig. A21.49
For rear beam P = 3841 x 12.5/8 = 6000 lb.
Vertical position of shear center: -
Fig. A21.52 shows these flange loads
Assume section bends about centroidal z applied to bay (1). These loads are dissipated
axis without twist under a load of Vx = 2400 lb.
qcb
v
=- ~
Iz
l:xA =-
= -4x20xO.5
4 l:xA
=- 40 Ib./in. ~
71 1----_
L /
qba = -40 - 4 x 20 x 0.5 = - 80
Qad = -80 - 4 (-10) 1 = -40 r§

ja/ ItdL
I
Rib A
'""
Bay (1)
Fig. A21. 52
Fig. A21.50 shows the shear flow results. ~I /
~
R=2400~ ~
~=--4"
~t~-;::=----:;8p::,0-=---- b
l40 40~ uniformly in bay (1) over a distance of 30
d inches, or the shear flow per inch produced by
Fig. A21. 50
c these flange loads equals ~P P/30, whence =
~Pa = ~Pd = 4000/30 =133.3 and ~Pb = ~Pc
The vertical distance Z from point (a) to
the line of action of the resultant which = 6000/30 = 200 lb.
locates the vertical location of shear center
is, Fig. A21.53 shows an element of bay (1)
one inch wide with these ~P loads. The shear
flow q assuming the front web cut equals l:~P.
Z = L:Ma /2400 = (40 x 8 x 30)/2400 = 4 in. The resulting static shear flows which equals
l:~P is shown in Fig. A21.53.
Fig. A21.51 shows the shear center location
and the external loads. The moment about the
shear center which equals the torsion on the
wing bay (2) equals, lOlL
Ms • c • =- 8800 x 9.55 - 2400 x 13 = -115240 ,0~-m3--~66.7

in. lb. c~\.~ld:: 2DD ~ Fig. A21. 53
" s.c.
1_ Ql~.554
8800
4
, t--
133.3
---
133.3 I
200
2400 I9 ~~
.n
d
Fig. A21. 51
The moment of this shear flow system about
point (d) =
133.33x30x12-66.7x8x30 = 31980.
For L:M = 0, we need a constant shear flow q
- 31980/2 x 300 = -53.3 Ib./in. Adding this
=
constant shear flow to that of Fig. A21.53
gives the shear flow system of Fig. A21.54.
This torsional moment must be resisted by These results represent the effect on bay (1)
front and rear beams. Hence shear load on
each beam =
115240/30 3841" lb. =
80
As in the previous example problem in- -~-
~ ~
volving torsion, the beams in bay (2) will be 53.3 t f 120 501.3 ~ '232
assumed to bend without rotation of their ends, Fig. A21.54 t +
f -So-"""- t
or in other words the bending moment at mid-
point of bay is zero. The flange loads at
points a, b, c and d on bay (1) from the differ-
of removing the bottom skin in bay (2). Adding 192 with sense as shown. Since this skin is
the shear flows of Fi~. A21.54 to those of Firs. missing we reverse this shear flow and find the
A21.46, we obtain the final shear flows in bay resisting shear flows on the other three sides
(1) as sho,vn in Fig. A21.55. of the bay cross-section. Fig. A21.58 shows
the section, with the 3 unknown shear flows
BOUNDARY LOADS ON RIB (A) Qab, qbc and Qad'
The bouncary loads on rib (A) will equal
f~l
the aif ference bet\veen th3 shear flows in bays
(1) and (2). Fi~. A21.56 shows a free body of
rib (A) with the shears flows obtained from qad=128 t~~t %c=288
Figs. A21.55 and A21.48.
- 192 - Fig. A21. 58
'::la'-'?>
p--=:=_l_~~ _ ~ (1) Rib A To find qab use ZFx = 0, 192 x 30 - 30 Qab =0

~t-_·_- I Whence, qab = 192
J --
, 80 : \232
~~---=:::---_~--c-. _
I, ~ ~'l2-----=--~-i
/-.~~ 640
., ZNd = -30 x 192 x 12 + 8 qbc x 30 =0
o Whence, qbc = 288
Bay (2) Fig. A21. 56 ZF z = 4 x 192 - 8 x 288 + 12 qad =0
Whence, qad = 128
The resulting applied boundary forces to
Adding the sh3ar flows of Fig. A21.58 to
the rib equal the alGobraic sum of the shear
those of Fig. A21.46 gives the final shear flows
flows on each side of the rib which ~ives the
in bay (2) as sho,vn in Fig. A21.59. These re-
values in Fii"'. A21.57.
80
320 I[_-_-----=------J-t 640

o Fig. A21. 59
Fig. A21. 57
suIts check the results in Fig. A21.48 obtained

With the rib boundary loads known, the
in solution method (1).
stresses in the rib can be found as previously
illustrated in this chapter.
Solution No.2
This method of solution first finds the

shear flow in all bays assuming bottom skin is
not removed in center bay (2). This f",i ves a
shear flow in the bottom skin. However, the
skin in bay (2) is actually removed so a
corrective set of shear flows on bay (2) along
the boundary lines of the bottom skin must be
applied to eliminate the shear flows found in
the bottom skin. The problem then consists of
finding the influence of these corrective shear
_
flows upon the shear flows as found for bays
(1) and (2) when bottom skin in bay (2) was Fig. A21. 60
not removed.
.--- --- 192 --- ........ ..- -
The first step is to find the shear flows
in all bays ass~~ing bottom skin in bay (2) is
not removed. The calculations would be exactly Fig. A21.60 shows the corrective shear
like those in solution (1) and the shear flow flows of Fig. A21.58 applied to bay (2). On the
in all bays would be those in Fig. A21.46. The bottom skin the corrective shear flow is shown
bottom skin in Fig. A21.46 has a shear flow of on the boundary of the cut-out. These shear
A21. 17
flows cause differential bending of the front

and rear beams in bay (2). If we make the
assumjJtion that the beam end suffer no rotation,
the bending moment is zero at midpoint of the
bay and thus the flan~e loads at points a, b,
c and d of bay (1) equal the al~ebraic su~ of
the shear flows on each side of a flange times
half the span of bay (2) or 12.5 inches. Thus
from Fig. A21.60,
Pa = (192 + 128)12.5 = 4000 lb. compression
Pb = (288 + 192)12.5 = 6000 lb. tension
Pc = (288 + 192)12.z) = 6000 compression
Pd = (128 + 192)12.5 = 4000 tension
Referring to Fig. A21.52, we find that the

P valueS above are the same as the P values
obtained by solution (1). Thus the remainder
of solution (2) would be identical to that in
SoDItion (1), and therefore the calculations
will not be repeated here.
A21. 10 Fuselage Frames
Frames in a fuselage serve the same pur-

pose as ribs in wing structures. Ribs are
usually of beam or truss construction and can
be stress analyzed fairly accurately by statics.
Fuselage frames however, are of the closed
ring type of structure and are therefore static-
ally indeterminate relative to internal
stresses. Once the applied loads on a frame
are known the internal stresses can be found
by the application of the elastic theory as
covered in Chapters AS, A9, AIO and All. The PHOTO. A21. 5
loads on fuselage frames due to discontinuities Fuselage-Wing Portion
in the fuselage structure, such as those due to of "Martin" 404 Transport
windows and doors, can be approximately de-
termined by the procedures previously presented A21. 12 Calculation of Frame Boundary Supporting Forces.
for wing ribs.
Example Problem 1
The photographs on pa~e 32 of Chapter A15
show some of the fr&~e construction of the Fig. A21.61 illustrates a cross-section of
Douglas DC-S airliner. Other Dictures of a circular fusela~e. Two concentrated loads of
fuselage construction are given in Chapter A20. 2000 lb. each are applied to the fuselage frame
Photographs A21.4 and 5 illustrate typical at the points indicated. The problem is to
determine the reacting shear flow forces in the
frame construction and arrangement.
fuselage skin which will balance the two ex-
A21. 11 Supporting Boundary Forces on
ternally applied loads. This fuselage section
Fuselage Frames. might be considered as the aft portion of a
medium size fuselage and the loads are due to
When external concentrated loads are air loads on the horizontal tail surfaces. To
applied to a fuselage frame through a suitable make the numerical calculations short the
fitting or connection, the frame is held in fuselage stringer arrangement has been assumed
equilibrium by reacting fuselage skin forces sym~etrical.
which are usually transferred to the frame
boundary by rivets which fasten fusela~e skin Solution:
to frame. Since the fuselage shell is usually
stress analyzed by the beam theory, it is In this solution the fuselage skin resist-
therefore consistent to determine the distri- ing forces will be assumed to vary according to
bution of the supporting skin forces by the the general beam theory. The general flexural
same theory. shear flow equation for bending about the Y
axis is,
= -31.82-6.275x .30x13.5 = -57.22

q =_Vz ~zA , where Vz = 4000 lb.
Iy
= -57.22-6.275x .15xlO = -66.62
The moment of inertia I y of the fuselage
cross-section is required. In this simplified = -66.62-60275x.l.5x5 = -71032
illustration, the area of each stringer plus its
effective skin will be taken as .15 sq.in. The Due to symmetry of effective cross-section,
student should of course realize after studying the shear flow is symmetrical about the Y axis.
Chapters A19 and A20 that the true effective
area should be used on the compressive side and As a check on the above work, the summa-
that the skin on the tension side of the fuse- tion of the z components of the shear flow on
lage is entirely effective. These facts would each skin panel between the stringers should
tend to make the effective cross-section unsym- equal the external load of 4000 lb.
metrical about the Yaxis. Since the only pur-
~F z of skin shear flow equals the vertical
pose of this illustrative solution is to show
how the frame loads are balanced, the section projected length of each panel times the shear
being assumed as symmetrical which will greatly flow q on that panel, or
decrease the amount of calculations required.
~Fz = [1.4x16.57+2.7x31.S2+3.5X57.22
2000 ~ 2000 + 5x66.62+5x71.32] 4 = 4000 lb.
~
(check)
Double angles at 11. 5 , 4 - 11. 5 ' ]
3,3',9,9'. 1 I' Fig. A21.62 shows the frame with its
2 , bulb angle balanced load system. The internal stresses
stringers can now be found by the methods of Chapters
31~~:f::::::=ff:::::::::::::::::::-. AS to All.
2000 2000
5' t f
6' Y
11 11'
Fig. A21. 61
Moment of inertia of fuselage section

about Y axis which is the neutral axis under
our simplified assumptions.
Iy = .15 (17.6"+16.2'"+13.5'"+13.5'"+10'"+
5")4 = 637 in.4
Fig. A21. 62
Due to symmetry of effective section and
external loading, the shear flow in the fuselage
skin on the z axis or between stringers 1 and 1 Example Problem 2. Unsymmetrical Vertical Loading
and 11 or 11 will be zero. Thus starting with
stringer (1) the shear flow in the skin resist- In certain conditions in flying and land-
ing the external loads of 4000 lb. can be ing unsymmetrical concentrated loads are applied
written around the circumference of the section. to 'the fuselage or hull structure. For example,
Fi s . A21.63 shows the same section and frame as
wa~ used in Problem 1. Due to an unsymmetrical
q = _VI z ~zA = _ 4000
637
~zA = _ 6.275 ~zA load on the horizontal tail, the reactions from
y the tail on the fuselage are as illustrated in
the figure. The total load in the z direction
= -6.275x .15x17.6 = -16.57 Ib./in. is still 4000 lb. but the loads are not sym-
= -16.57-6.275x .15x16.2 = -31.82
metrical about the z axis. For analysis pur-
poses, consider the loads as transferred to the
A21. 19
e.g. of the section as indicated in Fig. A21.64.

The moment of the two loads about the e.g.
1500 x 11.5 - 2500 x 11.5 =
-11500 in .lb. The
= I+ - 8"+- 8" +- 8" ~
500
,-.--
shear load Vz =
4000 produces the same shear / h f
(3)
c
3"
flow pattern as Fig. A21.62. To balance the /
moment of -11500, a constant shear flow q J. / (1) e l- f.- 500
b
/' (2)
around the frame is necessary. / ;.'Ieb
(4)
- M-
q - 2A - 2
11500
x 11 x 18" =. 5 65 Ib /i
. n.
; /"
tg }d
I
600 500
(A = area of fuselage cross-section)
Fig. A21. 66
Adding this constant force system to that
of Fig. A21.62, gives the final boundary (2)
supporting forces on the frame as illustrated
in Fig. A21.65. The elastic stress analysis
of the frame can now proceed.
1500 2500
l .....~l- ~......,+
Fig. A21.67 shows a wing rib inserted in a

Fig. A21. 63 Fig. A21. 64 3 flange single cell wing beam, which is sub-
jected to the external loads as shown.
(1) Find rib flange loads at (c) and (d).

(2) Find rib web shear flow on each side
of stiffener cd.
(3) Find rib flange and web loads at
section 5" to left of line abo
Fig. A21. 68
r 600
I 7 1000.. of<.
10
A21. 13 Problems.
Fig. A21.68 shows a 3 stringer single cell
wing beam. A rib is inserted to distribute the
(1) Fig. A21.66 shows a cantilever beam loaded concentrated loads as shown.
as shown. Find the shear flow in each of the
4 web panels. Draw axial load diagram for (1) Find shear flows in rib web panel (1)
each of the vertical web stiffeners and also (2) and (3).
the horizontal stiffener be. Plot axial load
diagram for beam flange members as obtained (2) Find rib flange loads at sections dc
from web shear flows. and abo
(4) Fig. A21.69 (8) Same as problem (7) but with top skin re-
shows a 2 stringer, moved instead of lower skin.
2 cell wing beam.
A rib is inserted (9) Same as (5) but with read spar web removed
.03 to transfer 1000 lb. instead of bottom skin.
load to beam struc-
ture. (10) Same as problem (7) but with rear spar web
Fig. A21. 69 b removed instead of bottom skin.
r-- 10
"--1 Find shear flow
in rib web in each
cell adjacent to
(ll) In Fig.
A21.71 the external
line abo Also rib bulkhead loads Pl
flange loads ad- and P" equal 4000
jacent to points lb. each and P3
(a) and (b). equals zero. The
fuselage stringer
material consists
of four omega
~71:I
sections with an
area of .25 sq.
'-= .,. in. each. Deter-
J:l:;~ ~ I mine the skin re-
sisting forces on
~ :-/-------:
~ _....,0 I Rib B
the bulkhead in
balanc ing the
above loads. Fig. A21. 71
~ ~ § ,/ Neglect any effec-
J:l:;~ 4," ~ _ tive skin in this
1--I-~~'~P4
problem.
[/1, I (12) Same as problem (11) but make P~ = 4000

~
A,~ II Rib A
_I
i
p
and P"
(13)
=
6000.
Same as problem (12) but add P3 -- 3000 lb.
T-!
~
I r\
a c
c:rner Stringer Areas:-
(a) = 1 sq. in.
(14) In a water
landing condition
---,o'-r-__ p"
(b) =.8 sq. in.
(c) = • 5 sq. in. the hull frame ~15"-1
-l. ;b of Fig. A21. 72
b~
d (d) = 0.4 sq. in.
is SUbjected to a
toj~ i
normal bottom
pressure of 200
Fig. A21. 70
lb. per in.
The area of the ~I 40"
~I
bulb angle
(5) Fig. A21.70 shows 3 bays of a cantilever

single cell, 4 stringer wing beam. The bottom
skin in bay (2) is removed. Find the shear
flows in all bays and boundary loads on ribs
(A) and (B) when the external Wing loads are as
stringers is
.ll sq. in.
each and they
are 7/8 in. deep.
The area of the
Z stringers is
d
I I 8.5"
~
1
follows: T =
56000 in.lb., P~ = 0, P" = 0, .18 sq. in. each
P 3 = 2000 lb., P 4 2000 lb., P e O. = = and the depth 1.5 Fig. A21. 72
in. The area of
(6) Same as problem (5) but upper skin in bay the stringers a,
(2) is removed instead of the lower skin. b, c, d and e is
.20 sq. in. each. Neglecting any effective
(7) Same as problem (5) but with the following skin determine the skin resisting forces on
external loads. the frame in balancing the bottom water pres-
sures.
T = 56000 in.lb., P~ = 5000 lb.,
(15) Same as problem (14) but consider that
P" = 2000 lb., P3 =P 4
= 0 and Pe = 1000 Ib the water pressure is only acting on one side
of the bottom of the frame.
CHAPTER A22
ANALYSIS OF SPECIAL WING PROBLEMS
ALFRED F. SCHMITT
A22. 1 Introduction. the one finally analyzed, are treated only

lightly; additional references are cited where
In a previous chapter (A19) analyses of appropriate.
wing beams were carried out using the engineer-
ing theory of bending and rational modifications The analyses shown are strictly applicable
thereof. As discussed there, wing configura- to (reasonably) thin-skinned Wings only, where-
tions which depart radically from the usual in the "constant shear flow" assumptions are
conception of a "beam" present the engineer valid, Viz.
with the choice of making approximate and/or
empirical corrections to bea~ theory, or of i-the sheet carries shear stresses only
following a complete analytic treatment of the ii - normal (direct) stresses are carried
structure. in the flanges (spar caps and stringers) with
effective areas of skin lumped in. In all
This chapter illustrates the latter ap- cases handled here the skin was assumed fully
proach to several special problems associated effective (stresses below skin buckling stress -
with aircraft wing structures, viz., * see Art. A19.ll, Chapter A19).
Art. A22.2 - stresses around a panel cutout To enhance the usefulness of these problems,
Art. A22.3 - Shear lag problem all the structures chosen for analysis were
Art. A22.4 - cutout in a box beam taken from referenced NACA (National Advisory
Art. A22.5 - swept wing box beam Committee for Aeronautics) publications wherein
the reader may find detailed discussions of the
Aside from presenting one analytic treatment problems, other methods of analysis and data
of these problems, a discussion is given of the obtained from tests upon the specimens. Where
physical nature of each phenomenon. An under- available, these data have been used herein for
standing of the nature of the problem is of , comparison.
prime importance, Since no one analytic
technique can be all-powerful in the solution A22. 2 Stresses Around a Panel Cutout
of stress problems. The analyst must exercise
judgment and ingenuity in approaching each new "Cutouts in Wings and fuselages constitute
situation. one of the most troublesome problems confront-
ing the aircraft designer. Because the stress
In this chapter all analyses are made concentrations caused by cutouts are localized,
using the matrix formulation of the Method of a number of valuable partial solutions of the
Dummy Unit Loads (Chapters A7, AS), a famili- problem can be obtained by analyzing the be-
arity with which is assumed. havior, under load, of simple skin-stringer
panelS" (1)**
Such problems as those listed above are
too unWieldy to be studied here in great detail; ThUS, in the case of a wing beam With a
hence no attempt at exhaustive analyses has panel cutout of the upper surface (Fig. A22.1),
been made. To bring into relief the main it would be feasible to analyze the section
features of each problem, the structure selected immediately around the cutout as a flat sheet-
for analySis is one which is simple in con- stringer panel under the action of axial
struction and so loaded as to exhibit clearly stringer loads and edge shears (coming from the
the phenomenon under study. Many practical spar webs). The axial stringer forces could be
detailS, such as the effects of sheet wrinkling, computed with sufficient accuracy by the en-
rivet and fitting "give", stress concentrations, gineering theory of bending (E.T.B.) since
etc., have been side-stepped so as not to be- these are removed sufficiently far from the
cloud the objective. Further, the problems of cutout proper. The edge shear flows are readily
idealization of the original structure, into computed by those elementary considerations
which give the spar-web shear flows.
* One other important special problem - the so-called "bend-
ing stresses due to torsion" - is not treated here specifically.
As indicated in Chapter A8, the general box beam analysis
presented there encompasses this problem (Example Prob. ** Numbers in parentheses refer to the bibliography at the
15, p. p. A8. 24 through A8. 27). end of the chapter.
A22.1
A22.2 ANALYSIS OF SPECIAL WING PROBLEMS
Fig. A22.1
Fig. A22. 2 Idealization of the half panel by
use of substitute stringers
The sheet-stringer panel may, in general,
contain a large number of longitudinal ele- Fig. A22.3 gives the geometry of the ideal-
ments (stringers). The labor involved in ized panel.
treating this multi-element structure in detail A 1 = .703 in 2
is prohibitive, and thus an appropriate ideal- I-Tr~~-n-----,-----,
ization must be made. First, it is likely that I A I
F==H:=;:!£ ~~-~
the panel may be considered to be symmetric 2
about a longitudinal axis, so that only the A",= 1.045 in
J
half-panel need be handled. Second, the com- , ) AR = 0.25 in 2
plex, multi-stringer structure is replaced by I=::H=~-~---"~
one having but three stringers. As indicated I
I
I 1
t 1 = 0.0331 in
in Fig. A22.2, these stringers are: L_ _ ~ ~
I I " t 2 = 0.0331 in
,..b1 -b2 "; iL
#1, a substitute stringer having for its
area all the effective area of the fully con- ate
A1 A
1--0
A",
b 1 = 5.96 in
2
tinuous members to one side of the "combing 1 t1 ' t2 b2 = 7.56 in
stringer" (the stringer bordering the cutout)
and placed at the centroid of the area of Fig. A22.3 L = 15.0 in
material for which it substitutes. The stress
which this stringer develops is then the SOLUTION:
average stress for the material it replaces.
Fig. A22.4 is an exploded view of the half-
#2, the combing stringer, being simply the panel showing the placement and numbering of the
main continuous stringer bordering the cutout. internal generalized forces (Art. A7.9, Chapter
A7) and the external loading. Note that the
#3, another substitute stringer, this applied axial stresses were assumed to be con-
one replacing all of the effective material stant chordwise, giving stringer loads pro-
made discontinuous by the cutout. It is located portional to the stringer areas; their sum is
at the centroid of the material it replaces, P 2 , one of external loads.
and its stress is the average stress for this
same material. * The applied edge shear flows, coming from
the spar web, were assumed constant spanwise,
The sheet thicknesses used are the same as from a constant shear load. Other load
as those of the actual structure.** distributions may be handled by allowing these
applied shears to vary from panel to panel.
For very extreme load variations additional
* An alternate idealization, in which stringers #1 and #3 are transverse members could be inserted to create
located along the lines AB and CD, respectively (Fig. A22.2), more spanwise panels allowing a better fit to
was used in Reference (2) for a box beam loaded in torsion.
the spar shear variation. The applied shear
** When the longitudinal members themselves contribute to flows were considered as the other external
the shear stiffness of the cover (as is the case for "hat" section load and designated P 1 •
stringers riveted to the skin so as to form small closed
cross sections), an effective thickness must be used. This Panels on the centerline have zero shear
point is discussed in Reference (3). In this source, however, due to symmetry (Fig. A22.3).
the increase in shear stiffness is accounted for, not by
increasing skin thickness, but by decreasing the panel width -
an equivalent procedure.
A22.3
The half panel was twice indeterminate.

Member loads q sand qa were selected as redun-
qJ.7=' 359 (P 2 +3PJ.L)
dants. With these set equal to zero, successive
qJ.s=·108 (P2 +3PJ.L) applications of loads PJ. and P2 were made.
qJ.g=' 533 (P2 +3PJ.L)
Also, successive applications of unit values of
qs and qe were made. The results:
~~ 1 2 6 8
1 4.755 .316 -1.68 - .840

2 9.510 .316 - .840 -1.68
3 -19.76 -1.316 1.68 .840
4 -39.51 -1.316 .840 1.68
5 0 - .268 1.00 0
6 0 0 1.00 0
Fig. A22.4 7 -12.09 - .268 0 1.00
8 0 0 0 LOO
9 - L317 - .0450 .112 .056
Experience has shown that for symmetric 10 - 1.317 0 - .056 .056
panels symmetrically loaded it is satisfactory
11 .711 .0450 - .056 - .112
to consider transverse members to be rigid in
their own planes (4). Thus, in thiS problem, 12 0 .0355 0 0
member flexibilities for forces q6' qs' q7 and
q S may be taken to be zero. In the actual 13 - 1.599 - .0355 0 0
NACA test specimen with which results are to 14 0 .359 0 0
be compared, those transverse members bordering
15 0 .108 0 0
the cutout appeared to have been heavily re-
inforced (to an extent unknown to the writer). 16 0 .533 0 0
Hence it is logical to take their stiffnesses
17 16.16 .359 0 0
as great.
18 4.86 .108 0 0
Member flexibility coefficients were
19 23.98 .359 0 0
collected in matrix form as below using the
formulas of Chapter A7. Note that the co-
efficients for sUbscripts 5, 6, 7 and 8 were
set equal to zero (rigid transverse members).
O<ij, MEMBER FLEXIBILITY COEFFICIENTS (E = I) (Voids denote zeros)
1:\ 1 2 3 4 5 6 7 8 9 10 11 12 13 14
-3.56
15 16 17 18 19
1 14.22 3.56
2 3.56 14.22 -3.56
3 47.17 11. 79 -11.79
4 11.79 47.17 -11. 79
5
6
7
8
9 7 024
10 7,024
11 7 024
12 8 909
13 8 909
14 -3. 56 7.11
15 -11. 79 23.58
16 4.785
17 -3.56 7.11
18 -11. 79 23.58
19 4.785
The "off diagonal" values have negative sign because the sense of those internal generalized forces having subscripts (14), (15),
(17) and (18) was taken opposite to that used in the derivation in Art. A7.10. A change in sense requires a change in sign in
off -diagonal coefficients only.
The following matrices were formed:
Per eq. (17) of Chapter A8:
I.: ] 1=-5206 -262.8l

La<: rn = l.:6633 -262.~
Per eq. (18) of Chapter A8:
r=
~rs
] = [392.1
293.5
293.5l
392.l.J .5
The inverse of this last was found; CALCULATED MEASURED
c.rs -1] = 10 -(3 [ 5.802 -4.343J

Fig. A22. 5 Comparison between calculated and
measured stresses (psi) on the
C- J -4.343 5.802 half-panel.
Finally, per eq. (23) of Chapter A8, the

unit load stress distribution was,
inserted in theOCij matrix (in place of the
zeroes used above) were
~ 1 2
1 -10.93 -.65 ~ 5 6 7 8
2 -18.34 -.65 5 18.03 5.04

3 - 4.08 -.351 6 5.04 10.08
4 -11.66 -.351 7 18.03 5.04
5 1.40 .115 8 5.04 10.08
6 1.40 .383 The problem was solved retracing the same
7 3.79 .115 steps as before, but using the mOdified~i~
matrix, to yield the stresses:
8 15.88 .383
9 - .270 .019 IX 1 2
10 - .510 0 1 -11.11 -.621

11 - 1.15 -.019 2 -16.83 -.621
12 0 .0355 3 - 3.90 -.379
13 - 1.60 -.0355 4 -13.17 -.379
14 0 .359 5 2.14 .104
15 0 .108 6 2.14 .372
16 0 .533 7 2.52 .104
~i~ FLEXIBLE
17 16.16 .359 8 14.61 .372
18 4.86 .108 = 9 - .26 .017
RIBS
19 23.98 .359 10 - .62 0
The above analytical results are compared 11 - 1.05 -.017
with NACA test data (1) in Fig. A22.5 for the 12 0 .036
loading P 2 = 1, P l. = O. Agreement is seen to
be good. 13 - 1.60 -.036
14 0 .359
EFFECT OF RIB FLEXIBILITY:
15 0 .108
To investigate the influence of rib 16 0 .533
flexibility, the problem was reworked assuming
aluminum rib caps, of constant area A = .25 in.~ 17 16.11 .359
as the transverse members bordering the cutout. 18 4.86 .108
The appropriate member flexibility coefficients
19 23.98 .359
A22.5
Comparison of this result with the pre- shear. The action may be comprehended readily
vious one for the rigid ribs reveals that the by visualizing an extreme case: a large degree
most important effect of rib flexibility was of "lag" would occur if the load transferring
to increase the concentration of stresses in skin were made of a highly flexible material
the combing stringer bordering the cutout. It such as a plastic sheet or even rubber. In such
should be noted, however, that for this sym- a case the inside stringers would be out of
metric panel, the use of a very flexible rib action almost entirely! With the inside stringer
as compared with a rigid rib led to stress in- stresses lagging, the outside stringers and spar
creases of the order of only 10% in the combing caps must carry an over-stress to maintain
stringer. Thus, the "rule of thumb" that trans- equlli bri um ("actual" in the figure).
verse flexibilities may be neglected in sum-
metric panels is re-affirmed. Fig. A22.7 shows the beam analyzed herein.
A22.3 Shear Lag Analysis of Box Beams

f=· 636(A=. 5432
"The bending stresses in box beams do not
-1 3@8"1l---. ~
always conform very closely to the predictions -\
of the engineering theory of bending. The 12"
deviations from the theory are caused chiefly
by the shear deformations in the cover of the
box that constitutes the flange of the beam. 12"
The problem of analyzing these deviations from
the engineering theory of bending has become
known as the shear lag problem, a term that is
convenient though not very descriptive." (3)
Fig. A22.6 illustrates the basic problem.

The beam cover sheet is loaded along the edges
by shear flows from the spar webs. These shear 12"
flows are resisted by axial forces developed
in the longitudinal members (spar caps and
stringers). According to elementary consider-
ations, the stringer stresses should be uniform
chordwise at any given beam station ("ele- A= .3497
mentary theory" in the figure). Actually, the
central stringers tend to "lag behind" the Fig. A22.7
others in picking up the load because the
The beam is an idealization of one tested
by the NACA and reported in reference (3). Note
that the beam has no lower cover sheet a~d that
it is symmetric about a vertical axis. Trans-
verse bulkheads are located at stations 12", 24",
I 36" and 48" from the root.
i I I
-;:/ I The actual beam specimen had three more
~ Imirmm'lITiTlff'"T1'l~l'i"il-;:/ :
stringers than shown in the idealized structure,
-0 -;:/ '1 I these being located one each midway between the
I~~~~LI p~I_-----~~'1 pairs of longitudinals shown on the beam cover
I Ih' ~h j of Fig. A22.7. In the idealizing process, these
I # #; extra stringer areas were divided equally be-
I p I ~ I tween adjacent longitudinals. The stringer
: ~ I # areas shown are the effective areas, with those
Ip I d in the top cover tapering linearly from root to
I ' - - - - - - ; r ,../ tip. All skin was considered effective in
--Elementary carrying direct stresses.
Theory
- - -Actual Stress Some detailed discussions of the techniques
Fig. A22.6 of idealization of practical beams are given in
references (5) and (5a).
SOLUTION:
intermediate sheet, which transfers the loads To permit the handling of the calculation
in from the edges, is not perfectly rigid in in a limited space, it was elected to analyze
the structure for a single transverse (vertical) to zero, the stress distribution due to a one-
tip load symmetrically placed. In that case, half pound load at the tip ( a unit load
because of symmetry, it was necessary to treat divided equally between beam halves) was readily
only one-half of the structure. In addition, computed.
no shear flows could appear in the middle .07692
panels. Further, it is known that the influence
of rib flexibility on shear lag is slight for 0
symmetric systems, so that the ribs were con- -.9230
sidered rigid in their own planes; hence no
generalized forces were needed on the ribs to .9230
describe their strain energies. 0
Fig. A22.8 shows the placement and number- .07692
ing of the generalized forces on the half-beam. 0
-1.8461
1.8461
0
~imJ = .07692
0
-2.769
2.769
0
.07692
0
Fig. A22.8 Choice of generalized forces for shear -3.692
lag problem.
3.692
0
Member flexibility coefficients were
computed with the formulas of Chapter A7 and
arranged in a matrix. Next, the unit redundant stress distribu-
tion was computed. Fig. A22.9 illustrates a
The shear flows q", q7' q1." and q1.7 were typical calculation, showing the stresses in the
selected as redundants. Setting these equal tip bay for q2 = 1, q7 = q1.2 = q1.7 = O.
0<:1.. , MEMBER FLEXIBILITY COEFFICIENTS (E - 1)
I~ 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
1 3271
2 16 640
3 10.74 2.684
4 20.05 4.724
5 17.99 4.202
6 3271
7 16 640
8 2.684 8.939 1. 785
9 4.7M 17.84 4.255
10 4.202 15.78 3.708
11 3271
12 16 640
13 1. 785 7.142 1. 785
14 4.255 16.084 3.840
15 3.708 13.89 3.324
16 3271
17 16 640
18 1. 785 3.571
19 3.840 7.518
20 3.324 6.439
A22.7
q = -12
9 '" q =+12
~
\
+12
(Constant on in
__-12
to Root
Ecrj
= f2175 }
-1848
-1260
~ 1\ \ \; '\ - 460.7
~l\\~~~~\,~ t,210 3'M~

16,810 9,536
16,810 32,160 9,536 3,041
E=r~ = 9,536 9,536 25,030 3,041
tv~ l.~.l.b \~
3,041 3,041 3,041 18,650
q•• f\ The inverse of the latter was formed:

~~ ~ End Rib I ~\j
~ ~'2B' "~
-1.505 - .656
q1=0
10-6 -1.505 4.219 -1.000 - .279
Crs-j = - .656 -1.000 4.687 - .495
Fig. A22.9 Application of a self-equilibrating unit - .183 - .279 - .495 5.519
redundant stress q...
and, finally, per eq. (23) of Chapter A8,
The complete redundant stresses were:*

.07692
.03453
~ 2 7 12 17 - .9230
.5086
.4144
1 0 0 0 0
.07692
2 1 0 0 0 .03136
-1.846
3 0 0 0 0
1.055
4 -12 0 0 0 .7907
.07692
5 12 0 0 0
.02401
6 0 0 0 0 -2.7691
1.690
7 0 1 0 0
1.079
8 0 0 0 0 .07692
.01005
9 -12 -12 0 0
-3.692
= 10 12 12 0 0 2.493
1.199
11 0 0 0 0
12 0 0 1 0 Fig. A22.10 shows the above computed
stresses and those reported by the NACA as ob-
13 0 0 0 0
tained by test. Agreement is seen to be qUite
14 -12 -12 -12 0 good. 4. 589 ~ ('1. 885
15 12 12 12 0 1. 888
16 0 0 0 0 CALCULATED
17 0 0 0 1 3.416
18 0 0 0 0 2.364
1. 278
19 -12 -12 -12 -12
20 12 12 12 12 5. 1
The following matrix products were formed

(per eqs. (17), (18) of Chapter A8):
* This is an obvious place in which to use combinations of MEASURED

redundants to decrease the structural coupling (reference
Chapter A8, pp. A8.29, 30). (Recommended as an exercise
for the student. ) Fig. A22. 10 Comparison between calculated and measured
stresses (psi) in a box beam
A22. 4 Stress Analysis of a Box Beam With a Cutout. The calculation for step (2) above, will
now be carried out by considering a unit shear
In Article A22.2, one technique was em- flow, q:l.2 = 1, applied while all other loads
ployed for computing the stresses around a cut- are zero. Under this loading condition the
out. In that analysis the effect of the cutout relative displacements at redundant cuts 2, 7
was presumed to have been localized about the and 17 are equal to OC 2 :1.2' e-e 7 :1.2 and =:1.7 :1.2
, 1
cutout region; consequently, the problem was respectively, where these coefflcients have '
treated by isolating the affected panel. Quite been computed previously in Article A22.3.
often, when the cutout is placed well inboard
on the wing, its influence on the root stresses To restore continuity (reduce the relative
is appreciable. Therefore, it is desirable to displacements to zero) three redundant forces
be able to consider the overall problem of the are applied, one each, at the cuts 2, 7 and 17.
box beam with a cutout for such cases. The appropriate equations specifying continuity
are
"The most convenient and the most rapid
method of analyzing structures with cutouts is
the indirect, or inverse, method. The analysis
by the indirect method is made in two steps.
First, the structure is analyzed for the basic
condition that exists before the cutout is
made. The results of this basic analysis are Note that these equations say simply that
used to calculate the internal forces that
exist along the boundary of the proposed cutout. the deflections at the redundant cuts due to
the (unknown) redundant forces must be equal
External forces equal and opposite to these
internal forces are then introduced; these and opposite to the deflections due to q:l.2 = 1.
external forces reduce the stresses to zero All coefficients in the above equation
along the boundary of the proposed cutout, and
consequently the cutout can now be made without were computed in Art. A22.3. Specifically,
disturbing the stresses." (3)*
3'04~
~
0'210 16,810
9,536}
It is desired to modify the calculation of 16,810 32,160 3,041 {q2}
q7 = - 9,536
the previous article (A22.3, "Shear Lag { 3,041
AnalYSiS") to allow for the presence of two 3,041 3,041 18,650 q :J.7
cutouts symmetrically placed. Panel "q :1.2" was
removed while the single tip load remained. The above matriX is the [ocrsJ of Art.
Since the unit remained symmetric, the data A22.3, with the "q:l.2 row" and "q:l.2 column" re-
from the previous analysis, in which the trans- moved.
verse rib stiffnesses were taken to be
infinitely great, shOUld still yield satis- The equations were solved** to give the
factory results. values of the redundants for a unit applied
load, q:l.2 = 1, as
SOLUTION:
q2 - .1399
The calculation was accomplished in three
steps: q7 - .2134
1) The stress distribution was found in

{G s ,:J.2} = q:l.2
=
l.00
the "basic structure" (no cutout). This work q :1.7 - .1055 q:J.2 =1
was carried out in Article A22.3 where it was
found that q:J.2 = .02401 Ibs ./inch. The q:l.2 force is included in the above for
later convenience. Then the complete stress
2) A stress distribution was found for distribution, due to applying a unit shear flow
an "applied load" of q]'2 = l. Such a loading q]'1i =1, was
has zero external resultant.
3) .02401 times this last stress distri-

bution was subtracted from the first.
* The procedure described here is quite generally useful for ** See Appendix for a method of "extracting" the inverse of
studies of the effect of removing one or more members; this matrix from that preViously found for the complete
such might be required for an analysis of the effects of
structural damage. ~rJ matrix.
A22.9
where the row matrix l.?u,mJ is simply the "12

row" of @-im]BASIC
o COMPARISON WITH TEST DATA:
- .1399
o Reference 3 reports test data for the case
1.679 analyzed, the stringer stresses being plotted
-1.679 below in Fig. A22.11.
o
- .2134 In the actual test specimen a stringer
o passing through the cutout was severed, it hav-
4.240 ing zero stress at stations 12 and 24, there-
-4.240
= o fore. However, during the idealization process
discussed in Art. A22.3 (for the beam without
1.00 cutout) the area for this stringer was placed
o partly with the combing stringer and partly
-7.760 With the spar cap. In the same way, some
7.760 effective sheet from the midportion of the
o panel, now made discontinuous by the cutout,
- .1055 was added also to the spar cap and combing
o stringer.
-6.494
6.494 It follows then that the fUll idealized
areas of the combing stringer and spar cap
( [girJ taken from Art. A22.3) ShOUld not be used in figuring the stresses at
stations 12 and 24 (produced by forces qg'
To obtain the stresses in the loaded beam ql0' Q14' q16)' With these areas reduced by
the appropriate sUbtractions, the stresses
with the cutout, 0.02401 times the above were computed and are plotted in Fig. A22.11.
stresses were subtracted from the "basic" Agreement With test data is seen to be quite
stress distribution of Art. A22.3, as pre- satisfactory.
viously explained.
.07692
.03789
- .9230
.4683
CALCULATED
.4547
.07692
.03648
-1.8461
.9536
l = fa l
{ a imjcUToUTl -.02401{a
imjBASIC i, u
J= .8925
.07692
o
-2.7691
1.8763
.8927
MEASURED
.07692
.01258
-3.692 Fig. A22. 11 Comparison between calculated and measured
2.649 stresses (psi) in a box beam with cutouts.
1.037
Note that q~2 is now zero and the cutout A22.5 Analysis of a Swept Box Beam.
panel may be "lifted out".
"Experimental investigations of swept box
In the case of a structure under a variety beams have shown that the stresses and distor-
of external loadings (m = 1, 2, 3 ---), the tions in a swept wing can be appreciably differ-
more general equation, corresponding to the ent in character from those that would exist if
above, is the root were normal to the wing axis. The
principle effect of sweepback on the stresses
l
l imJBASIC -fal i,~~1 L
= la l
la a a ,mJ occurs under bending loads and consists in a
Li~CUTOUT concentration of bending stress and vertical
shear in the rear spar near the fuselage. With found in reference (8). One method of analyz-
regard to distortions, the effect of sweep is ing such wings is given in Chapter A23.
to produce some twist under loads that would
produce only bending of an unswept wing and THE STRUCTURE:
some bending under loads that would produce
only twist of an unswept wing." (6) The structure shown is Fig. A22.l2 is an
idealization of the NACA test beam of refer-
In the following example a swept box beam ences (6) and (9), in which a single substitute
is analyzed by the matrix methods of Chapter A8 stringer has been employed along the cover
and, in particular, by the specific techniques sheet to allow for the antitipated shear lag
of reference (7). The method accounts for the effect. The figure shows only one-half of the
interaction between the swept cover panels and complete unit, which was built symmetrically
the longitudinal members. It is this action about the axis corresponding to the longitudinal
that is responsible for the distinctive axis of the airplane.
structural characteristics of the swept box
beam. Only tip loads were to be applied (at
points A and B). The outer section of the beam
Again, we emphasize that the method used was assumed to carry stresses which could be
here is strictly applicable to thin-skinned calculated reasonably well by the engineering
wings of beam-like proportions only. Consider- theory of bending (E.T.B.). For this purpose
ing the wide variety of structural layouts it was judged satisfactory to consider the
which may be employed in swept wing configura- outer 66" of the beam as a single bay (A-B-D-C).
tions, a comprehensive treatment cannot be given If loads were to have been applied inboard of
here. An excellent review of methods better the tip, it would have been necessary to con-
adapted to thick-skinned construction and to sider additional bay divisions between A-B and
"plate-like" (very thin, wide) wings, may be C-D (that is, insert additional ribs at stations
1-
I
52"
.050 .050
I
I
I
C H D~I
!
i
1
yA=.539
I
I
Section C-C '?
.050 .050 (Steel) ~=. 125
I
66"
.078 t
0--""0
Ac..l. 373
Sectio~ A-A
I-'r
I...,A=l. 121
A A
t J A=l. 966
o-----....;""~o-----,>
A=l. 430
A (1) (2) B - - '

078 ·f Section B-B
0
r-:[.
0 -L
~ 15"-+-15"---~
Fig. A22.12 Fig. A22.13
A22.11
of load application). Rib C-D was located at affects total deflections only and can be
one of the actual rib locations in the NACA omitted in a stress analyses where deflections
test specimen and was assumed rigid in its own are not sought.
plane.
Since only symmetric loadings were con-
The choice of bay C-D-F-E as a single bay sidered in this analysis no shear was trans-
was somewhat arbitrary. For improved accuracy, mitted by the carry-through bay and hence no
additional ribs inboard of C-D could have been shear flows were shown in that portion.
used. Note that any ribs placed inboard of
point F will produce triangular skin panels in Sets of additional axial forces (q~6
the cover sheets. Examples of treatments for through q~e) were applied to the ends of the
such panels may be found in references (9), flanges and stringers adjacent to the obliquely
(10) and (ll). cut ends of the cover sheet panels in bay
C-D-F-E. These forces are necessary to account
Rib E-F was considered flexible in its own for the interaction between the swept covers
plane, it being known that the flexibility of a and the longitudinals. As shown in Fig. A22.14,
rib is important at a location where a structure the pure shear flow on the oblique edge is ob-
changes direction.* Note that this rib was tained by superposing onto the panel a zero-
made of steel in the test specimen. resultant system consisting of a uniform tensile
stress of intensity 2q plus a pair of concen-
Effective areas of longitudinals as shown trated balancing loads. The balancing loads
in Fig. A22.12 were computed by considering all must be contributed by the bordering longitudi-
of the skin to be effective. The spar cap nals and hence react on these as tensile loads
areas are equal to the sum of the areas of the (Fig. A22.14c). The balancing loads applied to
angle member at the cap location, plus one- the stringers are shown dashed since they are
half of the effective area of material between internal forces within the bay and are not to
the cap and the sUbstitute stringer (this area be entered into the equilibrium equations for
inclUdes several stringers as well as skin) the structure.
plus one-sixth of the attached spar web area:*
The substitute stringer area was collected in
like manner from the half-panels to either side.
The method used in calculating the

-
P=qL
~611 -
U
P p .. ~-==
LJ
~
effective areas of the rib caps (E-F) is given

in detail in reference (9), from which the
value used here was taken. The "carry-through
bay" cover sheet thickness is equal to that
-
2qm
P=qL
- 1
(a)
P_""'--
(b)
p~--!
...
(c)
"=
used on the specimen (.050") plus a weighted Fig. A22.14 Showing how the uniform shear stress on
increase to allow for the presence of splice an oblique panel end (b) is created by
plates along the plane of symmetry (see refer- superposition of a uniform tensile stress
ence 9). plus two balancing forces (a). The
balancing forces react as tensile loads on
the bordering longitudinals (c).
INTERNAL GENERALIZED FORCES:
Fig. A22.13 shows the choice and numbering From an energy viewpoint, these dashed
of the generalized forces. forces account for the additional strain energy
stored by the axial components of shear flows
The beam was rigidly supported at pOints in the non-rectangular panels. This energy is
E, F and at the two corresponding points on the stored in the cover panels themselves (and is
other beam half. These might correspond to the accounted for in this manner since the longi-
fuselage ring attach points in an airplane. tudinals contain the effective area from the
The vertical end caps on rib E-F were considered cover sheet***) and in the longitudinals which
rigid axially, so that no flexibility co- react against these components.
efficients were associated with the reactions
q,,~ and q" e' Flexibility of these members "Dashed loads" are applied to the longi-
tudinals adjacent to any obliquely cut panel
end. Similar dashed loads would be applied to
* The effect of neglecting this rib's flexibility is demon-
strated later in this example.
*** This much of the energy could be accounted for in another
** The factor of 1/6 is used so that the effective area con- fashion by modifying the member flexibility coefficient for
tribution of the web results in a structure haVing the same the sheet panel. See Reference 12, where this was done.
moment of inertia about a horizontal axis as the original. However, that referenc e incorrectly neglects the additional
Some of the problems of idealization are discussed in energy stored in the longitudinals, as was demonstrated in
reference (2), p. 16. Reference 13.
A22 12 ANALYSIS OF SPECIAL WING PROBLEMS
the outboard ends of the panels in bay C-D-F-E Stresses in inner bays:
if they too were cut obliquely. Such panel
configurations arise often in swept wing con- According to the discussion of Art. A8.12,
struction having ribs parallel to the air- Chapter A8, the determinate stress distribution
stream. Formulae for more general quadrilateral
panels are given in Reference 7. ~i~' may be any stress distribution in equil-
ibrium with the applied loads, and preferably
THE STRESS DISTRIBUTIONS one close to the final true stress distribution
The magnitude of the redundant forces is re-
For the symmetric loadings considered duced by use of a satisfactory estimate of the
here the structure was indeterminate only two true stresses.
times since the outer bay was assumed to be
determinate by the E.T.B. The stresses in the two inner bays were
determined for both gim and gir simUltaneously.
Stresses in outer bay by E.T.B.: Since this inner portion of the structure is
two times indeterminate we can estimate two
Flange stresses at rib C-D (for both loads. For this purpose the two flange loads
P~ = 1 and P" = 1) q ~" and q1.4 were written as
M = 66"# qu = q~" + q~"
I = 88.57 in. 4 ; C = 3.5" q1.4 = q~4 + q~4
fb =IMC = 2.608 psi where the (single) primed values are approxi-
mate values determined by the E.T.B. and the
Therefore, double primed values are the unknown corrections
(the redundants). Using My/I at stations 66"
and 118" from the tip gave*
q6 = q7 = 2.608 x 1.121 = 2.924 Ibs.
q6 = 2.608 x 1.373 = 3.581 Ibs. q~" = 5.228 P~ + 5.228 P"
For a unit transverse load at the shear q~4 = 3.899 P~ + 3.899 p..
center (midpoint, because of symmetry)
The equilibrium equations for the elements
of the structure were written next by summing
1#
Qa
=3.581 = 02713
2 x 66·
Ibs
in forces and moments.
q~t --->-.qa
qb~~~qb
Joint F
2.924
Qb =.02713 +~
~ ~
q"3 + q"6 =0
~ ~ = .07143 Ibs
in q"3 -+ q"6 = 5.513
1.414 qu
= P~ + 5.513 P" + 1.414 q~ 4
The unit load was shifted 15" to either side
by application of a torque, T = 15 in.lbs. Joint E
The uniform shear flow superposed was
q..4 + q"o =0
q = JA = 2 ~5 210 = .03571 ~b; q"4 - q"o = 1.414 q~"
= 7.392 P1. + 7.392 P" -+ 1.414 q~
"
Finally, superposition gave the stresses m about E-F
for the outer bay as
q"4 + q"6 + q"6 = 11.918 P1. -+ 8.888 P"
!~ 1 2
1 .1071 .03572
2 .06284 - .00858
[gimJ
= 3 .00858 - .06284
4 .03572 .1071
* This is a rather crude way to estimate these loads and is
5 2.924 2.924 used here only for simplicity. The analyst is generally better
advised to exercise a little more ingenuity in making these
6 3.581 3.581 estimates, even to the extent of being guided by other swept
wing solutions.
7 2.924 2.924
A22.13
E 1 0 0 0 -1 0 0 qs
~
dlJJ
G 1 -.7115 0 0 0 0 0 qe
'" G
c'
'" 0 0 21.2 0 21.2 0 1 qJ,o
~J
c'
t~... .
0)
uo
uo tco.
N
c:c
...
0 0 1.682 1 0 0 0 q1.1. =
0 0 0 1 1 0 0 q 1.e
C H H ~
qe qJ,o 0 -21.2 0 0 -21.2 1 0 q 21.

Shear Flows around the non- 0 0 0 0 0 -1 1 q 22
rectangular panels. (Check
by summing moments about
E, G and F.) - .1403 - .2111 0 0
.04431 .04431 .01923 0
Cap EC P1.
2.756 2.756 0 .707
P2
.04431 P1. + .04431 P2 + = .04432 .04432 0 .04545
q"1.2
.01923 q'~2 .2831 .3540 0 0
q"1.4
-3.696 -3.696 -.707 0
Rib Vertical at E
-5.466 -2.436 .707 .707
Cap EG After inverting the matrix of coefficients

on the left hand side of this last equation,
and mUltiplying through thereby, the stresses
were obtained as
Joint G
qs .1006 .0295 .00676 -.00676
q25 = 0 .0791 - .0208 -.01753 -.00950
qs
q21. + q22 = 0
q 1.0 .0013 - .0411 .00401 .0230 P1.
Cap GF .1135 -.00675 .00675
P2
q1.1.= .0422
q"1.2
q 1.e .2409 .2409 .00676 -.00676
q1.Q + q1.e + .04717 (q22-q23) = 0 q"1.4
q 21. 3.087 .964 -.9355 -.3444
Rib Vertical at F
q 22 -2.379 -1.473 -.2284 .3625
q1.1. + q1.e = .2831 P1. + .3540 P2
Cap DF The complete determinate and unit-
redundant stress distributions, using the re-
sults up to this point, were therefore:
q1.1. + 1.682 q1.0 = .04432 (p1. + P 2 )
+ .04545 q~4
The first five of the above equations

were readily solved by sUbstitution, yielding:
q23 = q"s = 2.756 P1. + 2.756 P 2 + .707 q"1.4

-q20 =q24 = 3.696 P1. + 3.696 P2 + .707 q"
1."
q25 =5.466 P1. + 2.436 P2 .707 q"1.2
- .707 q"1.4
Also, from equilibrium of joint G:
The remaining seven equations were

arranged for solution thus:
A22 14 ANALYSIS OF SPECIAL WING PROBLEMS
"'m r
1 2 12" 14" q:L6 = "PD" = 2 x 52.3 x 15 x (-) qe
125.8
= -12.5 qe
i\
= -.989 P:L + .260 P 2 + .2191 q~2 + .1188 q~ .
1 •1071 .03572 0 0
2 .06284 - .00858 0 0 q:L e = tIp c " =2 x 73.5 x 15 x (-) q"
125.8 = -17.5 qe
3 .00858 - .06284 0 0
= -1.384 P:L + .360 P2 + .3068 q"1.2 + .1662 q'1.I ..
4 •03572 .1071 0 0
Similarly, for the panel HGFD;
5 2.924 2.924 0 0
6 3.581 3.581 0 0 q:L7 = -.0146 P:L + .460 P2 -.0449 q~2 - .2576 q~ ..
7 2.924 2.924 0 0 q:L6 =
-.0244 P:L + .773 P2 -.0754 q~2 - .4324 q~ .
The following matrix products were formed:
8 .1006 .0295 .00676 - .00676
(cf. eqs. 17, 18, 23 of Chapter A8)
9 .0791 - .0208 - .01753 - .00950
10 .0013 - .0411 .00401 .0230 [e><rJ =[ 17.70
-173.7
250.2J
- 50.43
11 .0422 .1135 - .00675 .00675
12 5.228 5.228 1.0 0 [coerJ = [114.5
42.96
42'~IJ
83.54
m:gi~ 13 7.729 3.444 -1.0 -1.0
[:<rs -~ =[: 0.01082 - .00556~
14 3.899 3.899 0 1.0 - .005563 0.01483
15 - .989 .260 .2191 .1188
Finally, the true stress distribution was found
16 -1.384 .360 .3068 .1662 as,
17 - .0146 .460 - .0449 - .2576
1
~ 2
18 - .0244 .773 - .0754 - .4324 1 .1071 .03572

19 .2409 .2409 .00676 - .00676 2 .06284 - .00858
20 -3.696 -3.696 - .707 0 3 .00858 - .06284
21 3.087 .964 - .9355 - .3444 4 .03572 .1071
22 -2.379 -1.473 - .2284 .3625 5 2.924 2.924
23 2.756 2.756 0 .707 6 3.581 3.581
24 3.696 3.696 .707 0 7 2.924 2.924
25 5.466 2.436 - .707 - .707 8 .0747 - .0051
26 2.756 2.756 0 .707 9 .0740 .0112
10 .0581 - .0039
DASHED LOAD CALCULATIONS: 11 .0681 .1481
In the above matriX, loads q:L6' q:Le' q:L7 12 4.070 2.241
and q:L6 were obtained from qe and q:LO following
formulas given in reference 7 for general ~imJ = 13 6.213 4.292
quadrilateral panels. The equations applicable 14 6.573 6.038
to a parallel-sided panel are: 15 - .925 - .140
16 -1.295 - .201
'< yl~
d 17 - .6514 .0431
Pc=C+d qLx 2
18 -1.093 .073
/ c
PD=c + d qL x 2 19 .2150 .2060
PC7,/ T
d Iq L Balancing Load 20 -2.877 -1.584
Po.-.- / ---L (= Dashed Load)
21 3.249 3.022
V Formulae from Ref. 7
22 -1.145 - .015
For the panel CEGH; c 52.3", d = = 73.5", 23 4.646 4.268

c +- d =
125.8" L = 15". Hence 24 2.877 1.584
25 4.394 3.036
26 4.652 4.268
A22.15
MATRIX OF MEMBER FLEXIBILITY COEFFICIENTccij (E = I) (Formulas from Chapter A7 and Reference 7)
I 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26
15400
1 .~~L
102960
15400
70.170 15.46 15.46
50.01 8.984 8.984 8.984
52.33 6.542 6.542
12 130
69400
46000
5 130
15.46 30.92 30.92
8.984 17.97 17.97 17.97
6.542 13.08 13.08
15.46 30.92 30.92
8.984 17.97 17.97 17.97
8.984 17.97 17.97 17.97
6.542 13.08 13.08
2130
9.046 4.523
4.523 9.046
9.046 4.523
4.523 9.046
19.58
14.24
19.58
COMPARISON WITH TEST DATA: requirement. More details concerning the test-
ing techniques and method of data presentation
4.6 would probably resolve this conflict. Both test
and calculated values clearly exhibit the
characteristic build-up of stresses in the rear
spar of a swept wing.
RIB E-G-F RIGID:

As a matter of interest, it was decided to
investigate the effect upon stresses when rib
E-G-F is taken to be rigid. Such a calculation
is readily achieved by putting the member
flexibility coefficients for the rib equal to
zero.
/'
Thus, in the matrix ~ ij], those co-
Station {OJ /'
CALCULATED MEASURED
efficients with subscripts 19, 20, 21, 22 and
Fig. A22.15 Comparison Between Calculated and 23 were set equal to zero and the complete
Measured Stresses (psi) in a Swept
Box Beam. calculation was repeated.
The results, for spar cap loads at the wing

Fig. A22.15 shows a comparison between root (a most sensitive point), were:
the calculated stringer stresses and those
measured by the NACA as reported in Reference 6.
The stresses shown are for a unit tip load,
centrally placed (p~ ~ p.. ~ 1/2 lb.). ~ 1 2
::: 12 3.824 1.792

Considering the limitations on the analysis
the agreement is generally satisfactory. Thus, 14 6.993 6.610
the discrepancy in the My/I stresses at station
65 may be attributed to a lack of precise A comparison with the flexible rib calcu-
knowledge of the test parameters. The calculations follows:
lated stresses in the leading edge spar between
stations 65 and 118 cannot duplicate the ex- SPAR CAP
perimental variation since only a single bay LOADING FORCE FLEX. RIB RIGID RIB %DIFF.
was employed in this region in the idealization. Simple
The fact that the calculated root stresses run q1 .. 6.311 5.616 11
Bending;
consistently above the test values is difficult P1=P.. =1 q14 12.61 13.60 8
to explain. Inasmuch as the calculated stresses
satisfy equilibrium, the test values, all being Torsion; q12 1.829 2.032 11
lower, would seem to defy this fundamental P~=-P .. ~l q 14 0.535 0.383 28
Considering that rib EGF was relatively (7) Wehle, L. B., and Lansing, W., A Method for
rigid to begin with - being made of heavy Reducing the Analysis of Complex Redundant
gage steel - it may be seen that neglect of Structures to a Routine Procedure, Journal
the flexibility of a corresponding all-aluminum of Aero. Sci., 19, Oct. 1952
rib could lead to serious errors.
(8) Williams, M. L., A Review of Certain
REFERENCES Analysis Methods for Swept Wing Structures,
Journ. of Aero. Sci., 19, p. 615, 1952
(1) Kuhn, P., Duberg, J. E., and Diskin, J. H.,
Stresses Around Rectangular Cut-Outs in (9) Heldenfels, R., Zender, G., and Libove, C.,
Skin-Stringer Panels Under Axial Load - II, Stress and Distortion Analysis of a Swept
NACA WR L368 (ARR 3J02), Oct. 1943 Box Beam Having Bulkheads Perpendicular to
the Spars, NACA TN 2232.
(2) Rosecrans, R., A Method for Calculating
Stresses in Torsion-Box Covers with Cut- (10) Bisplinghoff, R., and Lang, A., An In-
outs, NACA TN 2290, Feb. 1951 vestigation of Deformations and Stresses
in Sweptback and Tapered Wings with Dis-
(3) Kuhn, P., and Chiarito, P. T., Shear Lag continuities, Mass. Inst. of Tech. Rept.,
in Box Beams: Methods of Analysis and JUly, 1949
Experimental Investigations, NACA TR 739, (See also, Journ. of Aero. Sci. 18, p. 705,
1942 1951)
(4) Hoff, N. J., and Libby, P. A., Recommenda- (11) Denke, P., The Matric Solution of Certain
tions for Numerical Solutions of Reinforced Non-Linear Problems in Structural Analysis,
Panel and Fuselage-Ring Problems, NACA TR Journ. of Aero. Sci., 23, 1956
934, 1949
(12) Levy, S., Computqtion of Influence Co-
(5) Kuhn, P., Approximate Stress Analysis of efficients for Aircraft Structures with
Multi-Stringer Beams with Shear Deformation Discontinuities and Sweepback, Journ. of
of the Flanges, NACA TR 636, 1938 Aero. Sci., 14, Oct. 1947
(5a) Kuhn, P., Deformation Analysis of Wing (13) Islinger, J. S., Stress Analysis and
Structures, NACA TN 1361, July 1947 Stress Measurements for a Swept Back Wing
Having Ribs Parallel to the Airstream,
(6) Zender, G., and Libove, C., Stress and McDonnell Aircraft Corp. Report 1127, April
Distortion Measurements in a 45° Swept Box 1949.
Beam Subjected to Bending and Torsion,
NACA TN 1525
CHAPTER A23
ANALYSIS BY THE METHOD OF DISPLACEMENTS

JOSEPH A. WOLF. Jr.
A23. 1 Introduction Of course both metrlOds can be used to des-

cribe structural behavior. For example, some
The displacement or stiffness method (1-4)* of the disadvantages of the force method can be
is a structural analysis procedure using dis- overcome by means of progranulled automatic re-
placements as the independent variables (and dundant selectors or "structure cutters." The
unknown quantities) rather than the forces. As engineer must be fa'niliar wIth both methods in
applied here, the displacement method will be order to perform the analysis function effi-
used in conjunction with a structural modeling ciently.
using finite elements. Typical elements used
are rod and beam segments, thin sheet and plate The formulation of the displacement method
sections, and conical shell frusta. The number has been perfected within the last 15 years,
of unknown displacements is determined by the prompted by the availability of large electronic
number of elements chosen and the lcwel of digital computers. In 1950, the solution of a
sophistication or detail used in modeling. It system involVing ten or twelve independent co-
is assumed that the description given by dis- ordinates represented a massive effort for an
placements of selected nodal points is adequate analyst working by hand With a desk calculator.
to represent the deformation of the structure. Present computer programs routinely handle
If a better representation is needed, the structures represented by thousands of coordi-
structure may be modeled by more nodes and nates. structural analysis can now be used in
hence more elements. The actual loading is re- the design process for preliminary or parame-
placed by a set of nodal forces calculated tric calculations because of this increase in
using the method of virtual work. capability and speed. With on-line graphics
eqUipment, the design engineer can even mani-
The displacement methoQ has the following pulate his design by supplying and receiving
advantages: information directly during computer calcula-
tions.
1. The coordinate transformation matriX is
easy to visualize and easy to generate, being We note here, as an aside, that the Moment
determined solely by the geometry of the system, Distribution Method of Chapter All, the Slope
Deflection Method of Chapter A12 and the Method
2. There is no difficulty in deciding how of Successive Corrections of Chapters A6 and A15
to choose redundant reactions to yield a well- are all examples of the Displacement Method of
conditioned problem, structural analysis. In each of these methods
the displacements are taken as the independent
3. Matrices tend to be banded ones when variables, and these are adjusted to achieve
using the displacement method, resulting in eqUilibrium of the loaded structure. The ad-
economies during computer storage and matrix justment may appear as a systematic relaxation
manipulations, of artificial constraints (Moment Distribution
and Successive Corrections) or it may be done
4. The treatment of structural dynamics mathematically in one stroke by the solution
and structural stability are simplified because of a set of simultaneous equations (Slope De-
the inertial and geometric properties of ele- flection Method). The latter approach-solution
ments and systems transform from coordinate of a set of simultaneous equations-is essential-
system to coordinate system using the same ly that followed herein, the solution usually
matrices as are used in transforming the ele- being effected by matrix inversion or Gaussian
ment and system stiffness properties, el1minat i on.
5. The number of unknowns is based on the A23.2 Structural Coordinates

number of coordinates chosen, While in the
force method redundant forces are the unknowns. It would be impossible to consider the de-
Thus the displacement method is favored for Slections of each particle of a structure. In-
analyzing highly-redundant structures such as stead, we characterize the system by a finite
frames. number of quantities which describe in a gross
* Numbers in parentheses designate References at end of
way (but a sufficiently useful way) the deformed
Chapter. shape of the system. These quantities are called
A23. 1
A23.2 ANALYSIS BY THE METHOD OF DISPLACEMENTS
coordinates, and are of two kinds, discrete and Here again, an exact representation cannot be
distributedo Both kinds involve a simplified given because, for practical reasons, the series
modeling of reality. used is of finite extent 0
Discrete coordinates are used if the When a collection of elements are joined
principal modes of deformation can be identified into a system, there are certain connectivity or
in terms of displacements of a finite number of constraint conditions and boundary conditions
points. which characterize the system, and which distin
gUish it from a mere pile of elements.
For example, with a beam element, we as-
sume that the deformed shape can be adequately To fix these concepts, we consider the ex-
described in terms of the motion of the ends of ample of a simple portal frame whose deformation
the element. For each end there are six com- is assumed to be restricted to the plane of the
ponents of motion, three for linear displace- figure (Fig. A23.3).
ment and three for rotation (Fig. A23.1). The
deformation of the element therefore is des-
cribed only to the degree possible by defining c\
B
t) 03 Oat )Oe
-----+ -----+---'--------' -----+
On t')o.:" ) 0~6
C-
the deflection and rotation of ends This 0
O" 07 15,.0 0H
description is obviously completely adequate if 3 elements
no external loads are applied along the element.
18 coordinates
If loads are applied along the element and
greater detail is desired, nodal points can be
defined at interior locations, effectively
breaking each member into two or more elements.
Fig. A23.3
The constraints and boundary conditions are
1. Fixed end conditions (no deflection) at

points A and D.
2. Kinematic constraints at points Band

Fig. A23.1 C, such that deflections and rotations of ele-
ments 1 and 2 are the same at point B, and of
Distributed coordinates are space functions elements 2 and 3 are the same at point C.
which simultaneously approximate the displace-
ments of all points of an elastic member. For After eliminating" duplicate coordinates, we
example, if we have a simply supported beam, we are left with six system coordinates as shown
might describe it using the ~ functions shown in in Fig. A23.4. Element coordinates are denoted
Fig. A23.2. by ° and structural system coordinates by u,
respectively. (The symbols used generally fol-
w (x) low those of Rubinstein, Refs. 1 and 2. Thus
... - ... - - - .... _ .... ..... --~"
the reader interested in stUdying the theory
'~ x
in greater detail will not need to learn a new
L glossary) .
r==-=-------------------=~<I> ~
u'
u~
--+
ti U'3
u~ t u's u'
'\
~ rB~--=-----C=- - 4
_____- <1>" structure

6 coordinates
u'
--- "-- _ . / ------------ 'Y
th 3
Fig. A23.2 Fig. A23.4
Thus As a practical matter, the kinematic con-

straints are taken into account when compiling
the system stiffness matrix. The boundary con-
ditions may be treated in the same way or, if
A23.3
desired, may be applied all at once in a subse- Various methods have been used; we will illus-
quent step. A further simplification is common trate the procedure by employing the theorem of
in many structural problems. The stresses and virtual work in the case of an elastic rod, and
strains due to axial loads within an element the differential equation of flexure in the
are often much smaller than those produced by case of a beam. The forces and displacements
bending loads. For such cases one may assume of elements will be denoted by P and 0, re-
that the frame members are inextensible. Thus, spectively, as distinguished from forces and
in Fig. A23.4, we may take u~ = u~ = 0, and u~ = displacements at nodal points of a structural
ui, and are left with three generalized coordi- system, denoted by F and u, respectively.
nates for this level of modeling abstraction.
Fig. A23.5 shows the structural deformation pat- undeformed
terns corresponding to these generalized co- rod
ordinates.
(
I
I
I
-I
I
I
(
/
,
\
I
L
\
I
/" '\
I
I
,
\
l?
O~
P...----..I
_x _ ______..,
L~
lJ
[-
Pa
deformed
rod
= U'l.= u~ I I U,a = U'3 U3 = u'e
I " til. I
Fig. A23.6
-,,>- 71~ -,;\-- 77- -,1,-- For a uniform elastic rod SUbject to axial
disolacements O~ and O2 at the ends, the dis-
Fig. .A23. 5 placement at any section x is given by
Three Generalized System
Coordinates for Portal Frame - - - - - (3)
A23.3 Theoretical Basis of the Method of Displacements The total axial strain is uniform, and given by
1
To perform a static analysis of a given ex = L (0 2 - o~) - (4)
structure, we must derive relationships between
external, applied forces JFI and the nodal If the rod is SUbjected to an increase in tem-
points' generalized deflections luI. In par- perature of T degrees as well as an axial stress
ticUlar, since the deflections are to be treated (load), the total strain ex is equal to the
as independent quantities, we seek to derive an elastic strain EX, plus the thermal strain~T.
equation The elastic strain is the quantity needed to
calculate the strain energy, that is,
lF~ = [KJ lu~ - - - - - - - - - - - - - (2)
- - - (5)
wher~ the matriX [KJ is the system stiffness
matrix. (Throughout this chapter, matrix no- Rewriting Eq. 2, page A7.1, in terms of strain,
tation has been employed. The symbols used are using the substitutions GxA = Sand ExE = Gx
defined in Appendix A, except where noted). for the elastic rOd, we find
To give the elements of the stiffness

matrix physical meaning, we may consider an ex-
V = A~L E~ = ~~ [( O 2 - O~) - O(TL r -- - (6)
periment in which one of the u's, say us, is Using the theorem of virtual work (Eq. 14, page
given a unit deflection. The elements of the A7.5)
sth column of [KJ are equal to the forces neces-
sary to give the structure this deflected con- Pr - au - - - - - - - - - - - - - - - (7)
- aOr
figuration, or equally, the elements of the sth
column are the forces developed at each of the so
nodal points when coordinate s is deflected a
p _ AE
aU = - L
unit amount with all other coordinates held ~ -a6~ [(0 2 - O~) - O<TLJ - - -(Ba)
fixed. Within this sth column, the element Krs
is the force developed at coordinate r in that and
deformed configuration. The system stiffness
matrix is compiled by adding together the indi- p
2
= a0
at! = AE
L
[(0 -0 )
2 ~
-o(TL] - - -(Bb)
vidual element stiffness matrices, as will be 2
demonstrated later. where the forces have the same positive direc-
tional sense as O~ and O2 , This information
Stiffness properties of structural elements can be eXhibited in matrix form as
are defined by applying the theory of elasticity
to the particular configuration of interest.
C[1 -lJ {o~} l} __ (9) L2
{ P~}
3
P ,L
~
P
_2_ = EI
:= A + AEo<T { -6- - 2 (14a)
P2 -1 1 02 -1
(14b)
and we recognize the square matrix as [kJ, the _ 12EI
element stiffness matrix for axial deflection which have the solution P~ - - ~ and
of an elastic rod. The elements of [kJ may 6EI
P2 =- 1)"' The values of P 3 and P4 , found
also be calculated directly by rewriting the
Theorem of Virtual Work. Recalling the defi- from the equations of static equilibrium, are
nition of k rs ' it is the force Pr per unit de- 12EI 6EI
P 3 = 13 and P 4 = - L2' These results estab-
flection in coordinate os' Thus
lish the third column of the stiffness matrix.
a2 V - - - - - -(10) The remaining terms are found by a similar pro-
cedure, with the result
This result is of interest also because it 6L -12

demonstrates the s~netry of the stiffness 2
4L -6L
matrix, k rs k sr '=
-6L 12
As a second illustration, we will use the
2L 2 -6L
solution of the differential equation of flex-
ure to derive the stiffness matrix of a beam. The method of superposition provides an alter-
Consider a unit end deflection, 03 = 1, in Fig. nate derivation of the stiffness coefficients
A2:3.7. (ref. Fig. A23.8)
= P 4 L/EI
2
= P 4 L /2EI
Fig. A23.7
The differential equation for a beam is (page

A17.2)
Fig. A23.8
Using the boundary conditions

subject here to the boundary conditions P3L3 P4L2_
°3 = °33 + °34 = 3EI + 2EI - 1 (16a)
P3L 2 + P4L _ 0
at x 0 - - - - - - - - (12a) °4 = °43 + °44 = 2EI EI - - (16b)
we find, as before, P3 = 12EI 6EI

~ and P4 - - ~.
In this case, P~ and P 2 are determined from

static equilibrium. The remaining coefficients
at x =L - - - - - - - - (12b)
are determined in the same way. Note also that
the reciprocity manHest in the symmetry of the
stiffness matrix is demonstrated by the result
After integrating tWice with respect to x, the 043 / = 0 34/ =~, a restatement of Max-
result is P3 P4 2EI
well's reciprocity theorem.
EIw = P~:3 _ P~2+ C~x + C2 -(13)
A23.4 Tabulation of Element Properties
Using the boundary conditions at x = 0, we
find that C~ = C2 := O. The conditions at x = L In each case, the generalized force is
provide two equations in the unknowns P~ and P 2 , taken to have the same positive sense as the
A23.5
associated generalized displacement coordinate P~ A BL -A CL O~

shown.
POl BL 2V~11 -BL L2¢ij °2
A. Bar with Axial Forces = 3E1i
2DL" -BL A -CL
P" -A 0"
,--o~
f---+ I - 0
~r--2
I I P4 CL L2¢ij -CL 2L2¢jj °4
where
A 2 (¢ii + ¢jj + ~ij)

B. Bar with Torsion B = 2¢ii + ¢ij
-~-,-I --.JI-~-
C = ¢ij + 2~ jj
O~ O2
D = ¢ii ¢jj - i ¢ij
{::} and ¢ii' ¢ij and ¢jj are given in Fig. A7.34c
as functions of the ratio Ii/1j, rather than
Ai/Aj' Note that these functions have been de-
C. Beam rived for the case of a beam whose moment of
inertia varies linearly.
E. Bar with Linearly Varying Axial Force
P~ 12 6L -12 6L o~ °2 ..........1 1-0
POl _ EI 6L 4L2 -6L 2L2 O2 l ~ L .1 "
- L"
P" -12 -6L 12 -6L 0"
P4 6L 2L2 -6L 4L2 °4
Note that the stiffness matrix of a beam element

sUbject to a general (three-dimensional) case of
loading at the ends (a 12 x 12 matrix) may be
compiled from the above information. Referring note that the shear flow P~ and the associated
to Fig. A23.1, we associate coordinates 2 and 8 deflection o~ have the units of pounds per inch
with coordinates o~ and O2 of case A, above, and square inches, respectively (Ref. section
and coordinates 5 and 11 with coordinates o~ A7.5). It is difficult to give a physical in-
and 02 of case B. Further, assuming that the terpretation to o~, but we note that the product
transverse coordinate axes are chosen to coin- of P~ and o~ has the units of strain energy,
cide with the principal axes of the beam cross as expected. As an exercise, the reader may
section, the results of case C provide the re- show that the flexibility matrix for this ca"se
maining information for the general case. (Sec.A7.10) is the inverse of the stiffness
Specifically, we associate coordinates 3 and 4 matrix.
(or 1 and 6) with coordinates o~ and 02' and
coordinates 9 and 10 (or 7" and 12) with coordi- F. Shear Panel
nates 0" and 04 of case C. The parentheses ~
indicate the coordinate correspondence for the
t 1
other transverse direction.
04
D. Tapered Beam 1
t0 1
fO" ------r
02e Ij
L ] Ii ) 04 l :5~
a
J
I·
L
~
P1. 1 1 -1 -1 01. o o 0-
1
1 o o
P2 Gt
1 -1 -1 °2
- ab o 1 o
P" -1 -1 1 1 0" o o o
P4 -1 -1 1 1 °4 o o o
o o o
In this case, also, the units are pounds psr
inch and square inches for the "forces" and 1 o o
"displacements," respectively. Note that there
o o o
FJ --
is only one independent coordinate here, which
~
is represented physically by the change in angle o 1
at one of the corners due to the shearing action
of the applied forces. 1 o (18)
o o
G. Other Elements
o o 1
Only the simplest elements have been shown
in this section. Fortunately, many practical o o o
problems can be worked using this information, 1 o o
since a comprehensive catalog of elements could
easily fill a separate chapter. Other elements o o 1
having widespread application are two-dimen- o o o
sional elements with membrane forces (including
triangular, square, and trapezoidal elements), o o o
two-dimensional plate bending elements (includ- o o o
ing triangular, square and trapezoidal elements),
thin shell elements (including conical frusta
and doubly curved elements), axisymmetric thick For convenience, the system [e] matrix may be
shell elements (including ring elements having partitioned into element [?~ matrices, as
various cross-section configurations), and shown. These smaller matrices may be used
general three-dimensional elements (including one at a time to transform element properties,
as will be shown subsequently.
tetrahedron and hexahedron elements). These
and other elements are being developed and per-
fected by structural analysts. It is worth Similarly, we will want to transform
noting that lack of interelement compatibility forces from system to system. Tne method for
is an imDortant s!1ortcoming with the more com- doing this is derived from the expression for
Virtual work.
plex finite elements, and provides the moti-
vation for much continuing work. For details,
the reader is referred to the abundant liter- Consider forces F, associated with the u
ature in this field. coordinate system. For a set of virtual dis-
placements (6 U} , the virtual work is*
A23.5 Transformation of Coordinates
6\-1 = (F} T 16U} - - - - - - - - - - - (19)
We will often need to transform displace-
ments between original coordinates (or coordi- For the ° system,
nates in which it was convenient to define
element properties) and generalized or inde-
pendent coordinates. For this purpose, we use
We also have the compatibility relationship
geometry to define a transformation matrix G?J
given by
- - (17)
derived from the coordinate transformation.
For the portal frame of Figs. A23.3 and A23.5, Now, since virtual work is a scalar, and since
by inspection, we have the pIS and F's are in equilibriwn, and the
o's and u's are compatible, the Virtual work is
the same in either system.
Thus
* \ ) T and [JT denote transpose of column and rectangular

matrices respectively.
A23.7
and since the ~u' 3 are independent so tF*l = [K u ] {U*} + [Ka ] tuo 1 - -(32a)
and rO) = [Ke~] {u*J + [Kee ] tuo } -(32b)
which is the desired relationship. or ~u01 = - [Kee ] -~ [Ke~] [u*1 - - (33)

We next ask how the stiffness and flexi-
bility matrices are transformed from one set of SUbstituting back into Eq. 32a,
coordinates to another. The appropriate formu-
la can be derived from the expression for the
strain energy in an element.
Therefore we may define a reduced stiffness
matrix
with to} =
[fJ lUI' so the first expression may
be rewritten and, once the lU*} are known in any given prob-
lem, the {u o} may be found immediately from
Eq. 33.
Similarly, the flexibility method is use-

and, since the u's are independent, ful in statically indeterminate structures
- (26)
where the opposite situation pertains - that
is, where forces are associated with (redun-
dant) zero displacements.
For transforming the flexibility matrices (de-
noted by [a] and [a] for element and system, We extend fF} to include a set of self-
respectively) a further result is needed. From equilibrating forces corresponding to the
equilibrium, we can find a relationship between number of redundants in the structure. For the
element forces {PI and external nodal point portal frame, there are three redundants, as
forces F 1t shown in Fig. A23.9. The disDlacements at the
redundants are actually zero.
IP} = [bJ {Fl- - - - - - - - - - - - - (27)
which defines [bJ (for an indetenninant struc-
ture, see below). Writing the strain energy of
an element
we can then substitute for }p} using Eq. 27 and Fig. A23.9
derive the final expression
Thus we have
[a] = [bJT [a] [b] - - - - - - - - - - (29)
where we have used the fact that the ~Fl forces
are independent.
and by analogy with the ca8e of zero forces
Often there will be displacements in ~u1
which are associated with zero applied forces - - - (37)
in those· coordinates. Let
where
- - - - - (30)
- (38)
where the ru*} are displacements on coordinates
with applied forces tF*], and the {u o} are dis- The flexibility method is useful also in
placements on coordinates with zero valued deriving a relationship between system forces
a;:Jpl1ed forces. and element forces for a statically indetermi-
nate structure.
Then U:: r!-.L~~ ! K
r } = LKe
0 I
~h!l.J
ee u
f.-u-~-1- - - - - - - - (31)
For the extended set of forces, including Example Problem 1

redundants, we may write
Fig. A23.10 shows the singly redundant
- - (39)
three member frame previously analyzed in sec-
tion A8.4
based upon static equilibriwn. At this point A B C

the (FO} are not known.
)45 0 30°t,.
However, from Eq. 36, we found
T
100"
,
"
.2
or
so
(40)
1
Fig. A23.1O
which is the desired transformation. We have The coordinates chosen are shown in Fig. A23.11.
seen how to transform information about forces,
displacements, stiffnesses and/or flexibilities
from one coordinate system to another. Our next
task is to apply these methods to transform from
element properties to system properties. Rather
than repeat the several derivations, the results
are tabulated below. U2
Table A23.1 u~
Stiffness Method Flexibility

Method
Fig. A23. lIa
System Coordinates
Element
Properties
System
Properties
6"'~ t6,
Transformation
DisPla~~ments (c)~ [0] {uJ

and Forces [FJ ~ [B r lP} o~ "'J/ ° 5
03
[K]~["r[k][~] [a]~[br[Q]H
Transformation
of
Structural
Fig. A23. lIb
Properties or or Element Coordinates
111 T In
[K]~L [~e] [ke][~e] [a]~ L [be][Qe][be] Note that system coordinates at the boundary
e~l e~l points A, Band C could have been included, but
where m = number of elements
would have dropped out when the boundary con-
in structural system ditions were applied. For simplicity, these
coordinates have been deleted at the beginning
In the case of zero force coordinates of the analysis. To compile the system stiff-
ness matrix, the contribution from each element
~F*3 = ([KuJ - [KaJ [K 22] -~ [K2~]) (u*\ must be found. For example, for element OC,
the element stiffness is given by
= [K*J (u*}
In the case of zero displacement (redundant) [kocJ = AocE [ 1
L oc -1
-1] = 1~3[2
1 -2
- - (42)
coordinates
~u*} :: ([auJ - [aaJ [a22]-~[a2~]) IF*1 and the transformation from element to system
= [a*J fF*l coordinates is given by fooc} = [?oc] fu 1

A23.9
or, from Figs. A23.10 and A23.11,
Thus element OC contributes the term,
[K oc ] = [Boc]T [k oc ] [~oc] = 103

E [1.5 0.866J
0.866 0.5
- - - - - - - - - - - - - - - - - (44)
to the system stiffness matrix. Similarly, the Fig. A23. 12b
following information may be compiled for ele- Element Coordinates
ments OA and OB.
Table A23. 2
[K] = [~]T [kJ [~J ' where
r---------,------------------
Element In general,
Quantity
OA OB
'"1 0 0
~[-: -: ]
Element Stiffness [ 1. 414 -1. 414]
E 0 1 0
in Local Coordinates
W -1.414 1. 414
[ keJ 0 0 0
[: ;]
0 0 0
Coordinate
Transformation [-:.707 :.707] 1--- --- ---
0 0 0
[ @e]
0 1 0
Element Stiffness
in System Coordinates
[K ]
e
E
W
[0.707 -0.707]
-0.707 0.707
[ 0
~ 0
[B] =
[::] =
0
r--
0 0
0
---
0
1
---
- - - - (47)
The system stiffness matrix is given by 1 0 0

0 0 1
0 0 0
0 0 0
or '- -
[KJ = ~3
10
r
2 . 207
LO.159
0.159]
3.207
(46) and
which provides the relationship between system

forces and deflections. = - - - - (48)
Exa~ple Problem 2
Assemble the system stiffness matrix for with each eke] equal to the result given in
the portal frame of Fig. A2:3.12a, aSS11,'1l11g uni- case C, section A23.4.
form element properties and neglecting axial
deformation. The element coordinates are shown The final result, after carrying out the
in Fig. A23.12b, and the element stiffness indicated operations (using either the complete
matrix is given by case C of section A23.4. matrices of Eqs. 47 and 48, or the partitioned
U3'---"_ _~ _
form of Eq. 45) is
u~
24
[KJ = EI
L3
6L - - - - - - (49)
L [ -6L
7~
A23.6 Calculation of Displacements and Stresses
I
There are several methods for solving for
Fig. A23.12a the deflections resulting from the application
System Coordinates of given forces. The most straightforward is
to invert the stiffness matrix (for details, We wish to systematize the solution by first
see Appendix A) triangularizing the left-hand side of the
80 that equations, resulting in fewer unknowns in each
successive equation, and culminating in a final
equation in the single unknown x n . The solu-
{u I = [KJ -'- [FJ - - - - - - - (50)
tion is completed by solving first for the last
or for the three member frame of example prob- unknown, xn, then using this information to
lem 1, section A23.5, solve for x n _ 1, and so on. In the triangu-
larized form, each unknown, x r say, depends
only on the xm' With n ~ m> r.
UJl.} = l~a [ 4.547 -0.225]{F'-J - - - (51)
{ Ua -0.225 3.129 Fa The first step in the triangularization
consists of calculating new coefficients for
Given that F,- = 0 and Fa =-lOOOlb., we find the second equation and all subsequent equa-
that the horizontal deflection U1 22500/E = tions using
inches and the vertical deflection Ua == -312900/E
inches. The final step is to compute element i=2, , n
forces. To do this, the element deflections j =2, , n+l - - (55)
must be found using the transformation matrix
[B] or [ee]. For member DC, using Eq. 43, which brings the equations to the form
{:} [0.:55 o -312900

J
0.500]{2~00 ={-13~000}
0
- (52)
E
and {p} - [k] fO}

or
Pel = 1~3 [2 -2] {137EOOOl ={-274 1 __(53) The final form is

{ P6 f -2 2 0 J 274f + C1nXn = C1n+1
(1) (1)
Thus the force in member OC is 274 lb. tension + Can Xn=t;an+1
which checks With the result found previously +C(a)X =C(a) 1
in section 8.4. Similarly, the forces in mem- 3n n 3n+ (57)
bers OA and OB are 335 lb. and 626 lb. respec-
tively.
A23.7 Method of Gaussian Elimination

where at any step, the coefficients are given
For large systems, involving solution for by
tens, hundreds, or, often, thousands of un- k =1,2, ,n-l
knowns, the matrix inversion method is more j = k+l, ,n+l (58)
time-consuming than other techniques. Basical- i = k+l, ,n
ly, the alternate solution methods are divided and (0) _
into two categories, elimination methods, and Cij - Cij
iteration methods (5,6).
The back substitution then takes the symbolic
The method of Gaussian elimination or vari- form
ants such as Cholesky decomposition are widely
used as practical solution methods -for large
systems of equations. The basic Gaussian elimi- Ln Cij
(i- 1) ] _
Xj ' i-n, ... ,1
nation will be described and used here. Con- j=ih
sider a system of linear equations,
- - - - - - - - - - - - (59)
C1 1X1 + C1SXS + C'-3 X3 + + C,-nXn = c,,-n+ 1
CS,-X1 + CsaXs + CS3 X" + + CsnX n = /Csn+;L
C3,-X1 + C3S XS + C33 X" + + C3nXn = C3n+1
- - (54)
A23.11
~xample Problem 3 and the total system stiffness matrix is
Fig. A23.13 shows the sin~le bay pin con-

nected truss which was discussed on page A8.7. 1.7071 -0.7071 -1 -0.7071 0.7071 0
---.--,.--------------.." __ 1000# 1. 7071 0 0.7071 -0.7071 0 -1

e d
1.~)303 0.5303 0 -0.5303 -0.5303
1.b303 0 -1 -0.5303 -0.5303

CKJ = 10' (61)
1.7071 -0.7071 -1
30
1.7071 0
Symmet,icCll
1.5303 0.5303
1 •.5303
Fig. A23.13 Equation 2 may be rewritten as
This truss will now be solved by the displace-

ment method, using Gaussian elimination. The
initial step of the solution is to compile the
{L} R
=0 [1\-"-"_ l1\-"1'
KZ:L Kzz
I
I
I
_'J {_u~l- 0
----- (62)
system stiffness matrix. The required element

where R denotes the forces at the constraints
information is tabulated below. (boundaries), and uC represents system coordi-
Table A23. 3 nates remaining after applying the boundary
cOYlntions. The constrained statics problem
AE is thus obtained simply by deleting rows and
Member eosQ e sinQe columns corresponding to the fixed coordinates.
L
ab 1. 0 x 10 6 1.0 0.0
For this example, the boundary conditions are
u e = u., Us = 0 for all loading conlHtions.
=0
bd 1.0x10· 0.0 1.0 The problem formulation may be simplified by

eliminating the constrained coordinates,
de 1.0x10 6 1.0 0.0 leaving
6
ea 1.0x10 0.0 1.0
6
eb 1. 4142 x 10 0.7071 -0.7071
where
6
ad 1.607 x 10 0.7071 0.7071
1.7071 -0.7071 -1 o -0.7071
Note that G is the angle between the global -0.7071 1. 7071 0 o 0.7071
(system) x axis and the st"uctural axis of a (64)
member. [K"J :: 10 8 -1 o 1.5303 0.5303 0
o o 0.b303 1.5303 0
tyz U4 -0.7071 0.7071 0 o 1.7071
U
~",-u-=-:L t
--,;_ U"
If desired, the reaction forces may be found

from
The solution procedure outlined above can be

Ue modified to account for yielding supports, etc.
UB ~~ ---,,~6 ~ i f desired.
U 7
To solve for the displacements, the method
Fig. A23. 14 of Gaussian elimination will be used. The re-
System Coordinates sult of the triangularization step, correspond-
ing to Eq. 57, is the matrix equation
The transfor~ation matrix for an element is then
given by
Cos Ge Sin Ge]
- - - - - - (50)
=[ -Sin Go Cos Ge
A23.12 AN ALYSIS BY THE METHOD OF DISPLACEMENTS
1.7071 -0.7071 -1 o -0.70'71 A23.8 Gauss-Seidel Iteration
o 1.4142 -0.4142 0 A second class of solution methods involves

0.4142
the use of iteration. Iterative methods of so-
10 6 0 o 0.8232 0.5303 -0.2929 lution are particularly appealing for structures
whose description results in sparse matrices,
o o o 1.1887 0.1887 that is, matrices containing relatively few
nonzero elements. Large space frames and two-
o o o o 1.1588 and three-dimensional solid bodies modeled with
finite elements are examples of such structures.
One widely used iteration technique is the
Gauss-Seidel method. In this method, an in-
itial solution vector is generated, often by
guessing plausible values. This trial solution
(66 ) is then systematically improved, by solving the
linear equations in sequence, using the current
F4 - 0.6442 F" - 0.1887 F2 - 0.4.555 F:I. values for all unknowns other than the one
being evaluated. As soon as an unknown has
Fe - 0.1587 F.. + 0.4580 F" - 0.1588 F2 + 0.6168 F:I. been determined, it is incorporated in the
values used for subsequent steps within a
For the problem to be analyzed, F" is the only given iteration cycle, rather than waiting until
force to be considered. Tnus the right-hand the cycle has been completed. As an illustra-
side for the back substitution process is tion, consider the portal frame of Example
Problem 2, with an applied force F:I. = 24000 lb.
o Asswne values of EI ~ lOa lb. in. 2 and L = 100
inches, and calculate the deflections using
o Gauss-Seidel iteration. After making the
indicated SUbstitutions, and dividing through
F" 1 by the common factor of 8000, the problem may
be brought to the form
-0.6442
~5
75
0.4580 10000 -75 U2 ={3J
2500 ] {u:l.} 0 - - - - (67)
Each row of the matrix equation represents [ 2500
-75 10000 u" 0
an algebraic equation in the unknown u values.
These are solved in inverse order, with the For the iteration, a further reduction is
following results worthwhile, wherein only one unknown appears
on the left-hand side of each scalar equation.
ue = 0.3952 X 10- 6 F" FollOWing this manipulation, the equations to
be solved are
u4 = -0.604'7 X 10- 6 F"
U" = 1.7450 X 10- 6 F" -(68a)
u2 =
0.3953 X 10'-6 F"
u 2 = -0.0075u:l. - 0.25u" -(68b)
u:l. 1.349'7 X 10- 6 F"
=
u" = 0.0075u:l. - 0.25u 2 -(68c)
The static deflections are found by SUbstituting
the given value of F" = 1000. As a trial solution, we take U:l. = 1, u 2 = U" =
0, inasmuch as we expect sidesway to be the
Member forces are determined in the usual way, dominant deflection. The results at selected
by means' of the appropriate U3eJ and eke] steps of the procedure are given in Table
matrices as before. The results are shown in A23.4. The convergence does not appear to be
Fig. A23.15. rapid, but methods exist for accelerating the
O.395x 10-"" solution. For further details, the reader is
~",-,=-1.:..::3--=-5-=-0.:.:x--=-1-=-O~-"_'_'-"r- 1. 745 x 10 -"" referred to a text on numerical analysis.
395.3#
'11* 0.605x 10-""
fCO'r.>'r.>· ;::
'S:Q"", ""
g
'~.jf"'
395.2# 0.395 x 10 -""
Fig. A23.15
A23.13
Table A23. 4
Step of Solution u~ u2 u" (a)

Initial Tr ial 1 0 0
Initial Result of Eq. 68a 1
Ibl~
0 0
Initial Result of Eq. 68b 1 -0.0075 0
Initial Result of Eq. 68c 1 -0.0075 0.0094

(End of First Iteration)
Second Result of ECt. 68a
Second Result of Eq. 68b

1. 4225 -0.0075 0.0094
(c) r = 3 (~)2
m" L\~LJ
-
2 (~\"
~ ~----u,,=l
t
1. 4225 -0.0130 0.0094
[
Second Result of Eq. 68c 1. 4225 -0.0130 0.0139
(End of Second Iteration)
End of Fifth Iteration 1. 8925 -0.0187 0.0189

(d)
End of Tenth Iteration 1. 995 -0.0199 0.0199
Exact Solution 2.0 -0.02 0.02

f (X)
A23.9 Calculation of Nodal Forces for Loads Acting (e)~

Between Nodal Points
In many practical situations, the structure

is loaded at locations other than the nodal
points. Of course, in the case of concentrated
b x
L .1
loads, this complication may be resolved by in- Fig. A23.16
troducing nodes at the loading sites. However,
for distributed loads, such as pressure, there By analogy, the total virtual work is also
is no way of avoiding the problem. Therefore, given by calculating the work done by the nodal
we wish to derive a method for calculating forces, Pi' acting through the compatible nodal
equivalent nodal forces which are consistent virtual displacements, ~ui' The result is
with the finite element formulation of the prob- 4
lem. To do this, we employ the method of vir- ~W = L Pi~ui - - - - - - - (72)
tual work. As an example, consider the case of i=l
an elastic beam element. Associated with each
nodal coordinate, there is a deflected shape or and since the Llui are independent, the right-
distributed coordinate for the entire finite hand sides of Eqs. 71 and 72 may be equated
element, found from the solution of the differ- term by term, giVing the desired equation,
ential equation of beam bending. These shapes
are shown in Fig. A23.16.
Pi = jL f(X)~i (X)dx, i = 1, ... , 4 - (73)
ThUS, from Eq. 1, the connection between o
nodal coordinates and the deformed shape of the In the case of moments acting between nodal
element is points, a parallel development gives the re-
sult
Pi =jLm(X)~i (X)dx, i = 1, ... , 4 - ('74)

Let the applied load be given by f(x) as shown o
in Fig. A23.16e. Tnen the total virtual work
produced by the applied load acting through a where m(X) is the applied moment and ¢i =
virtual displacement ~w is d¢/dx.
tM =j L f (X) ~w (X) dx - - - - - - - - (70) Several loading cases occur often enough

o to justify tabulating the associated nodal
forces. Table A23.5 contains the nodal forces
or substituting from Eq. 69, appropriate to the loads shown in Fig. A~3.17.
4 L
I1W= [ j f(X)~i(X)dx~ui - - - - - (71)
i=l 0
model is required as well, relating to both the

t P~ conceptual and detail models, and reducing the
~~
(a) structure to a properly formulated mathematical
L!2 problem. The nwnbers produced by the solution
of the mathematical problem are then used to
verify the design concept and to provide a
rational basis for redesign, where necessary.
I Finite element modeling is one method for pro-
(b)
~.----'-----~J
L!2 ducing a mathematical description of the struc-
ture. The next two sections comprise several
representative exmnple problems, intended to
(c)
provide the reader with an appreciation for the
nature of typical finite element models and the
kinds of results which can be obtained with them.
2PX
I~
Example Problem 4
(d)
Reconsider the truss of Fig. A7.16, repro-
I. L ~ duced in Fig. A23.18. A truss is defined as a
pinned structure wherein moments are not trans-
Fig. A23.17
Table A23. 5 1000# 1000#
Case P", Pz P" P4

~-----r----~l
tc ~D
(a) Con::entrated P!2 PL!S P!2 -PL!S
Force at
Midspan
40"
~~----jJ
(b) Concentrated 3M M 3M M
Moment at -2L -4 2L -"4
Midspan
(c) Uniform
Load
P!2 PL!12 P!2 -PL!12
~30"+30"~30" H
(d) Triangular 3P!1O PL!15 7P!1O -PL!lO 2000# 2000#

Load
Fig. A23. IS
Note that these nodal forces are related to the mHted between members at the joints. This
fixed end moments used in the moment distri- idealizati on is seldom met, and t he truss is,
bution method of Chapter All. The moments P2 in reality, a frame composed of relatively
and P4 are fixed end moments, while the re- slender members. The structure shown will be
actions P", and P" may be obtained by statics. reanalyzed as a frame, and the results com-
For other cases of interest, the reader is re- pared with the truss ~nalysis presented in
ferred to the information displayed in Chapter section A7.7. The previous problem statement
All, Table All-l, herein, from which tile ap- does not include data on section properties
propriate nodal forceR can be derived. Loadings for the members of the structure. Calculated
not appearing in the table can often be obtained values have been supplied, based on a plaus-
by superimposing the given cases. ible design philosophy. Assuming that the
material is alliminum, having a modulus of
A23.1O Structural Modeling Using Finite Elements elasticity E = 107 psi, member areas can be
found from Table A7.3. A typical aircraft I
The process of structural modeling occurs section is shown in Fig. A23.19, along with
at several levels of abstraction during the dimensions given in terms of the material
course of any design. Typically, one first has thickness. For this section, we have
a conceptual model, selecting the type of struc-
-.r
ture necessary to transmit the applied loads
from among many possibilities, including trusses,
ca~les, frames, arches, and shells. A more de-
tailed model is necessary when making the design
'------'1 20t
drawings. One must detail connections and
brackets, size members based on available sheet
metal gages, and rolled or extruded sections,
actu.al material properties, etc. An analytical Fig. A23. 19
'--1
4
--2-0t-~(Tt
J
A23.15
the relationships analysis, a difference of 58 per cent. The

structural designer must be aware of the possi-
A = 60 t 2 - - - - - (75) bility of this effect, and examine its sig-
nificance for the particular structure being
and I = 5080 t 4
_ -(76a) analyzed.
or I = 1.4.'39 A2 - - - - -(76b)
Example Problem 5
which have been used to calculate the moments of
inertia given in Table A23.6. Tne redundant portal frame of Fig. A23.20
is to be analyzed by the method of displacements.
Table A23. 6
L A I 1.50
Struss Sframe Mframe
Member . 2 4
in. m. in. lb. lb. in. -lb.
AB
BC
CD
30
30
30
0.4785
0.3074
0.3074
0.323
O. 130
O. 130
10500
2250
2250
10500
2262
2244
208
692
95
-
1000# B
2.25 2.25
I
C
50"
EF 30 0.5365 0.406 -5250 -5239 732 A Dl

FG 30 0.5365 0.406 -5250 -5225 204
Fig. A23.20
GH 30 O. 348 O. 171 0 10 355
The relative bending stiffnesses of the segments
BE 50 1. 015 1. 45 -8750 -8728 805 are given on the figure. The frame is modeled
using beam elements, whose stiffness properties
BG 50 0.348 0.171 5000 4939 120 are given by case C of section A23.4. The curved
portion of the frame is apprOXimated by 30
DG 50 0.5365 0.406 -3750 -3723 86
straight beam element~ (chordS), connecting
BF 40 0.3074 O. 130 2000 1981 348 equally spaced nodal points located on the
elastic axis of the curved portion. Here the
CG 40 0.3074 0.130 -1000 - 965 387 model contains 32 elements, and, after accounting
for connectivity and boundary conditions, there
DH 40 0.3074 O. 130 2000 1980 252 are 93 generalized coordinates in the system.
A computer solution was performed, again using
A finite element model of the structure the Gaussian elimination method. As a check,
has been used, with each member represented by the computed values of reactions at point A have
a six by six stiffness matrix. This stiffness been compared with results of an analysis given
matrix is found by combinIng results of cases A in section A8.2.2. The equations obtained in
and C of section A23.4 to provide the relation Chapter A8 are satisfied by the values M =
between forces and displacements in one ro- -27650 in.-lb., V = 801.3 lb., and T = 114.6 lb.,
tational and two translational coordinates at with the sign conventions given in Fig. A8.3g,h,
each end of the member. Tnere are 12 elements i. The comparison is given in Table A23.7,
in the structure, connected at eight nodal along with results for three other models, in-
points. Thus the connectivity conditions at volving 15, eight, and four beam elements to
the joints produce a reduction from 72 element approximate the curved portion of the portal
coordinates to 24 system coordinates. Finally, frame, and one element for each column.
applying two boundary conditions each at nodes
A and E, one finds there are 20 generalized Table A23. 7
coordinates for this analysis. The system
M VA T
description and loads were read into a digital Model
A A
computer, programmed to solve the problem using in. -lb. lb. lb.
the method of Gaussian elimination.
Analytical
The results of the analysis are shown in -27650 801. 3 114.6
Table A23.6. The effect of neglecting end fix- (Sec. A8. 2. 2)
ity is slight for the axial forces. However, 32 elements -27627 801. 13 114.92
the inclusion of induced bending moments in the
calculation of outer fiber stresses produces 17 elements -27589 800.68 115.34
large differences for several members. For ex-
ample, the maximum stress in member BC is given 10 elements -27461 799. 19 116.78
by (S/A) = 7320 psi for the truss analysis and
by (S/A) + (Me/I) = 11540 psi for the frame 6 elements -26955 793.29 122.78
The bending moment is plotted as a function of

position on the frame in Fig. A23.21. This
gives a good approximation for the spatial vari-
ation of outer fiber stresses, since bending is
the predominant type of deformation in this
problem.
-4801 In.-lb.
12429 in. -lb.
Fig. A23.23
12 @ 2" ~I
-27627 in. lb. ///// 10881 in. -lb. kk/
T
3 @ 2"
Case 1
36 elements
52 nodal circles
Fig. A23.21
~
Example Problem 6
f-- - - - - - - - - - - - - -
The thick walled pipe of Fig. A23.22 is to
be modeled using finite elements. The dimen-
sions of the pipe are) = 24 in., ri = 4 in., ~ Case 2
=
and r o 10 in. 6 @ 1" 144 elements
175 nodal circles
~
x
x
24 @ 1" .1
Fig. A23.24
plane of the free end. These end stresses

were set equal to the theoretical values at
the centers of the element faces for a re-
sultant shear load of 4200n lb., equivalent to
an uniform vertical shear stress of 50 psi.
The interior stress distributions have been
Fig. A23.22 derived from the theory of elasticity (7)
and are plotted along with computed values in
An applied shear load acts at the free end Fig. A23.25 and A2:3.26. Reported values for
of the pipe with the resultant force parallel each station represent averages of values for
to the x axis. The model consists of annular the adjacent elements, referred to the mid-
elements having a squ"cre cross section. A point of the bounding face. In the derivation
typical element and its associated coordinates of these theoretical values, the nature of the
are shown in Fig. A23.23. The stress components end condition at the wall is neglected. Thus,
are also defined in Fig. A23.23. (For a more the two results should show good agreement at
detailed discussion, see Chapter A24). Two intermediate stations, but not near the wall.
grid sizes are used in the analysis, as shown These features are observed in Figs. A23.25
in Fig. A23.24. and A23.26. Near the wall, the finite element
model gives more realistic stress values,
The loading is applied by means of equiva- based on the actual boundary conditions. This
lent shear stresses (~zr and ~ze) acting in the is often true, especially in cases where the
A23.17
150,-----,----,------,--------,-----,------,
system is sufficiently complex to preclude an
Solution Symbol Distance from Load analytical solution in closed form. Note that
Elasticity the coarser nodal grid gives results comparable
Theory in accuracy to those achieved with the refined
DiSPlacementE 6 Inches division using 144 elements. The decision re-
Method ~
36 Elements
1--- - -1- - -
1:'------)1,
-
12 Inches
18 Inches
garding how many elements to use must be based
on experience with comparative studies of this
sort.
100
A23. 11 A Delta Wing Example Problem*
The idealized delta wing structure of Fig.

A23.27 will be analyzed to determine the influ-
~ .. ence coefficients and the internal stresses as
functions of the applied loads. This problem
will also illustrate the use of SUbstructures
in analysis.
50
10
Radial Distance - Inches
Fig. A23.25
Fig. A23.27
150 ,---~---~---~---~----r----...,
Idealized Delta Wing Structure
Solution Symbol Distance from Load
Elasticity
Theory IDEALIZATION
Displacement { 6 Inches
Method 1---- - -1-- - - , 12 Inches
The delta wing has been idealized into a
", , 144 Elements "t--~- ~ 18 Inches grid-like structure having grid points numbered
as in Fig. A23.28. Grid numbers increase to
the rear and outward.
100
4 7 10
~ -~
~
~,~--~~-
'"'" ',------
Q)
H
1i'i
2 r-------;Y
50
~_---"--..... ......... ~ jTzratg=oo 1 Fig. A23.28

Grid Numbering Scheme
"
The torsional stiffness of the wing's shear-
carrying cover sheet is accounted for by the
presence of quadrilateral "torque boxes" con-
necting points of intersection of the spars and
Radial Distance - Inches
ribs. The resistance of a torque box having a
corner at, say, point 6 of the grid (e.g., box
Fig. A23.26 * By Alfred F. Schmitt
6-7-8-9) is a result of the tendency for the STRUCTURAL PROPERTIES SUMMARIZED

box to keep all four corners in the same plane.
Obviously, three-cornered "boxes" have no such The properties of the structure (Fig.
resistance; hence none are shown in Fig. A23.27. A23.27) after idealization, are summarized as
follows:
In the structure under consideration mem-
ber 1-2-3-4 lies on the airplane centerline. Net Point Beam Depth
The bending stiffness of this member includes (see Fig. A23.28) (inches)
that of the half-fuselage plus one-half of the
"carry-through structure." More detailed 1 5.56
representations of the structure in this region 2 9.12
are generally desirable, the oversimplified
model used here being employed to limit the 3 7.80
amount of data to be handled. Some techniques 4 - - - 3.16
which may be applied in idealizing the structure
in this region are given in Reference (II) 5 - - 4.88
6 7.26
Idealization, particularly with regard to
effective skin areas, has been discussed in 7 - - - - 3.06
some detail by Levy (8). The complexities of 8 4.02
this phase of the problem are too great to per-
mit an expansive treatment here.* Briefly, 9 2.74
Levy's recommendations are: 10 - 0.72
(a) all skin may be considered effective All cover skin t = .051"
between spanwise spars when computing the cap
areas of such spars. This assumption is sub- BEJU1 ELEMENT PROPERT IES
ject to modification, of course, if spanwise
stresses are anticipated which will buckle the (Moments of inertia, in (inches 4 ) , are
skin. assumed to vary linearly between numbered
points)
(b) for streamwise ribs an effective width
of 0.362L, where L is the rib length, may be 1/2 Fuselage Beam (1-2-3-4) Spar 2-5
taken as acting with each rib cap (Fig. A23.29a).
I 3@48"---l t--
I
48"----'
,
(c) for the leading edge spar an effective
width of skin of .181 of the span between span- F = -1 ! 1 1 11
I = 2000 2700 3400 2200 64.91 18.25
wise spars is taken as acting with the cap of
each such spar segment (Fig. A23.29b).
Rib 5-6-7 Spar 3-6-8
r-48"-----r- 48"-::
I ! f= . I i
20.79 3.59 47.23 40.82 12.3~
(b)
Rib 8-9 Spar 4-7-9-10
Fig. A23.29
Effective Widths of Cover Sheet for Rib and I-- 48"- "'---3@48"
Oblique Spar Caps (After Levy, Ref. 8) I I
£ 1 L is },
I = 6.23 2.87 5.09 4.78 3.82 0.26
Leading Edge Spar (1-5-8-10)
I
[email protected]'1--'l-----1
* A rational, systematic means of treating cover sheet 1 I £ !
panels is given in Reference 10).
I = 16.02 12.29 8.30 0.26
A23.19
SIGN CONVENTION 0.003954 0.1136 -0.003954 0.07616
The sign convention adopted in conjunction 0.1136 4.015 -0.1136 1.438

with the grid numbering scheme of Fig. A23.28 is
as follows: 0.003954 -0.1136 0.003954 -0.07616
Element Forces, Pm 0.07616 1.438 -0.07616 2.217
Transverse Force at a joint, where we have taken E = 1 for convenience. It

is understood that forces and displacements of
+UP
the beam element (spar 2-5) are applied in posi-
tive sense and arrayed in numeric order as in
or, Pitching Moment at a joint, Case D, Sec. A23.4, when the beam element is
viewed with grid numbers increasing left to
+ NOSE DOWN right.
or, Rolling Moment at a joint, In like manner, the element stiffnesses for
the remaining fourteen (14) beam segments (spar,
+ RIGHT WING DOWN rib and leading edge spar segments) are com-
puted. They are not tabulated here for reasons
or, Moment on an Oblique Member, of space conservation.
acting in the plane of that member, Torque Box Stiffness Matrices
+ COUNTER CLOCKWISE WHEN THE MEMBER IS
In the case of the torque boxes, the stiff-
VIEWED WITH ITS JOINT NUMBERS IN- ness matrices may be computed directly (they are
CREASING LEFT TO RIGHT. determinate). The following approximate analy-
sis is suggested as being satisfactory for most
Element Displacements, Om torque boxes. A more detailed analysis may be
in order for boxes with extreme geometries. An
Follow same numbering and sign convention alternate method is presented in Reference (8)
as do corresponding forces. and an appropriate discussion may be found in
Reference (12). A completely different treat-
With the above sign convention any beam ment of the cover skin is given in Reference
element (spar or rib) which is viewed so that (10) .
its joint numbers increase left to right will
have forces and displacements taken in a sense Consider the quadrilateral box of Fig.
compatible with the beam element tabulations of A23.30.
Sec. A23.4.
\'1----- L n - - - - :
CALCULATION OF ELEMENT STIFFNESSES
The element stiffness matrices are com-

puted next for the beam segments and torque
boses. For the beam segments the equations of
Case D, Sec. A23.4 are used. The torque boxes
are treated expressly, below.
=-'-\-----,
-- Lm
'.
Beam Segment Stiffness Matrices Representative
EqUivalent
Root Section
For the beam segments the spar between grid
points 2 and 5 typifies the calculation. From Fig. A23.30
the beam element properties given earlier, and
Fig. A7.34C For purposes of this immediate analysis, the
box is considered cantilevered from one end,
Ii/Ij = 0.281 such as r-s. An elastic axis (e.a.) is as-
sumed to exist midway between the long sides
~11 = 0.670 and an "effective root" is employed. The
<!lij = 0.480 ~ij/2 = 0.240 torque on the box is
~jj = 0.370 T = Pm bm - Pn bn
Then D2_ 5 = (0.670) (0.370) - (0.240)2 = 0.1903
An (assumed) average rat;'J of tWist, fl, is com- 17 of this chapter. Then using equation 26 we
puted approximately by using the GJ at a repre- obtain the square stiffness matrix
sentative section half way along the box.*
mnrs
_ T _ Pmbrn - Pnbn
fl- GJ - GJ
Then the deflection on and Om are given by
= Pmbm GJ- Pnbn Lnbn
H+ GJ Ln/Lm 1
(Lnbm+Lmbn)(bn+bm) -Lm/Ln
In the above derivation the torsional

stiffness GJ at a representative section is
By summing moments about the effective root, Pn used. The stiffness is obtained from Bredt's
is found in terms of Pm and ttwn eliminated equation for the tWist of a single cell thin-
from the above equations to give walled tube (Eq. 18, Chapter A6)
T ,[ ds
fl = 4A 2
G ) T
~Il = PmLmbm Lnbm + Lmbn from which, by definition
~ Ln
_ 4A 2 G
Since it is only the deflection of one
corner (say, point m) which is to be related to
GJ
-¢- ;s-
the corner reactions, the box is now rotated
about the effective root axis to reduce on to Here A is the area enclosed within the tube
zero. The resulting total deflection of point cross section by the median line of the tube
m is wall and the integration is carried out around
the tube perimeter (index s gives distance along
0illTOTAL ::: ~n + ~ on the perimeter). For the torque boxes encoun-
tered in delta wings it is probably satisfactory
= PmLm (Lnbm + Lmb n ) (b n + bm) to neglect the ds/t contribution from the verti-
GJ Ln cal webs, it being small compared with the cor-
responding contribution from the cover sheets.
Re-solving this, we write
GJ Ln In the example wing three boxes (2-3-5-6,
Pm ::: Lm(b n + bm) (Lnbm + Lmbn) 3-4-6-7 and 6-7-8-9) are to be used. These
boxes are each 48 inches square in plan and
This last expression relates the load at have average depths (assumed here to be the
point m to the deflection at that point, with uniform depths of the representative sections)
the other three corners undeflected. Thus the of 7.26", 5.32" and 4.27", respectively. Fig.
deflection of point m as defined was relative A23.31 shows the assumed representative cross
to the plane mnr s. This relative displacement sect Lon of box 2-3-5-6 and its GJ calculation.
may be related by geometry to the case where
all four corners are displaced (in effect, re-
moving the rigid body displacements). r '-t=.051 i ~ r~AV=7.26"
fc~~
~mrelative
::: L 1 L~/L
-~f'l n 1--- 48" ~- ~. _~ I
A = 7. 26 x 48 = 348 in 2
Note:
f ds 48
t = 2 x . 051 = 1. 88 x 10
" G=~
2.6
_ 4(348)2. 1 _ . 2
where GJE=l- 1.88 x 10"x 2.6 - 99.3lb.lil.
_ Lnbm + Lmb n + bs (Lm - Ln ) (ds/t contribution of vertical webs neglected)
A
Ln(b r + bsl
Fig. A23.31
A
-(Lnbm + Lmb n ) + br (Lm - Ln ) Calculation of GJ for the Representative
- Ln (b r + bs)
Section of Box 2-3-5-6
This last relation corresponds to Equation

In similar fashion one finds:
For box 3-4-6-7
* Calculation of this GJ is discussed later. GJE=l = 53.3
A23.21
For box 6-7-8-9 One now writes the relations which trans-
form the displacements. As an example, one
GJE=l =34.3 writes
Finally, the stiffness matrices for the u
1 2 3 4 5 6 ----- 11 12 13 14 15 16 ----- --29 30
three boxes of the example become: 6
0 1 0 0 0 o ----- 0 0 0 0 0 0
t3 1f- " ;;;:
For box 2-3-5-6 2 0 0 0 0 0 o ----- 0 0 0 0 0 0

3 0 0 0 0 1 0 ----- 0 0 0 0 0 0 0 0
o -----
l
-~ -~]
4 0 0 0 0 0 0 0 0 0 1 0 0 0
-1 -1
1 1
0.000898 Then
-1 1 1 -1
1 -1 -1 1
For box 3-4-6-7

2 5 12 15
[ -~ -~~
-1 -1
2 .003850 -.003850 .n08 .07404
1 1
.5 -.003850 .003850 -.n08 -.07404
[KJ = 0.000482 -1 1 1 -1
12 .n08 -.n08 3.929 1.387
1 -1 -1 1
1.5 .07404 -.07404 1.387 2.167
For box 6-7-8-9 Here for economy we have written only the non-
zero terms of the transformed element stiffness
[ -~ -~]
-1 -1
matrix. These simple transformations, involving
1 1 no coordinate rotations may be written by in-
0.000310 spection since obviously, for element 2-5
-1 1 1 -1
1 -1 -1 1 6~ ----.. U 2
62 ----.. u~ 2
SYSTEM DISPLACEMENTS/TRANSFORMATION OF 63~U6
COORDINATES
64 ----.. u~ 6
The element displacements of the foregoing

element stiffness matrices may now be related For the leading edge spar there are geo-
to (transformed to) the system displacements metric transformations. ThUS, for element 1-5,
chosen for the overall wing structure. Fig.
A23.32 shows the system displacements which may
be chosen. )z 1 5 n 15 21 25
1 1 0 0 0 0 0
2 0 0 .707 0 .707 0
3 0 1 0 0 0 0
4 0 0 0 .707 0 .707
Where, again, only displacements haVing non-

zero entries are shown explicitly.
Then
Fig. A23.32 i~ 1 5 11 15 21 25
Delta Wing System Displacements. 1 .0005295 -.0005295 .01326 .01214 .01326 .01214
All rotations positive by right-hand rule. = 5 -.0005295 .0005295 -.01326 -.01214 -.01326 -.01214
11 .01326 -.01326 .4361 .2004 .4361 .2004
15 .01214 -.01214 .2004 .3821 .2004 .3821
21 .01326 -.01326 .4361 .2004 .4361 .2004
25 .01214 -.01214 .2004 .3821 .2004 .3821
Proceeding in the above manner, stiffness matri- found on span 9-10. This situation suggests
ces for all 15 bea~ elements are transformed. that a single beam, continuous across grid
The torque boxes are already in "system" co- point 9, be used instead of the separate ele-
ordinates since each was designated by the grid ments.
numbers of its four corners.
The sUbstructures selected for this prob-
ANALYSIS BY SUBSTRUCTURES lem, along With their element deflections, are
shown in Fig. A23.33. They are the rather
From this point on the analysis proceeds in natural choices of the continuous wing beams
a very direct manner as previously illustrated. (spars and ribs) with only the minimum neces-
However, as is also apparent, any large struc- sary deflections considered. While the fore-
ture having many nodal points is going to yield going reasoning had dealt largely With forces,
a very high order matric equation when analyzed it is the corresponding element displacements
in such a completely elemental fashion. Which are used as the variables and are so-
shown in the figure.
It is possible, however, to reduce the
order of the problem by some preliminary re-
combining of the elements into sUbstructures in t 0" tO 2 t O~ to.
the course of which one can Beam 1-2-3-4 I I I
(a) use certain a priori information to t0 6 t0 7
Rib 5-6-7 I
eliminate some loads (and corresponding dis- I
placements) t Os
Rib 8-9 I
(b) extract some portions of the problem
out of the main calculation, putting them into
sUbproblems solved on the substructures.
These devices will be employed in the following
section. Spar 2-5 (e-I_--II) 0".
0 ••
Certain of the external loads associated f O~ to 6 t0 6
with the displacements of Fig. A23.32 are known

Spar 3-6-8 (l-I_--'-_-----'
I 1)6. 6
O.~
to be zero by nature of the load modeling used
in this problem. Thus, only transverse forces, to. , 07 t Os t 0"0
Spar 4-7 -9-10 t I t
F 6 through F~o, are to be applied as loads at
the wing grid points, reacted by transverse
loads F~ through F4 and rolling moments F~~ Fig. A23.33
through F~4 along the airplane centerline.
Substructures and Their Element Displacements.
Note numbering of displacements has been kept
For this structure and this loading it is consistent with system numbering of Fig. A23.32.
then known a priori that many element loads are
zero. For example, there can be no couples on
the beam elements at grid point 10*. It is, in The SUbstructure of spar 4-7-9-10 illus-
effect, as if this joint was a pin joint and trates the motivation in the SUbstructure
only its translation was considered. variable choices. It was modeled as shown
since it was known there were no applied
Many element loads are not zero but need couples at grid points 7, 9 and 10 and hence
not be solved for explicitly in the main prob- the rotational displacements at those points
lem if they can be computed sUbsequently by need not be considered. Elimination of
analysis of a substructure within which they are variables at no-load points can, of course be
subsumed. For example, the bending moments on done on the final complete system stiffness
the beam elements 7-9 and 9-10 meeting at grid matrix (obtaining a reduced stiffness matrix
point 9 are equal and may be found easily once per Equation 35). Such a reduction may in-
the transverse element load at point 10 is volve manipulation of very large matrices.
Using substructures this reduction process is
divided into several smaller, more tractable
* The implication here is that beam torsional stiffnesses are problems while simultaneously reducing the
not considered in such interactions. This assumption is order of the main problem. The rotational
probably quite satisfactory in general, bending stiffnesses
being much greater than torsional stiffnesses for most displacements thus eliminated can be easily
beam elements. On occasion, a beam will have considerable obtained via a subcalculation, if desired,
torsional stiffness and it will then be necessary to account (ref. Equation 33).
for it. The leading edge spar might be such a beam if it
replaces the "0" nose section of the Wing in the idealization. Couples must be considered to act wher-
ever three or more nonorthogonal beam elements
A23 23
intersect to account for their interaction; Adding these together gives

hence the rotational displacements at grid
points 5 and 8 on all ribs and spars intersec- 3 6 8 13 16 18
ting there and at grid point 1. (The other wing
half provides the third structure at that point.) 3 .004746 -.004746
/
.1166 .1112
6 -.004746 .007304 -.002558 -.1166 -.0382 .04974
Looking ahead we see a complete system Ka _ a _ a 8 -.002558 .002.558 -.07305 -.04974
stiffness matrix of 19th order (u~ through U~6
13 .1166 -.1166 3.775 1.819
plus u~a, U2~' US6 and u 2a ) instead of 30th
order. This matrix can be subsequently re- 16 .1112 -.0382 -.07305 1.819 6.095 .9325
duced in order because of five known zero ex- 18 .04974 -.04974 .9325 1.455
ternal loads (there are no externally applied
moments at the points where the nonorthogonal (Voids denote zeroes)
beams intersect). This reduction will require
inversion of a 5th order matrix rather than the The constraint of generalized force u
16th order matrix otherwise required to reduce being zero is now applied using Equation 35.
the 30th order problem to an equivalent system The inverse "Ks;~" in this instance is merely
of equations. 1/6.095 and the matrix "Kss " is the row
(.1112, -.0382, -.07305, 1.819, 0.9325).
SUBSTRUCTURE STIFFNESS MATRICES Applying the equation leads to
Calculation of the substructure stiffness

matrices follows the procedure established for 3 6 8 13 18
any structure built up of elements. For ex- 3 .002718 -.004049 .001332 .0834 -.0170
ample, consider beam 3-6-8 which comprises ele-
ments 3-6 and 6-8. The sUbstructure is shown K*3-e-e = 6 -.004049 .007065 -.003016 -.1052 .05558
below With all displacements indicated, in- 8 .001332 -.003016 .001683 .02179 -.0385 7
cluding 6~a, even though it is to be eliminated 13 .0834 -.1052 .02179 3.232 -.2782
because the applied load P~a is known to be
18 -.0170 .05558 -.03857 -.2782 1.312
zero.
The foregoing procedure is followed to

obtain reduced stiffness matrices for each of
the remaining SUbstructures.
COMPLETE WING STIFFNESS MATRIX
The stiffness matrix is collected from
that of each of the elements 3-6 and 6-8. The stiffness matrix for the composite wing
now may be obtained by forming the sum of the
3 6 13 16 complete stiffness matrices for the beam sub-
structures and the torque boxes. For this pur-
3 .004746 -.004746 .1166 .1112 pose a large matric table is laid out and en-
tries from the individual stiffness matrices
6 -.004746 .004746 -.1166 -.1112 are transferred into the appropriate locations.
K
a-a = 13 .1166 -.1166 3.775 1.819 Wherever mUltiple entries occur in a box these
are surmned .
16 •1112 -.1112 1.819 3.521
Before the large wing matrix is laid out
6 it is necessary to observe first that the
8 16 18 matrix which would be obtained as just indicated
6 .002558 would be singular, i.e., its determinant would
-.002558 .07305 .04974 be zero, and hence it could not be inverted to
8 -.002558 .002558 -.07305 -.04974 yield the flexibility influence coefficients
K
a-a = 16 .07305 -.07305 2.574 .9325
(see Appendix A). This condition arises due to
the fact that the equations represented (19 in
18 .04974 -.04974 .9325 1.455 number in the exa~ple problem) are not inde-
pendent; three of these equations can be ob-
Adding these together gives tained as linear combinations of the others.
That there are three such interdependencies

may be seen from the eXistence of the three
equations of statics which may be applied to
the wing (surmnation of normal forces and sum-
mation of moments in two vertical planes):
hence three of the reactions expressed by the ficients are those for only the half-wing
structural equations in the matrix may be found acting alone and supported by constraints as
from the others by the equations of statics. asswned above. To account for the presence of
To remove the "singularity" from the stiffness the other half of the wing, 'it is necessary to
matrix it is only necessary to drop out three specify additional geometric conditions along
equations - achieved by removing three rows and the airplane centerline. This step is accom-
corresponding columns (so as to retain a sym- plished by asswning the following deflections
metric stiffness matrix). zero (eliminating their corresponding rows and
columns from the matrix):
The act of removing the three equations
selected is also equivalent to assuming the u~~ for symmetric loadings (zero
corresponding deflections to be zero. In this lateral slope or rolling
way a reference base for the deflections is also u~"
established. The choice of reference base is
somewhat arbitrary, but, following a suggestion
of Williams (13), a triangular base will be
U~4
U"
J rotation along the airplane
centerline)
employed as shown in Fig. A23.34.

u4
for anti symmetric loadings (zero
U2~
\ I ' transverse deflections and pitch-
U22 ing rotations along the airplane
centerline
u 2"
U 24
It will be seen that in both cases addi-

tional equations are eliminated from the 16 x
16 matrices further reducing them. In general
Fig. A23.34 there is no reason Why the matrices for the
Deflection Reference Base symmetric and antis~nmetric cases have to be
the same size since the modeling of applied
Here the deflections at points 1 and 5 are zero, loads is not the sane in these cases (for ex-
fixing the reference triangle, since the point ample, no forces F~, F 2 , F" or F4 are applied
corresponding to 5 in the other half of the wing in an antisymmetric loading).
(say, 5' ), will have zero deflection also due to
symmetry. The third condition is applied to Written below is the 13 x 13 wing stiffness
point 2: point 2 will be assumed to have zero matrix for the symmetric case (u~ = u e = u~~ =
rolling rotation (U~2 = 0) for symmetric wing U~2 = u~" = U~4 = 0). As explained earlier,
loadings and to have zero transverse deflection each entry therein is the sum of all correspond-
(U2 = 0) for antisymmetric loadings.* ing stiffnesses for all elements meeting at the
point. A typical mUltiple entry occurs for
Hence the following equations (rows and joint 6 - (row Fe, column u e ). These comprise:
columns) are to be omitted from the wing stiff-
ness matrix: 0.007065 from spar 3-6-8
0.001730 from rib 5-6-7
symmetric loadings 0.000310 from box 6-7-8-9

0.000898 from box 2-3-5-6
0.000482 from box 3-4-6-7
J for antisymmetrie loadings

0.010485 TOTAL
The 16 x 16 wing stiffness matrices thus

obtained can now be inverted as they are non-
singular. However, i f this step is carried out
one finds that the resultant influence coef-
* The antisymmetric loading pattern is one wherein the wing

is loaded equally, but in opposite sense, on corresponding
points of the two wing halves. Any general loading may be
resolved into the sum of one symmetric and one anti-
symmetric loading.
A23.25
U2 Uo U. U. U, U. U. U. O U.. U.. U •• U •• u ••
0.4299 -.2670 .07314 . 000898 .07616 -5.393

-.2670 .2753 -.1027 -.005429 .000482 .001332 -.01700 1.449
.07314 -.1027 .04629 .000482 -.0009328 .0001586 -.00002122 -.2450
.000898 -.005429 .000482 .010485 -.001380 -.003326 .000310 .05558 -.02658
.000482 -.0009328 -.001380 .001776 .000310 -.0006789 .00008089 .004706
.001332 -.003326 .000310 .002564 -.0004515 -.0000499 -.009686 -.04470 -.009686 .000664
Kwing .~ F. .0001586 .000310 -.0006789 -.0004515 .0007662 -.0001045 -.006796

Symrn. Fl.o -.0000212 -.00008089 -.0000499 -.0001045 .0000947 -.002394 -.002394
F.
· .07616
-.0170 .05558
-.009686
-.04470 -.002394
2.9213
.1425
.1425
1.6937
.2004 .7043
.1425
.1425
.3817
• -5.393 1.449 -.2450 .2004 155.44 .2004
F2 , -.02658 .004706 -.009686 .7043 .1425 .2004 1.5285 .1425
· .000664 -.006796 -.002394 .1425 .3817 .1425 .7078
Concerning the numerical solution, the

REDUCED WING STIFFNESS MATRIX stiffness matrix derived from the displacement
method is a positive definite symmetric matrix,
The wing stiffness matrix as derived above for which nontrivial, unique sol~tions can be
contains all the necessary geometric constraints guaranteed a priori. The real difficulty lies
to represent correctly the wing behavior under a in finding a computationally efficient algorithm
symmetric loading. It may, however, be further for determining these solutions.
reduced by the application of certain zero-
loading constraints which exist purely because Several methods have been illustrated in
of the external load modeling used. Specifi- the example problems of this chapter, including
cally, the load modeling sets the couples F15 , direct solution by inverting the stiffness
F1B, F21, F26 and F 2B equal to zero. Hence matrix (example problem 1) or by Gaussian
Equation 35 may be used to derive a reduced elimination (example problem 3), and solution by
matrix. This reduced matrix offers the ad- iteration with a Gauss-Seidel algorithm (example
vantages of simpler handling if the further problem 2). As a rule of thumb, direct solution
computing tasks of deriving wing influence co- is best for small, fUlly populated matrices,
efficients (by inverting the wing stiffness while both types of solution are used for large,
matrix) and internal wing stresses (i.e., spe- sparsely populated matrices. The preferred
cific element loads) are undertaken. method of solution, and the demarcation between
large and small problems varies from authority
A23.12 Accuracy, Convergence and Error Bounds to authority, and in fact, the choice of method
may be overriden at a given facility by a re-
We have to deal with two kinds of accu- liable, debugged procedure of either type.
racy; first, the accuracy associated with the Problem bookkeeping should be optimized as well.
finite element model; and second, the accuracy For example, where possible, modes should be
associated with the numerical solution, once a numbered to produce a banded stiffness matrix
~odel has been chosen. It is worth pointing to minimize storage reqUirements.
out that the results although given by the com-
puter to many decimal places may be useless, Certain practical measures may be employed
due to a conceptual defect in the model, or a to facilitate numerical calculation. In the
computational defect in the solution. case of one or more extremely stiff elements,
associated coordinates can be eliminated (as
Of the two processes, that of choosing a in example problem 2) to yield better con-
finite element model is more dependent on ex- ditioning of the problem.
perience and judgment. For some types of ele-
ments, convergence to an exact (theory of Computational errors arise from two
elasticity or continuous) solution is assured sources, termed roundoff error and truncation
by taking a sequence of models with more and error. Roundoff error occurs because a finite
more (smaller) elements. Exmnple problems 5 number of digits are carried in the calcula-
and 6 illustrate this situation. It may be tions; double precision arithmetic is often
worthwhile to bound the desired result by con- necessary to avoid loss of significant figures.
sidering limiting, simpler problems with similar In some algorithms, the order of operations is
boundary conditions. In regions of high stress altered by interchanging pairs of rows and
and strain gradients, extreme care must be columns, again to avoid loss of significant
taken in modeling and interpretation of results, figures in subtraction and achieve better ac-
to assure obtaining useful values of stress and curacy. Truncation error is typically as-
deflection. sociated with iteration methods, whose
accuracy is assured only by completing an in- production runs can be performed. Use of on-
finite number of steps. By terminating the line terminals operating on a time-sharing
calculations after a finite number of iterations, system 1s an efficient way to debug, as it
we produce a truncation error. Further error eliminates most of the waiting time between
may be produced if the matrix coefficients are runs. This is especially useful during the
obtained empirically. Placing bounds on the initial runs to discover spelling and syntax
errors is a difficult but important part of the errors. Often a machine failure will terminate
calculation procedure. For details, the reader a run before completion. It is desirable to
is referred to texts on the SUbject of numeri- output intermediate results so that something
cal analysis (5,6). is learned from every run. The amount of
running time for each portion of the program
A23. 13 Management of Problem for the Computer can be checked easily, and can pinpoint those
parts of the calCUlations where reprogramming
A discussion of the management of a would be most productive.
structural problem for solution on an electronic
digital computer could easily occupy a separate Once a program is ready for production,
book. We will be content here to give the it should be compiled and stored on tape or
reader a checklist of topics Which can be con- disk in machine language rather than FORTRAN
sidered as a guide to the questions he should or PL/I. This will save considerable time per
consider and discuss with a computer consultant. run. Where possible, established, widely used
It is assumed that the desirability and feasi- programs should be used; with a new program
bility of performing a computer analy8is have debugging represents a significant proportion
already been determined, and the analytical of the work required to produce results. How-
approach has been selected. ever, one must beware of poorly documented
programs; the cost of the learning cycle could
At this point the key factor is the cost of easily exceed the cost of writing and debugging
the calculation for a given degree of accuracy. a new program.
The method employed, and thus the cost will be
intimately related to the rate structure at the REFERENCES
local computer facility. Often costs not cen-
tral to the computation are the controlling (1) Rubinstein, M. F., structural Systems -
ones. A typical charge rate includes central Statics, Dynamics and stability, Prentice-
processor unit time, number of input-output Hall, Englewood Cliffs, N. J., 1970.
requests, and required storage region size.
Other factors, such as cost of cards and printed (2) Rubinstein, M. F., MatriX Computer Analysis
output, tape and disk storage charges, graphic of Structures, Prentice-Hall, Englewood
plotter usage, etc., may be prorated or charged Cliffs, N. J., 1966.
directly. For a given computer, one of the
elements of the rate structure generally domi- (3) Przemieniecki, J. S., Theory of MatriX
nates the programming strategy. Earlier in the structural Analysis, McGraw-Hill, New York,
evolution of computers, processing time and 1968.
storage were the limiting factors. With im-
provements in these areas, current computer (4) Zienkiewicz, The Finite Element Method in
technology is often bound by input-output limi- Structural and Continuum Mechanics,
tations. In addition to cost considerations, McGraw-Hill, Ltd., London, 1967.
the possibility of committing a blunder in data
inputting has led to use of special data gen- (5) Crandall, S. H., Engineering Analysis,
eration routines and checking programs. MCGraw-Hill, New York, 1956.
Graphical output is especially useful in de-
tecting errors in nodal point location and in (6) Ralston, A., A First Course in Numerical
checking element definitions. Because o~ in- Analysis, MCGraw-Hill, New York, 1965.
put-output costs, it is desirable to keep as
much of the program in core storage as possible. (7) Sokolnikoff, I. S., Mathematical Theory
The cost of storing ann fetching information of Elasticity, McGraw-Hill, New York,
may exceed the cost of regenerating data by a 1956.
factor of ten, twenty, or more. When moving
data in and out of the central processing unit, (8) Levy, S., Structural Analysis and Influ-
the format used can have a critical influence ence Coefficients for Delta Wings, Journal
on cost. It is often advantageous to transfer of the Aeronautical Sciences, 20. 1953.
many record lines at a time using the simplest
format definition available. (9) Schuerch, H., Delta Wing Design Analysis,
Society of Automotive Engineers National
Every program contains errors or bugs which Aeronautic Meeting Preprint No. 141,
must be discovered and corrected before reliable September, 1953.
A23.27
(10) Turner, M. J., Clough, R. W., Martin, (12) Wooley, Ruth, Check of Method for Computing
H. C., and Topp, L. J., Stiffness and De- Influence Coefficients of Delta and Other
flection Analysis of Complex Structures, Wings, National Bureau of Standards Report
Journal of the Aeronautical Sciences, 3655, 1954. (Available as ASTIA No.
23, September, 1956. AD46866).
(11) Kroll, W., Effect of Rib Flexibility on (13) Williams, D., Recent Developments in the
the Vibration Modes of a Delta Wing Air- Structural Approach to Aeroelastic Prob-
craft, Institute of Aeronautical Sciences, lems, Journal of the Royal Aeronautical
Preprint No. 585, 1956. Society, 58, 1954 (see also, Aircraft
Engineering, 25, 1954).
INTRODUCTION TO
THE THEORY OF ELASTICITY AND THERMOELASTICITY
BY
R. ]. BOLLARD, Ph.D.
CHAPTER A24
THE 3-DIMENSIONAL EQUATIONS OF THERMOELASTICITY
There are two most important

A24. 1 Assumptions: that a deformed body returns to its original
assumptions in the classical theory of elasti- shape upon removal of the deformation source,
city concerning the state of an elastic body be it external loads or temperature. If one
throughout its volume. considers the simple tension stress strain
curve, this assumption reqUires that at no point
(i) The material is homogeneous. in the body can the stress level be above the
yield point. If this stress level is exceeded
(ii) The material is isotropic. the problem again becomes non-linear since the
stress-strain relations no longer follow Hooke's
The first requires that the elastic body Law. This non-linear response is idealized in
possess the smne material properties at all the theory of plasticity. It is usual, also,
points throughout its volume and on the sur- in the application of the classical theory of
face. When homogeneity does not eXist, such elasticity to ignore the visco-elastic behavior
as in the case of an inclusion of foreign of materials but the extension to this case
material or of a laminated structure of where the stress-strain relations contain only
different material in each Imnination, the linear time derivatives of stress and strain
mechanical and elastic constants become is qUite straightforward. Since this type of
functions of position in the body. This is behavior, the best known of which is creep, be-
also the case where the material properties comes more pronounced as the temperature in-
become functions of temperature in a non- creases, this possible extension in the linear
uniformly heated body. case will be mentioned in Chapter A27.
The second requires that the elastic The violation of these last two assumptions,
properties are the smne in all directions at small deformations and linear stress-strain re-
any point in the elastic body. If a material lations, leading to the non-linear problem
is non-isotropic, such as in extruded graphite Where, of course, the rules of superposition no
or laminated plastics, the relationships be- longer apply, is mentioned not to cause con-
tween stress and strain are no longer the fusion but to serve as a warning of the limita-
simple equations containing two independent tions of the theory presented herein.
elastic constants. In the most general non-
A24.2 Formulation of the 3-Dimensional Equations of
isotropic material it requires 21 elastic Thermoelasticity.
constants to define the stress-strain re-
lationship. The existence of ~lanes of elastic The problem of thermoelasticity is to
symmetry reduces this number.(4j determine at any point within and on the bound-
ary of a heated elastic body subject to external
A third important assumption is that the loads and constraints the stresses, strains and
displacements are so small that the position displacements. The stress condition at any
of a point in the body does not require re- chosen point within or on the boundary of the
definition after deformation and that the body must be such that equilibrium of forces
higher order terms of displacement derivatives and moments exists at that point. Consequently,
are negligible. If these assumptions are not the expression of the equilibrium state of the
made the relationship between strains and dis- stress components will be independent of the
placements become non-linear and the much more material properties, the deformations, the
difficult theory of large deformations must be temperature and temperature distribution, the
employed. stress-strain law, etc. The relationship
between stress components must be only such as
The final assumption is that the material to satisfy the equations of equilibrium. It is
behaves elastically. 'I'his assumption requires well known that the stress at a pOint, being the
force intensity in a chosen direction, can be
(4) Numbers in parentheses refer to the bibliography at the represented by a single vector, haVing magnitude
end of the chapter. and direction, which is related to a chosen cut
A24.1
A24.2 AN INTRODUCTION TO
plane surface of infinitesimal area at that solution when all 18 equations are simUltan-
point. Thus the definitions of the stress at eously satisfied. The development of the
the point requires more than the statement of solution to a specific problem reqUires a set
the stress vector. It requires in addition of boundary condition equations for evaluation
the definition of the cutting plane through of the unknown coefficients existing in the
the point of interest to which the stress general solution.
vector is referred. The stress at a point can
be fully defined by choosing a Sllfficient nllill- As in the theory of elasticity, the solu-
ber of cuttl~g planes and determining the tion to a chosen physical problem in thermo-
stress vectors, usually defined by its compon- elasticity must satisfy the defined 18 equation
ents, associated with eaGh. In 3-diJlensional set and, as well, the boundary conditions.
bodies a minImum of 3 cutting planes which are The equations of thermoelasticity are derived
mutually at right angles to eaGh other (orthog- in the following sections in the following
onal) is reqUired. Thus, vlhen eacIl stress sequence.
vector associated with one of these cutting
pla~es is defined by its 3 comnonents, there (A) Equilibrium Equations
will exist 9 stress components to be deter-
mined. These 9 components of the stress at a (B) Strain - Displacement Relations
point mak c3-up the stress tensor(4) which is
the most convenient representation for the (c) Stress-Strain Relations
more difficult problems in theoretical analysis.
The three cutting planes are usually chosen (D) Boundary Conditions
parallel to a chosen space fixed cartesian co-
ordinate system, the choice usually clearly It should be noted at this time that the
indicated for the physical system under only change that is necessary in the equations
analysis, and thus, the stress components will of the linear theory of elasticity to the linear
all be parallel to this axis system. Other theory of thermoelasticity is in the stress-
axis systems such as cylindrical axes, can be strain relations.
chosen so long as there is provided the 3
mutually perpendicular planes at the point of A24.3 The Equilibrium Equations.
interest. These 9 stress com~onents are
related by the equilibrium equations as shown Consider the point 0 in an elastic body
in the next section. There will be 6 equations which is SUbject to external loading and intern-
of eqUilibrium and thus there eXists more un- al, or body, forces. It is desired to deter-
knowns (9) than there are equations which fact mine the state of stress at O. This could be
prevents any possible general solution. done completely by defining the resultant stress
tensor with respect to some spatial reference
Under the influence of this stress field but is most commonly done in the following man-
the elastic body will deform in a manner de- ner. Take the point of interest 0 as the origin
pendent upon the stress strain relations. The of three mutually perpendicular axes oriented
stress strain relations for the isothermal parallel to a chosen spatial cartesian axis
problem (the same temperature througllOut) are system. The choice of this reference axis
well known from the linear Hook's Law. (2,3,4) system is usually clearly indicated by the
The introduction of temperature modifies these geometry of the system under study.
relations by the addition of another strain
component due to the thermal expansion z
properties of the material being studied. 3-Dimensional Body
~,
These stress-strain relations provide 6 more
equations but at the same time 6 additional Point 0 within the
unknown strain components are introduced. The body
system of equations is still inc~nplete.
x
The strains Give rise to displacements.
)----------~y
The relationship between strains and displace-
ments which are independent of temperature and
temperature distribution requires 6 equations
for definition and at the same time introduces x
only 3 unlmowns which are the displacement
components. Assume that the body is cut by the planes
xy, xz, yz, having the cownon point of inter-
The number of unknowns is now 18 and there section O. Associated with each of these cut-
exists 18 equations so that the equation set is ting planes there will be a reSUltant stress
complete and a general solution is possible. vector at 0 reqUiring three components for
This general solution will be the unique definition. It is logical to choose the di-
THE 3-DIMENSIONAL EQUATIONS OF THERMOELASTICITY A24.3
rections of these components as those of the equation, surface areas over which the stresses
axis system. (being force intensities) are acting must be
provided. This is most conveniently done by
For example, consider the cutting plane considering the stress components acting over
yz. The resultant stress vector at 0 associ- the faces of a small parallelepiped constructed
ated with this cutting plane can be defined by with faces in or parallel to the cutting planes
the three components ax; ~xy; ~xz where contained in the cartesian coordinate system
with the point of interest 0 as the origin.
ax component normal to the zy plane This parallelepiped is considered arbitrarily
and therefore parallel to the small and in the limit as the edge dimensions
x - aXiS, hence the subscript x approach zero the equations resulting from
equilibrium of the element will describe the
~xy = oneof the stress components in the stress conditions at the point O. Consider
yz plane which is identified by the then such a parallelepiped which is already
normal direction x, hence the first considered infinitesimally small with edge
subscript, and in the y direction dimensions then correctly specified as dx, dy
as indicated by the second sub- and dz. The stress components on those faces
script. This component could be z
written as ~(yz)y but since the
axes are mutually perpendicular the
yz plane is more economically
identified by the subscript x as
shown.
~xz = the other in-plane component at

°Y"'~f--__I
.}-_+-='-+...l- y
right angles to ~xy where again the
first subscript identifies the
cutting plane yz and the second the
direction.
With the adoption of a sign convention

these vector components can be considered z
x
scalar quantities. Associated with the cutting
plane xz will be the components Oy, ~yx, ~yz contained in the cutting planes xz, xy and yz
and with the cutting plane xy the components are shown according to the positive sign con-
0z, ~zx, ~zY' Any other of the infinitely vention. According to this same sign convention
many possible cutting planes passing through 0 the stress components on those faces parallel
can be defined in terms of coordinates associ- to the cutting planes are shown in the next
ated with the three mutually perpendicular figure. It is these components which are of
planes xy, xz, yz. Thus, definition of the 9 interest from the physical interpretation of the
stress components ax, Oy, oz, ~xY' ~xz, ~yx, stress condition at 0 since, because of the
~yz, ~zx' ~zy will completely define the stress limiting process mentioned above, it is these
condition at the point 0 in the body. The directions which are indicated by positive
magnitudes of these vector components will be values in the solution to the differential
obtained from the solutions of the differential equations of equilibrium. It is observed that
equations of equilibrium, stress-strain and positive values of ax, Oy and Oz indicate
strain-displacement relations and the boundary tension of the elastic material at the point O.
conditions, wherein the components are treated z
as scalar quantities since the sign convention
illustrated in the next section is always
implied and the direction is that of the
chosen cartesian coordinate system.
To establish the equations of equilibriun dz

the fundamental laws of statics are employed, a~yz
viz; +--dy
ay ao y
r - t - - ' - - Oy + - - dy
ZFx = ZF y = ZF z = ZMo = 0 ay
.y
where Fx indicates a force in the direction x,
etc., and Mo a moment about O.
Since it is forces which satisfy this

x
When in doubt during the physical interpreta- 'Txy = 'Tyx ; 'Txz = 'Tzx ; 'Tyz = 'T zy - - - (5 )
tion of solutions to the equations of elasti-
city one should return to these two figures The equations 2, 3, 4 and the three equations
which define the positive sense of the stress of 5 make up the 6 eqUilibrium equations. In
components. the following discussions equations 5 will be
accepted and thus the preceding discussion on
The stress components have all been in- the equation set of 18 must be modified to the
creased an amount equal to the rate change of set of 15 with a similar reduction of 3 in the
magnitude in the direction of movement times number of unknowns.
the di~tance moved. This increase is, of
course, the change in magnitude between the It should be noted that these equations
components on the planes passing through 0 2, 3 and 4 define the equilibrium state of the
and those displaced an infinitesimal distance volume element and their derivation requires
away in the positive direction. In this no limiting assumptions as to the elastic
volume element, dxdydz, it will also be nature of the material. The equations are thus
possible to have body forces X, Y, Z defined qUite general and will be the same for non-
as the force per unit volume and which can linear systems and those in which displacements
arise as gravitational forces, magnetic are created by temperature differences. For
forces, etc., acting in the direction x, y convenience, these eqUilibrium equations are
and z respectively and through the centroid written in cylindrical coordinates r, Q and z.
of the volume element. It should be borne
in mind that these are internal forces within dar + ~ d'Trg + d'Trz + or - 0g
the element considered and are distinct from dr r dQ dZ r
the external forces which will appear in the z + R ;:: 0
equations defining the boundary conditions.
The equations of equilibrium are obtained d'Trz + ~ d'TQr + dOz + 'Trz

directly from the static equilibrium con- dr r dQ dz r
ditions: + Z =0
ZFx =0 x d'Lrg + ~ dog + d'TQz + 2'Tr Q
ZFy =0 - - - - - - - - - - - - - (1) dr r dQ dZ r
ZF z =0 + I-Q-l =0
To illustrate the procedure the forces in the where the body force components in the r, z and
x direction are summed and equated to zero. Q directions are identified as R, Z and~
respectively, and in spherical coordinates r, Q
(Ox + OOX) O'TVY and 0
ax
dx dy dz - 0xdydz + ('Tyx +~dY)dZdX
dar + ~ d'TrQ + 1 d'Tr¢
- 'Tyx dz dx + ('Tzx + O'Tzx dZ) dy dx - 'Tzx dy dx + cir r dQ r sin Q d0
oZ
Xdxdydz =0 z + r1 (2or - Og - 0'1 + 'Trg co t '""'I
which, on simplification, gives + R = 0
Q
oOx + ~ + O'Tzx + X =0 _ (2 ) )f--+-~Y

d'TrQ + ~ dog + 1 d'L~,0 +
oX oy oZ dr r dQ r sin g d
Similarly
x
cot Q + 3'Tr~
o'TXy + OOy + O'Tzy + Y + ~ = 0
oX oy oZ
o- - - - - - (3)
d'Lr0 + 1.. d'Tg,0 + 1 ~+
o'Txz + O'Tyz + ooz + Z =0 _ _ (4)
dr r ciQ r sin Q d0
ox oy oZ
; ~'Tr0 + 2'Tg0 cot gJ
It can be readily shown, by taking moments
about the volume center and satisfying the + ~ = 0
equation of static equilibrium, ZM o = 0 that
the complementary components of the shear where the body forces in the r, Q and 0 direc-
stress are equal, ie. tions are given by R, ~ and respectively. p
A24.4 Strain - Displacement Relations. The shear strain 7f xy =a + #

Equations 2, 3, and 4 being three = du + Ov
equations with six unknowns are insufficient dy oX
to establish solutions for these unknowns within the
ox, Oy, Oz, ~xy, ~xz, ~yz. One looks there-
fore to equations relating the stresses to the assumptions stated above. In similar fashion
deformations of the deformable elastic body the remaining two components of shear strain
for the further equations necessary to provide can be obtained to give the complete set
a complete set. Consider first the strain
displacement relations. The discussion of -t xy = ou + .:!..y (9)
the relation between these strains and the dy dX
stresses will follow.
l5'Xz = dw
oX
+ du
oZ (10)
If a fictitious fiber of unit length lies
in the positive x - direction before deforma- "lf yz = dZ
dv + Ow (11)
tion, one has only to determine the increase oy
in length of this fiber after deformation to
obtain the strain at the point of origin of Again, for future convenience, these strain dis-
the fiber. With the assumption that this placement relations are written in cylindrical
change in length is sufficiently small that and spherical coordinates below.
second order terms of this change can be
neglected and further, that rotation of this In cylindrical coordinates,
fiber away from the x - direction is
small, it is easy to show by simple trigo- _ du . E _ dw. _ u + 1 dv
nomE:try that the strain EX is given by Er - dr' Z - dZ ' EQ - r r dQ
7f
rQ -
-
r1 Ou dv v..J!
dQ + dr .- r ' 0 rz
= dz
du + ~
dr
dU
EX = oX (6)
7f zQ = dv
dz
+ 1.r ow
dQ
By considering fibers originally in the y
and z directions the same development, given where u, v and w represent displacements in the
in all texts in elasticity, gives r, Q and z directions respectively. And in
spherical coordinates,
_ dv
Ey - dy - - - - - - - - - - - - - - (7) _ du u
Er - ar EQ =r + r1 ddQv
EZ = oZ
dw (8) = ~ + :!... cot Q + 1 Ow
E0 r r r sin Q 00
where u, v and ware the displacements of the v
point of interest, the origin of the three r
fibers considered, from the undeformed body
position. The shear strains are readily ob-
tained, when the same restrictions on the "If 1 ou + .~ _ w
magnitude of length change and rotation of the
r0 =r sin Q 00 dr r
fiber are enforced, by consideration of the
two fibers lying in the same plane. Consider
those two projections in the x-y plane start-
ing from the point of interest 0 in the body.
where u, v, w represent the displacements in
y
the r, Q and 0 directions respectively.
v t Position of Fiber
Ou in Undeformed Body It must be noted that these equations apply
~. __0:J _ - - 7 only under the restrictive assumptions on the
dy I / order of magnitude of displacements. These are
i OJ;-
1
Position of Fiber
in Deformep' Body
the linear strain-displacement relations of the
classical theory of elasticity. An obvious
extension to the classical theory immediately
I Lov presents itself as the theory of large deforma-
~ -:Tcix tions wherein the higher order terms of the
(-==--7<-'----i-p-----x
0_ -du- displacement derivative are considered. This
---+
dx u difficult non-linear theory is being actively
developed since many structural problems, such Ox = Ae + 2~EX

as the buckling and post-buckling behavior of
structures, involve large deformations. Oy = Ae + 2fl£y
We now possess a set of 9 equations, the Oz = Ae + 2fl£z
3 equations of equilibrium and the 6 strain-
displacement relations, with the 15 unknowns ~xy = ~dxy
ox, Oy, oz, ~xy, ~yz, ~xz, Ex, Ey, EZ, ~xy,
~yz, t xz , u, v, w so the equation set is even ~yz = I!Syz
more incomplete than before. However, by ~zx = \-I-7fzx
means of the stress-strain relations, Hooke's
Law, the equation set will be completed to 15. where the Lame constants
A24. 5 Stress-Strain Relations. liE
A.= (1 + )/) (1 - 2)1)
In accord with the limiting assumptions
of the linear theory of elasticity the general E _
statement of Hooke's Law requires 36 constants \-I- = 2(1 + Y) - G
in the isothermal problem. For isotropic
materials, wherein the elastic properties at and
any chosen point throughout the body are
the same for all possible directions of ob- e = EX + Ey + EZ = the dilatation of the
servation, these 36 constants are readily volume element.
shown to be dependent(l, 2, 3, 4) and can be re-
duced to the two most commonly used, E, It should be further noted that
Young's modulus and V, Poisson's ratio, or the
Lame constants A and~. For the isothermal 3..\ + 2\-1- = 3k
problem the stress-strain relations can be
written as; where
k = -..,.---=E=------,-
EX = i ~x - Y(Oy + oz~ 3(1 - 2V)
= i ~y -
In the case of the heated elastic body
Ey y(oz + oxTI these strain expressions must be modified by
Ez -If G
- 1 Oz - Y(Oy + oxTI
the addition of a component arising from the
thermal expansion property of all structural
materials. This component will be proportional
1 to the temperature level, T, above or below
txy =Cf'T;xy some reference temperature (usually room temper-
_ 1 ature). The constant of proportionality is the
t yz -G~YZ familiar linear coefficient of thermal expan-
sion, which, for isotropic materials is the
1 same in all directions of observation at the
lI'xz =Cf~xZ point O. It is obvious that the direct strain
components, Ex' Ey , EZ' being the change in
where E = Young's modulus = slope of the uni- length per unit length of a fictitious fiber
axial stress strain oriented in the x, y, z directions respectively,
curve are directly affected by the thermal expansion.
The shear stress components will not, however,
be affected for isotropic materials since the
G = shear modulus = 2(1 ~ Y) angular deformation of a small volume element
located at 0 will not be changed during the
v = Poisson's ratio = 0.3 for steel uniform expansion of this volume element. The
stress-strain relations for thermoelasticity
= 0.25 -. 0.50 for thus become:
all elastic mater-
ials at room EX =i ~x - y(Oy + oz] + aT - - - - (12)
temperature.
It is often more convenient to express the

stress strain relations with the Lame con-
Ey = i ~y - )/(ox + oz] + aT - - - - (13)
stants of linearity. In three-dimensional

elasticity these relations become
Ey =~
dy /YZ =dZ
dV + oW
oy
dw du oW
- - - - - - - - - - - - (15) EZ =dz- =-+--
oZ dx
1 6 strain-stress relations
7JXY = G 'txz - - - - - - - - - - - - (16)
EX = ~ ~x-V(OY+oz~ + aT
If yz =~ 'tyz - - - (17)
These equations can be inverted to express Ey =~ Ey-V(ox+ozj + aT

the stresses as functions of strains. This 1s
left as an exercise for the student but these EZ =~ ~z-Y(ox+oY~ + aT
relations containing the Lame constants are
given here for completeness of treatment.
t xy =G1 'txy
Ox = )..,e + 2fJ,E x (3)'' + 21-l)aT
oy = Ae + 21-l Ey (3).. + 2fJ,)ctT
oz = Ae + 2fJ,Ez (3).. + 2fJ,)aT
'txy = fJ,tXY
The equation set is now complete since the
'ryz = fJ,2lyZ number of independent equations is equal to the
number of unknowns.
'tzx = fJ,tzx A24.7 The Compatability Equations.
where the definitions of).., fJ, and e remain un- It will be noted in the strain displacement
changed from those previously given in the equations that 6 strain components are necessary
isothermal equations relating stress and to define the 3 displacement components. Thus,
strain. if one were to choose at random a set of 6
strain components this does not mean that there
A24.6 The General Equations of 3-Dimensional eXists a possible set of 3 displacement compon-
Thermoelasticity. ents which will allow satisfaction of the strain-
displacement equations.
The unknowns are: Consequently the choice of these 6 strain
components cannot be arbitrary and, in fact,
6 stress components ox,oy,oz,'rxy,'txz,'tyz must be such that the compatability equations,
6 strain components EX,Ey,Ez,txy,txz,tyZ which are equations relating the derivatives of
3 displacement components u, v, w the strain components, are satisfied. These
15 compatability equations thus ensure the exist-
ence of continuous single-valued displacement
The necessary 15 equations are given by functions. This, of course, ensures the unique-
ness of the solution when, as is possible, the
3 equations of equilibrium analysis is formulated in terms of stresses and
strains only. These equations are automatically
dox + d'rxy + d'rxz + X satisfied if the complete 15 equation set is
dX dy oZ =0 satisfied.
dOl + d'rxy + d'rlz + y These equations are derived in the follow-
dy dx dz =0 ing manner.
From equation 9
doz + d'txz + d'ryz + Z =0 J _ du dv
dZ dx dy °xy - -
dy +dx-
6 strain - displacement relations therefore

dfqt d 3u d3v
XY
dx dy =dxdy"
- - - +dx"dy
---
or from equations 6 and 7 ing these boundary conditions

d"'"/ xy = d 2E:x + ~ would be a general 3-dimensional
oxoy ~ dx 2 elastic body subjected to hydro-
static pressure or the same body
The remaining 5 compatability equations non-unifonnly heated. In the
are derived in similar manner and the complete first example the surface
set appears as follows tractions are the pressure levels
d 2t yZ _ d 2 E:y + d 2 E:z (18) acting normal to the surface and
dXdy - ~ d'X" in the latter the surfaces are
traction free (the boundary
d2i5xz _ ~ + d 2E:z - - - - - - - - (19)
tractions all zero). If the sur-
oxdz - OZ2 ~ face tractions at any point on
the surface are represented by
the vector components X, Y, Z
- - - - - - - - (20)
parallel to the x, y, z axes
respectively, the equations of
eqUilibrium at the boundary give
- - - (21) the boundary condition equations:
2 2 2 X = Gx l
d lf xy _ d t xz + d t yz = 2~ (22) + 'Lxyill + 'Lxzn
oydz ~ oxdy dxdz
Y = Gym + 'Lyzn + 'Lxyl - - - (24.)
2 2
2
_ d tJxy + d t xz + d 6'yz =
2 d2 E:Z (23)
Z = Gzn + 'Lxzl + 'Lyzm
OZ2 oYoz oxdz dXdy
As will be noted later, these compata-
bility equations are automatically satisfied where 1, m, n are the direction
i f the problem is formulated in terms of d.is-
cosines of the surface normal
placements and then the general solution must drawn outward from the surface.
satisfy only the equilibrium equations with
these equations written in terms of displace- Case 2. All external boundaries are con-
strained by geometrical restraints
ments. In comparison, if the problem is
formulated completely in terms of stresses and thus, all boundary displace-
then both the equilibrium and the compatability ments are specified. This case is
relatively rare in engineering
equations must be satisfied. The 3 eqUilibrium structures. A simple example
and the 6 stress-strain relations togetrler with would be that of a heated insert
the 6 compatability equations again provide a
fully contained within a rigid
complete set of equations for the general body. If the boundary displace-
soLltion to a problem formulated in terms of
ments are specified ov~r the whole
stresses and strains only.
boundary, the boundary conditions
It now remains only to obtain the specific can be written as:
solution to this complete set of equations for
the chosen physical problem. This reqUires u = f(s)
the use of the equations which state the
boundary conditions into which the general v = g(s) - - - - - - - - - (25)
solution is substituted to obtain a set of
algebraic equations from which the unknown w = h(s)
constants of the general solution can be
evaluated. where s is the surface point and
A24.8 Boundary Conditions f, g, and h are known functions
prescribed by the specified dis-
It is obvious that the boundary conditions placements.
can be prescribed in three different possible
ways. Consider each case in turn. Case 3. The boundary conditions are speci-
Case 1. All external boundary conditions fied by stresses over part of the
are prescribed in terms of ex- surface and by displacements over
ternal surface loads with no the remainder. A simple example
geometrical restraints imposed. is that of a heated cantilever
Since the derived equations are wing where the assumption of a
in terms of stresses, these loads rigid root fixity prescribes zero
are converted to load intensi- displacements at the root while
ties (lb. force/unit area) and over the remainder of the structure
are then termed surface tractions. surface the surface tractions are
Simple examples of a system hav- zero, for the unloaded wing, or
prescribed by the external load- theory of elasticity and thermoelasticity

ing. This problem is discussed contain the 15 unknown stress, strain and dis-
in some detail later in the text placement components. If the boundary condi-
of Chapter A25. Both sets of tions are, as in Case 1 of the preceding
equations stated in the above section, given by surface tractions only, the
two cases must then be used boundary condition equations contain 6 stress
where appropriate. components. It is then desirable to formulate
the 15 equations in terms of 6 stress equations.
A24.9 General Notes. These will be the 3 equilibrium equations and
the first 3 compatability equations written in
Only the linear theory of thermoelasticity terms of stresses by use of the stress-strain
is being considered herein and consequently and strain-displacement relations. The remain-
the rules of superposition apply. This means, ing 3 compatability equations must also be
of course, that one can determine the stresses, satisfied to ensure a unique solution. The
strains and displacements of an elastic specific solution is obtained in terms of
structure due to one of several causes and stresses only from the set of 9 equations and
add the results of each cause treated inde- the 3 boundary condition relations. This
pendently to determine the stresses, strains formulation is the one employed in the thermal
and displacements due to all the causes acting stress analysis where all surfaces are con-
simultaneously. In the thermoelastic analysis sidered traction free. It is important to
of heated loaded structures it is frequently remember that in this stress formulation the
most convenient to determine the stress, strain equations of equilibrium and compatability
and displacement distribution due to tempera- (written in terms of stresses) must be satis-
ture alone and to then add to this result the fied simultaneously by the general solution.
isothermal analysis of the loaded structure
according to the well documented theory of If, on the other hand, the boundary condi-
elasticity to obtain the final distribution. tions are those of Case 2 then the boundary con-
In determining the stress, strain and dis- dition equations contain only 3 known displace-
placement due to temperature one can rightly ments. The problem is then formulated in terms
consider all surfaces to be traction free so of .displacements. Since, by the choice of
long as, in the isotl18rmal analysis, deforma- displacements over the boundary, the equations
tions at geometrical restraints due to the of compatability are automatically satisfied it
temperature are restored to the prescribed remains only to obtain a solution of the 3
values by such boundary condition equations as: eouations of equilibrium which have been re-
written in terms of displacements, again by the
u = f(s) use of the stress-strain and strain-displacement
relations. Unfortunately, as stated before,
v = g(s) - - - - - - - - - (26) this most desirable case rarely exists in
engineering analysis.
w = h(s)
If the boundary conditions are of the mixed
The functions uT, vT, wT will be the type of Case 3 then no one formulation is the
boundary displacements calculated from the more desirable.
thermoelastic analysis considering all sur-
faces free of tractions.
This technique has the distinct advantage References:

that the thermoelastic problem can then always
be formulated in terms of stresses only, 1. "Theory of Elasticity" by S. Timoshenko,
leaving the mixed boundary condition problem McGraw-Hill Book Co., Inc., New York,
to the isothermal analysis. It will be seen N. Y.
later, however, that this is not always ad-
vantageous and it is often desirable to retain 2. "Mathematical Theory of Elasticity" by
the prescribed boundary displacement conditions A.E.H. Love; Dover Publications, New
in both the thermal and isothermal analyses. York, 1944.
Unfortunately no rules of choice can be given
and the decision must be made by the analyst 3. "Analysis of Stress and Strain" by A.
based on his own experience. The examples to J. Durelli, E. A. Phillips and C. H.
be considered in the next chapter will assist Tsao, McGraw-Hill Book Co., Inc., New
in this choice however. York, N.Y.
Mention is made in the above of a stress 4. "Mathematical Theory of Elasticity" by

formulation. This point deserves clarification. I.S.Sokolnikoff, McGraw-Hill Book Co.,
The fifteen equations of the 3-dimensional Inc., New York, N.Y.
BOEING
NORTH AMERICAN
LOCKHEED
Proposed Designs For Supersonic Airliners. Boeing won the competition but later changed design to
fixed wing type. Later Congress killed the SST project, a bad decision in this author's opinion.
CHAPTER A25
THE 2-DIMENSIONAL EQUATIONS OF ELASTICITY
AND THERMOELASTICITY
z
A25.1 mtroduction:In any elastic body wherein Assume: Oz = ~xy = ~yz = 0
the stress components at a point of interest w, oxz, ~yz small
all lie in the one plane and, due to symmetry
in the body itself or in the external loading, Unknowns: OX> Oy, ~xy, Ex' Ey , E z
the stresses throughout the body all lie in oxy, u, v
this one plane, the stress analysis becomes Plane of middle
2-dimensional. All stresses normal to this surface parallel
plane are zero. The analysis is then referred to xy plane
to as one of plane stress.
In the same manner, if all strains lie in \-----.-y

one plane (all strain components normal to the
plane being zero) the problem is again two
dimensional and the analysis is referred to as
~
one of plane strain.
All forces in
plane of middle
The restriction to a planar surface is, surface
of course, not necessary and the statement of Temperature a function
the 2-dimensional problem of stress or strain x of x and y only
should be generalized to that wherein the
general 3-dimensional equations can be reduced
to a 2-dimensional form by the possible neglect Figure A25.1 Plane Stress
of the stress or strain components parallel to
one of the axes of reference.
restrained and where the temperature does not
The equations for plane stress are differ- vary through the thickness as illustrated in
ent from those of plane strain and, conse- Fig. A25.1. The problem of the thin cantilever
quently, each in turn is treated in the follow- wing with no temperature variation through the
ing development. thickness is a physical system satisfying this
statement as is the isothermal problem of a
In the following development the equations simple cantilever beam. Both of these examples
of thermoelasticity are derived and discussed. will be discussed in some detail in Chapter A26.
The equations of elasticity follow directly by
deletion of all terms containing the temper- In equation 25.1 the unknowns are reduced
ature, T. to the following:
A25.2 Plane Stress. Ox EX U (w is assumed negligibly

small and the results
For illustrative purposes only, the dis- Oy ey v must therefore be ap-
cussion is limited to a physical problem of ~xy "JXY proximate if w does
such geometry that the system is best defined exist at all.)
by a cartesian coordinate system. The ex-
tension to a system defined by cylindrical or Where the statement of there being only the two
spherical coordinates follows directly. displacement components u and v is an assumption
consistent with that of the thin slice where the
With the above limitation, the plane edge loadings are symmetric about the middle
stress problem requires that all stress com- surface so that no bending displacements out of
ponents parallel to one of the axes are zero. the plane occur. This required symmetry of
Let this axis be the z-axis. Then loading for the plane stress formulation must
be carefully considered in those systems defined
o z -- ~xz = ~yz =0 25.1 by cylindrical or spherical coordinates such as
in the problems of thin cylindrical and spher-
is the equation which defines completely the ical shells.
state of plane stress. One simple example to
which this equation applies is that of a thin The 8 equations, necessary for a general
slice where the faces are unloaded and not solution, can be derived directly from the pre-
A25.1
ceding chapter by sUbstitution of equation 25.1 Y = O.

into the equation set of A24.7.
Boundary Conditions:
The two equations of equilibrium
The boundary condition equations from
aox a1:xy Chapter A24 reduce to the following in each
-ax
-+--
ay+ x = o 25.2
case.
aOy a1:
- - + - -xy
+y=o 25.3
Case 1, traction boundary conditions at
ay ax the slice edges,
The three stress-strain relations are: - x = 0xl + 1:x yn
25.4 Y = oym + 1:xyl 25.14
Z =°
25.5
where X, Yare the components of the external
surface forces per unit area in the directions
t XY = lG .,..'xy, 25.6
x and y respectively.
Case 2, displacement boundary conditions

°
where T = for the elasticity problem and where in the thermal stress analysis it is
and the three strain-displacement relations are, usual to assume X = Y = o.
EX
au
= ax 25.7 u = f(s)
v = g(s) 25.15
av
Ey 25.8
ay w= °
t xy au + av The discussion of Chapter A24 applies to
= ay ax 25.9
these boundary conditions and should be studied
again if necessary. This discussion outlines
The equation set is complete for the the usual procedure of treating the thermal
general solution. There is still required an stress problem in stress formulation with all
equation for the strain components EZ and the surfaces traction free and then determining the
compatability equation required for uniqueness effect of external loads and removal of thermal
of solution from a stress formulation. The displacements where prescribed by means of an
expression for EZ is not usually included in isothermal analysis.
the completed set since it results directly
from the solution of this set and is given by, A25.3 Derivation of Equations.
For those to whom this sUbject is new it

25.10
will be well to derive, in this section, the
complete equation set in some more detail.
The compatability equation is given by These derivations can be found in any text in
the theory of elasticity and are repeated here
sa y sa ",2
a uxy a EX Ey merely for the convenience of the reader. In
---
axay =- -+
ay2
0
---::--lr
ax
25.11
many two-dimensional stress problems it is
necessary to develop the strain-displacement
in .terms of strains, or by relations from fundamental consideration of
possible deformation in space of a chosen fiber
25.12 so that special attention should be paid to the
development of the most simple case presented
in terms of stresses where here.
25.13
The equations of equilibrium.
Consider the figure shown in Chapter A24
x =y= 0, on which the positive stress components are
shown for an element extracted from a three-
and E and a are constants (independent of dimensional body. If one were to view this
temperature and position in the bOdy). See figure from the z-axis in a positive z-direction
note in section A25.4 for case where X = 0, the following figure would be observed.
THE 2-DIMENSIONAL EQUATIONS OF ELASTICITY AND THERMOELASTICITY A25.3
y Oy + aa~Y dy Now, if the material is isotropic it follows

that the directions of principal stress are
'Lxy + a~f dy also the directions of principal strains. The
relationship between principal stresses and
strains can then be written as
+-
a'L
i y 'Lxy+~dx
OX---:-j dy L X Ox + a~XX dx
'L xy ~Jf:::::::==-~dx~.-=-=1 x
In matrix form, these equations can be written 1
'Lxy----r as
~oy
For the plane stress problem the stress com-
ponents shown are all that will be acting
since, by definition, the stress components and from Cramer's Rule it therefore follows that
out of the x-y plane are zero. The equations
of equilibrium are then readily derived by
summing the forces in the x and y direction,
remembering that the figure actually repre-
sents a plan view of a three-dimensional ele-
ment of depth dZ, and equating this algebraic and that
sum to zero.
+ __....:0..!!2_a. "-J.=-J._ _
Consider ZFy = O. aJ.:l. a 2" - a,,:l. aJ."
aoy Again, from the assumption of isotropy, which

(Oy + oy dy) dxdz - 0y dxdz + states that the elastic properties are the same
in all directions at a point in the elastic
O'Lxy body, it follows that since the principal
('Lxy + ax dx) dydz - 'Lxy dydz + Ydxdydz = O. direction SUbscripts can be arbitrarily pre-
scribed a complete interchange of subscripts
can be made. Thus it follows that
Thus, upon dividing through by the element
volume, dxdydz, and cancelling,
From the equation ZFx = 0,

oox O'Lxy
-- + - - + X
ax ay = o. and the constant a-=.:l.2!:'!. - = E)/
a:l.:l. a 22 - a,,:l. a:l. 2
If the elastic material is to be in Thus the direct extensions are related by
equilibrium the variation of stress components the simple formulae, wherein the 9 general con-
and the body forces must satisfy these stants have been reduced to two,
equations.
The stress-strain relations. EJ. = E1 (OJ. - )/°2 )

Consider the elastic isothermal problem. 1
The generalized statement of Hooke's Law pro- E2 = E (0 2 - vOJ.) •
v~des that each of the independent components
of stress (Ox, 0y, 'Lxy) may be expressed as a If, however, as is generally the case, the
linear function of the strain components coordinate directions are not the principal
(EX' Ey, Cxy) and visa versa. Thus follows directions then shear stresses and strains will
the mathematical expressions, exist simultaneously with the direct extensions.
The shear strain 1s independent of the direct
Ox = aJ.J. EX+ a:l." Ey + 1\3 txy extensions and is related to the shear stress
by the relation
0y = a 2:l. EX + a"2 Ey + a 23 tJxy
'Lxy = a 3:l. Ex + a 32 Ey + a 33 r/xy.

where, it can be shown The figure shows an unstrained fiber O_OJ.

located in the elastic body at the point 0 and
E at known orientation With respect to the axes
G=2(1+'Y) x and y. Upon deformation of the body the fiber
takes up the new position OJ.-O~. The original
Thus, in the general case where the co- length of the straight fiber is dr and the
ordinate axes are not the principal axes the length in the displaced position becomes dr +
two-dimensional equations relating stress and edr. The quantity e is known as the strain of
strain can be written the fiber. The end displacements of the fiber
are u, v and uJ.' vJ. as shown. These end dis-
placements are related by the equations
au 1 dr + au m dr
uJ. = u + ax ay
av
VJ. =v + - 1 dr + av m dr
ax ay
txy = ~ 't"xy
where 1 and m are the direction cosines of the
If the elastic body is heated it should be fiber. The coordinates of <\are u and v. The
obvious that the small element under consider- coordinates of OJ. are 1 dr and m dr. The co-
ation will expand uniformly and the resulting ordinates of O~ are
deformation due to temperature T will be equal
direct extensions aT, where a is the linear 1 dr + uJ.
coefficient of thermal expansion, in both the and m dr + vJ.
x and y directions. The shear strain component
will be unaffected. Thus the thermoelastic Thus the projection of the strained fiber
stress strain relations in the plane stress on the x-axis is given by
problem can be written as,
_ 1 ldr + uJ. - U =ldr + ~~ ldr + ~~ mdr
- E (ax - Yay) + aT
and along the y-axis by
av av
mdr + (vJ. - v) = mdr + ax ldr + ay mdr
J _1
6xy - G 't"xy From the right angle triangle it follows
that
Strain - Displacement Relations.
Consider a point, 0, in an elastic two- (dr + edr) '" au ldr + au

= (ldr + ax aymdr) "
dimensional body. Under the influence of
external loads and temperature differences av av
+ (mdr + ax ldr + ay mdr )
'"
this point will be displaced to a new position
O~ through the displacement components u and v.
As discussed in Chapter A24, it is a necessary which leads to
assumption of the linear theory of elasticity
that these displacements be sufficiently small
e = 1'" ax
au + m'" av + 1m (au + av )
ay ay ax
that no redefinition of the point 0 is re-
quired after deformation or that the squares
and products of their derivatives may be
neglected. The strain-displacement relations Thus the strain of any fiber of a given
are derived by consideration of the deformation orientation can be found if the derivatives of
of a fiber oriented originally in the direction displacements are known. Consider the strain
of the chosen component direction. in the x-direction. This is obtained by con-
sidering a fiber originally aligned with the
Consider the figure below, x-axis. In this case 1 = 1, m = 0 and
Y or.
0~://J/ vJ.
:'>~/ oJ.
~/ U
, 1 Similarly the strain in the y-direction is
0. / ~'t given by
O_v
u x
av aOx a'txy_
ey = ay ax + --a:y -
when the body
au av aoy a'txy_ =
forces X = Y 0
The term ay + ax is known as the shear --+---
ay ax 25.16
strain. The physical meaning can be seen by
consideration of a fiber oriented at an angle and
of 45 0 to the x-y axis system and therefore
subtending an angle of 90 0 at the origin.
It is left as an exercise to prove then that
au + av is the reduction in this subtended
ay ax =0 25.17
angle when the fiber is strained. This
physical picture is, perhaps, better under- when both E and a are independent of the
stood by reference to the figure below. temperature, T, and position in the body. The
derivation of this compatability equation in
terms of stresses is left as an exercise for
the stUdent.
au The specific solution to a problem in this

Jay --::::--1 formulation reqUires those values of ox, Oy and
~. - - I applied shear 'txy which satisfy the above 3 equations and the
,,1./ ( t r e s s 'txy. boundary condition equations of 25.14.
r,'II I
I
~original element =
When X = Y 0 (body forces not present)
/ _ -/....-- . shape
or when these body forces can be represented by
o ----~ II!
a potential function, V, in the following manner
av Y=
X = - -ax , av
- ay 25.18
ixy = reduction in original 90° angle.
the inverse solution technique usually proves
au av the most profitable. Since, in the linear
=-
ay+ax
-
problems of elasticity and thermoelasticity, the
stress formulation occurs most frequently this
It should be noted that the general ex- inverse approach is illustrated here in some
pression for strain, e, can now be rewritten detail. This approach is not restricted to the
in terms of ex, ey and txy as follows, plane stress problem now under discussion and
can be applied equally as well to plane strain
e = 12 ex 2
+ m ey + 1m txy or any two-dimensional stress formulations.
which is exactly analogous to the direct stress In general, one can choose a stress function
expression on a cutting plane oriented at an ¢ defined by the relations
angle to the x-y axis system, viz:
Ox =~
a2 ¢
a = 12 0x + m 2 0 y + 1m 'txy • ay
It is readily verified, therefore, that _ a2 ¢

the directions of principle stress coincide Oy - ax 2 25.19
with the direction of principal strains.
A25.4 Formulation of Plane stress Problem.
(a) stress Formulation - All boundary This function ¢ is commonly referred to as the
conditions given in terms of surface tractions. Airy stress function since Airy first applied
this well known mathematical technique to
The 2 eqUilibrium equations are al- problems in elasticity.
ready written in terms of the 3 stress com-
ponents OX> Oy, 'txy • The compatability equation It is obvious that the equilibrium
is given by 25.12. The equations which must equations are automatically satisfied by 0 When
therefore be satisfied by the three unknowns chosen according to 25.19. To obtain a general
ox, Oy, 'txy are: solution for the stresses it remains, therefore,
to solve only the compatability equation, 25.17,
which, when written in terms of 0, becomes
2.5.20 r
lil
=X lil
+ Y
lil
when X =Y =0 9
and
and the compatability equation from which the
general solution is obtained (equivalent to
equation 25.20 when the body forces are con-
when the body forces X & Yare defined by sidered zero).
25.18.
The stress function, 0, is defined by the
In both 25.20 and 25.21 equations
and E and a are not functions of temperature.

The general solution for 0 is obtained by
solving 25.20 or 25.21 for a given temperature
distribution. The general solution for
stresses follows directly from 25.19. The
specific solution is obtained by substitution
into 25.14, the boundary conditions. = _~ (l.-a0)
ar r a9
In this case, where the boundary conditions
are those of traction free boundary conditions, and the compatability equation becomes
the equation 25.14 can be written in terms of
o as
0=0
and a0
an
=0 25.22
This reduction of the 3-dimensional case
where n is ·the outward drawn surface normal at to the 2-dimensional plane stress problem can
all points on the boundary. ThUS, the specific be generalized to an eqUivalent problem which
solution can be obtained in terms of 0 directly includes the case where the temperature varies
and the specific solutions for ox, oy, and through the plate or slice thickness, t, by
~y by direct substitution into 25.19. employing the equations
It should be noted that in this case of

traction free boundary conditions there will
exist no thermal stresses at all if the 'Lxz = 'Lyz = 0 25.24
temperature distribution is such as to satisfy
the equation at the upper and lower surfaces. The equations
are analogous to 25.1. This equation requires
v"T =0 25.23 that the stress components, the temperature,
the displacements and the edge loadings be
and body forces do not eXist. ThUS, if T is taken as average values across the thickness.
constant or a linear function of x and y, no For example, the temperature would be written
thermal stresses will reSUlt. as
It is wise to repeat at this time that if T = ft/2 Tdz

there is any displacement restraint placed on -t/2
the boundaries, stresses due to the restraint
of thermal strains will result even though T The quantities 'Lxz, 'L yz ' 0xz' 0YZ and w must be
is such as to satisfy the equation V"'T = O. SUfficiently small across the thickness that
(See discussion in Chapter A24). they can be neglected. This is not the exact
plane stress problem and some care must be taken
It is convenient, at this pOint, to pre- in its application as will be demonstrated in
sent the definition of the stress function for Chapter A26. The results will be average values
cylindrical coordinates (equivalent to equation of ax, 0Y' 0xy' EX' Ey , U, v across the thick-
25.19 and obtained therefrom by the simple ness.
transformation from the x-y plane to r-9 plane
With z the axial direction. (b) Displacement Formulation - The bound-
ary conditions given in terms of boundary dis- grain, the rocket nozzle) atomic reactors
placements. By choosing the displacement (cylindrical fuel rods, heat exchanger tUbes)
formulation the compatability equation is auto- etc.
matically satisfied. The general solution for
the two displacements u and v are obtained ~he question of whether a system under
from the two equilibrium equation written in analysis can be considered one of plane stress
terms of displacements as follows: or one of plane strain is frequently raised and
should be answered at this point. One must
E a (au + av) E 'V,. consider the system carefully and check as to
2(I-Y) ax ax ay + 2(I+V) u whether all the requirements pertinent to one
or the other states has been satisfied. If any
aT
-aE
- (l-Y)
--+ ax
x=o have not then the problem is essentially a 3-
dimensional one and the 15-equation set must be
25.25 employed. However, in some cases, this decision
is not a clear cut one. Consider, for example,
the case of a cylindrical shell of finite
length h, of diameter b, and which is loaded
normal to its axis, body forces are zero and
temperature is independent of the axial co-
where, again, E, a and Yare constants. These ordinate. If one considers the ratio of length
equations are derived directly by the substi- to diameter, hlb, the problem can be classified
tution of equations 25.4, .5, .6, .7, .8 and .9 as follows
into equations 25.2 and .3.
h » 1 plane strain
The specific solution is obtained by b
sUbstituting the general solution into the
boundary condition equations, 25.15, and h
solving for the unknown constants. b «1 plane stress
A25. 5 Plane Strain. h 1 3-dimensional

b
The plane strain problem requires that no
displacement exists in one of the coordinate It should be noted, that in actual fact, the
axes directions, say the z-direction. Thus plane stress problem is a three-dimensional
w = 0 and u and v are functions of x and y strain problem and the plane strain problem one
only. This is equivalent to requiring that of three-dimensions in stress due to the
the following strain components are zero, possible presence of EZ in the first and Oz in
the secon<Q
25.26
It follows that the 8 unknowns will be ox, Oy,
~XY' txy, EX' Ey, U and v. In this case of z Assume 1:~(X,y)
plane strain Oz will be a function of ox and oy v=v(x,y)
and is, therefore, not considered as an un-
known. The substitution of 25.26 into the 15 '?>XZ=OYZ=EZ=O
equations of A24, reduces the number of inde-
~xz=~yz=O
pendent equations to 8, the two equilibrium
equations, the 3 stress-strain relations and EX=Ex(X,y)
the 3 strain-displacement relations. Thus the
equation set is complete. Ey=Ey(X,y)
Unknowns
Oxy=oxy(x,y)
The condition that w = 0, displacements ------------y
normal to the x-y plane do not eXist, requires u °x=ox(x,y)
~-----------~
that any cut surface originally lying in the 0Y=Oy(x,y)
x-y plane remains parallel to this plane after
~XY=~Xy(x,y)
deformation. One such physical system in Figure A25. 2
which this situation will exist is that of a
long cylindrical body subjected to external Oz = f(ox,Oy) Known.
loads and temperature which are independent of The distinction between plane stress and
the z, or axial, component and all loads and plane strain problems is clearly portrayed by
body forces act normal to the axial direction comparison between figures 25.1 and 25.2.
(i.e. all loading lies in the x-y plane and is
independent of axial position in the body.)
Such systems are common in missiles and rockets By direct substitution of 25.26 into the
(the cylindrical shell body, the solid fuel general equations of Chapter A24 the following
8 equations of the plane strain problem result: resulting stress, strain and displacement com-
ponents would then be carried out and super-
2 equations of equilibriU~ imposed on the preceding solutions to obtain
the actual solution.
aox + a~XY + X =0
ax y 25.27 This problem arises from the fact that 0z
is prescribed by the solution of the complete
a~XY + aOy + y = 0 equation set. A similar problem would arise in
x cry 25.28 the plane stress problem if surface strains
were prescribed as boundary conditions which,
3 stress-strain relations (where EZ =0 of course, is never done.
• 0z = )} (ox+Oy) . Case 2. Displacement boundary conditions.

-EaT
The boundary condition equations become
1_)12 V
= --(Ox
E
--oy)+(l+).I)aT
1-))
25.29
u = f(s)
25.36
Ey = -E-1_)/2 ( 11
°Y-l_))oy)+(l+)I)aT 25.30
v = g(s)
and the third displacement w must be zero since
V' _ 2(1 +V) this was the governing equation of the plane
Oxy - E 'Lxy 25.31
strain problem.
3 strain-displacement relations Case 3. Mixed boundary conditions.
If the surface tractions are prescribed

25.32
over part of the surface and displacements over
the remainder then the equations of Case 1 and
_ av Case 2 must be applied where appropriate.
Ey - ay 25.33
However, it is repeated here for emphasis
Oxy - au + av that the linear thermoelastic analysis is usually
- ay ax 25.34
broken down into a thermal stress analysis
assuming all surfaces unrestrained and traction
Boundary Conditions free, and upon the results of this part is
superimposed an isothermal analysis in which
The boundary condition equations from the effect of external loading and satisfaction
Chapter A24 reduce to the following in each of prescribed boundary displacements is taken
case; into account.
Case 1. Traction boundary conditions. A25.6 Mathematical Equality of the Plane Stress and Plane
Strain Problems.
x = 0x l + 'Lxyill
It will be noted that if one makes the
Y = 0yill + 'Lxyl 25.35 following sUbstitutions,
In this plane strain problem there will result E
from the solution an apparent surface traction El. =1 _v 2
at the free ends given by
lJ
v~ =1 - )I
25.37
This is, of course, non-zero and therefore this

boundary condition is not satisfied under the
assumption of traction-free surfaces and only the strain-stress equations, 25.29, 25.30 and
by the most rare chance would this coincide 25.31, can be written
with a prescribed applied surface traction.
This apparent surface traction must be removed
EX
by the application of an equal and opposite
traction the results of which, as derived from
an isothermal analysis, must be superimposed 25.38
on the thermal stress analysis. If a surface
traction is prescribed over this and the other
surfaces a further isothermal analysis of the
These equations are exactly of the same form (b) Displacement Formulation
as the same relations 25.4, 25.5, 25.6 in the
plane stress problem. In fact the two equation In this case again only the equilibrium
sets are now identical except for the change equations expressed in terms of displacements
in the elastic constants as given above. Thus must be satisfied for the general solution.
the mathematical problem is identically the These two equilibrium equations are:
same and consequently the discussion of section
A25.4 pertaining to the mathematical formulation
is directly applicable to this plane strain
problem so long as the reader changes the
elastic constants E, Y and a to E1 , Y1 and a 1 •
To avoid any confusion these discussions are
briefly repeated here and the pertinent 25.43
equations rewritten.
A25. 7 Formulation of Plane Strain Problem.
As in the preceding case, the plane-strain

problem can be formulated in terms of stresses
or displacements. or, using E and Y, are:
(a) Stress Formulation:

The equations which must be satisfied by
the three stress components ox, Oy and ~xy are aE aT + X = 0
1 - 2)1 ax
aox + a~xy + X = 0 25.44
dX ay E a (au + av) + E v" v
equilibrium 25.39 2(1+)1) (1-2JJ) ay ax ay 2(1-)12)
a~y + a~xy + y
y ax
=0
aE .2T-+ Y =O
- (1 - 2V) ay
v" (ox+Oy)+ E1 a v" T = 0 compatabili ty
1
The specific solution is obtained by use of the
25.40 boundary condition equations just discussed.
=v" (ox+Oy)
+ Ea o"T A25.8 Concluding Discussion.
1-)1 v ,
where E, )J and a are constants, for the general The 2-dimensional formulation of plane
solution, and for the specific solution, the stress and plane strain just presented applies
boundary conditions just discussed. to the majority of structural systems with
which the engineering stress analyst will have
= =
If X Y 0 (body forces zero) the in- to deal. There will, however, arise heated
verse formulation using the stress function 0, structural shapes in which the stress, strain,
as defined in 25.19, is the most profitable and displacement components will not obviously
approach requiring only the solution to the be such as to comply with the restrictions im-
equation posed by the plane stress and plane strain
idealization. In such cases the general equa-
v· 0 = -~ a 1 'V" T tions of 3-dimensional thermoelasticity must be
25.41 employed. These equations should be simplified
= - ~V"T
1 -V
as much as possible by assuming appropriate
stress, strain, and displacement components
either zero or negligible. In all such problems
for the general solution. it is adVisable to investigate the thermal
stress analysis under the assumption of un-
If X and Y are non-zero and derivable restrained and traction free boundary conditions
leaving the evaluation of effects of external
f rom th e re 1a ti ons X -- - ax'
av Y -_ - av
ay as di s- loading and boundary displacement restraint to
cussed in A25.4a, the equation becomes an isothermal analysis. This approach will be
illustrated in the later chapters where the
v"0 + (1- )l1)V"Y =- E1 a 1 V"T. thermoelastic analysis of shells of revolution
is discussed.
25.42
or 0.'"11
Y
+ 1 - 2)/ o"Y
l_vv
= _~ o"T
l_))v,
A25.10
A25.9 Problems. body forces X and Y derivable from the function

V as defined in 25.18.
(1) By direct substitution and use of the
equilibrium equations 25.2 and 25.3 prove that (7) If a body force, R, acting only in
the compatability equation 25.12 the radial direction and independent of Q is
defined as
~2(OX + Oy) + Ea~2T =0
follows from the equation
where X and Yare body forces in the x and y
a 2-"
oxy a 2£x a 2£y directions respectively and defined by
- - - = ~ + --::--lOr
axay ay ax
(2) From the stress diagram in section
X =- av
ax
A25.3 derive the equilibrium equations
aox a'txy Y = av
- ay
--+--+x=o
ax ay
prove, from the simple transformation from
aoy a'txy cartesian to cylindrical coordinates, that
-+--+y=o
ay ax
(3) For the inverse solution technique

R = ar
av
establish that the condition of zero surface
traction can be written in terms of the stress and thus, for this special case, the equilibrium
function ¢ as equations become
¢ =0 aOr + ~ a'trQ + or - oQ + av = 0
ar r aQ r ar
a¢ _
an - 0 ~ aOQ + a'trQ + 2 'trQ =0
r aQ ar r
where n is the outward drawn normal.
(4) Prove that the equilibrium equations

in cylindrical coordinates are given by
aOr + ~ a'trQ + or - oQ = 0
ar r aQ r
1. aOQ
r aQ
+ ~ + ~
ar r
=0
(5) Prove that the definition of the
stress function in cylindrical coordinates is
given by the equations
'trQ =- -ara (--)

1 a¢
r aQ
REFERENCES
when the r - Q plane corresponds to the x-y (1) "Elasticity in Engineering" E. E. Sechler,
plane and body forces are absent. Galcit Aeronautical Series, John Wiley &
Sons, Inc., 1952.
(6) Establish the plane stress compat-
ability equation as given in the text and, by (2) "Applied Elasticity" Chi-Teh Wang, McGraw-
direct substitution for the elastic constants, Hill Book Co., Inc., 1953.
write the compatability equation for the plane
strain thermoelastic problem. Modify both (3) All references of Chapter A24.
these equations to include the eXistence of
CHAPTER A26
SELECTED PROBLEMS IN ELASTICITY AND THERMOELASTICITY
A26.1 Introduction.The preceding chapter pre- initial temperature, To, are completely pre-
sen~ed the general fonnulation of the two- scribed by the cross-sectional area A and the
dimensional equations of elasticity and thenno- length L. The only restraint is that imposed
elasticity. This chapter is devoted to the on the length L by the rigid walls. The
development of solutions of simple problems in problem is to determine the stress distribution
elasticity and th3nnoelasticity employing these throughout the bar and the change in shape when
equations rather than the Strength of Materials temperature is increased to T.
approach which will be discussed in the chapter
on bea~s. These problems h~ve been selected Assume the bar is unrestrained during the
to illustrate clearly the fonnulation of the heating. The change in length will be
governing equations for the chosen mathematical
model, the sO~ltion of these equations and the l\L a(T - To)L
interpretation of the results. A large store
of such problems exist in the literature, or, the new heated total length will be
especially in the case of the isothermll
analysis, which is referenced at the close of L~ = L [! + arT - To)]
this chaptero Consequently the problems chosen
for this cl1apter are to illustrate the technique T To '= Temperature change
only and, while they may appear trivial to
some readers, they are intended to provide a where a = Coefficient of linear thennal
reasonably sound knowledge of the tec~niques expansion
so that the literature can be read easily and
with confidence. In all of the problems of = Change in length per unit length
this chapter the basic assumntions of Chapter (at a chosen reference tempera-
A24 are observed and consequently the dis- ture level) per degree change in
cussion is restricted to elastic materials temperature.
whose properties are not affected by temper-
ature. In engineering it is common practice to
measure temperature chan~e in OF and to use the
The chapter opens with a selection of reference temperature as 32 0 F. So long as a is
simple one-dimensional problems which are inter- a constant the reference temperature could be
jected at this point merely to provide the as well chosen as the original temperature
student with an understanding of the behaviour since, by the definition above, the value of a
of metals when they are heated and the elastic would be the same. However, for most materials,
constants are not affected by the temperature a is a coefficient of the temperature and there-
rise. This section is then followed by a fore the values given in the following table are
selection of problems in which th3 fonnulations applicable only over the stated temperature
of the preceding chapter are employed.
TABLE A26.1
A26.2 One-Dimensional Problems. REPRESENTATIVE VALUES OF Q
(Original length measured at 32 0 F)
&"'{ample 1. Material Temp. OF Q
Consider a cylindrical bar of metal Aluminum (99.95%) 13.22 x 10- 8
68 - 212
mounted between two rigid walls as shown in Commercial 68 - 212 13.33 x 10- 6
Figure A26.1. Beryllium (98. 9%) 68 - 212 6.83 x 10- 6
Brass (66 Cu, 34 Zn) 32 - 212 10.42 x 10- 6
Bronze (93.5 Cu, 6. 5 Sn) 68 - 212 9.72 x 10- 6
Graphite 100 4.37 X 10- 6
Copper 70 - 212 9.33 x 10- 6
Glass 32 - 212 4.63 x 10- 6
Steel 32 - 212 5.88 x 10- 6
L Molybdenum 70 - 212 2.72 x 10- 6
Invar (Nickel Steel) 100 0.46 X 10- 6
Stainless Steel 70 - 212 5.33 x 10- 6
Fig. A26.1 Tungsten (99. 978%) - 212 2.33 x 10- 6
32 - 950 2. 56 x 10- 6
(powder metallurgy) 4000 5.65 X 10- 6
The geometrical properties of the bar at the Magnesium Oxide 75 - 212 5.6 x 10- 6
A26.1
A26.2 SELECTED PROBLEMS IN ELASTICITY AND THERMOELASTICITY
range. Furthermore, every alloy of the for each bar. The problem is to determine the
materials shown possess a characteristic Q. axial stress in each bar and the displacement
The table can, therefore, be used only as a of the interface under the assu~ption of no
guide to the appropriate values of Q and the restraints in the radial direction and no sur-
exact value obtained from a more extensive face tractions.
source.
Again we remove the restraint from one
Now, for the exa~ple under discussion, it end. The total expansion will be
can be seen that the bar will have to be com-
pressed an amou~t 6L to fit between the rigid
walls. This will give rise to a compressive
strain of 6L/L, or and the bars must be co~pressed by a common
force P so that this total displacement is re-
£ = - a(T - To) in the axial direction, moved.
and a compressive stress of
The stress in bar 1 will be a ~ -- A~
-P
a = - Ea(T - To)
Thermal expansion in the radial direction and the stress in bar 2 will be
as well as an elastic strain due to Poisson's
effect will also occur. However there will be The corresponding strains will be
no stresses in the radial direction since
P
there are no restraints or expansion in this
direction and no surface tractions are applied.
£~ =- A~E
Thus there will exist the compressive stress and
uniformly distributed over the cross-section £2 = - APE
2
throughout the length of the bar, if, and only
if, no restraints on the radial deformations The corresponding displace~ent, or change
exist at the ends against the walls. The in length, of each bar will be
total load reacting against the wall will be
P
u~ = - A E L~
~
where A~ is the area of the final cross-section

shape and thus the change in this cross-section
could be significant in the calculation of P if Now
the material possessed a large value for a and
the temperature rise was large. The Poisson
effect is usually negligible. It is left as an -P G~~E + A~~ = (a~L~ + a 2L2) (T - To)
exercise for the student to determine A~ in
(Q~L~ + a 2L2 ) (T - To)
Problem 26.1.
or P - - (J:.L+~)
Example 2. A~E A2E
Consider the system of two cylindrical It then follows that
bars supported between rigid, frictionless
walls and heated to temperature T above the
a~bient temperature To when the bars were a
perfect fit between the walls. The system is
shown below in Figure A26.2.
Bar 1 Bar 2
The final position of the interface will
be located at a distance x from the left hand
A~ A2 wall given by the original length of bar I plus
Q ~ Q2
L~ L2 its thermal elongation plus u~, or
E~ E2 (Young's modulus the
same in both bars. ) L~ (a~L~+a2L2)(T-To)
Fig. A26.2
x = L~ + a~L~ (or - To) - E (h + A~L2)
E A2 E
The geometrical properties and the linear It is interesting to note that this problem is
coefficient of thermal expansion are different essentially not this simple. The analysis re-
A26.3
quires the assu~ption that the interface stress and

distribution is un1i.orm across the section of
each bar, In the case of bar 1 this will
certainly not be the case, The extension to From the strain-displacement relations
the problem whsre this two-dimensional stress Where, in this case, u is a function of x only
field is taken into account is beyond the _ du~
scope of this chapter. It is sufficient to sx~ - dx
note that this effect will be negligible in
all cases except where the ratio of A~/ Az is
large and, again, where the value of P is large,
Besides this obvious extension of the It then follows that

problem the following meaningful studies are
apparent.
(a) The investigation of the stability of

the simple colu~n system and
u, {'t:,' + a,IT - To~ '"'
(b) The investigation of the flow
behavior at very high temperatures
for both or one of the materials.
These significant and difficult aspects of this

and u. iC·t:.. + a,IT - Toj '"'
seemingly simple problem will be returned to But u~ + U z =

0 if the interface is to rem"lin,
as such and no total elongation is allov,ed.
in later chapters.
Furthermore, since the same force is transmitted
through each bar,
Example 3,
The same problem will be analyzed in a
more direct manner Which d00s not require the
technique of considering the system completely
Com3equ!sntly
unrestrained to begin with and then detennining
the loads necessary to restore the original,
or defined, shape. \~h1le this st"lted technique
has been discussed in Chapters A24.4 and A24.5
as being the most desirable, the direct
approach is sometimes the more rational and,
therefore, must not be overlooked, This
Since ax, E, Ez , A~, Az , Q~ and Q z are constants
problem serves as a good comparative example
throughout the system the integration in this
at this st"lge in the development of the theory
of thennoelastic ity. case is straight forNard and since 0 -+ x~ = L~
an,i x ~ -+ X z =
Lz •
Consider the Figure A26.3 which is the
srune as Figure A26.2 but shows the positive
coordinate system, the boundary conditions with
a possible deformed position illustr"lted.
It then follows directly th"lt
~~
r
ul = 0
~ x
u
Al
Ql
n ,
x~
1
Az
Q z r
L~ Lz U z= 0
E~ Ez (Young's modulus
different)
Fig. A26.3 and
The total strain in the system will be
Which is the same as before. The inherent ad-

where
vantage in this direct approach is the gener-
alization of the problem to take into account
possible variations in cross-section shape and
A26.4 SEL ECT ED PROBLEMS IN ELASTICITY AND T HERMOELASTICITY
variation of the physical properties in the insertion of the pin will be

axial direction.
Example 4.
~,(T~-To)L-aa(Ta-To)L- (C~ +c a:2] - PL
E,A
A direct application of the preceding
examples is the determination of shear and and the motion of bar 2 will be EP~, Since
a a
tensile stresses developed in bolts of a bolted these motions must be the same it follows
connection. Consider the bolted connection directly that
between tVIQ dissimilar metals which are heated
to different temperatures T~ and T a above
ambient temperature To. Initial bolt clear-
ances are normally allowed in bolted connections
to relieve thermal stresses and such clear-
ances are shovm in Figure A26.4 as C~ and Ca'
In this chosen case, P will produce compression
in the bar 1 and tension in bar 2 since the ex-
pansion of bar 1 was initially greater than that
Bar 1
of bar 2 and it was assumed that this relative
expansion was greater than the total clearance,
T1 whiCh, of course, is necessary for the creation
of thermal stresses at all. The solution should
be generalized to
Bar 2
Fig. A26.4
Associated with bar 1 and bar 2 are the

and the appropriate signs chosen from an initial
material constants E~, u~ and E a , u a
estimate of the relative displacements. It is
respectively.
an interesting extension of this problem to
take into account the elastic deformation of
Determination of shear load transmitted: -
the bolt in an array of bolts.
Due to symmetry we need consider only one
The tension load on the bolt, additive to
side of the above system. Assume T~ > Ta and
any bolt prestress, can be deduced in the
u~ > u a (the generalization to other combin-
follOl'Ving way. The total expansion in the
ations is straight forward and will follow the
bars in the bolt axis direction will be (due
same development).
to thennal expansion) in each sheet.
Assume the bolt removed.
Bar 1 expands an amount a~(T~ - To)L = There will be an elastic contraction however,
displacement of hole center line. due to the bolt increased tension force P which
will be
Bar 2 expands an amount ua(T a - To)L
Thus the relative displac~nent of the hole and respectively.

center line will be
where A, and Aa are assumed bar areas affected
by the bolt tension. A commonly used estimate
of this area is that contained in the circle
The allowable clearance will be C, + Ca so
of two bolt diameters about the bolt center-
that the relative displacement between the two
line. Consequently the total bar deformation
bars necessary to realign the holes so that the
bolt will just fit will be given by in the bolt axis direction will be
If all of this motion is obtained by compress-

Now, the total extension in the bolt, assuming
ing bar 1, insert ins the bolt and then releas-
it to be at a temperature T3 , will be
ing the system the two bars will expand to-
gether to an equilibriwn position wherein the
same force, P, exists in each. The total
motion of the precompressed bar 1 after the
A26.5
where the sUJscript B refers to the bolto 1 - - L - - - - - - - - - +-I t
These extensions must, of course, be the

~ Y(v) -I 1-
same and therefore
z
~
~ V
~VL..-_--'
/
X
(u)
1 DT
M
(w) ~ ELEVATION SECTION
The extension of this last formlla to multiple Fig. A26.5

layers, bolts of varyinG diameters wherein the
temperature v':lries along the length, etc., is the same geometrically as that shown in the
obvious from th,s above development and the Figure A26.5 but with the following assumptions:
discussion in example 3 as it applies to
varying temperat'ure along ths bolt axis. (1) There are no displacements out of the
x-y plane, i.e., w : O.
Exdm)le 5.
It is becoming relatively co@non practice (2) All deformations are small so that the
to design primary structures to avoid thermal equations of the basic fonnulation
stresses. One obvious way of doing so is to apply.
replace the usual flanged wing beam with a
statically determinant truss for no thermal (3) Since the beam is thin (t is small),
stresses can exist in such structures so long the applied moment lies in the plane
as no restraints are imposed on the thermal x-y, and there are no surface forces
expansion of ths whole truss system. on the sides of the beam, thus there
exists no stress components in the
The use of corrugated webs is one example z-direction.
of a structural technique designod to relieve
thsrmal stresses. In such structures, however, (4) There are no body forces present, i.e.
the thsrmal strains and hence, deflections, X = Y = Z : O.
become large and must be considered.
From assumption (3) it is obvious that all
Redundant structures will contain thermal stresses lie in the plane x-y and the problem
stresses whenever non-uniform heating is is therefore one of plane stress. (It so
present even though there are no geometrical happens that under assu~ption (2), which is
restraints imposed. In the case of statically applied merely as a convenience at this time,
indeterminant trusses or space frames wherein the problem is also one of plane strain.)
individual members are at different temper-
atures the analysis is the same as for the The question now is to determine stress
isothermal problem wherein the readily com- functions for ox' 0y and Txy which will satisfy
puted thennal strains and deflections are the equation of eqUIlibrium, compatability and
introduced as initial deformations. This the boundary conditions. The basic equations
problem is discussed in detail in article A8.9 are: -
of Chapter A8 and consequently is not treated
hsre.
Attention will be now directed to continu-

ous structures such as plates and beams, equilibrium
wherein the two-dimensional fonnulations of aoy + aTxy = 0
Chapter A25 will be required. ay ax
A26.3 Two-dimensional Problems.
compatability
Example 1.
As the first exa~ple let us consider the
problem of the simple beam subjected to an and the boundary conditions
externally applied moment and with the whole
system at room temperature. The physical
system is as shown beloN in Figure A26.5.
The problem is to determine the stress,

strain and displacement distribution throughout
the beam under the influence of the positive
1:: 2
dY =0 (applied end thrust is zero)
end moments shown. The mathematical model is

f(y) it is obvious that
(summation of ~xy =°
moments = 0)
(This could have been assumed directly, of
course, since there is no externally applied
shear present, but the matter of zero internal
From assumption (3) shear has been belabored since many students
appear to have some difficulty in grasping this
0z = ~Xz = ~yz =° simple fact.)
an,1 thus Thus the only stress component present is
Ox and the equations of equilibrium are satis-
fZ =tf xz = -Iyz = ° fied. It is obvious also that since
Since the external loading is bending only
and there is no variation of same in the
x-direction, it is reasonable to assume the ~ Ox
strain distribution to be of the following f y - - E
form:
Ixy=o,
f X = Co + C],y (plane sections
remain plane) the equation of compatability is also satisfied.
Thus we have the unique general solution to the
fy =f (y)
mathematical model chosen.
and thus
0y = ~xy = ~xz = ~yz =0z = 0.

0y - E f(y)
It now remains to obtain the specific
From the first equilibriun equation it then solution which satisfies the boundary conditions.
follows that ~xy must be a constant or a By direct substitution into the boundary con-
function of x only. Then, from the second dition equations, the constants Co and C], are
equilibriQm equation readily evaluated. as
Co = 0
M M
C], = = -EI
or E df(y)
dy
= 0, a constant or a function of x.
E
i+~
.sl t;' dy
Due to the symmetry of loading the last 2

possibility does not eXist, the second possi-
bility allows f(y) to be linear function of y and thus Ox = - ~ y, the well known simple
and the third allows f(y) to be a constant.
Neither of these are possible since there are beam formula where the negative sign indicates
no normal forces along the top and bottom compression. Now the strain distribution
edges to balance a 0y component. The only throughout the beam will be given by
solution for f(y), and therefore for Oy, which
will satisfy the equilibrium equations is
therefore
f(y) =0yE =0
Thus we have two possible stress components and the displacement distribution, from the
strain-displacement relations
_ au
fX -~ My
~xy = f(x) u = -Er .X + f],(y)
_ av
fy - a
The first equation of equilibrium is satisfied y V g],(x) + YMy2
=
but, from the second and the discussion on 2IE""
"6xy = au + ~
ay ax
A26.7
Since u 0 at x = =
0, it follows that equations. From the first, viz;
f~(y) =
O. Then, from the third strain-
displacement relation it follows that
dg,(x) Mx
dx E1
and therefore it follows, as before, that Co = O.

From the second and letting C,g(x) - C2 (x)
d
Since, at x = 0, v = 0 then it follows that
h = 0 at y = O. Thus the displacement distri-
bution will be
F££~Y'dY + peL - xl = 0
2
u = l'1yx
E1
and therefore
peL - x)
E1
w = O. so that
Example 2. ax
=- P(L - x)
I Y
Consider now the same problem but with
the external load a transverse force, P,
applied at the end. From the first equation of equilibrium it
follows that
~ r-- L --------->l~
d'Txy = Py
~ l__ _
y
--+l X dY I
/
By integration
Fig. A26.6
'Txy =_ Py 2
+ f(x)
The same basic asswnptions are made as in 21
the previous example and the basic equations
are the same. The boundary condition Now, since there must be zero shear on the
equations are the same when the applied moment, top and bottom faces of the beam, i.e., at
11, is given by + d
x = - 2 it follows directly that f(x) is a
11 = peL - x) constant given by
We will again assume that the strain distri-
bution is of the following form
EX = (Co + C~y)g(x)
Ey = fey) From the second equation of equilibrium it

EZ =0 follows that
The assumption for EX is more meaningful day = 0

in this case for it limits the discussion to oy
beams such that the bending action is pre- and thus, from the same argument as in the
dominant. This means that the solution previous example, it follows that
developed under this assumption applies only
to beams where L» d. Let us consider one ay = 0
station, i.e., one value of x, along the beam
and consider first the boundary condition The strain distribution follows directly
from Hooke's law and the assumptions to give It is interesting to note that this
solution differs from that developed from the
EZ = ~XZ = Oyz = 0 Strength of Materials approach, The term
PLd" .
BIG arlses from the shearing force and is
P
EX = - Ef (L - X)y effectively a rotation at the end x =0 given
by
Ey =+ ~ P(L - X)y
EI oV I
ax
x=O
y=0
If this rotation is prevented, as it is

These strain components obviously satisfy
in this case, then the tip deflection should
the cOffi)atability eq~ations.
b"·
The displacement distribution can be _ PL3
obtained from the strain-displacement relations
v(y=o) - 3EI
as follows:
X=L
Since EX -
- au
ax -- - EI
P (L - x)y which would have been the solution had we
imDosed the boundary condition at the station
x = 0
it follows that u = _P~rx + ~:! + f(y)
av I =0
av y ax x=O
and since Ey = oy = EI P(L - x)y y=O
)/ y" The fixed end is not free to distort and

it follows that v = k(x) + EI peL - x) -z- thus allaN the displacement distrlbution
obtained above and therefore the stress and
Since u = 0 at x = 0 it follows directly strain distribution will not be that obtained
that f(y) =
O. From the third strain- in the vicinity of this fixed end. The solu-
displacement relation, viz: tion will be valid throughout the beam for all
points x > d (Saint-Venant's Principle). It
should be mentioned also that for the solution
to be valid in the neighborhood of the free
end the shear load P must be applied according
it follows, by direct substitution, that to the shear distribution
P (d" 2) = PLx + px" + av -

P d"
( - - y")
2IG ""4 - Y EI 2EI ax 21 4
We are interested in the deflection of the If this is not done then the solution will
neutral axis only, that is the plot of not apply in the region of the free end.
v (x, Y = 0) and therefore we can rewrite this
equation as This matter of the fixed end boundary
dV Pd" PLx Px" conditions deserves more study and is therefore
-
dx
=- - +EI
BIG
- - -2EI
- discussed in the next exa~ple. For an even
more eXhaustive discussion the reader is
Consequently we are imposing the boundary referred to Reference 1.
dV
condition that at x = 0 , dx = O.
Example 3.
Upon integration and satisfaction of the In the previous example the stress distri-
boundary condition v
= 0 at x = 0, the general bution was obtained as
expression for the deformation of the neutral
axis will be obtained as: P(L - x)y
I
_ Pxd" Px PLx" 3
v = BIG - 6EI + 2EI

cry =0
and the deflection at the tip will be
- PL 3 PLd"
vx=L = 3EI + BD
A26.9
Therefore, from Hooke's Law and assuming P d2 ypy2

V ~ 0, the strain distribution will be f' (y) = 2IG (4 - y2) + 2E'!- C2
By direct integration it then follows that

EX = - ~ (L - x)y
Pd2x + Px 2L Px 3
Ey = Eel
llP (L - x)y g ()
x = 8IG 2EI - 6EI - CJ.X + C3
j( _ P d2
iJ xy - 2IG (4 - y2)
Thus, the complete expressions for u and v
ani then, from the strain-displacement relations will be
u =- il
P (L - ~) + f (y) u = _ Pxy (L _~) + ~ (d 2y _ y3) + ).IPy3
EI 2 2IG 4 3 6EI
v =~
2EI
(L - X)y2 + g(x)
Pd 2x Px2L Px 3
- YP
where f(y) and g(x) are functions respectively V - - x) Y2 + 8IG + 2EI - 6EI
2EI ( L
of y and x alone. It is the aim in this
example to evaluate these functions in a more
general sense than was done in Example 2.
Substitution of the expressions for u and v Now to fix the beam at the end x = 0 it is
necessary to prevent a translation in the u and
in the third strain-displacement relation, viz: v directions and to prevent rotation at this
end. Thus the end conditions can be written as
u(o,o) = v(o,o) = 0
gives
and au I =0
Pd2
2IG (""4 - y2)
P
= - E~ (L -1) + f' (y) ayx--0 , y=o
_ I1P y
2'2.:1
2 + g' (x) on ~I =0
axx=o, y=O
where f' (y) = df

dy
and g' (x) =~ There is a choice on these last two since
we need only 3 boundary conditions in addition
to the equation
This equation must hold for all values of
x and y and we can therefore investigate the
nature of the functions f' (y) and g' (x) by
alternately letting x and y be zero. For
y = 0,
to evaluate CJ., C2 , C3 and C4 • Thus the bound-
2 ary conditions are over prescribed and different
Pd- P
- EI (L -~)
8IG -- - 2. 2 + f' (0) + g' (x) solutions for the displacement distribution can
be obtained by choosing different sets of three.
The choice is only two, however, since no
For x = 0 f' (y) _lIPy2 + g' (0) translation is allowed. The solution for the
2EI set containing the condition
For x = 0, y = 0
au I =0
Pd
2
ayx=o, y=O
8IG -"- f' (0) + g' (0)
is that of example 2. Therefore, let us consider
Now f' (0) and g' (0) are constants CJ. and the set containing the condition
C2 respectively. Therefore
, Pd 2 Px X
av I
ax
=0
g (x) = 8IG + EI (L -"2) - CJ. x=o, y=o
Applying these conditions and using the
other available equation in CJ. and C2 provides
the following solution for the displacements.

~xy
_w
-
(L2EI- x) [d""42- y
2]
u =_ PXY
EI
(L _ x) + ~(d2y
2 2IG 4
_L)
3
+)lPY"
6EI and the displacements are determined as before.
Px 2 L Px" The actual problem is not as simple how-
V = 2EIP )I ( 2
L - x)y +2EI
---- -
6EI ever as for the above solution where the upper
and lower edges of the beam are considered
and we note now that the tip deflection at stress free. The applied load as shown will
y = 0, x = L becomes require that Oy be now-zero on one, or both,
of these faces. In fact, all three stress
components ox' Oy and ~xy will be now-zero
throughout the beam. The solution is not
simple but the reader has now the techniques
Example 4. to attack it. The solution follows the same
Let us now consider the same beam with an steps as before with careful observation of
externally applied distributed load as shown the boundary conditions along the top and
in Figure 26.7. bottom edges where Oy is now-zero on the loaded
edge. The most simple example of this problem
is where w is constant along the beam. It is
~==~==========~'I ~ t l- left as an exercise for the reader to show
D1 that for this case and where the constant load

is hung on the bottom surface giving rise to
the boundary condition
Oy = w, at y =- 2"d
Fig. A26.7
that the stress components are:
It is common engineering practice to wt 2 wt y" wt d 2

ignore the local effects in the stress field Ox = T y(L - x) - T 3 + T 20 Y
due to the applied load and to thus retain
the argument that cry = 0 for all values of y. _ wt y" wt 2 wtd"
The only non-zero stress components are thus Oy - I 6" - 81 d Y + 124
Ox and ~xy. The solution technique is there-
fore the same as in the preceding example but wt 2 wt d2
where now the moment M is given in the form, ~xy = T (L - x) Y2 - T (L - x) ""8
for the coordinate system shown,
and the deflection of the neutral axis is
given by
L-X
M(x) =
j o
W(x)UdU
GI
X2 2
2
2 8 EI 4
2 2
2
6
4
v = _wt{Lx_ }d _wt{L X _LX" +x }
24
2
wt d 2 J:: wt d 2 +1:. wt d
EI 40 x - E I 16 x G I ""8 Lx
The solution follows directly the steps of the
preceding example. In the majority of problems General Note: Since we are dealing with linear
the variation of w along the beam is linear, problems, it is possible to break down any
lumped into a sequence of discrete loads, or w loading given into components matching those of
is a constant. In any of these cases this the problems given above and by direct super-
integral is readily evaluated in the initial position of the separate solutions corresponding
steps of the problem solution. For example, to these component loadings obtain the final
if w = constant, solution for the complex combined loading states.
This general technique will be employed to
obtain solutions to general thermoelastic
problems in beams.
and the stress field is given by EX8.t'llple 5.

Consider the cantilever beam shown SUb-
w (L _ x) 2
Ox - - 2EI Y jected to a non-linear temperature distribution
in the depth direction but constant in the
Oy =0 spanwise direction, T = T(y). This general
distribution is an approximation to the
A26.11
temperature distribution in high speed aircraft

wings.
(ii) The resultant moment on these faces

must be zero or
Assumptions: (i) The assumptions governing the

basic theory as discussed in
Chapter A25.
(ii) Oy = ~xy = Oy O.
(iii) The restraint to deformation
L~xtYdY
2
d
= C
v at the root (x = 0) is
negligible. This assumption Solution:
is questionable when the root
fixing is considered. The The general solution is obtained by direct
effect of the inclusion of integration as
root restraint on v can be
evaluated by superposing a Ox = - aET + C~ Y + C 2
self-equilibrating isothermal
stress pattern to restrain v. Use of the boundary conditions allows the
The effect will only be evaluation of C~ and C 2 as
appreciable in an area local
to the root of a beam (L» d)
and therefore will be
neglected.
(iv) Material properties are
independent of temperature.
Basic equations: (Refer again to Chapter A25.)

Since there are no boundary tractions along the
edges and it is assumed that there is no
resistance to expansion at the root station, it
follows that ~xy =O. It is also now apparent
that Ox will be a function of y only (ox
constant in the spanwise direction) for this
case of thermal loading only.
Hence, the stress distribution at any
The compatibility equation (See eq. 25.12), station is given by:
thus reduces to
d d
Since Ox and T are functions of y only, this ox' [-T+~L :dy ~; yL;Yd
oE + Y]
equation further reduces to the ordinary
differential equation 2 2
Eq. (A)
and it follows directly from the general stress-
strain relations with Oy = 0;
Boundary Conditions.
The cantilever beam is replaced by a beam
of length 2L to satisfy the condition imposed
by symmetry of zero deformation in the
x-direction at the root. On this beam all
surfaces must be zero. Hence the boundary
conditions on Ox can be written as;
and
(1) The reSUltant force over the end
face at x + L is zero or
that 0 will be a function of y only. The

general equation
reduces to
+ aT (1 + )) )
t XY = O.
Integrating both sides, with respect to y, will
and the displacements obtained by direct inte- give
gration of the strain-displacement relations
with satisfaction of the boundary conditions
u = 0 at x = o.
+ C",y + C4
and the boundary conditions on 0, necessary for

the evaluation of C., C2 , C", and C4 will follow
Again we are interested in the deformation of
the "neutral" axis or, in this case, any line from Chapter A25 as ¢ = ~~ = 0 at the boundaries
at y =constant. Since 2fxy =
0 it follows from
y = ~ %. The stresses can then be obtained
the strain displacement relations that
di rectly from
av _ au
ax - - ay
a2
v a2 u
and ax 2 = - axay and the strains and displacements obtained as
in the preceding example.
d
L
2 Note that if T is linear, or satisfies
=- 12 V2 T = 0,
then ¢ m 0 everywhere and consequently
a ~ _£ Tydy no stresses will result. Strains, however, will
exist and therefore deformations will result.
2
Example 7. Numerical example.
=constant.
Consider a cantilever beam with the follow-
Thus the curvature is constant along the ing dimensions and material properties.
length of the beam and v can be considered due
to an effective end moment of value L = 10 ins. a = 7 x 10-6
d d = 2 ins. E 30 x 10 6 (assumed constant)

, (2 1 i
t -- 8" = 0.3
~~ Yj_~
ns. )J
aE Tydy
and subject to the following chordwise thermal
2 gradient
Thus the solution is complete.
~
Example 6.
The preceding problem can, of course, be
solved employing the inverse formulation in
terms of the stress function ¢ as outlined in lOOOOF
Chapter A25.
Y~ d d
The thermal gradient has been chosen as a
2 0 -2
function of y only, and since cry = 0, it follows
A26.13
Assuming a parabolic temperature distribution chordwise varying temperature distributions by

it follows that employing direct superposition of separate
solutions. For example, consider the case of
T(y) = 50y" + 100y + 1050 the simple cantilever beam where the external
loading is an end moment M and the temperature
J.I ~dY
distribution is linear across the depth and
Then = 2133 ; (~YdY = 66. given by T(g) = A + By. The stress components
will then be given by
l )-1
Ox = ox~ (due to moment M) + Ox" (due to
Then Ox = 210 (-50y" + 16) From Eq. A temperature)
Oy =0
Thus Ox / d = -6,740 Ib./ins." compression
'JIxy = 0
y=::!:"2
and the displacements by
u = u~ (due to moment M) + u" (due to

OX! = 3360 Ib.lins." tension. temperature)
y =0 v = v~ (1ue to moment M) + v" (due to
The two symmetric nodes will be at temperature)
It is left as an exercise for the reader to
y = show that
= + 0.57 ins.
The stress profile is therefore approxi-

mately (numbers in solution are approximate) Ox" = 0
symmetric as shown in the plot
Myx
u~ = -Ef"
y
6740 lb. /ins." d u" =a (A + By) x .

--
~_--+---- +2" Compression
3360 lb. lins. "

v~
_ M
- 2EI (x" + JJy")
---+--+---- x Tension
v" = J.y (A + ~ By) - 2"
1 aBx "
Concluding Remarks.
Station x. The examples given above illustrate how

problems in two-dimensional elasticity are
It should be noted that the boundary attacked and how problems involving complex
condition of zero normal stresses at the loading states can be simplified by consider-
boundary x = ! L is not satisfied. The solution ation of separate more simple components of the
obtained is valid only over the range x = loading state and obtaining the final solution
! (L-d) according to st. Venant's principle. by employing superposition. This technique is
If the solution is to be extended over the further illustrated in those problems dealing
range - L - x - L then a self-equilibrating with thermal stresses.
stress-distribution must be applied at the ends
to balance the normal stresses. This is a It should be borne in mind that the
difficult problem and beyond the scope of this simplification of the general three-dimensional
discussion. elasticity problem to the two-dimensional plane-
stress or plane-strain problem must be done with
Example 8. care and the consequences of this idealization,
In accordance with the discussion of as it is in the majority of problems, not over-
Chapter A25 it is possible now to obtain the looked.
solutions to the simple beam problem when the All of the illustrative problems presented
beam is sUbjected to external loadings and in this chapter are almost mathematically
trivial problems in the theory of elasticity the inclusion of the kinematical variables) and
and thermoelasticity. We have restricted our that the elasticity problem due to this
attention to simple beam problems where the temperature distribution can be considered
bea~ depth is small compared to the length and separately. This assumption is valid only if
where the temperature varies across the depth the rate processes involved are slow.
only. However, the basic equations and
solution techniques are basic to the solution REFERENCES
of more complex problems where the temperature
is a function of two spatial coordinates, 1. Boley, Bruno A. and Weiner, Jerome H.,
T = T(x,y), and the geometry is that of a "Theory of Thermal stresses". John Wiley
general two dimentional plate. In these & Sons, New York, 1960.
problems all the stress components will exist
and displacements in the z-direction may occur 2. Gatewood, B. E., "Thermal stresses" with
in the general two-dimensional plate problem. Application to Airplanes, Missiles,
Generalization of the teclLniques illustrated Turbines and Nuclear Reactors. McGraw
here in this class of problems is beyond the Hill, New York, 1957.
scope of this discussion and the r~ader who
wishes to pursue this sUbject is referred to 3. Timoshenko, S. H. and Goodier, J. N.,
references listed in next column which can now "Theory of Elasticity", 2nd Ed., Chapter 14,
be read with ease. McGraw Hill, New York, 1951.
An important assumption basic to all of 4. Sokolinkoff, I. S., "Mathematical Theory of

the preceding discussion on thermoelasticity Elasticity", 2nd Ed., Section 101, McGraw
and inherent in the references cited is that Hill, New York, 1956.
there exists no "thermo-mec11anical coupling".
This allows the calculation of the temperature
field as a separate problem with no cross-
coupling with the mechanical response (without
PART B
FLIGHT VEHICLE MATERIALS AND THEIR PROPERTIES
CHAPTER Bl
BASIC PRINCIPLES AND DEFINITIONS
Bl. 1 Introduction. Failure in a different manner can result

from a structural unit being too fleXible, and
The flight vehicle structures engineer this fleXibility might influence aerodynamic
faces a major design requirement of a high forces sufficiently as to produce unsatisfac-
degree of structural integrity against failure, tory vehicle flying characteristics. In some
but with as light structural weight as possible. cases this flexibility may not be serious
structural failure in flight vehicles can relative to the loss of the vehicle, but it is
often prove serious relative to loss of life still a degree of failure because changes must
and the vehicle. However, experience has be made in the structure to provide a satis-
shown that if a flight vehicle, whether factory operating vehicle. In some cases,
military or commercial in type, is to be excessive distortion such as the torsional
satisfactory from a payload and performance tWist of the wing can be very serious as this
standpoint, major effort must be made to save excessive deflection can lead to a build-up
structural weight, that is, to eliminate all of aerodynamic-dynamic forces to cause flutter
structural weight not required to insure or violent vibration which can cause failure
against failure. involving the loss of the airplane.
Since a flight vehicle is sUbjected to To illustrate another degree of failure

various types of loading such as static, of a structural unit, consider a wing bUilt in
dynamic and repeated, which may act under a fuel tank. The stiffened sheet units which
wide range of temperature conditions, it is make up the tank are also a part of the wing
necessary that the structures engineer have a structure. In general these sheet units are
broad knowledge of the behavior of materials designed not to wrinkle or buckle under air-
under loading if safe and efficient structures plane operating conditions in order to insure
are to be obtained. This Part B provides against leakage of fuel around riveted con-
information for the structures design engineer nections. Therefore if portions of the tank
relative to the behavior of the most common walls do wrinkle in operation resulting in
flight vehicle materials under load and the fuel leakage, which in turn require repair or
various other conditions encountered in flight modification of the structure, we can say
environment, such as elevated temperatures. failure has occurred since the tank failed to
do its job satisfactory, and involves the item
Bl. 2 Failure of Structures. of extra expense to make satisfactory. To
illustrate further, the flight vehicle is
A flight vehicle like any other machine, equipped with many installations, such as the
is designed to do a certain job satisfactorily. control systems for the control surfaces, the
If any structural unit of the vehicle suffers power plant control system etc., which involves
effects which in turn effects in some manner many structural units. In many cases exces-
the satisfactory performance of the vehicle, sive elastic or inelastic deformation of a
the unit can be considered as having failed. unit can cause unsatisfactory operation of
Failure of a structural unit is therefore a the system, hence the unit can be considered
rather broad term. For example, failure of a as having failed although it may not be a
structural unit may be due to too high a stress serious failure relative to causing the loss
or load causing a complete fracture of the unit of the vehicle. Thus the aero-astro
while the vehicle is in flight. If this unit structures engineer is concerned with and
should happen to be one such as a wing beam or responsible for preventing many degrees of
a major wing fitting, the failure is a serious failure of structural units which make up the
one as it usually involves loss of the airplane flight vehicle and its installations and
and loss of life. Likewise the collapse of a obviously the greater his knowledge regarding
strut in landing gear structure during a land- the behavior of materials, the greater his
ing operation of the airplane can be a very chances of avoiding troubles from the many
serious failure. Failure of a structural unit degrees of structural failure.
may be due to fatigue and since fatigue failure
is of the fracture type Without warning indi- Bl. 3 General Types of Loading.
cations of impending failure, it can also prove
to be a very serious type of failure. Failure of a structural member is in~
fluenced by the manner in which the load is
Bl.l
B1.2 BEHAVIOR OF MATERIALS AND THEIR PROPERTIES
applied. Relative to the length of time in thus repeated loads are important in design of
applying the load to a member, two broad structures.
classifications appear logical, namely,
(1) Static loading and (2) Dynamic loading. Dynamic or Impact Loading. A dynamic or impact
For purposes of explanation and general loading when applied to a member produces
discussion, these two broad classifications appreciable shock or vibration. To produce
will be further broken down as follows:- such action, the load must be applied far more
rapid than in a static loading. This rapid
Continuous Loading. application of the load causes the stresses
Gradually or slowly applied in the member to be momentarily greater than
Static Loading loading. if the same magnitUde of load was applied
{ Repeated gradually applied statically, that is slowly applied. For
loading. example, if a weight of magnitUde W is
gradually placed on the end of a cantilever
Impact or rapidly applied beam, the beam will bend and gradually reach
Dynamic Loading loading. a maximum end deflection. However if this
{ Repeated impact loading. same weight of magnitude W is dropped on the
end of the beam from even such a small height
Static Continuous Loading. A continuous load as one foot, the maximum end deflection will
is a load that remains on the member for a long be several times that under the same static
period of time. The most common example is the load W. The beam will Vibrate and finally
dead weight of the member or the structure it- come to rest with the same end deflection as
self. When an airplane becomes airborne, the under the static load W. In bringing the
weight of the wing and its contents is a con- dynamic load to rest, the beam must absorb
tinuous load on the wing. A tank subjected to energy equal to the change in potential energy
an internal pressure for a considerable period of the falling load W, and thus dynamic loads
of time is a continuous load. Since a contin- are often referred to as energy loads.
uous load is applied for a long time, it is a
type of loading that provides favorable con- From the basic laws of Physics, force
ditions for creep, a term to be explained later. equals mass times acceleration (F = Ma) and
For airplanes, continuous loadings are usually acceleration equals time rate of change of
associated with other loads acting simultan- velocity. Thus if the velocity of a body
eously. such as an airplane or missile is changed in
magnitude, or the direction of the velocity
Static Gradually or Slowly Applied Loads. A of the vehicle is changed, the vehicle is
static gradually applied load is one that accelerated which means forces are applied to
slowly builds up or increases to its maximum the vehicle. In severe flight airplane
value without causing appreciable shock or maneuvers like pulling out of a dive from
vibration. The time of loading may be a high speeds or in striking a severe trans-
matter of seconds or even hours. The stresses verse air gust when flying at high speed, or
in the member increases as the load is in- in landing the airplane on ground or water,
creased and remains constant when the load the forces acting externally on the airplane
becomes constant. As an example, an airplane are applied rather rapidly and are classed
which is climbing with a pressurized fuselage, as dynamic loads. Chapter A4 discusses the
the internal pressure loading on the fuselage subject of airplane loads relative to Whether
structure is gradually increasing as the they can be classed as static or dynamic and
difference in air pressure between the inside how they are treated relative to design of
and outside of the fuselage gradually increases aircraft structures.
as the airplane climbs to higher altitudes.
Bl. 4 The Static Tension Stress-Strain Diagram.
Static Repeated Gradually Applied Loads. If a
gradually applied load is applied a large num- The information for plotting a tension
ber of times to a member it is referred to as a stress-strain diagram of a material is ob-
repeated load. The load may be of such nature tained by loading a test specimen in axial
as to repeat a cycle causing the stress in the tension and measuring the load With corres-
member to go to a maximum value and then back ponding elongation over a given length, as
to zero stress, or from a maximum tensile the specimen is loaded statically (gradually
stress to a maximum compressive stress, etc. applied) from zero to the failing load. To
The situation envolving repeated loading is standardize results standard size test
important because it can cause failure under specimens are specified by the (ASTM) American
a stress in a member which would be perfectly Society For Testing Materials. The speed of
safe, if the load was applied only once or a the testing machine cross-head should not
small number of times. Repeated loads usually exceed 1/16 inch per inch of gage length per
cause failure by fracturing Without warning, minute up to the yield point of the material
Bl. 3
and it should not exceed 1/2 inch per inch of ~--~----

gage length per minute from the yield point to
the rupturing point of the material. The
instrument for measuring the elongation must
be calibrated to read 0.0002 inches or less.
The information given by the tension stress-
strain diagram is needed by the engineer since
it is needed in strength design, rigidity Proportional Limit
design, energy absorption, quality control and
many other uses. (a) Material Having a Definite
Yield Point (such as some Steels)
Fig. Bl.l shows typical tensile stress- Strain - Inches Per Inch
strain diagrams of materials that fall in three
broad classifications. In the study of such
diagrams various facts and relationships have ___- - - - " 7 - -.
been noted relative to behavior of materials Ultimate Tensile Stress
and standard terms and symbols have been pro-
vided for this basic important information.
These terms will be explained briefly. (/)
(/)
Ql
1-0 Proportional Limit (b) Materials not Having a
Modulus of Elasticity (E). The mechanical t.l Definite Yield Point (such as
property that defines resistance of a material Aluminum Alloys, Magnesium,
in the elastic range is called stiffness and and Some Steels)
for ductile materials is measured by the value Strain - Inches Per Inch
termed Modulus of Elasticity, and, designated
by the capital letter E. Referring to Fig.
Bl.l, it is noticed that the first part of all
three diagrams is a straight line, which indi- I
Ultimate Tensile
cates a constant ratio between stress and ~ Stress
strain over this range. The numerical value of I
(/)
this ratio is referred to as the modulus of (/)
Ql
Elasticity (E). E is therefore the slope of 1-0
the initial straight portion of the stress- t.l

strain diagram and its numerical value is (c) Clad Aluminum Alloys
obtained by dividing stress in pounds per
square inch by a strain which is non-dimensiona Strain - Inches Per Inch
or E = fiE, and thus E has the same units as
stress, namely pounds per square inch. Fig. Bl.l
The clad aluminum alloys have two E values proportional limit stress.
as indicated in the lower diagram of Fig. Bl.l.
The initial modulus is the same as for other Tensile Yield Stress (Fty). In referring
aluminum alloys, but holds only up to the pro- to the upper diagram in Fig. Bl.l, we find
portional limit stress of the soft pure that some materials show a sharp break at a
aluminum coating material. Immediately above stress considerably below the ultimate stress
this point there is a short transition stage and that the material elongates considerably
and the material then exhibits a secondary with little or no increase in load. The
modulus of Elasticity up to the proportional stress at which this takes place is called
limit stress of the stronger core material. the yield point or yield stress. However many
This second modulus is the slope of the second materials and most flight vehicle materials do
straight line in the diagram. Both modulus not show this sharp break, but yield more
values are based on a stress using the gross gradually as illustrated in the middle diagram
area which includes both core and covering of Fig. Bl.l, and thus there is no definite
material. yield point as described above. Since
permanent deformations of any appreciable
Tensile Proportional Limit Stress. (Ftb). The amount are undesirable in most structures or
proportional limit stress is that stress which machines, it is normal practice to adopt an
exists when the stress strain curve departs arbitrary amount of permanent strain that is
from the initial straight line portion by a considered admissible for design purposes.
unit strain of 0.0001. In general the pro- Test authorities have established this value
portional limit stress gives a practical of permanent strain or set as 0.002 and the
diViding line between the elastic and inelastic stress which existed to cause this permanent
range of the material. The modulus of strain when released from the material 1s
elasticity is considered constant up to the called the yield stress. Fig. Bl.l shows how
B1.4 BEHAVIOR OF MATERIALS AND THEIR PROPE RTIES
it is determined graphically by drawing a line Figs. Bl.3 and Bl.4 compare the shapes of
from the 0.002 point parallel to the straight the tension stress-strain curves for some
portion of the stress-strain curve, and where common aircraft materials.
this line intersects the stress-strain curve
represents the yield strength or yield stress. B1. 5 The Static Compression Stress-Strain Diagram.
Ultimate Tensile Stress (Ftu). The ultimate Because safety and light structural weight
tensile stress is that stress under the maxi- are so important in flight, vehicle structural
mum load carried by the test specimen. It design, the engineer must consider the entire
should be realized that the stresses are based stress-strain picture through both the tension
on the original cross-sectional area of the and compressive stress range. This is due to
test specimen without regard to the lateral the fact that buckling, both primary and local,
contraction of the specimen during the test, is a common type of failure in flight vehicle
thus the actual or true stresses are greater structures and failure may occur under stresses
than those plotted in the conventional stress- in either the elastic or plastic range. In
strain curve. Fig. Bl.2 shows the general general the shape of the stress-strain curve
relationship between actual and the apparent as it departs away from the initial straight
stress as plotted in stress-strain curves. line portion, is different under compressive
The difference is not appreciable until the stresses than when under tensile stresses.
higher regions of the plastic range are Furthermore, the various flight vehicle
reached. materials have different shapes for the region
lIJ , . . - - - - - - - - - - ,
of the stress-strain curve adjacent to the
lIJ
Ql Actual Stres~ / straight portion. Since light structural
M
.....
00 '" weight is so important, considerable effort is
Ql made in design to develop high allowable
::::lIJ compressive stresses, and in many flight
5i
Eo<
Apparent j vehicle structural units, these allowable
..... Stress Ultimate design compressive stresses fall in
~ L..- -' the inelastic or plastic zone.
Fig. B1. 2
Unit Strain
Fig. 81.5 shows a comparison of the
stress-strain curves in tension and compres-
sion for four Widely used aluminum alloys.
240 Below the proportional limit stress the
modulus of elasticity is the same under both
0 200 tension and compressive stresses. The yield
...
0
0 stress in compression is determined in the
same manner as explained for tension.
...
~
00
160
P. Compressive Ultimate Stress (F cu )' Under a
lIJ~ 120 static tension stress, the ultimate tensile
lIJ
Ql
M stress of a member made from a given material
i15 80 is not influenced appreciably by the shape of
the cross-section or the length of the member,
40 however under a compressive stress the
ultimate compressive strength of a member is
.05 greatly influenced by both cross-sectional
shape and length of the member. Any member,
unless very short and compact, tends to
Fig. B1. 3
buckle laterally as a whole or to buckle
200
laterally or cripple locally when under
compressive stress. If a member is quite
0 short or restrained against lateral buckling,
160
...
0
0
~
then failure for some materials such as stone,
wood and a few metals will be by definite
Fip..l 120 fracture, thus giving a definite value for
the Ultimate compressive stress. Most air-
ai 80 craft materials are so· ductile that no fracture
lIJ
Ql
M
is encountered in compression, but the material
i15 yields and swells out so that the increasing
cross-sectional area tends to carry increasing
load. It is therefore practically impossible
Unit Strain to select a value of the ultimate compressive
stress of d uctlle materials without having
Fig. B1. 4 Stress-Strain Curves. Initial Portions.
Bl.5
70 70
24S-T3 SHEET 24S-T4 EXTRUSIONS
THICKNESSc O.2~O-IN THICKNESS < 0.2~0-IN. / E,
Ej'
-
60 60
~. X-COMPRES;:'O:'\
W-TENSION _ TENSI~N"""\ f ------,
I ;'
'" J.,..-- ....-r , I

50
-'\
<II """'- I/o ~NSION
50
"'- ........... Ii I
40 " "- ~ "'-W'COMPRESSION
........ 40
u::;:N
7"/(11. !
:!
J~r-;- ~ / r--... 51
.,
"
<II
<II
30 W 30
20
....rr<II
V 20
/
Sfrain e
10
/ 10 /
Fig. Bl. 6
0
0
/ 2 4 6 8 10 00
V The slope of this tangent gives the local rate
of change of stress With strain. The secant
e X 10' IN liN E X 10-· PSI
e
2 4
X 10' IN liN
6 8 10
E X 10-· PSI
modulus Es is determined by drawing a secant
(straight line) from the origin to the point
80 .---,---,--~:-:-c::-~-:--:-~--,
W- COMPRESSION ..I
I COMPRE SSI0J.
TENSION I _
in question. This modulus measures the ratio
.. X - etlMPRESSION --.. IL 80 t--.......t - - - t - - j - - - + '1;;,_'-"""",,! between stress and actual strain. Curves
701--"'" W-TENSION~rt: r---r---L~ which show how the tangent modulus varies
601--
~~~ENSIO~ 70 / .............
"\
With stress are referred to as tangent modulus
curves. Fig. Bl.5 illustrates such curves
50<11 I~~ 60<11 /

for four different aluminum alloys. It should
be noted that the tangent modulus is the same
as the modulus of elasticity in the elastic
40 ; / '50 ; -+---+V-+--+----+--I+----1 range and gets smaller in magnitude as the
~
VI ~
stress gets higher in the plastic range.
40 /
30 ~ ~ --+--+--t---t---t--+-t---t Bl. 7 Elastic - Inelastic Action.
V
<II 30
201--
/ f,\Lct.AD TIlS-n SHEET ~ 20/-----.f---t--t--+--t+---1
If a member is SUbjected to a certain
'0 L AND PLATE /
stress, the member undergoes a certain strain.
I If this strain vanishes upon the removal of
0
I' THICKNlESS 0·101I - 0 •4
I"-IN
•
10I-j-+---+----+----+---+--+I-----l
75S-T6 EXTRUSIONS the stress, the action is called elastic.
0 2 4 6 • 10 THICKNESS. 3-IN
Generally speaking, for practical purposes,
0
e X 10· INIIN E X 10" PSI 0 2 4 6 8 10
12 a material is considered elastic under stresses
X-Crossgrll.ln e X 10' IN liN E X 10-· PSI
up to the proportional limit stress as
Fig. Bl. 5 previously defined. Fig. Bl.7 illustrates
elastic action. However, if when the stress
is removed, a residual strain remains, the
some arbitrary measure or criteron. For action is generally referred to as inelastic
wrought materials it is normally assumed that or plastic. Fig. Bl.8 illustrates inelastic
Fcu equals Ftu' For brittle materials, that action.
are relatively weak in tension, an Fcu higher
!~:~~t
than Ftu can be obtained by compressive tests Elastic Action
of short compact specimens and this ultimate
: t~OP'Umit
compressive stress is generally referred to
as the block compressive stress.
Bl. 6 Tangent Modulus. Secant Modulus. ~

Modern structural theory for calculating Strain 1" • r- Strain
the compressive strength of structural members
Fig. Bl. 7 Fig. Bl. 8
as covered in detail in other chapters of this
book, makes use of two additional terms or
values which measure the stiffness of a member 81.8 Ductility.
when the compressive stresses in the member
fall in the inelastic range. These terms are The term ductility from an engineering
tangent modulus of elasticity (Et) and secant standpoint indicates a large capacity of a
,modulus of elasticity (Es )' These two modi- material for inelastic (plastic) deformation
fications of the modulus of elasticity (E) in tension or shear without rupture, as
apply in the plastic range and are illustrated contrasted with the term brittleness which
in Fig. Bl.6. The tangent modulus Et is indicates little capacity for plastic de-
determined by drawing a tangent to the stress- formation without failure. From a physical
strain diagram at the point under consideration. standpoint, ductility is a term which measures
B1. 6 BEHAVIOR OF MATERIALS AND THEIR PROPERTIES
the ability of a material to be drawn into a

wire or tube or to be forged or die cast. Stress-Strain
Ductility is usually measured by the percentage Curve"\.
elongation of a tensile test specimen after
failure, for a specified gage length, and is
usually an accurate enough value to compare
materials. f
dE
Percent elongation = (La L- 0L 0) 100 = measureof
ductility.
H I K
where Lo = original gage length and La = gage
length after fracture. In referring to
ductility in terms of percent alongation, it
is important that the gage length be stated, high and the strain at this stress is high,
since the percent elongation will vary with or from equation (2), when the proportional
gage length, because a large part of the total limit stress is high and modulUS of elasticity
strain occurs in the necked down portion of the is low.
gage length just before fracture.
In Fig. Bl.9 if the stress is released
B1. 9 Capacity to Absorb Energy. Resilience. Toughness. from point D in the plastic range, the recovery
diagram will be approximately a straight line
Resilience. The capacity of a material to DD parallel to AO, and the area CDE represents
absorb energy in the elastic range is referred the energy released, and often referred to as
to as its resilience. For measure of hyper-elastic resilience.
resilience we have the term modulus of
resilience, which is defined as the maximum Toughness. Toughness of a material can be
amount of energy per unit volume which can be defined as its ability to absorb energy when
stored in the material by stressing it and stressed in the plastic range. Since the
then completely recovered when the stress is term energy is involved, another definition
removed. The maximum stress for elastic would be the capacity of a material for
action for computing the modulus of resilience resisting fracture under a dynamic load.
is usually taken as the proportional limit Toughness is usually measured by the term
stress. Therefore for a unit volume of modulus of Toughness which is the amount of
material (1 cu. in.) the work done in stressing strain energy absorbed per unit volume when
a material up to its proportional limit stress stressed to the ultimate strength value.
would equal the avera~e stress f p /2 times the
elongation (E p ) in one inch. If we let U In Fig. Bl.9, let f equal the average
represent modulus of resilience, then stress over the unit strain distance dE from
F to G. Then work done per unit volume in
stressing F to G is fdE which is represented
U =(: p) Ep ----- --- -- ---- - (1) by the area FGHI. The total work done in
stressing to the ultimate stress f u would
Ek
But Ep = fplE, hence
then equal I fdE, which is the area under
the entire stress-strain curve up to the
=(f;) (f~)
ultimate stress point, or the area 0 A J K 0
U - fp2/2E (2) in Fig. Bl.9 and the units are in. lb. per
cu. inch. Strictly speaking it should not
include the elastic resilience or the energy
absorbed in the elastic range, but since this
Under a condition of axial loading, the modulus area is small compared to the area under the
of resilience can be found as the area under curve in the plastic range it is usually
the stress-strain curve up to the proportional included in toughness measurements.
limit stress. Thus in Fig. Bl.9, the area OAB
represents the energy absorbed in stressing It should be noted that the capacity of
the material from zero to the proportional a member for resisting an aXially applied
limi t stress. dynamic load is increased by increasing the
length of a member, because the volume is
High resilience is desired in members increased directly With length. However, the
subjected to shock, such as springs. From Ultimate strength remains the same since it
equation (1), a high value of resilience is is a function of cross-sectional area and not
obtained when the proportional limit stress is of volume of the material.
Bl. 7
Toughness is a desirable characteristic B1.11 Construction of a Stress-strain Curve Through a

when designing to resist impact or dynamic Given Yield stress by Using a Known Test Stress-
loads as it gives a reserve strength or factor Strain Curve.
of safety against failure by fracture when
over-loading in actual use should haDpen to Materials in general are produced to
cause the member to be stressed fairly high satisfy certain guaranteed minimum strength
into the plastic zone. properties such as yield stress, ultimate
tensile stress, etc. Thus in the design of
The property of ductility helps to produce important members in a composite structure,
toughness, but does not alone control toughness the minimum guaranteed properties must be used
as illustrated in Fig. Bl.lO, which shows the to provide the required degree of safety. In
general, most materials will give properties
slightly above the guaranteed values, thus
the test stress-strain curve of purchased
material cannot be used for design purposes.
The stress-strain curve passing through a
given yield point stress can be readily ob-
tained for a test stress-strain curve as
follows:
Unit Strain
Fig. B1.10 In Fig. Bl.ll, the heavy curve 0 B C
represents a known typical stress-strain curve
stress-strain curves for three different for the given material. Let the minimum
materials. Material (A) is strong but brittle, guaranteed yield stress be the value as shown
whereas material (C) is weak and ductile, and at point A, using the 0.2 percent method.
material (B) represents average strength and Then proceed as follows:
ductility. However, all three materials have
the same modulus of toughness since the areas (1) Draw a straight line through point 0 and
under all three curves is the same. point A which will intersect the typical
curve at point B. Point B may be above
B1.10 Poisson's Ratio. or below the typical curve.
When a material is stressed, it will (2) Locate any other point on the typical
deform in the direction of the stress and also curve such as point C, and draw line
at right angles to it. For axial loading and from 0 through C.
for stress below the proportional limit stress,
the ratio of the unit strains at right angles (3) Locate point D on line 0 C by the follow-
to the stress, to the unit strain in the ing ratio:-
direction of the stress is called Poisson's
ratio. It is determined by direct measurement OD = OB
OA x OC
in a tensile or compressive test of a specimen,
and is approximately equal to 0.3 for steel (4) Repeat step 3 to obtain a number of
and 0.33 for non-ferrous materials. In many points as shown by dots on Fig. Bl.ll,
structures there are members which are subjected and draw smooth curve through these points
to stresses in more than one direction, say to obtain desired stress-strain curve.
along all three coordinate axes. Poisson's
ratio is used to determine the reSUltant stress
and deformation in the various directions.
Poisson's Ratio in Plastic Range. Information
is somewhat limited as to Poisson's ratio in f
(PIll
the plastic range and particularly during the
transition range from elastic to plastic action.
For the assumption of a plastically
incompressible isotropic SOlid, Poisson's ratio
assumeS a value of 0.5. Gerard and Weldform
found from their research, that the transition ~ (1Dcbu/1IIcbl
of Poisson's ratio from the elastic value of Fig. B1.11
around 1/3 to 1/2 in the plastic range is
gradual and is most pronounced in the yield B1. 12 Non-Dimensional stress-strain Curves.
point region of the stress-strain curve.
The structural designer is constantly
confronted with the design of structural units
which fail by inelastic instability. The
solution of such problems requires information on the stress-strain curve. Table Bl.l gives
given by the compressive stress-strain curve. the values of F o • 7 , F o • s &, n, etc., for many
Since flight vehicles make use of many differ- flight vehicle materials. Notice that the
ent materials, and each material usually has shape parameter varies Widely for materials,
many different states of manufacture which being as low as 4 and as high as 90.
give different mechanical properties, the
question of time required to obtain certain B1. 13 Influence of Temperature on Material Properties.
design information from stress-strain curves
becomes important. For example, in the Before the advent of the supersonic air-
aluminum alloys alone there are about 100 plane or the long range missile, the aero-
different alloys, and when elevated temperatures nautical structures engineer could design the
at various time exposures are added, the number airframe of aircraft using the normal static
of stress-strain curves required is further mechanical properties of materials, since the
greatly increased. temperature rise encountered by such aircraft
had practically no effect on the material
Fortunately, this time consuming work was strength properties. The development of the
greatly lessened when Ramsberg and Osgood turbine jet and rocket jet power plants pro-
(Ref. 1) proposed an equation to describe the vided the means of opening up the whole new
stress-strain curve in the yield range. Their field of supersonic and space flight. The
proposed equation specifies the stress-strain flight environmental conditions were now
curve by the use of three parameters, the greatly expanded, the major change being that
modulus of elasticity E, the secant yield aerodynamic heating caused by high speeds in
stress F o • 7 , which is taken as the line of the atmosphere caused surface temperatures
slope 0.7E drawn from origin (see Fig. Bl.12), on the airframe which would greatly effect
and a parameter n which describes the shape the normal static material strength propertIes
of the stress-strain curve in the yield region. and thus temperature and time became important
In order to evaluate the term n, another stress in the structural desIgn of certain types of
F o s& is needed, which is the intersection of flight vehicles.
the curve by a line of slope of O.85E through
the origin (see Fig. Bl.12). B1. 14 Creep of Materials.
O.85E O.7E It is obvious that temperature can weaken
E/l Stress-Strain
Curve
a material because if the temperature is high
enough the material will melt or flow and thus
have no load carrying capacity as a structural
member. When a stressed member is subjected
to temperature, it undergoes a change of shape
F 0.86 in addition to that of the well known thermal
expansion. The term creep is used to describe
this general influence of temperature and time
f on a stressed material. Creep Is defined in
general as the progressive, relatively slow
change in shape under stress when subjected
to an elevated temperature. A simple illustra-
tion of creep is a person standing on a
E
bituminous road surface on a very hot summer
day. The longer he stands on the same spot
Fig. B1.12 the deeper the shoe soles settle into the road
The Ramsberg and Osgood proposed three surface, whereas on a cold Winter day the same
parameter representation of stress-strain time of standing on one spot would produce no
relations in the inelastic range Is:- noticeable penetration of the road surface.
High temperature, when used in reference
- - - - - (3) to creep, has different temperature values for
different materials for the same amount of
The equation for n is, creep. For example, mercury, which melts at
-38 0 F, may creep a certain amount at -750F,
n =1 + loge (17/7)/loge(F o • 7 /F o • u ) (4) whereas tungsten, which melts at 6l70 0r, may
not creep as much at 2000 0F as the mercury
Fig. Bl.13 is a plot of equation (4). The under -75OF. All materials creep under
quantities EE/F 0.7 are non-dimensional and may conditions of temperature, stress and time of
be used in determining the non-dlmenslonal stress application. The simplest manner in
curves of Fig. B1.l4. E, n, and F o • 7 must be whIch to obtain the effects of creep is to
known to use these curves in obtaining values study Its effect on the static stress-strain
diagram for the material.
Bl.9
Fig. Bl.13
50 .....
40 I'..
30
25
"- I"
20 ~
15
~
log (17/7)
"~t'\ n=l+ e
loge (FO. /F0.85)
", "-
10
9
8
7
n
6 "'-.
5
f'...
4
r-... ~
2
"""'"
'" r-...
....
1.5
1.0
1.01 1. 02 1. 03 1.04 1.051.06 1.08 1. 1 1. 15 1.2 1.3 1.4 1.5 1.6 1.8 2.0
FG• 7 /F O. 85
Fig. Bl.14
1.4
1.2
~ 3
1.0
.. ~ ~
0.8 ~~ ~
~
~
0.6
L4
~
~
0.4
V
0.2 III
o
V 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0
Bl.IO BEHAVIOR OF MATERIALS AND THEIR PROPERTIES
.
Table BI 1 Values of Ftu, F ev, E e, FO 7, FO 85, n , for Various Materials Unde r Room & Elevat e d Temperatures (From Ref 8)
Temp. F tu' Fey' Ee F O. 7 F O. 85
MATERIAL Temp. e,
Exp. OF n
Hr. % ksl ksl 106pSI ksl ksi
STAINLESS STEEL
AISI 301 174 Hard Sheet 1/2 RT 25 125 80 27.0 73 63 6.9
Transverse Compression
Longitudinal Compression 1/2 RT 25 125 43 26.0 28.2 23 5.2
AISI301 1/2 Hard Sheet 1/2 RT 15 150 118 27.0 116.5 105 9.2
Transverse Compression 1/2 400 118 108.5 23.2 108.5 97 8.6
1/2 600 110 107.5 20.9 108.5 96.5 8.2
1/2 1000 86 86 16.2 94.5 83.5 8.0
Longitudinal Compression 1/2 RT 15 150 58 26.0 48 37 4.4
1/2 400 118 53.3 22.4 45.5 36 4.7
1/2 600 110 52.8 20.1 44 31 3.5
1/2 1000 86 45.2 15.6 40 30.5 4.3
AISI 301 3/4 Hard Sheet 1/2 RT 12 175 160 27.0 163.5 151. 5 13.2
Transverse Compression 1/2 400 148 148 24.1 153 142.5 13.2
1/2 600 138 138 22.4 152 140 11. 2
1/2 1000 112 112 18.9 127 121 19.2
Longitudinal Compression 1/2 RT 12 175 76 26.0 70 61. 5 7.6
1/2 400 148 71 23.3 65 56 6.8
1/2 600 138 70.3 21. 6 65.5 56.5 6.8
1/2 1000 112 59.3 18.2 55 46 5.9
AISI 301 Full Hard Sheet 1/2 RT 8 185 179 27.0 183 172 16
Transverse Compression 1/2 400 168 168 25.1 174 164 16
1/2 600 159 159 23.8 172 162 16
1/2 1000 131 130 21. 6 141. 5 135.5 21. 5
Longitudinal Compression 1/2 RT 8 185 85 26.0 77.5 63 5.2
1/2 400 168 80.8 24.2 74 59.5 5
1/2 600 159 79.9 22.9 74 58 4.6
1/2 1000 131 66.3 20.8 58 42.5 3.9
17 -4 PH Bar & Forgings 1/2 RT 6 180 165 27.5 166 160 24
1/2 400 162 135 25.3 137 129 16
1/2 700 146 105.5 23.1 106 97 11
1/2 1000 88 62.6 21. 2 60 52 7.1
17-7 PH (THl050) Sheet, Strip & Plate, 1/2 RT 180 162 29.0 166 145 7.4
t = .010 to .125 In. 1/2 400 169 144 27.8 146 126 6.8
1/2 700 144 118 24.9 117 104 8.4
1/2 1000 88 61. 5 20.3 56 47 6
17-7 PH (RH950) Sheet, Strip & Plate, 1/2 RT 210 205 29.0 208 196 16.4
t = .010 to . 125 in.
19-9DL (AMS 5526) & 19-9DX 1/2 RT 30 95 45 29.0 36.5 32 7.6
(AMS 5538), Sheet, Strip & Plate
19-9DL (AMS 5527) & 19-9DX 1/2 RT 12 125 90 29.0 85 74 7.2
(AMS 5539) Sheet, Strip & Plate
PH15-7Mo (TH1050) Sheet & Strip, 1/2 RT 5 190 170 28.0 171 164 22.5
t = .020 to . 187 In.
PH15-7Mo (RH950) Sheet & Strip, 1/2 RT 4 225 200 28.0 218 189 7.3
t = .020 to • 187 in.
LOW CARBON & ALLOY STEELS
AISI 1023 & 1025 Tube, Sheet & Bar,
Cold Finished RT 22 55 36 29.0 32.7 31. 5 24
AISI 4130 Normalized, to>. 188 In. 1/2 RT 23 90 70 29.0 61. 5 53 6.8
1/2 500 81 61. 5 27.3 55 48 7.3
1/2 800 68 46.2 23.8 40 32.5 5.2
1/2 1000 46 30.8 20.6 28 22 4.7
AISI4130, 4140, 4340 Heat Treated 1/2 RT 23 125 113 29.0 111 102 10.9
1/2 500 113 98.3 27.3 96 88 10.9
1/2 850 88 68.9 23.2 66.5 61. 5 12
1/2 1000 64 49.7 20.6 45.5 41 9.2
AISI4130, 4140, 4340 Heat Treated 1/2 RT 18.5 150 145 29.0 145 140 25
1/2 500 135 126 27.3 126 122 29
1/2 850 105 88.5 23.2 88 83.5 18.5
1/2 1000 76 63.8 20.6 62 57 10.9
AISI4130, 4140, 4340 Heat Treated 1/2 RT 15 180 179 29.0 179 176 50
1/2 500 162 156 27.3 156 153 46
1/2 850 126 109.3 23.2 109.4 105 22
1/2 1000 92 77 20.6 75 68 9.8
AISI 4130, 4140, 4340 Heat Treated 1/2 RT 13.5 200 198 29.0 198 196 90
1/2 500 180 170 27.3 172.5 169 46
1/2 850 140 121 23.2 121. 5 117 25
1/2 1000 104 87.1 20.6 87 83 19
HEAT RESISTANT ALLOYS
A-286 (AMS 5725A) Sheet, Plate 1/2 RT 15 140 95 29.0 93 87 14
& Strip 1/2 600 129 88.4 24.4 87 81 13.5
1/2 1000 115 81. 7 19.8 81 75 12.5
1/2 1400 52 50.3 14.2 50 47 15.3
K-MONEL Sheet, Age Hardened 1/2 RT 15 125 90 26.0 88 82 13.5
MONEL Sheet, Cold Rolled & Annealed 1/2 RT 35 70 28 26.0 20 17 6.4
INCONEL-X 1/2 RT 20 155 105 31. 0 104 100 23.5
1/2 400 152 95.6 28.9 94 89 17
1/2 800 141 90.2 26.4 88.6 84 18.5
1/2 1200 104 83 23.2 82 78.6 21
Bl.ll
Table BI. I Values of Ftu, Fey, E e , FO 7, FO 85, n, for Various Materials Under Room & Elevated Temperatures (From Ref. 6) (Continued)
Temp. F tu' Fey' Ee F O• 7 F O. 85
MATERIAL Exp. Temp. e,
of n
Hr. % ksi ksi 106psi ksi ksi
ALUMINUM ALLOYS
2014-T6 Extrusions 2 RT 7 60 53 10.7 53 50.3 18.5
t,,; 0.499 in. 2 300 51 42.5 10.2 41. 5 40 24
2 450 28 21 9.2 20.5 19.5 25
2 600 10 8.0 7.4 5.5 4.5 5.4
1/2 300 51 43.5 10.2 44.0 42.5 25
1/2 450 31 26 9.2 26 25.2 29
2014-T6 Forgings 2 RT 7 62 52 10.7 52.3 50 20
t ~ 4 in. 2 300 53 41 10.2 40.5 38.5 19
2 450 29 22 9.2 21. 5 20 12.6
2 600 10 7.5 7.4 4.5 3.0 3.2
1/2 300 53 43 10.2 42.5 40 15.8
1/2 450 32 25.5 9.2 25.0 23.5 15.6
2024- T3 Sheet & Plate, 2 RT 12 65 40 10.7 39 36 11.5
Heat Treated, t,,;. 250 in. 2 300 37 10.3 35.7 33.5 15
2 500 26 8.4 24.8 22.8 10.9
2 700 7.5 6.4 6.2 5.5 8.2
2024-T4 Sheet & Plate, 2 RT 12 65 38 10.7 36.7 34.5 15.6
Heat Treated, t ,,; O. 50 in. 2 300 34 10.3 32.5 30.5 14.6
2 500 24 8.4 23 21 10.2
2 700 7 6.4 60 5.7 18.5
2024-T3 Clad Sheet & Plate, 2 RT 12 60 37 10.7 35.7 33 12
Heat Treated, t =. 020 to • 062 in. 2 300 34 10.3 33 30.3 11
2 500 24.5 8.4 22.7 20 7.9
2 700 6.5 6.4 5.8 5.5 18.5
2024-T6 Clad Sheet & Plate, 2 RT 8 62 49 10.7 49 45 11
Heat Treated, t '" O. 063 in. 2 300 45 10.3 44.3 40.7 11
2 500 22 8.4 31. 5 28 8.3
2 700 6 6.4 7.0 6.0 6.6
2024- T6 Clad Sheet & Plate, 2 RT 8 60 47 10.7 47 43 10.6
Heat Treated, t 0< O. 063 in. 2 300 43.2 10.3 42.3 38.7 10.8
2 500 21 8.4 29.5 26 7.8
2 700 6 6.4 5.00 4.0 4.9
2024-T81 Clad Sheet, Heat Treated, 2 RT 5 62 55 10.7 56 51. 6 11.2
to< 0.064 in. 2 300 50.5 10. :i 51. 2 46.5 10
2
2
6061-T6 Sheet, Heat Treated & Aged, 1/2 RT 10 42 35 10.1 35 34 31
to< 0.25 in. 1/2 300 29.5 9.5 29 28 26
1/2 450 20.5 8.5 19.3 17.7 10.9
1/2 600 7.5 7.0
.
6.6 6.2 15.2
7075-T6 Bare Sheet & Plate, 2 RT 7 76 67 10.5 70 63 9.2
t,,;; O. 50 in. 2 300 54 9.4 55.8 52.5 15.6
2 425 25.5 8.1 25.4 23.5 12.1
2 600 8 5.3 7.2 5.2 3.7
1/2 425 30 8.1 34.5 32.5 16
7075- T6 Extrusions, 2 RT 7 75 70 10.5 72 68 16.6
t :,,; O. 25 in. 2 300 54 9.4 58.5 54.5 13.4
2 450 22.5 7.8 21. 3 18.5 7.2
2 600 8 5.3 6.5 4.3 3.2
1/2 450 25 7.8 29 26 8.8
7075-T6 Die Forgings, 2 RT 7 71 58 10.5 58.5 55.1 15.2
t :,,; 2 in. 2 300 47.6 9.4 47.8 45 15.6
2 450 18.5 7.8 17.3 16 12
2 600 7.0 5.3 5.0 3.7 3.9
1/2 450 23 7.8 24 22 10.9
7075- T6 Hand Forgings, 2 RT 4 72 63 10.5 63.8 61. 5 25
Area:"; 16 sq. in. 2 300 51. 6 9.4 52.2 50 21. 5
2 450 20.2 7.8 20.3 19 13.7
2 600 7.6 5.3 6.0 5.0 5.8
1/2 450 24 7.8 26.5 25.3 19.5
7075- T6 Clad Sheet & Plate, 2 RT 8 70 64 10.5 64.5 61. 6 19.5
t:,,; 0.50 in. 2 300 50 9.4 54 51. 7 20
2 450 20.5 7.8 19.7 17.5 4.6
2 600 7.7 5.3 7.7 5.5 3.6
1/2 450 23 7.8 27.2 25.3 12.4
7079-T6 Hand Forgings, 1/2 RT 4 67 59 10.5 59.5 57.5 26
t :,,; 6. 0 in. 1/2 300 47 9.4 46.5 45 29
1/2 450 21 7.8 20 18.5 12
1/2 600 7.0 5.3 5.5 3.5 3.0
MAGNESIUM ALLOYS
AZ61A Extrusions, RT 8 38 14 6.3 12.9 12.3 19
t ,,; O. 249 in.
HK31A-0 Sheet 1/2 RT 12 30 12 6.5 10 8.4 6
t = 0.016 to 0.250 in. 1/2 300 20 11.1 6.16 8.9 6.9 4.5
1/2 500 15 9.3 4.94 7.5 5.6 4.2
1/2 600 10 4.9 3.77 3.3 1.6 2.2
HK31A- H24 Sheet, 1/2 RT 4 34 19 6.5 17.3 14.6 6.2
to< O. 250 in. 1/2 300 22 17.7 6.2 15.6 12.6 5.1
1/2 500 17 14.8 4.9 13.1 10.5 4.9
1/2 600 11 7.8 3.8 6.7 5.2 4.5
TITANIUM A LL YS
Ti-8Mn Annealed SMet, 1000 RT 10 120 110 15.5 119.5 102 13.7
Plate & Strip
Ti-6Al-4V Annealed Bar 1/2 RT 10 130 126 16.0 127 124.5 43
& Sheet, t::,,;. 187 in. 1/2 400 105 96 14.1 97 93 22
1/2 600 99 84.5 13.0 85.5 82 22
1/2 800 87 79.4 11.8 80.5 77 21. 5
1/2 1000 70 60.6 7.7 61 59.5 36
Bl.12 BEHAVIOR OF MATERIALS AND THEIR PROPERTIES
Fig. Bl.15 (curves A and B) show the Bl. 15. The General Creep Pattern.
stress-strain curve for a material. Curve (A)
is for a low elevated temperature condition A typical manner of plotting creep-rupture
and curve (B) that of a high elevated test data is illustrated in Fig. 81.16. For
temperature condition. The results were ob- metals tested at high value of stress or
tained by a normal testing machine procedure temperature, three stages in the creep-time
requiring a short time test period, hence the relation can be observed as shown in Fig. Bl.16.
results can be considered as independent of The initial stage, often called the stage of
time. primary creep, inclUdes the elastic deformation
and that region where the rate of creep de-
formation decreases rather rapidly with time,
which no doubt indicates an influence of strain
hardening. The second stage, often referred
to as the secondary creep stage, represents a
stage where the rate of strain has decreased
to a constant value (except for high stress)
for a considerable time period, and this stage
represents the period of minimum creep rate.
The third stage, often called the tertiary
creep stage, represents the period where the
reduction in cross-sectional area leads to a
higher stress, a greater creep rate and finally
rupture.
Transition Point. The inflection point between

the constant creep rate of the second stage and
INCREASING STRAIN _
the increasing rate of the third stage is
referred to as the transition point. Failure
Fig. Bl. 15 (Ref. 2) - Effect of temperature and time on the generally occurs in a relatively short time
strength characteristics of metals. after the transition point. Transition points
may not occur at very low stresses and may
The figure shows that the higher also not be 'definable at very high stresses.
temperature (curve B) reduces the ultimate Minimum creep rate is that indicated in the
strength, yield strength and modulus of second stage, where the creep rate is practi-
elasticity of the material as compared to cally constant.
curve (A) which is a test at a lower temper-
ature.
Initial Second Stage Third I
EFFECT OF TIME Stage (Approx. Constant) Stage ·1
I Rupture
If temperature and stress are of such Point
combination as to produce appreciable creep,
then time becomes an important effect. For
example, in Fig. Bl.15, if the material at low
temperature (curve A) is stressed to a value P
within the elastic limit of the material, and
this stress is maintained for a considerable , Transition Point
time period, very little creep, if any, will
be detected, and when the stress is removed, \... Slope = Creep Rate
the material will practically resume its Creep Intercept
original dimensions. However, if the material
is stressed to the same value P but under a
Time - Hrs.
high temperature condition, creep will occur
just as long as the stress P remains applied Fig. Bl.16 Typical creep-rupture curve.
and the stress-strain curve will take a shape
like curve (C) in Fig. Bl.15. This time Fig. 81.17 shows a plot of creep-time
dependent strain will follow a characteristic curves for a material at constant temperature
pattern. The material will never return to and several stress levels. It is noticed
its original shape after creep has taken place that increasing stress changes the creep-time
regardless if the stress is removed. If both relationship considerably.
stress and high temperature continue, rupture
produced by creep will finally occur. Bl. 16 Stress-Time Design Charts.
In many structural design problems in-

volving elevated temperatures such as power
B1.13
'00
e
,··J·+····I······· ~ 80
j
E 60
~
C 40
~
.r
'i: 20
u"
.r
Fig. Bl. 19 (Ref. 3) - Effect of exposure at elevated temper-

atures on the room-temperature tensile yield strength (Fty)
of 7079-T6 aluminum alloy (hand forgings).
TI"E~
100
Fig. Bl. 17 (Ref. 2) Creep curves for a material at constant

temperature and various stress levels showing the character-
istic stages of creep behavior.
I~
80
60
plants, the critical design factor is not C 40

strength but the permissible amount of creep
that can be allowed to still permit the
structure or machine to function or operate
satisfactorily. Extensive tests are usually 200 400 600 800 1000 1200 1400 1600
necessary to provide reliable creep design Temperature, F
information and such test information is often
recorded in the form as illustrated in Fig. Fig. Bl. 20 (Ref. 3) - Effect of temperature on the tensile
Bl.18. yield strength (Fty) of 5 Cr-Mo-V aircraft steels.
... stresses at a given temperature when heated

- from zero up to 200 0 F per second under constant
load conditions than when loaded after the
I-- t-- "' material has been 1/2 hour or longer at
I-- constant temperature. Increasing the temper-
"'" ature rate from 2000 F to 2000 0 F per second or
. , . .lo- .. more results in only a small increase in
strength. Figs. Bl.2l and Bl.22 (Ref. 4)
show the effect of temperature rates up to
Fig. Bl. 18 (Ref. 2) - Stress-time design chart at a single lOOoF per second upon the yield and ultimate
constant temperature for selecting limiting stress values.
70
B1. 17 Effect of Time of Exposure.

60
In general materials can be roughly
classified into those which time of exposure
to elevated temperatures has great influence
50
~
0
-- --...,
~
Yield stress-"" ..-YoeId_
on the mechanical properties and those where (stress-strain test) ~pid-heoting test)
such exposure produces relatively small effect. 40
This general fact is illustrated in Figs.

Bl.19 and Bl.20. The yield strength of the
5'.....
ksi
~~
aluminum alloy in Fig. Bl.19 is far more
influenced by time of exposure to elevated 20
~ ~'OO
0: '\
~
60-
temperature than the steel alloy as shown in
Fig. Bl.20. 0
\\
'02.'0 '2'
Bl. 18 Effect of Rapid Rate of Heating iemperOIUte rote,
"F/sec
Supersonic aircraft and missiles are

o 100 200 = Tern_OF
400 500 600 700 eoo
subjected to rapid aerodynamic heating. The Fig. Bl. 21 - Tensile yield stress of 2024-T3 aluminum
results of tests indicate that in general alloy for temperature rates from O. 20 F to lOOoF per
metals can withstand substantially higher second and of stress-strain tests for 1/2-hour exposure.
Bl.14 BEHAVIOR OF MATERIALS AND THEIR PROPERTIES
80
0
70 ~,
. . .,.z .-Ultimol~ $tress

(stress-slrainlest)
'-,
60
"-
\
~ i\.'
•.--f\lplurestress
(rapid..j-.eafing test)
,I
40
r~\~
I i
I i , •
·
.: 60
30
I
\\ •
~
i
i I "
I \
\ ..
0
\\ ~ NCO
",,~ l2i;;;~
-]00
~emp.roture rote,
""-
Fig. Bl.22 - Tensile rupture stress of 2024-T3 aluminum Fig. 1. 23 (Ref. 3) - Effect of Temperature on Ultimate
alloy for temperature rates from O. 20 F to lOOoF per second Strength (Ftu) of 2014-T 6 Aluminum Alloy.
and ultimate tensile stress of stress-strain tests for 1/2-
hour exposure.
vehicles. Fatigue failure is failure due to
strength respectively of aluminum alloy as being stressed a number of times. For example,
compared to values when loaded after the a beam may be designed to safely and efficient-
material has been exposed 1/2 hour at constant ly carry a design static load and it will carry
temperature. this static load indefinitely without failure.
However, if this load is repeated a large
Bl. 19 General Effect of Low Temperatures Upon Material enough number of times, it will fail under this
Properties. static design load. The higher the beam stress
under the static design load, the less the
The development of the missile and the number of repeated loadings to cause failure.
space vehicle brought another factor into the
ever increasing number of environmental con- To date no adequate theory has been
ditions that effect structural design, namely, developed to clearly explain the fatigue
extremely low temperatures. For example, in failure of materials. Fatigue failure appears
space the shady side of the flight vehicle is to begin with a crack starting at a point of
subjected to very low temperatures. Missiles weakness in the material and progressing along
carry fuels and oxidizers such as liquid crystal boundaries. A microscopic examination
hydrogen and oxygen which boil at -423 and of metals indicates there are many small
-297 oF respectively. In general, low cracks scattered throughout a material. Under
temperatures increase the strength and stiff- the action of repeated stress these small
ness of materials. This effect tends to cracks open and close during the stress cycle.
decrease the ductility of the material or, in The cracks cause higher stress to exist at
other words, produce brittleness, a property the base of the crack as compared to the
that is not desirable in structures because stress if there were no crack. Under this
of the possibility of a catastrophiC failure. repeated concentration of stress, the cracks
In general, the hexagonal closely packed will gradually extend across the section of
crystalline structures are best suited for the member and finally causing complete
giving the best service under low temperatures. failure of the member.
The most important of such materials are
aluminum, titanium, and nickel-base alloys. Fatigue testing consists of 3 types:-
Fig. Bl.23 shows the effect of both elevated (1) the testing of material crystals, (2) the
and low temperatures on the Ultimate tensile testing of small structural test specimens,
strength of 20l4-T6 aluminum alloy under and (3) the testing of complete composite
various exposure times. structures. A tremendous amount of test
information is available for the second type
Bl.20 Fatigue of Materials. of testing. More and more attention is being
given to the third type of testing. For
Designing structures to provide safety example, a complete airplane Wing or fuselage
against what is called fatigue failure is one is often subjected to elaborate fatigue
of the most important and difficult problems testing in order to insure the safe design
facing the structural designer of flight life of the airplane.
Bl.15
The presence of cracks initially is not magnitude of stresses in aircraft structures is

necessary to start a fatigue or progressive discussed briefly in Chapter A4. The limited
failure as irregularities such as slag discussion which follows will deal only With
inclusions, surface scratches, pitting, etc., the effect of impact loads on the behavior of
can cause corrosive action to start, thus materials.
supplying the condition for the promotion of
cracks and the resultant progressive failure. IMPACT TESTING METHODS
The strength of ferrous metals under There are in general two types of tests to
repeated stresses is often referred to as the determine the behavior of materials under
endurance or fatigue limit. The endurance impact loads. The usual impact test which has
limit stress is the stress that can be repeated been conducted for many years is referred to as
an infinite number of times without causing the notched bar test and consists of subjecting
fracture of the material. Non-ferrous notched specimens to aXial, bending and torsional
materialsjsuch as the aluminum alloys do not loads by the well known Charpy or Izod impact
have an endurance limit as defined above but testing machines. In both of these machines
continue to weaken as the stress cycles are an impact load is applied to the specimen by
increased. Due to this fact and also since swinging a weight W from a certain vertical
the required service life of structures and height (h) to strike and rupture the notched
machines vary greatly, it is customary to specimen and then stopping at a vertical
refer to the strength under repeated stresses height (h'). The energy expended in rupturing
as endurance or fatigue strength instead of the specimen is then equal approximately to
endurance limit. Thus the fatigue strength (Wh-Wh'). This type of test is primarily used
is the maximum stress that can be repeated for studying the influence of metallurgical
for a specified number of cycles Without variables.
producing failure of the structural unit.
The other type of impact testing is made
The results of testing a specimen under on unnotched specimens and the general purpose
repeated stresses such as tension, compression, is to obtain the stress-strain diagram of
bending, etc., is often plotted in a form materials under impact load or the load-
which is referred to as the S-N (stress versus distortion diagram of a structural member or
cycles) diagram, as illustrated in Fig. Bl.24. composite structure as the unit is completely
fractured under an impact load.
\ Bl. 22 Examples of Some Results of Impact Testing
~ ,, Q Q
n
of Materials.
Figs. Bl.25, 26 and 27 show the results

of impact tests upon the stress-strain curve
as compared to the static stress-strain
diagram (Ref. 5).
10 6 10 6 /00
i I
I
Number of Cycles for Failure, logN- I v- 64 fljsec i
Fig. Bl. 24 (' v v "'-'Stalic
: !
The problem of fatigue design of aircraft V- I !
airframes is covered in Chapter C13 of this I I
book. I
Bl. 21 Effect of Impact Loading on Material Properties.

o .04 08 ./2 ./6 .20 .24 .28 .32JO
.~ -~
An impact load when applied to a structure Untl elongation, in/in.

produces appreciable shock or Vibration. To
produce such action, the load must be applied
rapidly, that is, in a short interval of time. Fig. Bl. 25 - Stress-Strain curves, 24ST aluminum alloy.
The effect of impact loads differs from that
of static loads in that impact loads appreciably
effect the magnitude of the stresses produced
in a member and also the resistance properties
or behavior of the material under load. The
importance and effect of dynamic loads on the
Table Bl.2 shows additional impact testing

I results as compared to the static test results.
50
r Stotic v ~ 150 ff/sec
.t 40 1.1\
.!:?
0-
-Q 30
1/
-- l- V ~ 60 ft/sec Table B1. 2
Comparison of Strengths, Ductility and

S
C)
I Energy Absorption Under Impact
..... 20
I and Static Loads. xxx
~
1;," 10 Ratio of Impact to Static Value For,
Material Maximum Elong-
0 .02 .04 .06 .08 .10 ./2 .14 ./6 .18 Yield Energy
Unit elongation, in.Jin. Point Load ation
SAE No. 614()X 1. 020 2.205 2.865

Fig. B1. 26 - Stress-strain curves, Dow metal X. SAE No. 1015AA 1. 328 1. 285 1. 418 1. 376
SAE No. 1018 1. 995 1. 397 0.992 0.946
18-8 Allov 1.224 1. 212 0.682 0.600
17ST Alum. Al. 1. 294 1. 094 1. UUU U.885
Brassxx 1. 387 1.142 1. 163 1. 203
Aluminum 1. 522 1. 323 1.628 1.627
Copper 1.433 1. 390 1. 527 1. 597
240 x Oil quenched from 1620o F.

xx Cold-rolled.
h
xxx From "Stress-Strain Relations Under Tension and
200 Impact Loading" by D. S. Clark & G. Datwyler,
\
.S; 'f\-..- x
I
--- r--.
--- !/ 0 60 ItI ec
::::- -----...
Proc. A.S. T. M. 1938, Vol. 38.
0- 160
~ Stalic~
'\ I'---- REFERENCES
~
a
§120
I v = 15 II/sec 1""- r- Ref. 1. Ramberg & Osgood. Description of
«i ,
V) I Stress-Strain Curves of 3 Parameters.
~ 80
V)
NACA, Tech. Note 902.
Ref. 2. Time and Temperature Gremlins of
40 Destruction. By L. A. Yerkorvich,
Cornell Aero. Lab. Research Trends.
02 .04 .06 .08 10 ./2 .14
Sept. 1956.
0
Unit elongation, /n.Jin. Ref. 3. Military Handbook (MIL-HDBK-5 )
Aug. 1962.
Ref. 4. NACA Technical Note 3462.
Fig. B1.27 - Stress-strain curves, SAE 6140, drawn 1020o F. Ref. 5. NACA Technical Note 868.
Ref. 6. From Structures Manual, Convair
Astronautics .
.
CHAPTER B2
MECHANICAL AND PHYSICAL PROPERTIES OF METALLIC
MATERIALS FOR FLIGHT VEHICLE STRUCTURES
It would require several hundred pages to list the properties of the many
General Explanation.
materials used in flight vehicle structural design. The metallic materials presented in this
chapter are those most widely used and should be sUfficient for the use of the student in his
structural analysis and design problems. All Tables and Charts in this chapter are taken from
the government publication "Military Handbook, MIL-HDBK-5, August, 1962. Metallic Materials and
Elements for Flight Vehicle Structures". This publication is for sale by the Supt. of Public
Doc~~ents, Washington 25, D.C. The properties given in the various tables are for a static
loading condition under room temperature. The effect of temperature upon the mechanical
properties is given in the various graphs.
AISI ALLOY STEELS
Table B2. 1 (AISI) Alloy Steels
Alloy ..................... AISI 4130, 8630, AISI 4130, 4140, 4340, I 4140 AISI
and 8735 8630, 8735, and 8740 4340 4340
8740
I
Form ..................... Sheet, strip, All wrought forms
plate, tubing
Condition ................. N Heat treated (quenched and tempered) to

obtain F,u indicated
Thickness or diameter, in... :<0.187 >0.187
Ba~s ..................... ....... . . . . . . . . . ....... . ....... . . ....... . ....... (0)
Mechanical properties:
F,u, ksi ............... 95 90 125 150 180 200 260
F,y, ksi ............... 75 70 103 132 163 176 217
Fey, ksi ............... 75 70 113 145 179 198 242
F. u, ksi. .............. 55 55 82 95 109 119 149
Fb,u, ksi:
(eID=1.5) . ....... ....... . . ....... 194 219 250 272 347
(eID=2.0) ........ 140 140 251 287 326 355 440
F bry , ksi
(eID=1.5) . ....... ....... . . ....... 151 189 230 255 312
(eID=2.0) . ....... .. . .... . . ....... 180 218 256 280 346
e, percent ............. bL 10
See table See table 2.3.1.1(c)
2.3.1.1 (b) bT 3
I I
E, 10 6 psi. ............ 29.0
E c' 10 6 psi ............ 29.0
G, 10 6 psi. ............ 11. 0
Physical properties:
w, Ib/in. 3 . . . . . . . . . 0.283.
C, Btu/(1b) (F) ........ 0.114 (at 32° F).
K, Btu/[(hr) (ft2) (F) 1ft] . 22.0 (at 32° F).
a, 10- 6 in./in./F ....... 6.3 (0° to 200° F).
82.1
B2. 2 MECHANICAL AND PHYSICAL PROPERTIES OF METALLIC MATERIALS FOR FLIGHT VEHICLE STRUCTURES
AlSl ALLOY STEELS (Cont.)
100
e:>
g
e 80
Q.
...
Q)
Strength at temperature
Exposure up to 1/2 hr.
! 60
E
Q)
j
I- 120 ;; 40
E .P
o
o c 20
a: ~
~
C 100 200 400 600 800
u
W Temperature, F
"0 Fig. 82. 2. Effect of temperature on the compressive
c:
0 E yield strength (Fey) of heat-treated AISI alloy steels.
w
80
::>
u:-
,.,
e
i!
-
LL
c
Q)
60
Fft Ftu i
§
0
...
Q)
;;
Il. u.1-
~
40 .r
Temperature, F
Fig. 82. 3. Effect of temperature on the ultimate

shear strength (Fsu) of heat-treated AISI alloy steels.
20
o
-400 o 400 800 1200
Temperature, F
Fig. 82.1. Effect of temperature on Ftu> Fty, and E of
AISI alloy steel
Fig. 82. 4. Effect of temperature on the ultimate

bearing strength (Fbru) of heat-treated AISI alloy steels.
i§
80
60
;; 40
.i
~
20
Temperature, F
Fig. 82. 5. Effect of temperature on the bearing

yield strength (Fbry) of heat-treated AISI alloy steels.
82.S
5 Cr-Mo-V AIRCRAF'l' STEEL i t

100
E
00
~
1
:
TUle Ba.2 ~Ip .........tca1lU1d PbJa1cal Propert1ea of ICr-IIo-Y Aircraft Steel ..
i
"
..i i
Alloy..................... 5Cr-Mo-V aircraft steel. ..I
Form .................... All wrought forms.
Condition ................. Heat treated to obtain the Flu indicated.
Up to 12 in. diam. or equivalent. 1'Ig. B2. 8 Effect of temperature OIl
Section size ..... , .........
)
the tensile and compressive modulus
(E and Ed of 5 Cr-Mo-V aircraft
Basis ..................... (0) (0) (0) IIteela.
FlU, ksi. .............. 240 260 280
FlU, ksi. .............. 200 220 240
Feu, ksi. .............. 220 240 260
F. u, ksi ............... 145 155 170
F bru , ksi:
(ejD=1.5) . ....... ............. . ............. . . ....... ..... _
(e/D=2.0) . ....... 400 435 465

F bru , ksi:
(e/D=1.5) ........ .. ............. ............. . . ............ .... 82. 9 meet 01 temperature OD
(e/D=2.0) . ....... 315 340 365 1M compressive yield strength (Fey)
e, percent: .. I Cr-1Io- V aircnft Meels.
Bar, in 4D ........ 9 8 7
Sheet, in 2 in. (b) . 6 5 4
Sheet, in 1 in ...... 8 7 6
iI
11
E, 106 psi. ... " ........ 30.0
E e, 10 6 psi. ........... 30.0 .l
G, 10 6 psi. , ........... 11.0
1&>, Ib/in. 3 • • • • • . • • • . • . • • 0.281. J'Ig. 81. 10 Efteet 01 te. . .rature OIl
C, Btu/(lb)(F) ...... , .. O.l1(e) (32 0 F). the ultimate sbea'r IItrea&th (Pau) of
K, Btu/[(hr)(ft2)(F)/ft]. 16.6 (400° to 1,100° F). 5 Cr-IIo-Y aircraft...... .
a, 10-6 in./in./F ........ 7.1 (80° to 800° F); 7.4 (80°-1,200° F).
• Minimum properties expected when heat treated as recommended In

section 2.6.l.0.
b For sheet thickness greater than 0.01iO Inch.
• Calculated value.
PlI. Ba. 11 Effect of temperature OD

the ultimate beariDg strength (Fbru)
01 5 Cr-"'¥' aircraft ateels.
100
i 40
i
~ 20
I °o
Fig. B2. 6 Effect of temperature OIl Fig. B2. 7 Effect of temperature OIl PIg. B2. 12 Effect of temperature OD
the ultimate tensile strength (Foo) of the tensile yield strength (Fty) of 5 the bearing yield strength (Fbry) of 5
Cr-Mo-V aircraft steels. Cr-Mo-V aircraft steels. Cr-Mo-V aircraft steels.
B2.4 MECHANICAL AND PHYSICAL PROPERTIES OF METALLIC MATERIALS FOR FLIGHT VEHICLE STRUCTURES
17-7 PH STAINLESS STEEL

Table B2. 3 Design Mechanical and Physical Properties of 17-7 PH stainless Steel
Alloy ..................... 17-7 PH
Form ..................... Sheet, strip, and plate(a) Bars and Forgings(b)
Condition ................. TH 1050 RH 950 TH 1050 \ RH 950
Thickness or diameter, in .. 0.005 to 0.005 to 0.1878 to 6 and under

0.500 0.1874 0.500
Basis ..................... S S(e)
F 'u, ksi. .............. 180 210 200 170 185
F tv ' ksi. .............. 150 190 180 140 150
F ev , ksi ............... 158 200 189 147 158
F. u , ksi ............... 117 136 130 111 120
F bru , ksi:
(eID=1.5) . ....... 2'97 346 330 280 305
(eID=2.0) ........ 360 420 400 340 370
F brv , ksi:
(eID=1.5) . ....... 225 285 270 210 225
(eID=2.0) . ..... " 247 313 297 231 247
e, percent ............. See table 2.7.2.1(b) 6 6
E, 10 6 psi. ............ 29.0

E e , 10 6 psi ............ 30.0
G, 10 6 psi. ............ 11.0
w, Ib/in. 3 • • • • • . . • • . . • • • 0.276.
C, Btu/(lb)(F) ......... 0.11. (d)
K, Btu/[(hr)(ft2)(F)/ft] .. 9.75 (at 300 0 F).
a, 10- 6 in./in./F . 6.3 (70 0 to 600 0 F) for TH 1050.
6.8 (70 0 to 600 0 F) for RH 950.
• Test directiolIongitudinal for widths less than 9 in.; transverse for short transverse (thickness) direction.
widths 9 in. and over. c Vendors guaranteed minimums for Ftu., Fly, and e.
• Test direction longitudinal; these properties not applicable to the d Calculated value.
zoo 400 600 800 1000 IZOO 1400 JeOO

T~ture,F
Fig. B2. 13. Effect of temperature on the ultimate Fig. B2. 14. Effect of temperature on the tensile yield
tensile strength (Fro) of 17-7 PH (THI050) stainless steel. strength (Fty) of 17-7 PH (THI050) stainless steel.
B2.5
17-7 PH STAINLESS STEEL (Cant.)
IJ
-
..."
...
200 400 600 800 1000 1200 1400 1600 200 400 600 800 1000 1200 1400 1600
TemperolIn, F Tempe..." ... F
Fig. B2.15. Effect of temperature on the compressive Fig. B2. 18. Effect of temperature on the ultimate
yield strength (Fey) of 17-7 PH (TH1050) stainless steel. bearing strength (Fbru) of 17-7 PH (TH1050) stainless steel.
100
I
j
1;
.:
Fig. B2.16. Effect of temperature on the ultimate Fig. B2.19. Effect of temperature on the bearing yield
shear strength (Fsu) of 17-7 PH (TH1050) stainless steel. strength (Fbry) of 17-7 PH {TH1050) stainless steel.
200 400 600 800 1000 1200 1400 1600

Temperal.... F
Fig. B2. 20. Effect of temperature on the tensile and

compressive modulus (E and Ed of 17-7 PH (TH1050)
stainless steel.
B2.6 MECHANICAL AND PHYSICAL PROPERTIES OF METALLIC MATERIALS FOR FLIGIIT VEHICLE STRUCTURES
17-4 PH STAINLESS STEEL

Table B2.4 Design Mechanical and Physical Properties of 17-4 PH Stainless Steel
Alloy ..... , ....... , ....... 17-4 PH
Form ..................... Plate Bars and forgings (b)

Fig. B2. 23 Effect of temperature on
Condition ......... , " ..... H 900 H 900 the compressive yield strength (Fey)
of 17-4 PH (H900) stainless steel.
Thickness or diameter, in.... 8 and under '00
Basis ..................... Sea) S
Flu. ksi ............... 190 190 ~ 20
F l.. ksi ............... 170 170 !
Feu. ksi. .............. 178 178
Temperature, F
FlU. ksi. .............. 123 123
F bru • ksi: Fig. B2. 24 Effect of temperature on
(e/D=1.5) ..... ... 313 313 the ultimate shear strength (F su ) of
(e/D=2.0) ...... .. 17-4 PH (H900) stainless steel.
380 380
F bru • ksi:
(e/D=1.5) ...... .. 255 255
(e/D=2.0) ..... ... 280 280
e, percent:
In 2 in............ 10 ...
In 4 D ............ ... 10 1; 40
...1.
~ 20
E, 10 6 pSl . . . . . . . . . . . . • 29.0 ! 0
o 200 400 600 BOO 1000 1200 1400 1600
E e, 10 6 psi. . 30.0 Temperature, F
G, 10 6 psi. . 11.0 Fig. B2. 25 Effect of temperature on

the ultimate bearing strength (Fbru)
of 17-4 PH (11900) stainl........
w, lb/in. 3 .•. . . . . . . . . . .. 0.282.
C. Btu/(lbHF)...... ... 0.11 (32 0 to 212 0 F).
K, Btu/[(hr) (ft2) (F)/ft]. 10.3 (at 300 0 F); 11.2 (at 500 0 F); 13.1 (at 900 0 F).
a, 10- 6 in./in./F. . . . . . .. 6.0 (70 0 to 200 0 F); 6.1 (70 0 to 400 0 F); 6.5 (70 0 to
900 0 F).
Q Vendors guaranteed minimums (or Ftu, Ft'!l' and e.

, Test direction longitudinal; tllese properties not applicable to the short transverse (thickness) direction. Temperoture, F

the bearing yield strength (Fbry) of
17-4 PH (H900) stainless steel.
100 '00
i 80
! 60
J
1; 40
...2
~ ~
20
20
~ ~ 0
o 200 400 600 800
Temperature. F
Fig. B2.21 Effect of temperature on Fig. B2. 22 Effect of temperature on Fig. B2. 27 Effect of temperature on
the ultimate tensile strength (Ftu) of the tensile yield strength (Fty) of the tensile and compressive modulus
17-4 PH (H900) stainless steel. 17-4 PH (H900) stainless steel. (E and Ed of 17-4 PH (H900)
stainless steel.
B2.7
AM-350 STAINLESS STEEL
Table B2. 5 Design Mechanical and Physical Properties of AM-350 Stainless Steel
Alloy ................ " .... AM-350
Form ...................... Sheet and strip· Sheet and strip·

o 40
Condition ................. DA SCT .P

~ 20
Thickness, in............... 0.187 and under 0.187 and under
~ 0
o 200 400 600 800 1000 1200 1400 1600
Basis ..................... S S Temperature, F
Fig. B2.30 Effect of temperature on
the compressive yield strength (Fcy) of
Mechanical properties: AM-350 stainless steel (double-aged).
F tu , ksi .............. 165 185
F ty , ksi .... .......... . 135 150
Fey, ksi ................ 142 158
F. u, ksi. .............. 107 120 100
Fb,u, ksi:
(ejD=1.5) ......... 272 305
i 80
(ejD=2.0) .... ..... 330 370 !

Fb,y, ksi: j
0
(ejD=1.5) ..... .... 202 225
.:J
(ejD=2.0) ..... .... 223 247 ]
e, percent ............. 10 10
5'"
i Temperature, F
E, 10 6 psi ............. 29.0

En 10 6 psi. ......... .. 30.0 the tensile and compressive modulus
G, 10 6 psi .......... ... 11.0 (E and Eel of AM-350 stainless steel
(double-aged).
w, Ibjin. 3 . . . . . . . . . . . . . . . 0.282.
C, Btuj(lb)(F) ......... 0.12 (32 0 to 212 0 F).
K, Btuj[ (hr)(ft2)(F) jft] . .. 8.4 (at 100 0 F); 11.7 (at 800 0 F).
a, 10-6 in.jin.jF ......... 6.3 (70 0 to 212 0 F); 7.2 (70 0 to 932 0 F).
• Test direction longitudinal for widths less than 9 in.; transverse for widths 9 in. and over.
200 400 600 800 1000 1200 1400 1600

Temperature, F
}<·ig. B2.32 Effect of temperature on

the ultimate tensile strength (Ftu) of
AM-350 stainless steel (SCT).
100
~
e
!
j
80
I
j
0 40 0
"" ~-
~ 5
~ .r
Temperature, F Temperature, F Temperature, F
Fig. B2. 28 Effect of temperature on Fig. B2. 29 Effect of temperature on Fig. B2. 33 Effect of temperature on
the ultimate tensile strength (Ftu) of the tensile yield strength (Fty) of the tensile yield strength (Fty) of
AM-350 stainless steel (double-aged). AM-350 stainless steel (double-aged). AM-350 stainless steel (SCT).
AM-350 STAINLESS STEEL (Cant.)
100 100
~
!
.=e
e 80
~
80
"~
0!!- 0!!-
60
J
0
60
J
0
40 40
.;t
.P
'20
20
'" ~
~ cf
.f 200 400 600 800
Temperature. F
200 400 600 800
Temperature, F
1000 1400
Fig. B2. 34 Effect of temperature on the tensile and Fig. B2. 36 Effect of temperature on the compressive
compressive modulus (E and Ed of AM-350 stainless yield strength (Fey) of AM-350 stainless steel (SCT).
steel (SCT).
100 !OO
!~
!
e
I~ 80
"~
0!!-
80
g
a:
60
J
60
0 40 0 40
... 1;
.f
~ ~
20 20
~ cf
200 400 600 800 1000 1200 1400 1600
Temperature. F Temperature, F
Fig. B2.35 Effect of temperature on the bearing Fig. B2. 37 Effect of temperature on the ultimate
yield strength (Fbry) of AM-350 stainless steel bearing strength (Fbru) of AM-350 stainless steel
(SCT). (SeT).
100
i §
80
60
0 40
,}
~
u
20
cf
Temperature, F
Fig. B2.38 Effect of temperature on the ultimate

shear strength (F su ) of AM-350 stainless steel
(SCT).
B2.9
AISI 301 STAINLESS STEEL
Table B2. 6 Design Mechanical and Physical Properties of AISI 301 Stainless Steel
Alloy ............ , .... " . AISI 301 a
Form ................... Plate b, sheet, and strip
Condition ................ Annealed }~ hard }~ hard 7~ hard Full hard
Basis .................... S S S S S
Flu, ksi:
L ............... 75 125 150 175 185
T ............... 75 125 150 175 185
Flu, ksi:
L ................ 30 75 110 135 140
T ................ 30 75 110 135 140
F cu, ksi:
L ................ 35 43 58 76 85
T ................ 35 80 118 160 179
F,u, ksi. .............. 40 67.5 80 95 100
F bru , ksi:
(eID= 1.5) ........ ......... . ......... . . ......... ......... . . .........
(eID=2.0) . ....... 150 250 300 350 370
F bru , ksi:
(eID=1.5) . ...... . . . . . . . . . . ......... . ......... . . ......... . .........
(eID=2.0) . ...... 50 140 200 240 270
e, percent ............. (c) (c) (c) (c) (c)
E, 10 6 psi:
L ................ 29.0 27.0 26.0 26.0 26.0
T ................ 29.0 28.0 28.0 28.0 28.0
E c, 10 6 psi:
L ................ 28.0 26.0 26.0 26.0 26.0
T ................ 28.0 27.0 27.0 27.0 27.0
G, 10 6 psi. ............ 12.5 12.0 11.5 11.0 11. 0
w, Ib/in. 3 • . • . . . . . . . . . . . 0.286.
C, Btu/(lb)(F) . ........ 0.108 (at 32 0 F).
K, Btu/[(hr)(ft2)(F) 1ft] . 7.74 (at 32 0 F).
a, 10-6 in./in./F . ....... 9.2 (70 0 to 200 0 F).
" Properties for annealed condition also applicable to annealed AISI

302,303,304,321, and 347.
b Only annealed condition applicable to plate.
, See table 2.8.1.l(b).
N OTE.- Yield strength, particularly in compression, and modulus of

elasticity in the longitudinal direction may be raised appreciably by
thermal stress-relieving treatment in the range 500" to 800° F.
e AIBI 301 STAINLESS STEEL (Cant.) e

'00
~
I
j
80
80
J
j
80
80
1; 40 1; 40
~u
..:-
~ ~
20 20
'00
.r .l:
Temperatur8,F

!j
80
60
200 400 600 eoo
Temperature, F

1000 1200
the ultimate tensile strength (Ftu) of 1; 40 the compressive yield strength (F cy )

AISI 301 (half-hard) stainless steel. ~i; of AISI 301 (full-hard) stainless steel.
] 20
H'O
.l: 01 temperature
up 10 thr 1
n +:j ~l
Temperature. F

the bearing yield strength (Fbry) of
AISI 301 (half-hard) stainless steel. .; J'
!f
] 20
.r ! tL 1 '600
'00
Temperolur.,F
Temperature, F
Fig. B2.40 Effect of temperature on I Fig. B2.49 Effect of temperature on

J
00
the tensile yield strength (Fty) of the ultimate shear strength (Fsu) of
AISI 301 (half-hard) stainless steel. AISI 301 (full-hard) stainless steel.
j 80
1; 40
u
~ '"~
2 20
80
! 60
~
j .r
1; 40
~. Fig. B2.45 Effect of temperature on
".t.5 20
(E and Eel of AISI 301 (half-hard)
stainless steel.
Temperature, F Temperature,F
Fig. B2.41 Effect of temperature on Fig. B2. 50 Effect of temperature on

the compressive yield strength (Fcy) the ultimate bearing strength (Fbru)
of AISI 301 (half-hard) stainless steel. of AISI 301 (full-hard) stainless steel.
'00
:1 ~ st;~~gth~~,!;;;;perOlure
iJ
" ;E~posure up 10 t hr
80 •• 1" '
60
1; 40
~~
~
.l:
lJJlJ!\llilliM!\ill eoo 1000 1200 1400 1600
Fig. B2.46 Effect of temperature on Temperolure,F
Temperature, F
Fig. B2.42 Effect of temperature on AISI 301 (full-hard) stainless steel. Fig. B2. 51 Effect of temperature on
the ultimate shear strength (Fsu) of the bearing yield strength (Fbry) of
AISI 301 (half-hard) stainless steel. AISI 301 (full-hard) stainless steel.
'00 '00
j
2 ~ so
i 80
'j"
~
80
I J 80
80
j
80
j
1; 1; 40
40 1; 40
.:t
J; ..:-
~ 20
~ ]
20 20
'"
.t .l: ~
200 400 600 BOO 1000 .l: 200 400 600 eoo 1000 1200 1400 1600
Temperature, F
Temperature, F Temperoture,F
Fig. B2.43 Effect of temperature on Fig. B2.47 Effect of temperature on Fig. B2. 52 Effect of temperature on
the ultimate bearing. strength (Fbru) the tensile yield strength (Fty) of AISI the tensile and compressive modulus
of AISI 301 (half-hard) stainless steel. 301 (full-hard) stainless steel. (E and Ec) of AISI 301 (full-hard)
stainless steel.
B2.. 11
2014 ALUMINUM ALLOYS (SHEET & PLATE, EXTRUSIONS, FORGINGS)

NOTE: Values in (A) columns are minimum guaranteed values. Values in (B)
column will be met or exceeded by 90 percent of material supplied.
Table B2. 7 Design Mechanical and Physical Properties of 2014 Aluminum Alloy (Sheet and Plate)
Alloy. '" . . .. ... 2014
Form.... .. . '" . . ... Sheet and plate
Condition .. .... . " . .... . -T6 a
1.501- 2.001- 3.001-b

Thickness, in ............... 0.020-0.039 0.040-0.499 0.500-1.000 1.001-1.500 2.000 3.000 4.000
Basis. ..... . .. . ...... . ..... A B A B A B A B A A A

- - --- --- --- --- --- --- --- ----
F I", Icsi
L. . . . . . ..... . ... 65 67 68 70 68 70 67 6S 65 63 59
T .. .... . .... . . 64 66 67 69 67 69 67 68 65 63 59
ST. ... . . . .. . ·. ·. ·. · . .. · . ·. 58 54
F I 1/> lcsi
L .. . . . . . . . ...... . . 58 60 60 62 60 62 59 62 59 57 55
T .. ... . . . . . . . . . . 57 59 59 61 59 61 59 62 59 57 55
ST .. . ........ . . '" · . ·. ·. ·. ·. .. ·. 53 51
F e1/' Icsi
D .. ....... . .. . ·. ·. 60 62 60 62 61 64 61 59 57
T .. ........ . .... . ·. 61 63 61 63 61 64 61 59 57
ST .. ... . . . . . . . .. . . .. ·. · . ·. · . 59 57
F.", I<:s·i .. .... . . .. · . · . 41 42 41 42 41 41 40 39 37
Fbm, ks,t
(r/D = 1.5) . ·. 102 105 102 105 101 102 98 89 88
(eID=2.0) .. ·. 129 133 129 133 127 129 124 129 112
Fbrm ksi e
(rID=1.5) .. 84 87 84 87 83 87 83 80 77
(eID=2.0) ... · . 96 99 96 99 94 99 94 91 88
e, percent
L. ... . 8 6 6 ·. 6 4 3
7'. 6 ·. 8 ·. 6 .. 4 ·. 3 2 1.5
ST .. . . . . . . . . . ·. ·. ·. ·. .. ·. ·. 1 ..
E, 10' psi .. . . . . . . . . . 10.5
E e, 10' psi. . . . . . . . ..... 10.7
G, 10' psi. ............ ·L(l
w,lb/in. 3 . 0.101
C, Btu/(lb)(F). ....... 0.23 (at 212°F)
K, Btu/[(hr)(ft')(F)/ft] no (at 77°F)
a, 10-' in./in./F .... 12.5 (68° to 212°F)
B2.12 MECHANICAL AND PHYSICAL PROPERTIES OF METALLIC MATERIALS FOR FLIGHT VEffiCLE STRUCTURES
2014 ALUMINUM ALLOYS (SHEET & Ph4TE, EXTRUSIONS, FORGINGS) (Cant.)
Table B2. 8 Design Mechanical and Physical Properties of Clad 2014 Aluminum Alloy (Clad Sheet and Plate)
Alloy ......................... Clad 2014
Form ......................... Sheet and plate
Condition ..................... -T6 a
Thickness, in ................. <0.039 0.04~0.499 0.5~1.000 1.001-1.500
Basis.......................... A B A B A B A B
----
FlU, ksi
L .. ................... 64 64 65 67 65 67 64 65
T ..................... 63 63 64 66 64 66 64 65
ST .... ...............
FIll, ksi
L .. ................... 56 56 58 60 58 60 57 60
T ..................... 55 55 57 59 57 59 57 60
ST .... ...............
F cy, ksi
L .. ................... 56 56 58 60 58 60 59 62
T ..................... 57 57 59 61 59 61 59 62
ST .... ...............
F ou, ksi . .................. 39 39 39 40 39 40 39 40
Fbru, ksi b
(e/D=1.5) ... .......... 96 96 98 101 98 101 96 98
(e/D=2.0) ............. 122 122 124 127 124 127 122 124
FbTII' ksi b
(e/D=1.5) ............. 78 78 81 84 81 84 80 84
(e/D=2.0) ............. 90 90 93 96 93 96 91 96
e, percent
L .... ................. 7 8 6 6
T ..................... 7 8 6 4
E, 10' psi. ................. 10.5
Eel 10' psi. ................ 10.7
G, W. psi. ................. 4.0
B2.13
2014 ALTJMINm1 ALLOYS (SH~ET & PLATE, EXTRUSIONS, FORGINGS) (Cant.)
Table B2.9 Design Mechanical and Physical Properties of 2014 Aluminum Alloy (Extrusions)
Alloy .... ... . . . . . . . . . . . . . . . . 2014
Form. . . ... . ... Extruded rod, bar, and shapes
Heat treated
and aged
by user"
Condition. . . . . . . . . . ....... . -T6e -T62
Cross-sectional area, in.'. . . . . . . <25 >25, <32 Up to 32
Thickness, in. b ... ....... .... . 0.125-{).499 0.500-0.749 0.75(}-1.499 1.500-2.999 3.000-4.499 :;;:0.750 0.125-4.499
Basis........................ A B A B A B A B A B A A
------ ------------
Flu, ksi
L . ....... ........... 60 61 64 68 68 73 68 73 68 73 68 60
T .. .......... ....... 60 61 64 67 63 66 61 63 58 61 56 56
FIll' ksi
L ................... 53 57 58 62 60 65 60 65 60 65 58 53
7'........ ........... 53 57 55 59 54 58 52 55 49 53 47 47
Fell' ksi
L . ........ ......... . 55 59 60 64 62 67 62 67 62 67 ·. 53
Too ................. 53 56 58' 62 57 61 57 61 57 61 ·. 48
F,u, ksi .... .............. 35 35 37 39 39 42 39 42 39 42 39 35
Fbru, ksi
(eID=1.5) .. ......... 90 92 96 102 88 95 88 95 88 95 ·. ·.
(eID=2.0) .. " ....... 114 116 122 129 109 117 109 117 109 117 ·. ·.
Fbrll' ksi
(eiD = 1.5) ........... 74 80 81 87 78 85 78 85 78 85 ·. ·.
(eID=2.0) .. ......... 85 91 93 99 84 91 84 91 84 91 ·. ·.
e, percent
L ................. 7 .. 7 .. 7 .. 7 .. 7 .. 6 7
Td ................ 5 .. 5 .. 2 .. 2 .. 1 .. 1 ·.
E, 10' psi ... ............. 10.5
E e, 10' psi ............... 10.7
G, 10' psi . ............... 4.0
B2.14 MECHANICAL AND PHYSICAL PROPERTIES OF METALLIC MATERIALS FOR FLIGHT VEillCLE STRUCTURES
2014 AL~INUM ALLOYS (SHEET & PLATE, EXTffiJSIONS, FORGINGS) (Cont.)
Table B2. 10 Design Mechanical and Physical Properties of 2014 Aluminum Alloy (Forgings)
Alloy .. .... ......... . 2014
Die Hand forgings Hand forg:ngs

h'orm. . ......... forgings Length <3 times width Length >3 times width
-T6
Condition .. ...... . ......... -T4
I
Thickness, in ........... ... . <4 inches <6 inches
>16, >36, >144, >16, >36, >144,

Cross-sectional area, in.'........ <16 <36 <144 <256 <16 <36 <144 <256
- - - - - - - - - - - - - - - - - - ----
J3asis.. .................. .... . A A A A A A A A A A
--- - - - - - - - - - - - - - - - - - - - - - - - ---- ----
Flu, ksi
L .. .................. 55 65 65 65 62 60 65 65 62 60
1' .. .................. 52 62 63 6:3 59 58 63 63 59 58
S7'. .................. ·. ·. 60 60 56 55 60 60 56 55
FlU, ksi
L ................... 30 55 55 53 53 52 55 53 53 52
7'.................... 28 52 55 53 52 50 55 53 52 50
S7' .. .......... ...... . ·. ·. 55 53 52 50 55 53 52 50
Feu, ksi
L ... ......... ....... . 30 55 55 53 53 .52 55 53 53 52
T~ ................... 28 52 55 53 .52 50 5S 53 ·. ·.
ST .... ............... ·. ·. .. .. .. .. .. .. ·. ·.
F. u, ksi ......... . . . . . . . . . . 34 39 40 40 38 37 40 40 ·. ·.
Fbru, ksi
(e/D=1.5). ........... ·. ·. 91 91 87 84 91 91 87 84
(e/D=2.0). ........... ·. ·. 117 117 112 108 117 117 112 108
Fbru, ksi
(e/D = 1.5). ........... ·. ·. 77 74 74 73 77 74 74 73
(e/D=2.0) .. .......... ·. ·. 88 85 85 83 88 85 85 83
e, percent
L . ................... 11 7 10 9 7 5 10 9 7 5
1'.................... ·. 3 6 5 3 2 4 3 2.5 2
ST .. ................ ·. · . 3 2 1 1 2 2 1 1
E, 10' psi................. .0.5
E e, 10' psi ................ 10.7
G, 10' psi. ...... ......... . 4.0
B2.15
140
1+ +"-i
20
.,-' ...
.!L= ......
o o 100 200 300 00 500 600 700 lJoo
TEMPERATlJlE, of
Fig. B2. 53 Effect of temperature on the ultimate strength (Ftu) of 2014-T6 aluminum alloy (bare and clad sheet 0.020-0.039
in. thick; bare and clad plate 1. 501-4. 000 in. thick; rolled bar, rod and shapes; hand and die forgings; extruded bar, rod and
shapes 0.125-0.749 in. thick with cross-sectional area < 25 sq. in.).
+- ft
NI.HI-Li+H'·
i1
120 +
I j~ fH'
£11 + tt:j=+~.rOO··j.+·HSTO'H.TH
ltt:Hti' EXP05UO'
AT TEMPERATURE
UP TO 10 000 HO
-!- r ."' I+ J :~ i
L _ -/__)..
r+-,-
·tl-
.,~ '-
.L 1
J
+-
I 1/2 HR
1 10
H=j , t 100
HR
HR
1000 HR ~
10 000 HO .1
.]-
II'
1'j , jo
H,-l-
I t]fi.nil~t
j'Hitrm i .
I
o
1
,·,1
-400
1!!I hlH 1t
-300
iH
-200
t +
-100 o 100 200 300 00 500
t· tlP'
l-T·,-.. '·
+, f
I> 0 700 1O
TEMPERATURE, OF
Fig. 82.54 Effect of temperature on the tensile yield strength (Fty) of 2014-T6 aluminum alloy (bare and clad plate
3.001-4.000 in. thick; rolled bar, rod and shapes; hand and die forgings; extruded bar, rod and shapes O. 125-0.499
in. thick with cross-sectional area < 25 sq. in.).
EFFECT OF TEMPERATURE ON 2014 ALUMINUM ALLOYS (Cant.)
100 100
e }
~
~ I eo
'" '~"
J
0
E
0
60
40
~,
~.
~ ~
~ l
200 300 400 500 600 700 800 100 200 300 400 500 600 700 800
Temperature, F 100 •
Temperature. F
e
~
!l
Fig. B2.55 Effect of temperature on E Fig. B2. 58 Effect of temperature on
the compressive yield strength (Fey)
J'" the bearing yield strength (Fbry) of
60
of 2014-T6 aluminum alloy (all 0

2014-T6 aluminum alloy (all products
40
products except thick extrusions). except thick extrusions).
~.
20
tJ
100 ~ 0
~ 120
} 0
•i eo Temperature, F
j
'"
E
~
the ultimate bearing strength (Fbru) w
0
of 2014-T6 aluminum alloy (all §
~
w
products except thick extrusions). C 40
~ tJ
l ,:; 20 .
200 400 500 600 700 800 -600 -400 -200
Temperalure,F TemperQI<Jre, F
Fig. B2.56 Effect of temperature on Fig. B2. 59 Effect of temperature on

the ultimate shear strength (F su ) of the tensile and compressive modulus
2014-T6 aluminum alloy (all products (E and Eel of 2014 and 2017 aluminum
except thick extrusions. alloys.
2024 ALUMINUM ALLOY (BARE SHEET & PLATE, EXTRUSIONS, BAR, ROD & WIRE)
Alloy ... 2024

Form ... .......... Sheet and plate Coiled sheet
Heat treated
Condition. Heat treated by user" Heat treated b and rolled Heat treated
-T42 -T3 -T4 -T36 -T4
0.250- 0.501- 1.001- 2.001- 0.010- 0.250- 0.501- 1.001- 2.001-
Thickness (in.). <0.250 0.500 1.000 2.000 3.000 0.249 0.500 1.000 2.000
3.000 0.012-0.128 0.020-0.500
Basis ... A A
-
A
- ----
A A
-
A
- -
B
---
A B A B AA BB A B A B
Mechanical properties: - - - - - - - - -- - - - - - - - - - - - - - - -- - - - --- - - - ---
FlU/ ksi
L. 62 fJ4 62 60 56 65 68 65 67 63 68 61 64 .. 70 72 62 66
T .. 62 64 62 60 56 64 67 64 66 62 67 60 63 56 59 69 71 62 66
F ty , ksi
L. 40 38 38 38 38 48 51 46 49 44 48 42 46 .. 60 62 40 41
T. 40 38 38 38 38 42 44 40 43 40 44 40 44 40 44 52 54 40 41
Fey, ksi
L. ........ 40 38 38 38 38 40 42 38 41 38 42 38 42 .. .. 49 51 40 41
T. 40 38 38 38 38 45 47 43 46 43 47 42 46 .. 56 58 40 41
F su , ksi. 37 38 37 36 34 40 42 40 41 38 41 36 38 43 44 37 40
Fbru, k.',i c
(e/D=1.5) . 93 96 93 90 84 98 102 98 101 95 102 92 96 .. 105 108 93 99
(e/D~2.0) 118 122 118 114 106 124 129 124 127 120 129 116 122 .. 133 137 118 126
Fbry, ksic
(e/D=1.5) .. 56 53 53 53 53 69 71 64 69 62 67 60 66 .. 84 88 56 57
(e/D;=2.0) .. 64 61 61 61 61 79 82 74 78 70 77 68 75 .. .. 96 100 64 66
e, percent
T. (d) 12 8 (d) 4 (d) 12 .. 1 8 (d) 4 (d) (d)
E, 10 6 psi 10.5
Eel 10 6 psi. 10.7
G, 10' psi.. 4.0
B2.17
2024 ALUMINUM ALLOY (BARE SHEET & PLATE, EXTRUSIONS, BAR, ROD & WIRE) (Cont.)
Table B2. 12 Design Mechanical and Physical Properties of 2024 Aluminum Alloy (Extrusions)
Alloy ...................... 2024
Form ...................... Extruded bars, rods, and shapes
Heat Heat
treated treated
Heat treated by user b cold
worked
and aged
Condition .................. -T4 e -T42 -T81
Thickness, in ............... 0.050- 0.250- 0.500- 0.750- 1.500- 3.000:;;: All thick- ::<0.250
0.249 0.499 0.749 1.499 2.999 ness
Cross-sectional area, in. 2 ••••• ::<25 :;;:25, :;;:32 ::<32

I
Basis ...................... A B A B A B A B A B A B A A
--------------------
F,u. ksi
L .... ............. 57 61 60 62 60 62 65 70 70 74 70 74 57 64
T ................. 57 61 60 62 60 62 58 61 54 57 50 53 50 . ........ .
F,u. ksi
L .. ............... 42 47 44 47 44 47 46 54 52 54 52 54 38 56
T ................. 42 46 43 46 42 45 41 44 38 41 36 39 36 . .........
Feu. ksi
L ................. 38 41 39 42 39 42 44 52 50 52 50 52 38 ..........
T ................. 38 41 39 42 39 42 42 48 42 44 42 44 38 . .........
F elD • ksi ... ............. 30 32 32 33 32 33 34 38 38 40 38 40 30 . .........
F bru . ksi
(ejD= 1.5) ......... 85 91 85 91 85 91 85 91 85 91 85 91 85 ..........
(ejD=2.0) . ........ 108 114 108 114 108 114 108 114 108 114 108 114 108 · . . . . . . . . .
F bru . ksi
(ejD= 1.5) ......... 59 66 60 66 60 66 61 66 62 66 62 66 53 ..........
(ejD=2.0) . ........ 67 75 69 75 69 75 71 75 73 75 73 75 61 · . . . . . . . . .
e, percent
L ................. 12 ... . 12 ... . 12 ... . 10 ... . 10 ... . 10 . ... 12 ..........
Ta ................ ... . ... . 6 ... . 6 ... . 5 . ... 2 ... . 2 ... . . ......... · . . . . . . . . .
E, 106 psi ... ........... 10.5
E e • 106 psi . ............ 10.7
G, 106 psi . ............. 4.0
-
2024 ALUMINUM ALLOY (BARE SHEET & PLATE, EXTRUSIONS, BAR, ROD & WIRE) (Cont.)
Table B2. 13 Design Mechanical and Physical Properties of 2024 Aluminum Alloy (Bar, Rod, and Wire);
Rolled, Drawn or Cold Finished; Rolled Tubing
Alloy ............... , ..... 2024
Form ..................... Bar, rod and wire; rolled, drawn or cold-finished Tubing
Heat Heat
treated treated
Condition ................. -T4 or -T351 Heat treated by cold
user' worked
and aged
Cross-sectional area, in ..... =<36 -T3 -T42 -T81
Thickness, in ......... .... Up to 1.001- 2.001- 3.001- 4.001- 5.001- 6.001- 0.018 to 0.500 0.018-
1.000 2.000 3.000 4.000 5.000 a 6.000 a 6.500 a 0.500
- - - --- - - - - - - - - - - - -
Basis ..................... A A A A A A A A B A A
---------------------
F tu , ksi
L. ............... 62 62 62 62 62 62 62 64 70 62 68
T .. .............. 61 59 57 55 54 52 ..... . ...... ..... . ....... . . .......
Ft., ksi
L ......... ....... 40 40 40 40 40 40 40 42 46 40 60
T . ............... 40 39 38 37 37 36 ..... . ...... ..... . ....... . . .......
F,., ksi
L ................ 32 32 32 32 32 32 ..... . 42 46 40 ........
T ................ ..... . ..... . ..... . ..... . . . . . . . ..... . ............ ..... . ....... . . .......
F,u, ksi ............... 37 37 37 37 37 37 37 39 42 39 .... . .' ..
F bTu , ksi
(e/D=1.5) ..... .... 93 93 93 93 93 93 ..... . 93 105 96 ........
(e/D=2.0) ......... 118 118 118 118 118 118 ..... . 118 133 122 . .......
'
F bT ., ksi
(e/D=1.5) .. oo 56
••••• 56 56 56 56 56 ..... . 59 64 56 ........
(e/D=2.0) ..... .... 64 64 64 64 64 64 ..... . 67 74 64 ........
e, percent
L .. .............. lOb 10 10 10 10 10 10 10 d 0 . .. .. .. . . . ... . . .
•••••
T ... ............. 10 8 6 4 2 2 .. . .. . ..... . ..... . . ....... ........
E, 106 psi. ............ 10.5

E" 10 6 psi. ............ 10.7
G, 10 6 psi ............. 4.0
w, Ib/in. 3 . . . . . . . . . . . . . . 0.100
C, Btu/Clb) (F) .........
K, Btu/[(hr) (ft 2) (F)/ft] .
6
ai, 10- in./in./F ........
70 (at 77 0 F) .
0.23 (at 212 0 F)
12.6 (68 0 to 212 0 F)

Table B2. 14 Design Mechanical and Physical Properties of Clad 2024 Aluminum Alloy (Sheet, Plate, and Coiled Sheet)
Alloy Clad 2024
Form Sheet and plate Coiled sheet Sheet and plate

Heat
Heat treated Heat treated' Heat treated by user • Heat treated Heat treated, cold worked, and aged
and rolled treated and
aged
Condition __ . -T3 -T36 -T4 -T42 -T4 -T6 -T81 -T86
0.010-1 0.063- 0.020-1 0.003-/ 0.250- 0.500- 1.001- 2.001- 0.008- 0.063- 0.250- O. 500- I. 001- 2.001- 0.010- 0.063- 0.010-10.063- <0.0030.020-10.063-10.250-1
Thickness, in. __ . 0.062 0.249 0.062 0.499 0.500' 0.499 1.000' 2.000' 3.000' 0.062 0.249 0.499 I. 000- 2.000' 3.000' 0.062 0.128 0.062 0.249 ;;'0.063 0.062 0.249 0.499 0.500' o
Width, in ..... --____________________ <48 <48 s;:
1-,--.--,---1--.--.--';----.----,-- - - - - - - - - - - - - - - - - - - - - - - -----,-1----1--- 1--.---.----,----,---,-,--,.-- t:::J
Basis . A nAB A BAil A B A BAn A BAn A B A B A A A A A B A B A A A BAn A B A B
-- --------- -------------------- - --- ----- [I)
o
Mechanical prol)('rtil's:
F, •• los; ~
L. m fi~ f>3 fi.' f>3 fifi fi, fi'l fi, fi! fi3 fi5 61 fi5 5~1 62 M 58 5~1 61 6~ 60 5X 5·1 5' 61 61 fi3 60 62 61 67 6, 6U 71 73 70 71 69 70 ~
~
7' 5" fil fi2 r,' fi2 fi.' fifi r" fill Ii' 62 61 60 61 5, 61 51 57 56 58 5U 61 62 60 5X 54 5' 61 61 fi3 60 62 62 6., 66 6, 70 72 70 71 6J 70
F,~. k.'Ji
~
I,. I., 4, 4r, 4; 55 5' 5' fiO 5' fir 46 4~ 42 4fi 40 44 31 35 36 ~ 38 36 36 3fi 3, ~ 3~ 40 47 49 5, 5U 63 65 Ii, 6U 65 66 61 65
7'. :1" 41 4/1 4~ 4.' 50 5/1 52 50 52 40 42 ~ 42 ~ 42 ~ 42 34 35 36 ~ 3S 36 36 36 3, ~ 3., 40 47 4U 51 56 62 61 66 5" 65 fm fil 65
F~", kid
L. 3, 3!1 3, W 4fi 4S 4s 48 4'1 4! 3' 40 36 40 36 40 34 35 36 3, 38 36 36 36 3, ~ ~ 40 47 4U 55 5, 63 65 6, 6U 65 66 61 65 ~
t-<
T. 42 H 13 4:, 51 .51 51 5f> .,1 56 43 45 41 45 10 41 34 35 36 38 as 36 36 36 3, 3, ~ 40 47 49 55 57 65 67 69 71 68 W 67 68 t-<
F•• , k&i.. 3' 3!11O 41 3!/ 41 41 42 41 42 40 41 37 31 34 35 35 3; 37 36 35 3, 3, 3S 36 37 38 39 39 40 41 42 41 42 41 41 Q
F.r •• k.,i d
IHI 1/3 !I.5 9' !15 9!1 101 1Il1 101 101 95 ~ 92 9S 89 93
~
(e/D)=1.5.... 84 87 89 92 93 90 8, 81 S7 92 92 95 90 93 96100 100 103 106 109 105 106 103 105 UJ
114 II~ 121112112.. 12.5127131127131120124116124112118
~
(e/D)=2.0. 100 110 112116 liS 114 110 103 1111 Ilfi 116120 114 liS 12212, 127131 13513!1133135131133
1-'6r". ksi"
(e/D)=1.5... . f>1 f>, f>1 6' ;; SI SI s4 SI S4 61 6,/ 5!1 61 56 62 4S 49 50 53 5.1 50 50 50 52 53 53 56 66 69 7s S3 &j 91 94 97 91 92 90 91 ~
(elD)-2.0... . 73 ,f> 71 ,; SS 93 93 96 93 96 74 7S 6, 74 61 70 51 56 58 61 61 5s 58 5X 59 61 61 64 75 7S IHI 91 101 1114 111,110104 1Of> 102104
e. flt'rC'l'nt "U
T. (et leI 8 9 lOllS (e) 4. (e) 15 . 11 8 (e) 4 (e) 15 8 S 5 5 3. 4 4 ... 4 ... !;
E,lII'psi >-3
1'1i.. 10.5 t"J
~
:r~~i ~I
o
10 o
R" !I.5 1 I 9.5 I 10.0 I 9.5 I 111.0 I 95 1 10 0 I 95
1 10 0 19 51 9.5 I 10 0 H
&;
Pri. 10.7 t:::J
Rl'r.... !I. 7 110.2 I 9.7 I 10.2 1 li.7 I 10.2 I 9., I 111.2 I 9.7 1 1 ~I
10. 2 9. 71 9.7 I 10.2
UJ
~
(;, 10' p.,i ..
~
Physic'ul pron'rlh's:
"', Ih/in'. ... . . . . . . . . . . . O.]()() 0.100
C, Btu/(I6)(F)... . . . . ..... , 0.23 (at 2120 (0') 0.23 (at 212 0 F)
K. Btu/[(hr)(ft ')(F)I. .• " 70 (at ; ; 0 Jo') 86 (at 77 0 F)
a 10-' in.jin.W.·. . ... 12.6 (f>,o to 212 0 F) 12.6 (68 0 to 2120 F)
• For the stress relieved temP<'r-T351 (platt' onl}'), all valtl('s fo1- th,· -T4 t"mp"r apply with tIl(' inasmuch as a round t('5t sI){'cimpn is rpfluin'd for testing. The valu('s given hefe fOf thicknl'ss 0.500
exception of ~' •• which may he som('what lo\\'l'r. inch and grl'atpr hav(' hepn adjustp(l to n'pn'spnt till' U\'l'ragc propl'rtil's across till' whole section,
... Heat treated hy user" ref"rs to all mall' rial solution hl'at tr"att'd hy th" USl'r n'garcll,'ss of the inrll1lling til(' dadding.
prior temper of the material. d S"l' Tahll' 3.1. I. I. I.
, Hpeclflcation minimums for elad mahorial 0.5IH) ind' t hi"k and Il('avil'r an- for th" rOr" mat"rial. • Tilt' ('longation valut's for t'W5(' rolumns an' Counll in TahIl' 3.2.3.0 (p).
...~
<0
100 100
!
~
~
E
~
~ '0
J 60
J 60
"
.,z "
,,-b
40
~ 20
~
or or
200 300 400 500 600 700 800
Temperature, F Temperature, F
Fig. B2. 60 Effect of temperature on Fig. B2. 64 Effect of exposure at Fig. B2. 68 Effect of temperature on
the ultimate tensile strength (Ftu) of elevated temperatures on the room- the ultimate bearing strength (Fbru)
2024-T3 and 2024-T4 aluminum alloy temperature ultimate tensile strength of clad 2024-T3 and clad 2024-T4
(all products except extrusions). (Ftu) of 2024-T3 and 2024-T4 aluminum aluminum alloy (sheet).
alloy (all products except thick
extrus ions).
100 100 100

! !
,!l
!
t
p
~ eo eo eo
!l E
E E
~ ~ ~
E
J 60
J 60
~
60
"..: 40
"
~
40
"
~i;
40
~ 20
~ 20
~
20
or or .r
200 300 400 500 600 700 800 200 300 400 500 600 700 800 100 200 300 400 500 600 700 800
Temperature, F Temperature, F Temperoture,F
Fig. B2.61 Effect of temperature on Fig. B2. 65 Effect of temperature on Fig. B2. 69 Effect of temperature on
the ultimate tensile strength (Ftu) of the compressive yield strength (Fcy) the bearing yield strength (Fbry) of
2024-T3 and 2024-T4 aluminum alloy of clad 2024-T3 and clad 2024-T4 clad 2024-T3 and clas 2024-T4
(extrusions). aluminum alloy (sheet). aluminum alloy (sheet).
i eo
~
~
E 60
~
"
-400 200 400 600 800 1000

Temperature,F Temperature,F Temperature, F
Fig. B2. 62 Effect of temperature on Fig. B2. 66 Effect of exposure at Fig. B2. 70 Effect of temperature on
the tensile yield strength (Fty) of elevated temperatures on the room- the tensile and compressive modulus
2024-T3 and 2024-T4 aluminum alloy temperature tensile yield strength (E and Eel of 2024 aluminum alloy.
(all products except extrusions). (Fty) of 2024-T3 and 2024-T4
aluminum alloy (all products except
thick extrusions).
100
eo
20
100 200 300 400 500 600 700 800 100 200 300 400 500 600 700 800 100 200 300 400 500 600 700 800
Temperature, F Temperature, F Temperature. F
the tensile yield strength (Fty) of the ultimate shear strength (F su ) of the elongation of 2024-T3 and 2024-T4
2024-T3 and 2024-T4 aluminum alloy clad 2024-T3 and clad 2024-T4 aluminum alloy (all products except
(extrusions). aluminum alloy (sheet). thick extrusions).
Alloy ........................ 7075

Form........................ Sheet and plate
Condition .................... -T6a
Thickness, in ............... 0.015-0.039 0.040--0.249 0.250--0.500 0.501-1.000 1.001-2.000 2.001-2.500 2.501-3.000 3.001-3.500 3.501-4.000
Basis........................ A B A B A B A B A B A B A B A A
-- ---- ---------- ------ --
Flu, ksi
L .... ............... 76 78 77 79 77 79 79 82 78 80 73 75 70 72 68 66 --.J
T .................. 76 78 77 79 77 79 77 80 77 79 73 75 70 72 68 66 o
--.J
ST ..... ............. ·. ·. ·. ·. ·. ·. ·. ·. ·. ·. 67 69 64 66 64 62 CJ1
Fly, ksi :r>

L .. ................. 66 69 67 70 67 69 69 72 68 71 62 65 60 62 58 56 t:
~
T .................. 65 68 66 69 66 68 66 69 66 69 62 65 60 62 58 56
ST ..... .............
Fey, ksi
·. ·. ·. ·. ·. ·. ·. ·. ·. 61 63 59 60 56 54
~:r>
L . .................. 67 70 68 71 69 71 69 72 68 71 65 67 63 65 62 60 t:-<
t:-<
T .............. ..... 70 73 71 74 69 71 69 72 68 71 65 67 63 65 62 60
ST .. ................ ·. ·. ·. ·. ·. ·. ·. ·. ·. ·. 64 66 62 63 62 60 Q
F. u, ksi . ................. 46 47 46 47 46 47 47 49 46 47 43 45 41 43 40 39 ~
Fbru, ksi b :xJ
tel
(eiD = 1.5) ........... 114 117 116 119 • 108 110 110 115 109 112 102 105 98 101 95 94 Ro
(e/D=2.0) .. ......... 144 148 146 150 139 142 142 147 140 144 131 135 126 130 122 119 ()
Fbry, ksi b ~
(eID=1.5) ........... 92 97 94 98 87 90 90 94 88 92 81 84 78 81 75 73 t::J
(eID=2.0) ........... 106 110 107 112 100 104 104 108 102 106 93 97 90 93 87 84 OJ
::r::
e, percent tel
L .... ............... . 7 · 8 ·. 8 ·. 6 ·. 5 ·. 5 ·. 5 ·. 5 .. ~
T ................... 7 ·. 8 ·. 8 ·. 6 ·. 4 ·. 3 ·. 3 ·. 3 2 Ro
ST .. ................ ·. ·. ·. ·. ·. ·. ·. ·. ·. ·. 1 ·. 1 ·. 1 ..
E, 10' psi . ..............
E e, 10' psi ........ .......
10.3
10.5
S
tel
0, 10' psi................ 3.9
w, Ib/in. 3 . . • . • . • • • • . . . • • . 0.101
C, Btu/(lb)(F) . .......... 0.23 (at 212°F)
K, Btu/[(hr)(ftt)(F)/ftj ... 76 (at 77°F)
lX, 10-' in./in./F . ......... 12.9 (68° to 212°F) .,
G For the .tress relieved temper -T661. all values for the -T6 temper apply with 2.001-2.600 L 62
the ex~eption of F ••• Applicable F •• values are B8 folio..... : 2.601-3.000 L 60
Tkic1me.. (in.) Direeno.. 0/ teat Fell (A "alue.) b See Table 3.1.1.1.1.
0.260-2.000 L 66
...
...
k,
tl:J
~
N
N
Table B2. 16 Design Mechanical and Physical Properties of Clad 7075 Aluminum Alloy (Sheet and Plate)
s::ttl
(')
Alloy .......................... Clad 7075 ~

Z
Form.......................... Sheet and plate n
;I>
Condition ...................... -T6G t"'
0.015- 0.040- 0.063- 0.188- 0.250- 0.500- 1.001- 2.001- 2.501- 3.001- 3.501- ;I>
Z
Thickness, in ................. 0.039 0.062 0.187 0.249 0.499 1.()()()b 2.()()()b 2.500 b 3.00<Y' 3.5o()b 4.000b t:1
Basis .......................... A B A B A B A B A B A A B A B A B A B
-- A
- -...J 'tI
-- - - - - ~ - - - - - - - - - - - - - - -- -- -- -- - - -- - - - --- 0
-...J
::I:
Mechanical properties: 0, ~
......
Feu. ksi-L ... .............. 70 73 72 74 73 75 75 77 75 77 77 79 76 78 71 73 68 70 66 64 :t> (')
;I>
T ... ............. 70
ST .... ............ ·.
73
·.
72
·.
74
· .
73
·.
75
·.
75
·.
77
·.
75
· .
77
· .
75
·.
77
·.
75
·.
77
· .
71
67
73
69
68
64
70
66
66
64
64
62 ~
H
t"'
'tI
:<l
FII/' ksi-L .. .............
Too ...... ........ 60
61 64
63
63
62
65
64
64
63
66
65
65
64
67
66
65
64
67
66
66
64
68
66
66
64
69
67
60
60
62
62
58
58
60
60
56
56
54
54 ~ 0
'tI
ttl
ST . . , ............. . ·. 63 59 60 56 54 :t>
FCI/. ksi-L . ................ 62
·. ·.
65 64
·. ·.
66
·.
65
·
67
·.
66 68
·.
66
·.
69
·.
66
·.
69
·.
66
·.
69
61
62 64 61 62 60 58
L"
L" ~
......
0 ttl
T .... ............. 64 67 66 68 67 69 68 70 66 69 66 69 66 69 62 64 61 62 60 58 >< Ul
ST ... ............. ·. ·. ·. ·. ·. ·. . ·. . ·. . 64 66 62 63 62 60 tIl 0

·. ·. · ·. · · :t> ""J
F.... ksic. .................. 42 44 43 44 45 45 46 45 46 46 47 45 46 42 43 40 41 39 38 ::0
44 t"j s::ttl
Fbru, ksi-(e/D=1.5} ... ..... 105 110 108 111 110 112 112 116 105 108 108 III 106 109 99 102 95 98 92 90 Ro ~
(e/D=2.0} ........ 133 139 137 141 139 142 142 146 135 139 139 142 137 140 128 131 122 126 119 115 ;I>
0 t"'
Fbrl/. ksiC-(e/ D = 1.5) ....... 85 90 88 91 90 92 91 94 84 87 86 88 86 90 78 81 75 78 73 70
~ t"'
(e/D=2.0} ... ..... 98 102 101 104 102 106 104 107 98 100 99 102 99 104 90 93 87 90 84 81 t::J n
8 . ..
~~
e, percent-L .............. 7 ·. 8 · . ·. 8 ·. 8 ·. 6 ·. 5 · . 5 ·. 5 · 5 [jJ
::r::
T .............. 7 ·. 8 · . 8 ·. 8 ·. 8 ·. 6 ·. 4 ·. 3 ·. 3 ·. 3 2 t'J
t'J
ST .... ......... ·. ·. ·. · . ·. ·. · . ·. ·. ·. ·. ·. ·. ·. 1 ·. 1 ·. 1 .. r-'J ttl
:<l
......
E, 10' psi-Pri.. ............ flo ;I>
10.:3 10.3 10.3 t"'
Sec.•........... 'l:I Ul
E c, 10' psi-Pri . ............
9.5 9.8 10.0 ~ ""J
10.5 10.5 10.5 r-'J 0
Sec............ t'J :<l
9.7 10.0 10.2
G, 10' psi . ................. ""J
0 t"'
......
Physical properties: 0
::::s 0
"" Ib/in. 3 . . . . . . . • . • . . . . • . . . 0.101 rt ::I:
~
C, Btu/ (lb) (F) .............. 0.23 (at 212°F) <
K, Btu/[(hr)(ft')(F)/ftj ...... 76 (at 77°F) ttl
IX, 10-' in./in./F ............ 12.9 (68° to 212°F) ei
(')
t"'
GFor the stress relieved temper -T661, all values for the -T6 temper apply with the b These values except in the ST direction have been adjusted to include the influence ttl
exception of F c.. Applicable F.. values are as follows: of the 1 1/2% per side nominal cladding thickness. Ul
~
Tmck....,sa (in.) Direction 0/ teat FCII (A .,aluea) c See Table 3.1.1.1.1. :<l
0.266-0.499 L 62 c::
(')
0.600-2.000 L 64
~
2.001-2.600 L 60 c::
2.601-3.000 L 68 :<l
ttl
Ul
Table B2. 17 Design Mechanical and Physical Properties of 7075 Aluminum Alloy (Extrusions)
Alloy ............................. 7075

Form ............................. Extrusions (rod, bars, and shapes)
CondItion ......................... -T6 a
Cross-sectional area, in. 2 • • . . • • • . . . . . <20 <20 I >20, <32 <32
Thickness b, in..................... Up to 0.249 0.250-0.499 0.500-0.749 0.750-1.499 1.500-2.999 3.000-4.499 --J
4.500-5.0
A
8
-B- - - -B- - - -B- -A- -B- -A- -B- -A- -B- -A- -B- A
Basis ............................. A A B en
Mechanical properties: >
F,u, ksi §
L ........................ 78 82 81 85 81 85 81 85 81 85 81 84 78 81 78 8
T ........................
F 'u, ksi
76 78 77 79 73 75 72 74 66 68 62 64 62 64 60 62
~
L . ....................... 70 74 73 77 72 76 72 76 72 76 71 74 70 73 68 7 >
T ........................ 64 67 66 68 63 66 62 64 56 58 54 56 54 56 53 55
~
Q
Feu, ksi
........
L . .......................
T ........................
71
71
75
75
74
74
78
78
73
73
77
77
72
72
76
75
72
69
75
72
72
65
75 68 72 ..... . .....
68 ..... . ...... . ..... .... . e
F ou , ksi . ......................
F bru , ksi
43 45 45 47 45 47 45 47 45 47 45 46 43 44 43 44
~
H
.~
(eID=1.5) . ............... 101 107 97 102 97 102 97 102 97 102 97 101 94 97 94 97
(eID=2.0) . ............... 125 131 130 136 130 136 130 136 130 136 122 126 117 125 117 12 5
F bru , ksi >xj
(eID=1.5) . ............... 91 96 80 85 79 84 79 84 79 84 78 81 77 80 75 78 @
(eID=2.0) . ............... 98 104 102 108 101 106 101 106 101 106 92 96 91 95 88 92 Q
H
~
e, percent
L ... ..................... 7 8 7 8 7 8 7 8 7 8 7 8 6 ..... . 6 ..... ~
T ........................ 5 ..... . 5 ..... . 4 ...... 3 . ..... 1 . ..... 1 . ..... 1 . ..... 1 . ....

E, 106 psi . ....................
B., 106 psi . ...................
10.3
10.5 ~
~
G, 106 psi . .............. ' ..... 3.9 :::0
o
Physical properties: ..
t:J
w, Ib/in. 3 • . . . . . . • , . . . . . . . . . . . . . 0.101 ::E:

H
C, Btu/(lb) (F) ................. 0.23 (at 212°F) :::0
i:'j
K, Btu/[(hr) (ft2)(F)/ft] . ......... 76 (at 77° F)
a, 10- 6 in./in./F . ............... 12.9 (68° to 212°F)
• For the stress relieved tempers -T6510 and -T6511, all values for the -T6 temper apply, with Thickne.. (in.) Area (sq. in.) Direction Of Test F••(A values)
Lhe exception of F ,.-L. Applicable F," values are listed below: 1.500-2.999 <20 L 70
Thickness (in.) Area (sq. in.) Direction oltest F '. (A values) 3.000-4.499 <32 L 66
<0.249 <20 L 70 4.50()-5.000 <32 L 65
0.250-0.749 <20 L 72 • For extrusions with outstanding legs, the load carrying ability of such legs shall be determined
0.750-1.499 <20 L 71 on the basis of the properties of the appropriate column correspondinl!" to the leg thickness.
~
....""
7075 ALUMINUM ALLOY (EXTRUSIONS, FORGINGS, BAR, ROD, WIRE) (Cant.)
Table B2. 18 Design Mechanical and Physical Properties of 7075 Aluminum Alloy
(Hand Forgings and Die Forgings)
Alloy .................... 7075
Form.................... Hand-forged stock, length Hand-forged stock, length Die

<3 times width >3 times width forgings
Condition................ -T64 -T6
Thickness, in............ <3
Cross-sectional area, in. t . . . <16 >16, <36 >36, <144 <16 >16, <36 >36, <144
Basis .................... A A A A A A A
F,u, ksi
L .. ............. 75 73 71 75 73 71 75
T ............... 75 71 69 73 71 69 71
ST .............. 72 68 66 70 68 66 ·.
F'll, ksi
L . .............. 64 61 60 63 60 59 65
T ............... 63 60 58 61 59 57 62
ST . ............. 63 60 58 61 59 57 ·.
Fell' ksi
L .. ............. 64 61 60 63 60 59 65
T . .............. 63 60 58 61 59 57 58
ST . ............. .. .. .. .. .. .. · .
F..., ksi .............. 45 44 43 45 44 43 45
Fbru, ksi
(e/D=1.5) ....... 97 95 85 97 95 85 ·.
(e/D=2.0) ....... 135 124 114 135 124 114 ·.
Fbrt/,ksi
(e/D=1.5) ....... 90 79 78 88 78 77 ·.
(e/D=2.0) ....... 96 91 90 94 90 88 ·.
e, percent
L .. ............. 9 7 4 9 7 4 7
T ............... 4 3 2 4 3 2 3
ST . ............. 2 2 1 2 2 1 ·.
E, 10- psi ............ 10.3
E e, 10- psi . .......... 10.5
G, 10- psi . ........... 3.9
B2.25
7075 ALUMINUM ALLOY (EXTRUSIONS, FORGINGS, BAR, ROD, WIRE) (Cant.)
Table B2. 19 Design Mechanical and Physical Properties of 7075 Aluminum Alloy
(Bar, Rod, Wire and Shapes; Rolled, Drawn or Cold-Finished)
Alloy ............ , .... 7075
Form ................. Bar, rod, wire and shapes, rolled, drawn or cold-finished
Condition ......... " .. -T6 or -T651
Thickness, in .......... Up to 1.001 2.001 3.001

1.000 a 2.000 a 3.000 a 4.000 a
Basis ................. A A A A
Flu, ksi:
L ... ......... 77 77 77 77
LT ........... 77 75 72 69
Ft., ksi:
L ............ 66 66 66 66
LT ........... 66 66 63 60
Fe., ksi:
L ... ......... 64 64 64 64
LT ........... ........... . ........... . ........... . . ...........
F ou , ksi. ......... 46 46 46 46
F bru , ksi:
(e{D=1.5) .... 100 100 100 100
(e{D=2.0) .... 123 123 123 123
F br ., ksi:
(e{D=1.5) .... 86 86 86 86
(e{D=2.0) .... 92 92 92 92
e, percent:
L ... ........ b7 7 7 7
LT .......... 4 3 2 1
E, 10 6 psi. ........ 10.3

E e, 10 6 psi. ........ 10.5
G, 10 6 psi. ......... 3.9
'00
1
J
80
60
.
'6 40
~
20
,t
100 200 300 400 500
'00 Temperolure,F '00
iJ 80
60
Fig. 82. 75 Effect of exposure at
elevated temperatures on the room-
temperature tensile yield strength
'6 40 (Fty) of 7075-T6 aluminum alloy
.t (all products).
5 20
,t
100 200 300 400 500 600 100 800
Tempera'ure, F Tempwah..... F
'00
Fig. B2. 72 Effect of temperature on Fig. B2.79 Effect of temperature on
the ultimate tensile strength (Ftu) of the bearing yield strength (Fbry) of
7075-T6 aluminum alloy (all products). 7075-T6 aluminum alloy (all products).
11 40
..t
~ 20
~
s '00
of
"
120
J 80 tOO 200 300 400 500 600 700 800 ~ 100

E
Temperature, F ~
E 60 j 80
~ Fig. B2. 76 Effect of temperature on

;; 40 the compressive yield strength (Fey)
.E of 7075-T6 aluminum alloy (all products).
~
20
,t 0
0 200 300 400 !lOO 700 800 -400 -200 0 200 400 600 800 1000
'00
Temperoture,F Temperature. F
'00

the tensile yield strength (Fty) of
7075-T6 aluminum alloy (all products).
IJ 80
60
Ptg.- B2~ 80 Effect of temperature Oil
(E and Ec) of 7075-T6 aluminum alloy.
'6
..~
40
~
20
,t
100 200 300 400 500 600 700 800
Temperature, F

fhe ultimate shear strength (Fsu) of
7075-T6 aluminum alloy (all products).
100 200 300 400 500 600 700 800 100 200 300 400 500 600 700 800
T.mperat...... F Temperature, F
,00
Fig. B2. 74 Effect of exposure at Fig. B2.81 Effect of exposure at
elevated temperatures on the room- elevated temperatures on the elongation
temperature ultimate tensile strength of 7075-T6 aluminum alloy (all products
(Ftu) of 7075-T6 aluminum alloy (all except thick extrusions).
products).
300 400 500 600 700 800

Tempera'ure, F
Fig. B2. 78 Effect of temperature Oil

the ultimate bearing strength (Fbru)
of 7075-T6 aluminum alloy (all products).
Table B2. 20 Design Mechanical and Physical Properties of AZ31B Magnesium Alloy (Sheet and Plate)
Alloy ........................ AZ31B
Form ........................ Sheet and plate
Condition ... _................ -0 -H24 -H26
Thickness (in.) ................ 0.016- 0.061- 0.250- 0.501- 0.016- 0.250- 0.375- 0.501- 1.001- 0.250- 0.375- 0.439- 0.501- 0.751- 1.001- 1.501-
0.060 0.249 0.500 2.000 0.249 0.374 0.500 1.000 2.000 0.374 0.438 0.500 0.750 1.000 1.500 2.000
- - - - - -- -- -- - - - - - - - - - ~
Ba.sis ........................ A B A B B A N
-- - - - - -- - - - - - - - - - - -- -- -- - - -- - CN
f-'
t:.d
F,., kBi.
L .................... 32
T ................... ... .
F,,, k8i.
36 32 36
... . ... . ... .
32 30
..... . ..... .
39
40
29
40
.., .
38
39
26
37
38
24
36
37
22
34
35
20
39
40
27
38
39
26
38
39
26
37
38
25
37
38
23
35
36
22
35
36
21
I
H
~
L .................... 18 19 15 19 15 15 30
~
T ................... ... . ... . ... . ... . ..... . ..... . 32 .... 29 27 25 23 30 29 29 28 26 25 24 L'
L'
Fell' kBi. Q
L .................... 12 13 12 14 10 10 24 25 20 16
13 10 22 21 18 17 16 15 14
Tb ................... ... . .. , . . , .. ... . ..... . ..... . ... . ... . •••• ..... . ..... .
0 • ..... . ..... . ..... . ..... . .. ... ..... . ...... ..... . ~
UJ
F••, ksi . ................. 17 23 17 23 17 ..... . 18 26 18 18 ..... . ..... . 18 18 18 ..... . ...... ..... . ...... :r::
F b,., k8i.
...... ..... . ...... ..... .
~
(e/D=1.5) ............ 50 52 50 52 50 ..... . 58 60 56 54 ..... . ..... . 58 56 56 Be
(e/D=2.0) ............ 60 61 60 61 60 ...... 68 70 65 63 ..... . ..... . 68 65 65 ..... . ..... . •••• ..... .
0 •
F b,.", k8i.
(e/D=1.5) ............ 29 30 29 31 27 ..... . 43 44 38 34 ..... . ..... . 40 39 36 ..... . ..... . ..... . ......
'"~
trl
(e/D=2.0) .......... .. 29 30 29 31 27 ..... . 43 44 38 34 ..... . ..... . 40 39 36 ..... . ..... . ..... . ......
e, percent
L .................... 12 18 12 19 12 10 6 8 8 8 8 8 6 6 6 6 6 6 6
T ................... ... . ... . ... . .... . ..... . ..... . 8 .... 10 10 10 10 8 8 8 8 8 8 8
E, 10° p8i . ............... 6.5
E., 10° p8i ................ 6.5
,
G, 10° psi ................ 2.4
til, lb/in.' ................. 0.0639
C, BTU/(lb)(F) ........... 0.25 (at 780 F)o
K, BTU/(hr)(ftl)(F)/ft ..... 56 (212 to 5720 F)
4, lO*t in./in./F ........... 14 (65 to 212 0 F)
• Estimated. • Transve1'llll F. r allowables are equal to or greater than the 10ngltudlDai F. r allowables.
~
N
"""
B2.28 MECHANICAL AND PHYSICAL PROPERTIES OF METALLIC MATERIALS FOR FLIGHT VEffiCLE STRUCTURES
AZ31B MAGNESIUM ALLOY (SHEET & PLATE) (Cant.)
100
IE
~
80
0
il
IL
~
20
~
Leoend 100 200 300 400 500 600 700 800
A:Ttnsil~train
B=Compressiye stress-strain Tomperoture, F
C:TensUt tanl;Jent modulus
O=Cor'1)rtllive ton;ent
modulus Fig. 82. lie Effect of temperature on the ultimate shear
strength (Fsu) of AZ31B-H24 magnesium alloy.
4 • 10
Strain,QOOI in./1n. 100
Tanoent Modulus. 10' 1'1;
Fig. B2. 82 Typical stress-strain and tangent-modulus

curves for AZ31B-O magnesium alloy at room
temperature (longitudinal).
100
e
~ 80
~ 100 200 300 400 500 600 700 800
!!
Temperature, F
j 60
0 40 FIg. 82. 87 Effect of temperature on the ultimate bear-

,
.r ing strength (!bru) of AZ31B-H24 magnesium alloy.
5 20
100
~
200 300 400 500 600 700 800
Temperature. F
Pig. 82. 83 Effect of temperature on the ultimate

tensUe strength (Ftu) of AZ31B-H24 magnesium alloy.
I
100
80
100 200 300 400 500 600 700 800
j 60 Temperature. F
o 40 FIg. 82. 88 Effect of temperature on the bearing yield

strength (FDry) of AZ318-H24 magnesium alloy.
50 ==o::r-----'L;;;E9EN=P"-------n
A-1inIiIe ,tresl-strain C- Tenailt tal'lQ8l'lt modulus
B- Co~reuiVi Ifrell-Straln 0- ~ive tonoent modulus
100 200 300 400 !500 600 700 800
Temperature, F 40
FIg. 82. 84 Effect of temperature on the tensUe yield

strength (Fty) of AZ318-H24 magnesium alloy.
100
10
4 I 10
S_,o.OOII~/1n.
200 300 400 500 600 700 800 TOIlQ8ll'-',ocfpsi

Temperature. F
Fig. 82. 89 Typical stress-strain and tangent-modulus
Fig. 82. 85 Effect of temperature on the compressive curves for AZ31B-H24 magnesium alloy at room
yield strength (Fey) of AZ31B-H24 magnesium alloy. temperature.
Table 82.21 Design Mechanical and Physical Properties of HK31A Magnesium Alloy (Sheet, Plate and Sand Castings)
Alloy .......................... HK31A
Form ....................... '" Sheet G and plate G Sand

castings b
Condition ...................... -0 -H24 -T6

~
~
Thickness, in. . .................. 0.016- 0.251- 0.501- 1.001- 0.016- 0.126- 0.251- 0.501- 1.001- 2.001-
Basis ........................... A
0.250
B
0.500 1.000 3.000
A
0.125
B A
0.250
B
0.500 1.000 2.000 3.000 ..........
(e)
I
r:n
H
- -- - §
~
Mechanical properties: t""
t""
F ,a, ksi ...... ............... 30 32 30 30 29 34 36 34 35 34 34 33 33 27 Q
F,~, ksi ...... ............... 18 19 16 15 14 26 28 23 24 21 18 15 15 13
F~, ksi ..................... 12 13 10 10 10 20 21 19 22 18 13 11 11 13 (j;"
F ,a , ksi ........ ............. 22 22 22 22 . ....... 23 23 23 23 23 23 . ....... ....... . ......... . ~
F b• a , ksi
(e/D=1.5) .............. 43 46 43 . ...... . ....... . 49 52 49 51 49 . ....... ....... . ....... . ......... . ..ga
....... . . ....... 57 . .......
Fbm ksi
(e/D=2.0) .............. 51 54 51 57 60 57 58 . ....... . ....... . .........
~
8
(e/D=1.5) .............. 24 25 21 ....... . ....... . 34 35 33 36 31 . ....... . ....... ....... . ......... . tr:l
(e/D=2.0) .............. 24 25 21 . ....... ....... . 34 35 33 36 31 . ....... ....... . ....... . . ......... Be
r:n
e, percent ................... 12 20 12 12 12 4 6 4 8 4 10 10 8 4 ~
E, lOS psi .... ............... 6. 5 o
Es, lOS psi ............... '" 6. 5 55
8
G, lOS psi ............ .... '" 2. 4 H
Physical properties: ~
101, lb./in.3 ••••••••••••••••••• 0.0647
C, BTU/(lb.)(F) ............. 0.25 (32 0 to 212 0 F)
K, BTU/(hr.)(ft.I)(F) ft ...... 60.0 (at 680 F)
a, lO-s in./in./F ............. 15 (68 0 to 3920 F)
• Properties for sheet and plate are taken parallel to the direction of rolllng. Transverse properties I Mechanical properties are based upon the guaranteed tensile properties from separately-east
are equal to or greater than the longitudinal properties. test bars. The mechanical propertIes of bars cut from castings may he as low as 75 percent of the
• Reference should he made to the specific requirements of the procuring or certificating agency tabulated values.
with regard to the use of the above values In the design of castjngs.
~
~
co
82.30 MECHANICAL AND PHYSICAL PROPERTIES OF METALLIC MATERIALS FOR FLIGHT VEffiCLE STRUCTURES
HK3lA MAGNESIUM ALLOY (SHEET, PLATE & SAND CASTINGS) (Cant.)
'00 '00
~
!
80
i. 80
~
0-
E
~ ~
J
15
60 E
8
0:
15
60
40 40
u=-
~ ,.
....
~ 20
u
.
1; 20
:f cf
200 300 400 500 600 700 800 '00 200 300 400 500 600 700 800
Temperature, F Temporalure, F
Fig. B2. 90 Effect of temperature on the ultimate Fig. B2.93 Effect of exposure at elevated
tensile strength (Ftu) of HK31A-H24 magnesium temperatures on the room-temperature tensile
alloy. yield strength (Fty) of HK31A-H24 magnesium
alloy.
'00 '00
!
I~
E
eo
60
i0-
E
{!!
§
0:
80
60
15 40
0 40
,,! ~
1;
u
. 20
u
Ii 20
cf ~
'00 200 300 400 500 600 700 800 '00 200 300 400 500 600 700 800
Temperature. F
Temperature, F
Fig. B2.91 Effect of temperature on the tensile Fig. B2.94 Effect of temperature on the ultimate
yield strength (Fty) of HK31A-H24 magnesium tensile strength (Ftu) of HK31A-T6 magnesium
alloy. alloy (sand casting).
100 '00
! !
2
t 80 .
2
0-
E
eo
E
{!! ~
E 60
§ 60
~
0:
15 15 40
40
,.
....2 'L
1; 20 ~ 20
~ U
~ cf
200 300 400 500 600 700 800 '00 200 300 400 500 600 700 800
Temoeroture, F Temperature, F
Fig. B2. 92 Effect of exposure at elevated Fig. B2.95 Effect of temperature on the tensile
temperatures on the room-temperature ultimate yield strength (Fty) of HK31A-T6 magnesium
tensile strength (F tu ) of HK31A-H24 magnesium alloy (sand casting).
alloy.
B2.31
AZ6lA, AZ63A, AZ80A MAGNESIUM (EXTRUSIONS, FORGINGS, CASTINGS)
Table B2. 22 Design Mechanical and Physical Properties of AZ61Aa Magnesium Alloy
(Extrusions and Forgings)
Alloy . AZ61A
Extruded bar, rod, and Extruded Extruded

Form . solid shapes Hollow shapes Tubes Forging
Condition . -F
Thickness, in . <:0.249 0.250-2.499 0.028-0.750
Basis .
F,u, ksi
L . 38 39 36 36 38
T .
F'II' ksi
L . 21 24 16 16 22
T .
Fell' ksi
L . 14 14 11 11 14
T .
F , ksi . 19 19 19
F/wu,ksi
(e/D=1.5) . 45 45 50
(e/D=2.0) . 55 55 60
FIn-t/,ksi
(e/D=1.5) . 28 28 28
(e/D=2.0) . 32 32 32
e, percent . 8 9 7 7 6
E, 10< psi . 6.3
E e, 10< psi . 6.3
G, 10< psi . 2.4
w, lb/in. ' . 0.0647
C, Btu/(lb)(F) . 0.25 (at 78°F) b
K, Btu/[(hr) (ft 2) (F)/ft) . 46 (212° to 572°F)
ac, 10-< in./in./F . 14 (65° to 212°F)
a ProPertif.S for extruded bars, rods, shaPes, tubes, and forgings are during fabrication.
taken parallel to the direction of extrusion or maximum metal 1I0w b Estimated.
B2.32 MECHANICAL AND PHYSICAL PROPERTIES OF METALLIC MATERIALS FOR FLIGIIT VEffiCLE STRUCTURES
AZ6lA, AZ63A, AZ80A MAGNESIUM: (EXTRUSIONS, FORGINGS, CASTINGS) (cont.)
Table B2. 23 Design Mechanical and Physical Properties of AZ80Aa Magnesium Alloy
(Extrusions and Forgings)
Alloy .... . . . . . . . . . . . . . . . . . . . . . AZ80A
Form ......................... Extruded bars, rods, and solid shapes Forgings
Condition ..................... -F -T5 -F -T5
Thickness, in ................. <:0.249 0.250- 1.500- <:0.249 0.250- 1.500-

1.499 2.499 1.499 2.499
Basis..........................
F,u, ksi
L .. ................... 43 43 43 47 48 48 42 42
7'.....................
F'I/' ksi
L ..................... 28 28 28 30 33 33 26 28
T ... ...................
p'el/' ksi
L ..................... 17 17 28 27 18 25
T .....................
J1'.u, ksi . .................. 19 19 19 20 20 20 20 20
Fb...., ksi
(e/D = 1.5) ............. 48 48 48 50
(e/D=2.0) ............. 56 56 56 70
Fbrl/' kb'i
(e/D=1.5) ............. 36 36 36 42
(e/D=2.0) ... .......... 40 40 40
e, percent.................. 9 8 6 4 4 4 5 2
E, 10' psi .... .............. 6.5
E e, 10' psi . ................ 6.5
G, 10' psi... ............... 2.4
B2.33
AZ6lA, AZ63A, AZ80A MAGNESIUM (EXTRUSIONS, FORGINGS, CASTINGS) (Cant.)
Table B2. 24 Material Specifications for AZ63A

25
Magnesium Alloy L·V· nd
A-Tensile stress· strain
B' Compr...ive .Ir... ·.Iraln
C· T.n.lI. lanv.nl· modulu.
20 O·Comp....ive laRV.nl· modulu.
Specification Type of product
C A
0 B
~~-~-56 . Sand' castings "II. 15
~~~-55 Permanent-mold castings
. §
TABLE 4.2.3.0 (b).DetJign M echa.nica,l and Ph1ltJica,l

! 10
Properties of AZ68A MagnetJium G Alloy (Castings)

5
Alloy . AZ63A
o
Sand and permanent- o 2 4 6
Strain,Oool in./in.
• 10
Form . mold castings Tongenl Modulus,IO'PSl
Fig. B2.96 Typical stress-strain and tangent-modulus

Condition . -F -T4 -T5 -T6 curves for AZ63A-T4 magnesium alloy (sand casting)
at room temperature.
Thickness (in.) .
L89.~d
A-T.n.ll••Ir...·.lraln
Basis b . B -Compres.lv••Ir••• ·.lraln
C- T.n.lI. laRV.nl modulu.
------------1--- --- --- --- 20 O-C~ve Ianvenl·modulu.
~echanical properties:
F,u, ksi . ............. 24 34 24 34 A
F'll, ksi ............... 10 10 11 16 B
F CI/, ksi . ............. 10 10 11 16
FlU, ksi . ............. 16 17 19
Fbru,ksi
(e/D=1.5) ...... ,. 36 36 50
(e/D=2.0) . .... " 50 50 65
Flny,ksi
(e/D=1.5) ....... 28 32 36
(e/D=2.0) ....... 30 36 45
e, percent ............ 4 7 2 3
E, 10' psi . ........... 6.5
E e, 10' psi ... ......... 2 4 6 6 10
6.5
Strain,Oool in./in.
G, 10' psi . ........... 2.4 TORVenl Modulus, 10' psi
Fig. B2.97 Typical stress-strain and tangent-modulus

curves for AZ63A-F magnesium alloy (sand casting) at
room temperature.
Table B2. 25 Design Mechanical and Physical Properties '00 tll
i
~
of 8Mn Titanium Alloy
i c.o
lI>-
! 80
i!::
Alloy \ 8Mn
j J 60
t:I:J
(')
t;
~
o
,!
40 ~
(')
~
Form 1 Sheet, plate, and strip >-
~
~ 20 t"
~'tl
~ ,f
400 600 800 1000 '200 '400 1600 600 800 '000 '200 1400 1600
Condition. Annealed 200
Temoerature. F Temperature, F
Fig. B2.101 Effect of temperature on the ultimate

=i
00
Fig. B2. 98 Effect of temperature on the ultimate (')
Thickness, in . shear strength (Fs u) of 8 Mn annealed titanium
tensile strength (Ftu) of 8Mn annealed titanium >-
...,~
alloy. alloy. t"
'tl
Basis . A ::tl
H 0
'tl
100 '00
~ t:I:J
iJ ~
~
Mechanical properties: 2 H ~
~
60 § l;;j
Flu, ksi ~ :J>
00
L . 120 t""' 0
7' . 120 60 j t""' 'oj
Fly, ksi t; 40 0 ~ i!::

t:I:J
~ >-3
L . 110 ,£ IL ill >-
T . 110 ~ 20
'E
tJ ~ t"
...
t"
Fey, ksi
L . 110
~ .t .~ (')
~
200 400 .. 600 800 1000 200 400 600 800 '000 '200 '400 '600
T .
~...,
110 Tempera'ure; f Temperature, F
>-3
F BU, ksi . 84 t:I:J
~t"
Fig. B2.99 Effect of temperature on the tensile Fig. 2. 102 Effect of temperature on the ultimate t<:l
Fbru, ksi yield strength (Fty) of 8 Mn annealed titanium bearing strength (Fbru) of 8 Mn annealed titanium
(e/D=1.5) . 170 Be
alloy. alloy. I:/)
00
(e/D=2.0). >-'3 'oj
Fbry, k.~i ::0
H @
(e/D=1.5) . 130 '"U 'oj
f
(e/D=2.0) .
If
100
t
100
...0
t"
e, percent . 10
E, 10' psi . 15.5 80 ~ 80
~
E
E e, 10' psi . 16.0
F
8 60 <
G, 10' psi .
j 60
It:
;;
t:I:J
...IJ::
'""
(')
t; 40 40 t"
,.. g t:I:J
IL"
Physical properties: S 20 '"ij 20

00
>-3
w, lb/in.>
C, Btu/(lb)(F)
.
.
0.171
0.118 (at 68°F) .t
u
,f El
(')
200 400 600 800 1000 '200 '400 1600 200 400 600 800 '000 '200 '400 '600 >-3
K, Btu/[(hr) (ft')(F)/ft) 6.3 Temperature, F Temperature, F c:::
ct, 10-' in./in./F . 4.8 (at 200°F) ::tl
t:I:J
Fig. B2. 100 Effect of temperature on the Fig. 2.103 Effect of temperature on the tensile 00
compressive yield strength (Fcy) of 8 Mn and compressive modulus (E and Ed of 8Mn
annealed titanium alloy. titanium alloy.
B2.35
6Al-4V TITANIUM ALLOY (BAR & SHEET)
Table B2. 26 Design Mechanical and Physical Properties

of 6Al-4V Titanium Alloy
Alloy . 6AI-4V
Form . Bar Sheet
Condition . Annealed
Thickness, in . <1.5 <0.187

Temperature, F
Basil!
--_. __.--------- ----- ------
. A A Fig. B2. 104 Effect of temperature on Ftu, Fty and E of
6Al-4V annealed titanium alloy (sheet and bar)
Mechanical propertielS:
F,u, hi
L . 130 130
T .. 130
1"/1, ksi
L 00', 120 120
T . 120
[<'Cy, hi
L . ]211 126
T . 126
[< , ksi . 80 76
1'" ksi
(eID=1.5) . 196 191 Fig. B2. 105 Effect of temperature on the compressive
(eID=2.0) . yield strength (Fey) of 6Al-4V annealed titanium alloy
248 244 (sheet and bar).
1'",.", ksi
(eID=1.5) . 174 163 100
(eiD =2.0) . 205 198
e, percent . 10 10
E, 10- psi
L 00 .. ]6.0 15.4
T . 16.4
E c, 10- psi
L .. 16.4 16.0
T . 16.9
G, 10- psi . 6.2
Physical properties: Fig. B2.106 Effect of temperature on the ultimate

Col, lb/in.•............. 0.160 bearing strength (Fbru) of 6Al-4V annealed titanium
C, Btu/(lb)(F) . 0.135 (at 68°F) alloy (sheet and bar).
K, Btu/! (hr)(ft")(F)/ft). 3.8 (at 63°F)
GI, 10-- in./in./F . 4.6 (at 200°F)
Fig. B2.107 Effect of temperature on the I1ltimate shear

strength (Fsu) of 6Al-4V annealed titanium alloy (sheet
and bar).
INCONEL X NICKEL ALLOY (SHEET)
Table B2. 27 Design Mechanical and Physical Properties

of Inconel X Nickel Alloy
Alloy. . . . . . . . . . . . . . . . . . . .. Inconel X
Form . Sheet
Condition . Precipitation hea t-treated Fig. B2. 108 Effect of temperature on

precipitation heat treated Inconel X
Thickness, in . nickel alloy.
Basis .
:\Iechanical properties:
Flu, ksi
L . 155
Temperature,F
T . 155
Fly, ksi Fig. B2. 109 Effect of temperature on
the tensile yield strength (Fty) of
L . 100
T . 100 nickel alloy.
Fey, ksi
L .. 105
T .. '" 105
F. u , ksi. 108
Fbrl" ksi
(elD=1.5)
(e/D=2.0). 286 .! zo
Fbry, ksi ~ 0
o
(e/D=1.5)
(e/D=2.0) 186 Fig. B2. 110 Effect of temperature on
e, percent . 20 the compressive yield strength (Fey) of
E, 10' psi . 31.0 precipitation heat treated Inconel X
Eel 10' psi . 31.0 nickel alloy.
G, 10' psi .
Phyaical properties:
w, lb/in. 3 .. 0.304
C, Btu/(ib)(F) ... 0.109
K, Btu/[ (hr)(ft 2)(F)/ft] 8.7 (80 0 to 212°F)
(x, 10-' in./in./F . 6.4 (100 0 to 200°F)

the ultimate shear strength (F su ) of
nickel alloy.
S I
1j
r,mperotUAl, F Temperotu.... F
the tensile modulus (E) of Inconel X the ultimate bearing strength (Fbru) the bearing yield strength (Fbry) of
nickel alloy. of precipitation heat treated Inconel X precipitation heat treated Inconel X
nickel alloy. nickel alloy.
PART C
PRACTICAL STRENGTH ANALYSIS &
DESIGN OF STRUCTURAL COMPONENTS
CHAPTER Cl
COMBINED STRESSES. THEORY OF YIELD AND ULTIMATE FAILURE.
Cl. I Uniform Stress Condition ~l
Aircraft structures are subjected to many

types of external loadings. These loads often
cause axial, bending and shearing stresses
acting simultaneously. If structures are to be
designed satisfactorily, combined stress re-
lationships must be known. Although in practi- ~l
cal structures uniform stress distribution is Fig. Cl. 3
not common, still sufficient accuracy for
design practice is provided by using the stress stresses.
relationships based on uniform stress assump-
tions. In deriving these stress relationships, Fig. Cl.4 shows a free body after the
the Greek letter sigma (0) will represent a block has been cut along a diagonal section.
stress intensity normal to the surface and thus
a tensile or compressive stress and the Greek For equilibrium the sum of the forces along
letter tau (~) will represent a stress intensity the x-x axis equals zero.
parallel to the surface and thus a shearing
stress. 0(1) + 2 (~l cos 45 0
cos 45
) =0
Cl. 2 Shearing Stresses on Planes at Right Angles.
0 0
/-dx/" hence, 0 = 2 (~ 1 cos 451 ) cos 45 -- ~

(2)
~y Therefore when a point in a body is sub-

2 1
T~:; ~1
;' /'
, , jected to pure shear stresses of intensity ~,
normal stresses of the same intensity as the o
~zjb:t~z A Po, to
shear stresses are produced on a plane at 45
dz ~LJ---- ~
with the shearing planes.
2 1
L '':xdxdy - -
Cl. 4 Principal Stresses
Fig. Cl.I Fig. Cl.2
For a body SUbjected to any combination of
stresses 3 mutually perpendicular planes can be
Fig. Cl.l shows a circular solid shaft found on which the shear stresses are zero. The
subjected to a torsional moment. The portion normal stresses on these planes of zero shear
(A) of the shaft exerts a shearing stress ~z on stress are referred to as principal stresses.
section (1-1) and portion (B) exerts a resist-
ing shearing stress ~z on section (2-2). Fig.
Cl. 5 Shearing Stresses Resulting From Principal
Cl.2 illustrates a differential cube cut from Stresses.
shaft between sections (1-1) and (2-2). For
equilibrium a resisting couple must exist on In Fig. Cl.5 the differential block is
top and bottom face of cube. Taking moments SUbjected to tensile principal stresses Ox and
about lower left edge of cube: Oz and zero principal stress 0Y' The block is
cut along a diagonal section giVing the free
~x dxdy (dz) - ~z dzdy (dx) =0 body of Fig. Cl.6. The stresses on the diagonal
section have been resolved into stress compon-
hence, ~x = ~z - - - - - - - - - - - (1) ents parallel and normal to the section as shown.
For equilibrium the summation of the stresses
Thus if a shearing unit stress occurs on along the axes (1-1) and (2-2) must equal zero.
one plane at a point in a body, a shearing unit
stress of same intensity eXists on planes at
right angles to the first plane.
Cl. 3 Simple Shear Produces Tensile and Compressive
On dudy - Ox dzdy cos Q - Oz dydx sin Q = 0,
Stresses. whence on = Ox dzdy cos Q + 0z dydx sin Q
dUdy dudy
Fig. Cl.3 shows an elementary block of
unit dimensions SUbjected to pure shearing
Cl.I
C1.2 COMBINED STRESSES. THEORY OF YIELD AND ULTIMATE FAILURE.
o z dydx Cl.6 Combined Stress Equations

Fig. Cl.7 shows a differential block sub-
jected to normal stresses on two planes at right
angles to each other and with shearing forces on
the same planes. The maximum normal and shear-
ing unit stresses will be determined.
Fig. CI.8 shows a free body diagram of a
portion cut by a diagonal plane at angle Q as
shown.
z dydx
Fig. Cl. 5
1
/
/
/
n dudy
Ox dzdy
/
Fig. Cl. 7
l'
o z dydx
Fig. Cl. 6
dzdy _ dydx _
But dudy - cos Q and dudy - sin Q, whence
(2a)
The normal stress on at a point is always
less than the maximum principal stress Ox or
Oz at the point.
ZF._. =0 Fig. Cl. 8
- 't dudy - Ox dzdy sin Q + Oz dydx cos Q =0

For equilibrium the sum of the forces in the z
and x directions must equal zero.
But ~~~; = cos Q and ~~: = sin Q
ZFx =0
=
hence, 't Oz sin Q cos Q - Ox cos Q sin Q
or, 't =
(Oz - ax) sin Q cos Q an dudy cos
=0 - -
Q + 't dudy sin Q - ax dzdy - 'txzdxdy
- - - - - - - - - - - - - - - -(4)
or, 't =
(l/2)(az - ax) sin 2 Q, where Oz is
maximum principal stress and ax is minimum ZF z = 0
principal stress.
an dudy sin Q - 't dudy cos Q - a z dydx - 'txzdzdy
Since sin 2 Q is maximum when Q 45 a , = =0 - - - - - - - - - - - - - - - - - -(5)
't = (amax - a min )/2 - - - - - - - - - (3)

By dividing each equation by du and noting that
Stated in words, the maximum value of the shear- ~~: :: cos Q and ~~: =sin 9, we obtain:
ing unit stress at a point in a stressed body
is one-half the algebraic differences of the
maximum and minimum principal unit stresses. (an-(Jx) cos Q + ('t - 'txz) sin g =0 - - - -(6)
C1.3
(an - Oz) sin Q - (~ - ~xz) cos Q =0 - - -(7) Cl. 7 Mohr's Circle for Determination of Principal
Stresses.
It is sometimes convenient to solve
The maximum normal stress an will be maximum
when Q equals such angle Q' as to make ~ zero. = graphically for the principal stresses and the
Thus if ~ = =
0 and Q Q' in equations (6) and
maximum shear stress. MOhr's circle furnishes
a graphical solution. (Fig. Cl.9a). In the
(7), we obtain,
Mohr method, two rectangular axes x and z are
chosen to represent the normal and shearing
(an - ax) cos Q' - ~xz sin Q' =0 - - - - -(8) stresses respectively. Taking point 0 as the
origin layoff to scale the normal stresses ax
(an - Oz) sin Q' - ~xz cos Q' =0 - - - - -J9) and Oz equal to OB and OA respectively. If ten-
In equations (8) and (9) an represen~£he sion, they are laid off to right of point 0 and
principal stress. Dividing one equation by to the left if compression. From B the shear
another to eliminate Q' , stress ~xz is laid off parallel to Oz and with
the sense of the shear stress on the face DC of
Fig. Cl.9b, thus locating point C. With point
an - ax = ~xz
whence, E the midpoint of AB as the center and with
~xz an - Oz radius EC describe a circle cutting OB at F and
G. AD will equal BC and will represent the
an" - (ax - Oz) an + 0xOz = ~xz .. , or shear on face AB of Fib. b. It can be proven
that OF and OG are the principal stresses Omax
a
n
= ax + Oz + _ '(ax - Oz)
2 - ·V 2
il ..
+ ~xz -
_ (10) and Omin respectively and EC is the maximum •
shear stress ~max .• The principal stresses occur
on planes that are parallel to CF and eG. (See
In equation (10), tensile normal stress is
plus and compression minus. For maximum an use Figs. c and d). The maximum shear stress occurs
plus sign before radical and minus sign for on two sections parallel to CH and CI where HEI
minimum an. is perpendicular to OB. Ifox should equal zero
then 0 would coincide With A.
To find the plane of the principal stress- z
es, the value of Q' may be solved for from
equations (8) and (9), which gives: Oz °n(max)
~ 1 ~xz
tan 2 Q' =ax2 -~xzOz (ll )
Pl---
or-[]I~ox
~xz-~-
Q' is measured from the plane of the
largest normal stress ax or oz. The direction
(b) Oz ~m1
of rotation of Q' from this plane is best de-
termined by inspection. Thus if only the Oz
shearing stresses ~xz were acting, the maximum
principal stress would be one of the 45 0 planes,
L~xz 0
~~
E
the particular 450 plane being easily deter- (0) I
mined by inspection of the sense of the shear /
stresses. Furthermore if only the largest 61\1..t1-\'30i'l Ox I ;
~
normal stress were acting it would be the maxi- I;
mum principal stress and Q' would equal zero.
Thus if both a and ~ act, the plane of the
principal stress will be between the plane on
(d)
~xz
ZT
~
J;foX; Oz
0x+ Oz
--2-
~x-Oz
H
(a)
which a acts and the 45 0 plane. As stated 2
before a refers to either ax or Oz whichever Oz ax
is the largest. Z Fig. Cl. 9
Maximum Value of Shearing Stress. (~max.)
Cl. 8 Components of stress From Principal stresses by
Mohr's Circle.
The maximum value of ~ from equation (3)
equals, In certain problems the principal stresses
may be known as in Fig. Cl.9 and it is desired
~ax. = (On(max.) - On(min.))/2 - - - - - (12) to find the stress components on other planes
designated by angle Q. In Fig. Cl.ll the axes
Substituting the maximum and minimum values of x and z represent the normal and shear stresses
an from (10) in (12), we obta1n maximum shear- respectively. The principal stresses are laid
ing stress as follows: off to scale on ox giving points D and E respec-
tively. Construct a circle with A the midpoint
~max. =+ of DE and With diameter ED. Draw angle CAB equal
CL4 COMBINED STRESSES. THEORY OF YIELD AND ULTIMATE FAILURE.
to 2Q. It can be proven that OB represents the

normal stress on the plane defc of Fig. Cl.lO,
~xz= 5000
.
oX"9~ejD ;x. 1O~
and CB represents the shear stress ~ on this
plane.
Oz
""_+-_ _,,,c • to
Fig. CL 12 ~xz =5000
d
Z
-t--_Ox
Oz, Ox are principal

Fig. CLIO stresses. X--~=--I*-'------"«-----'..--+'..--'-'--X
z
On on plane (defc) ~
I _I
Fig. CL 13
~-r
~ on plane 0n(min)
= -2070
1 -+--{1,
Z
E
(defc)
x 0 X
SUbstituting values,
Oz on = 1000~ + 0 ~ ~ (1000~ + 0)2 + 5000 2 = 5000

Oz + Ox
+ 7070 hence, On( max. ) =5000 + 7070 = 12070 psi
2
Ox ---------1 On(min.) = 5000 - 7070 =-2070 psi
z Fig. CLll ~max. = (1/2)(On( max. ) - On( mi n. )) (Ref.Eq.12)

= (1/2)(12070 - (-2070)) =7070 psi
CL 9 Example Problems.
~max.can also be computed by equation (13),
Example Problem 1.
The maximum normal and shear stresses will be whence,

determined for the block loaded as shown in . /(1000~ + 0)2 + 5000 2 =~ 7070 psi
Fig. Cl.12. '"max. = + V
The graphical solution making use of Mohr's

circle is shown in Fig. Cl.13. From reference
tan 2Q" =Ox2 ~xz
- Oz
= 2 x 5000
10000 - 0
=1
axes x and z thru point 0, the given normal
= Q~ = 22.5
0
stresses Ox 10000 is laid off to scale on ox hence, •
and toward the right giving point B. From B the
shear stress ~xz =
5000 is laid off parallel to Example Problem 2.
oz to locate point C. With E the midpoint of
OB as the center of the circle and with radius The maximum normal and shear stresses will
EC a circle is drawn which cuts the Ox axis at be determined for the block loaded as shown in
F and G. The maximum and minimum principal Fig. Cl.14.
stresses are then equal to of and oG which o
g Oz = 20000
equals 12070 and -2070 respectively. The maxi-
mum shear stress equals EC or 7070. : ~. ~xz = 12000
Algebraic Solution:
On
__ Ox + Oz + .
2 -
From eq. (10),
YI(ox 2- Oz\
-I
2+
~xz
2
Ox"~ ~jDI~'OX'lO~
~xz = 12000
t
Oz = 20000
Fig. CL 14
Cl.5
Fig. Cl.15 shows the graphical solution 0" - (ox + 0y + 0z)O· + (0xOy + 0yOz + 0xoz
using Mohr's circle. From point 0, Ox 10000 =
and 0z =
-20000 are laid off equal to OB and OA -r yz • _Yxy2)0 - (0xOyOz + 2YyzYxzYxy - 0xYyz·
respectively. ~xz equal to 12000 is laid off - 0yYxz • - 0zYxy • =0 - - - - - - - - - - -(14)
parallel to OZ at B locating C. With E the
midpoint of AB as the center of a circle of
radius EC a circle is drawn which cuts the ox
axis at F and D. The maximum normal and shear
stresses are indicated on the figure.
~
ITYX-
/ 'l;
Fig. Cl. 15 / Yxy
/0
Y
/
- -- t ~~
Y Fig. Cl.16
...I?,
.- ...~-1-
~~ Fig. Cl.17 shows the principal stress
X X system which replaces the system of Fig. Cl.16.
E 0 B
It can be shown that the maximum shear stress
max. is one of the following values.
I
I '(max. = +-"2
1
(01 0.)
+ 1
or Y max. = -2 (02 0,,) -- - - - -(15)
or ."., max. = +- 21 (0" - 01 )
The planes on which these shear stresses

act are indicated by the dashed lines in Fig.
Algebraic Solution Cl.lS, namely, adhe, bdge and dcef. The
r
largest of the shear stresses in equations (15)
On = Ox 2+ °z ± -J (Ox 2 °z + ~xz
•
depends on the rr.agnHude and signs of the
principal stresses, remembering that tension
is plus and compression is minus when making
- 20000 1: ~(10000-~-20000))·+ 12000· the substitution in equations (15).
= 10000 2
°l.
- - 5000 :!: 19200 "0,,
.
Iffi
, I
,
hence, °n(max. ) = -5000 =
19200 -24200 psi
0,,'
-- ,/,- -
I
0.
°n(min. ) = -5000 + 19200 = 14200

q, °1 Fig. Cl.18
2 x 12000
~max. = 1: 19200, tan 2Q'
1000 - (-20000) = .S Fig. Cl.17
Q' = ISO - 50' Cl. 11 Principal Strains
Cl. 10 Triaxial or Three Dimensional Stresses The strains under combined stresses are
usually expressed as strains in the direction
For bodies which are stressed in three of the principal stresses. Consider a C8,se of
directions, the state of stress can be defined simple tension as illustrated in Fig. Cl.19.
completely by the six stress components as The stress 01 causes a lengthening unit strain
illustrated in Fig. Cl.16. Using the same e in the direction of the stress 01' and a
procedure as was carried out for a two-dimen- shortening unit strain e' in a direction at
sional stress system, it can be shown that there right angles to the stress 01,.
are three principal stresses 01,' 0. and 0e,
whose values are the three roots of ° in the The ratio of e' to e is called Poisson's
following cubic equation. ratio and is usually given the symbol~. ThUS,
~ =e'/e
C1.6 COMBINED STRESSES. THEORY OF YIELD AND ULTIMATE FAILURE.
~,_LJ~_~~ The strain energy can be expressed in

terms of stress by SUbstituting values of E: in
I I
I
I
!
I
terms of 0 from equations (17) into equation
I
I (19), which gives,
I
f
10' -.: e
2" i: "2
:--rr-- J
- - - - - - - - - - - - - - - - - -(20)
O~ Fig. Cl. 19
Since 10 = o~/E, we obtain, For a two dimensional stress system,
Os = 0 and equation (20) becomes
10' = j.L o~/E - - - - - - -(16)
Now consider the cubical element in Fig.
Cl.20 subjected to the three principal stresses
O~, 0a and 0 8 , all being tension. The total Cl. 13 Structural Design Philosophy. Limit and Ultimate
unit strain e~ in the direction of stress o~ Loads. Factors of Safety. Margin of Safety.
will be expressed. Obviously, o~ tends to
stretch the element in the direction of o~ The basic philosophy governing the
whereas stresses oa and 0 8 tend to shorten the structural design of a flight vehicle is to
element in the direction of O~, hence, develop an adequate light weight structure that
will permit the vehicle to accomplish the
operations or missions that were established
as design reqUirements. The job of a
commercial airliner is to carry passengers and
cargo from place to place at the lowest cost.
To carry out this job a certain amount of
flight and ground maneuvering is reqUired and
and similarly for lOa and 10 8 ,
the loads due to these maneuvers must be
- - - - -(17) carried safely and efficiently by the structure.
lOa = °i - t(08 + o~) A military fighter airplane must be maneuvered
in flight far more severely to accomplish its
desired job as compared to the commercial air-
liner, thus the flight acceleration factors
For a two-dimensional stress system, that for the military fighter airplane will be
is, stresses acting in one plane, 0 8 0 and = considerably higher than that of the airliner.
the principal strains become, In other words, every type of flight vehicle
will undergo a different load enVironment,
which may be repeated frequently or infre-
E:~ =~ (O~ - .lJO"a ) quently during the life of the vehicle. The
load enVironment may involve many factors such
as flight maneuvering loads, air gust loads,
E: a = E1 (Oa - ~~) - - - - -(18) take off and landing loads, repeated loads,
1 high and low temperature conditions, etc.
E: 8 = E (O~ + °a)
Limit Loads. Limit loads are the calculated
Equations 17 and 18 give the strains when maximum loads which may be subjected to the
all the principal stresses are tensile stresses. flight vehicle in carrying out the job it is
For compressive principal stresses use a minus designed to accomplish during its life time
sign when SUbstituting the principal stresses of use. The term limit was no doubt chosen
in the equations. because every flight vehicle is limited
relative to the extent of its operations. A
Cl. 12 Elastic Strain Energy flight vehicle could easily be designed for
loads greater than the limit loads, but such
extra strength which is not necessary for
The strain energy in the elastic range for safety would only increase the weight of the
the unit cube in Fig. Cl.20 when subjected to structure and decrease the commercial or
combined stresses is equal to the work done by military payload or in general be detrimental
the three gradually applied principal stresses to the design.
o~, oa and 0;. These stresses produce strains
equal to E:~, E: a and E: s and thus the work done Factor of Safety. Factor of safety can be
per unit volume equals the strain energy. Thus defined as the ratio considered in structural
if U equals the strain energy, we obtain, design of the strength of the structure to
U = ¥ + ~ + ~ - - - - - - -(19)
the maximum calculated operational loads,
that is, the limit loads.
C1.7
Yield Factor of Safety. This term is defined incapable of performing its required function.
as the ratio of the yield strength of the Failure may be due to rupture or collapse or
structure to the limit load. due to excessive deflection or distortion.
Ultimate Factor of Safety. This term is Cl. 15 Determination of the Ultimate Strength of a Structural
defined as the ratio of the ultimate strength Member Under a Combined Load System. Stress
of the structure to the limit load. Ratio-Interaction Curve Method.
Yield Load. This term is defined as the limit Since the structural designer of flight
load multiplied by the yield factor of safety. vehicles must insure that the ultimate loads
can be carried by the structure without failure,
Ultimate Load. This term can be defined as the it is necessary that reliable methods be used
limit load multiplied by the ultimate factor to determine the ultimate strength of a
of safety. This resulting load is often structure. Structural theory as developed to
referred to by engineers as the design load, date is in general sufficiently developed to
which is misleading because the flight vehicle accurately determine the ultimate strength of
structure must be designed to satisfy both a structural member under a single type of
yield and ultimate failure and either one may loading, such as axial tension or compression,
be critical. pure bending or pure torsion. However, many
of the members which compose the structure of
Yield Margin of Safety. This term usually a flight vehicle are subjected simultaneously
expressed in percent represents the additional to various combinations of axial, bending and
yield strength of the structure over that torsional load systems and thus a method must
strength required to carry the limit loads. be available to determine the ultimate strength
of a structure under combined load systems.
Yield Margin of Safety - Yield Strength - 1 A strictly theoretical approach appears too
- Limit Load difficult for solution since failure may be
due to overall elastic or inelastic buckling,
Ultimate Margin of Safety. This term usually or the local elastic or inelastic instability.
expressed in percent represents the additional
ultimate strength of the structure over that The most satisfactory method developed to
strength required to carry the ultimate loads. date is the so-called stress ratio, inter-
action curve method, originally developed and
_ Ultimate Strength presented by Shanley. In this method the
Ultimate Margin of Safety - Ultimate Load - 1 stress conditions on the structure are repre-
sented by stress ratios, which can be con-
Cl. 14 Required Strength of Flight Structures. sidered as non-dimentional coefficients
denoting the fraction of the allowable stress
Under Limit Loads:- or strength for the member which can be
The flight vehicle structure shall be developed under the given conditions of com-
designed to have sufficient strength to carry bined loading.
simultaneously the limit loads and other
accompanying environmental phenomena for each For a single simple stress, the stress
design condition without undergoing excessive ratio can be expressed as,
elastic or plastic deformation. Since most
materials have no definite yield stress, it is
common practice to use the unit stress where a R = stress ratio = ~ -(22)
.002 inches per inch permanent set eXists as
the yield strength of the material, and in where f is the applied stress and F the
general this yield strength stress can be used allowable stress. The margin of safety in
as the maximum stress under the limit loads terms of the stress ratio R can be written,
unless definitely otherwise specified.
1
Under Ultimate Loads:- M.S. = R- 1.0 - - - - - - - - - -(23)
The flight vehicle structure shall be
designed to withstand simultaneously the Load ratios can be used instead of stress
ultimate loads and other accompanying environ- ratios and is often more convenient.
mental phenomena without failure. In general
no factor of safety is applied to the environ- For example for axial loading,
mental phenomena but only to the limit loads.
R - P/Pa , where P = applied axial load
Failure of a Structure:- and Pa the allowable load.
This term in general refers to a state or
condition of the structure which renders it
Cl. 8 COMBINED STRESSES. THEORY OF YIELD AND ULTIMATE FAILURE.
For pure bending, 1. Maximum Principal Stress Theory

2. Maximum Shearing stress Theory
R = MIMa , where M = applied bending 3. Maximum Strain Theory
moment and Ma the allowable bending moment. 4. Total Strain Energy Theory
5. Strain Energy of Distortion Theory
For pure torsion, 6. Octahedral Shear Stress Theory
R = T/Ta , where T is applied torsional The reader may review the explanation and
moment and Ta the allowable torsional moment. derivation of these 6 theories by referring
to such books as listed at the end of this
For combined loadings the general chapter.
conditions for failure are expressed by
Shanley as follows:- Test results indicate that the yield
strength at a point in a stressed structure
R~ + R~ + R~ + ------ = 1.0 - -(24) is more accurately defined by theories 5 and
6 followed in turn by theory 2. Since
In this above expression, R~, R 2 and R3 theories 5 and 6 give the same result, they
could refer to compression, bending and shear might be considered as the same general theory.
and the exponents x, y, and z give the In this chapter we will only give the resulting
relationship for combined stresses. The equations as derived by theory 6, since
equation states that the failure of a theories 5 and 6 appear to be the theories
structural member under a combined loading will used in flight vehicle structural design.
result only when the sum of the stress ratios
is equal to or greater than 1.0. Cl. 17 The Octahedral Shear Stress Theory.
For some of the simpler combined load Since this theory gives the same results
systems, the exponents of the stress ratios in as the well known energy of distortion method
equation (24) can be determined by the various it is often referred to as the Equivalent
well known theories of yield and failure that Stress Theory. The octahedral shear stress
have been developed. However, in many cases theory may be stated as follows:- In elastic
of combined loading and for particular types action at any point in a body under combined
of structures the exponents in equation (24) stress action begins only when the octahedral
must be determined by making actual failure shearing stress becomes equal to 0.47 fe'
tests of combined load systems. where f e is the tensile elastic strength of
the material as determined from a standard
Since the stress ratio method was pre- tension test. Since the elastic tensile
sented by Shanley many years ago, much testing strength is somewhat indefinite, it is common
has been done and as a result reliable inter- practice to use the engineering yield strength
actiop equations with known exponents have Fty. In this theory it is assumed that the
been obtained for many types of structural tensile and compressive yield strengths are
members under the various combined load the same.
systems. In a number of the following chapters,
the interaction equations which apply will be Figs. Cl.21 and Cl.22 illustrate the
used in determining the Ultimate strength conditions of equilibrium involving the octa-
design of structural members. hedral shear stress. In Fig. Cl.21, the cube
is SUbjected to the 3 principal stresses as
Cl. 16 Determination of Yield Strength of a Structural shown. A tetrahedron is cut from the cube
Member Under a Combined Load System. and shown in Fig. Cl.22. Three of the sides
of this tetrahedron are parallel to the
As explained in Art. Cl.14, the flight
vehicle structure must carry the limit loads
without yielding, which in general means the
yield strength of the material cannot be
exceeded when the structure is sUbjected to
the limit loads. In some parts of a flight
vehicle structure involving compact unit or
pressure vessels, biaxial or triaxial stress
conditions are often produced and it is
necessary to determine Whether any yielding 2
will occur under such combined stress action
when carrying the limit loads. For cases
where no elastic instability occurs, the
following well known theories of failure have
been developed.
Fig. Cl. 21 Fig. Cl. 22
C1.9
principal axes, while the normal to the fourth For a triaxial stress system,
side makes equal angles with the principal
axes. The octahedral shear and normal stresses _1/ 2 2 2 2
are the resulting stresses on the fourth side. f= ~v (fx-fz) +(fz-f y ) +(f,~fx) +6(fs +f s +f s )
y2 J D ~ P
- - - - - - - - - - - - - -(31)
The equation for the value of the normal
octahedral stress is, For a biaxial stress system, f y , f syz '
f sp = 0
1
foct = ~ (f~ + f 2 + fa) - - - - - -(25) f = J f~ 2
+ f~ - fxf z + 3f sxz - - -(32)
The equation for the octahedral shear C1. 18 Example Problem 1.
stress is,
A cylindrical stiffened thin sheet fuse-
lage is fabricated from 2024 aluminum alloy
sheet which has a tensile yield stress
Fty = 40000. Find the yield margin of safety
Now the octahedral shear stress is 0.47 under the following limit load conditions.
of the normal stress.
(1) A limit bending moment produces a bending
Let t be the effective axial stress in stress of 37000 psi (tension) at top
uniaxial tension or compression which results point of fuselage section. The flexural
in the given octahedral shear stress. shear stress is zero at this point.
(2) Same as condition (1) but pressurization

- -(27) of fuselage produces a circumferential
tension stress of 8600 psi and a longi-
Therefore mUltiplying Eq. (26) by tudinal tension stress of 4300 psi.
3/~ we obtain for a condition of principal
triaxial stresses, (3) Same as condition (2) but a yawing
maneuver of airplane produces a limit
torsional shearing stress of 8000 psi
in fuselage skin.
SOLUTION: Condition (1)
Let F equal the allowable tensile or
compressive stress. If the yield strength is
being determined then This is a uniaxial stress condition for
point being considered.
F
Margin of Safety M.S.= -=- - 1 -(29) Fty 40000
f Yield M.3. = ft - 1 = 37000 - 1 = .08
For a biaxial stress system taking fa = 0, SOLUTION: Condition (2)
we obtain,
t = J f~ + f: - f~f2 - - - - - - -(30)
There are no flexural shear stresses at
the fuselage point being considered. Since
It is often more convenient to use the x, y no torsion is being applied
and z component of stresses instead of the to fuselage no torsional
principal stresses. Fig. Cl.23 illustrates shear stresses exist. The
the various component stresses. stress system at the point
being considered is thus a
fz biaxial stress system and
f~ and f 2 are principal
stresses.
f~ = 37000 + 4300 =41300 psi
f 2 = 8600 psi
From equation (30),
Fig. C1. 23 =J 41300 2 + 8600 2 - 41300 x 8600

Cl.I0 COMBINED STRESSES. THEORY OF YIELD AND ULTIMATE FAll.URE.
whence f =37700 psi The vessel wall is to be stressed to the

yield stress of 42000, thus f 42000. =
_ F 40000
M.S. - f - 1 = 37700 - 1 :: .06 Whence
SOLUTION: Condition (3) (42000)s
Since a torsional
shear stress has now Solving, p = 970 psi.
been added, the new
stress is still two PROBLEMS
dimentional, however
the given tension (1) The combined stress loading at a point
stresses are not in a structure is as follows:- f z =
principal stresses -1000, f x =-2500, f s =
2000. Determine
due to the addition the magnitude and direction of the
of the torsional shear principal stresses. Determine the
stress. maximum shearing stress. Solve both
analytically and graphically.
f x =41300 psi. fz =8600 psi. fs =8000 psi.
(2) Same as Problem 1, but change f z to 4000
Instead of finding the principal stresses and f x to -3000 and f s to 2500.
and using Eq. (30), we will use the f x and f z
stresses and use Eq. (32) (3) A solid circular shaft is subjected to a
limit bending moment of 122000 inch
-
f =VIf xs+ s
fz - fxf z + 3f ss pounds and a torsional moment of 250,000
inch pounds. If diameter is 4 inches
= .,; 41300 s+ 8600 s- 41300 x 8600 + 3 x 8000 s and the yield tensile stress is 42,000,
what is yield Margin of Safety.
_ 40000 _
f = 40200 psi. M.S. - 40200 - 1 - -.01 (4) A thin walled cylinder of diameter 6
inches is subjected to an axial tensile
Thus yield is indicated since M.S. is load of 15,000 pounds, and a torsional
negative. moment of 12,000 inch pounds. What
should be the wall thickness if the
permissible yield stress is 30,000 psi.
Example Problem 2.
(5) A closed end cylindrical vessel is 15
A cylindrical pressure vessel is 100 inches in diameter and a wall thickness
inches in diameter and 1 inch thick. The of 0.25 inches. The vessel is subjected
vessel is made of steel with Fty 42000 psi. = to an internal pressure of 10,000 psi,
Determine the internal pressure that will and a tensile load of 22,000 pounds. If
produce yielding. the yield tensile stress of the material
is 75,000 p i, what torsional moment can
SOLUTION: This applied stress system is be added without causing yield.
biaxial with no flexural or torsional shear.
REFERENCES: -
Let: p equal internal pressure
t = wall thickness = 1 in.
d :: diameter 100" = Nadai, Theory of Flow and Fracture of Solids.
Timoshenko, Strength of Materials.
fs = circumferential stress due
to pressure p FreUdenthal, The Inelastic Behavior of
Engineering Materials and Structures.
f s -- 2t
pd
an
d f
J.
- pd
- 4t
Marin, J., Engineering Materials.
From Eq. 30
Seely & Smith, Advanced Mechanics of Materials.
CHAPTER C2
STRENGTH OF COLUMNS WITH STABLE CROSS-SECTIONS
C2.1 Methods of Column Failure. Column Equations. The range AB in Fig. C2.1 is for a range
of Lip values of below 20 to 25, and repre-
In Chapter AlB, the theory of the elastic sents a range where failure is due to plastic
and inelastic instability of the column was crushing of the column. In other words, the
presented. The equations from Chapter AlB column is too short to buckle or bow under
for a pin end support condition are:- end load but crushes under the high stresses.
This column range of stresses is usually
For elastic primary failure, referred to as the block compression strength.
A column, however, may fail by local

Fc = - - - - - - - - - (1) buckling or crippling due to distortion of
the column cross-section in its own plane.
For inelastic primary failure, The horizontal dashed line in Fig. C2.1
represents the condition where the primary
column strength is limited by the local weak-
Fc = - - - - - - - - - - - - (2) ness. This line moves up or down according
to the value of the local weakness. The
determination of the column strength when
Where Fc =
compressive unit stress at failure is due to local weakness is covered
failure = piA stress. in another chapter.
E = Young's modulus
C2.2 Column End Restraint. Fixity Coefficients.
Et = tangent modulus Column Effective Length.
L = column length The column strength is influenced by the
p = radius of gyration of cross- end support restraint against rotation and by
section any lateral supports between the column ends.
The letter c is commonly used to indicate the
Fig. C2.1 shows a typical plot of Fc end fixity coefficient, and c = 1.0 for zero
versus Lip. If the column dimensions are such end restraint against rotation, which can be
as to cause it to fail in range CD in Fig. produced mechanically by a pin or ball and
C2.1, the primary failure is due to elastic socket end support fitting. Thus including
instability and equation (1) holds. This the end restraint effect equations (1) and (2)
range of Lip values is often referred to by can be written,
engineers as the long column range.
A B C D
, Fc - - - - - (3)
(Lip )B
I I I I
Local Crippling Limit Let L' = effective length of the column
which equals the length between inflection
points of the deflected column under load.
Then L' = Live (4)
Thus equation (3) can be written as,

o L'/p
- - - (5)
Fig. C2.1
The range BC represents the range of LiP If we let P = failing or critical load,
values Where failure is due to inelastic in- equation (5) can be written as equation (6)
stability of the column as a whole and equation
(2) applies. This range BC is often referred by realizing that p = FcA and p = vIlA.
to as the short column range. n 2 EI nBEt I
p = --
(L' )B
p =- - ------
(L' )2
(6)
C2.1
C2 2 STRENGTH OF COLUMNS WITH STABLE CROSS-SECTIONS
Fig. C2.2 shows the deflected column TABLE C2.1

curve under the load P for various end and
17-7 PH(TH1050) Stainless Steel Sheet
lateral support conditions. The effective
lengths L' and the end fixity coefficients Ftu = 180,000 E = 29 X 10 6
Fey = 160,000
are also listed.
L'/p=n~
6
20,000 29 x 10 119.56
30,000 29 x 10 6 97.62
40,000 29 x 10 6 84.54
50,000 29 x 10 6 75.62
60,000 29 x 10 6 69.03
70,000 29 x 10 6 63.91
80, 000 29 x 10 6 59.78
89, 600 29 x 10 6 56.49
96,000 27.55x 10 6 53. 19
102,000 26.10 x 10 6 50.22
I'f;- 107,400
112,200
117,000
24.65 x 10 6
23.20 x 10 6
21. 75 x 10 6
47.57
45.15
42.81
C=2.05 20.30 x 10 6
121, 500 40.58
125,800 18.85 x 10 6 38.43
Fig. C2.2 130,000 17.40x10 6 36.32
134,000 15.95 x 10 6 34.25
C2.3 Design Column Curves for Various Materials. 137,700 14.50 x 10 6 32.22
147,000 11.60x10 6 27.89
For routine design purposes it is con- 158, 100 8.70 x 10 6 23.29
venient to have column curves of allowable 167, 600 5.80 x 10 6 18.47
failing column stress Fc versus the effective 173,800 2.90 x 10 6 12.82
slenderness ratio I' /p. In equation (5) we
will assume values of Fc , then find the This equation is plotted in Fig. C2.16.
tangent modulus Et corresponding to this For a given material, n, F O • 7 and E must be
stress and then solve for the term L' /p. known. Then assuming values of F, we can
Table C2.1 shows the calculations for find corresponding values of Et/E from Fig.
17.7 PH (THI050) stainless steel sheet at C2.16. For values of E, F O • 7 and n refer to
room temperature. The results are then plotted Table Bl.l in Chapter Bl.
in Fig. C2.7 to give the column strength curve.
Similar data was calculated for the material C2.5 Non-Dimensional Column Curves.
under certain exposure time to different
elevated temperatures and the results are also Quite useful non-dimentional column
plotted in Fig. C2.7. Figs. C2.3 to C2.15 curves have been derived by Cozzone and
give column curves for other materials under Melcon (See Ref. 3).
various temperature conditions. Use of these
curves will be made in example problems later The Euler column equation is
in thiS chapter. The horizontal dashed line F = n2Et/(L'/p)2, which can be written,
is the compressive yield stress. Values
above these cut-off lines should be sUb- Et (L'/p)2
stantiated by tests. F - n2
C2.4 Tangent Modulus Et from Ramberg-Osgood
Equation.
The problem therefore resolves itself
into obtaining and expression for Et/F from
The basic Ramberg-Osgood relationship the non-dimensional relationship. To do this
multiply both sides of equation (7) by F o • 7 /F
for Et is given as follows: (See Ref. 1) and equate to B 2.
Et 1
- - - - - - - (7)
E - 3
n
(_F_)n-l. - - (8)
1 +"7 FO• 7
Et = tangent modulus of elasticity Fig. C2.17 shows a plot of this equation

as taken from Ref. 3, and shows F/F o • 7 versus
E = modulus of elasticity B for various values of n.
For definition of other terms see Article The shape of the knee of the stress-
Bl.12 of Chapter Bl. strain curve is given by the shape parameter
n and the abscissa B incorporates the
C2.3
120
. Fig. C2.3
Steel, AJSI 4130, 4140, 4340
70
I'" 0.188
Tell!l'. !lql(l6\Ire up to 1/2 SR.
Fill = $0000 (R. T.)
Fey ~ '10000 (R. T.)
60 i
50
50
r:: 40
40
30
30
20 ;.. " . . . . ..-~.' ... .
10
ViP
10 Figure C2. 4
20 30- 40 50· 90
Vip
Figure C2. 3
150 '.
30
20
10
o 20 30 4050· 60 70 80 90 100
L'If'
Figure C2. 5
F c is in ksi
C2.4 STRENGTH OF COLUMNS WITH STABLE CROSS-SECTIONS
180
170
160
150
140
130
120
100
r.."
90
80
70
60
50
40
30
VIP
~ W W ~ ~ 00 ro 00 00 ~
Figure C2. 7 L'/P Figure C2. 8
70
60
50
40
r.."
r.. "
20
10
L'lp
Figure C2. 10
L'/P
Figure C2. 9
Fe is in ksi
C2.5
50 90
L'/p
Figure C2.12
Figure C2.1l
Fe is in kai
L'/('
Figure C2.13
C2.6 STRENGTH OF COLUMNS WITH STABLE CROSS-SECTION
F c is in ksi
~1.00
....
If
g:
o
n
.... .90
I"" "'"
""'.
---.... ...........
-
~ --........ ~
-......... r--:: r-...... -.......::: ~~ ~
.........
~~
~
1'-.." ~~ .10
~
\ ·20
\. -n- O
.. n F-
1. 00
.90
.........
"," [\.~
~
N
..3 .80
~
--........
""'.
""
.......... ~ r-....
"-
i"--. r-....
"'" '"
'"
'"'
~ "~" "'- r\!\
~\
\\ 9 1\ 1\1\
,\\~ \
-\ 1\\\ \
.80
~.O. >-.....
""
~~ K 3 1'(\ ,\I\~~
2~ ~ ~. ",'\ \\ ~\\ 1\
.70
.60
~ ~l:\~ \ .60
~
~ H--
Et
E .50
~~
I--
--'---
-........
I---
- -"----
.50
\~ I'-. r----- I-- r-- ~' 2.0
\ ~ I'-- r-.... ---.... -

.40
~ I'-..
'-..., ............. --- -.... ---.... r--
- 2:.5
.40
~ .~
...........
r----- f--. -
.30 Et
E =1
1
3 ( F )n- ~
~~ ",," "'- "" --........
~
....... -- 3- r--
.30
.20
I
~ Fig. C2.16 Dimensionless tangent modulus stress curves.
I
+ifn F o • 7
I I I I I
~~
\\
~~ "'"'
""-I'--
~~ "'-- I'--
""-
""""
r-----
-
-........ ~
- ! r-
.20
---- --
'6 .10
.10
\""- R E=::: ~, r-- r:-r ---.... r--
o . 10 . 20 . 30 . 40 . 50 . 60 . 70 . 80 .90
-.1. 00
~ ""
1. 10
~
lit>
1. 20
11
1. 30 1. 40 1. 50
F/F o • 7 Fig. C2. 16 (Ref. NACA T. N.902) 0 = s/s~
0.3
O. 2 ~~...:._~. --.
0.1·~
1.8 2.0 2.2 2.4

Fig. C2.17 B = *JF~. 7 (L'/p)
...
()
-J
C2.8 STR_ENGTH QXCOI.lJ~tNS WITH §,!,_~BLE CR.OE!S-SECTIONS
particular properties of the material EO.7/F. Thus in Fig. C2.18, the deflection of
point (0) away from tangent at midpoint c
Inserting value of F - nBEt/(L'/p)B in equals unity in our assumed conditions and it
equation (8), also equals the first moment of the area of
the M/EI diagram between (0) and (C) about (0).
Et ~
( n B Et/(L'!p)B)
(F-E-j=
0.7'\ BB (Fig. C2.19).
The value of the ordinate for M/EI diagram
P 1l X
_ 1 at any point x from 0 is EI sin L
or B --
n
- - - - - - (9)
The total area under the curve is,
The use of the curves in Fig. C2.17 will area =-EIP 10L sin -x
1l dx =-P [-L-n cos -nx] L
be illustrated later in the example problem L EI L 0
solutions.
C2.6 strength of Columns with Variable Cross-Section
or Moment of Inertia.
= :1 - {[~ (- l)J [- (l)J}* = ~ (~ + *)
To save weight in a built up column or
hence area =nEI
2PL and half area =nEI
PL
forged column, the member is tapered or is
made with a non-uniform cross-sect1on. To The center of gravity of the half area is
find the ultimate strength of such columns,
it is usually necessary to use a trial and Ai =1 x da
error method. The general method of solution
involving a consideration of column deflection P L/2
will be illustrated for a case of a long ( PL) - =-1 x sin [- x dx,
nEI x EI
column with uniform cross-section. 0
Fig. C2.18 shows a pin ended column in a L_ L/2

deflected neutral equilibrium position when - x =1 x sin ~ x dx
carrying the ultimate or critical load P. 1l 0
Assume that the shape of the deflected column
follows a sine curve relationship with the Integrating this simple expression and solving
deflection at midpoint equal to unity (see for x we obtain:
Fig. C2.18).
x = L/n
The equation of the deflected column
curve is y = sin nx I f P is the end load, the
Taking moments about point (0) of the M/EI
L diagram between 0 and C about 0:
bending moment at any point = M = Py = P
n x
sin L · Yo = 1 = PL L = PL B hence, P = nBEI
nEI n nBEI LB
By the well known "moment area" principle
(see Chapter A7; Art. A7.14), the deflection which is the Euler equation, and thus the
of a point (A) on the elastic curve away from assumed sine curve was the proper one for the
a tangent to elastic curve (B) equals the first deflected elastic curve of the column.
moment of the M/EI diagram between (A) and (B)
about (A).
Tak·i·ng moments:: about.. 0 of the areabe·tween" 0

PL and C;
Fig. C2.18 'll'EI Fig. C2.19 _ PL 5 _ 5PL B
Yo =
1 - 3EI 16 L - 48EI hence,
C2.9
48EI 9.6 EI (9) Solve for load P by writing an expression

P=5L"= L" which compares with P = for the deflection at the center point
9.9EI/L" of the Euler which equals unity. This is done by
equation or an error of using the moment area principle as was
3 percent. done in the previous example problem in-
volving a column with uniform section.
In the above outlined procedure, E has

been assumed constant or, in other words, the
column failure is elastic or failing stresses
p ~o C • P are below the proportional limit stress of the
~..~
material. The practical problem usually
involves a slenderness ratio where failure is
due to inelastic bending and thus E is not
constant. For this case, a trial and error
Pl. C CI '6 C;O"l6TANT
method of solution in necessary using the
.3Bl tangent modulus of elasticity which varies
Fig. C2.20 with stress in the inelastic stress range.
We can now apply the same procedure to C2.7 Design Column Curves for Columns with Non-
Uniform Cross-Section.
a column with non-uniform cross-section. The
steps in this procedure for a column Figs. C2.21 and C2.22 give curves for
symmetrical about the center point are as rapid solution of two types of stepped columns.
follows:- Figs. C2.23 and C2.24 gives curves for the
rapid solution of two forms of tapered columns.
(1) Assume a sine curve for the deflected Use of these curves will be illustrated later
column with unit deflection at center in this chapter.
point.
C2.8 Column Fixity Coefficients c for Use with Columns
(2) Plot a moment of inertia (I) curve for with Elastic Side Restraints and Known End
column cross-section. Bending Restraint.
(3) Find the bending moment curve due to end Figs. C2.25 and C2.26 give curves for
load P times the lateral deflection. finding fiXity coefficient c for columns with
one and two elastic lateral restraints and
(4) Divide these moment values by the EI Fig. C2.27 gives curves for finding c when
values to obtain M/EI curve. The modulus restraining moments at column ends are known.
of elasticity E is considered constant. Use of these various curves will be
illustrated later.
(5) Find the deflected column curve due to
this M/EI loading. C2.9 Selection of Materials for Elevated Temperature
Conditions.
(6) Compare the shape of the derived column
deflection curve with that originally Light weight is an important requirement
assumed as a sine curve. This can be in aerospace structural design. For columns
done by multiplying the computed de- that fail in the inelastic range of stresses,
flections by a factor that makes the a comparison of the Fcy/W ratio of materials
center deflection equal to unity. Since gives a fairly good picture of the efficiency
the assumed sine curve is not the true of compression members when subjected to
column deflection curve, the computed elevated temperature conditions. In this
deflection will differ somewhat from the ratio Fcy is the yield stress at the particular
sine curve. temperature and w is the weight per cu. inch
of the material. Fig. C2.28 shows a plot of
(7) With the computed deflection curve, Fc /w for temperature ranges up to 600 0 F.
modified to give unity at center point, Wirh 1/2 hour time exposure for several im-
repeat steps 3, 4, 5 and 6. The results portant aerospace materials.
this time will show derived deflection
curve still closer to the assumed C2.10 Example Problems.
deflection curve.
PROBLEM 1.
(8) To obtain the desired accuracy, the pro-
cedure in step (7) will usually have to Fig. C2.29 shows a forged (I) section
be repeated again. member 30 inches long, which is to be used as
CRITICAL LOADSJNON-UNIFORM COLUMNS
Fig. C2.21
CRITICAL LOADS-NON-UNIFORM COLUMNS
Fig. C2.22 ....
o
Sin~le Stepped - Pin Ended Double Stepped - Pin Ended
10 r;:;::-- _
10~ __
9
&/L
8
·9
8
7
-------~..:.-- ::
• 5 • 5
·7
.8
.6
2 2
o
6 7 12 13 1~ 15 1b 3 6 7 8 91OU12~
(U),/ (U)2
00
CRITICAL LOADS-NON-UNIFORM COLUMNS CRITICAL LOADS-NON-UNIFORM COLUMNS
>-:l
Constant Thickness - Taper in Plan-Form Solid Prisms Tapering in Wi4th and Thickness
Solid Cylinders Tapering to Cones Fig. C2.24 >
tll
t"'
t>.l
(j
::tl
o
00
00
I
00
t>.l
(j
>-:l
....
o
z
00
•
2 7 8 9 w U 12 ~ " ~ ~
(U),j(ul?,
Fig. C2.23
Reference - N. American Avaiation Structures Manual

SIMPLY SUPPORTED COLUMNS WITH ONE ELASTIC RESTRAINT
4.0 FIXITY CONSTANT FOR COLUMN
3.8 WITH END SUPPORT HAVING A
3.6 140 KNOWN BENDING RESTRAINT
3.4 K' = (lbs/in) 4.0
3.2 120
3.0 3.8
100 For l..L ,~=, C = 4.0
2.8
2.6
3.6 '"
Both Ends Equally Restrained
80
C 2.4 q
2.2 3.4
60
2.0
1.8 3.2
40
1.6
1.4 20 3.0
1.2 ~~I! 2.8 C FiXity Constant

1.0 0.1 0.2 0.3 0.4 0.5 C
Fig. C2.25 x/L E Modulus of Elasticity
2.6
I Moment of Inertia
SIMPLY SUPPORTED COLUMNS WITH TWO ELASTIC RESTRAINTS
9 2.4 L Length of Column
l..L = Bending Restraint
8 Coefficient - Spring
2.2 Constant (in. - it per Radian)
7 2.0
For j..L = 9Q, C = 2.05
6 1.8
C One End Restrained
5 1.6
4 --=-- 1.4
3' 1.2
2 1.0
1
o 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 o 10 20 30 40 50 60 70 80 90 100 110
Fig. C2.26 xh Ref. Convair structure Manual l..LL ....
"EI
Fig. C2. 27 ....
Fig. C2.28 Solution:

(1) AISIsteel, F tu =180,000
Since the column may fail by bending about
(2) 17-7PH. stainless steel, Ftu=210,000 either the X or Y axes, the column strength for
(3) 7075-76 Alum. Alloy bending about each of these axes will be calcu-
(4) A2:31B Magnesium Alloy lated. Since the column strength is a function
of the radius of gyration of the cross-section,
(5) 6AL-4V Titanium Alloy the first step in the solution will be the cal-
culation of Ix and I y , from which p x and Py can
be found.
Calculating Ix: In Fig. C2.30 the section will

be first considered a solid rectangle 2.5 x
2.75 and then the properties of portions (1)
and (2) will be subtracted.
1
Ix (rectangle) =I2x 2.5 X 2.75 3 = 4.32
Portions (1) and (2)
1 x 1.5 x 1.25 3 - 4 ( .625 x .25 x 1.292 " ) =
Ix = -12
-1.29
(1 0 of (2) about its x centroidal axis is
negligible)
Ix = 4.32 - 1.29 = 3.03 in. 4
100 200 :100 400

TEMP. of
500 600
Px = j'!i.. , Area A = 2.5 x 2.75 - 2 x .75 x 1. 25 -
A
4x .25 x .625) = 4.375 sq.
in.
a compression member. Find the ultimate
strength of the member if made from the = ,; 3.03/4.375 = .83 in.
following materials and subjected to the given Px
temperature and time conditions. Calculation of I y :
Case 1. Material 7079-T6 Alum. Alloy hand 1
forging and room temperature. I y (solid) = 12 x 2.75 x 2.5 3 = 3.58
Case 2. Same as Case 1, but subjected 1/2 Portion (1) = -(1.25x .75x .875 2 )2 -(1.25x
hour to a temperature of 300 oF. .75 3 /12)2 =- 1.52
Case 3. Same as Case 2, but for 600 oF. Portion (2) = -(.25x .625x .833")4-4(.25x
1.25 3 /36) =- .488
Case 4. Material 17-4 PH stainless steel,
hand forging at room temperature. Iy =3.58 - 1.52 - .488 =1.58 in. 4
Fig. C2.29 ¥- -- -->=0

------
Py = J 1.58/4.375 = .60"
Column strength is considerably influenced
~~)
by the end restraint conditions. For failure
by bending about the x-x aXiS, the end restraint
against rotation is zero as the single fitting
~ bolt has an axis parallel to the x-x axis and
thus c the fiXity coefficient is 1. For
failure by bending about the y-y axis we have
end restraint which will depend on the rigidity
of the bolt and the adjacent fitting and
structure. For this example problem, this
restraint will be such as to make the end
fiXity coefficient c = 1.5.
se;,.A-A
C2.13
For failure about x-x axis, Thus we make use of the curves in Fig.
C2.17.
L' = L/v'C= 30/v'l= 30, L'/px = 30/ .83
Case 1. Material 7079-T6 Alum. Alloy forging.
= 36 Table Bl.l of Chapter Bl summarizes certain
material properties. The properties needed
For failure about y-y axis, to use Fig. C2.17 are the shape factor n, the
moduls Ec and the stress Fa.?' Referring to
L' = 30/.,[1:5= 24.6, L'/Py = 24.6/.60 Table Bl.l, we find that n = 26, Ec =
= 41 10,500,000 and Fa.? = 59,500.
Therefore failure is critical for bending The horizontal scale in Fig. C2.17 in-
about y-y axis, with L'/p = 41. volves the parameter,
Case 1. The material is 7079-T6 Alum. Alloy

hand forging. Fig. C2.14 gives the failing B= l
n
~ Fa.?
E
(L' /p)
stress Fc for this material plotted against c
the L'/p ratio. Thus using L'/p = 41 and the
room temperature curve, we read Fc =
50500 psi. Substituting:-
Thus the failing load if P = FcA = 50500 x
4.375 = 220,000 lbs. 1
B -_ n j 59,500
10,500,000
(41) 1 01
=.
Case 2. Using the 300 0 F curve in Fig. C2.14
for the same L'/p value, we read Fc = 40,400, Using Fig. C2.17 with 1.01 on bottom
and thus P = 40,400 x 4.375 = 177,000. scale and projecting vertically upward to
n = 26 curve and then horizontal to scale at
Case 3. Using the 600 0 F curve, Fc reads 6100 left side of chart we read Fc/F a .? = .842.
and thus P = 6100 x 4.375 = 26700 lbs. Thus
subjecting this member to a temperature of Then Fc = 59,500 x .842 = 50,100, as
600 0 F for 1/2 hour reduces its strength from compared to 50,500 in the previous solution
220,000 to 26,700 lbs., which means that using Fig. C2.14.
Alum. Alloy is a poor material for carrying
loads under such temperatures since the Case 2. From Table Bl.l for this material
reduction in strength is quite large. subjected to a temperature of 600 0 F for 1/2
hour, we find n = 29, Fc = 9,400,000 and
Case 4. Material 17-4 PH stainless steel Fa • 7 = 46,500.
forging. Fig. C2.8 gives the column curves
for this material. For L' /p = 41 and using Then B = l
n
vi 9,400,000
46,500 (41) = .917
the room temperature curve we read Fc =
135,200 and thus P = 135,200 x 4.375 =
591,000 lbs. From Fig. C2.17 for B = .917 and n = 29,
we read Fc/F a .? =
.88, thus Fc =
46,500 x .88
C2.10 Solution Without Using Column Curves. = 40,900 as compared to 40,400 in the previous
solution.
When primary bending failure occurs at
stresses above the proportional limit stress, EXAMPLE PROBLEM 2.
the failing stress is given by equation (5)
which is, Fig. C2.31 shows an extruded (I) section.
A member composed of this section is 32 inches
long. The member is
Since Et is the tangent modulus of
braced laterally in the
x direction, thus
I---lf--I
elasticity, it varies with Fc ' and thus the
relation of Et to Fc must be known before the
equation can be solved. To plot column curves
failure will occur by
bending about x-x axis.
The member is pin ended
= 12" x---
t - --x
1 - :_1
for all materials in their many manufactured and thus c 1. The
forms plus the various temperature conditions material is 7075-T6 I 8
would require several hurldred individual Extrusion. The problem r--~I
column charts. The use of such curves can be is to find the failing I y
avoided if we know several values or parameters stress Fc under room
regarding the material as presented by temperature conditions. Fig. C2.31
Ramsburg and Osgood and expanded by Cozzone
and Melcon (see Arts. C2.4 and C2.5) for use This (I) section corresponds to Section
in column design. 15 in Table A3.15 in Chapter A3. Reference
to this table gives, is due to the fact that the stress existing
under a L' /p value of 51.7 is near the pro-
A = .594 sq. in. Px = .618 portional limit stress or Et is not much
different than Ec , the elastic modulus.
L' =L, sin c = 1, L'/px = 32/.618 = 51.7
To illustrate a situation where the 7075
Fig. C2.11 gives the column curves for material becomes more efficient in comparison
this material. For L' /p =
51.7 and room to the 2014 alloy, let us assume that our
temperature we read Fc = 38,500 psi. member has a rigid connection at its end which
will develop an end restraint equivalent to a
Solution by using Fig. C2.17, fixity coefficient c 2. =
From Table Bl.l, n =8.8, Ec = 7,800,000 Then L' = 32/y'2:= 22.6 and L'/p =
and FO • 7 = 29,000. 22.6/.618 =
36.7.
B =~ yl7,~~O~ggo (51.7) =1.00 For the 7075 material from Fig. C2.11, Fc =
58,300
From Fig. C2.17 we find Fc /F o • 7 = .74 For the 2014 material, we use Fig. C2.17
Then Fc =
.74 x 29,000 =
21,450 psi
B =~n vi 10,700,000
53,000 (36.7) = .823.
Consider the member is SUbjected to a
temperature of 450 0 F for 1/2 hour.
From Fig. C2.17 for B .823 and n= =
From Fig. C2.11, Fc = 21400 psi 18.5, we read Fc /F o • 7 =
.87, when Fc .87 x =
53,000 =
46,100 as compared to 58,300 for the
Using Fig. C2.17:- 7075 material, thus 7075 material would permit
lighter weight of required structural material.
From Table Bl.l, n = 8.8, Ec = 7,800,000 The student should realize that if the
and FO • 7 = 29,000.
stress range is such as to make Et =
Ec ' then
the bending failure is elastic instead of
B- nj
_ 1 29,000
7,800,000 (51.7) = 1.00 inelastic and equation (5), using Young's
modulus of elasticity Ec ' can be solved
directly without resort to column curves or
From Fig. C2.17 we find Fc /F o • 7 = .74 a consideration of Et, since Et is equal to
Then Fc = .74 x 29,000 = 21,450 psi Ec '
A very common aluminum alloy in aircraft The student should realize that equation
construction is 2014-T6 extrusions. Let it (5) is for strength under primary column
be required to determine the allowable stress failure due to bending as a whole and not due
Fc for our member when made of this material. to local buckling or crippling of the member
or by tWisting failure. The subject of
Since we have not presented column curves column design when local failure is involved
for this material, we will use Fig. C2.17. is covered in a later chapter.
From Table Bl.l, for our material, we In example problem 2, we have assumed
that local crippling is not critical, which
find n = 18.5, Ec =
10,700,000 and F O • 7 = calculation will show is true as explained
53,000
and covered in a later chapter.
- 1 / 53,000 (51 7) 1.10
Then B - nv10,700,000 . = C2. 12 Strength of Stepped Column.
The use of curves in Fig. C2.22 will be

From Fig. Cl.17 for B 1.16 and n= = 18.5, illustrated by the solution for the strength
we read Fc /F o ' 7 =
.71, hence Fc .71 x = of two stepped columns in order to illustrate
53,000 =
37,600.
both elastic and inelastic failure of such
The result shows that the 2014-T6 material columns.
gave a failing stress of 37,600 as compared to
Case 1. Elastic failure.
38,900 for the 7075-T6 material which has a
Fcy of 70,000 as compared to Fcy = 53,000 for Fig. C2.32 shows a double stepped pin
the 2014-T6 material. The reason for the
7075 material not showing much higher column ended column. The member is machined from a
1 inch diameter extruded rod made from
failing stress Fc over that for the 2014 alloy
C2.15
7075-T6 material. The problem is to find the This is a relatively short column so the
maximum compressive load this member will failing stress should fall in the inelastic
carry. range where E is not constant, therefore the
solution is a trial and error procedure. We
Portion 2 Portion 1 will base our first guess or trial on an
~3/4" Dia. j t1"Dia. j PJrtion 2 average L/p value.
, t
~b=15'~_a=30"_I~b=15"
: t ; =j' p for portion 1 is 0.25 inches
( L=60" ) p for portion 2 is 0.1875

Fig. C2.32 Average p = (6 x .25 + 6 x 0.1875)/12
= 0.22
PORTION 1 PORTION 2
Then L/p = 12/0.22 = 54.5, use 55.
AJ. = .7854 in. 2
A2 = .4418
I J. = .0491 in.
4
12 =
.0155 Fig. C2.11 is a column curve for 7075-T6
Ec = 10,500,000 Ec = 10,500,000 Alum. Alloy extruded material. With L/p =
55, we read allowable stress Fc = 33,500 psi.
From Fig. C2.22) Pcr = B(EIJ.) • This is Therefore
L2
the Euler equation for failure under elastic P = Fc A = 33,500 x 0.7854 = 26,300 lb.
bending. If the ratio aiL equals 1 or a
uniform section, B becomes n 2 or 10 as shown f 1 = 33,500 and f 2 = 26,300/.4418
in Fig. C2.22. The curves in Fig. C2.22 = 59,500 0
apply only to elastic failure. Since the

member in Fig. C2.32 is rather slender we The stress f in portion 2 is above the
will assume the failure is elastic and then proportional limit stress so a plasticity
check this assumption. correction must be made in using the curves
in Fig. C2.22.
EIJ. 10,500,000 x .0491
EI 2 = 10,500,000 x .0155 =3.17, aiL = 30/60
= 0.5
Referring to Table Bl.l in Chapter Bl,
we find the folloWing values for 7075-T6
extrusions:- n =
16.6, F O • 7 72,000, =
From Fig. C2.22 for aiL = 0.5 and EI 1 /EI 2 Ec =10,500,000.
= 3.17 we read B = 7.0
The tangent modulus Et will be found for
Whence, Pcr = BEll: = 7 x 10,500,000 x .0491 = the stresses fJ. and f 2 •
L2 60 2
For Portion 1, f 1 /F o • 7 =33,500/72,000
1000 lb. = .465
The stresses in each portion are, Referring to Fig. C2.16 and using 0.465
and n = 16.6, we read Et/E =
1.0, thus Et =
f1 = 1000/0.7854 = 1280 psi E and thus no plasticity correction for
Portion 1.
f2 = 1000/.4418 = 2270 psi
= 59,500/72,000
For Portion 2, f 2 /F o • 7
These compressive stresses are below the = .826
proportional limit stress of the material so
Ec is constant and our solution is correct. From Fig. C2.16, we obtain Et/E = .675
whence, Et = .675 x 10,500,000 = 7,090,000.
Case 2. Inelastic Failure.
EI 1 10,500,000 x .0491
The column has been shortened to the
dimensions as shown in Fig. C2.33. The EI 2
= 7,090,000 x .0155 =4.7
diameters and material remain the same as in From Fig. C2.22 for aiL = .5, we obtain
Case 1.
Portion 1 Portion 2
B =5.6.
.J j
: : I
Then Pcr =B EI 1 = 5.6 x 10,500,000 x .0491
r-b=3" --¥-- a=6" ------¥-b=3""} L 2
144
I~· L=12" ----~ = 20,000 lb.
Fig. C2.33
Our guessed strength was 26,300 lb. OUr C 1-1/4 -.058 A 1-1/4 -.058 B 1-1/4 -.058 D
guessed strength and calculated strength must
be the same so we must try again. T "
~'O
~
..,.
a>
0
I
Trial 2. Assume a critical load P = 00

'-
.....
23500 lb.
.....I
f 1 = 23500/.7854 = 29900 E
I~ 30" ~I~
G
30" ~I
H
f 2 = 23500/.4418 = 53100 Fig. C2.34
Portion 1. f /F o • 7 = 29900/72000 = .415 The member AB is welded to three adjacent

tubes at joints (A) and (B). Since these
From Fig. C2.16 for n = 16.6, we read tubes are the same at joints (A) and (B), the
Et/E = 1.0. fixity at the ends (A) and (B) of member AB
is the same.
Portion 2. = 53100/72000 = .738
f 2 /F o • 7
Referring to Fig. C2.27, the term ~ is
From Fig. C2.16, Et/E = .90, whence defined as the bending restraint coefficient-
Et = .90 x 10,500,000 = 9,450,000. spring constant expressed as inch pounds per
radian. In Fig. A,
10,500,000 x .0491 the moment M re-
= = 3.52. quired to rotate u"-+
lV\. A EI (constant)
__ B
9 , 450 , 000 x .0155 end (A) through 1
~=------=:":"':'---
radian when far end

t~
From Fig. C2.22 for a/L = .5, we read
B =6.62. is fixed is 4EI/L.
For derivation of
Fig. A
this value refer to Art. All.4 of Chapter

The Per
= 6.62xlO,500,000x .0491 = All. If the far end (B) is pinned in
144 =
(Fig. A), a moment M 3EI/L will rotate end
(A) through one radian. To be slightly
23,650 lb. conservative, we will assume the far ends of
members coming into joints (A) and (B) as
This practically checks the assumed value, pinned. Thus ~ =
3EI/L. The sum of ~ =
thus the answer is between 23,500 and 23,650 3EI/L will be computed for the 3 members
and if further accuracy is desired another which form the support of member AB at end
trial should be carried through. (A) •
The other types of columns with non- Member AC:- I =
.03867, I/L .001289 =
uniform cross-sections as shown in Figs.
C2.21, C2.23 and C2.24 are solved in a similar Member AE:- I =
.02775, IlL .00071 =
Member AF:- I = .02402, I/L = .000962
manner. These charts are to be used only with
pin ended columns. The end fixity coefficient
c for tapered columns is not the same as for ~ =Z 3EI/L
uniform section columns.
= 3( .001289 + .00071 + .000962) 29,000,000
C2. 13 Column Strength With Known End Restraining
Moment.
~ = 258,000
In Fig. C2.27 we need term ~ L/EI. The
.Fig. C2.27 shows curves for finding the L/EI refers to member AB. Thus ~ L/EI =
end fixity coefficient c for two conditions (258,000 x 30 )/29 ,000,000 x .0367 = 7.28.
of known end bending restraint.
We use the upper curve in Fig. C2.27
To illustrate the use of these curves, since restraint at both ends of member AB
a simple problem will be solved. is the same. Thus for ~ L/EI 7.28, we =
Fig. C2.34 shows a 3-bay welded steel
read end fixity coefficient c 2.58. =
tubular truss. The problem is to determine Then L' = L/v'C = 30!V 2.58 = 18.6.
the allowable compressive stress for member
AB. This strength is influenced by the fixity p for member = .422 inches.
existing at ends A and B. The diameter and
wall thickness of each tUbe in the truss is
shown on the figure. The material is AIBI L'/p = 18.6/.422 = 44.0
steel, Ftu =
90,000, Fty 70,000, E = = From column curve in Fig. C2.3, we read
29,000,000.
C2.17
allowable failing stress to be Fc = 55,200 C2.15 Problems.

psi.
(1) 6061-T6 Aluminum Alloy Sheet, heat-treated
If the far ends of the connecting members and aged has the following properties:
were assumed fixed instead of pinned, then
~ =
4EI/L, or we can multiply previous value (a) Under room temperature:- F O • 7 =
of 7.28 by 4/3, which gives 9.7 which, used in 35,000 psi, Ec =
10,100,000 psi, and
Fig. C2.27, gives c = 2.80. n = 31
L' /p =
30/v'"2:8 x .422 = 42.5. Then from (b) For 1/2 hr.exposure at 300 oF:- Fo 7 =
Fig. C2.3, Fc = 56,600 psi. Since the far 29,000, Ec = 9,500,000 and n 26: =
ends are less than fixed, the assumption that
far ends are pinned gives fairly accurate For the above two cases (a) and (b),
results. determine Et (tangent modulus values) from Fig.
C2.16 and then calculate and plot column curves
In a truss structure all members are for these 2 material conditions.
carrying axial loads and axial loads affect
the ability of members to resist rotation of (2) Fig. C2.36 shows
their ends. Art. All.12 of Chapter All the cross-section
explains how to take account of the effect of a compression
of axial load upon the stiffness of a member member. Calculate
as required in calculating the end restraint the failing com-
coefficient ~. pressive load under
the following
C2.14 Columns With Elastic Lateral Supports. cases:-
Figs. C2.25 and C2.26 prOVide curves for =
Case 1. L 25 inches.
finding the end fixity coefficient c to take Material AISI Steel 4140, I-z.
yl
care of elastic lateral supports at points Ftu =
180,000. Take end
midway between the column ends. fixity coefficient c 1 = Fig. C2.36
for bending about x-x
To illustrate the use of these Charts, a axis and 1.5 about axis y-y.
round bar 0.5 inches in diameter and 24 inches
long is braced laterally as shown in Fig. (3) Same as Problem (2) but member is sub-
C2.35. The bar is made of jected to a temperature of 8500 F for
AISI Steel, heat treated to 1/2 hour.
Ftu =
125,000. The spring
constant for the lateral (4) Two extruded channel sections identical
support is 775 Ibs. per to Section No. 50 in Table A3.11 in
inch. ~ L=24 Chapter A3, are riveted back to back
and used as a column member. If member
Moment of Inertia of 1/2"
rod = .003068,
Radius of Gyration =
±1 is 26 inches long and end fixity is
=
C 1 and material is 7075-T6 extrusion,
what is the failing compressive load.
.125 inches. Fig. C2.35 If member is fastened rigidly to adjacent
}("L 3 775 x 24 3
structure which provides a fixity c 2, =
what will be the failing load.
q = ~ = 29,000,000 x .003068 = 120
(5) Consider the column in problem (4) is
=
From Fig. C2.25 for x/L 10/24 = .416 made from 20J4-T6 Aluminum Alloy extrusion.
=
and q 120, we find c 2.92. = Find failing load.
Then L' = L/ve= 24/..) 2.92 = 14.08 (6 ) The pin ended single stepped column as
shown in Fig. C2.37 is made of AISI-4130
L'/p = 14.08/0.125 = 113 normalized steel, Ftu =
90,000, FCY =
70,000. Determine the maximum compressive
Fc =
2
= n x 29,000,000 = 22,500 load member will carry.
(113)2 psi I" Sq. Bar 3/4" Square Bar
r -----*"'----'----.tL----,
If the stress is above the proportional I I
limit stress fo~ the material, then the trial a = 21" ~_1'__9_"=4
and error approach must be used as illustrated L = 30" --+l
in the problem dealing With a tapered column. Fig. C2.37
(7) Same as Problem (5) but member is exposed (11) The cylindrical tapered member in Fig.
1/2 hour to a temperature of 500 oF. C2.39 is used as a compression member.
If member is made from AIBI Steel 4140,
(8) Same as Problem (5) but change dimension Ftu = 125,000, what is the failing load.
(a) to 10 inches, and L to 14.28 inches.
(9) Find the failing compressive load for til" Dia. Rod .1/2" Dia. Rod
the doubly stepped column in Fig. C2.38 c:::==: ~T
i f member is made from 7079-T6 hand
forging. t b
=9 ~a;~~~b=91
1-3/8 Dia. Rod Fig. C2.39

I 1-1~8 Dia. Rod
I (12) Same as Problem (7) but change dimensions
/:-=b=12 -+--a=12~b=12-1 to a = 6", b = 4.0", L = 14 inches.
~---- L=36 ----~
References:
Fig. C2.38 (1) NACA Technical Note 902.
(2) Non-dimensional Buckling Curves, by
(10) Same as Problem (7) but change dimensions Cozzone ~ Melcon, Jr. of Aero. Sciences,
a = 6", b =
4", L = 14". October, 1946.
(3) Chart from Lockheed Aircraft Structures
Manual.
CHAPTER C3
YIELD AND ULTIMATE STRENGTH IN BENDING
C3. 1 Introduction. compression. In this example solution, we will

find the internal resisting moment when we
Members subjected to bending alone or in limit the unit strain at the extreme edge on
combination with axial and torsional loads are the compressive side of the beam section to
qUite comnon in flight vehicle structures. 0.010. Now plane sections remain plane after
The limit design loads on a structural member bending in both elastic and inelastic stress
must be carried without permanent distortion conditions when member is in pure bending.
~nd the ultimate design loads must be carried We will guess the neutral axis as located
without rupture or failure. The well known 0.0375 inches above the centroidal axis as
bending stress equation f b = Mc/I, assumes a shown in Fig. C3.1b. Having assumed the
linear variation of stress with strain or, in maximu~ unit strain as .010, we can draw the
other words, the equation holds for stresses strain diagram of Fig. C3.1b. We now divide
below the proportional limit stress or, in the cross-section in Fig. C3.1a into 20
general, the elastic range. Failure of a narrow horizontal strips. Having the strain
member in bending, unless there is local curve in Fig. C3.1b, we can find the unit
weakness, does not occur at stresses in the
elastic range but occurs at stresses in the
inelastic range. Since the ultimate strength
of a member is needed to compare against the
ultimate design load to be carried, a theory
or procedure is n3cessary which will accurately
determine the ultimate and yield strength of a
member in bending.
C3.2 Basic Approach to Finding the Bending Strength
of Members.
The problem is to determine the internal

resisting moment of a beam section when sub-
jected to stresses which fall in the inelastic
range of stresses. This stress can be taken
as the ultimate tensile or compressive stress
of the material or limited to some stress or
deformation in the inelastic range. To obtain
the true internal resisting moment, we must
knryN how the normal tension and compressive
stress varies over the cross-section. The
stress-strain cu~ve for the material provides
the source for obtaining the true stress
picture. If a material has a different shape
in the tensile and compressive inelastic . 01 . 008 . 006 . 004 . 002 o . 002 . 004 . 006 . 008
zones, the neutral axes does not coincide with
the centroidal aXiS, thus adding some difficult
to an analysis method. The analysis procedure (0 )
Fig. C3.1
for determining the true internal resisting Te~SILf
TENSION 5TRl!5S
moment is best explained by an example
solution. ~,.
C3.3 Bending Strength of a Solid Round Bar.
Fig. C3.1a shows the cross-section of a

round solid bar made of aluninum alloy. The
stress-strain curve up to a unit strain of
.010 in. per inch is given in Fig. C3.2. Note
that the shape of the curve in the inelastic .010" CO~Pli!e:;SI"E
zone is not the same for both tension and COMPRESSION UNIT
STIU.55
:;T2AIN
C3.1
C3.3
The value of f o is determined by making Fig. C3.7 k Factor for Some Typical Shapes
the requirement that the internal moment of Shape K Shape K
the true stress system must equal the moment

of the assumed trapezoidal stress system -1- 1 -()-
1.2~
to
1.7
which results from the assumed stress-strain

curve as shown in Fig. C3.5. Fig. C3.6 shows -x- 4/3
1
-e- 1.7
the trapezoidal stress pattern drawn to a ~-fQl~- to

1.~ -0- 2
larger scale and showing only one half of the

sy,rrmetrial beam section. The trapezoidal -El- 1.~
stress pattern has been divided into a

rectangle (r) and a triangle (b) as shown in
N,t,:
the figure. k Villi"
"'tX-X"'1
.,pt, '0 biMin,
K • lOcI!
IIc' S.eU.. modlllvi
Let, Mb = total internal resisting moment. Qo Slotlc "',_IItobo\lt
1IlI1I1'lloIII.'.,.. 0110,..
" • •r n'lItr,l.sl,
mr = internal moment developed by
portion (r).
mb = internal moment developed by
portion (b).
Since fb varies linearly from zero to fb' \

the stress is elastic and thus the beam
equation holds, or J
V-
I-'
mb = fbllc for entire bea~ section. The 0.1 0.2 0.3 0.4 O.~ 0.6 0.7 0.8 0.' 1.0
stress variation on portion (r) is constant or m
rectangular, thus
Fig. C3.8 Section Factor K for 1,0, and C Sections (Ref. 2)
C3. 5 Design Curves for Finding Modulus of

Then mr = f o2Q for entire bea~ section Rupture (Fb).
But fb = fm - f o (from Fig. C3.6) The modulus of rupture Fb may be a yield

modulus, that is, in equation (3) the value of
c
ThUS, Mb = (fm-f o ) I + 2f oQ, or f m is equal to the yield stress of the material.
It may also be the Ultimate modulus of rupture,
Mbc in which case the value of f m in equation (3)
- f m + f o (2 Q - 1) - - - - - - - (1) equals the ultimate strength of the material.
I Ilc
The modulus of rupture may be limited to a stress
between the yield and Ultimate stress of the
Let k = Ilc
2Q - - - - (2) material because of local crippling or by
excessive distortion. Regardless of what value
k is a beam section shape factor. is used for f m in equation (3), the corres-
ponding value of f o must be kno'WTI before the
Let Fb = Mbc/I, then from equation (1) value of Fb can be determined. Figs. C3.9 to
C3.23 give strain curves for various material
Fb = f m + fo (k - 1) - - - - (3)
and the corresponding f o curve. The use of
these two curves permit the determination of
Fb is a fictitious Mc/I stress or the Fb if the k shape factor for the particular
modulus of rupture for a particular cross- beam section being considered is known. In
section at a given maximum stress level. deriving the values of f o ' the following
assumptions are made.
The values of k vary between 1 and 2.0.
If calculated value of k is greater than 2 (1) The stress-strain curve is assumed
use 2.0. Fig. C3.7 shows the value of the the same in tension and compression.
shape factor k for several typical shapes.
Fig. C3.8 shows curves for the rapid (2) The neutral axis is assumed to coin-
determination of the k factor for 3 common cide with the centroidal axis.
beam sections.
(3) During bending plane sections remain
plane.
(4) The cross-section is not subject to

local or torsional instability.
C3.4 YIELD AND ULTIMATE STRENGTH IN BENDING
(5) Beam-column, curvature and shear lag to give the value of f o . The value of f o from
effects are considered negligible. this chart operation gives f o = 29000. Then
from equation (3)
C3.6 General Accuracy of Method.
Fb
yield
= 65000 + 29000 (1.5 - 1) = 79500 psi.
(1) It is exact for a rectangular section
under pure bending with moment vector parallel Then yield bending moment = Myp = FbI/c
to a principal axis.
Thus Myp = 79500 x .0938 = 7460 in. lb.
(2) For double s~metric sections under
pure bending and moment vector parallel to a For finding the ultimate resisting
principal axis, the accuracy should be within bending, we use Ftu which is 75000 as the
5 percent. value of f m in equation (3). Again going to
Fig. C3.17 to stress of 75000 on stress-strain
(3) Single s~metric sections will vary curve and the vertically d01m to f o curve, we
from practically exact to definitely uncon- obtain f o = 70500.
servative (moment vector normal to axis of
sYilunetry) . Then Fb(Ult)= 75000 + 70500(1.5-1)= 110250 psi
(4) For sections subject to combined Then MUlt = FbI/c = 110250x.0938 = 10370 in. lb.
bending and axial load, the results will vary
from practically exact to conservative. Let us asswne that is is desired to limit
the strain in the extreme fiber to .03 inches
(5) For unsymnetrical bending, with and per inch. What would be the bending moment
without axial load, the results will vary from developed under this limitation.
practically exact to conservative.
From Fig. C3.17 for a unit strain of .03
C3.7 Example Problems in Finding Bending Strength. the corresponding stress from the stress
strain curve is 74700 and the f o stress is
KXAMPLE PROBLEM 1. 61200.
A rectangular bean section is 0.25 inches Then Fb = 74700 + 61200 (1.5-1) = 105300
wide and 1.5 inches deep. What yield and
ultimate bending moment will the section Then M =FbI/c = 105300 x .0938 = 9900 in. lb.
develop when made from 7075-T6 extruded
alLLminum alloy. EXAMPLE PROBLEM 2.
Solution: The modulus of bending stress is The symmetrical I beam section in Fig.
given by equation (3), (a) is subjected to an ultimate design pure
=
bending moment M 14000 in. lb. What is
Fb = f m + f o (k - 1) - - - - - - (3) the margin of safety if
the beam is made of
k = 2Q =2 x 0.75 x .25 x .375 x = .1408 magnesium forging AZ6lA
=
I/c 1
(12 ) .25 x 1. 5 3( 1/ . 75 )
.0938 which has Ftu 38000
and Fty = 22000.
k = 1.50 Ix = J:....
12
x 1.375 x 2 3
-
The value of k could also be found in Fig. 1 Fig. (a)

12 x 1.25 x 1.75 3
C3.7.
= .916 - 0558 = 0358.
Material is 7075-T6 aluminum alloy. From
Fig. C3.17, Ftu = 75000, Fty 65000. = I x / c = .358/1 = 0.358
To find the yield bending strength, the
value of f m in equation (3), the maximum
Q = 1.375 x .125 x .9375 + .875 x .125 x
.4375 = 0.209
stress permitted on the most remote fiber is
65000, the yield stress of the material. To _ 2Q _ 2 x .209
find f o , we go to Fig. C3.17 and find the k - I x / c - 0.358 1.17
point on the stress-strain curve that
corresponds to a stress of 65000. This point
is projected vertically downward to intersect k could have been from Fig. C3.8 When
the curve f o • This point is then projected using m = .125/1.375 = .091, and n .125/ =
to the stress scale at the edge of the chart 2.00 =
.062.
C3.5
From Fig. C3.19, for f m Ftu 38000 we = = bending strength was developed. From Fig.
find in projecting vertically downward to f o C3.19, the unit strain when stress is 38000
curve gives f o = 23700. Then sUbt. in is .035. Then since plane sections remain
equation (3) plane after bending, the unit strain at point
.50 inch from neutral axis is (.5/1) (.035) =
=38000 + 23700 (1.17 - 1) = 40770
Fb .0175. From Fig. C3.19, the stress existing
at this strain is f 31000 psi. A linear =
MUlt = FbI/c =40770 x .358 = 15000 in. lb. variation of stress as used in the flexural
equation would give half the maximum stress
Margin of Safety = M~lt _ 1 = 14000
15000 - 1 = _.-07 or 38000/2 =
19000 psi as against the true
11 stress of 31000.
Assume we desire the stress intensity at

a point 0.5 inches from neutral axis if full
50
60
~ 40
t" '" '
~ 30
~ 40 J'
10
20
0.02 0.04 0.06

E in/in
o
0.02 0.04 0.06 0.08 0.10 0.12
E In/In
60
60
.~ 40
A
~
~024"'I'4 Alum. Alloy Clad ~024-T81 Alum. Alloy Clad Sheet
Sheet &. Plate, Heat Treated s t < 0.Q64
t=0.25tQO.50 - 20
20 F = 62000
F tu =62000 tu
Fty" 5iOOO
Fty-= 40,000 Fig. C~.ll Fig. C3.l~
0.02 0.04 0.06 0.08 0.10 0.11

o 0.01 O.O~ 0.03 0.04 0.05
e in/in E in/in
60
- Stiress ~ strain. ,Curve t k:q 1
40
'"
~
_0
j3 2024-T3 Alum. Alloy $ll'lt & Plate

20 t:S 0.250
Ftu =64000
Fig. C3.13
Fty =42000
0.02 0.04 0.06 0.08 0.10 0.12 0.02 0.04 0.06 0.08 0.10 0.12
E In/In E In/In
Curves for finding Fb. Fb = f m + f o (k - 1).

70 b
'iii
~
S 40'
~Ei J'
20
0.01 0.02 0.03 0.04 0.05 0.055 0.01 0.02 0.03 0.04 0.05
E: in/in E in/in
'iii
~
~ 40
S ~
J s
J
E in/in
40
30
~ 20
""ji
10 AZ61A lIiIas:ne&ium AlluY Forgings (Lol>g)
Ftu = 38000
Fty .22000
Fig. C3.19
.005 . DID .015 .020 .025 .030 .035

E in/in 0.01 0.02 0.03 0.04 0.05 0.06
E in/in
StresB'-stram Curve k--;II: 1
160
~ 120
~ Fig. C3.22
17-4 PH Stainless Steel
Flu = 180,000 psi
_" 80 FlY = 165,000 psi
E = 27,5 x 10 6 psi
Elongation = 6%
40 Ileal Treated Bars & Forgings
o 0.01 0.02 0.03 0.04 0.05 0.008 0.012 0.016 0.020 0.024 0.028 0.032 0.035
e in/in E (inches/inch)
Curves for finding Fb ,

C3.7
portion on each side of the neutral axis of the

entire section.
Label the beam portion below the neutral

axis as (1) and that above by (2).
Portion 1. Fig. (d) shows

how the lower portion (1) --+I I+- O. 1"
is made a symmetrical t i1 I I
"T
section about neutral axes
by adding the dashed
portion. The internal
bending resistance will be
2 78 " I I 1. 39
_
i
I--{--N.A.
1. 39
-=l
-:.r-u
found for this entire
Fig. (d)
section in Fig. (d). One
o 0.01 0.02 0.03 0.04 0.05
half of this amount will
E: in/in
then be the true moment developed by portion (1).
EXAMPLE PROBLEM 3. Unsymmetrical Section.
I =~
12 bh = ~
12 x
3
0.1 X 2.78
3
= 0.178
Fig. (b) shows a tee beam section,
symmetrical about the vertical axis. If the I 1 /c 1 = 0.178/1.39 = .128
material is 17-4 PH stainless steel, what
ultimate bending moment will be developed if Q1 = 1.39 x 0.1 x .695 = .0965
bottom portion is the tension flange.
---:l. 1~1-1/2~ E2=·0154

_ _
k1 = 2n~1 /1 1/ C1 = 0.193
0.128 = 1 •5
1
Lo. 1"
...No" - - ----i------
to. 610
From Fig. C3.22, Ftu = 180000 which equals
y = 1. 39 f o from curve = 156000

____
1 ___l
0.1"-+1 I+-
.~
E1 =.03
Then Fb 1 = f m + f o (k - 1)
Fig. (b) Fig. (c) = 180000 + 156000 (1.5 - 1) =258000
M = (Fb1I1/C1)~ = 258000 x .128 x
The neutral axis will first be determined.
0.5 = 16550 in. lb.
- _ ZAy _ 2 x .1 x 1 + 1.4 x .1 x 1.95 = 1.39
Y - ZA - 2 x .1 + 1.4 x .1 The factor 1/2 is due to the fact that
portion (1) is only one half the beam section
Fig. (c) shows the unit strain picture. in Fig. (d).
The lower edge of the beam section is strained
to the maximum value of .035 as shown on the Portion 2. Fig. (e) shows the developed
stress-strain curve in Fig. C3.22. Since symmetrical section for the upper portion (2)
plahe sections remain plane the unit strain of the beam section.
e 2 at the upper edge of section is e 2 = .035 x
.61/1.39
Solution 1
= .0154.
Equation (3) was derived for a symmetrical

section about the neutral axis. The equation
--r.:"!: I - 1. &---+1 _
0.61 -r.t"fl C
_. N A, - - - - 1. 22
+
0.61
i
•• I I
I I
C:::_::':::J _
I e =.0154
involves finding the resisting moment Fig. (e)

developed by one half the beam section and
multiplying by 2. This is permissible since
the unit strain at both top and bottom edges The unit strain picture is shown in Fig.(f)
is the same. In this solution we will continue
to use equation (3). To do this it is
necessary to make symmetrical sections for the Ia = l~ x 1.5 X 1.22
3
- 1~ x 1.4 X 1.02
3
= .104
I 2/C2 = .0514/.61 = .0842,

Q2 = .61 x .1 x .305 + 1.4 x 0.1 x 0.56 = .097 Q2 = 0.61 x 0.1 x .305 + 1.4 x .1 x .561
= .0971
2Q2 = 0.194
k2 = 0.194/.1704 = 1.14 k2 = .0971/.0842 = 1.15
The stress for a unit strain of .0154 As explained in solution (1), for e =
from the stress strain curve in Fig. C3.22 is .0154, f m = 172000 and f o = 129400.
172000, and f o = 129400.
Fb 2 = 172000 + 129400 (1.15-1) = 191400
Then Fb 2 = 172000 + 129400 (1.14-1) = 190100
M2 = 191400 x .0842 = 16100
M2 = 21 (Fb2 I 2/c 2) = 190100 x .1704 = 16200 Mtotal = M~ + M2 = 16620 + 16100 =
32720 in. lb.
Total resisting moment = M~ + M2 = 16550 +
16200 = 32750 in. Ibs. EXAMPLE PROBLSM 4. Fig. C3.24 shows an un-
s~~etrical I beam section. The material is
Solution 2. 7079-T6 aluminum alloy die forging. The
upper portion is in bending compression. It
Instead of making each portion a will be assumed that the compressive crippling
s~rrnetricalsection as was done in solution (1) stress for the outstanding upper legs of the
and dividing the results by two, we will find section is 65000 psi. (The theory and method
the internal bending resistance of each portion of calculating crippling compressive strength
as is when bending about the neutral axis of is given in another chapter.) The Ultimate
the entire section. Equation (3) now becomes design bending moment is 16500 in. lb. Find
for each portion of beam section, M.S.
Fb =f m + f o (k - 1) - - - 3' i4-- 1. 5 ----+I e=.009
The section modulus of each portion refers

1
2 -N.A.- - -l- 01"
to neutral axis of entire beam section. Fig.
~
'10
I
(g) shows lower portion (1). _
I~
1
= "3bc~ = "31
3
x .1 x 1.39
3 . l·lO
TN.A.
e~=. 0129--
= .0895
= .0895/1.39 = .0645 b
lJ
---+1
\ 39=c ~
I+- O. 1
1--- 3/4--1
Fig. C3.24 Fig. C3.25
I~/c~
Q~
..= 1.39 x .1 x .695
Fig. (g) Solution:
= .0965 - = 2xO.lxl+l.4x.lxl.S5+.65xO.lx.05 = 1.178

Y 2xO.l+l.4xO.l+0.65xO.l
The maximum compressive stress permitted

From equation (3'), Fb = 180000 + 156000 is 65000 ps 1. From Fig. C3. 23, using the
(1.50 - 1) 258000 = stress-strain curve, we obtain a unit strain
of .009 for this stress. The unit strain at
M~ = Fb~ x I~/c~ - 258000 x .0645 = 16620 the bottom edge of section is .009 (1.178/
0.822) =
.0129 as shown in Fig. C3.25. From
Fig. (h) shows u)per Fig. C3.23 this strain causes a stress of
portion (2) 67500 psi.
1 2 = ~ x 1.5 x .61
1
3
3
_
~1~1.5~
1~ Y.61=C2
N.A~
Upper portion:
Fig. 1)
(See
T II--l.
0.822
---+l0~1
50
3"x 1.4 x .51 _1

I NA -"3 x 1.5 x .822 - 3'
'3 1
= .0514
Fig. (h)
~N.A.-- ---N.A.
x 1.4 X 0.722 3 = .102 Fig. (i)
C3.9
I/c = .102/.822 = .124 section will develop in bending about x

axes and Mx the design bending.
Q = 0.822 x 0.1 x 0.411 + 1.4 x 0.1 x 0.772
(2) Using My, carry out the same procedure for
= .1418 bending about y axes and find,
k = .1418/.124 = 1.142
f max = 65000. From Fig. C3.23, fo = 30000 Then the moment ratio Rb for combined
bending is,
Then fb = 65000 + 30000 (1.142-1) = 69260
m 1 = fbIlc = 69260 x .124 = 8580 in. lb.
_ 1
Lower Portion. Then margin of safety M.S. - Rb - 1.
INA = %x 0.75 x 1.178 3

C3.9 Section with One Axis of Symmetry.with Moment
'I~N'A'
Vector not Parallel to Either Axis.
1
- 3" x 0.65 X 1.078 3 1. 178 Since the symmetrical axis is a principal
INA = .137, I/c .137/ = 1 1 aXiS, the procedure in this case is the same
as for the double sYillTIetric case.
1.178 = .1164 1+-0. 75--+1 to. 1
Fig. (j) Rb = Rbx + Rb y
k = .1425/.1164 = 1.222
f max = 67500 M.S. =Rb1 - 1
From Fig. C3.23, f o = 44500 C3.10 Unsymmetrical Section with No Axis of Symmetry.
fb = 67500 + 44500 (1.222-1) = 77370 psi

Fig. C3.27
m2 =.1164 x 77370 = 9000 in. lb. shows an unsym-
metrical section
M
Total internal allowable resisting subjected to the

moment = m 1 + m2 applied moment
vector M.
or Ma = 8580 + 9000 = 17580 in. lb.
For this Fig. C3.27
= 0 • 94 case the pro-
Rb -- M
M -- 17580
16500 (L oa d ra t i 0 )
cedure is as
a
follows:
M.S. = Rb1 1
- 1 = 0.94 - 1 = .06 (1 ) Determine
principal axes
C3. 8 Complex Bending. Symmetrical Section. Moment location by equation,
Vector Not Parallel to Principal Axis.
2I xy
Fig. C3.26 shows a y
tan 2G = Iy-I x
double sym~etric section.
The x and y axes are M where x and yare centroidal axes, Ix
therefore principal and I y are moments of inertia about
axes. these axes and I xy the product of
inertia.
The following pro-
cedure can be followed I (2) Resolve the given moment M into compon-
which is quite con- I Fig. C3. 26 ents Mxp and Myp '
servative. y
(3) Follow the same procedure as before.
(1) Resolve the given moment into components
about x and y axes, or Mx and My' (4) The stress ratio Rb = Rbxp + Rbyp
(2) Using Mx follow the procedure as given
in the example problems and find Rb x = M.S.=..l.-l
Mx/Max ' where Max is the internal moment Rb
C3.11 Alternate More Exact Method for Complex a shear V-=600 Ibs. The problem is to find
Bending. the margin of safety under this combined
loading.
A beam section when resisting a pure
external bending moment bends about an axis For pure bending only the stress ratio is
that is called the neutral axis. No matter Rb =lL = 14000 = .933 (the value of 15000 is
what the shape of the beam cross-section for Ma 15000
any given external moment, there is an axis obtained from example problem 2).
about which bending takes place. The general
case involves an unsymmetrical beam cross- The stress ratio in shear is Rs =
fs/F su ,
section and material which has different where f s is the flexural shear stress and Fsu
stress-strain curves in compression and the ultimate shear stress of the material.
tension in the inelastic range. The neutral The problem therefore is to find the value of
axis therefore does not pass through the
fs'
centroid of the cross-section and thus the
method of solution is a trial and error The eqUivalent trapezoidal and triangular
approach. The solution procedure is outlined bending stress distribution will be determined
in Chapter A19, Article A19.17, and therefore for the design bending moment of 14000 in. lbs.
will not be repeated here. Also the chapter
dealing with the design of beams with non- For a triangular stress variation,
buckling webs explains and illustrates how
the ultimate bending resistance of an entire Fb =Mc/I =
= 14000/0.358 39150 psi.
beam section is determined.
From example problem (2) the shape factor
k was 1.17. On Fig. C3.19, the curve for
C3.12 Strength Under Combined Bending and k = 1.17 has been plotted. Starting with the
Flexural Shear. Fb stress of 39150 at the left scale, run
horizontal to an intersection with the k =
The previous part of this chapter has 1.17 curve, the projecting vertically down-
dealt with the determination of the strength ward to intersections with the stress- strain
of a beam section in pure bending. The usual curve and the f o curve to give 35800 for f m and
beam design problem involves flexural shear 19700 for f o ' The stress results are shown
with bending. In finding the true internal graphically in Fig. C3.28a and Fig. C3.28b.
resisting moment, the Cozzone simplified
method derives a trapezoidal bending stress 1 3 39150
distribution which will produce the same
internal resisting moment as the true internal 8;!;118---'-=-~~~rr~~-_--:~
bending stress system. A triangular stress
J__ T Lil~l~ . _
system is then derived which will also give
the true bending moment.
N.A
Now the equation for flexural shear stress -+1 1- -+1970/)
for a triangular bending stress distribution 1/2 Beam Section Fig. C3. 28a Fig. C3. 28b
is Case 1 Case 2
f
s
=VQ
It - (A)
The flexural shear stress is a function
of the rate of change of the bending stress.
Thus we can obtain a shear correction factor
Thus to use equation (A) for a C by comparing the bending stresses in the
trapizoidal bending stress, a correction factor two stress distribution diagram.
(C) must be applied or equation (A) becomes
The shear stress is maximum at the
f - CVQ neutral axis in this particular problem.
s - It - - - - - (B)
The total normal force on the cross-section
of beam above the neutral axis equals the
To illustrate how the correction factor stress times the area.
(C) can be determined, the I beam section used
in example problem (2) will be used. For simplification, the beam section will
be divided into the two portions labeled (a)
We will assume the ultimate design moment and (b).
of 14000 in. lbs. is produced by a load of
600 lbs. acting on a cantilever beam at a
point 23.30 inches from the fixed end of the
beam. Thus the beam section at the support
is subjected to bending moment of 14000 and
C3.11
For Case 1:- the cross-section will be determined. The

correction factor for this point is found by
Load on portion (a) is, comparing the loads on portion (a) for Case 1
and Case 2 stress patterns.
19700 x 1.375 x 0.125 =3380 lb.
Case 1. Load on (a) = 3380 + 2600 = 5980
16100 + 14100
2 x 1.375 x 0.125 = 2600 lb. Case 2. Load on (a) = 6300.
Portion (b) Hence, C = 5980/6300 .95
19700 x 0.875 x 0.125 = 2155 C3. 13 Strength Under Combined Bending Flexural Shear
and Axial Compression.
14100 x 0.5 x 0.875 x 0.125 = 771
The subject of the ultimate strength
Total Force 8906 lbs. design under combined loads is treated in
detail in a later chapter.
For Case 2:-
A conservative interaction equation for
Portion (a) combined bending, shear and axial load is,
39150 + 34250 (Ra + Rb)2 + R~ = 1

2 x 1.375 x 0.125 = 6300
1
Portion (b) or M.S. =J (Ra + Rb) 2 + R~ - 1
1874 Rb includes effect of secondary bend-

34250 x 0.5 x 0.875 x 0.125 = lbs. ing moment due to axial load times de-
8174
flection.
Thus the correction factor is,
C3. 14 Further Values of f m and fo •
C = 8906/8174 = 1.09
Table C3.2 gives the yield and ultimate
Then f s at neutral axis is, values of f m and f o for a number of other
materials common to the aerospace field of
structures. The yield and ultimate modulus
in bending Fb is found by SUbstituting in
the equation.
The Ultimate shear stress for this
particular magnesium material is 19000 psi.
(See chapter on material properties.)
TABLE C3.2 Values of f m and fo for Finding Fb
fs 3060
Stress Ratio Rs = F
sw
= 19000 = 0.161 Fb = f m + f o (k - 1) Yield Ultimate
Material Size fm=Fty fo fm=Ftu fo

The interaction equation for combined 2014-T6 Alum. Al. Die Forgings t'" 4 in. 52 18.8 62 57.5
bending and flexural shear is 6061-T6 Alum. Al. Sbeet t'" 0.02 35 17.5 42 40.5
7075-T6 Alum. Al. Die Forgings t'" 2.0 65 25.5 75 70.0
7079-T6 Alum. Al. Hand Forgings(L) t'" 6.0 62 33.5 71 67.0
Rb + R~ =1 A-Z61A Magnesium Al. Extru. (Long)
HK31A-0 Magnesium AI. Sheet
t'" 0.25
.016-0.25
21
18
7.6
4.0
38
30
15.7
22.6
1 ZK60A Magnesium Al. Forgings (Long) 26 7.5 42 30.8
A1S1 Alloy Steel (NormaliZed) >0.188 70 43.2 90 83.1
Whence, Margin of Safety = JRb + R~
- 1 AlSl Alloy Steel (NormaliZed)
A1S1 Alloy Steel (Heat Treated)
t '" O. 188 75
103
47.1
51. 6
95
125
86.0
118.0
A1S1 Alloy Steel (Heat Treated) 132 66.0 150 146.0
AISI Alloy Steel (Heat Treated) 163 53.0 180 172.0
1 AISI Alloy Steel (Heat Treated) 176 65.0 200 192.0
or, M.S. = J.9332+ .1612 - 1 = .05 17-7PH Stainless Steel
PH15-7 MO (RH950) Stainless Steel
150
200
62.0
55.0
180
225
161. 0
160.0
Ti-8MN Titanium Alloy 110 36.0 120 116.0
Thus the effect of the flexural shear

stress was to reduce the margin of safety of PROBLEMS
.07 in pure bending to .05 in the combined
stress action. (1) A round tube 1-1/2 inches in diameter
has a wall thickness of .095 inches.
As further calcUlation of the shear It is made of aluminum alloy whose
correction factor C, its value for the shear stress-strain curve is shown in Fig.
stress at the upper edge of portion (b) of C3.2. If the maximum unit strain in
compression is limited to .008, what Fig. 1

bending resisting moment will the section ~1_1~1
develop. (Note, since stress-strain
curve has different shape in tension and
compression, neutral axis does not
!'L O1
.
(2)
coincide with center line axis, thus use
trial and error method.
Same as Problem (1) but use a square tUbe

with 1-1/4 inch outside dimension and
t -'" +-0.1
t~:14~1
1
rD~81Il:
.081 inch wall thickness.
8 1
Use the Cozzons method for solving the ~ +-16
following problems.
~I .!. I~
(3) Find the ultimate bending moment that 2
Fig. 4 Fig. 5 Fig. 6
each of the following beam sections will
develop when bending about the principal
axis and made from each of the following (4) A simply supported beam has a span of 24
materials, inches. Depth of beam limited to 2 inches.
It must carry an ultimate load of 4000
(a) 7075-T6 Alu~inum Alloy Extrusion. Ibs. located at midpoint of beam. Material
Ftu = 75000, Fty = 65000, is 7075-T6 aluminum alloy extrusion.
Design an I shaped section to carry this
(b) Ti-6Al-4V Titaniwn Alloy, load, Neglect areas of corner fillets
Ftu = 130,000, Fty = 120,000. that would be used in extruded shapes.
REFERENCES.
Ref. 1. Bending Strength in Plastic Range. By

F. P. Cozzone, J. Aeronautical Sci.,
May, 19~'3.
Ref. 2. Vought Structures Manual;
CHAPTER C4
STRENGTH AND DESIGN OF ROUND, STREAMLINE, OVAL
AND SQUARE TUBING IN TENSION, COMPRESSION, BENDING,
TORSION AND COMBINED LOADINGS.
C4. 1 Introduction. steel and 2024 aluminum alloy. Observation of

Fig. C4.1 shows that for temperatures below
Before the advent of the stressed-skin 350 oF, a1Qminum alloy is lighter unless the
structure for aircraft, or during the period steel is heat-treated to Ftu = 180,000 or
when fuselage and wing were fabric covered, above. Above 350 oF, the ultimate strength of
round, oval and square tubing were used in aluminum alloy falls off rapidly, but steel
designing the major structure of the fuselage continues its rather uniform decrease in
and wing. If the wing and tail were externally tensile strength.
braced, streamline tubing was used. The
development of the metal covered structure The graph explains why aluminum alloy
eliminated the use of tubing in fuselage and cannot be used entirely for the surface of
wing design, however, tUbing continued to be supersonic airplanes flying at speeds around
used for landing gear structure, engine mounts, 2000 miles per hour, as aerodynamic heating
control systems, fixed equipment such as would produce surface temperatures in the
passenger seats, etc. region where the strength of aluminum alloy
decreases rapidly.
With the opening of the space age, tUbing
as a structural unit in space vehicles is
again being widely used because drag in space
is not an important factor. Round tubing is 700
the best shape for transmitting torsional
forces and thus is widely used in control 600
systems. Round and square tUbing permit simple
connection or end fitting design. The metals 0
0
industry has made available a large number of 0
,...., 500
diameter and wall thicknesses and thus the ~
structural designer has a large number of sizes ~ 4'00

to select from. ')
r<.
300
C4. 2 Design for Tension.
200
In general the strength design require-
ments are that the limit loads must be carried
without exceeding the tensile yield stress 100
100 200 300 400 500 600 700 800
(Fty) of the material and the ultimate design Temp. of. 1/2 Hr. Exposure
load which is equal to the limit load times a
Fig. C4.1
factor of safety must be carried without
failure, which means the tensile stress cannot Practically every structural tubular
exceed the (Ftu) of the material. In general member in a flight vehicle structure must be
for aircraft, the factor of safety is 1.5. fastened or attached to another adjacent
For unmanned missiles and space vehicles the member. The connection can be made by using
factor of safety may be as low as 1.2. Since some sort of end fitting which is fastened to
the ratio of Ftu to Fty for materials varies the tube by rivets or bolts or by welding. If
widely, sometimes the yield under limit loads rivets or bolts are used, holes are cut in the
is more critical than failure under the tube walls which means the tube is weakened
ultimate design loads, thus the student should since tube area is cut away, thus net area on
always be sure he has the critical situation. any tube cross-section must be used in calcu-
lating the tube tensile strength. If welding
Since elevated temperature and time of is used in the connection, the welding heat
exposure effect the yield and ultimate strength causes grain growth in the tube material
of materials, the problem of material selection adjacent to the weld area, which decreases the
relative to light weight becomes an important tube strength, thus a welding correction factor
design factor. Fig. C4.1 shows a plot of the must be used in calculating the ultimate tensile
Ftu/w ratio versus elevated temperatures up strength. Designing for tension will be
to 800 degrees, where w is the density of the illustrated later in this chapter.
material. The tube materials are AISI alloy
C4.1
STRENGTH & DESIGN OF ROUND, STREAMLINE, OVAL AND SQUARE TUBING
C4.2 IN TENSION, COMPRESSION, BENDING, TORSION AND COMBINED LOADINGS.
C4. 3 Design for Compression. any tendency of the test curve to rise rapidly
for very short lengths where failure is by
The strength of members With stable cross- block compression. Table C4.l shows the
sections when acting as columns can be cal- reSUlting short and long column equations
culated by Euler's equation if the bending after values of Fco and E have been SUbstituted
failure is elastic, or E is constant (eq. C4.1) in equations C4.3 and C4.4 and E in the Euler
and for inelastic bending failure, Euler's equation. The column headed transitional
equation With the tangent modulus Et replacing L' /p, represents the value of L' /p where failure
E (eq. C4.2) checks test reSUlts. (The stUdent change from inelastic to elastic failure or, in
should refer to Chapters A18 and C2 for theory other words, it is the dividing point between
on column strength.) the so-called long and short column range.
Thus if the equations are used, the L'/p value
- - - - - - - - - - - (C4.1) must be known in order to select the proper
equation.
(C4.2) C4. 5 Column Formulas for Aluminum Alloy Tubes.
L' is the effective length and equals From (Ref. 1), the basic short column
L/~ where c is the column end fixity equations for aluminum alloys are:-
coefficient.
FC - Fco [1 - 0.385 (L' /p)/n) E/FcoJ (C4.5)
Long and Short Columns
Fc = Fco [1 0.333 (L'/p)/nv E/FcoJ (C4.6 )
For many years the problem or subject of
inelastic column strength or failure was Fc - Fco [1 - 0.272 (L'/p)/n"; E/F co ] (C4.7)
treated almost entirely from a consideration
of test results. That is, sufficient tests For long columns:-
were made to establish the shape of the failing
stress curve in the region where the failure Fc = n lil E/(L'/p)2 ------- - (C4.8)
was at stresses above the proportional limit
stress of the material. Mathematical curves The equations for determining Fco are
were then derived to fit the test results. given in Table C4.2 (from Ref. 1). The table
Engineers referred to the columns which failed also indicates which of the three short column
by inelastic bending as short columns, and equations to use for the various aluminum alloy
thus referred to the equations that fit the materials.
test data as short column equations. The
columns that failed by elastic bending were To illustrate the use of Table C4.2, the
then referred to as long columns. The test column formula for 2024-T3 aluminum alloy
curve for long columns would follow the Euler tubing will be derived:-
column equation (C4.l) and thus tests were
not necessary to establish allowable failing From Chapter B2, we find the following
stresses in the so-called long column range. strength properties for 2024-T3 tubing,
Thus over the years short column equations
based on test results have been presented by Ftu = 64000, Fcy = 42000
official government agencies for use in
structural design. The official publication From Table C4.2, the equation for Fco is,
for the aerospace field is the Military
Handbook MIL-HDBK (Ref. 1). Fco = Fcy (1 + ..; Fcy/lOOO SUbstituting,
C4. 4 Column Formulas for Round Steel Tubes. Fco = 42000 (1 + ..; 42/1000 = 42000 (1 + .205)
From (Ref. 1) the basic short column = 50600

equations are:- From Table C4.2, we note that short
column equation C4.5 applies.
Fc = Fco [l-F co (V/p)lil/4n lil EJ - - - - - (C4.3)
SUbstituting in this equation,
Fc = Fco { 1- 0.3027 [(L' /p )/n
J E/F co J 1.0 IS} - - - (C4.4)
Fc = 50600 [1 - 0.385(L' /p/n"; 10,500,000/50600]
Where Fco is the column yield stress or Fc = 50600 - 431 L'/p - - - - - - - - (C4.9)
(upper limit of column stress for primary
failure). It can be determined from test C4. 6 Column Formulas for Magnesium Alloys.
reSUlts by extending the short column curve to From (Ref. 1) the following short column
a point corresponding to zero length, ignoring equations for various magnesium alloy materials
C4.3
Table C4. 1 Column Formulas for Round Steel Tubes
Short columns (a)

Transi- Local
Material Ftu, ksi Fty, ksi Fco, ksi tional (C) Long columns (b)
Basic failure
Column formula (b) V/p
equation
1025 ...•,••.•..•.... 55 36 36 36, 000-1. 172(L'/p)'" .•• C4.3 122 286 x 1Q6/(V/p) 2 .... (d)
4130 ••.....•••.•••• 95 75(e) 79.5 79,500-51. 9(V /p) 1.6•• C4.4 91 286 x 10 6/(V/p)'" .... (d)
Heat-treated e alloy
steel ••....•..... 125 103 113 113,000-11. 15(L'/p) '" .• C4.3 73 286 x 10 6/(V/p)2 .... (d)
Heat-treated alloy
steel ............ 150 132 145 145,000-18. 36(V/p)'" .. C4.3 63 286 x 10"/(V/p)'" .... (d)
Heat-treated alloy 2
steel ..•••..•.•.. 180 163 179 179, 000-27. 95(L'/p) .. C4.3 56 286 X 10 6/(V/p)'" .... (d)
a Equation C4.1 may be used in the short column range if E d Not necessary to investigate for local instability when
is replaced by Et obtained from the combined stress- D/t<: 50.
strain curves for the material. e ThiS value is applicable when the material is furnished
b L'/p = L!p.,;C: L'/p shall not exceed 150 without specific in condition N (MIL-T-6736) but the yield strength is
authority from the procuring or certificating agency. reduced when normalized subsequent to welding to 60
c Transitional L/p is that above which columns are "long" ksi.
and below which they are "short." These are approximate
values.
are given in Tables C4.4 and C4.4a. or inelastic instability of the column as a
whole. As the slenderness ratio L'/p gets
Table C4. 4 Column Formulas for Magnesium- smaller, the Fc stress increases. Now if the
Alloy Extruded Open Shapesa diameter of the tube is relatively large and
GENERAL FORMULA the wall thickness relatively small or, in
n other words, if the diameter/thickness (D/t)
K(Fcy) ratio is large, failure will result by local
F c = (V/p)m crippling or crushing of the tube wall and
this local failing stress is usually repre-
(Stress values are in ksi) sented by the symbol Fcc. The values of Fcc
in general have been determined by tests (see
Alloy K n m Max. Fc design charts for Fcc versus D/t ratio).
MIA .• ...... ... 180 1/2 1.0 0.90 F cy C4. 9 Design Column Charts.
AZ3lB, AZ61A, AZ80A. 2,900 1/4 1.5 Fc~
AZ80A-T6, ZK60A-T5 • 3,300 1/4 1.5 O. 6 F cy In design, column strength charts are a
a Formulas given above are for members that do not fail by great time-saver as compared to SUbstituting
local buckling. in the various column equations, thus a number
of column charts are presented in this chapter
to facilitate the strength check of columns
Table C4. 4a Column Formula for AZ31B-H24 and the strength design of columns. Fig.
Magnesium-Alloy Sheet C4.2 is a chart of L'/p versus Fc for heat
treated round alloy steel tubing. Fig. C4.3
is a similar type of chart for aluminum alloy
round tubing. Fig. C4.4 gives column charts
Max. F c = F cy for magnesium alloy materials. All three
charts are taken from (Ref. 1). Figs. C4.5
and C4.6 represent a further simplication for
the design of steel and aluminum round tubing.
C4. 7 Short Column Equations for Other Materials.
C4. 10 Section Properties of Round Tubing.
For other metals for which short column
equations are not available, the use of Euler's Table C4.3 gives the section properties
equation, using the tangent modulus Et can be of round tUbing. A tube is designated by
used (eq. C4.2). Refer to Chapter C2 for giving its outside wall diameter (D) and its
information on how to construct column strength wall thickness (t). Thus a 2-1/4 - .058 means
curves using this equation. a tube with 2-1/4 inch outside diameter and a
wall thickness in inches of .058. Since a tUbe
C4. 8 Column Failure Due to Local Failure. is symmetrical about any axis, the polar moment
of inertia, which is needed in torsion problems,
The equations as presented give the equals twice the rectangular moment of inertia
allowable stress due to failure by bending of as given in Table C4.3. For weight comparison,
the column as a whole and the action is elastic the weight of steel and aluminum tubing is
Table C4. 2 Column Formulas for Aluminum Alloys
Alloy and Temper Product F•• Short Transitional Long

Columns L'/p Columns
2014-T3, T4, T451 Sheet and Plate·;

2024-T3, T35l, T36, T4, Rolled Rod, Bar and
T42 Shapes;
Drawn Tube
5052-All Tempers
5083-All Tempers
5086-All Tempers
5454-All Tempers All Products Fcu(l +.../ F.u/lOOO) Equation 1.732r.../E/F•• Equation
5456-All Tempers C4.5 C4.8
6061-T4, T451, T451O,
T4511
All Cast Alloys and Sand and Permanent

Tempers Mold Castings
-
2024-T3510, T3511, T4, Extrusions
T42
2014-T6, T651
2024-T6, T81, T86, T851 Sheet and Plate·;
7075-T6, T651 Rolled Rod, Bar and F eu(l + .../ F eu/1333) Equation 1.346r.../E/F•• Equation
7178-T6, T651 Shapes; Drawn C4.6 C4.8
Tube
6061-T6, T651 Sheet and Plate·
2014-T6, T651, T6510,

T6511, T652
2024-T6, T81, T851O,
2024-T6, T81, T851O,
T8511, T852
7075-T6, T6510, TtiStl, Extrusions, Forgings
T652
7079-T6, T6510, T6511 ,
T652
7178-T6, T651O, T6511
Rolled Rod, Bar and F.u(l +..; F .u/2OOO) Equation 1.224r.../ElFe. Equation
Shapes; C4.7 C4.8
6061-T6, T651, T651O, Drawn Tube;
T6511 Extrusions
X2020-T6, T651 Sheet and Plate·
·Includes clad as well as bare sheet and plate. Equation C4. 8 may be used In the short column range If E' Is
Transitional L'lp Is that above which the columns are "long" and replaced by E, obtained (rom the compressive stress-strain curve (or the
below which they are "short". material.
C4.5
given in the last two columns of the table. Table C4. 5
C4. 11 Some General Facts in Tubing Design. Tension Allowables Near Welds in Steel TUbing (X-4130)
Normalized Welded after HT HT after

1. For a given area, the larger the tube Type of Weld
Tube Welded or Norm. after Weld Welding
diameter, the greater the column strength
if failure due to local crippling is not *Tapered Welds
critical. of 300 or Less .947 Ftu 90,000 psi .90 Ftu
All others .841 Ftu 80,000 psi .80 Ftu

2. The higher the D/t ratio of tube the lower
the crippling or local failure strength. * Note: Gussets or plate inserts considered 00 "taper"
with ~ .
3. If columns fall within the long column ** For (X-4130) Special, comparable, values to the Ftu,
category, the use of higher strength alloy equal to 90, 000 and 80, 000, are stresses 94, 500 and
steel or aluminum alloy will not increase 84, 100 psi, respectively.
strength of column since E is practically Ref. Anc-5
constant for all chrome-moly steel alloys
and likewise for all aluminum alloys.
Failure is due to elastic buckling of the C4. 13 lllustrative Problems in Strength t;:hecking
column as a whole and is therefore a and Design of Round Steel Tubes as Columns
function only of I, L' and E. and Tension Members.
4. The column end restraint affects the neces- PROBLEM 1

sary tube size. Consult the design require-
ments of the Army, Navy, and C.A.A. in this Tube size = 1-1/2 - .058, Length L = 30 in
matter. In general with welded steel End fixity coefficient C 1. =
tubular trusses a coefficient of C = 2 is Material:- Alloy steel, Ftu = 95000. The
permissible except for engine mount and tube is welded at ends.
Nacelle structures. For trusses with Ultimate design loads are:- P -14,500 =
riveted joints a value of not over 1 5 is lbs. compression, and P = 18500 lbs. tension.
generally permissible. Required the Margin of Safety (M.S.)
5. The student should realize that practical Solution: The compressive (M.S.) will be
limitations such as clearance requirements determined first. As the simplest solution,
may determine the diameter of the tube we can use the column curves in Fig. C4.5.
instead of strength-weight considerations. =
For a length of 30 and C 1, from the upper
Thus design can consist of checking the right chart we project upward to the inter-
tubes available under the given section with the 1-1/2 diameter tube and then
restrictions. horizontally to the left hand scale to read
the column strength of 14800 lbs. which we
C4. 12 Effect of Welding of steel Tubes Upon the Tension will call the allowable failing Pa.
and Column Strength.
S = Pa _ 1 = 14800 _ 1 = 02
Since welding affects the grain structure
11..
M
P 14500 •
of the tube material adjacent to the weld, The tube strength could also be found by
tests show the strength of the material using Fig. C4.2 as follows:
adjacent to the weld is decreased as compared
to the unwelded material. If a tapered weld L' = L/Y-C- = 30/y-r- = 30
is used, the effect of the weld is decreased.
Table C4.5 shows the allowable stresses in L'/p =
30/.5102 =
58.7. P is found from
tension to use when tension loads are carried. Table C4.3 as well as the tube area 0.2628
sq. in. Using 58.7 for L'/p on lower scale
In short columns, the primary column and projecting upward to the Ftu = 95000
failing stress may be greater than the local curve, which is the lower curve, and then
crippling strength of the tube adjacent to horizontally to left hand scale we read
the weld at the end of the tube. This local Fc =56500 psi.
failing stress due to welding is referred to
as the weld cut-off stress and the column Whence, Pa = FcA = 56500 x .2628 =
compressive stress Fc should not exceed this 14850 lb.
value. This cut-off weld stress is shown by
the horizontal lines in Fig. C4.2 and C4.5. The solution obViously could be made by
SUbstituting in the short column equation for
steel having Ftu =
95000, or
180
.... , I I
Tobie of Weld Cutoffs
I
\
170 Heat Treated After Weldina I--
~.Ftu= 180 ksi
Heat Treated Cutoff,ksi Reference
160
\ to Ftu. ksi
\\ 180
150
144
120
REF. 1
150 125 95 I--
~ \ Normalized
140 ' ... '-\ Cutoff heat treated 100 76.6
f--
95 67.5
after weld ing
130
\ \ Ftu= 150 ksi
120 \' 60 ~ 50
on
on
~
, r\\~/
- - - _ Unwelded column
~ ii> 40
110
r- ...... aliowobles .A 0'
C
on
~
:c.on
Ftu=125 ksi
\ ii>
c
S 30
100 E u
~ \ ::0 IE
-0 30
1\ u
90
'"
\'
:0
0
'J 20
E
80
, <i
,,-0
10
"'-",,\
70
-"weld cutoff after heat
treat 4130 0
60 0 20 4Q 60 80 100 120 140
I L'lp
f t Y=75 ksi
50
Ftu=95 ksi Fig. C4.3 2024-T3 and 6061-T6
~~ round aluminum alloy tubing.
40
30
20
0
I0 20 40 60
""
80 100 120
"-I'-
140 160 180 200 220
L'/p
Fig. C4.2 Allowable column stress for heat-treated

alloy-steel round tubing.
50
Euler curve, E: 6,5 xl0 6 psi

40 -
.s
VV
AZ80A-T5 and ZK60A-T5
III extruded shapes (except hollow)
III
~ 30 _ 0 96Fcy I I I -
~
in AZ80A-F extruded shapes
c
E
.2 20
o
:/ (except hollow)
I I I
/' Round tubes, Fcy : 10 ksi
"'t2(
U'
0
IJ...
10
Fcy
:'--- ~
o I
o 20 40 60 80 100 120 140 160
LIP
Fig. C4.4 Allowable column stresses for magnesium-alloy columns.

::t:
Eo<
t:J 4
Z
\i::l
ll::
Eo<
rn
2
~p
,.J
0
t.l
C=2
C=l
rn
~ 20
Po<
r-.. o 20 40 60 80
0
COLUMN LENGTH - INCHES
~ 16
~
g 12 36
Eo<
20 40 60 80
rn
~
Po<
50
S
~ 40
8 4
~ 30
::t:
Eo< C=2
~ 20 C=l 20 40 60 80 100
\i::l
~
rn
~ 10
Fig. C4.5
6
t.l STRENGTH OF CHROME MOLY ROUND STEEL TUBES
C=2 FTY = 75, 000 PSI FTU = 95, 000 PSI
C=l 40 80 120 160
7000 "'\"
,
(J)
24000 ".'- ...... ", ,'''" ' ,A.., 1/4
~ 6000 (J) ll- 3/B
'C ~1/2;
5
Po< ~ 0- 5/B
E - S/4
I 5000 ~ 20000 F -7/S
0-1
~o-l A.:1/4 H - 1-1/s
, I _ 1-1/4
1'<1 4000
a- 3/8
C - l/Z'
J) • 5/8'*
~o-l 16000 J - I-S/S
K - 1-1/2;
o-l L - 1.5/B
, E - S/•• 1'<1
~ F - 7/8' o-l
N- 1-3/4
N -1-7/B
~ 3000
0-'1·
H-l-1/B
~ 12000 0-2
P - 2;-1/4
o-l
o-l
1 -1-1/4'
J - l-S/B
~ Q - 2-I/Z"
< K - 1-1/2
o-l R - 2;-3/.'
:;i 8-30.0"
[:~.
8000
1'<1
Eo<
'0 g 4000
.'
~ C ' Eo<
=3
rmARD SIZES
o 10 20 30 40 50 o 20 40 60 80 100 120
COLUMN LENGTH - INCHES (C = 1) COLUMN LENGTH - INCHES (C = 1)
35000
14000 A - S/8
(J)
B - 1/2;
(J)
~ 30000 C - 5/S
~ 12000
A- 1/4 Oc 3/4
B - 3/B 0 E -7/S
0 C - 1/2 Po< Y-I
Po< 0- 5/B I G· 1-1/S
E - 3/4 ~ 25000 Ii - 1-1/4
.049 F - 7/B I - l-:VS
~ 10000 G -'1 0 J - 1-1/2
H-l-1/B o-l , K- 1-5/s
0
, L - 1-3/4
o-l I - 1-1/."
J - 1-3/8 ~ 20000 M- 1-7/B
1'<1 iII
o-l 8000 K - 1-1/2' N-Z
iII
<
L - 1-5/S
M- 1-S/4* ~ 0- 2-1/4
P - 2;-1/2
~
o-l 6000
N - 1-7/S
0-2
315000
o-l
'Q.2-3/4 '
R- s* 0.0.
o-l P -2-1/4 <
Q - 2-1/2 0. D.
< 1'<1
1'<1
Eo<
~ 10000
g 4000
~
o-l
Eo< ~ 5000
o-l
~
*AN STArfuARDSIZES
*AN STANDARD SIZES o 20 40'" 60 8() 10() 120
o 20 . 4()6() 80 100 120 COLUMN LENGTH - INCHES (C = 1)
COLUMN LENGTH - INCHES (C = 1) 35000
24000 A- 1/4 (J) A-IVa

B-l/2
~
(J)
B - 3/8
30000
~~ 20000
C - 1/2 C - 5/S
0-5/8
, E - 3/4
5Po< 0- S/4
E-7/B
F - 7/B F-l
0-1
H-1-1/B
~ 25000 G~ 1-1/8
H - 1-1/4
q
I - 1-1/4 o I - 1-3/8
<516000 J - 1-3/S o-l J - 1-1/2
o-l K - 1-1/2 K - 1-5/S
1'<1 L - 1-5/S ~ 20000 L - 1-3/4
o-l M- 1-3/4 N-I-7/8
~ 12000 N - 1-7/S
0-2' ~ N·Z
0-
~o-l P-2-1/4"
Q - 2-1/2
o 15000
o-l
p - 2;
Q-2
R - 2-3/4 o-l R-30.,0.
:;i 8000 8 -3 0.0. <
1'<1 ~ 10000
g
Eo<
4000
g
Eo<
Eo< o-l
o-l ~
~
* AN STANDARD SIZES
o 20 40 60 80 100 120
COLUMN LENGTH - INCHES (C = 1) o 20 40 60 80 100 120
COLUMN LENGTH - INCHES (C = 1)
Weight Weight
Lb/l00 ins. Lb/l00 ins.
Dia. Gage A I I/Y D/t Steel Dural Dia. Gage A .? I I/Y D/t Steel Dural
.?
1/4 .022 .01576 .0810 .000103 .000825 11.38 .45 .16 1 7/8 .035 .2023 .6507 .08565 .09136 53.60 5.73 2.04
.028 .01953 .0791 .000122 .000978 8.3a .55 .20 .049 .2811 .6458 .11720 .12500 38.25 7.95 2.84
.058 .3311 .6427 .13677 .14589 32.30 9.38 3.35
3/8 .028 .03063 .1231 .000462 .002466 13.39 .86* .31 .065 .3696 .6404 .15156 .16166 28.80 10.47 3.74
.035 .03739 .1208 .000546 .002912 10.72 1. 06 .38 .083 .4673 .6342 .18797 .20050 22.60 13.25 4.73
.049 .06018 .1166 .000682 .003636 7.65 1.43 .51
2 .049 .3003 .6900 .14299 .14299 40.80 .50 3.04
1/2 .SQ8 .04152 .1672 .001160 .004641 17.85 1.17 .42 .058 .3539 .6869 .16696 .16696 34.45 10.03* 3.58·
.035 .05113 .1649 .001390 .005559 14.28 1.45* .52* .065 .3951 .6845 .18514 .18514 30.75 1l.19 4.00
.040 .06943 .1576 .001786 .007144 10.20 1.96 .70 .083 .4999 .8783 .2300 .2301 24.10 14.16 5.06
.095 .5885 .6744 .2586 .2586 21.05 16.11 5.76
5/8 .028 .05252 .2113 .002345 .007503 22.30 1.49 .54
.035 .06487 .2090 .002833 .009065 17.85 1.84* .66* 2 1/4 .049 .3388 .7783 .2052 .1824 45.90 9.59 3.43
.049 .08867 .2044 .003704 .011852 12.77 2.51 .90 .058 .3994 .7753 .2401 .2134 38.80 11.30* 4.05*
.058 .10331 .2016 .004195 .013425 10.79 2.93 1. 05 .065 .4462 .7728 .2665 .2369 34.60 12.64 4.52
.083 .5651 .7667 .3322 .2953 27.15 16.01 5.72
3/4 .028 .06351 .2555 .004145 .011052 26.80 1.80 .65 .095 .6432 .7626 .3741 .3325 23.70 18.22 6.51
.035 .07862 .2531 .005036 .013429 21.42 2.23* .80*
.049 .10791 .2485 .006661 .017762 15.30 3.06 1. 09 2 1/2 .049 .3773 .8667 .2834 .2267 51.00 10.68 3.82
.058 .12609 .2455 .007601 .02027 12.94 3.57 1. 28 .058 .4450 .8635 .3318 .2655 43.10 12.60 4.50
.065 .13988 .2433 .008278 .02208 11.53 3.96 1. 42 .065 .4972 .8613 .3688 .2950 38.45 14.09* 5.03*
.083 .6302 .8550 .4607 .3686 30.10 17.85 6.38
7/8 .028 .07451 .2996 .006689 .015289 31.23 2.11 .76 .095 .7178 .8509 .5197 .4158 26.30 20.34 7.27
.035 .09236 .2973 .008161 .018653 25.00 2.62* .94*
.049 .12715 .2925 .010882 .02487 17.85 3.60 1. 29 2 3/4 .049 .4158 .9551 .3793 .2759 56.10 11.78 4.20
.058 .14887 .2896 .012484 .02853 15.10 4.22 1. 51 .058 .4905 .0521 .4446 .3233 47.40 13.90 4.96
.065 .16541 .2865 .013653 .03121 13.47 4.66 1. 68 .065 .5483 .9496 .4944 .3596 42.30 15.50* 5.55*
.083 .6954 .9434 .6189 .4501 33.15 19.70* 7.04
1 .035 .10611 .3414 .012368 .02474 28.56 3.01* 1.07* .095 .7924 .9393 .6991 .5084 28.95 22.48 8.03
.049 .14640 .3367 .016594 .03319 20.40 4.15 1. 48
.058 .17164 .3337 .019111 .03822 17.25 4.86 1. 74 3 .058 .5361 1. 0403 .5802 .3868 51. 70 16.18 5.42
.065 .19093 .3314 .020970 .04193 15.38 5.41 1. 93 .065 .5993 1. 0380 .6457 .4305 46.20 16.95 6.06
.083 .7606 1. 0318 .8097 .5398 36.15 21. 55* 7.70*
1 1/8 .035 .1l985 .3856 .01782 .03168 32.10 3.40* 1. 21 .095 .8670 1. 0276 .9156 .6104 31. 58 24.56 8.78 o
.02402 22.95 4.68* 1. 68 .120 1. 0857 1. 0191 1.1276 .7518 25.00 30.76 1l.00 'z:1
.049 .16564 .3808 .04270
.058 .19442 .3780 .02775 .04933 19.40 5.51 1. 97
.065 .21650 .3755 .03052 .05425 17.30 6.14 2.20 3 1/4 .058 .5816 1.1287 .7410 .4560 56.10 16.47 5.89
.065 .6504 1. 1263 .8251 .5077 50.00 18.40 6.58
1 1/4 .035 .13360 .4297 .02467 .03948 35.70 3.78* 1. 35* .083 .8258 1.1201 1.0361 .6376 39.15 23.38* 8.35*
.049 .18488 .4250 .03339 .05342 25.50 5.23* 1. 87* .095 .9416 1.1160 1.1727 .7217 34.20 26.66 9.52
.058 .2172 .4219 .03867 .06187 21.55 6.15 2.20 .120 1.1800 1.1074 1.4472 .8906 27.10 33.43 1l.95
.065 .2420 .4196 .04260 .06816 19.22 6.86 2.45
3 1/2 .065 .7014 1. 2147 1.0349 .5914 53.80 19.85 7.09
1 3/8 .035 .1473 .4739 .03309 .04814 39.25 4.17 1. 49 .083 .8910 1. 2085 1.3012 .7435 42.20 25.20 9.01
.049 .2041 .4691 .04492 .06534 28.05 5.78* 2.07 .095 1. 0162 1.2043 1.4739 .8422 36.85 28.70* 10.25*
.058 .2400 .4661 .05213 .07583 23.70 6.80 2.43 .120 1.2742 1.1958 1. 8220 1.0411 29.15 36.00 12.89
.065 .2675 .4638 .05753 .08367 21.15 7.58 2.70
3 3/4 .065 .7525 1. 3031 1.2777 .6814 57.60 21. 30 7.60
1 1/2 .035 .1611 .5181 .04324 .05765 42.80 4.56 1. 63 .083 .9562 1. 2968 1.6080 .8576 45.20 27.06 9.67
.049 .2234 .5132 .05885 .07847 30.60 6.32* 2.26* .095 1. 0908 1. 2927 1.8228 .9722 39.50 30.84* 11. 04*
.058 .2628 .5102 .06841 .09121 25.85 7.45 2.66 .120 1. 3685 1. 2841 2.2565 1. 2035 31.25 38.70 13.82
.065 .2930 .5079 .07558 .10079 23.05 8.30 2.97
.083 .3695 .5018 .09305 .12407 18.08 10.47 3.74 4 .065 .8035 1. 3915 1. 5557 .7779 61. 50 22.75 8.12
.083 1.0214 1. 3852 1. 9597 .9799 42.20 28.95 10.32
1 5/8 .035 .1748 .5622 .05528 .06803 46.40 4.95 1. 77 .095 1.1655 1. 3810 2.2228 1.1114 42.10 32.95 11. 78
.049 .2426 .5575 .07540 .09279 33.15 6.87* 2.46 .120 1.4627 1. 3725 2.7552 1. 3776 33.33 41. 40* 14.80lt
.058 .2855 .5544 .08776 .10801 28.00 8.09 "2.89
.065 .3186 .5520 .09707 .1l948 25.00 9.05 3.23 4 1/4 .134 1.7327 1. 4557 3.6732 1.7408 31. 75 49.10* 17.55*
.083 .4021 .5459 .11985 .14751 19.58 11.40 4.06
4 1/2 .156 2.1289 1. 5369 5.0282 2.2347 28.80 60.40* 21. 5511-
1 3/4 .035 .1885 .6065 .06936 .07927 50.00 5.32 1. 91
.049 .2618 .6017 .09478 .10832 35.70 7.42* 2.65* 4 3/4 .188 2.6944 1. 6143 7.0213 2.9563 25.25 76.25* 27.20*
.058 .3083 .5986 .11046 .12624 30.20 8.73* 3.12
.065 .3441 .5962 .12230 .13977 26.90 9.75 3.48 *AN STANDARD TUBING
.083 .4347 .5901 .1fi36 .17299 21.10 12.32 4.40
C4.10 IN TENSION COMPRESSION BENDING TORSION AND COMBINED LOADINGS
Fe = 79500 - 51.9 {L' /p)~'~ L' = L/...;G = 40/v'2: = 28.4

= 79500 - 51.9 {58.7)~·~ = 56500 psi In Fig. C4.5 we can use L = 40 and c = 2
scale at bottom of chart or use c 1 scale =
The short column equation applies since, and L' = 28.4. Reading the chart we obtain
as shown in Table C4.1, the transitional L'/p Pa = 9200 Ibs. Thus the c = 2 fixity
is 91 and the value for our tube is 58.7. increased the strength of the tube from 6000
to 9200.
Tensile Strength
Case 4. Same as Case 3 but heat treated to
Since the tube is welded, the tUbe Ftu = 150,000 after welding.
material adjacent to the weld is weakened.
The weld correction values are given in Table L'/p = 28.4/.425 = 66.8
C4.5. We will assume a weld other than
tapered. Let Pa ~ allowable or failing tensile From Fig. C4.2 using 150,000 curve, we
strength of tube. read Fc =
63000, whence
Pa = Ftu (weld factor) (area of tUbe) Pa = FcA = 63000 x .1849 = 11650 lb.
= 95000 x .841 x .2628 = 21000 Ibs. In this case heat treating produced
additional strength, whereas in Case 2 it did
M.S. = {Pa/p)-l = (21000/18500) - 1 = not. The reason for this is that failure
0.13, thus compression is critical. occurs in the inelastic stress range and heat
treating raises the material propertiBs in the
PROBLEM 2 inelastic range. The end fixity changed the
column from a so-called long column to a short
Case 1. Tube size 1-1/4 - .049, L = 40 in. column.
c =1
Material: Alloy steel, Ftu = 95000 The strength could be found also by sub-
Find ultimate compressive load it will stituting in the short column equal for
carry. 150,000 steel as given in Table C4.1,
Solution: From Fig. C4.5, Pa = 6000 Ibs. Fc = 145000 - 18.36 {L'/p)2

Case 2. If tube was heat treated to Ftu = = 145000 - 18.36 {66.8)2 =63000 psi.
150,000, what compressive load would
it carry. PROBLEM 3
Solution: Fig. C4.5 cannot be used since Case 1. Tube size 2 - .065, L 24, c = 1.5 =
Ftu = 150,000, thus we will use Fig. C4.2. Material Ftu = 95000. Welded at ends.
L' = L/...;G = 40/VT = 4. From Table C4.3, Ultimate design load = 25000 Ibs.
p = .425 and area (A) .1849. = What is M.S.
L'/p = 40/.425 = 94. From Fig. C4.2, Solution: L' = L/ve : 24/~ = 19.7
using the 150,000 curve, we find Fc 32500. = = =
Then Pa = FcA =
32500 x .1849 = 6000 lb. Thus From Fig. C4.5 for L 19.7 on c 1
scale, we project upward to the 2 inch tube
heat treating the tUbe from 95000 to 150,000
for Ftu did not increase the column strength. and note that it intersects the horizontal
For a L'/p = 94, it is a long column and weld cut-off line which gives an allowable
failure is elastic and E is constant. column load at left scale of Pa = 26700 lb.
Failure in this case is local crippling
The strength could also be calculated by adjacent to welds at the tube ends.
Euler's equation from Table C4.I.
M.S. = Pa / p = 26700/25000 - 1 = .07
Fc =286,000,000/{L'/p)2 Case 2. Assume tUbe is heat treated to Ftu
= 286,000,000/(94)2 = 32500 psi, the = 125000 after welding. What is tube
same as previously calculated.------ strength.
Case 3. Same as Case 1, but assume tube is L'/p = 19.7/.6845 = 28.8

welded to several other tUbes at its
end and that the end fixity developed Using Fig. C4.2 With L'/p = 28.8
and pro-
is c = 2. jecting up to 125000 curve, we again note that
horizontal weld cut-off line is intersected
C4.11
giving Fc = 95000, whence Pa = 95000 x .3951 The results show that 1-1/4 - .035 is the
= 37500 Ibs. lightest. Since there is danger in welding
.035 thickness to the other heavier tUbe gauges
If the tube had not been welded at ends particularly the engine mount ring which is
the dashed part of the column curve could have usually relatively heavy for this size engine,
been used, thus giving additional strength. a minimum tUbe thickness of .049 will be used,
hence the 1-1/8 - .049 tube will be selected.
PROBLEM 4
Consider Member (2)
Fig. C4.7 shows a steel tubular ,engine
mount structure for a 1050 H.P. radial engine. Design loads 11650 tension and 4250
The ultimate design tension and compressive compression. Since the tension load appears
load in each member as determined from a stress critical, the tube will be designed for the
analysis for the various flying and landing tension load and then checked for the
conditions are shown in ( ) adjacent to each compressive load. The Ftu of the material
member. The true length L of each member is equals 95000 psi. Since the engine mount in a
also shown. Using chrome-moly steel tUbes, welded structure, the strength of the tube ad-
Ftu = 95000, select tube sizes for the given jacent to the end welds must be reduced to .841
loads. It is common practice to assume the x 95000 = 80000 psi (see Table C4.5).
column end fixity c = 1 for engine mount
members, since the mount is subjected to Hence tube area required = 11650/80000 =
considerable vibration and the true rigidity 0.146 sq, in. From Table C4.3, which gives
given by the engine mount ring is difficult the section properties of round tUbes, we
to accurately determine. select the following sizes:
1- .049, Area = .146, M.S. = ( .146/ .146) - 1 = 0

1- .058, Area = .172, M.S. = ( .172/ .146) - 1 = .19
1-1/8 .049, Area = •166, M.S • = .14
1-3/8 •035, Area = .147, M.S . = 0
To obtain a reasonable margin of safety, the
1-1/8 - .049 will be selected.
Front View
Fig. C4.7
Many structural designers prefer to have
large margins of safety on engine mount members
as conSiderable trouble has been encountered in
Consider member (3). Ultimate design load = the failure or cracking of engine mount members.
- 9250. Referring to the column charts of
Fig. C4.5, we find for C = 1 and L = 31.4 the The strength of the 1-1/8 - .049 tube as a
following tube sizes for a strength near column for length = 31.4 and c = 1 equals -6700
- 9250 Ibs. Ibs. from Fig. C4.5 which gives a margin of
safety of (6700/4250) - 1 = .58 on the maximum
1-1/2 - .035, Pa = 8850 (weak). (Note - Pa = compressive load. The student should select a
allowable load). tube size for member (1).
1-3/8 - .049, Pa - - 10350,weight=5.78 Ib./100", C4. 14 TIlustrative Problems Using Aluminum Alloy
M.S. = (10350/9250) - 1 = .12 and Magnesium Round Tubes as Columns and
Tension Members.
1-1/4- .058, Pa = -10000, weight = 6.15, M.S.
= 10000/9250) - 1 = .08 In general alloy steel round tUbes must
be heat treated to around 180,000 to 200,000
Thus use 1-3/8 - .049 since it is the lightest ultimate tension strength before they can com-
as well as the strongest. pare favorably with aluminum round tubes on a
material weight basis. However, aluminum
Consider member (4), Load = - 5470, L = 30, alloy as used for tubes cannot be welded
c 1= satisfactorily and thus in a truss structure
the end connections involving riveted and
From Fig. C4.5: bolted connections add weight and design
difficulties as compared to welded connections
1-1/8- .049, Pa = -7100, wt.= 4.68, M.S.= .30 in steel trusses.
1-1/4- .035, Pa = -6500, wt.= 3.78, M.S.= .19
1" - .058, Pa = -6000, wt.= 4.86, M.S.= .10

PROBLEM 1 C4.15 Strength of Streamline Tubing.
Case 1. Tube size 1 - .049 round. If a round tube is exposed to the air-
L = 24, c =
1, Material 2024-T3 stream, the air drag is about 15 times greater
Find failing compressive load. than if it were given a streamlined shape, thus
streamline tubes are used when the member is
Solution: The col1nnn curves in Fig. C4.6 are exposed to the airstream.
slightly conservative because the equation used
was slightly different from the equCLtion now Streamline tubes are drawn from round
specified in (Ref. 1). tubes. In designating a streamline tUbe, the
round tube from which it was made is used and
Use L = 24, we read for 1-049 tube a then the fineness ratio is also given. The
failing load of 2800 lb. fineness ratio is the ratio L/D, which dimen-
sions are shown in Fig. C4.8. The most common
As a second solution, we will use Fig. fineness ratio used is 2.5 to 1. Table C4.6
C4.3. L' = L/y'C: = 24/v-r- = 24. L'/P = shows the section properties of streamline
24/.3367 =
71.3. From Fig. C4.3, we read tubing haVing a fineness ratio of 2.5 to 1.
Fe = 20000. Then Pa FcA = =
20000 x .1464 = Figs. C4.9 and C4.10 give curves for finding
2930 lb. the column failing stress Fe and the local
crushing stress Fcc.
The answer could be obtained by substi-
tilting in equation C4.9, Fig. C4.8
Fe = 50600 431 L'/p B = . 944L R = . 19D

= 50600 431 (71.3) = 19900 D = • 5714d L = 2. 5D
d = Equiv. Round Dia.
Pa 19900 x .1464 = 2920 lb.
A column may also fail by local crushing PROBLEM 1. Case 1.

or crippling of the tube wall, thLlS the
crushing stress Fcc should be determined to A streamline tube made from a basic round
see if it is less than the primary bending tUbe of 2-1/2 - .065 size has a fineness ratio
failing stress for the column. of 2.5 to 1. The length L is 30 in. Take
c = 1. Material is alloy steel Fty 75000. =
For our tube the diameter over thickness Find the ultimate compressive load the member
ratio D/t = LO/ .0'19 = 20.40. Values of D/t will carry.
are given in Table C4.3.
Solution: From Table C4.6 for 2-1/2 - .065
Referring to the small chart in the u~per size we find the follOWing section properties:-
right hand corner of Fig. C4.3, we find for a Area (A) = .4972, P (major axis) = .5137 in.
D/t of 20.4 that Fcc =
47500 psi. Since this Then L' = L/V-C- =
30/v-r- = 30, and L'/p =
stress is greater than the bending failing 30/.5137 = 58.5
column stress of 20,000, it is not criticaL D/t value for tube = 2.5/.065 38.5 =
Case 2. Same as Case 1 but use c = 1.5 and From Fig. C4.9 for L'/p = 58.5
and D/t =
chan~e material to 6061-T6 alwninum 38.5, we read Fe = 46500 psi. For D/t 38.5 =
alloy. and reading from small chart in upper right
hand corner of Fig. C4.9, we read Fcc 66500. =
L' = L/V-C:= 24/~ = 19.7, L'/p = 19.7/.3367 Thus Fc is critical and Pa =
46500 x .4972 =
23000 lb.
Case 2. Same as Case 1 but change material

From Fig. C4.3, Fe = 22500 to 2024-T6 aluminum alloy.
Whence Pa = 22500 x .1464 = 3300 lb. For this material we use Fig. C4.10.
For L'/p = 58.5, we read Fc 26000 psi.
Fcc for D/t = 20.4 from Fig. C4.3 = 38500 (not
For D/t = 38.5, we read Fcc = 37500 (not
critical) . critical). Thus Pa = 26000 x .4972 = 12900 lb.
Case 3. Same as Case 2 but change material to C4. 16 Strength of Oval and Square Shaped Tubes
magnesium alloy, Fcy = 10,000. in Compression.
For L'/p =
58.5 and using lower curve on Tables C4.8 and C4.7 give the section
Fig. C4.4, we read Fc = 7600. Then Pa FcA = = properties for square and oval shaped tubes
7600 x .1464 = 1110 lb. respectively. For the design of these shaped
C4.13
Streamline ·TABLE C4. 6 SECTION PROPERTIES OF STREAMLINE TUBING (Fineness Ratio 2. 5to 1)
Equiv. Wall Axes Wt. I Equiv. Wall Axes
Round Z p I Z p
Area per fr. Major Major Major Round Wt. I Z p I Z P
O.D. Oec'l Gage Major Minor Minor Minor Minor Minor Area per ft. Major Major Major Minor Minor
O.D. Dec'l Gage Major Minor
'l4 .035 20 1.0114 .4286 .0786 .2673 .0017 .0079 .1486 .0073 .0140 .3046 2% .049 18 3.3713 1.4285 .3773 1.283 .1018 .1425 .5194 .4063 .2343 1.0377
.049 18 .1079 .3668 .0022 .0103 .1435 .0097 .0187 .2998 .058 17 .. ..
..
.4450 1.513 .1186 .1660 .5163 .4764 .2747 1.0346
)08 .035 20 1.1800
.. .5000 .0924 .3140 .0028 .0112 .1753 .0118 .0195 .3572
.065 16
.. .4972 1.690 .1312 .1837 .5137 .5300 .3057 1.0325
.049 18 .. .1272 .4323 .0037 .0148 .1703 .0158 .0261 .3525
.083 14 .. .. .6302 2.143 .1621 .2269 .5071 .6640 .3832 1.0265
.058 17
.. .1489 .5061 .0042 .0168 .1671 .0182 .QJOI .3495
.095
.120
13
11 .. .. .7178
.8972
2.440
3.050
.1815
.2188
.2541 .5028
.3063 .4938
.7504
.9226
.4333
.5329
1.0224
1.0141
1 .035 20 1.3485 .5714 .1061 .3607 .0043 .0151 .2020 .0178 .0257 .4100 2}i .049 18 3.7085 1.5714 .4158 1.414 .1366 .1739 .5732 .5434 .2848 1.1432
.049
.058
18
17
.. .. .1464
.1716
.4977
.5835
.0057
.0065
.0200
.0228
.1972
.1939
.0239
.0278
.0345
.0402
.4044
.4023
.058 17 .. .. .4905 1.668 .159.3 .2028 .5699 .6372 .3340 1.1398
.065 16
.. .. .5483 1.864 .1765 .2247 .5674 .7099 .3723 1.1379
17S .035 20 1.5170 .6428 .1199 .4074 .0063 .0196 .2292 .0257 .0329 .4626
.083 14
.. .. .6954 2.364 .2188 .2785 .5609 .8907 .4673 1.1318
.049 18
.. .. .1656 .5631 .0083 .0258 .2240 .0347 .0445 .4573
.095
.120
13
11
.. .. .7924
.9915
2.694
3.371
.2455
.2973
.3125
.3785
.5566
.5476
1.0081
1.2427
.5289
.6523
1.1279
1.1196
.058
.065
17
16
.. .. .1944
.2165
.6609
.7359
.0094
.0103
.0292
.0320
.2205
.2182
.0402
.0443
.0516
.0569
.4549
.4525 3 .058 17 4.0455 1.7142 .5361 1.822 .2085 .2433 .6236 .8318 .3997 1.2456
1}4 .035 20 1.6857
.. .7143
.. .1336 .4542 .00118 .0246 .2559 .0355 .0409 .5154
.065
.083
16
14
.. ..
..
.5993
.7606
2037
2586
.2312
.2872
.2698
.3352
.6211
.6145
.9265
1.1645
.4452
.5599
1.2434
1.2373
..
.049 18
.. .. .1849 .6285 .0116 .0325 .2509 .0482 .0557 .5107 .095 13 .. .. .8670 2.947 .3229 .3768 .6103 1.3184 .6338 1.2332
.058
.065
17
16
.. .. .2172
.2420
.7384
.8226
.0133
.0145
.0372
.0406
.2477
.2451
.0560
.0618
.0647
.0714
.5077
.5053
.120
.156
11
%~
.. .. 1.0857
1.)959
3.691
4.746
.3925
.4824
.4580
.5630
.6013
.5879
1.6289
2.0497
.7839
.9869
1.2249
1.2118
1% .035 20 1.8543 .7857 .1473 .5009 .0118 .QJOO .2828 .0476 .0549 .5683 3)4 .058 17 4.3828 1.8571 .5816 1.977 .2668 .2873 .6773 1.0617 .4708 1.)511
.049 18 .. .. .2041 .6939 .0157 .0400 .2777 .0648 .0680 .5635 .065 16
.. .. .6504 2.211 .2961 .3189 .6747 1.1830 .5248 1.)487
.058 17 .. .. .2400 .8158 .0181
.0198
.0461 .2745
,2.719
.0754 .0791 .5604 .083 14
.. .. .8258 2807 .3689 .3973 .6684 1.4886 .6604 1.)426
.065 16 .2675 .9094 .0504 .0833 .0875 .5581 .095 13 .. .. .9416 3201 .4150 .4470 .6639 1.6873 .7489 1.)386
111 .035 20 2.0228

.. .8571 .1611 .5476 .0154 .0359 .3096 .0621 .0597 .6209
.120
.156
11
%2
.. .. 1.1800
1.5186
4.011
5.163
.5062
.6251
.5452
.6732
.6549
.6416
2.0887 .9275
2.6353 1.1707
1.)304
1.3173
.049 18
.. .. .2234 .7593 .0207 .0483 .3045 .0848 .0815 .6162
..
.058 17 .. .. .2627 .8932 .0239 .0555 .3014 .0988 .0951 .6131 311 .065 16
..
4.7200 2.0000 .7014 2.385 .3723 .3723 .7286 1.4831 .6108 1.4541
.065 16
.. .. .2934 .9962 .0262 .0611 .2989 .1093 .1052 .6108 .083 14
.. ..
..
.8910 3.029 .4645 .4645 .7220 1.8686 .7699 1.4481
.083 14 .3695 1.256 .0316 .0737 .2923 .1351 .1302 .6047 .095 13 .. 1.0162 3.455 .5234 .5234 .7177 2.1194 .8733 1.4442
1% .035 20 2.1913
.. .9285 .1748 .5943 .0198 .0426 .3365 .0793 .0704 .6736
.120
.l56
11
"
/32
.. .. 1.2742
1.6414
4.332
5.580
.6400
.7934
.6400
.7934
.7087
.6952
2.6269 1.0828
3.3221 1.3705
1.4358
1.4227
.049 18 .. .. .2426 .8248 .0266 .0573 .3314 .1085 .0964 .6689
16 5.0570 2.1428 .7525 2.558 .4604 .4297 .7822 1.8299 .7033 1.5594
.058 17 .. .. .2855 .9707 .QJ08 .0448 .3282 .1266 .1124 .6659 3)4 .065 ..
..
.065 16 ..
..
.. .3186 1.083 .OJJ8 .0728 .3256 .1402 .1245 .6634 .083 14
..
.. ..
.9562
1.0908
3.251
3.708
.5754
.6490
.5371
.5777
.7757
.7713
2.3073
2.6195
.8871
1.0075
1.5534
1.5497
.083
.095
14
13
.. .4021
.4566
1.367
1.552
.0410
.0453
.0883
.0976
.3192
.3149
.1738
.1949
.1545
.1732
.6574
.6541
.095
.120
13
11
.. .. 1.)685 4.652 .7955 .7425 .7624 3.2505 1.2502 1.5412
.156 %2 1.7641 5.997 .9894 .9235 .7489 4.1192 1.5855 1.5281
1~ .035 20 2.3600 1.0000 .1886 .6411 .0249 .0498 .3633 .0995 .0820 .7263
2.2856 .8035 2.732 .5614 .4912 .8359 2.2270 .8025 1.6648
.049 18
.. .. .2618 .8902 .0336 .0672 .3583 .1363 .1123 .7216 4 .065 16 5.3940
..
..
.. 1.0214 3.472 .7024 .6146 .8293 2.8161 1.0148 1.6604
.058 17 .. .. .3083 1.048 .0389 .0778 .3551 .1592 .1312 .7186 .083
.095
14
13 .. .. 1.1655 3962 .7933 .6941 .8250 3.1917 1.1506 1.6549
.065 16
.. .. .3441 1.170 .0428 .0856 .3525 .1765 .1455 .7163
.120 11
.. .. 1.4627 4.973 .9739 .8522 .8160 3.9661 1.4303 1.6467
.083 14 .. .. .4347 1.478 .0521 .1042 .3461 .2192 .1809 .7101
.. .. 6.414 1.2153 1.0634 .8026 5.0332 1.8164 1.6333
".
.095 13 .4939 1.679 .0577 .1154 .3418 .2463 .2034 .7061 .156 %2 1.8867
.188 2.2457 7.635 1.4109 1.2345 .7926 5.9364 2.1431 1.6259
1)08 .035 20 2.5285 1.0714 .2023 .6878 .0308 .0575 .3902 .1227 .0943 .7789
.065 16 5.7313 2.4285 .8546 2.905 .6761 .5568 .8895 2.6784 .9082 1.7704
.049 18 .. .. .2811 .9556 .0417 .0779 .3851 .1685 .1296 .7743 4)4
.083 14
.. .. 1.0866 3.694 .8473 .6978 .8830 3.3817 1.1471 1.7641
.058 17 .. .. .3311 1.126 .0482 .Q900 .3815 .1969 .1515 .7712
.095 13 .. .. 1.2400 4.216 .9574 .7885 .8787 3.8426 1.3035 1.7604
.065 16 .. .. .3696 1.257 .0532 .0993 .3794 .2186 .1683 .7690
.120 11
.. .. 1.5570 5.293 1.1779 .9701 .8698 4.7789 1.6216 1.7519
.083 14 .. .. .4673 1.589 .0650 .1214 .3729 .2719 .2093 .7624
..
".
6.0748 2.0628 1.7387
.095 13 .5312 1.806 .0722 .1348 .3687 .3598 .2770 .7523 .156
.188
%2 .. 2.0095
2.3930
6.832
8.135
1.4729 1.2131
1.7144 1.4120
.8573
.8464 7.1732 2.4365 1.7314
2 .035 20 2.6970 1.1428 .2161 .7345 .0376 .0658 .4170 .1494 .1076 .8313 .7860 4.0256 1.2894 1.8696
411 .083 14 6.0683 2.5713 1.1517 3.915 1.0105 .9367
.049 18
.. .. .3003 1.021 .0510 .0893 .4120 .2053 .1480 .8269
.095 13 .. .. 1.)147 4.469 1.1429 .8889 .9324 4.5759 1.4662 1.8656
.. ..
.058
.065
.083
.095
17
16
14
13
..
..
..
..
.3539
.3951
.4999
.5685
1.203
1.)43
1.699
1.933
.0591
.0652
.0799
.0889
.1034
.1141
.1398
.1556
.4088
.4062
.3998
.3954
.2402
.2668
.3325
.3744
.1732
.1925
.2399
.2703
.8238
.8217
.8156
.8115
.120
.156
.188
11
%,
%.
..
..
..
..
..
..
1.6512
2.1322
2.5403
5.613
7.249
8.636
1.4082 1.0953
1.7650 1.3728
2.0580 1.6007
.9235
.9098
.9001
5.6952
7.2508
8.5740
1.8254
2.3247
2.7498
1.8572
1.8441
1.8372
.219 %2 2.9421 10.Q! 2.3221 1.8061 .8884 9.8001 3.1451 1.8251
2X .049 18 3.0343 1.2857 .3388 1.152 .0735 .1144 .4657 .2946 .1887 .9325
.058 17 .. .. .399~ 1.358 .0854 .1329 .4625 .3450 .2212 .9294
.065 16
.. .. .4462 1.517 .0944 .1469 .4600 .3835 .2458 .9271
80
LLU-LW.~'-
o .: 80,000
.083 14
.. .. .5650 1.921 .1161 .1807 .4533 .4794 .3075 .9223 ...;
.095
.120
13
11
.. .. .6432
.8030
2.186
2.730
.1298
.1556
.2020
.2421
.4492
.4402
.5409
.6629
.3470
.4255
.9170
.9086
rn
P<
j-
+
~ ~~-
c :I::
rz. D/t 20 rn
~
0 70
rn ~ rn 70,000
&:l u rn
riI
Z 30 I
OJ p:;
4; OJ
rn Eo-<
60 ~ 50
:::> 60 rz. rn 60,000
li
\
40
0
~ 1
:I:: "-n1J
til 4 If
50~
Eo-< 50
~ '"
t:
"0-, ~
10 ~ 50,000
V> ::> rn 50
40f----203O~ rn 0 25 50
V>
U, 30 f
~ Iitt', I I IIIII
""'-
riI "
til p:; D/ of Basic Round
,
U
t:
D/t 4 u. U
t
~I Eo-< Tube
E
.2
30
"'" 20
0 2040 60 80 100
rn 40
z
..
0
"'" I
u D/t t
~
:0 20 :::>
0
........ ..:l
3 0
51
<i 10 ~ U
riI
30
NOTE: Higher values of allowable
u.U """'"- t--- ..:l stress can be used in short column
D' BaSIC round tube diameter
I:l:I range if substantiated by tests
0 I I II 4;
20
0 20 40 60 80 100 120 140 :::
0
"-
L'lp ..:l Fig. C4.9 tHttltt
..:l ~
<>::
I 10 ALLOWABLE COLUMN AND CRUSHING STRESSES
Fig. C4.10 Streamline 2024-T3 Tubing OJ CHROME MOLYBDENUM STREAMLINE TUBING.
rz. F ty = 75000 PSI
0
o 20 40 60 100 120 140
Squares TABLE C4. 8 SECTION PROPERTIES OF STREAMLINE TUBING

Size Wall W.Jl Wt.
Dist. Area
Wt. z
Side Area
per ft. z
Dec'I Gage
A,,,,,, Dec'I Gage per ft.
Flats
1~ .035 20 .1926 .6549 .0661 .0881 ·5859
.022 24 .0294 .0998 .0006 .0031 .1415 .049 18 .2675 .9095 .0902 .1203 ·5807
.028 22 .0368 .1252 .0007 .0038 .1396 .058 17 .31SO 1.071 .1051 .1401 ·5776
.035 20 .0453 .1589 .0009 .0045 .1372 .065 16 .3516 1.195 .1162 .15SO ·57SO
.049 18 .0612 .2081 .001I .0057 .1318 .083 14 .4443 1.510 .1436 .1915 ·5686
.022 24 .0396 .1348 .0016 .0065 .2023 .095 13 .S050 1.717 .1608 .2144 .5643
.028 22 .0499 .1697 .0020 .0079 .1991 .120 11 .6284 2.136 .1940 .2586 ·5556
.035 20 .0616 .2096 .0024 .0094 .1956
.049 18 .0841 .2861 .0030 .0120 .1892 1~ .049 18 .3134 1.065 .1451 .1658 .6804
.058 17 .0980 .3330 .0033 .0134 .1853 .058 17 .3693 1.255 .1691 .1932 .6766
.065 16 .1084 .3684 .0035 .0141 .1804
.065 16 .4124 1.402 .1877 .2146 .6747
.022 24 .0483 .1641 .0033 .0104 .2600 .083 14 .5220 1.774 .2331 .2663 .6682
.028 22 .0630 .2143 .0040 .0129 .2526 .095 13 .5938 2.019 .2618 .2992 .6640
.035 20 .0780 .2652 .0048 .0155 .2489 .120 11 .7307 2.484 .3175 .3628 .6592
.049 18 .1071 .3640 .0063 .0201 .2420
.058 17 .1251 .4253 .0071 .0227 .2379
16 .0077 .0245 .2348 .049 18 .3592 1.221 .2180 .2180 .7790
.065 .1388 .4717
.058 17 .4236 1.440 .2557 .2557 .7769
.028 22 .0761 .2588 .0064 .0168 .2890 .065 16 .4732 1.609 .2838 .2838 .7744
.035 20 .0944 .3209 .0078 .0206 .2866 .083 14 .5996 2.038 .3536 .3536 .7679
.049 18 .1300 .4419 .0103 .0274 .2816
.058 17 .1522 .5175 .0118 .0313 .2784 .095 13 .6827 2.321 .3981 .3981 .7636
.065 16 .1692 .5751 .0129 .0342 .2760 .120 11 .8530 2.900 .4859 .4856 .7548
.083 14 .21I3 .7184 .0154 .0408 .2700 .156 'h 1.0912 3.710 .6010 .6010 .7422
.028 22 .0892 .3033 .0103 .0236 .3390 .3080
20 .1I08 .3766 .0125 .0288 .049 18 .4051 1.377 .2738 .8721
.035 .3364
.049 18 .1528 .5193 .0168 .0386 .3320 .058 17 .4778 1.624 .3607 .3207 .8689
.058 17 .1794 .6098 .0193 .0445 .3283 .065 16 .5341 1.816 .4000 .3556 .8654
.065 16 .1996 .6785 .0212 .0487 .3528 .083 14 .6773 2.302 .5007 .4451 .8598
.083 14 .2502 .8504 .0255 .0588 .3194 .095 13 .7716 2.623 .5640 .SOB .8549
.028 22 .1023 .3479 .0154 .0309 .3883 .120 11 .9652 3.281 .6918 .6149 .8466
.035 20 .1271 .4322 .0190 .0380 .3867 .156 li• 1.2372 4.206 .8765 .7791 .8418
.049 18 .1758 .5978 .0256 .051I .3813
.058 17 .2065 .7020 .0295 .0591 .3782 .058 17 .5321 1.809 .5074 .4035 .9757
.065 16 .2300 .7819 .0324 .0649 .3756 .065 16 .5949 2.022 .5643 .4488 .9739
.083 14 .2890 .9824 .0394 .0788 .3693
.3272 U12 .0436 .0873 .3652 .083 14 .7549 2.566 .7071 .5625 .9678
.095 13
.095 13 .8605 2.925 .7981 .6530 .9631
.035 20 .1435 .4878 .0275 .0489 .4340 .120 11 1.0775 3.663 .9809 .7806 .9541
.049 18 .1988 .6757 .0369 .0657 .431I .156 1.3831 4.702 1.2256 .9760 .9413
.058 17 .2336 .7942 .0427 .0759 .4265 'h
.065 16 .2604 .8852 .0471 .0838 .4241 .188 H• 1.6479 5.602 1.4257 1.3183 .9301
.083 14 .3278 1.1I4 .0576 .1024 .4191
.095 13 .3716 1.263 .0640 .1I38 .4151 .058 17 .5864 1.993 .6792 .4940 1.0763
.120 II .4600 1.564 .0760 .1351 .4064 .065 16 .6557 2.229 .7559 .5497 1.0737
1" .035
.049
.058
20
18
17
.1599
.2217
.2608
.5435
.7536
.8865
.0378
.0513
.0595
.0604
.0821
.0952
.4891
.4831
.4778
.083
.095
.120
14
13
11
.8326
.9494
1.1898
2.830
3.227
4.045
.9481
1.0724
1.3213
.6896
.7799
.8653
1.0671
1.0628
1.0541
.065 16 .2908 .9886 .0657 .1051 .4753 .156 li. 1.5291 5.198 1.6569 1.1121 1.0410
.083 14 .3666 1.246 .0806 .1290 .4960 .188 H• 1.8238 6.200 1.9334 1.3264 1.0296
.095 13 .4161 1.414 .0899 .1438 .4647
.120 II .5161 1.755 .1074 .1718 .4S61 .9860
.065 16 .7166 2.436 .6573 l.l730
1% .035 20 .1763 .5992 .0506 .0736 .5360 .083 14 .9102 3.094 1.2394 .8263 l.l669
.049 18 .2446 .8316 .0689 .1003 .5309 .095 13 1.0383 3.530 1.4032 .9355 l.l625
.058 17 .2879 .9787 .0801 .1I66 .5276 .120 11 1.3021 4.426 1.7325 1.1550 l.l544
.065 16 .3212 1.092 .0886 .1288 .5251
.083 14 .4055 1.378 .1091 .1587 .5187 .156 li. 1.6750 5.694 2.1793 1.4529 l.l406
.095 13 .4605 1.S66 .1219 .1773 .5145 .188 H• 1.9997 6.798 2.5S03 1.7002 l.l381
.120 II .5723 1.946 .1464 .2130 .5058 .219 H. 2.3081 7.847 2.8873 1.9249 l.l184
00
r:il
....
Eo<
~
r:il
Po.
o
~(
o.} ~ ~~ ~~ ~~ ~~~ ~~~ ~~~~ ~~~ ~~~ ~~~~ ~~~~~ ~~~~ ~~~~~ ~~~~~ ~~~~~~ ~~~%~~ ~~~~~~
Po. z ,:ll--t-------------------------------------i
Z ....
01Il
.... ::>
Eo< Eo<
U,.J
r:il..:
00>
<-0
"':r..
uO
~: ....
r:il
,.J
oi
III
..: ~: ~:
Eo<
.
~ ~?a~ ~~ ~~ ~~~ ~~~ ~~~~ ~~~ ;!::!::;;e ~::;:!::~ ~::;:!::~~::;:s~:::;'!:;:!::;!;";::;';::: :!::;!;"~:::ooI:~;!;.::~-{:oo!=:!::;!;~:::_="!::!::;!;~;:::ooI:~
~I-t-----------------------------------------;
~ ~ ~~ ~~ ~~ ~~~ ~~~ ~~~~ ~~~ ~~~ ~~~~ ~~~~~ ~~~~ ~~~~~ ~~~~~ ~~~~~~ ~~~~~~ ~~~q~~
C4.15
tubes the primary column strength can be found C4. 18 Problems InvolVing Bending Strength of Tubes.
by using the curves in Figs. C4.9 and C4.10.
The crushing stress Fcc for oval shaped tubes PROBLEM 1
can conservatively be taken as that for
streamline tubes as given in Figs. C4.9 and A 1-1/4 - .058 round tube is used as a
C4.10. For square tubes the local crushing simply supported beam with the supports at the
stress Fcc can be taken as the crippling stress ends. The span or length of the beam is 24
of a flat plate. For this stress refer to the inches. It carries a uniform distributed load
chapter which covers the buckling and crippling w in pounds per inch. Find the value of w to
stress of flat plates with various widths, cause the tube to fail in bending if the tube
thickness and boundary edge conditions. is made from the following materials:- alloy
steel Ftu =
95000, 6061-T6 aluminum alloy, and
ULTIMATE BENDING STRENGTH 6AL-4V titanium.
OF ROUND TUBES Solution: The maximum bending moment occurs

at the midpoint of the span and equals wL 2 /8 =
C4. 17 Charts for Finding Modulus of Rupture Stress. w x 24 2 /8 =
72 w in. lb.
Chapter C3 was concerned with the theory The section properties for the tube are
and methods of determining the ultimate yield D/t =21.55, and I/y =
.06187 obtained from
and failing stress of a section in pure Table C4.3.
bending. It was concerned with finding a
fictitious stress Fb which, when substituted The beam equation involving the modulus
=
in the well known beam formula M FbI/c, would of rupture Fb is M = FbI/y. SUbstituting 72w
give the value of the bending moment which for M, we obtain:
would cause failure.
FbI
w = 72y
Consider the alloy steel, Ftu =
95000. From
Fig. C4.11 for D/t = 21.55, we read Fb =
=
117000. Then w (117000 x .06187)/72 100.2 =
lb. per inch. Consider tube material as -----
6061-T6 which has a Ftu = 42000 psi.
From Fig. C4.12 for D/t =
21.55, we read
that Ft/Ftu = l.06. Thus Fb = 42000 x l.06 =
44500.
Then w =
(44500 x .06187)/72 =38.1.
Tube material 6AL-4V titanium. ----
\ From Fig. C4.14 for D/t = 21.55, we read

\ •
Fb = 191600. Then w =
(191600 x .06187)/72 =
\ .. _.- 118.7 Ibs. per in.
PROBLEM 2
A beam simply supported at its ends has
a span of 30 inches. The ultimate design load
consists of two equal loads of 2000 Ibs. each.
The beam is symmetrically loaded with each
load located 12 inches from the ends.
Select the lightest round tUbe when made

from the following material: (1) alloy steel
Ftu =220000, (2) 7075-T6 aluminum alloy.
Compare the resulting tube weights.
10 15 20 25 30 35 40 45 50 Solution: The maximum bending moment is

O/t
constant and occurs between the load points.
Fig. C4. 14 Bending modulus of rupture for round M = 2000 x 12 =24000 in. lb.
6A1-4V tubing. (Titanium)
1.7
1.6 ~
15 ~\
1.4
\~0..
\\ ~
_- -...
3
',.~
12 ,,
'" ..........
-. .. 450
.... .... --
u..&>':
I I
I ...
.... """ ............... ........ 6061-T6
[1
1.0
t--... . . . ---i.. . . . . . . _ 400
IT
0.9
~-T6ond7075-T6
0.8
0.7
2024-T3,T4 -- ..~ \
"\
0.6
0 10 20 30 40 50 60 70 80 90
350
\ \. "' I'. r...
Oil
I'..
:\ ~
Fig. C4. 12 Bending modulus of rupture for aluminum-alloy round tubing. .....
] 300
I'.
..... ..... ......
I':. ....
" .....
....
r-..
I'
r-..
......~
...
'0 250
-.... t'--
-
- - -
.... 10.. I"'- Flu, 260 ksi
on
:>
:;
........ -10.. Ftu. 240 ksi
"0
Flu. 220 ksi
o I"-
:::E ....... Flu, 200 ksi
-r-.
\ '"
c 200
.
"\ " ........
... -r-. Flu. 180 ksi
- - --
"0 I'"
"iii 60 C
....
\\
.>r. CD
01' .... ...
Flu. 150 ksi
~ r-..
Q
:l
:l
a::
50
\'\
150
"'.....I".:
... ........ - -- Flu ,125 ksi
" ::::-----.---
'0 40
....
<Jl
:l
\\l~ Fey
30 ksJ
100 Flu ,95 ksi
Flu. 90 ksi
--------
. /
-~
"3 ..........
"8 30
25 ksi_
........ I I I
~............... ~
20 ksi
~ .............. I I
--
15 ksi
g' .......... 50
'i5 20 o 10 20 30 40 50 60
..........-. 10ksi -
~ ........ ~
c
ell
CD
10
- 5 ksi
Fig. C4. 11 Bending modulus of rupture for round alloy-steel tubing.
Oil
lL..o i
10 20 30 60 70 80 90
Fig. C4. 13 Bending modulus of rupture of magnesium-alloy round tubing.

C4.17
Since the allowable or failing bending Therefore we will select a tube from Table
stress is a function of D/t, and since we do C4.3 that has a I/y value in the region of .30
not have a tube size, the design or solution to .37.
procedure is by trial and error.
Try 2-3/4 - 058. I/y =
.3233, D/t = 47.4.
Observation of the modulus of rupture From Fig. C4.12, Fb/Ftu =
0.90. Then Fb = .90
curves show that as D/t increases Fb decreases. x 77000 =
69300 psi. Then Ma Fb I/y =69300 =
This is due to the fact that failure in bending x .3233 =
22400. This is less than the design
is a local failure and the thinner the wall and bending moment of 24000 so this tube is weak.
larger the diameter, the lower the buckling or
crushing stress. However, the larger the D/t Try 3-058. I/y =
.3868, D/t 51.7 =
value the greater the section modulus I/y of Fb/Ftu =
.885 , Fb =
.885 x 77000 68000 =
the tUbe, which means increasing bending =
Ma Fb I/y = 68000 x .3868 26300 =
resistance. Thus we have two influences which M.S. =
(26300/24000)-1 = .09.
act oppositely relative to effecting the bending
strength. A study of other tubes in Table C4.3
shows that no other tube would be lighter in
There are many ways of guessing a tube weight.
size for checking purposes. In this example
problem we will assume two values for D/t and Tube weight =
30 x .5361 x .101 1.70 =
see what I/y would calculate to be. The two lbs., as against 2.22 lbs. for the alloy steel
values of D/t will be 45 and 25. heat treated to 220,000. Thus aluminum alloy
tubes from a weight standpoint usually yield
Consider the material alloy steel Ftu=220000:- results better than most materials. This
conclusion applies to only low temperatures,
For D/t = 45 from Fig.C4.11, Fb=232000 below 2500 F, as aluminum alloys lose strength
Then I/y = M/Fb = 24000/232000 = .103 rapidly for temperatures above 250 0 to 3000 F.
For D/t = 25, Fb = 266000 The student should calculate the lightest
Then I/y = 24000/266000 = .089 titanium tube and the lightest magnesium tube
using Figs. C4.14 and C4.13 respectively and
Therefore we will refer to Table C4.3 and compare the weight results with the steel and
select tubes that have an I/y value near the aluminum as found above •
•089 to .103 range and then find their true
bending strength. Table (A) shows the ULTIMATE TORSIONAL
selection and the necessary calculations,
using Fig. C4.11. STRENGTH OF ROUND TUBES
C4. 19 Torsional Modulus of Rupture.
Table A
In Article A6.2 of Chapter A6, the torsion
Tube Size IIY Area D/t Fb 1n..I/y (Ma/M)-l
M=~u M. S. = formula for circular sections, f s = Tr/J, was
1-7/8 -.035 .09136 .2023 53.6 227000 20800 -0.13
derived. This equation assumes the maximum
shear stress on the cross-section of a round
1-3/4 -.049 .1083 .261835.7 248000 26400 +0.10 bar or tube does not exceed the proportional
1-5/8 -.049 .0928 .2426 33.15 250000 23200 -0.04 limit of the material, or the stress variation
is linear as shown in Fig. C4.15 and this
The lightest available tube with a situation exists under the flight vehicle limit
positive margin of safety is 1-3/4 - .049 and loads. Before a round bar made of ductile
its weight for a 30 inch length is 30 x .2618
x 0.283 = 2.22 lbs.
Consider the tube made from 7075-T6 aluminum
alloy material which has a Ftu = 77000.
From Fig. C4.12:-
For D/t =
60, Fb/Ftu = .84, thus Fb = .84 x
77000 = 64700
I/y =M/Fb = 24000/64700 = .37
Fig. C4.15 Fig. C4.16
For D/t = 30, Fb/Ftu = 1.045, thus Fb = 1.045 x
77000 = 80500 material fails in torsion, the shear stresses
I/y = 24000/80500 = .30 fall in the inelastic or plastic range and
C4.18 IN TENSION. COMPRESSION BENDING, TORSION AND COMBINED LOADINGS.
thus the internal shear stress distribution is 2~

Specification WW - T- 825
~
similar to that indicated in Fig. C4.16. The
shear stress-strain curve is similar in shape 'i 20 ""'- ~'36
I .r---t----1--i
kSI
to the tension stress-strain curve.and is ~

a: 1'----
equal to approximately 0.6 of the ordinates. '5
~
I~
Thus to find the ultimate internal torsional .,~

resisting moment, we can divide the cross- 0
::0 10
section into a number of circular elements. g
0
From the stress curve in Fig. C4.16, the stress .~
~
~
at the midpoints of the circular elements can ..?
be found. Multiplying the area of the element 0
0 ~ 10 I~ 20 2~ 30 3~ 40
by this stress times the distance of the
l},
element from the center axis gives the moment
developed, and adding up the moment of the Fig. C4.31 Torsional modulus of rupture for magnesium-alloy
shear force on all the circular elements gives round tubing.
the total allowable or ultimate internal
resisting moment Ta. will it develop if made of the following
materials: (1) alloy steel Ftu = 95000,
If this moment Ta is to be used in the (2) aluminum alloy 2024-T3.
torsion equation, we must replace the true
stress curve in Fig. C4.16 by a triangular Solution: D/t = 25. J/r = 2(I/y) = 2 x
stress curve With maximum value Fst which will 0.11948 - 0.2390. L/D = 16.25/1.625 = 10.
produce the same internal resisting moment as For alloy steel, Ftu = 95000, we use Fig.
the true stress curve. Thus C4.17 where for D/t = 25 and L/D 10, we read =
Fst/lOOO =
47.3. Thus Fst 47300. =
Ta = Fst J/r, or Fst = Ta r/J - - - - (C4.10)
Ta = Fst J/r = 47300 x 0.2390 = 11300 in. lb.
Fst is called the torsional modulus of
rupture stress. It is a fictitious stress, For 2024-T3 aluminum alloy, we refer to Fig.
being higher than the ultimate shear stress C4.26 where we read Fst = 29000.
of the material. TUbes SUbjected to torsion
usually fail by inelastic or plastic instability Then Ta = Fst J/r = 29000 x 0.2390 =
or by elastic instability if the D/t ratio is 6930 in. lb.
large. To obtain reliable values of Fst,
resort is usually made to tests, and since the Problem 2.
round tUbe or rod is the most efficient and
most available shape, much testing has been A round tube 10 inches long is to carry
done over the years thus design curves are an Ultimate torsional moment of 7000 in. lb.
readily available for round tubes of the most Select the lightest tube if made from aluminum
common flight vehicle materials. alloy 2024-T4 and alloy steel Ftu = 200,000
and compare the resulting weight of each.
C4. 20 Torsional Modulus of Rupture Curves.
Solution: Since the modulus of rupture
Figs. C4.17 to C4.24 inclusive present depends on D/t and L/D and since the tube size
curves for finding the modulus of rupture is unknown, we will use the trial and check
stress Fst for steel alloys from Ftu = 95000 approach. The design calculations can con-
to 260,000 psi. It should be noted that the veniently be made in table form as follows:
torsional strength is influenced by the D/t
and the L/D values of the tube. Figs. C4.25 Trial T _ M.S. =
Tube Area D/t J/r L/D Fst a - Ta
to C4.30 inclusive give curves for the various FstJ / r 1'-1
Size
aluminum alloys and Fig. C4.31 gives informa-
tion for one magnesium alloy. All of these Alum. Alloy 2024-T4 (Fig. C4.27)
curves were taken from (Ref. 1).
1-1/2 -.065 .2930 23.05 .2016 6.7 30000 6050 -. 14
The curves are based on a theoretical 1-3/4 -.058 .3083 30.20 .2525 5.7 28000 7060 .01
investigation by Lee and Ades (see Ref. 2), 1-7/8 -.049 .2811 38.25 .2500 5.3 26000 6500 -.07
and have been found to be in good agreement
with experimental results. 2.0 -.049 .3003 40.30 .2860 5.0 26000 7440 .06
C4.21 Problems lllustrating Use of Torsional Modulus

Alloy Steel Ftu = 200,000 (Fig. C4.21)
of Rupture Curves. 1 - .035 .1061 28.56 .0495 10 98000 4850 -.31
1 - .049 .1461 20.40 .0664 10 106000 7050 .01
Problem 1. A 1-5/8 - .065 round tube is 16.25
inches long. What Ultimate torsional moment 1-1/4 -.035 .1336 35.70 .0790 8 94000 7430 .06
I I I I I 1
150 ~u=200 ksi '-+- 1\
1\ "iu=220 ksi f-+-
150I
I'\. (")
I\. 'f"
r-...
I~
130 LID ~ LID
130 ......
o ...... I- 0
t--. r-t- 1-1- I- 1/4 :-.... r- 1/4
I'..: f' ~ l- 1'-1- 8 r-I-
110
...... .... .... I- .... 1/2
0
) 110 I- r-I-I- 1/2
"" 1"-1- t-I-- i""- 1-1-

§ I--- I- 1-1- l"- I- 1
) 90
t-.... ..... I--- 1-1- 1"-1-
l- 90 r.. r"-I- 2
.~~
1'-.
I
I"
I---
1"'-
I'-
l-
l-
l-
'" t-
0-3~
tzJ
70
70 IZO
z
t-.. 5
~0-3
1'-
" 0D::
50 " '" I'-
10 50
1'0 r-..
~
10
~Z~
I(")t::l
OtzJ
I t--. S::fJ)
I I- r"-I-... 'dO
I~Z
20
I 30 20
30
10 20 30 40 50 60 70 80 o 10 20 30 40 50 60 70 80 ~O
.... ":1
D/, oil @:xl
Fig. C4. 22 Torsional modulus of rupture -
~ g
Fig. C4.21 Torsional modulus of rupture - IJ:lZ
alloy steels heat treated to Ftu = 200 ksi. alloy steels heat treated to Ftu = 220 ksi. tzJt::l
Z~
t::lfJ)
190 "iu=240 ksi f--+- 190 ~u= 260 ksi - ~0-3
~
O:xl
tzJ
\
0-3>
-I- oS::
1\ 1'\
170 170 ~t'"
~
lo-
-- - LID I.. .
0 52
Z~
tzJ
150 LID 150 N' o >0
.\
I\. "- ...... t-....
e3~
"
-l-
0 (,,)t'"
r.... l"- t--. I-... l-
...... l- 1/4
I" I" 1'-1-... 1/2 I~~
- ....
130 130
t-.... I---
r-.. r..... r... 1/2 80
...... ....... ~ ~fJ)
l"- I-... tzJ.tl
80
110
I-- FI-
) 110 -""'~ t::lS;
...... ~
-
I--- r-....
) r-.... I- " I'
t"':xl
otzJ
I'
90
"
2
90 " t;0-3
5251
I, r.....
......
" " r-..
.......
I
·0
~52
70 5 70 5
I' 1'0
I' r-.... 1'0
50
r-.... 50 ....... 10
10 .....
30
o 10 20 30 40
Oil
50 60
I-....
70
~
80
20 30
o 10 20 30 40
0/,
50 60
j--.
70
- 80
20
Fig. C4.23 Torsional modulus of rupture - Fig. C4.24 Torsional modulus of rupture -
alloy steels heat treated to Ftu = 240 ksi. alloy steels heat treated to Ftu = 260 ksi.
C4.21
2014 e T6 FORCING so 10H-TJ TUBING

Ftu '65 kSi F'tu=64 ksi
lro .0
0
II.
112
'14
~
112
30 L2' 30
20 20
10
10
,. NOff: THE CURVE REptESfNl'ING MA11RIAl

FAILURE 11l'D~ IS COMPunD FOR
F .39.000, AND DOES NOT AllOW fOR
2.
10
2D
rm: POSSIBILITY OF RmUCEO sn(NGTH
• t~f'£lOlN~GETH~'~'''~T13N~GE'!l'3N~'E'E~fGf§f§iEiElE~f§3
• ,. 2. 30
0/.
. so . 7• 10
0
0 10 20 ,0 '0
0"
'0 60 70 .0
Fig. C4.25 Torsional modulus of rupture - Fig. C4.26 Torsional modulus of rupture -
2014-T6 aluminum alloy forging. 2024-T3 aluminum alloy tubing.
101A-,. ruetNG
Ftu '62 kSI
L'D
~
'14
.: '0
'14
112
,,,
1/2
2D
20
,.
10 ID
'0
,0
10
o
o
o '0 20 30 ..
010
so 60 70 BO
o 10 20
'0 .0 so 60 70 '0
D/I Fig. C4.28 Torsional modulus of rupture -
6061-T6 aluminum alloy tubing.
Fig. C4.27 Torsional modulus of rupture -
2024-T4 aluminum alloy tubing.
7075-T6 ROLLED RO~} 7075·T6 FORCING

Flu '17ks; F1u '75 ksi
lro 50 lro
• ,I.•
. '14
,/2 .. ,/2
20
,. I.
20
2.
10
o tmtEtmtE±t8EttmlttmlE::E±tttE
o 10 20. JO 40 30 eo o
o I. . 7D
0/,
70 20
'" 50
Fig. C4. 29 Torsional modulus of rupture - Fig. C4.30 Torsional modulus of rupture -
7075-T6 aluminum alloy rolled rod. 7075-T6 aluminum alloy forging.
The lightest aluminum tUbe with a positive equation C4.14 is not used.
margin of safety is 2 - .049. The weight of a
10 in. length = .3003 x .101 x 10 =
.303 lbs.
The lightest steel tube is 1-1/4 - .035 and its 1.1
weight is 0.1336 x .283 x 10 = .378 lbs.
Although the steel is heat treated to Ftu = 1.0
200,000, it still is heavier than the aluminum
alloy tube. .9
.8 " . , , , , ,
STRENGTH OF ROUND TUBES r» .• " , •.• " '1
.7
UNDER COMBINED LOADINGS
Reference should be made to Artiale Cl.15 .6
Rc
of Chapter Cl for general explanation of
.5
stress-ratios and interaction equations as
used in determining the ultimate strength of .4
structural members under combined loadings.
.3
C4.22 Combined Bending & Compression.
.2
In tUbes subjected to combined bending .1
and compression, the stress due to compression
is uniform over the cross-section whereas the
bending stress is not uniform over the cross- .1 .2 .3
section. The following stress ratio equation
is possibly somewhat conservative but is C4.23 nIustrative Problem InvolVing Combined Bending
recommended by (Ref. 1). and Compression.
Rc + Rb' =1 - - - - - - - (C4.11) Fig. C4.33 shows a round 2024-T3 tube

acting as a beam-column. It is supported at
which can be written pin A and by the inclined strut BC at point C.
Let 1t be reqUired to determine the margin of
~ + fb' = 1.0 - - - - - - - - - - - - (C4.12) safety when carrying a uniform lateral load of
10 lb./in.
Fc Fb
To find the axial load in tube AC, take
Where f b ' equals the maximum bending stress moments about point (B).
inclUding the secondary moments due to axial
load times beam deflection
~MB =- 40 x 10 x 20 + 20 RAH = 0, hence
Fb = Bending modulus of rupture stress RAH = 400 lb.
fc = Compressive column axial stress Thus axial load in tube is 400 lb. compression.
Fc =Allowable column stress
The column strength of a 2024-T3 1-1/4 -
Equation C4.12 could also be written, .049 round tube 40 inches long with end fixity
=
C 1 is obtained from Fig. C4.6 and equals
P M
-P + -
- 2100 lb., hence stress ratio
a Ma= 1 . 0 - - - - - - - - (C4.13)
R - Column load -400
Where P = column load c - Column strength ::> -2100 = .191
Pa = allowable column load at failure
The bending moment will be maximum at the
M =
bending moment center of span because of symmetry of loading
and the value of the moment is obtained from
Ma =allowable ultimate bending moment the following equation which includes the
secondary moments due to deflection. (Refer-
Fig. C4.32 shows a plot of equation C4.11. ence table A5.1 of Chapter A5.)
The margin of safety equation is,
!'1max = wj" (1 - sec ...b..)
2j - - - - - - - - - - (A)
1
M.S. = Rc
Rb' - 1
- - - - - - - - - (C4.14)
lEI = j
+
Fig. C4.32 shows margin of safety curves
j =I~ 10,700,OOOx.0334
400
= V860= 29 3
•
for estimating closely the M.S. values i f
C4.23
40" bending and tension on round tubes, wilieh is

w = 10#/in. widely used, is
Rb + Rtl.· 6 =1 (C4.15)(Ref.2)
This equation is plotted in Fig. C4.34.

Th9 stress ratio Rt = ft/Ftu. The figure is
based on D/t = 10 and in general is conserva-
Fig. C4.33
tive for other D/t ratios.
1.1
Bt
1.0
wj2 = 10 x 860 = 8600 .9
L _ 40 L 1
2j - 58.6 = .683 sec 2T = L .8
cos 2j
.7
From Table A5.2 of Chapter A5 cos of .683 = I\ .6

.7757.
.5
1
hence cos L/2j = 1/.7757 = 1.29 .4
SUbstituting in equation (A) .3

.2
~nax. = 8600 (1-1.29) =- 2490 in. lb.
.1
From Fig. C4.12 for D/t =
25.5, we read
Fb/Ftu = 1.03. Hence Fb = 1.03 x 62000 = 64000 o .1 .2 .3 .4 .5 .6 .7 .8 .9 1.0 1.1
psi. Then M = FbI/y =
64000 x .0534 = 3420
in. lbs. Rt
C4. 24 Combined Bending and Torsion.

2490
Stress ratio in bending Rb = 3420 = .728 The interaction equation for combined
bending and torsion from (Ref. 1) is,
Rc + Rb = .191 + .728 = .917 which shows that
member is not weak. -(C4.16)
1 1 This equation is plotted in Fig. C4.35 for

Margin of safety M.S.= (Rc + Rb )-1 = (0191 + .728) M.S. = O. Curves showing various M.S. values
-1 = .09
1.1
The margin of safety could be read directly
from curves in Fig. C4.32. 1.0 ,--~
.9'--~
The student should notice that the maximQ~
bending moment of 2490 in. lb. is 24 percent
greater than the primary moment which equals
wL 2/8 = 10 x 40 2/8 = 2000. The lateral de- .7
flection at the midpoint of the tube thus equals
490/400 =
1.22 inch. of .6
.5
The secondary moments due to lateral de-
flection do not vary linearly, so if design
loads were increased the calculation of the .4
maximum bending should be repeated instead of
.3
assu~ing that the moment would increase directly
as the applied load to the beam. .2
C4.23a Combined Bending and Tension .1
The interaction equation for combined o .1 .2 .3 .4 .5 .6 .7 .8 .9 1.0 1.1

Rst
for various values of Rb and Rst are also shown Te = MIl + (TIM)" - -- - - - - - - - (C4.21)
on the figure.
Having the value of Te , select tube sizes
The expression for Margin of Safety is, that will develop this torsional moment Te as
was done in Problem 2 of Art. C4.21. These
sizes are then checked for combined bending and
1 - - - (C4.17)' torsion as illustrated above in the example
problem.
ILLUSTRATIVE PROBLEM 1 C4.24a Ultimate Strength in Combined Compression,
Bending and Torsion.
A 1-1/2 - .• 058 round steel tube (Ftu =
125000) is 30 inches long. It is subjected to The interaction equation for combined
an ultimate design bending moment of 10,000 compression, bending and torsion from (Ref. 1)
in. Ibs. and a torsional moment of 6000 in. is,
Ibs. Find the Margin of Safety.
Solution: D/t = 25.85, I/y = .0912, L/D = 20

Figs. C4.36 and 37 show a plot of this
To find Fb, we refer to Fig. C4.11 where equation. The expression for the Margin of
for D/t = 25.85 and Ftu = 125000 we read Safety is,
Fb = 148000. Then Ma = FbI/y = 148000 x .0912
= 13500 in. lb. 1
M.S. =----;:::::::;:==::::;::: - 1 - - - - - (C4.23)
M 10000 Rc + J R~, + R~t
Then Rb = M
a
= 13500 = .74
To illustrate the use of the interaction
To find Fst, we refer to Fig. C4.18 where curves, let us assume the following values for
for D/t = 25.85 and L/D = 20 we read Fst/1000 the three stress ratios: -
to be 61, whence Fst = 61000.
= .333, Rb = .20, Rst = .20
Rc
R - T - 6000 - 54
st - Ta - 11120 - • Then Rst/R c = .20/.333 = .60.
M.S. = In Fig. C4.36 locate point (a) at the

.J
1
Rb + Rs t
2
1
Intersection of Rc .333 and Rb =
.20. Since =
the intersection point (a) lies inside the
Rst/R c =
.60 curve, we know that a positive
=----;::====1==== -1 = _1__ 1 = .09 margin of safety exists. A line is now drawn
-I 2 2
.74 + .54 .917 through (0) and (a) and extended to an inter-
section with the Rst/R c = .60 curve at point
thus the ultimate strength of the tube has 9 (b). Projecting vertically downward to Rc
percent Margin of Safety under the combined scale, we read Rc =
.538. Then M.S. = (.538/
loading. .333) - 1 .62. =
In a design problem which involves a trial 1.0
and error procedure, using an equivalent
torsional moment Te which will produce the .9
same torsional stress as produced by the
.8
combined bending and torsional loads is quite
useful in shortening the trial and error .7 ~
procedure.
.6
Let f s (max ) =Ter/2I (C3.18) ~ .5
+-L
,.-t
f s (max ) also equals .j f2s + (f /2)" (C4.19 )

b .4
-J-'
Also f s = Tr/2I, and fb =Mr/I .3
SUbstituting these values in C4.18, .2
.1 +
- - (C4.20) t
1.0 .9 ,8 .7 .6 .5 .4 .3 .2 .1 0
Equating C4.20 and C4.18 and solving for Te , Rc
C4.25
Projecting horizontally from point (b) to L

2j
=2 25
x 12.43
= ~ = 1. 005
24.86
Rb scale, we read Rb .323. Then M.S. = =
(.323/.200) - 1 = .62, which checks value
previously found. If the ratio of Rst/Rc is sec ~
2j
= cos 1L/2j - 1 - 1 87 (See Table
- .5361 - . A5.2)
greater than one, we use the curves in Fig.
C4.37 and use the ratio Rc/Rst.
hence,
SUbstituting in equation C4.23,
1
Mmax. = 10x154.7 (1-1.87) = -1347 in. lb.
M.S. = ----;.========= - 1 = .62, as given
.333 + J .20" + .20 2 The column strength for a 25 inch length
from use of curves. and end fiXity c =
1 can be read from Fig.
C4.6 and equals 3700 lb. Then Rc P/Pa = =
1.0 1600/3700 = .432.
.9 To find the Ultimate bending strength,
we refer to Fig. C4.l2, where for D/t =
.8
1.125/.049 = 22.95, we read Fb/Ftu 1.04. =
.7 Then Fb =
1.04 x 62000 = 64500. Thus Ma =
FbI/y =
64500 x .0427 =
2760 in. lb.
.6 1347
R'b Therefore Rj) = 2760 = .487.
.5
To find the ultimate torsional strength
.4
+- we refer to Fig. C4.26 where for D/t = 22.95
.3
-i and estimating location of L/D = 25/1.125 =
22.2 line, we read Fst 27500 • =
.2
... --
Then Ta = FstJ/r = 27500 x .0427 x 2 = 2350
.1
+-
T 650
Whence Rst = T = 2350
a
= .276
o .1 .2 .3 .4 .5 .6 .7 .8 .9 1.0
Rst
M.S.
ILLUSTRATIVE PROBLEM
A 1-1/8 - .049 round tube made from = _ _-r==l==:~==::::- 1 =-.9-19-2

2024-T3 aluminum alloy carries the ultimate .432 + J 2 2
.487 + .276
design loads as shown in Fig. C4.38. Find the
margin of safety under the combined loading. -1 = .01
't~of:lQ
16~ + f • f +• • + f !++
w = 10#/in.
f " • • f '1 f ~OO#
C4.25 Ultimate Strength in Combined Bending and
Flexural Shear.
tfoe-- 25" 1-1/8 -.049 2024-T3 Rd. TUbe~ o.,,-Q"* The interaction curve from this type of
- .'t~"'J
combined loading from (Ref. 2) is,
Fig. C4.38
Rb + R~ =1 (C4.24)
Solution:
1
M.S. =--;:::.=:=== - 1 (C4.25)
The maximum bending moment due to symmetry J Rb + R;
will occur at midpoint of tube. For a beam
column carrying a uniform side load with no end The allowable flexural or transverse shear
moments, the maximum moment is given by the stress is taken as 1.2 times the allowable
following expression. (See Chapter A5, Table torsional stress of the tube (Fst).
A5.1) .
ILLUSTRATIVE PROBLEM
M
max. = Wj" (1
- s e c 2j
L)
A 1-1/4 - .058 round aluminum alloy
j = J¥ = 10,300,000 x .0240 =
1600
2024-T3 tube is used as a simple beam and
..; 154.7 = 12.43 carries an Ultimate design load of 600 lb.
as shown in Fig. C4.39.
C4.26 IN TENSION COMPRESSION BENDING TORSION AND COMBINED LOADINGS.
Interaction curves for this equation are

given in Fig. C4.40, where Rs the stress ratio
3001-/.,..-/.,.../.,.../.,.../.,..-/..,.....,,/
I'--!-'--!.....L.!.tz:::z::z::z:;/~ Shear Diagram
for flexural shear can be found as explained in
the previous article.
'"""'=--L.!..~'--!.....L-.L-t....I:"""Bend ing Moment
Diagram
J
Fig. C4.39
The critical section is adjacent to the
midpoint of beam where shear and bending forces
are largest.
The Ultimate bending strength can be found

from Fig. C4.26 where for D/t = 21.60 we read
Fb/Ftu =
1.07. Then Fb 62000 x 1.07 66300 = =
and therefore Ma = FbI/y = 66300 x .06187 =
4100. Then Rb = M/Ma = 3750/4100 .915. =
The allowable shear stress Fs can be taken
as 1.2 Fst (Ref. 2).
To find Fst, we refer to Fig. C4.26 where

for D/t =
21.6 and L/D 20, we read Fst = =
28400. Then Fs = 1.2 x 28400 = 34000.
The maximum shear stress in a round tube

which occurs at the centerline axis is given
by the equation,
where V = vertical beam shear load

A tube area =
D = outside diameter
d =
inside diameter
An approximate formula is
V
fs = 2 A' for D/t :> 10 error less than 1%.
Using the approximate equation,
fs = 2 .2172
x 300
= 2760 psi.
_ f s _ 2760 _
C4. 27 Ultimate Strength in Combined Tension and Torsion.
Rs - Fs - 34000 - .081
1 1
The interaction equation for this type of
M. S. =--;::::::::;::::=::;: - 1 = ---;::-==;;===~ 1 loading as presented in Ref. 3 is,
VRb + R~ V.915 2
+ .081
2
= .09 - - - - - - - - - - - (C4.27)
The effect of the shear stress is less Where Rt = ft/Ftu

than 1 percent. In using Fs 1.2 Fst, if the = 1
result is greater than Fsu for the material, M.S. =r=-~=~:-- 1 - - - - - - - - (C4.28)
use Fsu for Fs ' VRt + R~t
C4. 26 Ultimate Strength in Combined Compression, C4. 28 Ultimate Strength in Combined Tension, Torsion
Bending, Flexural Shear and Torsion. and Internal Pressure p in psi.
The interaction equation for this combined Ref. 3 gives the following interaction
loading as presented in Ref. 2 is, equation,
C4.27
Rt + R~t + R; =1 (C4.29) (3) If the truss of Fig. 2 is heat treated to

180000 psi after welding, how much weight
Where Rp = p/Pmax where p
max
= 2tFtu
d
could be saved over the results obtained
in problem (2).
where t = wall thickness and d =tUbe diameter.
(4) Same as problem (2) but change material
The expression for Margin of Safety is, to 2024-T3 aluminum alloy round tubes.
Members to be fastened together by rivets
M.S. = 1 _ 1 and gusset plates. For design of tension
- - - (C4.30) members assume that rivet or bolt holes
/Ft+F~t+R; cutout 10% of tube area.
PROBLEMS (5) Fig. 3 shows a typical tubular engine
mount structure. The engine is supported
(1) Fig. 1 shows a portion of a steel tUbular at points A and B. For design purposes
fuselage of a small airplane. The assume that engine torque is reacted 60%
critical tension and compression load is
shown adjacent to each truss member. at A and 40% at B. TUbe material is steel
Assuming and end fixity coefficient c = 2 Ftu = =
95000 psi. Use C 1 for all
for all members, select tube sizes for all members. Determine tUbe sizes for the
following design conditions:
members of the truss. The minimum size to
be used is 3/4 - .035. The top and bottom Condition I Vertical Load factor = 10
longerons should be spliced at least once (down)
using telescoping sizes. The material is
alloy steel Ftu =
95000 and truss is
Thrust Load factor = 2
welded. (forward)
Engine Torque Load factor =2
Condition II <Same as I except vertical

load factor is 5 up.
General Data: Weight of power plant in-

stallation = 440 lb.
Maximum engine thrust = 400 lb. Engine

H.P. =
120 at 2000 R.P.M.
Fig. 1
(2) For the cantilever welded steel tubular

truss of Fig. 2, select the lightest
members for the truss loading as shown.
~t- 22
4000#
T t
30"
T
20
1 -l
(6) The loads shown in Fig. 4 are to be
30" 30"
transmitted to the support at the left, or
1000# 2000#
in other words, a cantilever structure.
Fig. 2
The top and bottom longerons should be The problem is to design the lightest
truss configuration using round tUbes of
continuous members. Minimum size 3/4 - alloy steel Ftu = 95000 and welded to-
=
.035. Use C 2 for web members and C = gether at the truss joints. Use C 1.5 =
1.5 for longerons. Material (chrome-moly for end fixity of all members. There are
steel Ftu = 95000 psi).
C4.28 IN TENSION COMPRESSION BENDING TORSION AND COMBINED LOADINGS.
no restrictions on type of truss or

arrangement of members, however, the goal
is the lightest truss. Omit consideration
of weight of any gusset plates at truss C',D'
joints.
1500 1400 1200 A',B'
Fig. 6
Fig. 4 (12) The ultimate design load if 20,000 Ibs.

compression. L = 30 in. Use C = 1.
(7 ) Same as problem (6), but instead of a Design the lightest round tube from the
cantilever truss use a simply supported following materials and compare their
truss With supports at points (A) and (B). weights.
(8) Fig. 5 shows a front beam and front lift (a) Aluminum alloy 2024-T3
strut in an externally braced monoplane. (b) Alloy steel Ftu = 180,000
The wing beam and lift strut are in the (c) Magnesium alloy Fcy = 10,000
same vertical plane. The ultimate design (13) Same design load as in problem (10) but
loads on the beam for the critical design the lightest streamline tube from
conditions are w = 50 Ib./in. and w -30 = 2024-T3 aluminum alloy material.
Ibs. per inch. Minus means load is acting (14) A round tube is to carry an Ultimate pure
down. bending moment of 14000 in. Ibs. Select
(a) Design a streamline tube to act as the lightest tube size from the following
the lift strut. Material is 2024-T3 materials and compare their weights.
aluminum alloy. (a) Alloy steel Ftu =
240,000, (b) 2024-T3
(b) Same as (a) but made from alloy steel aluminum alloy, (c) Magnesium alloy
Fty= 75000. Compare the weights of Fcy = 30000, (d) Titanium 6AL-4V
the two designs. alloy.
(15) A round tube 20 inches long is to carry
an Ultimate torsional moment of 15000 in.
lb. Select the lightest tube size from
the following materials and compare their
weights.
(a) Alloy steel Ftu =
180,000,
(b) Aluminum alloy 2024-T3, (c) Mag-
=
nesium alloy Ftu 36000.
(16) Determine the lightest 2024-T3 aluminum
alloy round tube 10 inches long to carry
a combined bending and torsional design
load of 4500 and 3000 in.lbs. respectively.
(17) Same as Problem 16, but change material
to alloy steel Ftu = 95000.
(18) A 1-1/2 - .065 2024-T3 round tUbe 50 inches
(9) Fig. 6 illustrates the strut and wire long is used as a beam-column. The
bracing for attaching float to fuselage of distributed load on beam is 12 lb. per
a seaplane. Determine the necessary sizes inch and the axial load is 700 Ibs. What
for the streamline struts AC and BD for is the M.S. under these loads.
the following load conditions. (19) If the tUbe in problem (19) was also
subjected to a torsional moment of 1400
Condition 1. V = -32000 Ibs., in.lb., what would be the M.S.
H = - 8000 Ibs.
Condition 2. V =-8000 Ibs.,
H .. -28000 Ibs.
Material 2024-T3 aluminum alloy. Use C=l. References
(10) Tube size 2 - .065 round. L = 44 in.,
C = 1.5. Material alloy steel Ftu = (1) Military Handbook. MIL-HDBK-5. March,
95000, welded at ends. Design ultimate 1961.
loads equal 22000 lb. compression and (2) Lockheed Report 2072.
28000 Ibs. tension. Find margin of safety. (3) M. A. Sadowsky, "A Principle of Maximum
(11) Same as Problem 10 but heat treated to Plastic Resistance," Journal of Applied
Ftu = 150000 after welding. Mechanics, June, 1943.
CHAPTER Cs
BUCKLING STRENGTH OF FLAT SHEET IN COMPRESSION, SHEAR,
BENDING AND UNDER COMBINED STRESS SYSTEMS
C5. 1 Introduction. -It
Chapter A18, Part 2, introduced the

student to the theoretical approach to the COLUMN
problem of determining the buckling equation
SUCKLED FDRM-,
for flat sheet in compression with various
edge or boundary conditions. A similar AT STRIP
Fig. (a)
theoretical approach has been made for other
load systems, such as shear and bending, thus
the buckling equations for flat sheet have
been available for many years. This chapter
will summarize these equations and provide
design charts for practical use in designing
sheet and plate structures. Most of the
material in this chapter is taken from (Ref. 1),
NACA Technical Note 378l-Part I, "Buckling of Fig. (b)
Flat Plates" by Gerard and Becker. This report
is a comprehensive study and summary of
practically all important theoretical and
experimental work published before 1957. The
report is especially useful to structural design
engineers.
C5.2 Equation for Elastic Buckling Strength of
Flat Sheet in Compression.
From Chapter A18, the equation for the Fig. C5. 1 (Ref. 1) Transition from column to plate as
elastic instauility of flat sheet in compression supports are added along unloaded edges. Note changes in
is, buckle configurations.
2
n kC E t Fig. C5.2 gives curves for finding the
Ocr =----- (_)2 - - - - - - (C5.l) buckling coefficient k c for various boundary
12 (l-ZJ~) b or edge conditions and alb ratio of the sheet.
Where k c = buckling coefficient which depends The letter C on edge means clamped or
on edge boundary conditions and fixed against rotation. The letter F means a
sheet aspect ratio (alb) free edge and SS means simply supported or
hinged. Fig. C5.3 shows curves for k c for
E =modulus of elasticity various degrees of restraint (E) along the
Ve = elastic Poisson's ratio sides of the sheet panel, where E is the ratio
of rotational rigidity of the plate edge
b = short dimension of plate or loaded stiffener to the rotational rigidity of the
edge plate.
t = sheet thickness
Fig. C5.4 shows curves for k c for a
C5. 3 Buckling Coefficient k c flange that has one edge free and the other
With various degrees of edge restraint. Fig.
Fig. C5.l shows the change in buckled C5.5 illustrates where the compressive stress
shape as the boundary conditions are changed varies linearly over the length of the Sheet,
on the unloaded edges from free to restrained. a typical case being the sheet panels on the
upper side of a cantilever wing under normal
In Fig. (a) the sides are free, thus sheet flight condition. .
acts as a column. In Fig. (b) one side is
restrained and the other side free, and such a Fig. C5.6 gives the k c factor for a long
restrained sheet is referred to as a flange. sheet panel With two extremes of edge stiff-
In Fig. (c) both sides are restrained and this ener, namely a zee stiffener which is a
restrained element is referred to as a plate. torsionally weak stiffener and a hat section
C5.1
C5.2 BENDING AND UNDER COMBINED STRESS SYSTEMS
16 /5
I
I
I C
A C I I ss
8
C
I
14 f - - -
I
/4 1-----1 13
I I'
l-
,, I
I
I
Sf
c SS I IF 0
C
I
12
(:~::3
/2 \
\I 1--------. - /I
1\
....
l- "'''''''R')':¥\ \
~ FEE
I T
6 /0 r---- \ ~La4OED E06ES CLAMPED
\1 \r".
II
, \
J
1\ 9
V
,I \ \ ,
---- - LOADED ED6ES CLAMPED \
I- I'\ \
LOAOED £06E$ SIMPlY 1\ "- ......... ,
,
TYPE OF
""'" SlIPPORrED
S.fr~~T
8 :--- f'-_ I
~ 11\~
, I, '--
. . . --r--__ ----- II.WLOADED
E06ES
7 f-- fA /r-.... /'
-I aJ
I-
\
' ....,
A
~ I"---"
l'-- r:::- ~
~
50
20
W ~ ~-,
6
~
.~ --~
8
/0
I- ..., 5 5
~~
~'
4 /'0.. ~---" ---- ----- ., f

jj5
~\
\
I
I- \
,, \
\
\
I
2 \, "
I-
~ " t--'------
~'
- --_
-- ----- ----
D
2
,I I I
o /
" % 8 12 16 20 M 2.8 .J2 36 4.0
Fig. C5. 2 (Ref. 1) Compressive-buckling coefficients for flat

rectangular plates.
Fig. C5. 3 (Ref. 1) Compressive-buckling-stress coefficient
of plates as a function of alb for various amounts of edge
19 rotational restraint.
1 i \ 1\
18 I---+--ill-l--:+---t+
\ 1 \
7,---.,.---------------.,
I
, \ I \
I?
---++ \ ~ ·1
16
1.5 -.l- :\\~rl:::;~;::::

~ --\ \1
6 t--ic---r
,
14
--1-- I ,
5. 1-----111--+-
13
--1- 1
/2 1--+~+m\*-----,7-J'-A-~"k----'-~.-IC-~..,-----l
;
I--If----+-\\\-\--"'-+----+I~d_''"'"'__+-'''f=--+~ :
k.
II
5
10
.9
2
B
" ~ "-
.7
.5
D 2
:--
:--
.I
.5
0 1....,..._--T---3'i---4:T---5r--~<X>
o }t'-'......'........
"0 .5 1.0 2.0 2.5 30 :J5 4.0 4.5 50 o/b
a~
Fig. C5. 5 (Ref. 1) Average compressive-buckling-stress
coefficient for rectangular flat plate of constant thickness
Fig. C5.4 (Ref. 1) Compressive-buckling-stress coefficient with linearly varying axial load. k c n 2 E 11.-) 2
of flanges as a function of alb for various amounts of edge a av - av 'b'
rotational restraint. 12 (1 - V e 2 )
C5.3
which is a closed section and, therefore, a Substituting in Eq. C5.l,

relatively torsionally strong stiffener. Fig.
2
C5.6a gives the compression buckling
coefficients k c for isosceles triangular ocr -
- 11. x4.0xlO,700,000
12 (1 _ .3 2 )
(.04)2
5
= 2480 psi.
plates.
This stress is below the proportional
---r-----,.---...,..--.....,---,..----,
8 ...... limit stress for the material, thus equation
C5.l applies and needs no plasticity correction.
7 C5.4 Equation for Inelastic Buckling Strength of Flat Sheet
in Compression.
If the buckling or instability occurs at

a stress in the inelastic or plastic stress
51--'--_4------+----+----+---+-----t
range, then E and V are not the same as for
elastic buckling, thus a plasticity correction
factor is required and equation C5.l is
4 written,
2
31----+------+
Ocr = n'( 11. kc E
(~)2 __ - - __ (C5.2)
l2(1-Ve 2 ) b
Where T/ is the plasticity reduction factor

300 and equals ocr plastiC/Ocr elastic.
b/t The values of k c and Ve are always the
Fig. C5.6 (Ref. 1) Compressive-buckling coefficient for long
elastic values since the coefficient ~ contains
rectangular stiffened panels as a function of bit and stiffener all changes in those terms resulting from
torsional rigidity. inelastic behaVior.
3'Or----,-----,.----,------, A tremendous amount of theoretical and

experimental work has been done relative to
the value of the so-called plasticity cor-
rection factor. Possibly the first values
used by design engineers were T/ Et/E or =
=
T/ Esec/E. Whatever the expression for T/ it
2O'1----------t--+-- must involve a measure of the stiffness of the
material in the inelastic stress range and
since the stress-strain relation in the plastic
k c /5'1----H--\----+---+------1
range is non-linear, a resort must be made to
the stress-strain curve to obtain a plasticity
correction factor. This complication is
greatly simplified by using the Ramberg and
Osgood equations for the stress-strain curve
51----------+------"'.d---+----j which involves 3 simple parameters. (The
reader should refer to Chapter Bl for
information on the Ramberg-Osgood equations.)
oO!:-'-J...J..-'---:.5,----;/.'::"O--7;/5,----:!Z_O Thus using the Ramber-Osgood parameters (Ref.l)
a/b presents Figs. C5.7 and C5.8 for finding the
compressive buckling stress for flat sheet
Fig. C5. 6a (Ref. 1) Uniform Compression. panels with various boundary conditions for
both elastic and inelastic buckling or in-
Illustrative Problem. Find the compressive stability.
buckling stress for a sheet panel with (a) = 10 C5.5 Simple Problems to lllustrate Use of Curves in Figs.
and b = 5 inches, thickness t = .04 and all C5.7 and C5. 8.
edges are simply supported. Material is Ocr
2024-T3 aluminum alloy.
The sketch shows a 3 x 9
Solution: E = 10,700,000. V e = 0.3, inch sheet panel. The sides ss
alb = 10/5 = 2. The boundary or edge condition are simply supported. The
corresponds to Case (c) in Fig. C5.2. Thus material is aluminum alloy
using curve (c) for alb =
2, we read k c = 4.0. 2024-T3. The thickness is
.094". E = 10,700,000.
1.4 ..---,---,----r---,...--"""T"----,----r-----r-----,-----, n
121----t---+---t---+---+---+-----l---+---:;;~.,.jL'-------I
8 1------1----+----+----7f~~~~---+---+---+---+-----l
°crP O • 7
or
Fcc/F o. '7 6 I---t---t-------,~n,£-+_--+---+-----l
ff.1---+--:Q;"L--t----+---+----t-------I
.21--~L--+---t-----t----+----t----t------t---+_---I
OIlW.LJ.ll.uL-l..---l_.l.-....I---L--l_.l.-....I---l..--l_.l...-....I----L..--'_.l--....I---"--'
o .2 .4 .6 .8 LO L2 /4 2.0
2
kc n E t 2
12 (1 - Ve 2 )O (j)
0.7
Fig. C5.7 Chart of Nondimensional Compressive Buckling Stress for Long

Hinged Flanges. T/ = (E /E)(1 - U 2)1(1 - U 2).
s e
1.4 r---,.---,.---,.---,.---,...---,---,---,---,----, n
2
1.2 1---+---+----+---t-----j---j---+---+---*<!C---:::7'l 3
.8 1---+---+----+---.r,~,;£-..i_;."r;£--+----+----+----+----1
Ocr /o O.7
or
F ccrlF 0 • 7 .6 1---+---+-----¥-~hLf----j---+--___1
.4 1---+---""~--+---+-----1---+----1
.21-----,.p.---+----+---I------1r----+---+----t----+---i
O&J.l..uJ.UoU_...L--4_...L---J._.l-........._J..-...J.........l_..............l_...L-........_ ........--J._.l-.........- '

o .2 .8 1.0 1.2 2.0
2
kcn E t 2
12 (1 - Ve 2)0 • (j)
O 7
Fig. C5.8 Chart of Nondimensional Compressive Buckling Stress for Long
Clamped Flanges and for Supported Plates with Edge Rotational Restraint.
T/ = (E s /2E) { 1 + 0.5 [1 + (3Et/E s )J 10 12} (1 2

- 1)e )/(1 _ 1)2).
C5.5
Ve = 0.3. Find the buckling stress ocr. Thus a further correction must be made for
alclad sheets because of the lower strength
Solution: We use Fig. C5.8 since it covers clad covering material. Thus the buckling
the boundary conditions of our problem. The stress for alclad sheets can be written:
parameter for bottom scale is,
k n 2E (C5.3)
c
- - - - - - - - (A)
ft/2
~
t
For alb = 9/3 =
3, we find k c from curve ,~
(c) of Fig. C5.2 equals 4.0.
The use of Fig. C5.8 involves the use of J -*--
00.7 and n the Ramberg-Osgood parameters. tft/2 Cladding
Referring to Table Bl.l of Chapter Bl, we find Fig. C5.9
for 2024-T3 aluminum alloy that 00.7 = 39000
and the shape factor n = 11.5. Reference 1 gives simplified cladding
reduction factors as summarized in Table C5.l.
Substituting in (A):- Thus the buckling stress for alclad sheets is
determined for the primary strength properties
2
4.0 n x 10,700,000 (.0~4)2 = .98 as normally listed for such materials as
illustrated in the two previous example
12 (1 - .3 2) 39,000 problems. The resulting ocr is then reduced
by use of equation C5.3, using values of
From Fig. C5.8 using .98 on bottom scale
and n =
11.5 curve, we read on left hand scale ~ from Table C5.1.
that Ocr/crce? = .84.
Then ocr = 39000 x .84 = 32800 psi.

If we heglected any plasticity effect, then
we would use equation C5.2 with ~ 1.0, or,=
ocr =n 2
x4.0xlO,700,000 (.094)2 = 38400 psi
12 (1 _ .3 2 ) 3
Whereas the actual buckling stress was 32800,

°
or in this case the plasticity correction
factor is 328/384 .854.=
The sheet thickness used in this example
of .094 is relatively large. If we change the
sheet thickness to .051 inches the results
would be practically no correction within the £
accuracy of reading the curves, and the buckling
stress ocr would calculate to be 11200 psi, Fig. C5. 10 (Ref. 1) Stress.=strain Curves for Cladding, Core,
which is below the proportional limit stress and Alclad Combinations. O/Ocore = 1 - f +fJf; fJ
=OcVU core ·
and thus no plasticity correction.
C5. 6 Cladding Reduction Factors. Table C5. 1 (Ref. 1) Summary of Simplified Cladding
Reduction Factors
Aluminum alloy sheet is available With a -
thin covering of practically pure aluminum and Loading ° cl<Ocr<Opl °cr"" °pI
is widely used in aircraft structures. Such
1 + (3j?f/4) 1
material is referred to as alclad or clad Short plate columns
aluminum alloy. The mechanical strength 1 + 3f 1 + 3f
properties of this clad material is consider-
1 1
ably lower than the core material. Since the Long plate columns
clad is located at the extreme fibers of the 1 + 3f 1 + 3f
alclad Sheet, it is located where the strains
attain their highest value when buckling takes Compression and 1 + :Wf
place. Fig. C5.9 shows make up of an alclad shear panels 1 + 3f 1 + 3f
sheet and Fig. C5.l0 shows the stress-strain
curves for cladding, core and alclad combina-
tions.
BUCKLING UNDER SHEAR LOADS BUCKLING UNDER BENDING LOADS
C5.7 Buckling of Flat Rectangular Plates C5. 8 Buckling of Flat Plates Under Bending
Under Shear Loads. Loads.
The critical elastic shear buckling stress The equation for bending instability of
for flat plates with various boundary flat plates in bending is the same as for
conditions is given by the following equation: compression and shear except the buckling co-
n" k E efficient kb is different from k c or k s ' When
7 = __'__S (.!.-) "
- - - - - - - (C5.4)
a plate in bending buckles, it involves
cr relatively short wave length buckles equal to
12 (1 _ Ve ") b
2/3 b for long plates with simply supported
Where (b) is always the shorter dimension of edges (see Fig. C5.14). Thus the smaller
the plate as all edges carry shear. k s is the buckle patterns cause the buckling coefficient
shear buckling coefficient and is plotted as a kb to be larger than k c or ks '
function of the plate aspect ratio alb in Fig.
C5.11 for simply supported edges and clamped
edges.
If buckling occurs at a stress above the

proportional limit stress, a plasticity
correction must be included and equation C5.4
becomes
Fig. C5. 14 (Ref. 7) Bending Buckle Patterns
71 n" kE
7
cr
=_-=-s_ _-=-s__ (~) " - - - - - - - (C5.5) For bending elastic buckling the equation
12 (1 - V e ") is,
Test results compare favorably with the
results of equation C5.5 i f 71s = Gs/G where G - - (C5.6)
is the shear modulus and Gs the shear secant
modulus as obtained from a shear stress-strain For bending inelastic buckling,
diagram for the material.
Tl
b ,n" k
b E t
A long rectangular plate subjected to Ocr = - - - - - - ( - ) " - - - - - - - (C5.7)
pure shear produces internal compressive 12 (l-Ve ") b
stresses on planes at 45 degrees with the
plate edges and thus these compressive stresses Where kb is the buckling coefficient and
cause the long panel to buckle in patterns at is obtained from Fig. C5.15 for various alb
an angle to the plate edges as illustrated in ratios and edge restraint E against rotation.
Fig. C5.12, and the buckle patterns have a half In the alb ratio the loaded edge is (b).
wave length of 1.25b.
The plasticity reduction factor can be
obtained from Fig. C5.8 using simply supported
1---1. 25b~ 7 edges.
BUCKLING OF FLAT SHEETS

UNDER COMBINED LOADS
The practical design case involving the
use of thin sheets usually involves a combined
Fig. C5. 12 (Ref. 7) load system, thus the calculation of the
buckling strength of flat sheets under com-
Fig. C5.13 is a chart of non-dimensional bined stress systems is necessary. The
shear buckling stress for panels with various approach used involves the use of inter-action
edge rotational restraint. This chart is equations or curves (see Chapter Cl, Art.
similar to the chart in Figs. C5.7 and C5.8 in Cl.15 for explanation of inter-action
that the values 00.7 and n must be known for equations) •
the material before the chart can be used to
find the shear buckling stress. C5.9 Combined Bending and Longitudinal
Compression.
The interaction equation that has been

widely used for combined bending and longi-
C5.7
...----,,.....-...,.---r---,----r----r----r----r---,---,n
.6 t---+--+------i---+--+---i---+--+---:7'~~--i
.4t----+--_+_--+--~f7"""""''.A;r-_+--_+_--4---t-----1
'T:crl 0 0 • 7
or
Fsc/Fo. 73 1----1---+--~~~_+--_+--+-
2t---+----:#7>'-----if--_+---+---t--
./
.I .2 .4 .5 .6 .8 .9 1.0
k s n" E t"
2 (-)
12 (1- Ve )00 • 7 b
Fig. C5. 13 (Ref. 1) Chart of Nondimensional Shear Buckling Stress for Panels With
Edge- Rotational Restraint. 71 = (Es/E) (1 .. Ve ")1(1 - V').
60
15..------r---~----r---...,....----.
56 -
131------+-\-----+----+----+-----1
52 f------
48
~..
/
//////////////
////////
,
.
or
./ /
\\ J::.
~r;/
44
.......
~ ~ P::
--
'
/It-----t--f---'~+_----+---+_--______l 40
v
~\'- I "'-.
"- ..........
r--- ---
v-
CD
100
50
k. 36
t'--" ~ -
L./ "-
.~ ----
-
I
- 20
\, --- - -
9t-----fV----+--f---+----+--I----1 10
V) t--
./ 5
~~~
'--- 3
I
24 . o
7t-----+-~;:___+_----jL-+_--___j~--+_I i i
E.
._~--
I
16 I
Link to C5.8....
9 II 13 15 17 19 21 23
5 O~u..L..u..L..~/---..l:2~---:!:---~-----J5 3 5 7
14
o/b
Fig. C5. 11 (Ref. 1) Shear-Buckling-Stress Coefficient of Fig. C5.15 Bending-Buckling Coefficient of Plates as a Function
Plates as a Function of alb for Clamped and Hinged Edges. of alb for Various Amounts of Edge Rotational Restraint.
tudinal compression is, 1.1

• '>~'10 Fig. C5.16
Rb :1..76 + Rc :: 1. 0 1.0 Combined Bending
- - - - - - - - (C5.8) & Shear.
·00 2 ,.
.9 Rb+ R s=1
This equation was originally presented in '10
Ref. 2 and the interaction curve from plotting .8
this equation is found in many of the '20
structures manuals of aerospace companies. .7
"'10
. 0
Fig. C5.15 is a plot of eq. C5.8. It rf .6
also shows curves for various margin of safety
.5
values.
.4
Fig. C5.15 .3
1.0 Combined Bending
& Long. Compression .2
.9 Rb :1..76+ R
c .1
.8
o .1 .2 .3 .4 .5 .6 .7 .8 .9 1.0 1.1
.7
R s , (Rs + Rst), Rst
Rc .6
.5 Fig.·C5.17 is a plot of equation C5.11 •

If the direct stress is tension, it is
.4 included on the figure as negative compression
using the compression allowable •
.3
C5. 12 Combined Compression, Bending & Shear.
•2
.1 From Ref. 5, the conditions for buckling

are represented by the interaction curves of
Fig. C5.18. This figure tells whether the
o .1 .2 .3 .4 .5 .6 .7 .8 .9 1.0 1.1 sheet will buckle or not but will not give the
Rb margin of safety. Given the ratios Rc , Rs and
Rb:- if the value of the Rc curve defined by
C5.10 Combined Bending & Shear. the given value of Rb and Rs is greater
numerically than the given value of Rc , then
The interaction equation for this com- the panel will buckle.
bined loading (Ref. 1 & 2) is,
(C5. 9) .9
The expression for margin of safety is, .8

Rc=O
M.S. =-;::=1==_2 2
1 (C5.10) .7 .1
v'R +R .2
b s
.6 .3
Fig. C5.16 is a plot of equation C5.9. .4
Curves showing various M.S. values are also Rb .5
shown. Rs is the stress ratio due to .5
torsional shear stress and Rst is the stress .4
.6
ratio for transverse or flexural shear stress. .7
C5. 11 Combined Shear and Longitudinal Direct Stress. .3 .8

(Tension or Compression. ) 8
.2 .9
The interaction equation is (Ref. 3,4) .95.
.1
R + R2 = 1.0 (C5.11) .975
L S o
o .1 .2 .3 .4 .5 .6 .7 .8 .9 1.0
- 1 (C5.12) Rs
Fig. C5. 18 (Ref. 5)
C5.9
1.8
~C~~~~~~JI
1.6
.._ 4 .. ·4 ...... .., ... _ ...
«(~~~~~. -=~(
1.4
1.2
.... 1.0
'"
p:;
1-0
0
.8
'"
p:;
.6
.4
.2
o
-2.0 -1. 8 -1. 6 -1. 4 -1. 2 -1. 0 -.8 -.6 -.4 -.2 o .2 .4 .6 .8 1.0 1.2
RL Longitudinal Tension RL Longitudinal Compression
Fig. C5. 19 (Ref. 5) Combined Compression, Bending and Shear.

The margin of safety of elastically centroid of the cross-section, a bending moment

buckled flat panels may be determined from is produced about the x-x axis equal to 1400 x
Fig. C5.l9. The dashed lines indicate a 3.7 =
5170 in. lb. =
Mx , where 3.7 is distance
typical application where Rc = .161, Rs = .23, from load P1 to x-x axis.
and Rb = .38. Point 1 is first determined for
the specific value of Rs and Rb. The dashed
diagonal line from the origin 0 through point
1, intersecting the related Rc/R s curve at
point 2, yields the allowable shear and bending Area of Zee
stresses for the desired margin of safety Stringer = .18
calculations. (Note when Rc is less than Rs Area of Corner
use the right half of the figure; in other Member =
cases use the left half). 0.25 sq. in.
C5. 13 Illustrative Problems.
In general a structural component com-

posed of stiffened sheet panels will not fail
when buckling of the sheet panels occurs since
the stiffening units can usually continue to
carry more loading before they fail. However,
there are many design situations which require
that initial buckling of sheet panels satisfy
certain design specifications. For example,
the top skin on a low Wing passenger airplane
should not buckle under accelerations due to
air gusts which occur in normal every day
flying thus preventing passengers from
observing wing skin buckling in normal flying
conditions. Another example would be that no
buckling of fuselage skin panels should occur
while airplane is on ground with full load
aboard in order to prevent public from
observing buckling of fuselage skin. In many
airplanes, fuel tanks are built integral with
the wing or fuselage, thus to eliminate the
chances of leakage developing, it is best to Fig. C5.20
design that no buckling of sheet panels that
bound the fuel tanks occur in flying and The sheet thicknesses, stiffener areas and
landing conditions. In some cases aerodynamic all necessary dimensions are shown on Fig.
or rigidity requirements may dictate no C5.20. The total cross-sectional area of beam
buckling of sheet panels. To insure that section including all skin and stringers is
buckling will not occur under certain load 3.73 sq. in. The moment of inertia about x-x
requirements, it is good practice to be centroidal axis calCUlates to be 49.30 in. 4 •
conservative in selecting or calCUlating the
boundary restraints of the sheet panels. Since the beam section is symmetrical,
the top panels A, Band C are SUbjected to the
Problem 1. same stress under the P1 load system.
Fig. C5.20 shows a portion of a cantilever Compressive stress due to transferring

wing composed of Sheet, stiffeners and ribs. loads P1 to centroid of beam cross-section is,
The problem is to determine whether skin panels
marked (A), (B) and (c) will buckle under the fc = 2P 1 /area = 1400/3.73 = 375 psi
various given load cases. The sheet material
is aluminum alloy 2024-T3. Compressive stress due to constant bending
moment of 5170 in. lbs. is,
Load Case 1.
f
c
=
= MxZ/I x 5170 x 4.233/49.30 = 444 psi
P1 = 700 lb., P2 = 0, P3 =0
Total f =
375 + 444
c
819 psi. =
With only loads P1 acting, the one cell
stiffened cantilever beam is subjected to a The skin panels are subjected to compres-
compressive axial load of 2 x 700 1400 lb. = sion as shown in Fig. a. The boundary edge
Since the P1 loads are not acting through the conditions given by the longitudinal stiffeners
C5.11
and the rib flanges will be con- The Rc + R~ = .431 + .309 2 .526. Since =
servatively assumed as simply the result is less than 1.0, no buckling
supported. (F c is same as ocr) occurs.
cr
2
The M.S. = -----;.===== -
2
n kc E t 2 1
FCcr = 12 (1-V 2 ) ("b) Rc + J R~ + 4R~
e
(See Eq. C5.1) = .431 + J .43122 + - 1 = .69
4 X .309 2
alb of skin panel = 15/5 = 3
Fig. (a)
Load Case 3.
From Fig. C5.2 for Case C,
we read k c 4.0 = P~ = 700, P2 = 500, P3 = 100 lb.
2
F - n
ccr-
X 4.0 X 10,700,000 (.0535 )2
12 (1 - 0.3 2 )
= 1900 psi The two loads P3 produce bending and
flexural shear on the beam. The bending moment
produces a different end compressive stress on
Since Fccr the buckling stress is greater the three sheet panels since the bending moment
than the applied stress f c , the panels will is not constant over the panel moment. To
not buckle. simplify we will take average bending moment on
the panel.
M.S. = (Fccr/f c ) - 1 = (1900/819) -1 = 1.32
Mx(av) = 200 x 52.5 = 10500 in. lb.
Load case 2.
f c due to this bending =MxZ/l x = 10500 x
p~ = 700 lb., P2 = 500, P3 =0 . 4.233/49.3 = 903 psi.
The two loads P2 acting in opposite Total f c = 903 + 819 = 1722 psi.
directions produce a couple or a torsional
moment of 500 X 16.5 = 8250 in. lb. on the Rc = fc/F c cr = 1722/1900 = .906
beam structure, which means we have added a
pure shear stress system to the compressive The two loads P3 produce a traverse shear
stress system of Case 1 loading. =
load V 200 lb. The flexural shear stress
must be added to the torsional shear stress as
The shear stress in the top panels A, B found in Case 2 loading.
and Cis,
Due to symmetry of beam section and P3
fs = T/2At = 8250/2 x 138 x .035 = 854 psi. loading the shear flow q at midpoint of sheet
panel (B) is zero. We will thus start at this
(Where A is the cell inclosed area) point and go clockwise around cell. The shear
flow equation (see Chapter A15) is,
The shear buckling stress is
20 7.20
n 2
k E 41!50~~ 10.94
Fs - s (.!/ - - - - (See Eq. C5.4) (1 --+ 2 i 3)l / / !
~~.5"'tt::jI;J
cr 12 (1 _ V e 2 ) b
alb = 15/5 = 3. From Fig. C5.11, for 5~ 12.34

hinged or simply supported edges, we read
k s = 5.8. Fig. (b)
q =lZZA
Ix
x 5.8 x 10,700,000 (.035)2= 2760 si
12 (1 - .3 2 ) 5 P • q =- :~.03 = 4.05 ZZA
The sheet panels are now loaded in = 0 (Refer to Fig. b)
combined compression and shear so the inter-
action equation must be used. From Art. C5.12 =- 4.05 x 2.5 x .035 x 4.23 = - 1.50
the interaction equation is Rc + R 1. s= - - l.50-4.05x .18x3.69 = 4.20
Rc = fc/F ccr = 819/1900 = .431 - - 4.20-4.05x5x .035x4.23 = - 7.20
- - 7.20 - 4.05 x .25 x 3.69 = - 10.94
Rs = fs/F scr = 854/2760 = .309 =- 10.94 - 4.05 x .051 x 3.69 x 3.69/2 =-12.34
(See Fig. b for plot of shear flow)
C5.12 BENDING AND UNDI::R COMBINED STRESS SYSTEMS
The shear flow q on panel (A) varies from (4) A sheet panel 5" x 12.5" x .051 has all
4.20 to 7.20 or the average q = (4.2 + 7.2)/2 edges simply supported. The panel is
= 5.7. Thus the average shear stress is subjected to combined compression and
5.7/.035 =
163 psi. It is in the same shear loads which produce the following
direction as the torsional shear flow and thus stresses:-
is additive.
fc = 2400 psi, applied normal to 5" side.
Total f s =163 + 854 =917 psi
fs = 2800
psi. Will the sheet buckle
Rs = fs/F scr = 917/2760 = .332 under the given load system if made of
aluminum alloy 2024-T3 material. What
R; =1, Subt.:- .906 + .332 = 1.016,
2
Rc + is the margin of safety.
since the result is greater than 1.0, initial
buckling has started. The margin of safety is (5) If the material in problem (4) is changed
slightly negative and equals, to alloy steel Ftu = 95000 psi, what
would be the margin of safety. If sheet
M.S.
2
=------:;::===:0=====::::0 1 = - .01 was heat treated to Ftu =
180,000, what
.905 + J .905 + 4 x .332
2 2 would be the M.S.
In this example problem, the panels were (6) A 3" x 12" x .040" sheet panel is sub-
assumed simply supported, which is conservative. jected to the following combined stresses.
Reference to Fig. C5.6 shows that k c could be f c = 3000, fb = 10000, f s = 8000. The f c
assumed higher as the panel is riveted to a and fb stresses are normal to the 3" side.
zee shaped stringer which has some torsional If sides are simply supported, will panel
resistance and thus panel is not free to buckle if made of 7075-T6 aluminum alloy.
rotate at its boundaries. What is M.S.
Panel (C) is less critical because the What will be the M.S. if material is
flexural shear is acting opposite to the changed to Titanium Ti-8I1n.
torsional shear stress, thus f s total = 845 -
163 =
682 psi. The Rs = 682/2760 = .246. References:-
Rc + R; = .905 + .246 2 = .966. Since the (1) NACA Tech. Note 3781, 1957.
result is less than 1.0, panel will not buckle.
(2) ANC-5 Amendment 2. Aug. 1946.
Panel (B) carries a small shear flow,
being zero at center of panel and increasing (3) NACA ARR.No. L6A05
uniformly to 1.5 lb. per inch at the edges,
and flowing in opposite directions from the (4) NACA ARR.No. 3K13
centerline. Thus transverse shear will have
negligible effect. Thus Rs = 854/2760 = .309. (5) ANC-5 Revision of 1942
Rc + R; = .905 + .309 2 =
1.00, or panel General References on Theory
(B) is on the verge of buckling under the
assumptions made in the solution. (6) NACA Tech. Note 3781.
C5.14 Problems. (7) "Introduction to Structural Stability

Theory". By Gerard. Book published by
(1) A sheet panel is 3" x 9" x .051" in size. McGraw-Hill.
Consider all sides as simply supported.
Determine the buckling load if the (8) "Theory of Elastic Stability", by
compressive load is applied normal to the Timoshenko, McGraW-Hill Co.
3" sides. Do so for 3 different materials,
(1) aluminum alloy 7075-T6, (2) magnesium (9) "A Unified ·Theory of Plastic Buckling
HK3lA, (3) Titanium Ti-8I1n. of Columns and Plates", NACA Report
898. By Stowell.
(2) In Problem (1) if all edges were clamped,
What would be the buckling load. (10) "Plastic Buckling of Simply Supported
Compressed Plates", NACA T.N. 1817.
(3) In Problem (1) if the 9" sides were By Pride & Heimerl.
clamped or fixed and the 3" sides simply
supported, what would be the buckling (11) "Buckling of Metal Structures" by
load. Bleigh. McGraw-Hill Co.
CHAPTER C6
LOCAL BUCKLING STRESS FOR COMPOSITE SHAPES
C6. 1 Introduction. From Equation C5.1 of Chapter C5, the

buckling compressive stress for a long flange
Thin flat sheet is inefficient for carrying is,
compressive loads because the buckling stresses
are relatively low. However, this weakness or
fault can be greatly improved by forming the
flat sheet into composite shapes such as
angles, channels, zees, etc. Most of the many From Fig. C5.2 of Chapter C5, k c = .43,
composite shapes can also be made by the then
extruding process. Formed or extruded members
are widely used in Flight Vehicle Structures,
thus methods of calculating the compressive
strength of such members is necessary.
If the buckling stresses are above the
proportional limit stress, use Fig. C5.7 in
C6. 2 Compressive Buckling Stress for Equal
Flanged Elements. Chapter C5 to take care of the plasticity effect.
The simplest equal flanged member that can For formed angles, the flange width b
be formed is the angle shape. Other shapes extends to centerline of adjacent leg, but for
with equal flanges are the T section and the extruded angles, the width b extends to inside
cruciform section as shown in Fig. C6.1. edge of the adjacent flange or leg.
~rlange ~~lange ~
C6.3 Compressive Buckling Stress for Simple Flange-Web
Elements.
~
Angle
~
T-Section
~ciform
Section
The most common flange-web structural
shapes are channels, zees, and hat sections. A
flange has one unloaded edge free, whereas a
Fig. C6. 1 web has no free unloaded edge and thus has an
These sections can be considered as a group of unknown restraint on the boundary between the
long flanges, as illustrated, for the angle web and the flange. Fig. C6.3 shows the break
section in Fig. C6.2. Since the flanges which down of a Z section into two flange and one web
make up the section are equal in size, each plate elements.
flange will buckle at the same stress. There-
fore each flange cannot restrain the other and
thus it can be assumed that each flange is
simply supported along the flange junction as
illustrated in Fig. C6.2.
Q)
Q)
~ Flange
cr cr cr cr
Breakdown of Z into Flange

and Web Elements.
Fig. C6.3
The buckling strength of the web and flange

elements depends on the boundary restraint
between the two elements. If this restraint
which is unknown could be found in terms of a
known rotational restraint E as presented in
Chapter C5, the buckling coefficients could be
Fig. C6.2 found from charts in Chapter C5. Having the
C6.l
C6.2 LOCAL BUCKLING STRESS FOR COMPOSITE SHAPES
buckling stress for each element, the critical For the plasticity correction of shapes
buckling stress will be the smaller of the two. covered in Fig. C6.4 (Ref. 4), the plasticity
The buckling load based on the buckling stress correction for a flange free on one edge can
is not the failing load as more load can be be used with accuracy. Thus we can use chart
taken by the material in the corner regions in Fig. C5.7 of Chapter C5.to correct for
before local failure or crippling takes place. plasticity effects.
The subject of local crippling of formed and
extruded shapes is covered in Chapter C7. The parameter for bottom scale of Fig.
C5.7 is equation (A) divided by 0 0 . 7 , or
Using the moment distribution method or a 56100/39000 =
1.44. Using this value and the
step by step analysis procedure, several n = 11.5 curve, we read from Fig. C5.7 that
research studies have determined the restraint ocr/a 0.7 = 1.02.
factors between web and flange elements for
simple shapes like channels, Z, H, square tUbes Therefore the local buckling stress is
and formulated design charts for such shapes.
(Refs. 1 to 5 inclusive.) ocr =39000 x 1.02 = 39800 psi.
C6.4 Design Charts for Local Buckling Stresses of PROBLEM 2.

Some Composite Web-Flange Shapes.
If the member in Problem 1 is subjected
Figs. C6.4 to C6.7 inclusive give charts to a 300 0 F temperature for 2 hours duration,
for determining the local buckling stress of what would be the local buckling stress.
channel, Z, H, square tUbe and hat shaped
sections. For formed sections, the width b From Table Bl.l for this temperature
extends to centerline of adjacent element and condition,
for extruded sections the width b extends to
inside edge of adjacent element. 0 0. 7 = 35700, Ec = 10,300,000, n = 15
C6. 5 Problems lllustrating Use of Charts. = 2.9 n 2

x 10,300,000
12{1-.3 2 ) (35700)
PROBLEM 1.
1.51.
The Z section in Fig. (a)
1s formed from aluminum alloy
2024-T3 sheet. What compres- Using this value on bottom scale of Fig.
sive stress will start local C5.7 and n curve = 15, we read ocr/o 0.7 = 1.03.
buckling of an element of the Thus ocr = 1.03 x 35700 36800 psi. =
member. Fig. (a)
PROBLEM 3.
Solution:-
Same as Problem 1, but change material to
bw = 1.5 =
- .064 1.436 Titanium Ti-8Mn Sheet.
bf = 0.75 - .032 = 0.718 From Table Bl.l we obtain for this material:-
bf/bw = 0.718/1.436 = .50 Ec = 15,500,000, 0 0• = 119500, n = 13.7.
7
= .064/.064 = 1.0 2
tw/tf
ocr
=2.9 n12(1-.3
x 15,500,000 ( .064 )2 = 81200
2
1.436------ ps •
)
i
From Fig. C6.4, we read kw = 2.9
This stress is near the proportional limit
2
ocr
= 12k~ I-V
n E
2)
(~)2
bw - - - - (A)
stress so plasticity correction should be small
i f any.
e
2 ocr/a 0.7= 81200/n9500 = .68
Ocr = 2.9 n x 10,700,000 ( .064)2 - 56100 psi
12{1-.3 2 ) 1.436-
From Fig. C5.7, using n = 13.7 curve, we
This stress is above the proportional read ocr/00'7 = .68. Then ocr = 119500 x 0.68 =
limit stress of the material, thus a plasticity 81300, thus no plasticity correction.
correction must be made. The buckling occurs
on the flange. PROBLEM 4.
From Table Bl.l of Chapter Bl, we obtain The rectangular tube has the dimensions as
for 2024-T3 aluminum alloy:- 0 0 . 7 = 39000 and shown in Fig. (b). It is extruded from aluminum
n = 11.5. alloy 2014-T6. Determine the local compressive
buckling stress.
CG.3
7r------~-~____::___r--r_-_,
WEB
7
Kw
31---+-\\\'Irt--~rt+\-~rl~--+-~----i
2
"J;
~
WEB,b"
FLANGE,bf
L2
.2 /2
Fig. CG.4 (Ref. 2) Channel- and Z-section stiffeners.

2 2 Fig. CG. 5 (Ref. 2) H-section stiffeners.
k wn E tw
°cr -- 12 (1 _ 11e -bw 2
k wn E tw
2
°cr = 12 (1 - V
2 ) 2
2
e 2) bw
Yl
~:::::t:==::t:===i====
I- is
2.0
i'
i4
/('"
4
3 L----l.----+---+----'\--+-'\-""',+---j
kr 31---+---+---\-l-~-l-~--+~----"*...----+------t
~-+-~-k:---""'Ik;::-....-t~
./1
--.l.-.~
.8 10 i2 14
Fig. CG. G (Ref. 2) Rectangular-tube-section stiffeners.

Fig. CG.7 (Ref. 5) Buckling stress for hat-section stiffeners.
k Tn 2 E t 2
t = tf = t w =tT; Ocr = 2 2
(Data of Ref. 12.)
12 (1 - V e ) bT
~fT
Solution:- In the design of rectangular tUbes, the
designer should select the tube thicknesses for
b =1 .08 = .92 ~~~O both long and short sides so that buckling
occurs on both sides, thus giving the lightest
h =2 .08 = 1.92
1:- 2.0 '~I section for buckling strength.
b/h = .92/1.92 = .479
As pointed out previously in this chapter,
tb/th = 1.0 Fig. (b)
the load on the member which causes local
From Fig. Cl.6, we read kh = 5.2. buckling is not the failing or maximum load for
a short length of the member. This local failing
2 or crippling stress is treated in the next
Ocr - 5.2
-
TI x 10,700,000 ( .04)2
12(1-.3 2 ) 1.92
= 21900 psi
• chapter. Since the buckling stress may fall in
the inelastic stress range, the buckles will not
entirely disappear when load is removed. Since
As shown in Fig. C6.6, buckling occurs on the h limit or applied loads must be carried without
side of the tube. The computed buckling stress permanent distortion, it is thus important to
is below the proportional limit stress, thus know when local buckling starts. For those
no plasticity correction. missile and space vehicles that carry no
humans, the factor of safety on limit loads is
PROBLEM 5. considerably less than for aircraft, thus the
spread between local buckling and local failing
Same as Problem 4 but change the thickness
strength becomes important in design.
of the h side to .072, but leave the b side
.04 in thickness.
C6.6 Buckling of Stiffened Flat Sheets Under Longitudinal
Compression.
Solution:-
In supersonic aircraft, it is important
b =1 = .856
.144- that the surface skin, particularly that on the
h =2 .08 = 1.92 wing, not buckle under flight conditions since
a buckled surface could effect the aerodynamic
b/h = .856/1.92 = .44-6 characteristics of the airflow around the wing,
tb/th = .04/.072 = .555 thus it is important to know when the skin or
its stiffening units initially buckle in order
From Fig. C6.6, kh = 4.3 to design so that such buckling will not occur
under flight conditions.
2
_ 4.3 x 10,700,000 (.072)2
ocr -
TI
12 (1-.3
2) 1.92 = 5860 0 psi.
Gallaher and Boughan (Ref. 6) and Boughan
and Baab (Ref. 7) determined the local buckling
This stress is above the proportional limit coefficients for idealized web, Z and T stiffened
stress thus a plasticity correction is necessary. plates. The results of their studies are shown
(Ref. 4) gives no value for a plasticity in Figures C6.8 to C6.12 and were taken from
correction but recommends the correction for a (Ref. 4). The initial local buckling stress
clamped long flange which is a slightly con- for plate or stiffener is given by the equation:-
servative correction. This plasticity
correction should also be used for the hat shape ks TI 2 E (.!a) 2
as shown in Fig. C6.7. Ocr 12 (l-))e ") bs - - - - - - - - (B)
Thus Fig. C5.8 of Chapter C5 can be used If the buckling stress is above the pro-
to correct for effect of plasticity. portional limit stress of the material, correct
for plasticity effect by using Fig. C5.8 of
From Table Bl.l, for our material 00.7 = Chapter C5.
53000 and n 18.5. =
Problem Illustrating Use of Charts.
The value of the parameter for bottom
scale of Fig. C5.8 is Ocr/OO.7 58600/53000 = = Fig. C shows a plate with idealized Z
1.10. section stiffeners. The material is 2024-T3
/f=.5
From Fig. C5.8, using n = 18.5 curve, we tf =.0625
read Ocr/00'7 =
.91. Therefore ocr 53000 x =
.91 =
48100. t w =·0625
t s =·125
Thus by changing the long side of tube
from .04 to .072, the buckling stress was in-
creased from 21900 to 48100 psi.
~ b s =4.0 .1
Fig. (c)
C6.5
6
7,...-----r---...,---....,...--..,.---.------,
Bf./CKLING OF SKIN
RESTRAINED BY STIFFEN~~"
5
t•
6 BllCKLlNG OF $rtFFENER
RESTRAINED BY SKIN
.79
.50
t,
4 r---1---+--'\'~rr---"...-'~:__-1 b,;
k.
3
.~,
.5
2.
.5
b. .4
r
Is
'
's .5
2.0 b.
/8 0
0 .2 II 6 .8 1.0 1.2
/6
br
/~
/2
Ii;
/0
.9
.8
Fig. C6. 9 (Ref. 6) Z-section stiffeners.
.7.6 twits = 0.50 and 0.79 .
.5
.8 I1J /2
6 , - - - - , - - ,-------,----.----,---,
!6.
Is
5 /0 --=F'-=+---+----j
Web stiffeners. O. 5 < twits < 2.0
.65
Fig. C6. 8 (Ref. 7) Compressive-local-buckling coefficients for 4
infinitely wide idealized stiffened flat plates.
ocr = 12 (:
k n"E
x V ")
(tb:)\" 5 r----J---'----+-\----'-P~----I
.5
e
2. _--f',;:_""".5
4
.5
7,...----,----"T---,----,..--....,...----,
o0~=w..l,-----l,---.6L...-----l.8---I..L:o,------JI2.
b,
6 f--=-+-"""'=--t----+----j br ~
bs
0.3,.4, .5
Fig. C6. 10 (Ref. 6) Z-section stiffeners.
5 twits = O. 63 and 1. O•
.-' 7
~
t
15
k. bf
6
i-i-=r-='::t====t:==1 b.
10 0.4,.5,.6
3 .5
.-' 5
2 !I
tr k. .6
.5
3 f---+--r-----,------'l:+--'<--'!4
2.
.6 .8 lO l2
Fig. C6. 11 (Ref. 7) T-section stiffeners.

bf bw
b/b ..
twltf = 1. 0; - " > 10; - " > 0.25. Fig. C6. 12 (Ref. 7) T-section stiffeners.
tf bs twltf = O. 7; br/tf "> 10; bw/bs "> 0.25.
aluminum alloy. Determine the initial buckling References

stress under longitudinal compression.
(1) Lundquist, Stowell and Schuette: Principles
bf _ 0.5 _ 0
h:- - 1.5 - .333,
bw _ 1.5 - 375 ,
- - -0 -.
of Moment Distribution Applied to Stability
uw bs 4. of structures Composed of Bars and Plates.
NACA WRL-326, 1943.
tw _ .0625 _
tS - 0.125 - .50 (2) Kroll, Fisher and Heimerl: Charts for
Calculation of the Critical stress for Local
Using the above three values and referring Instability of Columns with I, Z, channel
to Fig. C6.9, we read the buckling coefficient and Rectangular Tube Sections. NACA
k s to be 4.2. SUbstituting in equation (B) WRL-429, 1943.
(3) Kroll: Tables of Stiffness and Carry-over
= 4.2 n
2
x 10,700,000 (.125) 2 = 39800
12(1-.3") 4.0
i
ps .
Factor for Flat Rectangular Plates Under
Compression: NACA WRL-398, 1943.
(4) Becker: Handbook of Structural Stability.
This stress is no doubt above the pro- Part II. Buckling of Composite Elements.
portional limit stress so a check for plasticity NACA TN.3782, July 1957.
effect will be made. For this effect we use
Fig. C5.8 of Chapter C5. (5) Van Der Maas: Charts for the Calculation
of the Critical Compressive Stress for
For our 2024-T3 material, we find from Local Instability of Columns With Hat
Table Bl.l of Chapter Bl, that 00.7 = 39000 Sections. Jour. Aero. Sci. Vol. 21, June
and the shape parameter n 11.5. = 1954.
The bottom scale parameter on Fig. C5.8 (6) Gallaher and Boughan: A Method for Cal-
culating the Compressive Strength of Z
is equation (B) divided by 00.7' thus it equals Stiffened Panels that Develop Local In-
39800/39000 = 1.02. Using this value and the
n = 11.5 curve on Fig. C5.8, we read on left stability. NACA TN.1482, 1947.
side scale that 0cr/OO.7 = .86. Therefore the (7) Boughan and Baab: Charts for Calculating
buckling stress ocr· .86 x 39000 = 33600 psi. the Critical Compressive Stress for Local
Instability of Idealized T Stiffened
Panels. NACA WRL-204, 1944.
CHAPTER C7
CRIPPLING STRENGTH OF COMPOSITE SHAPES AND
SHEET-STIFFENER PANELS IN COMPRESSION
SHEET EFFECTIVE WIDTHS. COLUMN STRENGTH.
C7. 1 Introduction. presented in (Ref. 1). In this method the
member section is divided into equal or unequal
Chapter C6 was concerned With the local angles as illustrated in Fig. C7.2. The
buckling stress of composite sections when strength of these angle elements can be
loaded in compression. Tests of short lengths established by theory or tests. The ultimate
of sections composed of flange-plate elements strength or failing strength can then be found
often show that after the section has buckled by adding up the strengths of the angle elements
locally, the unit still has the abilit~ to that make up the composite section.
carry a greater load before failure occurs. In
other words, the local buckling and local Needham made a large number of tests on
failure loads are not the same. For cases angle and channel sections. From a study of
where local buckling occurs at low stress, the these test results as well as other pUblished
crippling or failing stress will be higher. test data on channels, square and rectangular
When local buckling occurs at high stress such tUbes, etc., he arrived at the following
as .7 to .8 Fcy , buckling and crippling stress equation for the crippling or failing stress
are practically the same. Fig. C7.1 illustrates of angle sections.
the stress distribution on the cross-section
after local buckling has occurred but prior to
local crippling or failure.
-..r:-r.
a
--L
r=
I
b -.jJ...
Fig. a
Basic Angle
t
Basic Angle
A n g l e - e 4such
unit I
I
units
ilf·o-
I
Basic Angle
Unit. Two Unit. One Unit. No
edges free. edge free. edge free.
Fig. C7.1 Fig. C7.2
Fcs/(FcyE) 1/.. .. Ce/(b~ ) 0.711 (C7.1)
where,
As the load on the section is increased, Fcs = crippling stress (psi)
the buckles on the flat portions get larger but Fcy = compression yield stress (psi)
most of the increasing load is transferred to E = young's modulus of elasticity in
the much stiffer corner regions until the stress compression (psi)
intensity reaches a high enough value to cause b'lt = equivalent bit of section =
sufficient deformation to cause failure. (a + b)/2t
Ce = coefficient that depends on the
A theoretical solution for the local degree of edge support along the
crippling stress for all types of shapes has edges of contiguous angle units.
not been developed as the boundary restraint Specifically they are:-
between flange and plate elements is unknown Ce = 0.316 (two edges free)
and also the manner in which the stress bUilds Ce =
0.342 (one edge free)
up in the corner regions is not well understood. Ce = 0.366 (no edge free)
Consequently, the methods of solution are semi-
empirical in Character, and the results of such The crippling stress for angles, channels,
methods have been sufficiently proven by tests. zees and rectangular tubes can be determined
Two methods of calculating crippling stresses directly from use of equation C7.1. The
will be presented in this chapter. crippling load on an angle unit is then,
C7.2 METHOD 1. THE ANGLE METHOD, or the (C7.2)
Needham Method.
where A is the area of the angle.
This method which will be referred to as
the angle method or the Needham method was The crippling stress of other formed
C7.1
C7.2 CRIPPLING STRENGTH OF COMPOSITE SHAPES AND SHEET-STIFFENER PANELS IN COMPRESSION
structural shapes can be determined by dividing

(a) (b)
the shape into a series of angle units and
computing the crippling loads for these
individual angle units by use of equations C7.l
and C7.2. The weighted crippling stress for
the entire section is obtained from the
following equation:-
F - ~ (crippling loads of angles)_ (C7.3)

cs - ~ (area of angles)
C7. 3 Design Curves.
Fig. C7.3 gives curves for determining the

crippling stress of angle units as per equation Fig. C7. 5 Stresses and displacements of flat plates after
C7.l, and Fig. C7.4 gives curves for determining buckling under conditions of uniform end shortening
the crippling loads for angle units. Using (reference 4). (a) Straight unloaded edges, (b) stress free
unloaded edges free to warp in the plane of the plate.
these curves and equation C7.3, the crippling
stress of composite shapes other than angles,
channels, zees and rectangular tubes can readily C7. 6 The Gerard Equations for Crippling Stress.
be calculated.
The following equations are taken from
Illustrative problems, using this method, (Ref. 3).
will be given later and the results compared
With method 2. Crippling stresses for For sections With distorted unloaded edges
composite sections should be limited to the as angles, tUbes, V groove plates, multi-corner
values given in Table C7.l unless sUbstantiated sections and stiffened panels, the following
by test results. crippling stress equation applies within ± 10
percent limits:-
C7.4 METHOD 2. For Crippling Stress Calculation.
(The Gerard Method). Fcs/F cy = 0.56
Introduction: References 2 and 3 give L(gt"/A) (E/F Cy ):L/2] 0.815 _ _ (C7.4)
the results of a very comprehensive study by
Gerard on the subject of crippling stresses. For sections with straight unloaded edges
From a thorough study of published theoretical such as plates, tee, cruciform and H sections,
studies and most available test or experimental the following equation for crippling stress
results, Gerard has developed and presented a applies within ± 5 percent limits.
more generalized or broader semi-empirical
method of determining crippling stresses. In FCS/Fcy =0.67
one sense it is generalization or broader 40
application of the Needham method which was L(gt"/A)(E/F Cy ):L/"r. - - (C7.5)

presented as method 1. The student and
practicing structures engineer should refer to For 2 corner sections, Z, J, and channel
the above references for a complete discussion sections, the following equation applies Within
of how the resulting crippling stress equations ±10 percent limits.
were obtained and how these check the extensive
test results. In this short chapter we can Fcs/F CY = 3.2
only present the resulting equations, design
curves for same and example problems in the use [(t 2/A) (E/FCy):L/"] 0.715 __ (C7.6)
of the information, in the determination of
crippling stresses. Fcs = crippling stress for section (psi)
Fcy = compressive yield stress (psi)
C7. 5 Stresses and Displacements of Flat Plates t = element thickness (inches)
After Buckling Under Conditions of Uniform A =section area (in.")
End Shortening. E =Young's modulus of elasticity
Fig. C7.5 shows a picture of the resulting g number of flanges which compose the
stress distribution on flat plates after composite section, plus the number
buckling under conditions of uniform end of cuts necessary to divide the
Shortening as determined by Coan in (Ref. 4). section into a series of flanges. See
The Gerard method recognizes the effect of Fig. C7.6 for method of cutting compo-
distortion of the free unloaded edges upon the site sections to determine value of g.
failing strength of the member section.
The cut-off or maximum crippling stress Fcs
for a composite section should be limited to the
2.0 v-
1.8 /" ----- V- -- 0.09 I I I I

~
V
/' /" V ---
'-6
-
/" <'" TWO EDGES FREE _
0.08
- -
V V
1.6 /' "- ONE EDGE FREE _ I \
/
/' V
NO EDGE FREE ,7 60
a i
/ / /" 0.07 lI
.-;-T ._ _ -
/ /' /
1.4
/
/
1/
/
/
I 40
b-J
/ / 0.06 -
1.2 I / / I 20 \\
Pcc /
/
0.05 1\\\
2
t. .JFc71.0 /
I00 Fcc \'\
f
/
A ~ \l\\
"/ 0.04
0.8
f 80
t2
I\\C\ NO EDGE FREE
/ 0.0::' ~ '"'- ONE EDGE FREE
~~
TWO EDGES FREE
/
~-t
0.6 60
t::::: :::::--
'< I'-.::
~
/ 0.02
---- ::::::::: f2:: ~
0.4
/
/
~b---J
40
0.01
-
/
0.2 20
1/ o 0 10 20 30 40 50 60 70 80
1/ b a+b
o -:t="2t
o 10 20 30 40 50 60 70 80
5l = .9±.b.. (Note: Fee same as F es )
t 2t
Fig. C7.3 Dimensionless Crippling Stress vs. b'/t (Ref. 1)
(Note: Pee same as Pes)
Fig. C7.4 Dimensionless Crippling Load vs. by'lt (Ref. 1)
C7 4 CRIPPLING STRENGTH OF COMPOSITE SHAPES AND SHEET-STIFFENER PANELS IN COMPRESSION
Distorted Unloaded Edges C7. 7 Correction for Cladding.

Angle Plate Tube
I Since many formed sections are made from
~-+i-i-f++-
alclad Sheets, the clad covering acts to reduce
L
Bas ic Section 1 Cut
Cut
I
4 Cuts
the value of buckling stress and thus a
n
correction factor must be used to take care
of this reduction in strength. This correction
from (Ref. 3) is,
2 Flanges 8 Flanges
g=2
3=g 12 =g
where,
+
Straight Unloaded Edges
T-Section Cruciform H-Section °cl = cladding yield stress
ocr = buckling stress
T Icut f = ratio of total cladding thickness to
total thickness. f = 0.10 for alclad
2024-T3 and .08 for alclad 7075-T3.
Basic Section o Cuts 1 Cut
4 Flanges 6 Flanges C7.8 Maximum Values for Crippling Stresses.
g=3
4=g 7=g The cut-off or maximum crippling stress for
a composite section should be limited to the
values in Table C7.1 unless test results are
Fig. C7.6 Method of cutting simple elements to determine g. obtained to substantiate the use of higher
crippling stresses.
Table C7.1
values in Table C7.1 unless higher values can Type of Section Max. F cs
be EUbstantiated by test results. The cut-off Angles .7 Fey
values given in Table C7.1 are no doubt slightly
conservative. Design curves for equation C7.4, V Groove Plates Fcy
5 and 6 are given in Figs. C7.7, C7.8 and C7.9. Multi-Cqrner Sections, Including Tubes .8 F CY
Stiffened Panels F cy
Tee, Cruciform and H Sections .8 F cy
2 Corner Sections. Zee, J, Channels .9 F cy
1.5 I-FR::;:::::]cHW+f~q:,-.fIt1-*-:-,Fr' -~Yfi'P+H+*~~t!I+H-"W4'"

Fig. C7.7
Curve for Crippling Stress F cs for
1.0 Angles, Tubes, V-Groove Plates,
.9 and Multi-Corner Sections .
.8
.7
Fcs = O. 56[(gt2/A)(E/Fcy) >./2] o. a6
Fcy
.6
See Table C7. 1 for Cut-off Values.
.5
F cs .4
F cy
.3
.25
.20
.15
.1
.1 .15 7 8 9 10
C7.5
1.5
: Fig. C7.8
, .1 'tiCurve for Plates, Tees, Cruci Form'
}- and H Sections.
1.0
.9 F cs = 0.67[(gt'"!A)(E/FCy ):I./2]o.4o
.8 Fcy
.7
See Table C7. 1
.6 for Cut-off Values
.5
F cs .4
F cy
.3 -
.25
.1_
.2
.15
.1 .15.2 .25.3 .4 .5.6.7.8.91.0 1.5 2 2.5 3 4 5 6 7 8 910
-{"
Fig. C7.9
1.5 Crippling Stress Fcs for 2 Corner Sections,
Z, J and Channel Sections.
Fcs = 3. 2[(t 2/A)(E c /F cy ) :1./3] 0.76

1.0 F cy
.9
.8 See Table C7. 1
.7 for Cut-off Values
.6
.5
~ .4
F cy
.3
.25
.2
.15
.j
. 1 .t±±±tlttr:t:tdJ±tl±±!
.1 1.5 2 2.5 3 4 5 6 7 8 9 10 15 20 25 30 40 50 6070 8090100
(~) (~):I./3
2 t Ec
C7.6 CffiPPLING STRENGTH OF COMPOSITE SHAPES AND SHEET-STIFFENER PANELS IN COMPRESSION
C7.9 Restraint Produced by Lips and Bulbs. From Chapter C5, the compressive buckling
coefficient using Ve = .3 is 0.388 for a flange
Quite often in formed sections, the flange element and 3.617 for a plate element. There-
element which has a free edge is rather small fore,
in width as illustrated in Fig. a. Also for
extruded sections, a bUlb is often used as tL 2 > E(t f )2
illustrated in Figs. b. The question then 0.388 E (bL) = 3.617 bf (C7.10)
arises, is the lip or bUlb sufficiently large
enough to provide a simple support to the From equations C7.9 and C7.10, the follow-
adjacent plate element. Since the compressive ing relationship is obtained,
buckling coefficient for a plate element is
bL < bf
tL = 0.328 tf - - - (C7.11)
R
Lip Lip Bulb
J~_b Fig. b
Fig. C7.11 shows the results as a curve.
8
LIP PROVIDES AT LEAST
SIMPLE SUPPORT FOR FLANGE
w L~ ,~ ~
4.0 and 0.43 for a flange element, the use of a BUCKLIN/l
OF LIP v,uv, ~:;>Y
small lip or bUlb can increase the value of the b
f~V
6
L
coefficient considerably above 0.43 and thus t 5
produce a more efficient load carrying element. 4
If'
The problem of determining the dimensions of a
lip or bulb to give at least a simply supported 30 10 20 30 40 50 60 70 BO
edge condition to the adjacent plate element has bt
been investigated theoretically by Windenburg T
(Ref. 5). The results of his studies gives the
following design criterion. Fig. C7. 11 Minimum lip dimensions required for flange to
buckle as simply supported plate (Ref. 5).
In extruded sections, a circular bulb is
- - - - - (C7.8) often used to stiffen a free edge as illustrated.
in Fig. C7.10b. The moment of inertia of the
Where IL and AL are the moment of inertia and bulb area about the centerline of the plate
area of the lip or bulb respectively. (See element is,
Fig. C7 .10).
As for the case of the lip, the buckling

stress of the bulb must be greater or equal to
the buckling stress of the adjacent plate
(b) element, which gives,
Fig. C7.10
Fig. C7.12 shows design curve representing
From Fig. C7.10a for the lip, AL = bLt, the above equation.
and IL =
tbL 8 /3.
--
In substituting these values in equation
(C7.8), the dimensions of the lip are expressed
as V
~
V
0.910 (b L )8 _ btL = 5 b: (C7.9) V
t
V
10 20 b 30 40 50
To determine bL and t, an additional re- '!f
quirement is specified, namely, that the
buckling stress of the lip must be greater or
equal to the buckling stress of the adjacent Fig. C7. 12 Minimum bulb dimensions required for buckle as
plate element. simply supported plate (Ref. 5).
C7.7
C7. 10 lllustrative Problems in Calculating Crippling Solution Method 1.

Stresses.
Fcs = Fcc = •033oVllO, 000 x 15,500,000 = 43000 psi
Problem 1. Find the crippling stress for the
equal leg angle shown in Fig. a. The material Solution Method 2.
is aluminum alloy 2024-T3.
( .093 2)( 1l0,000 )1/2 = 1.565
,Solution by Method 1.
T~
..~31"-1 .05 2 x .05 15,500,000
Material properties are:- 1" 16 From Fig. C7.7, Fcs/F cy =

0.38, whence
Fig. a Fcs = 0.38 x 110,000 =
41,800.
Fcy = 40000, Ec = 10,700,000 1 Area=.093
Fig. a Problem 4.
For this method, we use Fig.
C7.3. The parameter for bottom scale of Fig. Find the crippling stress for the channel
C7.3 is (a + b)/2t, where (a) and (b) are leg section shown in Fig. b if the material is
lengths measured to centerline of adjacent leg aluminum alloy 2024-T3. Fcy 40,000, =
of angle. For our case a = b = 1 - .025 = Ec = 10,700,000.
0.975. Thus (a + b)2t 1.95/0.1 =19.5. =
Solution Method 1.
From Fig. C7.3 using 19.5 on lower scale, \-E- 1.50 ~
and the curve for two edges free, we read on As shown in Fig. c, 3i'" T
the left hand scale that Fcc/YFCyE .0330. = the channel is composed 16 .75
Since we have only one angle, the crippling of 2 equal angle units Area = . 137 1.
stress Fcs Fcc = .0330 x Y40,000 x 10,700,000 (1) and (2). Since they Fig. b --11--.05
= 21600 psi. are the same size, we I-b---.l
need only calculate the
Solution by Method 2 (Gerard Method) failing stress for one T, I
For angle sections we use equation C7.4.

angle. 1. "
Angle
Unit 1
Angle
Unit ~
A plot of this equation is given in Fig. C7.7. a+b .725 + .725 = 14.5 Fig. c
The parameter for lower scale is, 2t 0.1
A Fcy 1./2
(gt2)(~) , where A = section area and g From Fig. C7.3 for b'/t = 14.5, we read
equals the number of flanges plus cuts, or Fcc/vFcyE = .045 (for one ed~e free)
g = 2 for an angle section. Substituting,
Then Fcs = Fcc = .045 x
( .093 2) ( 40,000 )1. /2 = 1.138
2 x .05 10,700,000 )40,000 x 10,700,000 = 29400 psi
Using this value in Fig. C7.7, we read Solution Method 2.

~cs/FCY = 0.50. Therefore Fcs =
0.50 x 40,000
- 20,000 psi. For a channel section we use equation C7.6
which is plotted on Fig. C7.9. The paraneter
Problem 2. S&TIe as Problem 1, but change for bottom scale of Fig. C7.9 is,
material to aluminum alloy 7075-T6.
Fcy = 67,000, Ec = 10,500,000

Solution Method 1. From Fig. C7.9 Fcs/F cy .65, whence,=
Fcs = .65 x 40,000 = 26,000 psi.
Fcs = Fcc = .033oV67000 x 10,500,000 - 28,000 psi
Problem 5. Same as Problem 4 but change material
Solution by Method 2. to aluminum alloy 7075-T6. Fcy 67,000, =
Ec =
10,500,000.
( .093 )( 67,000 )1/2 = 1.485
2 x .05 2 10,500,000 Solution by Method 1.
From Fig. C7.7, Fcs/F cy = 0.40, whence Fcs = .045 x V67,000 x 10,500,000 =38200 psi
Fcs = .40 x 67,000 = 26800 psi.
Solution by Method 2.
Problem 3. Same as Problem 1 but change
material to Titanium Ti-8Mn. Fcy 110,000, = (.137)( 67,000 )1/3 = 10.17
Ec =
15,500,000. .05 2 10,500,000
From Fig. C7.9, Fcs/F cy = .57, whence .8 x 11,100 =

8900 psi. Thus unless tests
Fcs = .57 x 67,000 =
38,200 substantiate higher values, the crippling stress
should be taken as 8900 psi.
Problem 6.
Problem 8. Same as Problem 6, but change
Find the crippling material to stainless steel l7-7PH(TH1050),
stress for the square Fcy :: 162,000, Ec = 29,000,000
tube as shown in Fig. d.
Material is 2024-T3 '3 I ...........05 Method 1.
aluminum alloy. Fcy = 16 Cut
40,000, Ec = 10,700,000. f--Cut -Cut Fcs = .0392J162,000X29,000,000:: 85,000 psi
Area = .373
Solution by Method 1. I Method 2.
The square tube is I
Cut (12 .373 162,000 ~/2_
x .05 2 )(29,000,000) - .93
considered as made up
of 4 equal angles with no Fig. d
edge free. From Fig. C7.7, Fcs/F CY =
.595. Fcs =
162,000 x .595 =
96,200 psi.
b' /t = (a + b)/2t = 1.95/0.1 = 19.5
Problem 9.
From Fig. C7.3, using upper curve, we obtain
Fcc/VFcyE :: .0392. Whence,
Fcs :: Fcc:: .0392 x/40,000 x 10,700,000 ::
25,600 psi
l1.5
Area A = .373. g = number of cuts plus T
flanges or 4 + 8 = 12. 1 ~~Ic=l.o==:j:=:
.375
For rectangular tUbes we use equation

C7.4 or Fig. C7.7.
Fig. e
Determine the crippling stress for the

formed section shown in Fig. e if material is
aluminum alloy 2024-T3. Fcy :: 40,000,
From Fig. C7.7, Fcs/F cy = .70 Ec = 10,700,000.
Therefore Fcs = 40,000 x .70 = 28,000 psi
Solution by Method 1 (Needham)
Problem 7. Same as Problem 6, but change
material to magnesium HK31A-0 Sheet, subjected The section is divided into 6 angle units
to a temperature of 300 0 F for 1/2 hour. by the dashed lines in Fig. e. They are
numbered (1) to (3) since we have symmetry.
The procedure will be to find the failing
From Table Bl.l of Chapter Bl, load for each angle and add up the total for the
FCY = 11,100, Ec =
6,160,000. 6 angle units. The crippling stress will
then equal this total load divided by the
section area.
Fcs :: .0392/11,100 x 6,160,000 = 10,250 psi
Angle unit (1) (One edge free)
b'/t = (a+b)/2t:: [(.375- .02)+ (.5- .02)J/.08
= 10.44
From Fig. C7.7, Fcs/F cy = .96 From Fig. C7.3, Fcc/vFcyE :: .06
Fcs :: .96 x 11,100 :: 10,600 psi Whence, Fcc = .06 x /40,000 x 10,700,000 =
39,300 psi
From Table C7.l, the cut-off or maximum
crippling for rectangular tubes is .8 Fcy or
C7.9
Area of angle (1) = .0309 =A as Section No.2. The area from that table is
0.113 sq. in.
Pcc = .0309 x 39,300 = 1215 lb.
Solution by Method 2 (Gerard)
Angle unit (2) Area = .0459 (no edge free)
For this material Fcy = 53,000, Ec
b'/t = (a + b)/2t = (.48+ .73)/.08 = 15.1 10,700,000.
From Fig. C7.3 Fcc/~ = .0475 The first question that arises is the bulb
size sUfficient to give an end stiffness to the
Fcc = )40,000 x 10,700,000 x .0475 = 31,100 psi (a) leg so that the bulb may be equivalent to
the normal corner.
Pcc = 31,100 x .0459 = 1428 lb.
In Fig. f, bf =0.78, hence bf/t = 15
Angle unit (3) Area = .0509 (no edge free)
Referring to Fig. C7.l2, we observe that
b' /t = (.73 + .605)/.08 = 16.7 for a bf/t value of 15 we need a D/t ratio of
at least 3.8. The D/t value for our bulb
From Fig. C7.3, Fcc/~ = .046 angle is (7/32)/.05 = 4.4, thus bulb has
sufficient stiffness to develop a corner. The
Fcc = .046 x )40,000 x 10,700,000 = 30,200 psi next question that arises should the bulb
angle still be classed as an angle section
Pcc = .0509 x 30,200 = 1540 lb. for which equation C7.4 applies or be classed
as a channel or 2 corner section With the bUlb
Fcs = ~Pcc/area acting as a short thick leg of the channel.
For this case, equation C7.6 would apply.
= (2 x 1215 +2 x 1428 + 2 x 1540)/0.255 =
32,800 psi The crippling stress will be calculated
by both equations.
Solution by Method 2 (Gerard)
By equation C7.4 or Fig. C7.7:-
...!..- (~)J./"
gt" Ec If bulb is considered as a full corner
then g = 4 flanges plus 1 cut 5.
g = number of flanges plus number of
cuts = 12 + 5 = 17.
SUbstituting in the above term,
From Fig. C7.7,
(17 .255 ) ( 40, 000 ) J. /" _
x .04" 10,700,000 - .573 FCS/Fcy = .82, hence Fcs = .82 x 53000 = 43500
From Fig. C7.7, Fcs/F CY = .895, whence By equation C7.6 or Fig. C7.9,
Fcs = 40,000 x .895 = 35,600 psi.
From Table C7.l, it is recommended for

multi-corner sections that Fcs maximum be
limited to .8 Fcy unless tests can prove higher From Fig. C7.9, Fcs/F CY = .7, hence Fcs =
values. .7 x 53,000 =
37,100.
FCsmax . = .8 x 40,000 = 32,000 psi. Since Possibly the best estimate of the
crippling stress would be the average of the
this is less than the above calculated values, two above results or 40,300 psi.
it should be used.
In Table C7.l, the so-called cut-off
Problem 10. Find the stress for angles is .7 Fcy and channels .9
crippling stress for the Fcy • If we use the average value or .8 Fcy ,
extruded bulb angle shown
in Fig. f if material is it gives Fcs = .8 x 53,000 = 42,400 as
20l4-T6 extruded aluminum maximum permissible because of limited test
alloy. results on bulb angles.
This particular bulb For the case where the bulb or lip does
not develop the stiffness necessary to
section is taken from assume a full corner, then the bulb is only
Table A3.l6 of Chapter A3 Fig. f
considered as an additional flange and the g Fig. b Sheet stress distribution before buckling
count would be four instead of 5, thus reducing
the crippling stress.
Sheet stress distribution after buckling
EFFECTIVE SHEET WIDTHS
C7. 11 Introduction.
The previous discussion in this chapter

has dealt with the crippling stress of formed
or extruded sections when acting alone, that is I I
not fastened to any other structure along its I Equivalent sheet effective width
length. However, the major structure of
aerospace vehicles, such as the wing or body,
involves a sheet covering which is strengthened
by attached or integral fabricated stiffeners
such as angles, zees, tees, etc. Since the
sheet and stiffeners must deform together, the
T0000 I Fig. d illill
sheet will therefore carry compressive load and
to neglect this load carrying capacity of the
sheet would be too conservative in aerospace
structural design where weight saving is very
important.
C7. 12 Sheet Effective Widths.
Fig. C7.13a illustrates a continuous flat

thin plate fastened to stiffener and the entire
unit is subjected to a uniform compressive load. , i ( I " /
Up to the buckling strength of the sheet the Sheet-stiffener panel
compressive stress distribution is uniform over
both stiffeners and sheet as in (Fig. b) assum- Fig. C7.13
ing same material for sheet and stiffeners. As
the load is ':'ncreased the sheet buckles between From Chapter C5, the buckling compressive
the stiffeners and does not carry a greater stress of a sheet panel is,
stress than the buckling stress. However as
the stiffeners are approached, the skin being
stabilized by the stiffeners to which it is (C7.13)
attached can take a higher stress and
immediately over the stiffeners the sheet can If we assume that the stiffener to which
take the same stress as the ultimate strength the sheet is attached provides a boundary
of the stiffener, assuming that the sheet has restraint equal to a simple support, then k c
a continuous connection to the stiffener. Fig. = 4.0, and if Poisson's ratio Ve is taken as
c shows the general stress distribution after 0.3, then equation C7.13 reduces to,
the sheet has buckled. This distribution
depends on the degree of restraint provided by Fcr = 3.60E(t/b)2 - (C7.14)
the stiffeners and the panel dimension.
The Von-Karman-8echler method as first
Various theoretical studies (Ref. 6) have proposed consisted of solving equation C7.14
been made to determine this stress distribution for a width (w) in place of (b), when Fcr was
after buckling. In general they lead to long equal to the yield stress of the material
and complicated equations. To provide a simple since experiments had shown that the ultimate
basis for design purposes, an attempt has been strength of a sheet simply supported at the
made to find an effective width of sheet w edges was independent of the width of the sheet.
Which would be considered as taking a uniform Thus equation C7.14 changes to,
stress (Fig. d) which would give the same total
2
sheet strength as the sheet under the true Fcy =
3.60E(t/w) , whence,
non-uniform stress distribution of Fig. c.
w = 1.90t vE/FCY - - - - - - - - (C7.15)
The question of sheet effective sheet has
been considered by many individuals. The Since the crippling or local failing
names of VonKarman, Sechler, Timoshenko, stress of a stiffener can exceed the yield
Newell, Frankland, Margurre, Fischel, Gerard, strength of the material, equation C7.15 was
and many more are closely associated with the later changed by replacing Fcy by the stress
present knowledge on effective sheet width. in the stringer FST' thus giving,
C7.11
w = 1.90t JE/FST - (C7.16) The effective sheet width as calculated by

equation C7.16 assumes that no inter-rivet
Some early experiments by Newell indicated buckling of the sheet occurs or, in other words,
the constant 1.90 was too high and for light the rivet or spot welds are close enough together
stringers a value of 1.7 was more realistic, to prevent local buckling of the sheet between
thus 1.7 has been widely used in industry. rivets when the sheet is carrying the stiffener
crippling stress. The subject of inter-rivet
If we assume the stiffness of the stiffener sheet buckling is discussed later in this
and its attachment to the sheet as developing chapter.
a fixed or clamped edge condition for the
sheet, then Fig. C7.16 illustrates a procedure to
follow for determining the effective width w
Fer = 6.35E(t/b)z or w = 2.52t vE/FST when sheet and stiffener are integral in
manufacture.
For general design purposes, it is felt
that 1.9 or equation C7.15 is appropriate for Case 1 Case 2
determining the effective width w. If t s ;; tf -< 2t s tf ~ 2t s
stiffener is relatively light, use 1.7. Fig.
C7.14 illustrates the effective width for
sheet-stiffener units which are fastened Tee Section
together by a single attachment line for each
flange of the stiffener.
2 = Rivet Lines
I-- w
Fig. C7.16
One
Rivet Line
Fig. C7.14
L
I-w ~ f+ w--l
For Case 1, find the crippling stress for
the tee section alone, assuming the vertical
stem of the tee has both ends simply supported.
For value of t in equation C7.15, use
(t s + tf)/2. The effective stiffener area
equals the area of the tee plUS the area of the
The crippling stress is determined for sheet of width w.
the stiffener alone. This stress is then used
in equation C7.15 to determine the effective For Case 2, determine the crippling stress
Widths w. The total area then equals the for the I section acting alone. Calculate w/2
stiffener area plus the area of the effective from equation C7.15 to include as effective
sheet width w. The radius of gyration should sheet area. The column properties should
include the effect of the effective skin area. include I section plus effective sheet.
Fig. C7.15 illustrates the case where C7.12 Effective Width WJ. for Sheet with One
stiffeners are fastened to sheet by two rows Edge Free.
of rivets on each stiffener flange. In this
case, the rivet lines are so close together In normal sheet-stiffener construction,
that the effective width w for each rivet line the sheet usually ends on a stiffener and thus
would overlap considerably. A common practice we have a free edge condition for the sheet as
in industry for such cases is to use the illustrated in Fig. C7.16a. The sheet ends at
effective width for one rivet line attachment
as per equation C7.16 to represent sheet width
to go with each stiffener flange. However, Fig. C7.16a
in calculating the crippling stress of the
stiffener alone, the stiffener flange which is
attached to sheet is considered as haVing a
thickness equal to 3/4 the sum of the flange
thickness plus the sheet thickness.
a distance b' from the rivet line. For a sheet
free on one edge, the buckling coefficient in
equation C7.13 is 0.43, thus equation C7.13
Staggered reduces to,
Rivet Rows Fig. C7.15
Fer = .387E(t/b' )2, and replacing b' by w , we
obtain,
- - - - - - - - - - (C7.17)
Then the total effective sheet width for before the crippling stress of the stiffener
this end stiffener would thus equal w~ + w/2. is reached, which means the sheet is less
effective in helping the stiffener carry a
C7. 13 Effective Width When Sheet and Stiffener Have subjected compressive load. Thus, in general,
Different Material Properties. to save structural weight, structural designers
select rivet spacings that will prevent inter-
In practical sheet-stiffener construction rivet buckling of the sheet. In general, the
it is co~~on to use extruded stiffeners which rivet spacing along the stiffeners in the
have different material strength properties in upper surface of the wing will be closer to-
the inelastic stress range as compared to the gether than on the bottom surface of the
sheet to which the stiffener is attached. For wing since the design compressive loads on the
example, in Fig. C7.17 the stiffener material top surface are considerably larger than those
could have the stress-strain curve represented on the bottom surface.
by curve (1) and the sheet to which it is
attached by the curve (2). Now when the The following method is widely used by
stiffener is stressed to point (B), the sheet engineers concerned with aerospace structures
directly adjacent to the stiffener attachment relative to calculating inter-rivet buckling
line must undergo the same strain as the stresses. It is assumed that the sheet
stiffener and thus the stress in the sheet between adjacent rivets acts as a column with
will be that given at point (A) in Fig. C7.17. fixed ends.
This difference in stress will influence the
effective width w. Correction for this The general col~~ equation from Chapter
condition can be made in equation C7.17 by C2 for stable cross-sections is,
mUltiplying it by FSH/FST, which gives
- (C7.19)
w = 1.90t(FSH/FST) ( / / ) - - - (C7.18) Where C is the end fixity coefficient and
ST
varies from a value of 1 for a pin end support
Where fST is the stiffener stress and fSH is the to 4 for a fixed end support.
sheet stress existing at the same strain as
existing for the stiffener and obtained from a The effective column length L' = Live,
stress-strain curve of the sheet material. thus equation C7.19 can be written
70
I I
- - (C7.20)
I
Curve 2 f-
~
60 A
oW-- I I Let p the rivet spacing be considered the
c..
V> 50 IJ~ I Curve 1 column length L. Assume a unit of sheet 1 inch
o B..,. wide and t its thickness. Then moment of
Z
ji 40 I inertia of cross section = 1 x t 3 /12, and area
:::> f A = 1 x t = t. Then radius of gyration p =
o::c 30 0.29t. Then substituting in equation C7.20 to
.... / I
obtain the inter-rivet buckling stress Fir,
z
V>
20
I
V>
UJ I
~ 10 - - - - (C7.21)
V>
/ I
I
o1/
o .002 .004 .006 .008 .010 .012
For clamped ends C = 4, thus
STRAIN - IN./IN.
Fig. C7.17 (C7.22)

(p/0.58t)"
For a rather complete study and comparison
of the various effective widths theories as To Dlot this equation, the tangent modulus
compared, see article by Gerard (Ref. 7). Et for the material must be known. However, we
Equation C7.15 is in general conservative for can use the various column curves in Chapter C2
higher bit ratios. which show a plot of Fc versus L'/p and in
equation C7.22 the term p/0.58t corresponds to
C7.14 Inter-Rivet Buckling Stress. L'/p.
The effective sheet area is considered to The fixity coefficient C = 4 can be used
act monolithically with the stiffener. How- for flat head rivets. For spot welds it should
ever, if the rivets or spot welds that fasten be decreased to 3.5. For the Brazier rivet type
the sheet to the stiffener are spaced too far use C = 3 and for counter-sunk or dimpled rivets
apart, the sheet will buckle between the rivets use C = 1.
C7.13
Figs. C7.18 and C7.19 show a plot of

equation C7.22 for aluminum alloy materials.
If the inter-rivet buckling stress cal-

culates to be more than the crippling stress
of the stiffener, then the effective sheet area
can be added to the stiffener area to obtain the Fig. C7. 21
total effective area. This total effective Detail of
area times the stiffener crippling stress will Point (1) Fig. C7.20
give the crippling load for the total sheet-
stiffener unit.
When the sheet between rivets buckles

before the crippling stress of the stiffener is
reached, the sheet in the buckled state has the problem is to determine the crippling stress of
ability to approximately hold this stress as the stiffener, the effective skin area and the
the stiffener continues to take load until it total compressive load that the unit can carry
reaches the stiffener crippling stress. This at the failure point. Since the stiffener is
buckling sheet strength can be taken advantage braced laterally by the web and the skin, column
of by reducing the effective sheet area. Thus bending action is prevented and thus the
effective sheet width equals, crippling strength is the true resulting strength
of this corner member under longitudinal
- - - - - (C7.23) compression. (Additional stresses are produced
on these corner members if web buckles under
The area of the corrected effective sheet shear stresses and web diagonal tension forces
is then added to the area of the stiffener. are acting. This subject is treated in the
The crippling load then equals the crippling chapter on semi-tension field beams.)
stress of the stiffener times the total area.
Solution:
The use of sheet effective widths in
finding the moment of inertia of a wing or Area of stiffener = 0.84 sq. in.
fuselage cross-section is a widely used pro- For 7075-T6 extrusion, Fcy = 70,000, Ec
cedure in the analysis for bending stresses in 10,500,000.
conventional wing and fuselage construction.
Reference should be made to article A19.13 of The Gerard method will be used in calculating
Chapter A19 and article A20.3 of Chapter A20 the crippling stress. Equation C7.4, or the
for practical illustrations in the use of design curve of Fig. C7.7, applies to this multi-
effective Widths. corner shape. The lower scale parameter for
Fig. C7.7 is,
C7.15 lllustrative Problem Involving
Effective Sheet. ~ (F CY )1/2
gt 2 Ec
= 0.24 ( 70,000 )1/2
6 x .072 2 10,500,000
= 0 • 63
Conventional airplane wing construction
is illustrated in Fig. C7.20. The wing is Using this value, we read from Fig. C7.7 that
covered with sheet, generally referred to as Fcs/F cy = 0.82. Thus Fcs = .82 x 70,000 =
skin, and this skin is stiffened by attaching 57,400 psi. The g value of 6 was determined as
formed or extruded shapes referred to as skin shown in Fig. (a).
stiffeners or skin stringers. A typical wing
section involves one or several interior Effective Sheet Widths: f f
straight webs and to tie these webs to the
skin, a stringer often referred to as the web Equation C7.16 will be - [cut
flange member, is required to facilitate this used to determine the sheet
connection. Fig. C7.21 is a detail of the effective widths. f
Flanges (f) = 5
flange member and tne connection at point (1) No. of cuts = 1
in Fig. C7.20. For the skin t = .05. g=6
Material is 7075-T6 aluminum
The stiffener or flange member is an alloy. Ec = 10,500,000. Fig. a
extrusion of 7075-T6 aluminum alloy. The skin
and web sheets are 7075-T6 aluminum alloy. The W = 1.9t vE/F cs = 1.9 x .05 vlO,500,000/57,400
skin is fastened to stiffener by two rows of
1/8 inch diameter rivets of the Brazier head = 1.28 in.
type, spaced 7/8 inch apart. The web is
attached to stiffener by one row of 3/16 Thus a piece of sheet 1.28/2 =
.64 wide acts to
diameter rivets spaced 1 inch apart. The each side of the rivet centerline. Observation
""",==':III!
-==1:1111 ()
.w=:=',;,;
:~::::::
;.BId
...
:"
~
gag
~ .
()
~I:I:II;
~" ~
'"C
~o
P="-:
'"C
r-<
~ :: .. Z
Fig. C7.19 (;)
~:,,:i;; 70
..,
u.l
Material Alum. Alloy Clad
~ :::: ..
Curve Designation Thickness Basis i;J
~4R ...• (CLAD) Z
..,
... 2024-T4 'C. 064
(;)
r"l'i ::z:
I
I 60 2024-T3
~
u> 2024-T36
u>
~~
r"l 2024-T6
~ ()
u> 2024-T81
e
~
2024-T86 ~
i
:
~),,~
;"!"1
...u 50
..:I 7075-T6
7075-T6 ~
.==~::~::~ P
IQ ::3
"==->"'"
~ ... t%J
u.l
;::::=~ ::;
-..--;::;s
,
E
ll:l ~
I '"C
40 t%J
:"ii
; ;; ~ : ...f:l u.l
==i(j ~
.....
I
ra. 30
20
_ _ _ _ _:..i
i,= .:I•••
:;;;;~;'
-==., -, ,_._--,
----~,
_.::
10
30 40 50 60
pit
I:=::
§~
C7.15
of the dimensions in Fig. C7.21 shows the usually of the flush surface type, either
effective width with each skin rivet line would countersunk or dimpled. If we make the skin
overlap slightly thQS we use only the width rivets of the countersunk type, the end fixity
between rivet line (see Fig. b). coefficient must be reduced to 1 to be safe.
Then the corrected p/t ratio to use with
For the web t = .064. Since the web has Fig. C7.18 would be (14 / Vl)p/t = 2 x .875/.05
a free edge, the effective width calculation = 35. From Fig. C7.18 and curve 8, Fir = 29,000
will be in two steps. whi~h is far below the cal~ulated crippling
stress, thus the rivet spacing WOQld have to be
reduced. Use 9/16 inch spacing corrected
~ = \~9 x .064 v!Io,.500,000/b7,400 :: 0.82
p/t = 2 x .5625/.05 = 22. From Fig. C7.18,
Fir = 57,400 psi, which happens to be the
Wl from equation C7.18 = .62t vE/F cs = .62 x crippling stress and therefore satisf~ctory•
•064 )10,.500,000/57,400 = 0.55 inch.
C7. 16 Failing Strength of Short Sheet-stiffener Panels
in Compression.
Fig. (b) shows w/2 \
the effective sheet \-.64~ .875 Gerard (Refs. 2, 3) from a ~omprehensive
width as calculated.
Total effective
,
\
I
study of test results on short sheet-stiffener
sheet area = 1.28 T panels in compression, has sho\lVYl that his
x .05 + .875 x .05 wl = .55 equation C7.4, or Fig. C7.7, can be used to
+ ( 0 • 55 + O. 82 )
.064 = .195. Total w
t- Fig. b give the local monolithic crippling stress for
sheet panels stiffened by Z, Y and H at shaped
2=·82 stiffeners. The method of calcul~ting the
area = .195 + 0.24 value of the g f~ctor is illustrated in Fig •
•435. 1 C7.22. Fig. C7.23 is a photograph showing the
crippling type of failure for a short panel
The total
compressive load that entire unit can carry involving the Z shape stiffener.
before failure is then equal to AF cs = .435 x
57,400 = 25,000 Ibs. C7. 17 Failure by Inter-Rivet Buckling.
This result assumes that no inter-rivet Howland (Ref. 8) assumed that the sheet
buckling occurs under the stress of 57,400 psi acts as a wide column which is clamped at its
in the sheet between rivets. ends and whose length is equal to the rivet
spacing. The inter-rivet buckli.ng stress
Th8 skin rivets are Brazier head type equation is then,
spaced 7/8 inch apart or p = 7/8. cn 2 n n ~ 2
Fir = 12 (1 _ lie l!) ( p) - - - - - (C7.24)
As discussed under inter-rivet buckling,
the end coefficient c for this type of rivet The end fixity coefficient C is taken as
should be less than 4 or assumed as 3. 4 for flat head rivets and reduced for other
types as previously explained for equation
Fig. C7.18 gives the inter-rivet buckling C7.21.
stress versus the p/t ratio. This chart is
based on a clamped end condition or C = 4. T/ is the plasticity correction factor
Since C = 3 will be used for the Brazier type -
rivet, we correct the p/t ratio by the ratio 71 is the clad correction factor
14 / /3:: 1.16. lie is Poisson's ratio (use 0.30)
The corrected p/t = 1.16 x .87 5/.05 = 20.3 t s = sheet thickness, inches.
p = rivet spacing or pitch in inches.
From Fig. C7.18 using curve (8) which is
our materi~l, we read Fir = 60,000 psi, thus For non-clad materials the curves of Fig.
skin will not buckle between rivets as C7.24 can be used. This figure is the same as
crippling stress is 57,400 psi. Fig. C5.8 of Chapter C5. For the clad correction
see Table C5.1 of Chapter C5.
The web rivets are of the flat head type
and C = 4 can be used. Spacing is 1 inch. C7.18 Failure of Short Panels by Sheet Wrinkling.
Hence p/t = 1/.064 = 15.6 and from Fig. C7.18,
curve 8, we read Fir - 64,600, which is In a riveted sheet-stiffener panel, if the
considerably more than the unit crippling rivet spacing is relatively large, the sheet will
stress. buckle between rivets, such as illustrated in the
photograph of Fig. C7.25. This inter-rivet
In wing construction, the skin rivets are buckling stress was discussed in the previous
C
fiilDililD
c- - - -c
F I F I
FFF?ii g=I7 : g;:I7
C C
I I
J
I
g=i7 ; g=I7
Average g = 16.83
I
Ig=u;
5 Cuts
12 Flanges (c) Hat-stiffened panel.
17 = g
F F :
C
g=8 g=8 I g=8 g=8 I g=7
2 Cuts Average g = 7.83

6 Flanges
8-g (b) Z-stiffened panel.
Average g = 18.83
5 Cuts
14 Flanges (a) Y-stiffened panel.
19 - g
Fig. C7.22 Method of cutting stiffened panels to determine g.
Fig. C7.23 A 24S-T aluminum-alloy Y-stiffened panel (on the left) and its 75S-T counterpart after failure.
C7.17
article. This sheet buckling does not deform approximate criterion for rivet strength from
the flange of the stiffener to which the sheet Ref. 2 is,
is attached. However, if the rivet or spot
weld spacing is such as to prevent inter-rivet S :> 0.7 bs..E (F )"
buckling of the sheet, then failure often occurs rEst d d w - - - - - - - (C7.27)
by a larger wrinkling of the sheet as
illustrated in Fig. C7.26. The larger wrinkle The tensile strength of the rivet Sr is
shape sUbjects the flange of the stiffner to defined in terms of the shank area and it may
which the sheet is attached to lateral forces be associated with either shank failure or
and thus the stiffener flange often deforms pUlling of the countersunk head of the rivet
with the sheet wrinkle shape. This deforming through the sheet.
of the stiffener flange produc~s stresses on
the stiffener web, thus wrinkling failure is a For aluminum alloy 2117-T4 rivets whose
combination of sheet and stiffener failure. tensile strength is s = 57 ksi, the criteria
The action of the wrinkling sheet to deform are:-
the stiffener flange places tension loads on
the rivets, thus rivet design enters into the s = 57 kSi., de/t av ~ 1.67
failing strength of sheet-stiffener panels
under compression. (C7.28)
190 160
s= de I t -( deIt av )",de/tav >l.67
av
where t av . is the average of sheet and stiffener
thickness in inches. The effective diameter de
is the diameter for a rivet made from 2117-T4
material.
The effective diameter of a rivet of

another material is,
Fig. C7.25 Fig. C7.26
Inter-Rivet Buckling Wrinkling Failure - - (C7.29)
Several persons have studied this wrinkling where Sr is the tensile strength of a rivet
or forced crippling of riveted panels (Refs. 9 defined as maximum tensile load divided by
and 10). A rather recent study was carried out shank area in ksi units.
by Semonian and Peterson (Ref. 11), which is
reviewed and simplified somewhat by Gerard in C7.21 Problem 1. lllustrating Calculation of
(Ref. 2). The results as given in (Refs. 11 Short Panel Failing Strength.
and 2) are used to calculate the wrinkling
stress. Fig. C7.29 shows a sheet-stiffener panel
composed of formed Z stiffeners. The material
C7.19 Equation for Wrinkling Failing Stress F w '
is aluminum alloy 2024-T3. Fe = 40,000.
From Ref. 2 we obtain, Fa.? = 39,000, n = 11.5, Ec = 10,700,000. The
problem is to determine the compressive failing.
k n" 77 ~ E t strength of a short length of this panel unit.
Fw = l~ (l-V ") (b:)" - - - - - - - - (C7.25)
e
kw, the wrinkling coefficient is obtained
from Fig. C7.27. This coefficient is a function
of the effective rivet offset f which is
obtained from Fig. C7.28. Having determined
kw from Fig. C7.27, equation C7.25 can be solved
by use of Fig. C7.24.
C7.20 Rivet Criterion for Wrinkling Failure.
A criterion for the rivet pitch found from

test data which results in a wrinkling mode Fig. C7.29
failure is,
General Panel Data:
(C7.26)
=
t w .064 bw = 2.437 bA = 0.593
The lateral force required to make the
stringer attachment flange conform to the t s = .064 bf = 0.905 bo = 0.343
wrinkled sheet, loads the rivet in tension. An bs = 2.00
~ 2
1.2
/
~
V-- 5
l-
0
n
~ --::::::: -~~ .../0
LO aJ
I-
ld~ ~
s
I-
I'~~T Fig. C7.24

Curves for
I-
1.#V finding Fir
as per eq.
.4
.2
I--
,
~
.r.
~
tr
C7.24 and
F w as per
eq. C7.25.
o V
o .2
I
s
I
e
I
1.0
1
1.2
I I
1.$
I
I.a
I \
2.0
c n" E ts "
12 (1 - V e ")F,0 . 7 (p-) or
10
9
1\ \ 1\ 1\
11
\\ \ \
7
.\ \\ \ \
6
1\ \ 1\ 1\
\
f \ \ 1\\
.H- b;' O'f .5
\
l\ ~ \ r\
4
1'.25
1\\ 1\ 1\.\ 1'0. .2 J"'." J'... '"

\ ~ J"'." ~ ~ J'-... ,...... !'--- r----
.15
2
r-- r---
~
...........
~ t - r-=::: ~ t-- r-- I-- t -
I
"'"
o .2 If .6 .6 1.0 1.2 /,4 1.6 ll1 2.0 2.2 2.4 2.6
~
b, /1,
Fig. C7.27 Experimentally determined coefficients for failure in

wrinkling mode. (Ref. 2)
I
--- -- -
10
llz..L
tor -
r-- 10
I--
9 -II 9-
-:::::---
r- I-- ~ r-- 11
7-
11 - ' 0
!............
,,-.-- ~
I-- ~ ~ ~
_6
!1
.L 7 -9
t.
-,,/ V V
-;::: t::=-I--
" 7 /V" V V
~
s
V
1-5
6
/
6 7
I' >-; 1=
I
" 9 10 11 12 1$ 14 Its
p
7'
Fig. C7. 28 (Ref. 2) Experimentally determined values of effective
rivet offset.
C7.19
whence, bw/tw = 38, Fir = 39,000 x 1.1 =

43,000. This value is far
above the stiffener or panel crippling stress
as previously calculated so inter-rivet
buckling is not at all critical.
The rivets are 3/32 diameter Brazier head type
AN456, 2117-T3 material spaced at 0.75 inches. Failure by Sheet or Face Wrinkling. (Fw)
Area of Z stiffener = 0.252 in." The wrinkling failure stress by equation

Area of skin for b s = 2 inches = 2 x .064 C7.25 is,
0.128 =
Total area of sheet and stiffener = 0.380 in."
Crippling stress of Stiffener Acting Alone.

Fcs(ST) . To determine value of kw, we use curves in
Figs. C7.28 and C7.27.
Since we have a Z type of stiffener,
equation C7.6 and Fig. C7.9 applies. The lower p/d = .75/.0937 = 8, bo/t w = 5.36
scale parameter in Fig. C7.9 is,
From Fig. C7.28, we read f/tw = 6.5, whence
f = .064 x 6.5 = 0.416
f/bw = .416/2.437 = .17
(tw/tw)/bs/ts =
38/(2.0/.064) 1.21 =
From Fig. C7~9, Fcs/F CY =_0.6. Hence, Fcs =
.6 x 40,000 - 24,000 psi - Fcs(ST)' From Fig. C7.27, we read kw = 4.4.
Crippling Stress of Panel Considered as a To solve equation for Fw we use Fig. C7.24.
Monolithic Limit. FCS(M)' The lower scale parameter is,
Equation C7.4 or Fig. C7.7 applies for kw n" E (t s )" =

monolithic failure of sheet-stiffener panels. 12 (1 - /)e ") (F 0 • 7) bs
The lower scale parameter on Fig. C7.7 is, 4.4 n" x 10,700,000 (.064)" 1 15
12 (1 - .3") (39,000) 2.0 = .'
0.380
( 7.83x )( 40,000 )~/"
.064" 10,700,000 For n = 11.5, we read from Fig. C7.24 that
= 0.73 Fw/Fo . 7 = .9, whence Fw =
.9 x 39,000 35,100 =
psi. Thus wrinkling failure is not critical as
g = 7.83 (see Fig. C7.22b) Fw is larger than FCS(M) and Fcs(ST)'
The results show that the crippling stress
From Fig. C7.7, we read Fcs/F cy = .725, hence for the stiffener alone of 24,000 psi is the
Fcs = 40,000 x .725 = 29,000 psi = FCS(M)' smallest value, or the stiffener is unstable as
it fails first. The entire panel unit will
Inter-Rivet Buckling Stress (Fir) not reach its failing strength when stiffener
stress is 24,000 because the skin wrinkling
The rivet type is Brazier head and the stress f w is higher. An approximation suggested
spacing p is 3/4 inch. in (Ref. 2) is to assume stiffeners carry the
same stress as the skin up to Fcs(ST) and beyond
Equation C7.24 applies and Fig. C7.24 is this the stiffener carries no additional load.
used to solve the equation. The lower scale Thus the panel failing stress F(F) can be
parameter in Fig. C7.24 is, calculated from the following equation.
For Brazier head rivet C = 3. v= .3.

Substituting:-
3n" x 10,700,000 (064)~

- -
12 (1 - .3")(39,000) '. 75 ~ = 5. 45 The total load carried by each stiffener plus
its sheet is 27,800 x .380 =
10,600 Ibs.
The shape parameter n for our material is
11.5. Reference to Fig. C7.24 for a value of The failing strength of the riveted panel
5.45 on bottom scale which is off the scale, cannot exceed the monolithic panel failing stress
we estimate the Fir/Fo.7 as above 1.1. Thus FCS(M), which was 29,000 psi for our panel or
greater than the calculated failing stress of (5) plD ::= 8 - - - - - - - - - - Face wrinkling
27,800 psi.
(6) Tensile strength of rivet or spot weld
Check of Rivet strength. attachment per inch should be ~ 0.05 Fcyt s
in order to prevent failure in wrinkling.
From expression C7.26 (7) As a rough guide do not use bent up
stringers if bslt s < 30 in order to prevent
plb s < 1.27/kw1/2 face wrinkling weakness.
0.75 _ I C7. 23 Y Stiffened Sheet Panels•
2:0- •375 < 1.27/(4.4)1 2 = .603(satisfactory)
A Y shape cannot be formed from sheet,
The criterion for required rivet strength thus it must be extruded. To make the Y shape
to make the stiffener flange follow the efficient, the various parts usually have a
wrinkled sheet is from C7.27 different thickness. Furthermore, the extruded
material has different mechanical properties
in the inelastic stress range as compared to
rolled sheet that is used for the panel, thus
these effects must be considered in calculating
s ( 0.7 )( 2 )(0.75)( )2 the crippling stress of the stiffener and the
r::> 10,700,000 3/32 3/32 35,100 complete panel unit.
Sr :> 13.7 ps i . The effective thickness t w of the stiffener

is determined by the following equation (Ref. 3):
From expression C7.28
- - - - - (C7.29)
s = 57 ksi for de/t av . ::= 1.67
where bi and ti refer to the length and thick-
For given panel de/t av . = 3132/.064 = 1.46, ness, respectively, of the cross-section
thus rivets have plenty of tensile strength to elements.
produce a wrinkling failure.
When the yield stress Fcy of stiffener and
C7. 22 General Design Limitations to Prevent Secondary sheet are different and effective FCY can be
Failure in Sheet-Stiffener Panels. estimated as follows (Ref. 3):
Sheet stiffener units can be designed as Fcy = {FCY(S~) + FCY(ST) [(twits) -l]}/tw/t s
columns if the secondary forms of failure such
as inter-rivet buckling and face wrinkling are - - - - - - - - - - - (C7.30)
avoided. The following design rules referring
to Fig. (A) will usually avoid these secondary The monolithic crippling stress for the
weaknesses. sheet stiffener panel can be calculated from
equation C7.4 or by the curve in Fig. C7.7.
Equation C7.4 is
FCS(M)/F cy = 0.56 [( gt 2 IA)(E/F cy )1 / 2]°.86

The constant 0.56 in thiS equation applies
for Y and hat-stiffened panels when the ratio
twits = 1.0. For other ratios the correction
of this constant which is referred to as j?g
in Gerard's basic equation can be obtained from
curve in Fig. C7.30.
1.2
I I I
o HA r STIFFENED PANELS
Fig. A • r STIFFENED PANELS
.e, 1.0 ~ i-A-

(1) tf 1 Its ~ 0.5 - - Prpmotes overall crippling ~
- f".--
&OItIl£t:TION
(2 ) 0.4 ~ bf Ibw < 0.5 -- Rolling versus local

2 buckling 04 0.6 08 1.0 l2 1.4
~,/'.
(3) Make b o as small as possible -- Face
wrinkling
Fig. C7. 30 fig correction for twits effect on Y- or hat-
(4) plb s < 0.5 - - - - - - Inter-rivet buckling stiffened panels.
C7.21
C7. 24 Example Problem Y Stiffened Panel. unit. Ec will be taken as 10,600,000 which is
the average E for stiffener and sheet. Substi-
The compressive monolithic failing stress tuting in the above parameter:-
of a Y stiffened panel, as illustrated in Fig.
C7.22a, will be calculated by the Gerard method. ( .808 ) ( 64,700 ):1./2 = 6
Fig. C7.31 shows details of the panel unit. 18.83 x .0763 x .064 10,600,000 .68
The stiffener is extruded from 2014-T6 aluminum
alloy for which Fcy = 53000 and Ec = 10,700,000. From Fig. C7.7 we read
The skin or panel sheet is aluminum alloy
7075-T6 for which FCY = 67,000 and Ec =
10,500,000.
Thus the monolithic crippling a failing
~ 4. 21" ------l
stress FcsM = FCY x .76 = 64,700 x .76 =
j .136t
.068
- -r
~ 1+-1.111 -.\
~
I 48,200 psi.
The curve as plotted in Fig. C7.7 is for a

tits ratio of 1.0. The ratio for our panel is
.0763/.064 = 1.19. The correction factor from
Fig. C7.30 is 1.03. Therefore FCS(M) =
1.03 x
48200 =
49,600 psi.
Load carried by one stiffener-sheet unit
= 0.808 x 49600 = 40100 lb.
COLUMN STRENGTH
Fig. C7.31
C7.25 Column Curve for Members With
Unstable Cross-Sections.
Stiffener area = 0.538 in. 2
Sheet area = 4.21 x .064 = .270 in. 2 Chapter C2 dealt with the column strength
Total area (A) per stiffener unit = 0.808 in. 2 of members with stable cross-sections. For
Radius of gyration of stiffener alone = 1.123 in. example, if we took a round tube with relatively
heavy wall thickness and tested various lengths
Solution:- in compression to obtain the failing stress and
then plotted these stress values Fc against the
Since the elements of the Y stiffener have slenderness ratio L'/p of the member, the test
different thicknesses, the effective t w by results would closely follow the curves ABFC in
equation C7.29 is needed. Fig. C7.32. The type of failure would be
elastic overall bending instability at stresses
Z bit i between points B-C and inelastic bending
tw = =
instability at stresses for most of the range
Zbi
BA. EUler's equation as shown in Fig. C7.32
1.111 x .136 + 1.432 x .068+3.2x .064+ 1.188 x .064 can be used to determine the failing stress for
1.111 + 1.432 + 3.2 + 1.188 both elastic and inelastic bending instability
with the tangent modulus Et being used in the
or tw = .0763 inches inelastic stress range.
Since stiffener and sheet have different Now suppose we test various lengths of a
material properties, an effective Fcy from member composed of the same material as used for
equation C7.30 will be calculated. obtaining curve ABC, but use a member with an
open cross-section, such as a channel, hat
Fc= {67' 000 + 53000 [( .0763)_1] 0763
}/· .064 = 64700 section, etc., with relatively small material
y .064 thickness. The test results for such members
would often follow a curve similar to DEFC in
The curve in Fig. C7.7 will be used to Fig. C7.32. Thus it is obVious that EUler's
solve the equation for FcS(M)' The !ower s~ale equation cannot be used in the range DEF, as the
parameter for Fig. C7.7 modified by twand Fcy true failing stresses are far less than that
is, given by the Euler column equation. A short
length of the member with L'/p less than 20 will
fail at practically the same stress, thus the
failing stress for lengths up to L'/p = 20 will
be practically the same and this stress has been
The value of g from Fig. C7.22a is 18.83 given the name of crippling stress (F cs ) and the
as the average value for a 6 stiffener panel previous portion of this chapter has been con-
n 2E t The equation gives a parabolic curve

A Fe (U/p)2 starting from the crippling stress at L'/p :: 0,
and becomes tangent to the Euler curve at a
\ stress value equal to one half the crippling
\ stress. Fig. C7.33 shows a plot of equation
\ C7.31 for aluminum alloy material for various
\ values of the crippling stress which is the Fc
stress at L'/p :: O.
\
D
0EW",
~ Curve
n 2
E
This method is qUite simple to use as the
only additional calculation required is the
crippling stress of the column section which is
obtained by methods previously explained in
Crippling
Stress
I Fe = (V/p) 2 this chapter. The other methods usually involve
the buckling stress as well as the crippling
Region stress. Since the crippling stress is normally
I constant below L'/p :: 20, the assumption that
Fcs is zero at L'/p = 0 is slightly conservative.
I METHOD 2.
1-- Transition Region
The following method or slight variations
of it appears in the structural design manuals
C
of a number of aerospace com)anies. The method
or procedure for determining the column failing
L'/p stresses in the so-called transition range
involves the use of the basic column curve for
Fig. C7.32 stable cross-sections. The procedure can best
be explained by reference to Fig. C7.34.
cerned with methods of calculating this local
crippling or failing stress. At point (F) the A
elastic buckling of some part of the cross-
section begins. Between stress points F to E Fey 0
-+- -
the action for the member involves both overall
elastic bending instability plus local buckling I
which becomes more extensive as the stress I
increases. The portion EF of the column I
strength curve is often referred to as the I
transition range. At present no reliable I
theoretical theory has been developed for D
F es -. - - .-.J~
determining the failing stresses in this
transition range, thus resort is made to semi-
empirical methods which have been checked
against test results and found to give reason-
ably close results.
C7.26 Methods Used for Determining the Column
Failing Stress in the Transition Region.
METHOD 1. JOHNSON-EULER EQUATION. F

.9F er ... -
G
Poss.ibly the first method used in calcu-
lating the column failing stress Fc in the
transition range EF in Fig. C7.32 was the well C
known Johnson-Euler equation which involves the L '/p
crippling stress. The equation is,
F 2 Fig. C7.34
Fc :: Fcs - 4~~E (L'/p)2 (C7.31) The curve ABC is the Euler column curve for
a column with a stable cross-section and for a
where, Fc :: column failing stress (psi) given material. It involves using the tangent
Fcs :: crippling stress, assumed to occur modulus Et in the inelastic stress range. The
at L'/p :: 0, where L' = L/ve-. following steps are taken to determine the
column curve for the so-called transition range:-
C7.23
(1) Locate point (0) on the basic column METHOD 3.

curve by drawing a horizontal line through
an Fc value equal to Fcy , the yield stress Another method that is widely used also
of the particular material being used. uses a parabolic curve to represent the column
strength in the transition range (see Ref. 12)
(2) Draw a horizontal line starting at point
D. Point D is at an Fc value equal to the The parabolic approximation has the follow-
(F cs ) crippling stress for the column ing form:-
section being considered. Point (E) on
this line is determined by projecting
vertically downward from point (0). - - - - - - - (C7.32)
(3) Locate point (F) at a value of Fc = .9 Fcr
where Fcr is the buckling stress for the where
cross-section. Draw a horizontal line
through point (F) to intersect the column Fc is the column failing stress.
curve at point (G). Fcs is the crippling stress.
Fer is buckling stress for the column
(4) Connect points E and G with a straight cross-section.
line. The line EG then represents the FE is the EUler column stress for the
column failing stress Fc for values of particular column being considered as
L '/p betl'veen points E and G. found from equation FE = n2E/(L'/p) 2
This method requires the use of the column The equation applies for Fe > Fcr . For
curve for stable sections and the determination cases where Fer> FpL where FpL is the pro-
of the buckling stress and thus requires more portional limit stress for the material use FpL
calculations than Method 1. It also requires instead of Fer in equation C7.32.
a graphical construction.
C7.27 Example Problems Involving the Finding C7.33 for a value of L'/p = 50 and a value of
of the Column Strength of Columns With Fc = Fcs = 38,200 at L'/p = O.
Unstable Cross-Sections.
Column Strength by Method 2
PROBLEM 1.
This method requires a graphical
A rectangular tube construction which involves the column curve
21 inches long has the for the given material and for a stable cross-
cross-section as shown section.
in Fig. (a). The
material is aluminum Fig. C7.35 shows the column curve for our
alloy sheet 7075-T6, Fig. a material for a column with stable cross-section.
with Fcy = 67,000 and Ec = 10,500,000. If the It is identical to Fig. C2.10 of Chapter C2
member has a pinned end condition, what 1s the for the room temperature condition, except the
column failing stress. curve in Fig. C7.35 has been drawn to a smaller
vertical scale.
Solution. The area of cross-section = .234 in. 2
The least radius of gyration is 0.42 inches.
Before the column failing stress Fc for 70

the 21 inch length can be found, the crippling
stress Fcs and the buckling stress Fcr must be 60
determined.
50
The crippling stress will be calculated ~
by the Gerard Method. The parameter for use "00

with Fig. C7.7 is, C
t.l
r...
~ (FCY)~/2= .234 ( 67 ,000 )~/2 =0 976 30
gt 2 Ec 12 x .04 2 10,500,000 •
20
From Fig. C7.7, Fcs/Fcy = .57, hence Fcs =
.57 x 67,000 = 38,200 psi.
10 20 30 40 50 60 70 80
The initial buckling stress Fcr will be L'/p
determined by the theory of Chapter C6.
The graphical construction to obtain the
b = 1 - .08 = .92, h = 2 - .08 = 1.96 column curve in the region between the crippling
stress and the buckling stress is as follows:
b/h = .92/1.96 = .479, tb/th = 1.0
At a column stress Fc = Fc = 67,000, draw
From Fig. C6.6 of Chapter C6, Kh = 5.2 a horizontal line to intersect ~he basic column
curve at point (0). Draw a horizontal line
~ n 2 Ec b 2 from point (D) which equals the crippling stress
Fcr = 12 (1 _ V 2) ("h) of 38,200 psi. Locate point (E) by projecting
e
vertically downward from point (0). Locate
5.2n 2 x 10,500,000 ( .04 )2 point (G) at an Fc stress equal to .9 Fcr =
Fcr = 12 (1- .3 2 ) 1.92 = 21,500 psi. .9 x 21,500 = 19,350 psi. Connect points (E)
and (G) by a straight line. This line repre-
Column Strength by Method 1 (Johnson-Euler eq.) sents the column failing curve for the member
of our problem for L'/p values between points
L = 21 inches L' = L/-JC = 21//1 = 21 (E) and (G). Taking L'/p = 50, we project up-
ward to EG line and then horizontally to scale
L'/p = 21/.42 = 50 at left to read Fc = 27,700 psi.
The Johnson-Euler equation is, Column Strength by Method 3
Fcs 2 2 In this method equation C7.32 is used.
4 n 2 E (L'/p) , sUbstituting,
_F C _-
38,200 2 2 Fcs I - (1 - Fcr/F cs ) (Fcr/FE) - - - - - (A)
Fc = 38,200 - 4 n2 10,500,000 (50) = 29,400 psi
Fcr = 21,500, Fcs = 38,200
This value could be read directly from Fig.
C7.25
FE is the Euler column stress for column SUbstituting in Johnson-Euler equation,

with a stable cross-section. To find FE, we
project upward from L'/p = 50 to the curve
AOC in Fig. C7.35 and then horizontally to
read Fc = FE = 41,700 psi.
Column Strength by Method 2
SUbstituting in eq. (A),
On Fig. C7.35 point E' is located at
F~ Fc = Fcs = 30,600. Point G' is located at
38,~00 =1 - (1 - 21,500/38,200)(21,500/41,700) .9 x 20,900 = 18,800 psi. The line E'G' is the
column curve. For L'/p = 45.8, we project up-
Fc = 38,200 (1 - 0.225) = 29,600 psi ward to line E'G' and then horizontally to left
scale to read Fc = 25,000 psi.
Example Problem 2.
fo--- bF "1
\.-.75 -::l Column Strength by Method 3
A formed Z section as
shown in Fig. (b) is used .04 I ~ ~
To find FE, use Fig. C7.35 with L'/p = 45.8
as a column. The length and basic Euler curve, gives a value of FE
L = 30". The member is x--- -* bw 1.5 47,000 psi.
braced in the x-x direction
thus column bending failure
must occur about x-x axis. I- .75 ~
~=1 SUbstituting in equation (A):-
Material is aluminum alloy, 20,900/30,600)(20,900/47,000)

Fcy =
67,000, Ec 10,500,000. = Fig. b
30,600
1 - (1
Assume the end fixity co-
efficient c = 1.5. FC = 30,600 (1 - .141) = 26,300 psi
Find the column failing stress. C7.28 Column Strength of Stiffener
With Effective Sheet.
Solution:
Column members are often attached to thin
Area of Section = 0.117 in.'" sheet by rivets or spot welding. If the rivet
Px =.535 in. spacing is such as to prevent inter-rivet
buckling, then the sheet will assist the stiffener
The initial buckling stress is needed in to carry a compressive load and to neglect the
finding the column strength. sheet would be too conservative.
bw = 1.5 - .04 = 1.46, bf = 0.75 - .02 = 0.73 If the attached sheet is relatively thin,
that is, less than the stiffener thickness, the
bF/bw = .73/1.46 = 0.5, tw/tF = 1.0 method of using the effective sheet width as a
part of the column area is widely used by
From Fig. C6.4 of Chapter C6, we read structural designers and will be used in this
Kw = 2.9. example solution.
F
cr
= 2.9 12n 2
x 10,500,000 (~) 2
(1 _ .3 2 ) 1.46
= 20 ' 900 ps
i Example Problems.
For an example
The crippling stress will be calculated by problem, we will assume
the Gerard Method using Fig. C7.9, which applies that the Z stiffener in
for Z shapes. The lower scale parameter is, Problem 2 is one of
several stiffeners riveted
~{cY)~h = (.117)( 67,000 )~/" = 13.6 to a sheet of .025 thick-
t 2 Ec .04 2 10,500,000 ness and of the same
material as the stiffener.
From Fig. C7.9, we read, Fig. c
Fcs/F cy = .455, hence, Fcs = 67,000 x .455 = Solution.
30,600 psi.
The rivet or spot weld spacing is made
Column Strength by Method 1 such as to prevent inter-rivet buckling. Thus
the column area will be as shown in Fig. c,
L' = L/ve-= 30 ~= 24.5 namely, the stiffener area plus the area of
the sheet for the effective width w. Since
L'/p = 24.5/.535 = 45.8 the effective sheet width w is a function of
the stiffener stress and since the stiffener
stress is a function of the radius of gyration, Where, Po = radius of gyration of stiffener

the design procedure is of the trial and error alone.
category.
p = radius of gyration of sheet and stiffener.
First Trial.
S = distance from centerline of sheet to neutral
Assume the effective sheet width is based axis of stiffener.
on column strength of Z stiffener acting alone.
The average column failing stress by the 3 t = sheet thickness. w = sheet effective Width.
methods in the Problem 2 solution was (25,860 +
25,000 + 26,300)/3 =
25,700 psi. Equation C7.33 has been put in curve form
as shown in Fig. C7.36, which will now be used
The effective width equation to be used is, to compute the radius of gyration Po for the
stiffener plus effective sheet.
w = 1.9 t vE/FST
S = .75 + .0125 = .7625
= 1.9 x .025 x/L0,500,000/25,700 = .965 in.
Po for stiffener alone = .535
Effective sheet area = 0.965x .025 = .0242
S/Po = .7625/.535 = 1.425
Area of stiffener =Ao = .117
w = .965
Total area = .1412
wt/A o = .965 x .025/.117 = .2063
Adding the effective sheet to the stiffener
will change the radius of gyration. Mr. R. J. From Fig. C7~36 for the above values of
White (Ref. 13) has developed equation C7.33 S/Po and wt/A o , we obtain
which gives the variation in the radius of
2
gyration in terms of known variables for any (p/PO)2 = 1.05, whence p = .5352 x 1.05, which
stiffener cross-section. Since the failing gives p = .548 in.
stress of a column is directly proportional to
the radius of gyration squared, equation C7.33 Then L'/p = 24.5/.548 = 44.8.
can be equated to the ratio of the column
Use Method 1 for column strength:-
Sp:or)e:s__es. Fc __ 1 + [1 + (psot] ~ 30,600 2 2
- - - (C7.33) Fc = 30,600 - 4 n2 x 10,500,000 44.8 = 26,050 psi
+(~~r
(
FST 1
I I I I
Fig. C7. 36 (Ref. 13)
Curves for the I.,
3.0
--
Determination of
V
- e~-r-~
Po - FST
/ !""'-
.......
--
/
/
./
.... r..... ..... "'-
I J I'
I 1// r---. I'--- .... ~
2.0
V .... .....1' .... .... n-
I--,
....
01 ""
rz..(!;
II
U/
,,-
-- -
r-
....... r--..
r- ....
. . . . r--.
I'---
....... '-.
-- - - -- - - 01""
r-. ......
io-..
I""-
r- r--..
-.....;
io-..
r-
') I--
rz.. rz..r:Il
II
,.---. ....
- -- - -- -
til til
V 4.
~
Qlo l,.;"
-I""-
........
"'- l""- t- I""-
I""-
2 I""-
~
0
0-1
l""-
-- I-- t- l""- I""- I- ~
~1.0 ....1- '-.
-I-
r- r- r-
--- - -- r- I-
..... t-- I""- "",
l"""- I--
I-- .....:.
t-:.~Ft
STIFFENER ALONE
I- -I--
-
I""-
....1- t- I--
r- t-- I-
wt
r- I -
-l-
~
,-,
Ao - Area of Stiff-
ener Alone
p 0 = Radius of Gy-
(/J2 _ Fe _ 1+[I+{pSOr]
- FST- (1+~r
Ao
ration of 0 Q
Stiffener Alone 0 1.0
wt 2.0 3.0 4.0
Ao
C7.27
The revised effective width based on this (4) Same as (2) but subjected to 300 0 F for
stiffener stress is, 1/2 hour duration.
w = 1.9 x .025 VIO,500,000/26,050 = .99 in. t:= 1-1/2 --l I-- --I
1B·w 1[[ F~.

2
This value is only .03 inch more than the

value of .96 previously, thus the effect on the
radius of gyration p will be negligible. The
column failing stress for the sheet stiffener
combination is therefore 26,050 psi, and the
f l05
compressive failing load would be 26,050 x (4) Find the crippling stress for sections of
.1412 = 3675 Ibs. Figs. g and h when formed from 2024-T3
and 7075-T6 aluminum alloy material.
C7. 29 Sheet-Stiffener Panels With Relatively Heavy
Sheet Thickness.
~ 1"---l
For Z, Y and hat shaped stiffener-sheet
panels as previously discussed relative to local
failing stress, use entire sheet and stiffener
H·w lb r
area in computing radius of gyration or, in
other words, due not use effective sheet widths
but use entire sheet as effective.
r- t -l Fig. g
PROBLEMS (5) Find the crippling stress for the extruded

sections numbers 4, 8, 16, 22, 28 and 34
(1) Find the crippling stress for the angle as given in Table A3.16 of Chapter A3.
sections of Figs. (a) and (b) when formed Material 2014-T3 aluminum alloy.
from the following materials. Use both
Needham and Gerard Methods. (1) 2024-T3 (6 ) Fig. i shows a corner member, in a
aluminum alloy. (2) 7075-T6 aluminum stiffened wing section. The skin is
alloy. (3) 17-7PH(THI050) stainless steel. fastened to the stiffener by one row of
(4) Same as (3) but subjected to an 1/8 diameter rivets at 3/4 inch spacing.
elevated temperature of 7000 F for 1/2 hour The web is fastened to stiffener by 2
duration. staggered rows of rivets, With rivet spac-
ing in each of 1-1/8 inch, and rivets are
f
7/8"
1-
LLig.a
3
16 1..
7/8"
tli::
1
~
16
ig b
.
.072
1.
of the flat head type. Material is 2024-T3
aluminum alloy. Also use 7075-T6 aluminum
material.
Find crippling stress for stiffener.
I- 7/8 -J T064 I- 1-1/2 --i T Will inter-rivet buckling occur. Find
effective sheet widths and total failing
load unit will carry.
(2) Find the crippling stress for the two
channel shapes of Figs. c and d when .035
formed of following materials. (1) AISI
4130 steel, heat-treated to Ftu =
f::.r:~~~2~nrf\-Skin
180,000. (2) Same as (1) but SUbjected to
an elevated temperature of 500 0 F for 1/2
duration. (3) Ti-8Mn Titanium.
Tf
(4) 7075-T6 aluminum alloy.
7/8
~
16
1-3/.
Fig. c
Trcl
3/4
..l
I--l-l/4~
16
Fig. d
.04
Fig.
--l ~.064
(7 ) Determine the compressive failing strength
of a short Z stiffened sheet panel unit such
(3 ) Find the crippling stress of the as that illustrated in Fig. C7.29. The
rectangular tubes in Figs. e and f when various dimensions are as follows. Refer
formed from following materials. to Fig. C7.29 for meaning of symbols.
(1) 2024-T3 aluminum alloy. (2) HK3lA-0
magnesium. (3) Ti-8Mn Titansium. tw = .072 bF = 1.0 bs = 2.50
t s = .072 bA = 0.625
bw =2.50 bo = 0.375
Stiffener and sheet material is 2024-T3 (5) Windenberg, D. F.:- Proc. 5th Inter-
aluminum alloy. Rivets are 1/8 diameter national Congress for Applied Mechanics.
Brazier Head type 2117-T3 material and pp. 54-61, 1939.
spaced 7/8 inch.
(6) Experimental StUdy of Deformation and
(8) S&TIe as (7) but change material to 7075-T6 Effective Widths in Axially Loaded Sheet
aluminum alloy. Stringer Panels. NACA TN 684, 1939.
(9) Same as (7) but change material to Ti-8Mn (7) Gerard, F.:- Effective Width of
Titanium. Elastically Supported Flat Plates. Jour.
Aero. Sci., Vol. 13, Oct. 1946.
(10) In the Example Problem on a Y stiffened
panel as given in Art. C7.25, change sheet (8) Howland, W.L.:- Effect of Rivet Spacing
thickness from .064 to .081 and calculate on Stiffened Thin Sheet in Compression.
the resulting panel strength. Jour. Aero. Sci., Vol. 3, Oct. 1936.
(11) The hat stiffener in Fig. g of Problem (4) (9) Argyris, J. H. and Dunne, P.C.:-
is one of several stiffeners riveted to Structural Principles and Data. Handbook
skin of thickness .032 and of the same of Aero., 4th Edition, 1952.
material as the stiffener. If the length
of the panel is 20 inches, what will be (10) Bizlaard, P. D. and Johnston, G. S.:-
the column failing stress if end fiXity Compression Buckling of Plates Due to
c = 1.5. Also for c =
2.0. Use method Forced Crippling of Stiffeners. Preprint
involVing effective sheet Widths. No. 408. S.M.F. Fund Paper. Inst. of
Aero Sci. Jan. 1953.
(12) Same as Problem 11 but use Z stiffener of
Fig. h of Problem 4. (11) Semonian, J. W. and Peterson, J. P.:-
An Analysis of the Stability and Ultimate
References Compressive strength of Short Sheet-
Stringer Panels With Special Reference to
(1) Needham, R. A.:- The Ultimate strength of the Influence of Riveted Connection
Aluminum Alloy Formed Structural Shapes Between Sheet and Stringer. NACA TN 3431,
in Compression. Jour. Aero. Sci. Vol. 21 March, 1955.
April, 1954.
(12) Gerard, G.:- Handbook of Structural
(2) Gerard, G.:- Handbook of Structural Stability, Part V. Compressive Strength
Stability. Part IV. Failure of Plates of Flat Stiffened Panels. NACA TN 3785.
and Composite Elements. NACA TN 3784. Aug. 1957.
August, 1957.
(13) Thesis by R. J. White:- The Ultimate
(3) Gerard, G.:- The Crippling Strength of Compressive strength of Thin Sheet Metal
Compression Elements. Jour. Aero. Sci., Panels. Appendix A. C.I.T. 1935.
Vol. 25, Jan. 1958.
(4) Coan, J. M.:- Large Deflection Theory for

Plates with Small Initial Curvature Loaded
in Compression. Jr. Applied Mech., Vol.
18, June 1951.
CHAPTER C8
BUCKLING STRENGTH OF MONOCOQUE CYLINDERS
This chapter presents information on the buckling strength of Circular cylinders
under compressive, bending and torsional loads acting separately and in com-
bination, without and with internal pressure. Som3 information on the bU.;kling
strength of conical cylinders is presented.
ca. 1 Introduction. for the strength design of monocoque structures

indicated that the design curves given in
Before the advent of the high speed air- (Ref. 1) were widely used. It thus seems
craft and particularly the missile and space appropriate to give some design m~terial from
vehicle, the use of the unstiffened cylinder or (Ref. 1).
monocoque type of structure was qUite limited.
However, the arrival of the space age has Cylinders are usually given four classi-
caused the thin walled cylinder to become im- fications relative to length:-
portant in the design of missiles and space
structures. (1) Short Cylinders. Short cylinders tend to
behave as flat plate columns (cylinders
The classical small deflection theory with infinite radius). They develop
which has proved adequate for determining the buckles in a sinusiodal wave form similar
buckling strength of flat sheet structures as to flat plate buckling. The end fiXity
covered in Chapter A18, has not proved adequate has considerable influence.
for determining the buckling strength of thin
walled cylinders or curved sheet panels since (2) Intermediate or Transition Length of
the theory gives results much too high when Cylinders. Buckling involves a mixed
compared with the experimental or test results. pattern combining sinusiodal and diamond
shapes. The effect of end fixity and
A more recent large deflection theory has cylinder length are only of nominal
shown closer agreement with experimental import::mce.
reSUlts, however, the theory reqUires a prior
knowledge of cylinder imperfections. As a (3) Long Cylinders. Such cylinders buckle in
result of the theoretical limi~ations to date, a diamond shaped pattern, and the length
the strength design of such structures is based and ,end conditions are not of mu;h im-
primarily on best fit curves for experimental portanee.
or test reSUlts, using theoretical parameters
and curves when they appear applicable. (4) Very Lonrr, Cylinders. Such cylinders buckle
by over-all colu~n instability or act as a
A missile structure in handling operations Euler type column.
and flight maneuvers is SUbjected to tensile,
compressive, bending and torsional load Fig. C8.1 shows a photograph of the buckle
systems. The cylinders may be pressurized or failure of an intermediate length cylinder under
unpressurized. The structural design of such axial compression.
structures thus requires a knowledge of the
buckling strength under the various load The name of Donnell is prominent in the
systems, acting separately and in combination. development of a theory for the buckling strength
This chapter will be confined to giving test of cylinders. From (Ref. 2), Donnell's eight-
data and design curves for the buckling order differential equation can be used to
strength of cylinders under the various types provide a small-deflection theoretical solution
of stress systems. to the behavior of cylinders in the short to
long length range. A solution of Donnell's
ca.2 Buckling of Monocoque Circular Cylinders equation by Batdorf (Ref. 3) gives the critical
Under Axial Compression. stress in terms of the buckling coefficient Kc ,
which for sim)ly supported cylinder ends is
The term monocoque cylinder means a thin defined by the equation:-
walled cylinder without longitudinal skin
stiffeners or transverse intermediate frames Kc = [(m2+JS2)2/m2J+[12Z2m2/TI4(m2+J?2)J
attached to cylinder skin.
- - - - - - CB.I
where,
Investigation by the author of the pro-
cedure used by a llilffiber of aerospace companies m = number of half waves in longitUdinal
I direction
ca.!
ca.2 BUCKLING STRENGTH OF MONOCOQUE CYLINDERS
Fig. ca. I Type of Failure Axial Compression and No Fig. ca. la Type of Failure Under Axial Load With Internal
Internal Pressure. (Ref. 1.) Pressure. (Ref. 1.)
Fig. ca. lc Type of Buckling Failure Under Pure Bending Fig. ca. ld Type of Buckling F .iilure Under Pure Bending
Load. No Internal Pressure. (Ref. 9.) With High Internal Pressure. (Ref. 9.).
ca.3
fJ = L/A Observation shows that the theoretical results

are far above the test values. Fig. C8.7
A = half wave length of buckles in (from Ref. 1) is a set of design curves of Kc
circumferential direction. versus Z for various r/t values and for 90
percent probability.
The compressive buckling stress is given
by, ca. 3 Additional Convenient Design Charts for Determining
Compressive Buckling Stress.
- - - C8.2
Figs. C8.8a and C8.8b are more convenient
design charts as the buckling stress Fc can
where, t = wall thickness, L is the cylinder cr
length and V is Poisson's ratio. The term ~ is be read from the chart, thus avoiding the
the plasticity correction factor and equals 1.0 calculations involved in using curves in Fig.
for elastic buckling. C8.7. Fig. C8.8a is for 99 percent probability
and 95 percent confidence level and Fig. C8.8b
Minimization of Eq. C8.1 with respect to for 90 percent probability and 95 percent
the parameter (m 2 + jJ2) 2/m2, gives the critical confidence level. The curves are based on tests
buckling coefficient for long cylinders in the of steel and aluminum alloy cylinders only. The
following form:- accuracy of these curves when used for other
materials has not been sUbstantiated by tests.
Kc = (4-!3/n 2)Z = 0.702Z The 99 percent probability curves are recom-
- - - C8.3 mended as design allowables for structures whose
where Z = (L2/rt)~ failure would be highly critical. The 90 per-
- C8.4 cent curves can be used for less critical
structures.
Substitution of Eq. C8.3 into Eq. C8.2
reduces to the well known classical equation
ca.4 Plasticity Correction.
Fcr = CE(t/r) - - - - - - - - C8.5 The plasticity correction for cylinders in
the long range (L 2 /rt ~ 100) is given by the
where, C = 1/J3 (1- .3 2) = 0.605 for V = .30
curves in Fig. C8.9 taken from (Ref. 4). In
general most practical cylinders in aerospace
The buckling coefficient for simply
supported end conditions in the transition structures will fall in the long cylinder range.
range can be determined by substituting the
limiting values ofj3 = 0 and m = 1 in Eq. C8.1. to
±i
A similar solution for cylinders with fixed
i2
edges deviates from the solution for simply 2
supported cylinders only in the flat plate and ,

LO '".$
transition ranges. -!10
tJ
i-j-
'ft)
.8
Figs. C8.2 to C8.5 inclusive taken from F
Ccr
Reference 1, show the theoretical curve which ~ .8
shows the buckling coefficient as a function
of the geometrical parameter Z. Theoretically,
the short cylinder range would occur at Z = O.
The values of Kc of 1.12 and 4.12 for Z = 1
correspond quite closely to the buckling
coefficients of simply supported and fixed 8 10 IA /8 i8 If)
ended plate columns. The long cylinder behavior Kc n
2
E t 2
is represented by a 45 degree sloping straight 12(1 Ve 2 )F,.,(L)

line portion of the curve. The curve con- Fig. CB.9 (Ref. 4) Nondimensional Buckling Chart for Axially Compressed Long Circular
necting the short and long cylinder ranges is Cylinders. T/ = (Es/E) [(Et/E)(1 - lie 2)/(1 - V")] '/ 2. Applicable
referred to as the transition range. When L 2jrt ::>' 100.
Figs. C8.2 to C8.5 show the plot of ca. 5 Buckling of Monocoque Circular Cylinders Under Axial
extensive test data and a 90 per cent proba- Load and Internal Pressure.
bility curve derived by the author of (Ref. 1)
by a statistical approach. Fig. C8.6 from Experiments conducted many years ago
(Ref. 1) shows a plot of much test data on a definitely showed that the compressive buckling
logarithmic chart of r/t versus Kc for Z = strength of monocoque cylinders was increased
10,000. A best fit curve and 90 and 99 per- if internal pressure was added to the closed
cent probability derived curves are also seen cylinder. Since weight saving is very important
as well as the theoretical curve for C = .605. in missile design, the use of pressurized
ca.4 BUCKLING STRENGTH OF MONaCO UE CYLINDERS
o BALLERSTEDT 6 WAGNER tDONNELL .1.O, CRATE 81 SCHWARTZ o BALLERSTmT a .cJNEIt • DONNELL .1.O, CRATE 81 SCHWARTZ
D FUNG a SEQtLER "LUNDQUIST _BRIDGET CI FUNG .. SECHLER eLlN)QUIST
- BRIDGET
• BUOIY (NAA) oNAA 4 Rll8ERTSON • SUCHY (NAA) .NAA _ROBERTSON
.. a..ARK a HOLT • KANEMITSU a NOJtMA o WILSON 81 NEWMARK III CLARK a HOLT • KANEMrrsu .. NOJIMA • WILSON a NEWMARK
~
6
4
""e,
'I' Ke 12[1-,u2) L
'JT2 E [-IT I ,,"cr ,,-2 E
TJ • Ke 12[1';#2] L
['r
~r
2 z. ~v'1-,u2 Z· ~1IT-,li2 V
KJ38 6
rl
/"
KJ3
8
6
rl
4 4
2
1/ V */
THEORETICAL aJRVE
If7.
THEORETICAL CURVE
/" Ke 102
8 8
6
4 .
/
90% PROBABILITY
DESIGN CURVE (r/l.3OOl
--SIMPLY SUPPORTED EDGES
~~
6
4
V- /
90% PROBABILITY
DESIGN CURVE (,/l.r50)
2 --SIMPLY SUPPORTED EDGES
---CLAMPED EDGES V ---Cl.AMPED EDGES
10 10
8 8
6
4
1/
- 6
4 -
2
1.0
./V
1.0
.~....,v
4 68 4 68 4 •• 4 68 468
468KJ2
1.0 10 102 10 3 1.0 10 105
Z
Z
Fig. ca.2 r/t = 100 to 500. Fig. ca.3 r/t = 500 to 1, 000.
o BALLERSTEOT a WAGNER ODONNELL • Lo, CRIlrE 81 SCHWARTZ
_BRIDGET o FUNG 81 SECHLER "LUNDQUIST o BALLERSTEDT a WAGNER • DONNELL • Lo, CRATE 81 sc.-.rrz
.1lUCHY \NAAI _NAA aROBERT'SOfli _IlRIOGET D FUNG a SEOtLER
• LUMlQUIST
.. CLARK a HOLT • KANEMITSU a NOJIMA o WILSON 81 NEWMARK • BIJCHY (NAA) oNAA - ROllERTSON
III CLARK a HOLT • KANEMITSU a NO.JIMA • WILSON a NEWMARK
10~
6 ""cr
4
'JT2 E
TJ .Ke 12[ljl.L2j L
[I r I~
4
6 ""cr 'JT2 E
11'Kc 12[1",2]
[IL r -
~
Z. "VI-,u2
y ~
z' rt v'i7
• I
10 3 /
8
6
4
1\ I IIIII
THEORETICAL CURVE
~ THEORETICAL CURVE ---: ~
10~
90% PROBABILITY
DESIGN CURVE 6 . 90% lBAB LI Y
'.
DESIGN CURVE
(r/l.15OO) • (r/l.3OOO)
V
• --SIMPLY SUPPORTED EDGES V- SIMPLY SUPPORTED EDGES
10
1 : / ---CLAMPED EDGES
10
/ ---CLAMPED EDGES
8 8
5
4 - 6
4
~~ V 10
./" .-/
1.0 4 68 4 68

Bruhn + Supplement PDF

Uploaded by

Copyright:

Available Formats

Bruhn + Supplement PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Bruhn + Supplement PDF

Uploaded by

Copyright:

Available Formats

ANALYSIS AND DESIGN

TRI-STATE OFFSET COMPANY

Professor (Emeritus) of Aeronautics and Astronautics

Assisted by the following persons and

DR. R. J. H. BOLLARD, (Chapters A24, A25, A26)

LLOYD E. HACKMAN, (Chapters C12)

DR. GEORGE LIANIS, (Chapter A18)

WILLIAM F. McCOMBS, (Chapter CIO-Part 2, Chapter Cll-

DR. A. F. SCHMITT, (Chapters A7, A8, A16, A17, A22)

CLARENCE R. SMITH, (Chapter C13)

DR.JOSEPHA. WOLF,JR., (ChapterA23)

All rights reserved. including the right of

Portions previously published

Published and Distributed by

A1 The Work of the Aerospace Structures Engineer.

STATICALLY DETERMINATE STRUCTURES

(Loads, Reactions, Stresses, Shears, Bending Moments, Deflections)

THEORY AND METHODS FOR SOLVING STATICALLY

BEAM BENDING AND SHEAR STRESSES.

A13 Bending Stresses.

INTRODUCTION TO PRACTICAL AIRCRAFT STRESS ANALYSIS

A19 Introduction to Wing Stress Analysis by Modified Beam Theory.

THEORY OF ELASTICITY AND THERMOELASTICITY

A24 The 3-Dimensional Equations of Thermoelasticity.

FLIGHT VEHICLE MATERIALS AND THEIR PROPERTIES

B1 Basic Principles and Definitions.

STRENGTH OF STRUCTURAL ELEMENTS AND COMPOSITE STRUCTURES

C1 Combined Stresses. Theory of Yield and Ultimate Failure.

CONNECTIONS AND DESIGN DETAILS

Dl Fittings and Connections. Bolted and Riveted.

Appendix A Elementary Arithmetical Rules of Matrices.

Al THE ~ORK OF THE AEROSPACE RNGllriER

A2 RQUILlBRIUlI OF FORCE SYSTEliS. TRUSS STRUCTIJRRS

A3 PROP~RTIES OF SRCTIONS - CE.NTROIDS. MOH7.V!S 02 INERTIA ETC

GF.NE.RAL LOADS ON AlrtCRAIT

A6 TORSION - STRESSES AND DEFLECTIONS

A9 B&NDING MOKENTS IN FR.AXES AND RINGS (Elastic Centre)

Ala srATICALLY INDKTIlR!illlAT"E STRucruR£S - COLUMN A!lALOGY METBOD

All CONTINUOUS STRUCTURBS - HOKENT DISTRIBUTION MKTEOD

A'3 BENDING STRESSP.S

A17 BENDING OF PLATES

A19. 1 Typical ~ing Structural A.rangement

A21 ANALYSIS OF SPECIAL \lING PROBLEMS

A25 THE 2-DIXENSIONAL EQUATIONS OF ELASTICITY &~ THERHOKLASTICITY

A26 SELBCTED PROBLEMS IN ELASTICITY &~ TffER.~OELASTICITY

Cl cmrnINW STRESSES. THEORY OF nSLl) AND ULTIMATE FAILURE

Cl.i Uniform Stress Condition

c2 STRE.."tGT8 OF COLUXNS '11TH STABLE CROSS-SECTIONS

C2.1 He~hods of Column Failure, Co:~~n ~~uat~0;lS

CJ.l Illtroduc t ion

C4. 19 Torsional Hodulus of Rupture

C4.22 Combilled Bending and Compression

ClO OESrGN OF KETAL BRAMS. W'RB SBRAR RESISTANT (NON~BUClU.ING)

C1G.1 Introduction to Flat Sheet 'Jeb 'iith Venical Stiffeners

ell DIAGONAL S&XI-i&~SION FIELD DESIGN

e12. 1 Introdllct ion

D3 SQKE IMPORTANT DETAl LS IN STRUCruRAL DESIGN

ENGINEERING COLUMN ANALYSIS

A A Survey of Various Column Problems.

C2.1 He~hods of Column Failure, Co:n uat~0;lS