Ada 101657

A101657
AFWAL.TR-T1.3041
VOLUME 1
EFFECT OF VARIANCES AND

MANUFACTURING TOLERANCES ON THE
DESIGN STRENGTH AND LIFE OF
MECHANICALLY FASTENED
COMPOSITE JOINTS
VOLUME I - METHODOLOGY DEVELOPMENT AND
DATA EVALUATION
S.P. Garbo and J.M. O4onowski J
McDonnell Aircraft Cormnvey

McDonnell Douglas Corporation
P.O. Box 516
St. Louis, Missouri 63166 .
April 1981
Final Rnport for Period 15 February 1978- 15 April 1981
-pwnftr URbIftISS idt'IbUwid UUUI*
-__ T DYNAMICS LABGS : ?ORY

'In FORCE WRIGHT AERONAUTICAL LABSRA•C•IES
AIR FORCE SYSTEkMS COMMAND
8 72103
WRIGHT-PATTERSON AIR FORCE BASE, OHIG 4
81 7 21 009
-D' NOTICE
When Government drawings, specifications, or other data are

used for any purpose other than in connection with a definitely
related Government procuremant operation, the United States
Government thereby incurs no responsibility nor azrny obligation
whatsoever; anid the fact that the government may have formulated,
furnished, or in any way supplied the said drawing, specifica-
tions, or other data, is not to be regarded by implication or
otherwise as in any manner licensing the holder or any other
person or corporation, or conveying any rights or permission to
manufacture, use, or sell any patented invention that may in any
way be related theereto.
Thie report has been reviewed by the Office of Public

Affairs (ASD/PA) and is releasable to the National Technical
Information Service (NTIS). At NTIS, it will be available to the
general public, including foreign nations.
Thiis technical report has been reviewed and is approved for
publication.
ROBERT L. GALLO, Capt, USAF DAVEY L/ SAITH, Chief

Project Engineer Structural Integrity Branch
Structures & Dynamics Division
FOR THE COMMANDER
RALPH L. KUSTER, JR., Col, USAF

Chief, Structures and Dynamics Division
If your address "is changed, if you wish to be removed front

our mailing list, or if the addressee is no longer employed by
your organization please notify AFWAL/FIBEC, Wright-Patterson
AFB, OH 45433 to help urs maintain a cu:rent mailing list. Copies
of this report should riot be returned unless return is required
by security considerations, contractual obligations, or notice on
a specific document.
AIR FORCE/5678011 July 1981 - 500
SECURITY q r6,CATION Of THIS PAGE (WPhen Da. Entered)"
RPORT DOCU4ENTATION PAGE READ INSTRUCTIONS
BEFORE COMPLETING FORM
"GoV ACCESSION NO. IENT'S CATALOG NUMBER
AFAL R-8--3,4 1' V I h V±

tFFECT OF YARIANCES AND MANUFACTURIW OLEIRANCE [Final Je_. 15
THPSI6N §TRENGTHq At IFE OF rE&ANICALLY 115 Apr t@S8lflITT-
.,ASTENED OMPOýITE P, :NTS. REPORT NUMBE
AU CONTRACT O
Sý0IF-0 GRA~NT UMNER(a)
J. M. /Ogonowski -- 7
ibe'oSwPt) G ORGANIZATION NAME AND ADDRESS 10 PROGRAM ELEMENT.'PROJECT, TASK
McDonnell Aircraft Company AREA A WORK UNIT NUMBERS
P.O. Box 516 P.E. 62201F
St. Louis, Missouri 63166 W.U. 24010110
I.CONTROLLING OFFICE NAME: AND ADDRESS
Flight Dynamics Laboratory (AFWAL/FIBEC) j [ Ap
Air Force Wright Aej:onautical Lab. (AFSC) Ur PAGES
Wright-Patterson AFB, Ohio 45433 144
14 MONITORING AGENCY NAME & Ar)ORESS(It different fron. Controllin4 Olfice) 15. SECURITY CLASS. (of it . report)
7, 1
Jnclassif ied
IS. DE C ASSI FICATION/ DOWNGRADING
SCHEDULE
16 DISTRBUTION STATEMENT (of th
Approved for public release; distribution unlimited
17, DISTRIBUTION ST 4ENT (of abstract eentered in Block 20, If different froe. Report) 4
IS SUPPLEMENTARY IES " '
C'
19. KE V WORDS Zonllnuie or reverse side if neciiesli znd Identify iiy black numbwr)
Bolted Joints Orthotropic Ltad Distributions

Composite Stress Concentrations Sl:ress Analysis
Graphite-epoxy Methodology Failure CriterLa
Fatigue Life
0 ABSTRACT (Conlinu on reverse sde If necesesry and Identify by block number)
The subject of this program was structural evaluation of mechanically

fastened composite Joints. Program objectives were threefold: (1) develop-
ment and verification by test of improved static strength methodology,
(2) experimental evaluation of the effects of manufacturing anomalies on
joint itatic strength, and (3) experimental evaluation of joint fatigue life.-
DI I JAN 3o 1473
SECURITt' CLASSIFICATION OF THIS PAGE (When Data Entered)
Li"-T 7 . ! Cm11
- ... ,
SECURITY CLASSIFICATION OF THIS PAGE(When Data Enfl.;d_
,'Program activities to accomplish these objectives were organized under

five tasks. Under Task 1 - Literature Survey, a surivey wae perform4 to
determine the state-of--the-art in design and analysisi of bolted coa.6 ,site
joints. Expezimental evaluations of joint static strength were performed
under Tatsks 2 and 3. In Task 2 - Evaluation of Joint Design Variables,
strength data were obtained through an experimental, program to evaluate the
effects of twelve joint design variables. Tn Task 3 - Evaluation of Manu-
facturing and Service Anomaelies, effects of seven anomailies on joint strength
were evaluated experimentally and compared with Task 2 strength data. Belted
composite joint durability was evaluated under Task 4 - Evaluation of Crii:tcal
Joint Design Variables on Fatigue life. Seven critical design variables o0.
manufacturing anomalies were identified based on Task 2 and 3 strength data.
Under Task 5 - Final Analyses eind Correlation, required data reduction,
methodology development and correlation, and necessary documentation were
performed . -
This report documents all program activitics performed under Tasks 2, 3,

4 and 5. Activities performed under Task 1 - L-iterature Survey, were pre-
viously reported in AFFDL-TR-78-179. Static. strength methodology and evalua-
tions of joint static and fatigue test data are reported. Analytic, studies
complfment methodology developmeizt and illustrate: the need fcr detailed
stress analysis, the utility of the developed "Bolted Joint Stress Field
Model" (BJSFM) procedure, and define i-odel limitatLins. For static strength
data, correlations with analytic predicbtions are included. Data trends in
all cases are discussed relative joint strz',agth and failure mode. For
joint fatigue studies, data trends are discuesed relative to life, hole
elongation, and failure mode behavior.
Tbis final report is organized in the •oll]owin$ three volumes:
Vlolume 1 - M1ethodology Development ard asta Eval.uation

Vilume 2 - lest Data, Equlpment and Procedures
Volume 3 - Belted Joint Stress Field Model (BJSFM) Camputr
FrL'gram User's Manual
SECURITY LASSIFICATION OF THIS PArkir-'When Date Entered)

FOREWORD
The work reported herein was performed by the McDonnell

Aircraft Company (MCAIR) of the McDonnell Douglas Corporation
(MDC), St. Louis, Missouri, under Air Force Contract F33615-77-
C-3140, for the P'light Dynamics Laboratory, Wright-Patterson Air
Force Base, Ohio.. This effort was conducted under Project No.
2401 "Structural Mechanics", Task 240101 " Structural Integrity
for Military Aeroxzpace Vehicles", Work Unit 24010110 "Effect of
Variances and Manufacturing Toleranceii on the Design Strength and
Life of Meclianically Fastened Composite Joints". The Air Force
Project Engineer at contract go-ahead was Mr. Roger J.
Aschenbrenner (A•'WAL/FIBEC); in Dec,-.mber 1979, Capt. Robert L.
Gallo (AFWAL/FIBEC) assumed this assignment. The work described
was conducted during the period 15 February 1978 through 15 April
1981.
Program Manager was Mr. Raman A. Garrett, Branch Chief

Terhnology, MCAIR Structural Research Department.. Principal
Investigator was Mr. Samual P. Garbo, MCAIR Structural Research
Department.
Dr ýC T'.I_
Di st ribution /
AvailmbilitY Codes .
nn," or
iii9
IL
I
I••.••
"..
"-- .•
."•mm ••--4`*•• •.•.. '..L4"
-••`.-- • ••••.•
TABLE OF COI4rENTS
Section P_
I INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . 1
I1 SUMMARY AND COMCLUSIONS . . . . . . . . . . . . . . . . 2

IXcI METHODOLOGY DEVELOPMENT KID ANALYTIC STUDIES . . . . . 5
1. BACKGROUND ...... . . . . . . . . . . . . 5
2. METHODOLOGY DEVELCPMENT ........... ... 5
3. PREDICTIONS USING BJSFM APPROACH. . . . . . . ... 15
a. Effects of Anisotropy on Stress

SDistributions .... ......... 15
b. Hole Size Effects ... .......... . 16
c. Finite Width Effects .............. 17
d. Characteristic Dimension Sensitivity Study . . 19
e. Biaxial Loading Effects .............. 20
f. Pure Bearing Strength Study .... .......... . 24
g. Effects of Enviromment . . .......... . 25
IV TEST DATA EVALUATION ..... .................. 26

I. METHODOLOGY VERIFICATION: EVALUATION OF
JOINT DESIGN VARIABLES - TASK 2 ......... .26
a. Task 2 Test Plan . ....... .......... 26

b. Correlation of 3JSFM Predictions With
Experimental Results and Evaluations ...... .. 35
(i) Strength of Laminates With Unloaded

and Unfilled Holes . . . ........... 35
(2) Strength of Laminates With Loaded
Fastener Holes ......... ............. 44
(a) Layup Variation .. .. ........ . 50

(b) Load Orientation .... ....... . . . . 50
(c) Edgt Distance ........... ............ 55
ký
;1) Widthi . . . . . . . . 57
(e) Hole Size ....... .............. 57
(f) Load Interaction .... ........... .. 64
(g) Fastener Pattern . ....... ............ 73
(h) Torque-up ......... . ............ 76
(i) Fastener Countersink and
Laminate Thickness ... .......... .. 77
(j) Stacking Sequence................ 86
(k) Single-Shear Loading ........... .. 86
V
TABLE OF CONTENTS (Concluded)
Section age
2. EV\LIUATION OF MANUFACTURING AN.) SERVICE

ANOMALIES - TASK 3 .......................... 93
a. Task 3 Test Plan .......... ............... 93

b. Experimental Results and Evaluation ..... ...... 94
(1) Out of Round Holes .. .......... 94

(2) Broken Fibers on Exit Side of Hole 98
(3) Porosity . . . . . . . . . . . . . . . . 100
(4) Improper Fastener Seating Depth .. 102
(5) Tilted Countersinks. . . ............ 104
(6) Interference Fit .................... . 106
(7) Fastener Removal and Reinstallation .... .. 109
3. EVALUATION OF CRITICAL JOINT DESIGN VARIABLES

ON FATIGUE LIFE - TASK 4 ...... ........... . . . . IC
a. Task 4 Test Plan . . . . ............ . . 110

b. Experimental Results and Evaluation . . . . . . 115
(1) Layup Variation ...... .............. .. 117

(2) Stacking Sequence . . . . ........... 125
(3) Torque-Up.............. . . . . . . . . . 1.30
(4) Joint Geometry . . . . . . . ... ... 135
(5) Interference Fit ............. 138
(6) Single-Shear Loading .............. . .. 138
(7) Porosity .... ............ .142
V RECOMM>,NDATIONS ....................... 143
VI REFERENCES .............. ....................... 144
vi
-i
LIST OF ILLUSTRATrIoNS
Figure Page
1 Superposition of Linear-Elastic Stress Solutions . 6
2 Uniaxially Loaded Infinite Plate ........ .......... 9
3 Assumed Cosine Bolt-Load Distribution ....... ...... 10
4 Characteristic Dimension Failure Hypothesis . . .. 12
5 Failure Criteria Comparison ....... ........... .. 13
6 Superposition of Solutions to Account

for Finite Width. . ............ .* . . . . 14
7 CircuMferential Stress Solutions at

Unloaded Holes ....... ................. . . . . 15
8 Circ mferential Stress Solutions at

Loaded Holes ...................... 6
9 Effect of Hole Size on Stress Distributions .... 17
10 Correlation of Loaded Hole Analyses ..... ........ 18
11 Effects of Width on Stress Ccncentrations ....... 18
12 Correlation of Loaded Hole Analysis for Small

Edge Distance Cases ....... ..................... 19
13 Effect of Characteristic Dimension on

Predicted Strength - 30/60/10 layup .. ........ .. 20
14 Effect of Biaxial Loading on Predicted Strength 21
15 Effect of Biaxial Loading on Failure Location . . . 22
16 effect of Biaxial Loading on Ply Strain

Concentrations ............ ................... 23
17 Effect of Bearing Load Direction on
Predicted Strength ..................... 24
18 Effect of Moisture and Temperature on
Laminate Strength ........... ................. 25
19 Task 2 - Joint Design Variables - Test Matrix . . 27
20 MCAIR Evaluation of Unloaded Hole Specimens -

Test Matrix ............... .................... 29
vii
LIST OF ILLUSTRATIONS (Continued)
Frigure P _
21 MCAIR Evaluation of Loaded Hole Specimens
Test Matrix ............. .................... 30
22 Unloaded Hole Specimen Geometry ... .......... 30
23 Loaded Hole Specimen Configuration ... ......... .. 31
24 Specimen Loading Configurations ... .......... 32
25 Effect of Laminate Stiffness on Specimen

Bolt-Load Distributions ....... .............. 33
26 Task 2 - Layup Number and Stacking Sequence ... 34
27 AS/3501-6 Lamina Mechanical Properties ....... .. 35
2R Effect of Layup Variation on Unloaded

Hole Tensile Strength ......... .............. 36
29 Effect of Hole Size on Unloaded Hole

Tensile Strength .......... .................. 37
30 Effect of Variation of Rc on Predicted
Tensile Strength ............ .................. 38

31 Predicted Failure Orientations Were Verified . 39
32 Effect Off-Axis Loading on Failure Stress ...... 40
33 Efrect of Variation of R, on Predicted

Compressive Strength ........ ................ 41
34 Effect of Hole Size on Unloaded Hole

Compressive Strength ........ ................ 42
35 Effect of Layup Variation on Unloaded

Hole Compressive Strength ....... ............. 43
36 Effect of Temperature on Unloaded Hole
Compression Strength ........ ................ 44
37 Two-Bolt-In-Tandem Spezimen Raseline

Configuration ............. ................... 45
38 Variations of Single 7nstener Specimen

Configurations . . ..................... 46
39 Variations of Two-Rolt-In-Tandem
Specimen ConfiguratiLns ......... .............. 47
viii
Figure Page
40 Variations of Load-Interaction Specimen

Configurations ............ ................... 48
41 Fastener Pattern Specimen Geometry ... ......... .. 49
42 Effect of Layup Variation on Joint

Tensile Strength ............ ................. .. 51
43 Effect of Layup Variation on Joint

Compreasive Strength ........ ................ 52
44 Effect of Off-Axis Loading on RTD Pure

Bearing Strength. ......... ................. 53
45 Effect of Temperature and Moisture on

Off-Axis Strength ........... ................. 54
46 Effect of Temperature and Moisture on

Cow.preusive Strength ........ ................ 55
47 Effect of Edge Distance on Joint

Tensile Strength ............ ................. .. 56
48 Effect of Fastener Spacing on Joint

49 Effect of Edge Distance on Pure

Bearing Strength .......... .................. 59
50 Effect of Specimen Width on RTD Joint Strength . . . 60
51 Effect of Temperature and Moisture

on Joint Strength ........... ................. 61
52 Joint Failures Changed at ETW Test Conditions. . . . 2
53 Effect of Hole Size on Joint Strength ....... 63
54 Load Interaction Specimen and Test Setup ...... .. 64
55 Correlation of Bearing-Bypass Strength of a

50/40/10 Layup ............ ................... 65
56 Correlation of Bearing-Bypass Strength of a

30/60/10 Layup ............ ................... 66
57 Effect of Bearing Loads on Specimen

Load-Deflection Behavior ...... .............. 67
ix
Figure Pag~e
58 Predicted Critical Plies for Net-Section
Failures .............. ...................... 68
59 Predicted Critical Plies for Bearing-Shearout

Failures .............. ...................... 69
E10 Tests Verify Predicted Failure

Initiation Points ........... ................. 70
61 Load Conditions for Off-Axis

".~oad-Interaction Tests ........ ............... .. 71
62 Effect of Off-Axis Bearing Loads on
Bypass Strength ........... ................. 72
63 Fastener-Pattern Specimen Loading Configurations . . 73
64 Effects of Environmnint on Fastener-Pattern

Joint Strength ............ ................... 74
65 Joint Failures for Fastener-Pattern Specimens . . . 75
66 Corrosion Resulting From Salt-Spray Exposure . . .. 76
67 Effect of Lavup and Torque-Up on RTD Pure

Bearing Strength ............ ... .................. 78
68 Failures of Pure Bearing Specimens ........... .. 79
69 Comparison of Loaded and Unloaded Hole

Specimen Strength ........... ................. 80
70 Effect of Torque-Up on RTD Pure Bearing

Strength 50/40/10 Layup ....... .............. 81
71 Effect of Toraue-Up on RTD Pure Bearing
Strength 70/26/10 Layurp ....... .............. 82
72 Effect of Torque-Up on RTD Pure Bearing

Strength 30/60/10 Layup ....... .............. 83
73 Comparison of Effects of Torque-Up on

Joint Stru-':'th ................ . ................... 84
74 Effects of Thickness and Countersink on

Joint Tensile Strength ........ ............... 85
75 Effect of Stacking Sequence on Joint

x
Figure Paqe
76 Correlation of Joint RTD Tensile Strength

With Grouped 0 Plies ...... ............... 88
77 Correlation of Joint ETW Tensile Strength

With Grouped 0° Plies ...............
78 Correlation of Joint RTD Compressive Strength
With Grouped 0° Plies .... ................ 90
79 Comparison of Single and Double-Shear
Joint Strengths ......... .................. 91
80 Correlation of Joint Strength With Edge Distance-to-

Width Ratios ........................ 92
81 Task 3 - Evaluation of Manufacturing Anomalies -

Test Matrix ............ .................... 93
S2 Specimen Moisture Conditioning Histories ...... .. 95
83 Summary of Task 3 Strength Reduction Percents . . . 96
84 Out-of-Round Holes - Specimen Details ....... 97
85 Effect of Out-of-Round Holes on Joint Strength .7 . C
86 C-Scans of Laminates With Delaminations

at Fastener Holes ........ ................. 98
87 Effect of Delaminations on Joint Strength ...... 99
88 Panel Fabrication Procedures Used to

Produce Panel Porosity ........... ............... 130
89 Examples of Panel Porosity .... ............. .. 101
90 Freeze-Thaw Exposure Profile .... ............ 102
91 Effect of Porosity Around Hole on Joini Strength . 103
92 Effect of Improper Fastener Seating on

Joint Strength ......... ................... 104
93 rilted Countersink - Specimen Configuration .... 105
94 Effect of Tilted Countersink on Joint Strength . . . 106
xi
Figure Pag
95 Effect of Fastener Interference Fit on

Joint Strength ........... .................. 107
96 P'hotomicrographic Examination of Laminates

With Interference Fit Holes .... ........... 108
97 Effect of Fastener Removal and Reinstallation

on Joint Strength ........ ................. 109
98 Task 4 - Evaluation of Critical Joint

Variables on Fatigue Life - Test Matrix ...... ...... li1i
99 Task 4 - Fatigue Specimen Configuration .......... 113
100 Distribution of Hours and Exceedances .. ...... 114
101 Measured Mix-Truncated Spectrum ... .......... 114
102 Task 4 Layup Numbers and Stacking Sequence ..... .. 115
103 Summary of Task 4 Specimen Static Strength ..... .. 116
104 RTD Baseline Joint Fatigue Life

50/40/10 Layup ......... ................... 117
30/60/10 Layup ......... ................... 118

19/76/5 Layup .......... ................... 118
107 Comparison of R = +0.1 Joint Fatijue Life Trends . . 119
108 Comparison of R = -1.0 Joint Fatigue Life Trends . 119
109 Hole Elongation Fatigue Life Trends

50/40/10 Layup ................. ................... 120
110 Hole Elongation Fatigue Life Trends

30/60/10 Layup ................. ................... 121
iii Hole Elongation Fatigue Life Trends

19/76/5 Layup .......... ................... 121
112 Representative Specimen Fatigue Failures ...... .. 122
113 Summary of Effects of Environment on

Joint Spectrum Fatigue Life ...................... 123
xii
LIST OF ILLUSTRATIONS (Concluded)
Figure Page
11, Effect of R - +0.1Layup

Baeie50/40/10 Loading..............
on Hole Elorgation - 124
Baseline 0//lay..........................2
115 Effect of R = +0.1 Loading on Joint Spring Rate -

Baseline 50/40/10 Layup . . . . . . . 125
116 Effect of Stacking Sequence on Joint Life -
50/40/10 Layup .*. . ......... ........... 126
117 Effect of Stacking Sequence on Joint Life -

19/76/5 Layup ........ . ............... 127
118 Stacking Sequence - Comparison With R = -1.0

Baseline Life Trends ...................... . . . . 128
119 Su.acking Sequence - Comparison With R = +0.1
Baseline Life Trends . . . . . .............. 128
120 Effect of Stacking Sequence - Hole Elongation
Levels - 50/40/10 Layup ........... . . . . 129
121 Effect of Stacking Sequence - Hole Elongation
Levels - 19/76/5 Layup ............. ........ 130
122 Summary of Effects of Stacking Sequence on

Joint Spectrum Life . . . . . ............... 131
123 Effect of Torque-Up on Joint Fatigue "Lie
50/40/10 Layup ............ ................... 132
124 Effect of Torque-Up on Joint Fatigue Life
19/76/5 Layup ............. ................... 133
125 Specimen Failures for Torque-Up Conditions ..... .. 134
126 Effect of Geometry on Joint Fatigue Life
50/40/10 Layap . . .................... 135
127 Effect of Geometry on Joint Fatigue Life

19/76/5 Layup ................. ................... i36
128 Changes in Failure Due to Specimen Geometry . . . . 137
129 Effect of 0.005 Inch Interference Fit on Joint

Fatigue Life - 50/40/10 Layup .... .......... 139
130 Effect of Single-Shear Loading on Joint
Fatigue Life - 50/40/10 Layup ..... ........... .. 140
131 Failures of Single-Shear Specimens ... ......... .. 141
132 Effect of Porosity on Joint Fatigue Life -
50/40/10 Layuip ............ ................... 142
xiii
SECTIONI
INTRODUCT ION
The subject of this program was structural evaluation of

mechanically fastened composite joints. Program objectives were
threefold: (1) development and verification by test of improved
static strength methodology, (2) experimental evaluation of the
effects of manufacturing anomalies on joint static strength, and
(3) experimental evaluation of joint fatigue life.
Program activities to accomplish th'ese objectives were
organized under five tasks. Under Task 1 - Literature Survey, a
survey was performed to determine the state-of-the-art in design
and analysis of bolted composite joints. Experimental evalua-
tions of joii~t static strength were performed under Tasks 2 and
3. In Task 2 - Evaluatior~ of Joint Design Variables, baseline
strengt~h data were obtained through an experimental program to
evaluate the effects of twelve joint design variables. In Task 3
- Evaluation of Manufacturing and Service Anomalies, effects of
seven anomalies on joint strength were experimentally evaluated
and compared with baseline Task 2 strength data. Bolted compos-
ite joint durability was evaluated under Task 4 - Evaluation of
Critical Joint Design Variables On Fatigue Life. Seven critical
design variables or manufacturing anomalies were ide'nLified based
on Task 2 and 3 strength data. Under Task 5 - Final Analyses and
Correlation, required data reduction, methodology development and
correlation, and necessary documentation were performed.
This report documents all program acti-''ities performed under

Tasks 2, 3, 4 and 5. Activities performed under Task 1 - Litera-
ture Survey, were previously reported in Reference 1. The main
body of this report (Sections III and IV) documents developed
static strength methodology and evaluations of joint static and
fatigue test data. Associated analytic studies have been
included in Section III to complement methodology development
discussions and to illustý-ate: the need for detailed stress
analysis, the utility of the developed "Bolted Joint Stress Field
Model" (BJSFM) procedure, and to define model limitations. For
static strength data reported in Section IV, correlations with
analytic predictions, where applicable, are included. DataJ
trends in all cases are discussed relative to joint strength and
failure mode and compared to baseline data. For joint fatigue
studies, data trends are discussed relative to life, hole elonga-
tion, and failure mode behavior.
This final report is crganized in three volumes: Volume 1-

Methodology Development and Data Evaluation, Volume 2 -- Test
Data, Equipment and Procedures, Volume 3 - Bolted Joint Stress
Field Model (BJsFM) computer Program User's Manual.
I
SECTION II
SUMMARY AND CONCLUSIONS
Analytic methods were developed which permit strength

analysis of bolted composite joints with minimal test data
requirements. Methodology was based on anisotropic theory of
elasticity; classical lamination plate theory, and a character-
istic dimension (R ) failure hypothesis. The princAple of
elastic superpositioAc was used to obtain laminate stress distribu-
tions due to combined loadings of bearing and bypass. Test data
requirements for general method use were minimized by extending
the characteristic dimension failure hypothesis to a ply-by-ply
analysis in conjunction with known material failure criteria.
Unidirectional (lamina) stiffness and strength data were used
with an empirical value of R to predict stress distributions,
critical plies, failure locatfon, and failure load of arbitrary
laminates under general loadings. The developed analysis
procedre is entitled the "Bolted Joint Stress Field Model"
The BJSFM methodology was originally developed for unloaded

hole analysis by McDonnell Aircraft Company (MCAIR) under an in-
house research and development program. The methodology wax
further verified and extended to the analysis of loaded holea
under this program.
Under Task 2 - Evaluation of Joint Design variables, effects

of twelve joint design variables on static strength were experi-
mentally evaluated. The BJSFM procedure was used to predict the
effects of five of the twelve design variables (layups, load
interaction, off-axis loading, hole size, and width). Only exper-
imental evaluation of the remaining variables was possible. The
range of test conditions in eachý design variable was identified
in the Task I - Literature Survey.
Initial verification of analysis was obtained by correlating

strength predictions with data b-i-- iti Lcfsws of specimens
with unfilled fastener hole,. Specimenn were tested to failure
in tension and compressio[z. Values for R of .02 inch for ten-
sile strength predictions and .025 inch f8r compressive strength
predictions were empirically determined for one laminate, each of
which were fabricated using the Hercules AS/3501-6 graphite-epoxy
system. With these values of R &nd unidirectional ply (lamina)
mechanical properties, BJSFM predictions were correlated with
unloaded hole strength data for an extensive range of layup varia-
tions and, in general, were accurate to within + 10%. Predicted
failure initjAtion points were visually verified: Correlation of
strength predictions with data indicated that only knowledge of
temperature-altered lamina properties is required to predict
effects of temperature on general laminate strength; R remained
unchanged. C
2
Joint load-deflection plots for specimens with either shear-
out or bearing failures indicated nonlinear mechanical behavior,
with ultimate failure occurring at loads considerably above ini-
tial nonlinearities. Strength predictions obtained using the
BTSFM procedure correlated with initial points cf joint nonlinear
deflection behavior. Ultimate strength predictions using the
linear-elartic BJSFM procedure became increasingly conservative
as this joint nonlinear load-deflection behavior occurred. How--
ever, predicted linear-elastic failures of critical plies were
due to exceeding lamina fiber or shear strength and implied the
type of localized joint failures and load redistribution typical
of the failure modis observed in tests.
Under Task 3 - Evaluation of Manufacturing and Service Anom-

alies, effects of seven anomalies on joint static strength were
experimentally evaluated and compared with baseline (no anomaly)'
strength data. The seven anomalies were aelected based on MCAIR
experience and an available industry-wide survey (Reference 2).
Results fell into two groups: (1) anomalies which resulted

in strength reductions of more than 13 percent (porosity around
hole, improper fastener seating depth, and tilted countersinks),
and (2) anomalies which resulted in strength reductions of less
than 13 percent (out-of-round holes, broken fibers on exit side
of hole, interference fit tolerances, removal and reinstalla-
tion). For anomalies of the first group, current industry
inspection and acceptance criteria would have detected all three
and resulted in part rejection or required repair.
Unuer Task 4 - Evaluation of Critical Joint Design Variables

on Fatigue Life, the influence of seven design variables and ano-
malies on joint durability were evaluated. Tests were performed
to provide data on joint fatigue life performance, hole elonga-
tion, and failure mode behavior. Single-fastener pure bearing
specimens were cycled, under tension-tension (R = +0.1) and ten-
sion-compression (R = -1.0) constant amplitude fatigue loading,
and under spectrum fatigue loadings.
During fatigue testi'ng, load-deflection data were obtained

at specified increments of accumulated hole elongation. Several
tests were performed at environmental conditions previously
defined in Task 2. Static and residual strength tests were
performed on selected specimens at each test condition.
Joint fatigue life was defined to occur at specimen failure

or when hole elongations of .02 inch were measured. Based on
this definition, little difference was noted in relative fatigue
life between 50/40/10, 30/6G/10 and 19/76/5 layups (percent 00,
+450, 90° plies respectively) under tension-tension (R = +0.1)
cycling. Under fully reversed load cycling, (R = -1.0), hole
elongations of .02 inch occurred more rapidly in the matrix-
dominant 19/76/5 layup, followed by the 30/60/10 and 50/40/10
layups. Under spectrum fatigue tests of a'1 three layups, no
S i. II I
hole elongation occurred after 16,COO spectrum hours at test
limit loads (TLL) of 891 of static strength values. This agreed
with constant amplitude fatigue test zesults; 16,000 hours of
spectrum loading at tbese levels did not produce enough cycles of
high loads to produce hole elongation.
Effects of faistener torque--up on Joint fatigue life were
pronounced. Joints with torque-up Paltien exhibited increased
strength and life for all layups testad, inder both R - +3.1 and
R - -1.0 cycling. While failure modoas wore the same, aireas of
damage were more pronounced for specimern with torque-up and
failure occurred more abruptly (rates of ',ole elongation, once
initiated, were faster).
Effects of joint eccentricity and geometry also infLuenced
joint fatigue life characteristics. hingle.-shear specimens
exhibited lower fatigue life as compared to the double-shear
baseline configuration. Further reductions occurred wit-h
flush-head fasten-ars relative to pzotrtuding head fastenerii;
attributed to jincreased fastener f.tex;.bility and nspecimnn
bending. For tTi_ 19/76/5 layup, chianrtges in specieren width caused
changes in failuxe modes and marked changeii in both strength and
life. However, for the 50/40/10 layup, variationa in width and
edge distance did not alter failure modes, and strergth and life
were virtually unchanged.
Effects of remaining v_4riables on joint fatigue life were
minor, with failure modes and durability essentially .he satme as
for baseline configurations.
Ii
SECTION III
METHODOLOGY DEVELOPMENT AND ANALYTIC STUDIES
1. 0ACKGROUND - As described in the Literature Survey, analysis

of t-ei -omposite joints in aircraft structural components
throughout the industry proceeds from overall structural and
bolt-load distribution analyses, to assessment of stress distri-
butions and strength predictions at individual fastener holes
through utilization of joint failure analysis.
Methodi currently used to deternine detailed stress distribu-

tions in the immediate vicinity of the fastener hole include both
theoretical and empirical approaches. Theoretical approaches
include analytic, finite element, and strength of materials
approximation methods. Analytic methods, preferred because of
their potential generality, economy, and exactness, are princi-
pally formulated from two-dimensional anisotropic elasticity
theory. Empirical approaches lack generality at high cost.
Joint failure analyses, in current use, include: (1) empiri-

cal approaches, (2) elastic and inelastic failure analyses,
(3) phenomenological failure analysis, and (4) fracture mechanics
models. Physical variables considered for accurate solutions
were generally agreed upon throughout the industry. However, the
degree to which variables were accounted for was different in
particular methods. No single methodology accounted for all of
the important variables (e.g. orthotropy, finite geometry, non-
linear or inelastic material behavior).
In each joint failure analysis approach, after detailed

stress distributions are determined, strength is assessed by
using some material failure criteria; however, no single material
failure criterion is uniformly endorsed. Studies of utilization
of various material failure criteria for joint failure analysis
are very limited.
Detailed stress analysis performed at individual fastener

holes and associated application of failure criteria represents
the primary area of analytic development in this program. The
methodology developed requires only unidirectional material
properties and minimal laminate test data to calculate laminate
strength with an arbitrary in-plane loading. Because the method
uses closed-form solutions, parametric studies are easily and
inexpensively performed.
2. METHODOLOGY DEVELOPMENT - The goal of the methodology develop-

ment. was to provide a technique for predicting the strength of an
anisotropic plate with a stress concentration. A closed-form
analytic approach was developed to predict stress distributions
and perform failure analysis of an anisotropic plate with a
loaded or unloaded fastener hole. This section describes the
5
analytical development of the Bolted Joint Stress Field Model
(BJSFM). An associated computer program is described in Volume
3.
The method of analysis is based on (1) anisotropic theory of

elasticity, (2) lamination plate theory and (3) a failure
hypothesia. The principle of elastic e,uperposition is used t:
obtain laminate stress distributions due to combined loadings of
bearing and bypass (Figure 1). The leveloped analysis can be
used with various material failure criteria and the failure
hypothesis to predict laminate load carrying capability.
T - -- --
Problemr•InttmtUlZ)e Ol addHoeOnly
0 01
Flgut. 1. Suputpcehmdtn of Llnear-Elastkc Strses Solutions
2S6+ 21 $66+
The elastic solution for the stress field in a homogeneous,

anisotropic infinite plate with a stress concentration was solved
using two-dimensional anisotropic theory of elasticity (Reference
3). Equilibrium and compatibility requirements a~e satisfied by
a stress function, F, which satisfies the generalized biharmonic
equation for anisotropic materials,
04W&
$22 a4F 263 __4F Ed4 1X __42

Iy M2
IIewEa Fgr 1.L': SI.ipsto of i Stres Solutions..
_4_ a4 F
PaxaY' U
~l6
I
wher,- Sjk are laminate compliance coefficients. The general
expresafion for the function F depends upon the roots of the
associated characteristic equation. Solving the characteristic
equation yiel(s a set of complex conjugate roots (RI, RA, R 2 ,
R 2 ). The stress function can be expressed as
F - 2Re {F 1 (Z) + F 2 (Z 2 ))
where Fl(ZI), F2(Z 2 ) are analytic functions of the complex coordi-

nates Z 1 = X + RIY and Z 2 X + R 2 Y respectively. Introducing
the functions
aF(Z1 aF(Z)
lZ(z1I = #2
az2
__2(z2
general expressions for the stress components are obtained:
Ox= 2Re {R I O (Z ) + R2 O (Z 2 )1
C, = 2Re {1 j (z 1 ) + 4 ý (Z 2 )1
Oxy -2Re {R 1 i (ZI) + R2 ; (Z 2 )}
Superscript primes represent derivatives with respect to the com-

plex arguments. Displacements, ignoring terms for rigid body
rotation and translation which in this problem do r~ot affect solu-
tions, can be expressed as:
U = 2Re {P •I (zI) + P2 cb 2 (Z 2 ))
V 2Re {Q I I (z 1 ) + Q2 •2 (Z 2 )}
where
P 1 =SI R1I2 + S 12 - S16 R1I
P2 S 11 R22 + S 12 -S 16 R 2
2k.1 • .. Ii l I I Ill l l
S22 + S 1 2 RP . S2 6
Q22
2 + S12 R2 - S2 6
2
To obtain exact solutions for an infinite platk with a

circular hole (loaded or unloadad) and uniform stresses at
infinity, conformal mappinq techniques were used. A mapping
function was used to map the physical circular boundary of
radius, a, in the Zk plane (k - 1,2) onto a unit circle in the &k
pl e. The mapping function is given by
2 _72 a2
k 1,2
k a(l - iUV) k = 1,2
The sign of the square root is chosen such that the exterior of a
hole is mapped to the exterior of a unit circle.
The above equations contain unknown stress functions * 1 (Zl)

and ý 2 (Z 2 ). For infinite plate problems with a stress concentra-
tion, these functions will have the general form:
cc -M
(Z) B Z + A In 1+ A
M=I
+2 in2 + E A2 M 2-M
(Z 2 ) B2 Z A
M=I
Linear Z1 and Z2 terms are required for a uniform stress at

infinity. Terms with In &1 and In F2 are present whenever the
resultant of the applied stresses on the circular boundary are
nonzero. Boundary conditions on the circular hole are satisfied
by the AIM and A2M series coefficients.
To obtain the solution for a plate with an unloaded hole

subjected to a remote uniaxial in-plane stress field, P, at an
arbitrary orientation, a, with the X-axis (Figure 2), the imposed
boundary conditions at infinity and on the circular boundary
result in the stress functions:
8
2i
2
iPa (1-iR1 )
i(z)= x
1 4(R 1 -R 277
[R2 sin 2a + 2 co 2a+ (2-R2a2 i 0 J

2 2 2
iPa2 (1-iR2 )
¢2 (z 2 ) " 4(R 1 - R2 )
[Z2 + 2 coo 2 a + i (2R

R n2a sin 2a + sin 2a)]
+ 2 -a2 _R2 2 a2
2 2 22
y-
1< 2 2
900
400
--- 45k)
Ip
Figure 2. Unlaxiafly Loaded Infinite Plate j
9
7
Only the linear terms and first coefficient of the eummation are
used for the unloaded hole solut.ion,
Loaded hole analysis was performed by specifying a radial

stress boundary condition varyina as a cosine over half the hole
(Tigure 3). Boundary conditions at infinity required to satisfy
equilibrium result in stress free conditions since the finite
frorce required to balance the bolt load is applied to an infinite
)V.undary. Thus, the lirear terns are not required.
Y
PR _PCO$O
Figure 3. Assumned Cosine Bolt-Load Distribution
Since the specified hole loading is not self-equilibrating

on the boundary, single-valued displacement conditions are
imposed to determine the log term coefficients. The following
set of simultaneous equations are solved for the A1 and A2
complex coefficients.
A A 1 + A2 - A 2 = Py/2 Yri
R A1 R A + R A2- R A = -P/2 7i
11 22 2 2 X
10
2 Y siC: Px
2 - + - 2- S12
1 R1 R2 A2 R2 2 2w S2 2
S1 P + $26P
A /R A /R 1 + A
12 /R.2 _/ 2 2 SI2 2 xni S y
2 2
'The terms Px and P are net load resultants on the internal

boundary in the X anc Y directions respectively.
Expressing the radial stress boundary conditions on the hole

in terns of a Fourier series and equating the series representa-
tion of the solution, the unknown AlM and A2M coefficients are
obtained (Reference 4). The resulting expansion can be written
A12 = aPi (I + iR 2 )/[16 (R 2 - RI)]
A2 2 = -aPi (I + iR 1 )/[16 (R 2 - RI)]
for M = 4,6,8...
A A =0
IM 2M
for M = 1,3,5...
14M1/ 2 2_
AI =-aPi(-l)(M-)/2 (2+i MR )/[-'m,2 (M -4) (R -RI)]
lM 2 2 1
A2 M = aPi(-l) (Ni- )/2 (2+i MR)/UIIiM 2 (M 2 -4) (R 2 -R 1 )]
These equations give the complete elastic stress distribu-

ti.on in an infinite, two-dimensional, anisotropic material with a
circular hole. These solutions are valid only for homogeneous
media, but are assumed valid also for mid-plane symmetric lamin-
ates. Laminate strains are calculated using material compliance
constitutive relations. Laminate compliance coefficients S k are
derived using classical lamination plate theory with uniAirec-
tional material elastic constants, ply angular orientations, and
ply thicknesses. Assuming that laminate strain remains constant
throu]gh the thickness, sLrains for individual plies along lamina
principal material axes are calculated using coordinate transfor-
mations. Stress distributions resulting from an arbitrary set of
in-plane loads (bearing z bypass) are obtained using the prin-
ciple of supcrposition (F 1re
1).
11
To account for inelastic or nonlin,-ir material behavior at
the hole boundary, the "characteristic dimension" hypothesis of
Whitney and Nuismer was adapted (Reference 5). Their hypothesis
states that failure of. a composite material with a stress colncen-
tration can be correlated with analytical predictio-s of point
stresses at a characteristic dimension from the edge of a stress
concentration (Figure 4).
THEORETICAL STRESS DISTRIBUTION
FAILURE TEST POINT
CHARACTERISTIC
DIMENSION
0- OR
Figuie 4. Characteristic Dimension Failure Hypothesis
The application of the characteristic dimension failure

hypothesis was extended to permit strength predictions for aniso-
tropic laminates under general in-plane loadings, without requir-
ing extensive laminate test data. To do this, laminate failure
is predicted by comparing elastic stress distributions with
material failure criteria on a ply-b,,- ply basis at a characteris-
tic dimension away from the hole boundary.
Various material failure criteria can be used with the char-
acteristic dimension failure hypot'Aesis. Interactive (Tsai-Hill,
Hoffnman, Tsai-Wu) and noninteractive (maximum stress, maximum
strain) criteria were evaluated. Failure envelopes for each cri-
teria for the same set of graphite-epoxy (AS/3501-6) material
allowables ignoring matrix failure are illustrated in Figure 5.
Finite width effects have a significant influence on the

circumferential stress distribution around a loaded fastener
hole. A superposition of stress distributions from loaded and
unloaded hole infinite plate solutions can be used to evaluate
the effects of finite width (Reference 6). In the loaded hole
12
analysis, the bolt load, P, is reacted (at infinity) by tensil
and compressive loads oi P/2 (see Figure 6). By superimposing
the solution for an unloaded hole under a remote tensile loading
of P/2 (a stress of P/2Wt) the desired loading on the bolt and
overall equilibrium is obtained. The resulting stress distri-
but ion gives a good approximation of the state of stress in a
plate of finite width but differs from an exact solution in that
the superimposed normal and shear stresses at the "edge" of the
plate are nonzero.
02- ksi
ý--
30
- +20 -' ,
-- +10
-300 -200 -100 +100 200
S_,o
Tsai-Hil-
Tsai-Wu -20 00
~ HoffmanI
S-Maximum re -40
" •I ý -- Maximum Strain
Figure 5. Failure Criteria Comparison
13
V
- -.. f
Loaded Hole Unloaded Hole
Infinite Plate Infinite Plate
P/2 ti0lt
w - w
PI I
P/2WP/W
P/2Wt P/2Wt
SI I I w
P/2W P/2W P/wt
Figure 6. Superposition of Solutions to Account for FInife Width
This methodology has been combined and incorporated into a

computer program entitled BJSFM. Capabilities are programmed to
handle material anisotropy, general in-plane loadin~gs (tension,
compression, biaxiality, shear, bearing), multi-material (hybrid)
laminates, and arbitrary hole sizes. Only mechanical properties
for the basic lamina (unidirectional ply) are required to obtain
strength predictions. A detailed description of the computer
program is given in Volume 3.
14
I
3. PREDICTT'N3 UISING BJSFM APPROACH - Predictions made using the 1
BJSkN -ce presented in this section to indicate the scope and
flexibility of the developed methodology as well as indicate data
ard information pertinent to analysis of bolted joints in compos-
ile structures. The predictions shown are not intended to cover
all iaminates or loading combinations.
a. Effects of Anisotropy on Stress Distributions - Unlike

isotropic materials, stress concentrations in composite materials
are affected by layup and load orientation. Predicted circumfer-
ential Ltress plotted at an unloaded fastener hole boundary with
uniaxial bypass loads indicates important differences between
anisotropic composites and isotropic metals (Figure 7). Layup
variations which change laminate stiffness properties affect hole
boundary stress distributions and stress concentration factors;
however, metal distributions are independent of their stiffness
properties. Also, if loading shifts away from principal material
axes, shear-extensional coupling creates biaxial states of stress
in the laminate. Peak stresses no longer occur 90" to load direc-
tions and distributions shift. Loaded holes also show the same
dependence on layup and load orientation (Figure 8). These
complete stress distributions must be considered to determine
failure load, mode and location of failure initiation of
composite materials.
- ------ Isotropic
Composite
Y
Y
1_41
50/40/10 Lavup 70/20/10 Layup 70/20/10 Layup

0o C 00o
= 0 = 450
Figure 7. Circumferential Stress Solutions at Unioadod Holes
15
Isotropic material
Composite
Y
Y y
_x x
/ /
/
50/40/10 Layup 70/20/10 Layup 70/20/10 Layup

0 00 0° o = 450
GP1S.011544
Figure 8. Ciumfetrntlal Stress Solutions at Loaded Holes
b. Hole Size Effects - Hole size effects are also accounted

for analytically. Stress gradients at the edge of the hole vary
with hole size and laminate orientation. Smaller diameter holes
produce steeper stress gradients which decay rapidly as the dis-
tance from the hole is increased (Figure 9). Applying a failure
-riterion at a constant distance from the hole boundary, hole
sik.-. is accounted for due to the varying stress gradient with
hole z ize.
16
3.0
2.6 00-00-0 0 00
1:: xO 00° ply
2.2
-. °×• • .dia
1.8
5 di i
0.125 dia •• •••
C.FiieWdhEfcs-T
1.0- vlaiefcso iiego
0 0,02 0.04 0.06 0.08 0.10 0.12
Disctance from Hole Boundary - in. OpIli.0 74
Figure 9. Effect of Hole Size on Stress Distributions

I
c. Finite Width Effects - To evaluate effects of finite geo-•
metry on stress solutions and establish limits on the accuracy of

infinite plate solutions, a comparison of BJSFM solutions with
finite element solutions was performed. Two example cases are
illustrated: (1) a pure bearing specimen with a width-to-dia-
meter ratio (W/D) of 8 and an edge-to-diameter ratio (e/D) of 9,
and (2) a pure bearing specimen with a W/D of 6 and an e/D of 3.
The results for the first case are presented in Figure 10. Cir-
cumferential stresses normalized to the average bolt bearing
stress are plotted about the half-circle from directly in front
of the neat-fit bolt (0Q) to directly behind the bolt (1800).
The dashed line indicates the infinite plate BJSFM solution, the
solid line represents the BJSFM solution corrected using the
DeJong finite-width approximation method, and the triangular sym-
bols represent the finite element solution. .- nults shown in
Figures 10 and 11 indicate that for an e/D of 9, the BJSFM solu-
tions, corrected for finite width, correlate extremely well with
finite element solutions where W/D ratios were greater than 4.
17
| ' , . . .. .. . . . . . ..
1.2 -
w/d =81
id-S9
- 0.4 --- ]
GBRG .~
A Finite ale;,ient model

0 BJSFM - finite plate
- BJSFM - infinite plato
50/40/10 Lay pL
j ~~~~~-0.4 .........
0 20 40 60 80 101) 120 140 160 180
0 LOCATION ABOUT HOLE - DEG
Figure 10. Correlation of Loaded Hole Analysis
1.4_
_.1 1 1~
, - BJSFM Prediction
1.2
0) Finite element (,/d = 9.0)
n- • d - 0.25
-x
E0/40/10 Layup
E 1.0 NNW.•
•Z P.8-- _ _
0.6 - 20
0 4 8 12 16
Width/Diameter - w/d
OP15-O1•I46-
Figure 11. Effects of Width on Stress Concentrations
Sensitivity of solutions to small e/D ratios is illustrated

in Figure 12 by results obtained for the second case. At an e/D
value of 3, BJSFM infinite-plate solutions, approximately cor-
rected for finite widths (long dash line), are significant~ly
improved over the original BJSFM-infinite plate solution (short
dash line), but still differ from the correct finite geometry
solution (triangle symbol line).
41.8
'e/d = 3
1.2 - - w_
0.8
0.4.
- - -
BJF fnt wit
Finite element model
- -- ,
0BJSFM - finite width : "
BJSFM - infinite plate
50/40/10 Layup
-0.4 -• ,
0 20 40 60 80 100 120 140 160 180
t) Location About Hole - Deg ON 3411 &W
Figure 12. Correlation of Loaded Hole Analysis for Small Edge Distance Cases
d. Characteristic Dimension Sensitivity Study - Various

laminate failure criteria can be used with the BJSFM procedure.
To fully evaluate the characteristic dimension hypothesis, vari-
ous correlative and parametric studies were performed. A sensi-
tivity study was performed to determine the effect of various
characteristic dimensions on laminate strength predictions.
Loading configuration studies included unloaded and loaded fas-
tener holes,
Results for a 30/60/10 layup of graphite/epoxy with unloaded

hole sizes of .125, .25 and .50 inch are presented in Figure 13.
The Tsai-Hill material failure criterion was used, with strengths
predicted when first ply fiber or shear failure occurred. Each
curve initiates from a common predicted failure stress level
since the theoretical stress concentration at the hole boundary
(Rc = 0.0) is independent of hole size. Due to stress gradients
away from the hole boundary varying with hole size, predicted
failure stress levels change with the characteristic dimension.
The predicted failure stress curves each asymptotically approach
the predicted unnotched laminate strength with an increasing char-
acteristic dimension. These curves indicate that a 10% change in
the characteristic dimension yields a maximum 4% chang, n pre-
dicted laminate failure stress for the .125 inch diam, ,,r hole
and a maximum 2% change for the .50 inch diameter hole. Also, as
characteristic dimension changes occur, failure initiation angle
predictions change, as indicated in Figure 13.
19
70 100I
,Unotchd LaminateS ro
"c
007575
900 Failunt__ 4
60 -. - Initiation Angle 90
i ~ 0=.125--
,50 8
0 0.008 0.016 0.024 0.032 0 0.40 C.80 1.20
Characteristic Dimension, Rc- in. 0pw1§4l547
Flgur; I&3 Effect of Charectedstic Dimension on Predicted Strength

30/60110 Layup
e. Biaxial Loading Effects - To demonstrate the utility of

the Bolted Joint Stress Field Model, a parametric analysis was
performed to predict interactive effects of biaxiil loads on lam-
inates with unloaded fastener holes. Laminate strength, failure
location and critical plies were pi.edicted for a representative
composite wing skin laminate.
Figure 14 presents predicted failure stress and indicates

changes in failure modes as biaxial load ratics increase. Load
ratios of increasing (NY/Nx) for this tension-tension case pro-
vide relief at the fastener hole and laminate strength Increases.
Increasing shear load ratios (Nxy/Nx), however, intensify stress
concentration effects at the fastener hole; laminate strength
decreases and failure modes change from 0* Flies to 45° plies.
Figure 15 presents the predinted location at the fastener hole
boundary of first ply failure. Significant is the indication
that at certain biaxial load ratios, a wide arc of the hole bound-
ary may become critical. Figure 16 presents strain concentration
factors for critical plies within the lamirate with respect to
the strain developed under Nx alone.
20
55 1-7 00
Nxy Ply Fails
N~ x -oo,,•
5Op
Fails
05015
S = 0.250
I
0
CONN,
45
-496'5011 LayuIJ
o0 0.10 0.20 0.30 0.40 0.50

Ny/Nx
FIgure 14. Effect of Biaxial L*)adIng on Predicted Strength
21
130 I
46/50/4 Lavup
0
120
- 110
ol '50 Ply
4-.o•
LA.
> o1
w.10,/ - ~~~~0.25 -. 0 / .1
Fails
C
0
1001
-9
00 Ply
V. ,Fails
0 0.1 0.2 0.3 0.4 0.5

Ny
Figure 15. Effect of Blaxial Loading on Failure Location j

i
22
iU
39 ( 03
S-0 N
41!
3.7 ý
UL J
xV
Ic ________46/50/4 Layup
0
00
v
LO 35
'21 z-
Ply
+45 Faiils
0 0.1 0.2 0.3 0.4 0.5

NyV
Nx
Figure 16. Effect of Blaxial Loading on Ply Stiain Concentrations
Experimental and analytic studies have indicated that fail-

ure locations at fastener hole boundaries shift depending on
layup and biaxial stress states. The BJSFM methodology accounts
for this phenomenon and, further, reveals thc critical nly at the
failure location. Data shown in Figures 14, 15 and 16 show that
failure analysis must account for all locations around the fasten-
er hole, since the critical area is layup and loading dependent.
This is best indicated by Figure 15 which shows not only continu-
o'is shifts with biaxial loading, but also indicates wide arcs of
the hole perimeter may become equally sensitive to failure. This
in(licates the potential errors which can arise by using methods
which predict laminate failure by referencing unnotched laminate
strength data at "preselected" or "representative" loc-ations on
the hole boundary.
23
L.. m... .. . ... - wJ

--.
--
f. Pure Bearing Strength Study - Strength envelopes are pre-
dicted based on ply-by-ply analysis to determine critical plies
and failure initiation location. The results of a ply-by--ply
evaluation of strength based on first ply fiber rupture for the
50/40/10 layup under pure bearing loads is presented in Figure
17. The ply strength envelopes predict initial failure in 90*
pl.ies f. - all load directions with relatively constant laminate
strength up to a load angle of 25° and a gradual decrease from
25' to 90*. This predicts that while unloaded holes show a pro-
nounced sensitivity to off-axis bypass loading, bearing strengths
for this laminate are relatively unaffected by off-axis loadings.
Analytically, this behavior is a direct result of the differences
i ~ in
cond local
it ions. stress distributions resulting from the two loading
I,
!• ~ ~32o,0
280
50/40/10 LAYUP
STRESS 20 ,---__5°PLE
KSI 1240" .. /-O L p•--
80 ---
OPI).1PLY)
0 20 40 60 80 120
STRSDEG OFF PRINCIPAL MATERIAL AXIS
Figure 17. Effect of Bearing Load Direction on Predicted Strength
24
P
DOFF
g. Effects of Environment - Effects of environment on lami-
nate strength can also be accounted for with the BJSFM analysis.
Unidirectional material properties (stiffness and strength), as
affected by an environment, are used to redefine laminate beha-
vior. Bearing versus bypass failure envelopes were predicted for
a 50/40/10 layup at room temperature dry and at elevated tempera-
ture with moisture. For this layup, results indicate that little
change in predicted strength occurs for high bypass loads associ-
ated with fiber dominated modes of failure which are usually not
adversely affected by temperature and moisture (Figure 18). How-
ever, elevated temperature and moisture do affect matrix proper-
ties resulting primarily in lower shear modulus, shear strength
and lower compression strength. The lower shear stiffness causes
a redistribution of high bearing stresses to the stiffer 45*
plies in compression. This, coupled with decreased compression
strength, results in fiber compression failures at lower bearing
stresses than corresponding RTD properties (Figure 18). Although
the failures for this layup initiate in the same vicinity around
the hole, different phenomena cause the failures due to changes
in material mechanical properties with temperature and moisture.
80
RTD and ETW

00 Plies - OCR = 900 PB ---- Bypass
Fiber Critical/ ___
60 - -
c~40
> /--ETW F-RTD

+450 Plies
40OCR
OC =c 450
00/900 Plies
Fiber Critical Shear Critical
20- _ __ _
0 20 40 60 80 100
Bearing Stress - ksi 0p1S-OI 1B-M
Figure 18. Effect of Moisture and Temperature on Laminate Strength
25
' ~ ~~~~_.Mi•,
r .0 'l "- .
SECTION IV
TEST DATA EVALUATION
1. METHODOLOGY VERIFICATION: EVALUATION OF JOINT DESIGN VARI-

ABLES - TASK 2 - The primary objective of the Task 2 expe~rimental
program was to verity developed methodology through strength
tests over a range of bolted composite joint design variables.
Test data were also used to provide direction for development of
further improvements in the methodology. The Task 2 Test Matrix
is presented in Figure 19.
The selection of test variables was guided by information
gathered under the Task 1 - Literature Survey. Data obtained
from earlier MCAIR in-house test programs were identified under
this survey (Reference 1). MCAIR test programs outlined in
Figures 20 and 21 were used to supplement the Task 2 experimental
data. Ten different layups oi the 0°, +45', and 90° family of
ply orientations were tested under tension and compression load-
ings. Effects of hole size, thickness, countersinking, edge and
width distances, and joint eccentricity (single versus double
shear) on the strength of laminates with a single loaded or
unloaded fastener hole were evaluated. Specimen configurations
for unloaded fastener hole and puce bearing evaluations are
detailed in Figures 22 and 23.
These MCAIR data provided an initial basis for the effects
of layup variations on laminate strength for the two bounds of
bolted composite joint load-transfer, i.e., zero bolt bearing
(100 percent4 bypass+ strs an tota load
l ;
trn fir t-hrt'iuh a
single fastener (pure bearing). The availabili.ty c these data
permitted early verification of developed methodology, reduced
the scope of layup variations to be tested, and permitted Task 2
experimental efforts to include a more comprehensive range of
other bolted composite joint design variables.
a.Task 2 Test Plan - The effects of 12 design variables on
joint strength were evaluated. Tests were performed at three
environmental conditions: room temperature dry (RTD), room
temperature wet (RTW), and elevated temperature wet (ETW). Envi-
ronmental conditions were selected to be realistic with respect
to deployment profiles for multi-mission, high-performance super-
sonic aircraft. Tests were conducted at an elevated temperature
of 250°F and, when applicale, specimens were conditioned to an
equilibrium moisture content of approximately 0.8 percent by
weight. All specimens tested at 250°F were held at temperature
10 minutes before testing. Specimens were loaded in tension or
compression to failure. In general, a replication of four tests
were performed at each test condition.
Hercules AS/3501-6 Type II (ten mil thick) graphite-epoxy
was used as the principal material for fabrication of test
specimens. Limited testing was performed using NARMCO T300/5208
graphite-epoxy.
26
Z17
[ _ _ _ _
ClU
+ >'>I
m
>>
OocB-
>F>
fs , 0m
i
» ~~& ;
ot
N»
c
C2L L2hI!0
_j- __ "0I
Kl z caOO@~
L
-cow0
L-C~
CL
LU LU i lU0
CL C. CL CL CL I.
,- >-)CIL
B-B f - -B
ca c
LU LUZV B-j
Zf LUu
0j LU LUIE
ca cc -M
ot C3 0 ~-C )
U
,
LL cn n Ln L
27 0
o~11
LU
4.
- ci
z ~ in- »
m»a,>RRC
_:- C44"
44F , f 5
4.- 6
,, I [ <
* . - _ _ _ __a
.- SM IP
.5.L
0L* Il 'mII! ..; xL
I.1 c 00
-CC
SM IIJ SMSXM M
a-
0 N SU a 4~
z4-
aoU
Wn 0
____ _
m
_
W _____
_ _ _ @
_ __ _ _ _U _
cm.
Laminate % Layers of No. _ Hole Dia. Loadinr N of
Group 00 +45 900 Plies 0.01 0.125 0.250 0.500 Ten. Com. Tests
A 70 20 10 40 V V V 24
B so 40 10 40 V V V V V 24
120 V V 18
C 30 60 10 40 / V V V V V 24
108
S50 10 40 40 V/ V V 12
E 40, 20 40 40 V VI V/V 12
F 40 50 10 40 V V V V 12
G 30 30 40 40 IV' IV/V 12
H 20 40 40 40 V V V V 12
1 20 70 10 40 V V 12
J 10 80 10 40 V V V V 12
84
A 70 20 10 40 X X X 6
B 50 40 10 40 x X X 6
120 X X X 6
30 60 10 40 X X X 6
30
Total No. Tests 222
V Noncountersunk Specimens
X Countersunk Specimens 0Pl13111"iI
Figure 20. MCAIR Evaluation of Unloaded Hole Specimens • Test Matrix
29
Phase Layup %Layers No. Hole Die eld wid Single Double Total
No. No. 0O +450 0o Plies 0 0.250 0.375 0.500 0.250- 1.5 2.0 2.5 3.0 4.0 4 6 8 Sheer Shoat Specinons
1 170 20 10 20 0 V V/ V ••V V/ - 87
TI 2 50 10 40 20 0 V .-- -V- V- V - 9
I 3 50 40 10 20 0 VN/ V VV* V A 86
1[ 4 40 20 40 20 - V.-V-V- - V - 6
U 5 40 40 20 20 - V . . . .- V-- - V - 6
R 6 30 20 50 20 - V - - - - - V - 6
I 7 30 60 10 20 0 V V V VVVVVVV - 87
R 8 20 40 40 20 -- V - . . -V-V -- V-V - 6
1 9 20 60 2020- - - V
V/- V - -V -- 6
HI 1010 80 10 2P - V - . . .- V-- -V V - 6
11 11 55 40 5 20 0 V - - - V V V - V
- / - 15
---- 15 92 103 72 36----Total 323

Each test replication of 3 0P13-0115.121
Countersink
e Baseline
specimens in double shear
Figure 21. MCAIR Evaluation of Loaded Hole Specimens Test Matrix
11.000 in
d ' J.500T
900
0-/ _.1i in.
3.000
_ _d = 0.250
F 00 -/ 1.500 in.
1900 _j_
d = 0.1251
0- 00 --
O_____-
___ 0.75 in.
40 Plies 2000 T
0.375 in.
&---0Strain Gage
Figure 22. Unloaded Hole Specimen Geometry
30
ip
- t9 -' - i4'
eI
90° -- \--
450
Strain Gage
PTest Specimen Bload
Figure 23. Loaded Hole Specimen Configuration
Four test specimen c:onfigurations were used: (1) a single-

bolt pure bearing specimen, (2) a baseline two-bolt-in-tandem
bearing-bypass specimen, (3) a two-bolt load interaction specimen
and (4) a four-bolt fastener pattern specimen. Both single and
two-bolt-in-tandem specimens were loaded in either a single-shear
or double-shear configuration (Figure 24). All specimens were
strain gaged. Recorded data included: failure load and strain,
continuous load versus strain plots to failure, selected load
versus deflection plots to failure, thickness, width, hole diam-
eter measurements, weigh- gain of humidity exposure specimens,
and selective photographic documentation of failed specimens.
31
(a) Double Shear Face Configu ration
0.025 Max Gap
Strain GageSt
kClevis
P I e
Load
J --- P
Graphite/Epoxy
Specimen
0.025 Max Gap
(b) Single Shear Fac. Configuration
-Graphite/Epo'y Specimen
fSteel Load Block
OF13-OlI1-4t
Figure 24. Specimen Loading Configura\)Ions
Prior to specification of load-clL. ge. metry (thickness,

width, and material), a parametric study was pe4rformed to evalu-
ate the effect of clevis-to-laminate stiffness ratios on load
distributions for the baseline two-bolt-in-tanlem joint. Tbe
objective was to quantify the errors in load dio-tribution which
could arise due to material stiffness variability. This wou.Id,
in turn, provide a theoretical data base to aid in selection oi
load-clevis geometries and material. The re,3ults of this study
are presented in Figure 25. Based on theke results, a load
clevis-to-laminate stiffness ratio of 10 or more was selected.
This choice results in unequal bolt loads, but c'uarantees that no
distribution error greater than 0.5 percent sh-uld o.cur due to
laminate stiffness variations expected from specimen to specimen.
32
iJ
P1 P2
.TotaI _,__ Tota,

S~Clevis
Laminate
Load Clevis Stiffness Bnlline - 0 - 10

Specimen Stiffness Fraction of Total Load
0.6 Layup P1 P2
30/60/10 0.442 0.558
50/40/10
70/20/10 0.457
0.465 0.543
0.535
0.5
S"-ý-10 (Baseline)
I- 0.4
0..0
0.3
30/60/10 50/40/10 70/20/10
0.2
0 4 8 12
Ex - Laminate Stiffness Specimen - (106 psi)
f
16
I 20
P14k116.M
Figure 25. Edfect of Laminate Stiffness on Specimen Boll-Load Distributions
Individual test specimens were fabricated from 19 panels.

Stacking sequences are detailed in Figure 26. Panels were cut,
collated, and cured per standard MCAIR process specification.
Al!. specimens were tested to static failure according to Figure
19 under tensile or compressive loadings. Complete details of
all procedures, test results and support requirements are
contained in Volume 2 of this report.
33
__..p. •
a0 + Io00c0+
,, + I ISO0 Q ICC
zS
20
Cc + 1 00+ 10coo
2C.2
+0 la~oI +0 1080+0 IC +01080+0 lgo V
2- z
z ooooo~ i ~ ~ ~ a0
4L_ _ _ _ _ _ _ -j 06__
~0
_ _ _ _ _ _ _ _ _ _ _ - to
uJr
~ -NI~w~a~-0>
9L ~ .E
34j
b. correlation of Ik7SFM Ex erimental
Predictions With
Results luu experimental
G..Ctlit UumnhIpres
results ol-tained under Task 2 and predictions using the BJSFM.
Where IRJSFMl/test correlation was not possible, data trends are
?or aninitMiXl
JSF i-rccedur
veriicaionof
he fo th cae o anunloaded
fastener hole. Using the IV.SFM procedure, strength predictions
were made for all test layups and teati conditions. Lamrinate elas-
tic properties were calculated using unidirectional ply lamina
elastic constants. Laminate failure was assumed to occur when
first ply fiber or shear failure w'as predicted, based on the Tsai-
Hill or maximum strain material failure criterion applied at a
characteristic dimension away from the hole. Only laminaj
strength and stiffness data were required for this analysis;
these data are shown in Figure 27. RTD data are average values
obtained fromi MCA1R tests performed on unnotched unidirectional
laminate sandwich beams, arid 0*/9Q@ rail-shear specimens; RTW
and ETW data are estimatod.
Poete;ROOM RT Wet 250OF Wet

Temperature IMstimsaed)*11 (Estimafted)**
Elastic Constantsfj
is Eilt(106 psi) 18.85 18.85 18.54
6
"* Elc (10 psi) 18.20 18.20 17.80
" E2 (10'P psi) 1.90 1.72 1.27
* 12 (06 psi) 0.85 0.77 0.60
120.30 0.30 0.30

" f 1t (pin./in.) 12,206 12,206 12.735
"telu (pin./in.) 176012,6948, 6
"a~t (pin/in.) 5,380
* 712 (pin./inj* 20,352 22,470 18,461

is F 1tu(ksi) 230 230 236
e Ficu (ksi) 321 231 151
is F2 tu (ksi) 9.50 -
0 F 2cu (ksi) 38.90 35.20 26.00

to F 1 2 (ksi) 17.30 17.30 11.00
'Based on linear-elastic behavior GPSSIIS1-ttS

-Approximately 0.86% moisture
Figure 27. A5/3501-6 Lamina Mechanical Properties
35
Correlation of predicted tensile strerg9th with test dati
from ten layups is shown in Figure 28. Currelation of predicte.i
effects of hole size on laminate tensile strength with *-eat data
is shown in Figure 29. For the AS/3501-6 material system, a char-
acterietic dimension (Rc) of 0.02 inch was used for all tensile
failure analysis. This value was obtAineti b, comuparing .250 inch
hole-sire strength data from the 50/40/10 laminate with predic-
tions, as illustrated in Figure 30.
-- Theoretical * Test data A Predicted
120 I~~~
Rc 0.02 in.
100 -- %0 0 Plies
080
b
S6040 90/
'U.I
3x 400
0D
I10
20 60 60 8 10
20
% ± 450 Plies OpIzt..
Figure 28. Effect of Layup Variation on Unloaded Hole Tensile Strength
0!
36
180II ' -- I
fstu
4Uiross d
160 1 1"
140g
"__I I
-~120 -- Predicted
S 100
J .x 80 ' -2 70/ 0 /10 10 )

4 X
LLI
60 --
SI-- Wi/40/1 0 fA
40 • • •J
! ~30160/10 10)
20 "
14,000
12.000
0.250
.~10,000 -__ CSK
I Average
M
U ' Hole Size
(n 8.000
6,000 --- 70/60/1
4,000
2,000-
0 0.1 0.2 0.3 0.4 0.5 0.6
Hole Diameter - in. Qp Ml.5Iw
Figure 29. Effect of Hole Size on UnIo6ded Hole Tensile Strength
37
i! . ...... ... ~ . ,.
~~j .....- ,.
72 , I
Rc -d 0.250
FtU-
p66
-- ./ Predicte
0~ I
S64 - -
U-
62 --
0.02 Value of R. Used
in All Tension Analysis
58|
0.015 0.017 0 .019 0.021 0 .0 2 3 0.025 0.027
Characteristic Dimension, Rc - in.
Figure 30. Effect of Variation of Rc on Predicted Tensile Strength
Location of failure initiation depends on layup and hole

size. Predicted failure location occurred on the hole boundary
and at 90* to the applied load direction for all but two cases.
The exceptions are the 70/20/10 layup with 0.250 or 0.500 in.
diameter holes. Predicted origins of failure were at 65* for the
0.250 in. diameter hole and 70° for the 0.500 in. diameter hole.
in these two cases, failure was predicted in the +45* plies. All
predicted locations of failure correlated visually with test data
(e.g., Figure 31).
38
(30/60/10, 0.500 in. Dia Hole)
700/j
(70/20/10, 0.500 in. Dia Hole)

Piedicled f-jilure Ore,)jIjl~jl GPIe-Ot.117
Figure 31. Predicted Failure Orientations Were Verified
!ýtrength predictions for a biaxially loaded laminat., with an

open fastener hole were also verified by test. This loading con-
dition was achieved with uniaxial. loading at various angles rela-
tive to the laminate principal material axis. Due to orthotropic
material hehavior, off-axis uniaxial loading causes shear-exten-
sional coupling which produces a biaxial stress state in the test
speciiilen gage area. Using the 0.02 in. characteristic dimension,
correlation of tes;t data wit]h predictions is shown in Figure 32.
A characteristic dimension of Rc = .025 inch was used for

all compressive failure analysis. As in the tension case, Rc was
initially determined using only .250 inch hole size strength data
obtained from tests performed on the 30/60/10 laminate (Figure
33); however the final value of .025 inch was based on a best fit
of all hole-size strength data for the 30/60/10 layup. Predic--
tions and test data on the effects of hole size on compression
strength for this layup as well as the 50/40/1) and 70/20/10
layup are presented in Figure 34. A carpet plot of compressive
strength predictions for all ten layups of Figure 20 is corre-
lated with test data in Figure 35. Test data and predictions are
for the open hole, .250 inch diameter, specimen configuration.
7\ preliminary evaluation of the effects of temperature on

coMpressive strength of laminates with .250 inch diameter
unfilled fastener holes was also performed. Three different
laninates were tested at room temperature and 250°F. Using the
BJSFM procedure, strength predictions for both room temperature
and 250°F wJere based on a constant characteristic dimension of
0.025 inch and the Tsai-Ilill Failure Criterion. Correlation of
39
60-
50
. __Rc =
-48/48/4
0.02
,AYUP-
w/d = 6
Fu 40
FAILURE __REDICTED
STRESS 30
KSI 00 PLY REF
__ __ _
20 F()
10 d 0.250 F0
0 10 20 30 40 50 50 70 80 90 100
S- LU AD ORIENTATION - DEG
Figure 32. Effect of Off-Axis Loading on Failure Stress
40
-66_
0.025 Value Used for I
All Compression Analysis
-642
, I¶
d = 0.250 Data I
-• Predicted
•x -60 4
LL
KY
2 -58
_= •e" 30/60/10 Layup
u. .0Rc 0. 0d7 0 0.250
Chrateisi x-in
Dies, R
-54
IN
S~00 Ply Ref 1
-521 1i
0.019 0.021 0.023 0.025 0.027 0.029 0.031
Characteristic Dimension, Rc - in.
Figure 33. Effect of Variation of Rc on Predicted Compressive Strength
41
___ ___ I
I 2
-180 I
Cu u
-140 x Sso 6d
I
LL 120
-100
-80 __ 04/bA
S10/20/1 Hole Diameter in.

20,000 T
18,000
n Prediction
16,000 .
c!14,000 -
70/20/10 Weighted
Fgr 12,000
e or Hole Size t
x for 0.250 Csk
--- 0 0.1 0.2 - -'T

..... . 500.3Diameter
,,-Hole 0.4
/10g - in.
0.5 0.6 0.7 ...
PIM
Figure 34. Effect of Hole Size on Unloaded Hole CompreSSIve Strength
42
Theoretical 0 Te%t Data A Predicted
120 %o o 'ie ,
90• •r = 0.025
1007 'ioo ~80 _______

1 8 7060
"*80
050
0 60
U 40
U_
01
0 20 40 60 80 100
% +450 Plies
GPI"11t -101
Figure 35. Effect of Layup Variation on Unloaded Hole Compressive Strength
test data and predictions is illustrated in Figure 36. Results

indicate that effects of enemperature on unloaded hole laminate
strength for a wide range of layups can be satisfactorily
predicted by using the BJSFM procedure with appropriate
temperature-corrected values of lamina elastic constants and
lawnina strengths.
For all test conditions, compressive failure modes of open
hole specimens were similar. Laminate material in the net-
section area "broomed-out" symmetrically about the thickness
centerline, typical of compressive failures in unnotched
laminates.
Failure initiation points were predicted to occur on the
hole boundary 85-90' frow the specimen length axis in all but the
highly orthotropic layups. For these latter layups (70/20/10 and
50/10/40), failure initiation points were predicted to occur at
1250. However, due to extensive laminate compressive damage, vis-
ual verification of failure initiation points was not possible.
4
43
~- 4~-
II
Sw
SxCU d =0.25xCU
180
i 80 50/4b/10
Rc = 0.025 in.
"-60 Test data

.42/5 0 Predict3d
LL w/d =6
40f-m=
20/60/20
T
0 50 100 150 200 250 300
Temperature - OF
Figure 36. Effect of Temperature on Unloaded Hole Compression Strength
(2) Strength of Laminates With Loaded Fastener Holes -

In the Task 2 test matrix (Figure 19), evaluations of bolted
composite joint strength can be sub-grouped under four studies:
(1) effects of orthotropic mechanical properties, (2) effects of
joint geometry, (3) effects of loading and fastener pattern
interactions, and (4) effects of through-the-thickness variables.
A baseline 50/40/10 layup and joint geometry (single or two
fastener-in-tandem confiquration) detailed in Fiqure 37 is used
throughout the Task 2 test program as a reference point. Details
of specific specimen geometries are shown in Fiv,•res 38, 39, 40
and 41.
Evaluation of effects of orthotropic mechanical properties

included layup and load orientation variables. Effects of lami-
nate variations (70/20/10 and 30/60/10 layups) on joint strength
were compared to the 50/40/10 baseline strength data for the load
condition of bearing and bypass loads aligned with a principal
material axis. Effects of bearing load, off the principal
material axis, on ultimate bearing strength were evaluated for
the 50/40/10 layup under pure bearing loadings. As-manufactured
and moisture conditioned specimens were tesCed in tension and
compression at room temperature and 250 0 F.
44
15.00 -
1.000---- - - 0.750
;" , I,,,
I, ,, p o 1.500
I I /
d0G 2 +0.0022 _/ SS
0 Strain Gage -0.0000 20 Plies
Notes: T
Laminate: 50/40/10 Layup
Stacking Sequence: [+450, 00, -450, 00, 900, 00, +450, 00, -450, 00]s
Thickness (t): 0.208 in. Nominal (20 Plies)
Fastener Type: ST3M 453-4 (0.2495 + 0.0000/-0.0005 in. dia)
Torque Value: 50 in.-Ib (1/4 in. Fastener)
Load Configuration: Double-Shear
OPI"-116-47
Figure 37. Two-Bolt In.Tandem Specimen Baseline Configuration
45
,,I
wi
15.00
-3.50-
S--I --- i
Cut Off and d

0GStrain Re-Drill for t
Next Test-
II 1 .
Design w e d
VaribleVariations
Variable w
(in.) (in.) (in.) __________
Plies in.
Load Oreintation All 1.50 0.750 0.250 20 0.208
Countersink (h/t) 1 2.25 1.125 0.375 20 0., j8

(h/t) 2 2.25 1.125 0.375 40 0.416
(h/t) 3 2.25 1.125 0.375 60 0.624
Single Shear Baseline 1.50 0.750 0250 20 0.208
(Protruding Hd) e/d 1.50 0.500 0.250
U 20 0.208
w/d 1.00 0.750I 0.250 20 0.208 J
d 2.25 1.125 0.375 20 0.208
Stacking Sequence A:: 1.50 0.750 0.250 20 0.208
Fastener Torque All 1.50 0.750 0.250 2C 0.208
Thickness t1 2.25 1.125 0.375 40 0.416
t2 2.25 1.125 0.375 60 0,624
OP1lC-11546
Figure 38. Variations of Single Fastener Specimen Configurations
4.6
- 15.00
-,---- 3.50 S-" e
-4
00 Strain Cut Off and / -- d
Re-Drill for t
Gage
Next Test r i
Disgn Vaitos w & d S t

Variable Variations (in.) (in.) (in.) (_r.) (in.)_
Layup All 1.500 0.7500 0.2500 1.500 0.208 j

Edge Distance (e/d) 1 1.500 0.3750 0.2500 1.000 0.208
(e/d) 2 1.500 0.5000 0.2500 1,000 0.208
(e/d) 3 1.500 1.0000 0.2500 1.000 0.208
S1 1.500 0.7500 0.2500 0.500 0.208
S2 1.500 0.7500 0.2500 0.750 0.208
Widxh (w/d) 1 1.000 0.7500 0.2500 1.000 0.208
(w/d) 2 1.250 0.7500 0.2500 1.000 0.208
(w/d) 3 2.000 0.7500 0.2500 1.000 0.208
Hole Size D1 1.137 0.5685 0.1895 0.758 0.208
D2 2.250 1.1250 0.3750 1,500 0.208
D3 3.000 1.5000 0.5000 2.000 0.208
Single Shear All 1.500 0.7b00 0.2500 1.000 0.208
Fastener To'que All 1.500 0.7500 0.25U0 1 1.000 1 0.208
OP13-0115.45
Figure 39. Variations of Two-Bolt-IWn.Tandem Specimen Configurations
47
15.00
0.750/T
V ~0.7501
s00 Strain Gages
0.208
[I_ ;; -Vl
Design d W S
Variaale (in.) Variation (i.) (in.) (in.)
Load Intaraction 0.250 01 (100) 2.00 2.836 2.880

02 (22.50) 2.50 2.414 2,613
0 (450) 3.50 2.000 2.828
0.375 0 (00) 2.25 3.800 3.800
P0 P
Loading Configuration 0P151W5
Figure 40. Variations of Load-Interaction Specimen Configurations
48
,=8 p.
n o 1.000 2.500
3.50 - 1.000 0.750
-- 7.50
____ ______ ________ _________ ____ ___ji 0

208
GPIS.•IO11-6
Figure 41. Fastener Potterm Specimen Geometry
Inthe second group, effects of joint geometry, joint vari-

"ables evaluated include edge distance (e), width (w), and hole-
size (d) for the 50/40/10 laminate. Specimens were tested in
tension at room temperature dry and selectively at 250*F after
moisture conditioning.
Studies of loading and fastener pattern interaction

I. involved: (1) evaluation of the effect of bearing stress on
laminate bypass strength and (2) evaluation of stress concentra-
tion interactions in multi-fastener patterns. Bearing-bypass
interactions were evaluated for two layups (50/40/10 and 30/60/
10) with bearing and bypass 2oads aligned with the 0* principal
material axis and, for the 50/40/10 layup, when bearing loads
were oriented at 100, 22.50, and 450 relative to bypass loads
which were parallel to a principal material axis. Four-bolt
specimens were used to evaluate the interaction of fastener
stress concentrations for two layups (50/40/10 and 30/60/10) and
were tested selectively in tension and compression at room temper-
ature and 250°F after moisture preconditioning.
49
- -k - .,-._ • .
Five through-the-thickness joint variables - (1) torque-up,
(2) single shear (joint load eccentricity), (3V laminate thick-
ness, (4) countersinking fasteners, and (5) stacking sequence -
were evaluated to determine the possible three-dimensional inter-
actions at individual fastener holes. Single-fastener pure
bearing and two-fastener bearing plus bypass specimens were
tested in tension and compression at room temperature dry condi-
tions. Selective tests were performed at room temperature and
250°F with specimens moisture preconditioned. Strength data were
compared with the baseline joint configuration.
(a) Layup Variation - Effects of layup variation on

laminate bearing strength were evaluated using the two-fastener
in-tandem specimen. Three layups (70/20/10, 50/40/10 and
30/60/10) were tested in tension and compression. Tests were
performed at RTD, RTW, and ETW. Results of all tension tests are
presented in Figure 42. Results of all compression tests are pre-
sented in Figure 43. In both sets of data increases in gross
failure strain were linear with increases in percent of +450
plies contained in the layup. Parallel trends were oserved for
moisture conditioned specimens tested at 250°F (ETW), with
strains genezzlly lowered by 800-1000 win/in for all layups both
in tension and compression. Specimens of the 50/40/10 layup,
which were moisture preconditioned and tested at room temperature
(RTW) exhibited little strength changes. Associated ultimate
bearing strengths reflected similar trends between RTD and ETW
data with approximately a 30-40 KSI loss in bearing strength for
all layups tested at ETW.
Associated joint load-deflection data indicated the

presence of nonlinear mechanical behavior. Predictions, using
the BJSFM procedure, and joint loads at which initial nonlinear-
ity occurred are also presented in Figure 42 and 43. Predictions
were based on the previously determined Rc values of 0.02 inch
and 0.025 inch for tension and compression respectively.
(b) Load Orientation - Effect of material ortho-

tropy on pure bearing strength was evaluated using a single-
fastener specimen fabricated at various angles off the 0°
principal material axis. Specimens were tested selectively at
room temperature dry and at 250'F after moisture conditioning.
Strengths at room temperature dry conditions are presented

in Figure 44. Included are bearing stress values at which ini-
tial nonilinearities or abrupt changes were observed in joint
load-deflection data. All off-axis test specimens were loaded in
double-shear with fastener torque-up of 50 inch-pounds. BJSFM
predictions indicate first ply fiber rupture or shear failure,
based on a maximum strain failure criterion applied at a charac-
teristic dimension of 0.02 inch away from the hole boundary.
Correlation of this linear-elastic analysis with initial points
of nonlinear joint deflection behavior, indicates that lamina
shear or fiber failures are likely before ultimate bearing
failure occurs.
50
L I
I I Double-shear
"Gr3~oss 4d 3d Ultimate failure
F 50 in.-lb Torque-Up 0 Inlitial nonlinearity
160
120
.0 xPred.cted
40 ___ ____________
_______
6,000 30/60/10
'000
Ci
* 2,000
702/0 20 4Predicted
60 80to
Percent of 1450 Plies

OP13-O 115-340
Figure 42. Effect of Layup Variation on Joint Tensile Strength
51
-II
d -,0.250 -
PTotal I P2 6d
XGross 4d 3d
50 in.-lb Torque-Up U 0 A Ultimate strength

Double-Shear 0 A Initial non-linearity
170
• •, f •ETW
130-
U-
90
/
50"
7
a - ,000, r I
C -J
__-__--5,000 Predicted
S-5,000 -
.0.0,
IIl
--1,000,
0 20 40 60 80 100
Percent of ±450 Plies CIP13-0115,44
Figure 43. Effect of Layup Variation on Joint Compressive Strength
52
j
d = 0.250-\ Double-shear
0 Vo•\N
- _T 50 in,-Ib torque-up
P 6d 0Oin.-lb torque'up
0 [0 Initial nonlinearity
50/40/10 layup
3d
160,
40
Figu ein.Shenr FaitinrShte re
80
0 20 40 60 80 100
ditioned~~
spcmesat20e.Thoe reiced ictlped:cin
Off-Axis Loading Angle, 0i - deg
tbt T
a P13-0115i-
Figure 44. Effect of Off-Axis Loading on RTD Pure Bearing Strength
Strength of moisture conditioned specimens is presented in

Figure 45. Trends appear to be similar to RTD results with joint.
gross failure strains reduced by approximately 700-1000 kimi/ in
and bearing strength by approximately 40 KSI in the moisture con-
ditioned specimens at 250*F. Theoretical predictions at both RTD
arid ETW are indicated in Figure 45, versus strain data for ini-
tial. points of nonlinearity. Both ETW and RTD prediLctions were
based on the .02 inch v~ilue of Rc. Results from off-axis lciding
at 100 under compression loading ara presented in Figure 46.
53
pI
200 __
A* UUltimate
0 0 Initial nonlinearity -6d
160---
50 in.-lb rorque Up
Double-Shear 3
RTD3
120
-0XT W
80
40 ____
6,000
5,000
T D
4,000
Predicted RTD
2,01050
1,00 woo___ \ Predicted ETW
0 20 40 60 80 100 120 140

Off Axis Loading Angle, 0 -deg OI164
Figure 45. Effect of Temperature and Moisture on Off-Axis Strength
54
P p1 •
6d_
bl, Initial cu
Test Condition x Nonlinearity _ _ X
50/40/10
Layup ksi Prcent A ksi Percent A
Z an./in. Percent A
of IRTD of RTD of RTD
RTD 153 - - -2,640 -
RTW 157 +2.7 122 +10.4 -.2,830 +7.3
ETIW 113 -26.3 76 -3.8 -1,980 -25.11J

OP130115.42
Figure 46. Effect of Temperature and Moisture on Compressive Strength
Edge Distance
(c) - Effects of edge distance on
joint: strength were evaluated by testing two-fastener-in-tandem
double shear specimens. Tests were performed on dry specimens at
room temperature and on moisture conditioned specimens at 250°F.
3pecimen.-, were loaded to failure in tension. Additional RTD data
on the effects of edge distance on pure bearing (single-fastener)
streorgth were available from an earlier WCAIR in-house program.
Results of RTD and ETW tests are presented in Figure 47.

Specimen failure modes at all e/d ratios exhiLited shearout to
bearing-shearout failures. Joint bearing strenoth and gross
failure strain data exhibit a marked bilinear behavior. Reduced
joint strength with earlier insensitivity to e/d ratios was
exhibited in the ETW tests.
55A
d 0.250\
P 1 + P2 -d I
50 in.-Ib Torque-Up x 4d e
160
i RTD
120 .TV
C,,
80
40
5,000
RTD
4,000
S'• __ ETW -
3,000
2,000
1,000
0 1 2 3 4 5
Edge Distance/Diameter , e/d
OP134-11E-40
Figure 47. Effect of Edge Distance on Joint Tensile Strength
56
Effects of fastener spacing (S) on joir.t strength were
evaluated at RTD test conditions. Results (Figure 48) indicate
that for high-bearing to bypass joint loadings, strength reduc-
tins occur as S/d ratios decrease from 4.0 to 2.0. While
fastener spacing creates a beneficial "shadowing effect" at the
rn '.-section on stress concentrations, this is not the case in
1, gh bearing-to-bypass multi-fastener joints where bearing or
shearout failures prevail. The proximity of adjacent fastener
holes acts like an edge-of-part to internal fasterner holes,
reducing bearing strength capability and overall laminate strain
capability.
Single fastener (pure bearing) joints were tested in single-

shear at RTD for edge distance/diameter values ranging from 1.5
to 3. Effects of torque-up (30.0 increased to 50.0 inch-pounds)
are reflected in a 25 KSI increase in bearing strength (Figure
49) for the e/d ratio of 3. The stacking sequence used in these
tests was identical to the number 6 stacking sequences of Task 2
(Figure 26). The increased strength of the more homogeneous
stacking sequence (number 1) is also indicated in Figure 49.
(d) Width - Effects of width on joint strength were

evaluated using the 50/40/10 laminate in a two-fastener-in-
tandem double-shear configuration. Specimens were tested to
failure in tension at RTD and ETW. Results are presented in
Figures 50 and 51. For RTD tests, specimens failed in the net
section at the lower w/d ratios of 3 and 4, and failed in bearing-
shearout at the w/d ratios of 6 and 8. Effects of temperature
(250 0 F) and moisture content reduced strengths markedly at the
higher w/d ratios but not at the lower w/d ratio of 4. This
decreased sensitivity reflects a transition to fiber dominant net-
section failures which are insensitive to effects of moisture.
Failures for RTD specimens occurred in bearing-shearout at w/d
ratios of 6 and 8 and transitioned to net-section at w/d ratios
of 4 and 5. Decreased shear and compression properties at ETW
conditions caused bearing-shearout failures at all but the lowest
w/d ratios. Typical failed specimens are illustrated in
Figure 52.
(e) Hole Siz2 - Effects of hole size on joint

strength were evaluated using the two-fastener-in-tandem double-
shear specimen. Specimens were loaded in tension to failure at
room temperature. Strengths are illustrated in Figure 53.
Bearing strength and gross joint failure strain trends varied
linearly over the range of hole sizes tested; however, all torque-
up values for these tests were maintained at 50 inch-pounds. The
triangular •yrrbol in Figure 53 indicates the increased strength
of joints with 160 inch-pounds of torque-up. This indicates a
reduced sensiitivity to effects of hole-size on joint strength
when fastener torque-up values are increased. Predicted joint
strengths for a nontorqued configuration are also presented in
Figure 53. All specimens failed in bearing-shearout, regardless
of hole size.
57
k-s--s--i d-=0.25
P1 +P2 --- P2 - 1 .--4!

i
50 in.-Ib To'que-Up
Double-Shear
180
0,U • • ~ S= .75--• _•• •
SS= 0.50
.' 100
. . . . - Prediction
t601
* A n Ultimate
N
O A 03 Initial nonlinearity
50/40/10 layup
t 3,000 ___"-
2,000
2 3 4 5
Edge Distance/Diameter - e/d
Figure 48. Effect of Fastener Spacing on Joint Tensile Strength
58
LmLmmmmm
d 0.375-
(I50/40/10 layup
- I .
Baseline stacking sequence (no. 1)
Ct A@ Stacking sequence no. 6
A
X 3d
160
4L
120
nx
LL
Ci)
.E 80
40'•
3,000
C • Double-Shear
"• 2,000
nx Single
1,000
1 2 3 4 5
Edge Distance/Diameter - e/d OPI"115-37
Figure 49. Effect of Edge Distance on Pure Bearing Strength
59
'St
S--I_
d 0.250-
~Tota --e - 6d
Double-Shear
50 in.-lb Torque-Up 4d 3d
S~160 -
L~i Net Section

S~~~Failure Mode--\ = z
_Bearing/Shearou:
120 Failure Mode
120
U o
•J * Ultimate strcngth
o Initial n~onlinearity
40 10
Width/Diameter, w/d a•.~-
Figure 50. Effect of Specimen Width on RTD Joint Strength
(nto
tm 1100
60
Pi +P e I d- ,5
--we- 6d
50 in.-lb Torque-Up 4d 3d
160 ,---RTD Bearing
F I ro N et RT Nt . l " " Failure
Shearout
[
Section Failure
v4'N 120
.0 x
--*• -.. 1
S.......... .
-- - ETW
4C
6,000
5,000 . .. .. •
'. 4,000 --
_____D
3,000
0
SETW
2,000
1
0 A Ul'imate
A Initial nonlinearity ETW
__II__I_
IOOC,
0 2 4 6 8 10
Width/Diameter, w/d
OPt3-115-35
Figure 51. Effect of Temperature and Moisture on Joint Strength
61
rI
iI
/
Ia) Net-Section
at RTD
(b)
Bearing-Shearout
Fig(5re at
2. Joint ETW
Failures
Changed
at ETW
Test Conditions
I.
62{
I2
P1 + P2 -'- we 6d Double-hear
P2 P1 j Ultimate failure
tu uIII J 0 Initial nonlinearity

Ex
4d 3d
160
' ", 1'

12o Torqe-U
60 in.-Ib Torque-Up
120
C4
4--
0.0 in.lb Torque Up
40
6,000
O. 125 1 9.375 05 6 5
80, -df
4,000
-x
2,000
00 0.125 0.250 0.375 in.0500 0.6250,750
Figure 53. Effect of Hole Size on Joint Strength
63
i I
(f) Load Interaction - Effects of fastener bearing
and bypass stress interaction on laminate strength were evaluated
by two load conditions; (1) bearing and bypass loadings aligned
parallel to the 0' principal material axis, and (2) bypass load-
ings with bearing loads at angles of 100, 22.50 and 450 off the
00 principal material axis. Roth studies used a specially
designed, hydraulic-actuated, scissor mechanism (Figure 54) to
permit bearing loads to be applied to fastener holes independent
of bypass loads. In all interaction tests, hearing loads were
applied and maintained while typass loading was increased until
specimen failure occurred.
II
CIPI13-15.33
Figure 54. Load Interaction Specimen and Test Setup
The first evaluation provided preliminary data on the

effects of aligned bearing and bypass stress on laminate strength
for a 50/40/10 and 30/60/10 layup. Unloaded hole (zero bearing
stress) and pure bearing (zero bypass stress) data were included
64
, •
in the study. Predicted strengths are correlated with test data
in Figures 55 and 56. Predictions made using the BJSFM are based
on the maximum strain failure criterion applied at a characteris-
tic dimension of 0.02 inches away from the hole boundary. Solid
symbols indicate occurrence of ultimate static failure, open
symbols indicate initial nonlinear or discontinuous load-versus-
deflection behavior, if any. For the 50/40/10 layup, the triangu-
lar symbnl represents strength data obtained from an unfilled
single-hole test specimen. This lower strength value indicated
the presence of fastener hole interaction ("hole-shadowing") in
the bearing-bypass specimen (Figure 40) which tended to reduce
stress concentration effects by approximately 10 percent
H ~~80 7-n-- ____________ __
1
Pbypass Pbolt bypass
6o' -.- o- -- )6
60 0 y
45 plies @ OCr T:40 '
S40
0
00 plies @ Ocr 85-9
20
SUltimate strength data
0 Initial nonlinearity
I0I
- Single hole test specimen
0 20 40 60 80 100 120
Bearing Stress - ksi GP13-Ol15-114
Figure 55. Correlation of Bearing-Bypass Strength of a 50140/10 Layup
65
1..
80
Pbypatl • - Pbolt- • bypass
60 rii 204000 Ply 80
4U
00 plies @ cr
U e ale
Figu strengthf e 7
0 Isnitial non-linearitya
I . a
0 20 40 60 80 10O0 120
Bearing Stress - ksi OpIW.0IS-113 i
Figure 56. Correlation of Bearing-Bypass Strength of a 30/60110 Layup
Typical load-versus-de rlection trends are illustrated in

Figure 57. General trends for both layups were: (1) bypass
strength decreased as bearing stresses increased, and (2) initial
non-linear or discontinuous load-versus-deflection behavior was
present, appeared earlier, and became m-ore marked as bearing
stresses increased relative to bypass stress. Predicted failure
at low bearing stress was net-section fiber failure of (0 plies;
at high bearing stress, failure was predicted to occur in either
+45' oi- 900 plies due to ply shear stress. Figures 58 and 59
indicate graphic representations of predicted critical plies and
failure locations. Representative failed specimens in Figure 60
indicate verification of predicted failure modes.
66
32 __
50/40/10 LAMINATE
I INT
28 PTOTALP__ 7 PRBYPASS
24 -- 7 1,__ ___
UNLOADED HOLE
20 -- -____ - -
PTOTAL K---
1000 LB LOADED HOLE
+ BYPASS LOAD
12 -
I•NDICATOR OF FIRST
PLY FAILURE
PURE BEARING ___

8
_2% OFFSET
0 e "- 'ozo 0.0
0
0 C,02 0.04 0.06 0.08 0.10
Figure 57. Effect of Bearing Loads on Specimen Load-Deflection Prhavior
67
00 PLY FIBER CRITICAL -450 PLY SHEAR CRITICAL
3-
/ 2R
1 , I _.•I,,/
,- - /
//
NET.SFCTION FAILURE (LOW BEARING)
FBY-PASS FBy.PASS
,P13-01 I S103
Figure 58. Predicted Critical Plies for Net.Section Failures
68
' - - : " - • • ::..• o::.,:. " -7 -7 =' k.-•

" " "•• • •' i -:-.- :
00 PLY SHEAR CRITICAL -45 PLY SHEAR CRITCAL
IIle
2R
/10
/,/
SHEAROUT B
BEARING FAILURE (HIGH BEARING)
PBOLT
-FBY-PASS
FBY-PASS
OP13-0115-112
Figure 59. Predicted Critical Plies 'or Bearine-Shearout Fai!ures
69
I.
F , -
Low Bearirvi toads Highi Bearing Loads
OP13-116-111
Figure 60. Tests Verify Predicted Failure Initiation Points
The second evaluation examined effects of increasing off-

axis bearing stress on laminate bypass strength. For the 50/40/
10 laminate, bearing stresses of approximately 25, 27, and 35 KSI
were applied to a .250 inch diaireter fastener hole at off-axis
angles of 100, 22.50, and 450 respectively. Bypass loading para-
l.el to the 0' principal material axis was increased from 0.0
pounds until static failure occurred. The resultant bearing
loads, and angles of application, were selected to maintain a
constant 25 KSI bearing stress component (fxbr) parallel to the
bypass loading, while perpendicular bearing stress components
(f br) increased to values of 4, 10 and 25 KSI
y Io
Theory and test data are correlated in Figure 62. For these
bearing stresses, all failures were in a net-section mode.
70
fbr
+ 100
~ ~ ~ 1 1 by pass
fbr 25.4 ksi go° Ply Reference

0
"j d ""
fbypasI fbr f0
- fbr
fbr 0
fbr 27.1 ksi

fo =-
0 fbr fbr
_.........
br (deg) X y
f &< 450 (ksi) (ksi)
10.0 25 4.3
IN.'I, __fbypass22.5 25 10.4
1-y 45.0 25 25.0
fbbr = 35.4 ksi

fGp 1-0.31
Figure 61. Load Conditions fof Off-Axis Load Interaction Tests
71 j
L*-
r!
P]
P Double-shear
-Ft Ultimate failure
- Initial nonlinearity
70
*~60
"U. 50
Prediction J
40 _ - --- - --
E 6000
Zw 5000
"• 4000
Prediction
300
0 5 1015 20 25 30
Fbr - ksi Y 1
OP13&0115.32
Figure 62. Effect of Off-Axir Bearing Loads on Bypass Strength
For both load-interaction studies, at increased bearing

loads, ultimate s ngths occurred .Eter initial nonlinearities
were observed ill joint load-deflection plots. Predicted
strengths, based on the linear-elastic BJSFM analysis, correlated
with the initial nonlinear indications. High bypass loads with
zero to low bearing caused joint net-section failures with plies
predicted critical due to excessive fiber strain. Joint load-
deflection behavior was linear with catastrophic failure occur-
ring at predicted first ply fiber rupture. At higher bearing
stresses, plies were predicted critical due to excessire shear
stress, a typically nonlinear ply property. This correlated with
observed joint failures which were in bearing or bearing-shearout
with nonlinear load-deflection behavior. Failures were localized
at the fastener hole with load redistribution occurring prior to
failure.
72
(g) Fastener Pattern - Stress concentration inter-
actions were evaluated using a four-fastener specimen (reference
Figure 41). Edge distances and fastener spacings were 3 and 4
diameters respectively. Column spacing was 4 diameters. All
fasteners were .250 inch in diameter and specimens were tested to
failure at RTD and ETW conditions. Steel, aluminum, and titanium
load blocks for the four-fastener specimens were designed to
provide joint stiffness ratios between the metal load blocks and
the Gr/Ep specimen of approximately 10, 5, and 1 respectively.
Test setups are illustrated in Figure 63.
(a) Single-Lap (b) Double-Lap
Steel Load Block /Aluminum

Straps
Graphite/Epoxy Specimen
Graphite/Epoxy: Ti Stiffness (1:5)

Graphite/Epoxy
Specimen Titanium
Graphite/Epoxy: Ti Stiffness (1:10) Straps
-Graphite/Epoxy Spesimen
Graphite/Epoxy: Ti Stiffness (1:1)

OP13.011-30
Figure 63. Fastener Pattern Specimen Loading Configurations
73
Tension strengths are presented in Figure 64. No signifi-
cant variations in internal load distributions due to variations
in joint stiffness were evident. Predicted bolt-loads indicated
that the graphite-to-aluminum joint strength should be approxi-
mately 4% higher than the graphite-to-titanium joint; however, a
slight reduction (6-8%) was noted in test. Based on data from
single and two-fastener specimens, small variations in internal
bolt-load distribution are probably masked by joint nonlineari-
ties which occur prior to failure. Single-shear joint strengths
were similar to the double-shear strengths at RTD, and lower
(13-20%) than double-shear strength at ETW. Single-shear joint
failures were predominantly shearout. For double-shear joints,
failures tended to shift to net-section tension or tension-
cleavage as is illustrated in Figure 65. Strengths and failure
modes of the single-shear joints reflect joint eccentricity
effects. Because bolt-bending was minimal, a physical variable
which could account for strength and mode differences is the
double lap arrangement of the metal loading members. This
arrangement inhibits initiation of localized damage at the
fastener hole due to bearing or shearout and creates additional
friction load paths.
PP P2-G P
RTD ETW
Layup Configuration Fbru etu Fbru etu

x2 XGross X2 XGros
(ksi) (uin./in.) (ksi) (pin./in.)
30/60/10 Steel Load Block 123 5,470 110 5,110

A 90 4,700
50/40/10 Steel Load Block 130 4,350 101 3,380
50/40/10 Graphite/Epoxy-Ti 133 4,350 112 3,880

50/40/10 Graphite/Epoxy-Al 139 4,510 122 4,220
50/40/10 Graphite/Epoxy-Al - -- 123 4,580
(Salt Exposure)
A\ T300/5208; all others AS/3501-6 OP13O011-2t
Figure 64. Effects of Environment on Faatoner Pattern Joint Strength
74
OP13-0115-85
Figure 65. Joint Failures for Fastener Pattern Specimena

A limited examination of the galvanic effect of salt-water
exposure on graphite-to-aluminum joints was performed with the
foar-fastener specimen. Specimens were attached to 7075-T76
aluminun plates and exposed to a 5% NaCl salt spray at 95*F.
Setup was exposed for 34 consecutive days. Visual appearance
after the exposure is indicated in Figure 66. As expected, the
exposure resulted in corrosion of the aluminum parts. Because
composite member strength was the issue of question, specimens
were then attached to pristine aluminum members in a double-lap
configuration and tested to static failure at 250'F under tension
loadinqs. As indicated by strength results shown in Figure 64,
composite member strength was unaffected by the salt spray
exposure.
Figure 66. Corrosion Resulting from Salt-Spray Exposure
Two additional sets of specimens were fabricated from

T300/5208 graphite-epoxy in the 50/40/10 and 30/60/10 laminate
configurations. These specimens were tested to failure in
tension at room temperature. Relative to the AS/3501-6 data,
failure modes were similar, but strength values were lower by
approximately 14-23 percent (Figure 64).
(h) Torque-Up - Effects of torque-up on bearing

strength were evaluated for three layups. In addition, pure
bearing (single fastener) and two-fastener-in-tandem (bearing
plus bypass) double shear specimens were tested with torque-up
values ranging from 0 to 70 inch-poundL.
76
Pure bearinq strengths and associated failure strains are
compared in Figure 67. All layups failed in shearout-bearing.
Representative failures are illustrated in Figure 68. Layup
sensitivity to torque-up appeared to be related to percentage of
0* plies (or lack of + 450 plies). For the three layups of 30,
50, and 70 percent 05' plies, respective increases in bearing
strength at 50 inch-pounds of torque-up over pin conditions were
approximately 15, 21, and 30 percent.
Pure bearing and unloaded hole failure strain data are

compared in Figure 69. Higher percentages of 0* plies cause
increased strain concentrations which lower laminate gross strain
capability. Resulting unloaded hole gross strain concentration
factors (Kt's) range from 2.44 for the 70/20/10 layup, to 2.1 for
the 30/60/10 layp, with pure bearing Kt's respectively of 10.2
and 3.5 for fasteners baving 50 inch-p.ounds torque-up. Removing
torque-up entirely increased the effective Kt under pure bearing
to 13.6 and 4.1 respectively.
In acdition to bearing strength, joint load-deflection plots

were examined for nonlinear mechanical response. Points of ini-
tial nonlinearities are plotted with ultimate bearing strengths
in. Figures 70, 71 and 72. Although scatter bands are wider,
unlike ultimate strength data, the load values at initial non-
linearities appear to be less affected by torque-up. This
correlation of predicted first ply shear failure indicates that
joint nonlinearities are associated with localized failure of
plies at the hole boundary. For this calculation, the previously
determined <'alue of .02 inch was used as the characteristic dimen-
sion in conjunction with the maximum strain failure criterion.
Strengths of the two-fastener-in-tandem specimens are compared in
Figure 73 with pure bearing strengths.
(i) Fastener Countersink and Laminate Thickness -

Effects of countersinks on laminate bearing strength were
evaluated over a range of countersink denth-to-laminate thickness
ratios (.77, .38, and .26). Test specimens were of a single-
fastener pure bearing configuration having a .375 inch diameter
fastener with the laminate loaded in double-shear. As-manufac-
tured speci mens were tested in tension to failure at room
temperature. Layup stackings, detailed in Figure 26 were integer
multiples of the 50/40/10 baseline 20 ply stacking sequence.
These same 40 ply and 60 ply layups were further used to evaluate
the effects of thickness without countersirks. The do'.ble-shear
specimen configuration was used and specimens weie loaded to
failure in tension and compression. Test conditions included RTW
and ETW. RTD strengths are presented in Figure 74. Within data
scatter, effects of countersink versus noncountersink on strength
appear to be insignificant for the 50/40/10 laminate. However,
initial points of nonlinearity indicate a 50% lower onset for
countersink 20 ply laminates, becoming less pronounced at the 40
ply thickness and disaFpearing at the 60 ply thickness. These
77
L "
2 20 It
Double-Shear Ex d -00250
.2
LL
4,000
Ci
c - ' • "='.50/40/10
60
0 20 40 60 80 100
i Torque-Up - in.-Ih ~ G PI 3-115-2 5
Figure 67. Effects of Layup and Torque-Up on RTD Pure Bearing Strength
78
70/20/10 Layup 50/40/10 Layup
-i 0.250 -
• --,-- 1.5
30/60/10 Layup
GP13-011547
Figure 68. Failures of Pure Hearing Specirnans
79
14,000
13,000
12,000 " _"
Unnotched
Unidirectional
11,000 Tensile Strain
Capability
10,000
9,000
Ct
X /-d 0.250
*- 8,000
x 6d I
6Unloaded Hole-
F 6,000
5,000 .---
4,000 Torque-Up
0in.-Ib--•.II
Loaded Hole 5 n.b- ___,,__
3,000
6d
Double-Shear
(All layups contain 10 percent 900 plies) j
0 10 20 30 40 50 60 70 80 90 100
Percent of ±450 Plies
GP13-0115677
Figure 69. Comparison of Loaded and Unloaded Hole Specimen SAengths
80
Double-shear
U Ultimate failure
3d
Q3 Initial nonlinearity
200 --
120
-Prpcixtion
80 _
40 -- _
20 40 60 80 100
Torque-Up - in..lb GP13-o115.27
Figure 70. Effect of Torque-Up on RTD Pure Bearing Strength

50/40/10 Layup
81
200 F l T
Diuble-shear
( 0.250 Ult mate failure
_T Initial nonlinleirity
P - 6d Predicton
P
160
3d
120
-x
U_
- Prediction
40__ _ _jz
00 20' 40 60 80 1 GO
Toi1ue-Up - in.-l
OP13-VI15.26
Figure 71. Effect of Torque-Up on RTD Pure Bearing Strength

70/20/10 Lay up
82
(I 0.250p\
P 6d
Double-shear
Ultimate failure
r3 Initial nonlinearity
3d
200
160
120
.L -BJSFM
80 ,. •,..Prediction
-1
40
0 -
0 20 4J 60 80 100
Torque-Up- in.-Ib GPIS11Z
Figure 72., Effect of Torque-Up on RTD Pure Bearing Strength

30/60/10 Layup
P,3
-I
d= 0.250- d -0 .250
PT
PT 6dP _G 6d
, •0.25_..
0 Double-Shear _ d_,
Sex 3 ext 4d 3d
180
'- Single Fastener

Joint
2 24 asteners-
12
Sin-a-Row Joint
i 6,000
-4--
C-
2x
2,000r-= J-
0
0 20 40 60 80 100
Torque-Up - in.-lb
041340116-2
Figure 73. Comnpalds of Effects of Torque-Up on RTD Joint Stmngth
84
L ,t
trends indicate the expected effect of local countersink geometry
on linear-elastic stress distributions, with differences vanish-
ing due to subsequent joint nonlinear behavior prior to specimen
failure.
d- 0.375-
pp 6d
Double-Shear 3d
200
* U Ultimate failure
Initial nonlinear~ty
S150-
M -
TNon-CNk-
CE 50
1c8 ,,
40,
!~
1 v0
00
C.
2,400
2,200-
Qk-- IT"
i
2,000 f,
Thcnesin.t 0P3152
1,4oo01
0 .01 0.2 0.3 0.4 0.5 0.6 0.?
Thickness - in.
Figure 74. Effect of Thickness and Countersink on Joint Tensile Strength
85
- .
(j) Stacking Sequence - Five variations of the
50/40/10 layup w:'re tested in pure bearing. Spe imens were
loaded to failure at room temperature in tension and compression.
Effects of moisture content (approximately .8% by' w-ight) and
elevated temperatura (250*F) on bearing strength were evaluated
for tensile loaing. The 50/40/10 layup stacking sequences are
detailed in Figure 26. The primary variation in stacicing
sequence was separation of 0* plies.
Tension strengths are presented in Figure 75. Lowest

strength values are associated with layups having 60% and 100% of
their 0O plies grouped. This type of stacking increases intra-
laminar shear stresses between each group of 0° plies and the
remaining +45* or 90* plies. The dominance of this factor is
illustrated in Figures 76 and 77. Both tension RTD and ETW
strength trends are presented relative to the maximum percent of
grouped 0* plies contained in the 50/40/10 layup. Data trends
indicate a linear dependence on 0* ply groupings. Predicted
bearing strengths are shown in both p'..ots as a constant value.
Effects of through-the-thickness variables on stress distribu-
tions are not accounted for in current BJSFM methodology.
For RTD ccmpression data, strengths are plotted versus maxi-
mum percent of grouped 0* plies in Figure 78. While compression
strength data indicated a linear relation, fastener bearing
stresses are quite different t an stresses in the previous
tensile specimens. While groupings of 0* plies do affect joint
compression strength, it would not appear to be through a
presence or lack of intralaminar shear planes.
(,k) Single-Shear Loading - A comparison of joint
single shear versus double-shear strength is presented in Figure
79. Initial pcints of nonlinearity indicate the expected
decrease in joint strength associated with increased joint eccen-
tricity. However, laminate bearing strength at joint failure
exceeded bolt shear strength capability in the .250 inch
protruding-head single-shear configuration, and composite member
failure trends could not be quantified. Single-shear joint
strength versus variations in e/w ratios is plotted in Figure 80.
86
200
b~u 3&~1 6d
d 0.250 -
150
z x
uL
1T.
CE
50
•RTD • nUltimate
[0 ETW 0 E3 Initial nonlinearitv
2,600
T
2.200
.C
C I
U.
. 1,400
1,000
600
1 6 7
"acking Sequence Number
50/40/10 Layup
Figure 75. Effect of Stacking Sequence on Joint Tensile Strength
87
200 1
.. m4Ultimatq strength -*-025;-\
- ()initial non-linearityT
16 ___6d -
1050 in..bTorqu.e I.
Double-Shear 3d
4 120 -
LL
411111111
80
13JSF M
Prediction
.40
2,000 -- -__
0 04 60 f 80 10
I,,
88
d =0.250
-- 6d 50 in.-lb Torque-Up
(.• .. N Double-Shear
7- Ultimate failure
33d 0 Initial nonlinearity
120
100
80
LL 60 ,i.. i.---.-•,,-- -,
201
2,000
1,500 1 - ill
; 1,003
500 --- Stacking

Sequence
Number 1 8 9 6 7
01
20 40 60 80 100
0
Maximum Percent of 00 Plies Groupetd
aPIS-el11-U|.
Figure 77. Correlallion of Joint ETW Tensile Streng!h with Grouped 00 Plies
39
iI
(1 0.250
i=7I 50 n..lb Tc;que-Up
Double-Shaar
vX" 1-40 Ultimate Failure

Initial Nonlinearity
160 -'
150 --- - - -- _ _ _ _ _ -. _ _ _ _ _ _
140 - -
130 - - - -
___ _ _ -_U ___
120 • •• •--
2 1Prediction
.U -x go o Ply
_ Fiber Failure -- ____ __
Prediction
10 450 Plies - - --- ___
Shear Failure -"-
801 1__ _ _ _ _ _1
3,000 f--------
2,600
W
x 1,000 -Stacking - t _
Sequence
Nu0e 1 1
0 20 40 60 80 100
Maximum "'Nicent of Grouped 00 Flies
Figure 78. Correlation of Joint RTD Compressive Strength with Grouped 00 Plies
90
Clevis
Double- Shear
Ultimate
f@0 Csk Single-Shear
Ultimate
fA Csk
160 0 0.. 0 A Initial Nonlinearity
140
120
' I
C" 100 - _ __ _
8012
601 __ _ _ _ _ _ _
40
0.125 0.250 0.375 0.500
Diameter - in.gplal111
Figure 79. Comparldsc mof Single and Double-Shear Joint Strengths
91
18
o.
Ultimate failure
I
'.--49-
RTC)
-
1Y50/40/1o
0
ETW
-
Single-shear
Noncoutitersink
LA .UP
Initial nonlinearity - C -- , -,. 50 in.-lb torque-up
140
S100 - _ _
.-A-
60
20-
3,000 -
d0., 0.25
2,500 P
I!
P
2,50 Clevis
2,000
t--
'U.X
1,500 - _ __
1,000
e/d = 2 e/d 3 e/d= 3

5001 1
0 0.2 0.4 0.6 0.8 '.0
eWPt
Figure 80. Correlatiore of J,3lnt Strength with

Edge Distance-to-Width Ratios
92
2. EVAIIIATION OFMANUJFACTUIRING; AN!DSERVICE ANOMALIES -TASK 3
in thIS task te serviceK~
anomal ies Oil joint static strength were evaluated. The? selection
of the seven bolted joint anowF~lies was based on MCAIR experience
and results of an industry-wide survey conducted as part of the
"Advanced C,_omp)osite Serviceability Program" (Reference 2,. Data
fromn Task 3 weýre compared with baseline joint strengths (no
anomaly) determined from tests performed in Task 2.
a. Task 3 Test Plan - The objective of Taskz 3 was to evalu-

ate the effects of manufacturing and service anomalies on bolted
comrposite- Joint static strength. Only laminate quality or hole
quality was varied for Task 3 testing. Test Matrix details are
providedI in 1'iqure 81.
Number of Tests
Per Environment Total
Anomaly rSpecimen
RT (Dry) IRT (Wet) ET (Wet) Tests
olesTensio -n Compressior Compression
1. Ou-of-ound
"'I" Laminate (50/40/10) A - ~ -4
"2' Lamrinate (3 0/6 0/10) 4 - 4

2. Broken) Fibers on Exit Side of Hole
Severe Delamination 4 4 4 12
Moderate De~amination 4 4 4 12
3. Porosity arouind hole
Severe Porosity 4 2,2 4 12
Moderate Porosity - 2,2 J~4 8
4. Improper Fastener Seating Depth

80% of Thickness 4 -- 4
100%) of Thickness 4 -- 4
5. Tilted Countersinks ____ ___ _____
Away from Bearing Surface 4 -- 4 8

Toward Beeý,ng Surface448
6. Interference LayUP 1 4 -4 8
Fit Tolerances 1 4 4 8
0.003 in. Interference 2 4 -- 4 8
0.008 in. Interference 2 4 -4 8
7. Fastener Removal and Reir-stallation
100 Cycles 4 4
Totdl 116
A GOP13-115-110
zi\A fter f reeze' xhirs c,- Rit /2\Tenvsion tests
rigure 8". Task 3 - Evaluation of Manufacturing Anomalies-Test Matrix
93
ICi
Test procedures and specimen configurations (Figure 37) used
in this task were consistent with those establi3hed in Task 2 to
insure correlatable results. Specimens were tested to failure in
tension and compression at three environmental conditions: room
temperature dry (RTD), room temperature wet. (RTW), and elevated
temperature wet (ETW). ETW tests were conducted at 250"F with
specimen moisture contents of approximately .80 percent by
weight. Hercules AS/3501-6 graphite-epoxy was used as the mate-
rial for fabrication of test specimens. All specimens were
tested to failure according to Figure 81 under tensile or compres-
sive loadings. Absorption profiles representative of specimens
requiring moisture preconditioning are i-lustrated in Figure 82.
All specimens were instrumented with strain gauges. Docu-

mentation included: (I) failure load and failure stra in,
(2) load-strain plots to failure, (3) thickness, width and hole
size measurements of all specimens, (4) resin content of each
panel, (5) weight gain of humidity exposure specimens, (5) repre-
sentative photographs, and (6) selected load-deflection plots to
failure. Complete details of Task 3 testing, test procedures,
data and support equipment are provided in Volume 2.
b. Experimental Results and Evaluation - Results from all

T'Vsk 3 tests are summarized in Figure 83. Indicated percentages
of increase or decreased strength are based on a comparison of
gross failure strains with baseline strength data from Task 2.
Results can be grouped under two classifications: (1)

effects of anomalies which resulted in strength reductions of
more than 13 percent and 12) those which resulted in strenqith
L edtc L ions Less kII1II11 1,ret Tho fis .tol. n l..o
porosity around hole, improper fastener seating depth, and tilted
countersinks. Current project inspection and acceptance criteria
in the industry would have detected all three of these anomalies
and resulted in part rejection or required repair. Detailed
results are discussed in the following paragraphs.
(1) Out of Round doles - Effects of out-of-round holes on

joint strength were evaluated by testing specimens produced by
drilling offset (.004 inch) holes as shown in Figure 84. The
major diameter of the out-of-round hole was fabricated perpendi-
cular to the 00 fibers and the direction of loading.
Joint specimens from two laminates (V 1/40/10 and 30/60/10)

were tested to failure, in tension at roum temperature. Test
results were compared with baseline data from Task 2. Joint
strengths for both layups (Figure 85) indicate littLe or no
sensitivity to out-of- round holes. This is consistent with
unloaded hole studies where an out-of-round value of .004 irch
(modeled as an ellipt ½c hole) resalted in predicted strength
reductions of no more than 1.0 percent.
94
€I
1.2
1.0 -__
0.8
C
CU 0.4
0.2
98% RH
R +8 95% 9H I _I 55% RH
16 0 °F I 160 0 c '" I 8 •I 18 0 °F
2.4
95% RH
_-
RH
,,
(a) For Baaeline Specimens
98% RH55%
__
__
RH
1600F 180OF 180OF

18
2.0
. 1.6 ]- --
. 1.2 ---
w08
A
0.4 derv" delamination
04 n vere porosity
C severe delaminition
0
0 25 so 75 100 125 150
Days
(b) For DelanrAnated or Porous Specimens oPsI11-O1i6t
Figure 82. Specimen Moisture Conditioning Histories
95
• - • • .-. . .• ,
RrD COMPRESSION
TENSION RT 25&F
OUT-O, ROUND HOLES
* 50/40/10 LAMINATE 0 - -
* 30/60/1(0 LAMINATE -4.8 - -
BROKEN FIBERS EXIT SIDE OF HOLE
e SEVERE -9.5 -3.t -12.2
* MODERATE -4.9 -4.2 _
POROSITY AROUND HOLE

* SEVERE * -12.1 -32.6
& SEVERE WITH FREEZE-THAW -13.3 -
e MODERATE -5.4 -17.9
e MODERATE WITH FREEZE-THAW - -7.9 -
IMPROPER FASTENER SEATING DEPTH
* 80% THICKNESS -23.2 - -
* 100% THICKNESS -56.9 - -
TILTED COUNTERSINKS
e AWAY FROM BEARING SURFACE - -20.0
e TOWARD BEARING SURFACE -23.9- -22.7
INTERFERENCE FIT TOLERANCES (INCH) &
* 50/40110 @ 0.003 * - +14.7
@0.008 * - +11.2
* 30/60/10 0 0.003 * - +2.4
FASTENER REMOVAL AND REINSTALLATION

e 100CYCLES - -7.4
_
A0.86%moisture content & Tem;ile loading *Under 2% change - No test
Figure 63. Summary of Task 3 Strength Reduction Percents
96
L!
LT
O..
100-
two- 0.750--, I__
Note: Dimension in inche-, I 0Pt50115tS
Figure 84. Outof-Round Holes - Specimen Details
d =0.250
"I(
PTctal
P2 P1 6
- XGross
Task 2 Task 3 %
Test Condition Baseline
"u cData Anom-.aly Results Change
bU•. odn
Out-of -Round Holes Fbd C.u Loading
XGross x2 XGros Fbru eGross
_ (k•i (liin./in.) (ksi) (in./in.)
50/40/10 Layup • RTO 140 3,990 146 4,010 +4.3 +0.6 Ten
30/60/10 Layup - RTD 143 5,470 141 5,200 -1.3 -4.8 Ten
OPI4.01S-14
Figure 85. Effect of Out-of-Round Holes on Joint Strength
97
(2) Broken Fibers on Exit Side of Hole - Specimens were
tested for two conditions of severity, "moderate" delaminations
and "severe" delaminations. In order to obtain the various
degrees of delamination, drill procedures included: use of dull
bits without backup material, and improper drill and feed speeds.
Tests to failure were performed in tension and compression.
Compression tests were conducted at room temperature and elevated
temperature after specimen moisture conditioning.
Delaminations were detected visually .nd with ultrasonic

C-scan. Delaminations were defined as "moderate" when they
extended over 10-20 percent (2-4 plies deep) of the laminate
thickness on the exit side. Delaminations were defined as
"severe" when they extended over 20-30 percent (4-6 plies) of the
lamina.ie thickness on the exit side. Nondestructive C-scans of
holes with moderate and with severe delaminations are illustrated
in Fioure 86 and compared with a nominal condition.
1.0 0.75 1 1+ 1 1.00

0"
0.75
ExitSideDe mnatioExS D lamination

Ei Sides with' n the Last Exit Side within the Last
of Holes 1-20%o: of Holcs 20 - 30% of
Thickness Thickness
Specimen Number 6-21-3 Specimen Number 2-20-12 Specimen Number 2-20-2
I+ ' II'' l"...

ii~ !
*,.,.,.
***: .*.,., . •
.... ........ Pt I
T:++
T::::+.+
Il I!++
is ' '•t W::
i! +l ! i ;i
Itti:iitt(,, ,
Nominal Broken Fibers Broken Fibers

Condition Moderate Severe
Delamination Delamination
0p13-Ohlt.12
Figure 86. C-Scans of Laminates with Delaminations at Fastener Holes
98
Dry laminates at room temperature were tested to evaluate
joint strength in tension. Because of its sensitivity to environ-
ment, joint cormpression strengths were evaluated at RTW and ETW
test conditions. S:rengths are listed in Figure 87 along with
the baseline strength data from Task 2. Strength reductions of
2-5 percent and 7-10 percent occurred in RTD tension tests for
specimens with moderate and severe delaminations, respectively.
Compression 'strengths at 250°F indicated inc.:eased sensitivity
with reductions of about 12 percent for specimens with severe
delami~ations.
d 0.250-
Task 2 Task 3 %
Test Condition Baseline Data Anomaly Results Chanue
Hole Delaminations bru - u FbrU u Loading
(50/40/10 Layup) x2 XGross 2 Gross Fbru CGross
(ksi) (pin./in.) (ksi) (pin./in.)
Moderate:
RTD +140 +3,990 +138 +3,790 -2.0 -4.9 Ten
RTW -155 -4,740 -150 -4,450 -3.2 -6.2 Comp
ETW -120 -3,790 -115 -3,810 -3.7 +0.6 Comp
Severe:
RTD +140 +3,990 +130 +3,610 -7.3 -9.5 Ten
RTW -155 -4,740 -142
ETW -120 -3,790 -109 [ -4,580
-3,330
-8.4
-8.7
-3.5
-12.2
Comp
Comp
0PlS.411-la
Figure 67. Effect of Delamninatlons on Joint Strength
Effec-..s of exit side delamination are dependent on depth of

damage! and on stacking sequence. For uniaxial loadings, greater
strength reductions can be expected if delaminations interrupt
fibers parallel to loading direction. However, in highly ortho-
tropic layups, delaminations may reduce stress concentrations by
reducing the laminate stiffness around the fastener hole. For
this particular 50/40/10 stacking sequence (+45, 0, -45, 0, 90,
0, +45, 0, -45, 0)., "moderate" exit damage (of 2-4 plies)
affected 1-2 of the ten 0° plies, while "severe" exit damage (of
4-6 plies) affected 2-3 of the ten 0* plies. For stacking
sequences with more +450 plies towards the surface, considerable
differences in strength data would have occurred.
IL
.. . . ...... -
(3) Porosity - Two levels of porosity were evaluated;
"moderate" and "severe". Desired levels of porosity in the
50/40/10 laminate were obtained by ujing the altered collation
and curing procedures summarized in Figure 88. Specimens were
located within panels such that fastener holes occurred in areas
of desired porosity levels as indicated by photomicrograph and
nondestructive inspections (Figure 89). Specimens were tested
under "worst case" conditions. Under tension loadings, only the
effects of severe porosiuy on joint strength were evaluated at
room temperature. Under compressive loadings, effects of
moderate and severe porosity on joint strength were evaluated at
RTD, RTW, and ETW. Four specimens (two of moderate and two of
severe porosity) were loaded to failure in compression at room
temperature a•fter moisture conditioning and exposure to
freeza-thaw c•" 1 .n• (Figure 90).
Procass Used for ,. Proem Used to IProdue.

Curing Pro cedur e Good Panels Mod orat Po rosity S " Porosity
Vacuum Oebulk Yes None None•
intermediate Temperature Hold 1 hr @ 275°F None None

Bag Va-uum 0.05 in. Heading 0.8 in. Heading 1.5 in. Heading
Autoclave Pressure 100 psig 50 psig 50 psig
I
Moisture Induced & None Every 7th Ply Every Ply
OP1341101
L -lne mist sprayed betwmn plies
Figure 8o. Panel Fabrication Procedures Used to Produce Panel Porosity
100
5p
Spec No 3 22 4
4.1
(a) Moderate Porosity
100 C-Scan
0 Region
Spec No. 3-23-1
ib) Severe Porosity G -1I.1
Figure 89. Examples of Panel Porosity
101
160
140k
120 - C-ool Down from 180OF - 55% R• exposure
100
80---
- Room Temperature
LA.
o 40-/
S20 /
S0 --
S-20 1
_--40 --
-60 - _- -65 F
-80 _
-1I00 -- --- 30- 30 30 -
5--
-120 -1 Cyce
-140
Time in Minutes
FIgure 90. Freeze-Thaw Exposure Proffie
Strengths from all tests and baseline (no anomaly) strength

data are listed in Figure 0)l. Little sensitivity to severe
porosity was indicated under tensile loading. tinder compressive
loadings, strength reductions ranged from 4-18 percent for speci-
mens with moderafe porosity and 10-33 percent for specimens with
severe porosity. The greatest reductions occurred at the 250OF
test condition. Specimens with exposures to freeze-thaw cycling
indicated a slightly greater loss of strength than specimens not
similarly exposed; 3-4 percent added reduction and 1-2 percent
added reduction respectively for moderate and severe porosity
conditions. Specimens loaded in compression failed in bearing,
similar in appearance to specimens with no porosity.
(4) Improper Fastener Seating Depth - Effects of excessive

countersink depth on joint strengt'h were evaluated by testing
composite joint members having fasteiners seated too deeply in the
laminate. Two conditions of countersink depth were evaluated at
room temperature in tension.
102
I.
d 0.250-\
P~otalP 2 P1
lXross
Task 2 Task 3 %
Baseline Data Anomaly Results Change
Test Cnditio
Porosity bru u FbrU uoading
(50/40/10 Layup) Fx2 XGron x2 XGross Fbru I Gross
(kSi) (Win./in.) (ksi) (.in.!n.)
Moderate Porosity:
RTD +140 +3,990 ..-.. Ten
RTW -155 -4,740 -184 -4,480 -4.3 -54 Comp
RTW (F-T) --15 --4,740 -142 -4,370 --8.4 -7.9 Comp
ETW -120 -3,790 -104 -3,110 -12.8 -17.9 Comp
Severe Porosity:
RTD +140 43,990 +140 +3,940 0 -1.2 Ten
RTW -155 -4,740 --139 --4,170 -10.3 --12.1 Comp
RTW (F-T) -155 -4,740 -137 -4.110 -11.9 -13.3 Comp
ETW -120 -3,790 -83 -2,550 -30.3 -32.6 Comp
Note: (F-T) - ttxposed to freoze-thaw cyclas Opt3411tS4
FIgure 91. Effect of Porisity Around Hole on Joint Strength
Strengths are listed with nominal condition strengths in

Figure 92. Strengths for joints having improper seating depths
are compared with nominal countersunk (52% of laminate thickness)
strength data. For the 50/40/10 layup tested, joint strengths
for countersink vcr sus noncounters.ý nk larrinates (Task 2,
Reference Figure 74) indicated no significant reductions occurred
when fastener seating depth was nominal. The loss of bearing
material and the associated knife-edge condition for the 80% and
100% depths may account for the strength reduction. Based on
average bearing stresses, countersink depths of 80% and 100%
thickness reduce cylindrical bearing area by 28% and 48%
respectively, which approximates the percent changes in strength
listed in Figure 92. Additionally, the deeper countersink in-
creases the effective diameter of the hole which further reduces
laminate strength due to hole size effects.
103
J-
r\ !
d -0.25(j-
~Tot al'
XGross
Task 2 T-ak 3 %
Test Condition Baseline Data Anomaly Results Change
Fastener Seating FbrU u bru u Loading
(50/40/10 L.yup) Fx2 CXGross Fx CGross Fbru cGross
(ksi) ( pin./in.) (ksi) (C b rin./ino)
80% Depth - RTD 140 3,990 117 3,240 -20 -23 Ten
100% (Knifedge) - RTO 140 3,990 92 2,540 -53 -57 Ten
0P150115-?
Figure 92. Effect of Improper Fastener Seating on Joint Silrength
(5) Tilted Countersinks - Countersink perpendicularity was

investigated for two conditions of misalignment. The misaligned
countersink was tilted 100 away from the bearing surface for one
condition and tilted 100 toward it for the other. These tests
were conducted in tension &t RTD and in compression at 250*F
after specimen moisture conditioning. As illustrated in Figure
93, tilted countersinks cause interferen,:e prior to proper
fastener seating and create a gap under the fastener head reduc-
ing laminate bearing material. Experimental data indicate
approximately equal strength reductions (Figure 94), regardless
of direction and magnitude of countersink tilt, for moisture
conditioned specimens te~ted in compression at 250*F.
104
4.F
S - 0.20 Diame*,---te
Point of
+100 (Toward)
Interference-\ -100 (Away From)
0.108
-Lminate2 I Bearin_ _
S~~~~S u rf ac e ____'-
Note: Dimensions in inches 0P134. 1S4
Figure 93. 1 ilted Countersink - Specimen Configuration
105
__ _ _ _ _ _ _ __ _ _ _ _ __
__ _ _ _ _ _
d 0.250--T
_ _ _ _ _
XGtoss
LT,-k 2 Task 3 %
Test Condition Baseline Data Anomaly Reults Chanwe
Tilted Countersink Fbru M Fbru u Loading
(50/40/10 Layup) x2 XGross x2 XGross Fbru 'Gross
(ksi) (pJin./in.) (ksi) (/pin,/in.)
Away From Bearing

RTD +140 +3,990 +140 +3,910 +1.1 -2.J Ten
ETW 120 -3,750 -'100 -2,990 12.6 -20.2 Comp
Toward Bearing
RTD +-140 +3990 +110 •3,030 -18 -24 Ten
ETW -120 -3,750 100 -2,898 -14 -23 Comp
Figure 94. Effent of Tilted Countersink on Joint Strength
(6) Interfer."(ce Fit - The effects of fastener interference

fits on joint str.t',,jch were investigated in two different lami-
nates (50/40/10 and 30/60/10). Two-factener-in-tandem specimens
were tested to failure in tension at RTD and 1-TW conditions.
Strength evaluations were performed on specimens with .003 and
.008 inch levels of interference fit.
Results (Figure 95) for both layups indicate an insensitiv-

ity to interference at room temperature. Mixed results were
ohtairied at FTW. Joint strength of the more fiber-dominant
50/40/10 layup improved by 8-15% at both levels of interference.
Joint strengths of the matrix-dominant 30/60/10 layup showed
little or no increase (-1.8 to +4.6 percent) at the ETW test
corld it ioI.
106
d 0.250-\
PTetaI
!6d
P2 P16
( ;OSS K ±
Task 2 Task 3 %
Test Condition Baseline Data Anomaly Results Change
Tost Conditio
Interference Fit Fbru (U Fbru Cu G Loading
X2 XGross x2 XGross FbrU egress
(ksi) (pin./in.) (ksi) (pin./in.)
50/40/10 Layup:
0.003 Interference
RTD 140 3,990 140 4,000 +2.1 +0.4 Ten
ETW 110 3,080 120 3,530 +11.8 +14.7 Ten
0.008 Interference
RTD 140 3,990 140 4,030 +1.0 +0.9 Ten
ETW 110 3,080 120 3,420 ý8.1 +11.2 Ten
30/60/10 Layup:
RTD 140 5,470 140 5,460 +0.2 -0.1 1en

ETW 120 4,710 120 4,820 +4.6 +2.4 Ten
0.008 Interference
RTD 140 5,470 140 5,490 -4.6 +0.3 Ten
ETW 120 4,710 120 4,620 4-1.5 -1.8 Ten
OP13-O115.4
Figure 95. Effect of Fastener Interference Fit on Joint Strength
Laminate damage due to fastener installation at different

levels of interference fit were further evaluated. Fastfener
interference fits ranging from .002 to .0075 inch were examined.
Standard pull-through installation techniques and fastener types
were used. Each fastenrr area was sectioned and photomicrographs
taken. Representative results are shown in Figure 96. Photo-
graphs indicate little or no damage at a fastener interference of
.0035 inch or less; hov -2r, damage is indicated at the fastener
exit side as well as t rugh-the-thickness for interference fits
from .004 through .007 iich.
107
Pitu:tte At,,; ivv-
fntoiference Fits:
0.0010 in, 0.0034 in. .05n
IVI
Exit Side
Figure 96. Photomlcrographic Examination of Laminates with

Interteroaice Fit Holes
108
(7) Fastener Removal and Reinstallation - Specimens were
tested to failure in tension at RTD and in compression at ETW.
Fasteners were installed, torqued to 50 inch-pounds, and com-
pletely removed. This procedure was repeated 100 ti.mes prior to
strength testing. These tests were used to evaluate whether
repeated fastener torque-up associated with installation and re-
installation would locally damage the laminate hole area, and/or
affect joint strength. No attempt was made to simulate problems
encountered in project or field environments (e.g. part-to-part
misfit). While important, these problems require definition of
specific structural component and element details for proper eval-
uation and were beyond the scope of this program.
Visual appearance of all fastener holes after installation

and re-installation cycling was unchanged. Strength data listed
in Figure 97 indicates little sensitivity to RTD tensile test
conditions. Compressive strength values indicated an increased
sensitivity (7% reduction); however, test scatter for the compres-
sion tests was much larger. No significant effects on joint
strength were indicated.
Task 2 Task 3 %
Test Condition Baseline Dotta Anomaly Results Change
Fastener Reinstallation Fbru u FbrU u Loading
(50/40/10 Layup) XGross X2 XGross Fbru eGross
(ksi) (jin./in.) (ksi) (pin./in.)
100 Cycles
RTD -r140 +3,990 +140 +3,900 -1.3 -2.2 Ten
ETW -120 -3,790 -11,, -3,510 -9.3 -7.4 Comp
Figure 97. Effect of Fastener Removal and Reinstallation on Joint Strength
109
3. EVALUATION OF CRITICAL JOINT DESIGN VARIABLES ON FATIGUE
LIFE - TASK 4 - The objective of Task 4 was to evaluate the influ-
ence of selected design variables and manufacturing anomalies on
joint fatigue life. Seven variables or anomalies shown in Tasks
2 and 3 to have significant effect on static strength were chosen
for evaluation. Tests were to provide
performance, hole elongation, and failure data on
modes. joint fatigue life
A pure bearing test specimen was used to provide basic joint

fatigue data while minimizing test difficulties and permitting
correlation with Task 2 and Task 3 static test data. All tension-
tension (R = +0.1) and tension-compression (R = -1.0) constant
amplitude testing was performed at room temperature with speci-
mens in the as-manufactured condition. Spectrum fatigue testing
of the 50/40/10 layup was performed at three environmental condi-
tions. room temperature dry, room temperature wet, and elevated
temperature wet. Elevated test temperature (250°F) and the
moisture conditioning level (approximately .8 percent by weight)
were the same as in Tasks 2 and 3. Hercules AS/3501-6 graphite-
epoxy was used for fabrication of all Task 4 test specimens.
a. Task 4 Test Plan - The test plan summarized in Figure 98

was based on literature survey resulis which indicated that: (1)
little generic fatigue life data had been generated for 'olted
joints, (2) life prediction methods are essentially empirical,
and (3) theoretical approaches being pursued, such as the "wear-
out" model, require large residual strength data bases for
statistical evaluation. This data base would be impractical in
size for the seven variables to be evaluated in Task 4. Hence,
an empirical approach was adapted to determine the effects of the
selected variables on joint fatigue life. Testing of bcth fiber--
dominated laminates and matrix-dominated laminates was included
in order to identify differences in failure modes, with suffi-
cient tests under spectrum fatigue loading to indicate major data
trends.
The seven test variables selected for evaluation were:

(1) stacking sequence, (2) torque-up, (3) single-shear versus
double-shear, (4) fastener fit, (5) porosity, (6) layup, and
(7) geometry.
Emphasis was placed on generation of hole elongation data

and its relevancy to joint fatigue life. All con.;tant-amplitude
fatigue specimens were cycled t., failure, or 106 cycles, which-
ever occurred first. Constant amplitude fatigue testing was per-
formed at three stress levels for each specimen type. Selection
of the stress levels for fatigue testing was based on load-deflec-
tion data obtained from static tests. During fatigue testing,
load-deflection data were also obtained each time specified hole
elongations occurred. All constant amplitude tests were
performed at room temperature with as-manufactured specimens.
Selected specimens which did not fail in 106 cycles were tested
to determine residual strength.
110
Tests were performed to determine effects of spectrum load--
ing on joint life as a function of test limit load. If failure
had not occurred after four lifetimes of exp-sire, selected spe-
cimens were static tested to determine residual Strength. As in
constant amplitude testing, three test limit load levels were
selected for each test variable evaluated.
MAX NO. OF TESTS NO. OF TESTS

NO. TEST VAqIABLE FATIGUE CONSTANT AMPLITUDE SPECTRUM FATIGUE
STRESS R +0.1 R = -1.0 RTW RTD ETW RTW(TS)

1 BASELINE 01 3 3 3 3 3 3
50I40il0 u2 3 3 3 3 3
•3 333 3 333 3 3 33
____1 3 - ?• - 33 3 .
WSW -10 &&01 3- 3 .- 3 - -

02 3 - - -
03 3 3 -- 3 - -
19/76/5 _1 0,t_ 3 3 - 3 - -
023 - 3 - -
(03 3 - 3 - -
3. ....
2 STACKING SEOUENCE 01 3- - -
50/40/11C 02 3 3 - 3 - -
03 3 3 - 3 -
1&3 .. .
19:;5
; 3 3 -- 3 - -
02 3 3 - 3 - -
03 3 3 3 - -
3 TOROUE UP
T = 160 IN.-LB 01 3- - -
50/40/10/ 3 - 3 -
03 3 3 - 3 - -
_ .....
T = 160 IN.-LB 01 - - -
19/76/5 02 3- -
03 3 3 - 3 - -
L_ 3
SComplementing static tests d= 0.375 in., w/d= 6, e/d: 3 Torque up 0 in.-lb
Figure 98. Task 4 - Evaluation of Critical Joint Design

Parameters on Fatigue Life - Test Matrix
111
MAX NO. OF TESTS NO. OF TESTS
NO. TEST VARIABLE FATIGUE CONSTANT AMPLITUDE SPECTRUM FATIGUE
STRESS R - +0.1 R = -1.0 RTW RTD ETW RTW(TS)
4 I GEOMETRY
50/40/10 Z•al 3...
0 . ..
d -0.375 2 3-..
w/d -4 03 3 ... .
e/d -3 & 3
3..
19/76/5 a\1 -3 .. . ..
d - 0.375 02 3 .. . . .
w/d -4 03 3 .. . . .
e/d -3 & 3 .. . . ..
19/76/5 01 3 .. . . .
d - 0.375 02 3-.. . . .
wM -3 03 3 .. . .. .
e/d - 3 .. . ..
50/40/10 /•Ž 0 3 -.. . .
d - 0.375 02 3 .
w/d - 4 03 3 -.. . .
e/d =4 3 -.. . .
5 FASTENER FIT (0.003 - 0.008 INTERFERENCE)

50/40/10LA Li\ 01 3 .. . . .
02 3-.. ...
03 3 .. . ..
6 SINGLE SHEAR (PROTRUDING AND CSK)

50140110t2\ 01 3 .. . . .
T = 160 1N.-LB 02 3 .. . ..
02 3.. . .
d - 0.375 IN. CSK 01 3 .. . ..

wM -6 02 3-.. . ..
e/d 3 03 3 .. . ..
T = 160 IN.-LB 3A.. . . .
7 POROSITY
750/40/10 / 3 3 . . ..
02 3 3 . . ..
03 3 3 . . ..
TOTAL TESTS = 35k j

Complementing static tests A• d 0.375 in., w/d= 6, /'d 3 14\ Torque up 0 in,-lb
OP13.0116-262
Figure 98. (Continued) Task 4 . Evaluattion of Crit~cal Joint Design

Parameters on Fatigue Life-Test Matrix
112
L p...,
A baseline joint geometry (Figure 99) was tested to provide
a ftitigue life reference for all test conditions. Elevated
temperature exposure, moisture preconditioning, and thermal
spiking were included in this "baselir ." series of tests on the
50/40/10 layup. Three sets of the 50/40/10 specimens were
tested: (.) one set at room temperature with as-manufactured
moisture levels, RTD, (2) one moisture conditioned set at 2500F,
ETW, and (3) one moisture cond ioned set with thermal spikes
(TS) applied intermittently during -he humid aging process,
RTW(TS). The evaluation of a bass Ine joint geometry, also
included two additional layups (30/60/10 and 19/76/5), which were
tested under both constant amplitude and spectrum loading, RTD.
7.50
1.125-
1 00 STRAIN GAGE/
+__ 2,25
0.3745 +0.0022
-0.0000 1
] 1~
LJ..~CLEVIS(
SPEC IMEN-
LOADING CONFIGURATION
GP134115-107
Figure 99. Task 4 - Fatigue Specimen Configuration
The random load spectrum used in Task 4 testing was an "F-15

Measured-Mix Wing Spectrum-Truncated". This spectrum is a combi-
nation of three F-15 wing baseiine spectra (Air-to-Air, Air-to-
Ground, and Instrumentation and Navigation,, truncated to
eliminate all loads less than 55% design limit load. Truncation
was considered appropriate based on the relatively low failure
strains associated with static tests of pure bearing (single
fastener) joints. The distribution of hours and exceedances for
the Air-to-Air, Air-to-Ground, and Instrumentation and Navigation
in the Measured Mix spectrum are given in Figure 100. The
associated exceedances are illustrated in Figure 101.
113
MEASURED MIX
EXCEEDANCES OF
H
60% LIMIT STRESS
AIR-TO-AIR 700 3,1!

AIR-TO-GROUND 100 14U
INSTRUMENTATION
AND NAVIGATION 200 10
TOTAL 1,000 3,300
Figure 100. Distribution of Hours and Exceedances
10,000- -\ _-
1,000 -
EXCEEDANCE
,,, ~ 1OOR
~1,000
HR 100 - /
10
-30 -20 -10 0 10 20 30 40 50 60 70 80 90 100

PERCENT DESIGN LIMIT STRESS OP13-0115105
Figure 101. Measured Mix.Truncated Spectrum
Fatigue specimen joint load-deflection response was moni-

tored during the test. Evaluation of joint load-deflection
hysteresis data provided hole elongation and overall joint
compliance data as a function of cyclic exposure.
Indavidual test specimens were fabricated from 10 panels.
Stacking sequences are detailed in Figure 102.
114
4$
PLY NUMBERS LAY UP NUMBER AND PLY ORIENTATION
TO CENTERLINE 1 2 3 11 12
1 + + + + +
2 0 0 - - 0
3 - - 0 0 -
4 0 0 0 + 0
5 90 + 90 - +
6 0 90 0 0 -
7 + -+ + +
8 0 0 - - -
9 - + 0 -- +
10 0 -0-- I
CL CENTERLINE 90 90
NO. OF PLIES 20 20 20 21 21
PERCENT OF (BASELINE) (BASELINE) 50/40/10 19/76/5 (BASELINE)
00°/450 /90 0 50/40/10 30/60/10 1 1 19/76/5
Notes:
1) "+" and "-" reffers to ±450 phes.
Figure- 102- Task A. I.avup Numbers and Stacking Sequence
All specimens were tested to static failure or exposed to

fatigue loadings according to Figure 98. A double-shear loading
configuration (Figure 99) was used for all but the single-shear
testing described in a following section. Full detail - of all
Task 4 procedures, test data, and support equipment ai provided
in Volume 2 of this report.
b. Experimental Results and Evaluation - Static tensile

strength data obtained at each test condit-ic- for each unique
specimen configuration are summarized in Figure 103. I,oad-versus-
deflection curves obtained in these tests were used to select
fatigue stress levels.
115
d =0.375
00 ef6d
ly 0 in. Ab Torque-Up
1 P~ouble-Shear
tu
_ _ _ 3d _
Test Spring Stearngt Failure Strain Failu~re

o. N.Variation
Parameter (k - Rate
kips/in.) Srnt h
(fbru - ksl)
(tu
(gos
X9,81'
-i.i
ldn
Load
b)
1 Layup. 50/40/10 RTD 400 100 1,400 7,530

RTW 410 90 1,430 7,250
RTWV (TS) 400 90 1,390 7,020
ETW 330 50 790 4,210
I30/60/10RTD 350 100 1,910 7,480
___ 19/76/5 FITD 330 100 2,650 7,850
2 Stacking '50/40/10 .Z90 201,33%) 7,230

Sequence: 119/76/5 320 100 2,720 8,070
3Torque-Up: 50/40/10 400 120 1,770 9,350

(T =160 in.-gb) 19/76/5 300 150 4,160 12,570
4 Guieometry: 50/40/10 e/d =34 30 0 2,210 7,440
w/d 4 3330
bI50/40/10 w/d =4 390 100 2,160 7,610J
19/76/5 w/d 4 2 400 103, 960 I,9N0
e/d = 3
5 0.0045 in. 50/40/10 500 100 1,580 8,10IG

Interference Fit
6 Single Shear 50/40/10 (PROTR) 350 120 1,830 9,250
(T= 160 in.-lb) (CSK) 250 110 1,450 8,200
Pororsity: ____________ 400 _100 1,470 7,810
Figure 103. Summary of Task 4 Specimein Static Strength
Residual strengths were, in general, equal. or greater than

nonfatigued specimen static strengths; however, in moicst cases,
these specimens had acquired hole e long'at ions of .02 inch or
greater during fatigue. For structural applications, h,-d)e elonga-
tio~ns of .02 inch exceed the usual yield criteria for metall ic
joints.16
11----- -W
(1) Layop Variation - Three laminate variations were tested
to determine relatveý fatigue life; the fiber-dominant 50/40/10
layup and two matrix-dominant ] %yups, 30/bo/i0 rnd 19/76/5.
These laminates were selected for fatigue evaluation on the basis
of Task 2 results where changes were noted in a,-rength, failure
mode, and failure initiation point.
Results of tension-tension (R = +0.1) and tension.-comprvs-

sion (R = -1.0) cyclic loading on each laminate at room tempera-
ttire, dry (RTD) test conditions are summarized in Figures 104,
105, and 106, in terms of fatigue cycles required to produce an
0.02 inch elongation of the fastener hole. The results indicate
similar static and fatigue strength for all layups for tension-
tension (R = +0.1) cycling, as summarized in Figure 107. For
tensicn-compression (R = -1.0), the 19/76/5 and 30/60/10 matrix-
dominant layups sustained fewer load cycles prior to developing
an .02 inch hole elongation, as compared to the 50/40/10 layup
(Figure 108).
100,
4 "•"•R = +0.1
c60-
R =-1.0
3 7 5
4 P- P-
0 in -Ib Torque-Up -
Doublfe-Shear 3d
011 1 1 1 1 1 !1 -I 1II !1 1 I I I! I 1 I
1 101 102 103 105 106 107
N - Cycle-
Figure 104. RTD Baseline Joint Fatig'vle Life

50140110 Layup
117
L ,1
. , ,-". . . . ._I..
100
80
60
G1~O __ I
S 40 d - 0,375-"
C--,
20 0 in.-lb Torque-Up
Double-Shear 3d
0 WL llL 1 I L1 I I_ I I I I I I 1 1
1 101 102 103 104 105 106 107
N - Cycles
Figure 105. RTD Baseline Joint Fatigue Life

30/60/10 Layup
.+0.
20
""V- -
0 in.-lb Torque-Up
2
Double-Shear 3d
0 dJW
1 101 102 103 104 105 106 107
N - Cycles OP1S-0115ee
Figure 106. RTD Baseline Joint Fatigue Life

19/7615 Layup
118
100
1 /76/5 Laoup
ZL /-30/60 /10 Layup

80 50/40/10
"• •' Layup
w
S60
~40
20 0 in.-lb Torque-Up I. -
Double-Shear 3d
01 11 1 I !1 I II 1 1 , I I II
1 101 102 103 104 105 106 107
N - Cycles
Figure 107. Comparison of R : + 0.1 Joint Fatigue Life Trends
100
80
-- 0 30/60/10 Layup
U..60
o d 0.375 19/76/5 Layup

S40 -_
_
20 0 in.-lb Torque-Up 3
3d
Double-Shear
0
1 101 102 103 104 105 106
N - Cycles
Figure 108. Comparison of R = -1.0 Joint Fatigue Life Trends

119
In order to evaluate the rate of hole elongation in compos.-
ite joints, the hole elongation data at R -+0.1 is summarized in
Figures 109, 110, and 111 for three levels of hole elongation
(.005, .010, and .020 inch). These data indicate that the matrix-
dominant 19/76/5 layup exhibits earliest initiation of hole
elongation, but had the most gradual rate of accumulationi.
Cuoiversely, the fiber-dominant (50/40/10) layup exhibited the
most-delayed initiation of hole elongation, but had the most
rapid accumulation. The 30/60/10 layup exhibited an intermediate
performance. However, hole elongation arid associated damage in
all three layups resembled bearing failure modes as shown in
Figure 112, indicating a possible common matrix failure mechanism
with di.ffe:ences in initiation time and accumulation resulting
from fiber layup effects.
80
CL
l20 Ai020in.
1 101 102 10 10 15 06 io ,
N -Cycles OP13-011541
Figure 109. Hole Elongation Fatigue Life Trends

50/40/10 Layup
"120
60
CHole Elcongation 0.005-
*m 40
>P P 6d
20
0 in.-Ib Torque-Up [.
Double-Shear 3d
0 101 1 1o111
N - Cycles GP"14

30/60/10 Layup
100
60
Hole Elongation =0.005
(1 0.375
0 rin-lb Torque-Up1-
Double-Shear 3(d
111 i2 3 4o o6 7
N - Cycles PS014

19/76/5 Layup
121
50/40/10 Layup 30/60/10 Layup
7J 12
----1- -0
-I-
19/176/5 Layup
0P13.01 15-U
Figure 112. Representative Specimen Fatigue Failures
1 22
Spectrum fatigue tests were conducted for the 50/40/10,
30/60/10, and 19/76/5 layup with test limit load (TLL) defined at
89%, 93%, and 94% of static strength respectively. The results
of these tests showed no measurable hole elongation for any of
the tested layups after testing equivalent to 16,000 spectrum
service hours. Further, fo: tests with TLL defined at 73%, 87%,
and 89% of static strength, no hole elongation was observed after
32,000 spectrum service hour7s. The differences between constant
amplitude results and spectrum results are attributed I , the
infrequent application of loads high enough to produce.. hole
elongation in any of the three layups tested.
Spectrum fatigue life of the 50/40/10 layup was determined

under the RTW, ETW, and RTW(TS) test conditions to identify envi-
ronmental effects. These results are summarized in Figure 113.
Hole elongation was greater for the RTW condition as compared to
th2 RTD condition for test limit load levels above 95% of static
sti..-ngths with otherwise similar performance belcw that load
level. The RTW(TS) tests produced greater hole elongation than
the RTD tests for all TLL above 85% of static strengths; however,
the performance of the RTW and RTW(TS) tests is not. distinguish-
able from the test data. All residual strengths were greater
than RTD static strength.
Spectrum Hole Elongation (in.) After Fatigue Cycling Joint

Test etum (As a Percent of Static Strength)
H ours@TLL - Staticth
Streng
Conditions Cycled TLL@75% 75-85% 85-95% >95% Stet
50/40/10 <16,000 - (F) J (F)
RTW 16,000 - 0.000 0.0 "0 0.088 7,250
Residual Strength - 8,560 8,670 8,0110
50/40/10 <16,000 - (F)•

RTD 16,000 - 0.0002 - 0.0006 0.03 - 0.046
16,000- 32,000 - (F) I - 7,530
32,000 0.000 - 0.006 - 0.035
Residual Strength 8,980 - 8,080 9,240
50/40/10 <16,000 -
ETW 16,000 - 0.0006 - 0.0012 0.000 0.0008 4,210
Residual Strength* - 8,660 7,540 8,540
50/40/10 <16,000 - (F)$ (F)
RTW (TS) 16,000 - 0.003 - 0.004 0.03 - 0.059 0.053 7,020
Residual Strength - 9,140 8,150 8,650
OPIS-011544
(F) Specimen failure occurred in some joints tested
* Rusidual strength tests at RT
-- No test
Figure 113. Summari of Effects of Environment on Joint Spectrum Fatigue Life
123
The spring rates of the test specimens for the 50/40/10
layup were also determined at various times in the constant
amplitude fatigue testing to determine correlation with hole
elongation data. Hole elongation data for this layup at RTD,
shown in Figure 114, are similar in threshold points and trends
to joint spring rate data summarized in Figure 115.
0.05
0.04
Cycle Peak c 82 1 ksi 70.6 ksi 64.2 ksi

S 0.03 Bearing Stress -- ---- __-'-- __ • __-"-' _ __ __
UJ 0.02 6d
0.03
N - Cycles oles.
Figure 114. Effect of R =+ 0.1 Loading on Hole Elongation

Baseline 50140110 Layup
I/
124
1 1 1 1 1 1 1
500
400 -
- 3
Cycle Peak
Bearing Stress (ksi) = 82.1 ksi 70.6 k•, 64.2 ksi
d 0.375__ _ _ __ _ __ _ _
(n PP6
1L - Double-Shear
0 in.-lb Torque-Up 3d
I I 1 I IiII I I I I I I I l i I
101 102 10 10 10 10 10
N - Cycles
Figure 115. Effect of R = + 0.1 Loading on Joint Spring Rate

50/40/10 Layup
(2) Stacking Sequence - Tests were conducted to determine

the effect of stacking sequence changes on the joint fatigue
performance of the 50/40/10 and the 19/76/5 layups. Stacking
sequence evaluations compared joint life when laminates had +450
plies grouped in a balanced pair as compared to when +450 plies
were uniformly dispersed. In Task 2 testing of the 50/40/10
layup, all alternate stacking sequences which grouped either the
0° plies or the +45* plies exhibited lower static strength. The
0° ply effect was expected; the purpose of these tests was to
determine whether the unexpected effect of the +45° ply grouping
would extend to fatigue life.
Results obtained under constant amplitude cycling of R =

+0.1 and R = -1.0 are presented ir Figures 116 and 117 for the
50/40/10 and 19/76/5 layups respectively. For the 50/40/10
layup, a comparison of static strength data indicated the tested
stacking sequence (variation 3, reference Figure 26) failed at
approximately 6% lower strengths. No significant difference in
static strength was evident between the 19/76/5 baseline and
variation stacking sequence data. A comparison of joint fatigue
life with baseline data indicates no significant differences in
life under constant amplitude cycling (Figures 118 and 119).
125 i
10
fpm
_mJim
20•0 ,n.-Ib Torque-Up I-.4
N - Cycles 0P13-O16-76
Figure 116. Effect of Stacking Sequence on Joint Life

50/40/10 Layu p
126
120
80
60 p
20-
0 in.-lb Torque-Up i-,
Double-Shear 3d
0 - 1i1 1 I IH I I IH I i I IH J I I IH i I i i
104 105 106 107
1 101 102 103
oP13.o11s47
N - Cycles
on Joint Life
Figure 117. Effect of Stacking Sequence
19/76/5 Layup
127
100
80
(n 60 04/
40 \19/76/5 Baseline
________6d______
Double-Shear
0
1101 10 10 10 10~ 101
N -Cycles PS012
Figure 118. Stacking Sequence -Comparison with R = 1.0 Baseline Life Trends
100
80
60
M 40
Dc'ible-Shear 3
1 101 102 10 0 105 106 10

N - Cycles
Figure 119. Stacking Sequence - Comparison with R ++0.1 Baseline Life Trends
128
Detailed plots of hole elongation intervals of .005 and .010
inch are added to the .020 hole elcngation trend ne for both
stacking sequence variations in Figulres 120 and 121 and indicate
a closer groupinq of elongation levels. This indicates a faster
rate of fatigue damage for both of the stacking sequence varia-
tions, however in both cases .initiation thresholds occurred
later.
100
80
40 d .375--' 3d "-
6d
20
0 in.-lb Torque-Up
Double-Sh'.ar
0 1 1 1 1 1 1 lii 1 II I I IL L L1 1 I I 11 I I I
1 101 102 103 104 105 106 107
N - Cycles
0P13-011545
Figure 120. Effect of Stacking Sequence - Hole Elongation Levels

50/40/10 Layup
129
p. j
LI
100
80 _ _ _ __
HoHle Elongation 0.005
CO
P; P 6o
40ý
20
0 in. -IbTorque-Up
Double-Shear
0
103 104 105 106 107
1 101 102
N - Cycles
Figure 121. Effect of Stacking Sequence - Hole Elongation Levels

19176/5 Layup
Effects of spectrum fatigue loading on joint life were simi-

lar in all respects to baseline data. Results of hole elongation
and residual strength after cyclic exposure are listed in Figure
122. No sensitivity of life or strength was indicated.
(3) Torque-Up - To evaluate the effect of fastener torque-

up on joint life, double-shear specimens of the 50/40/10 and
19/76/5 layups were cycled at room temperature, dry, test condi-
tions under constant amplitude tension-tension (R = +0.1) and
tension-compression (R = -1.0) loadings, and under spectrum
fatigue loading. Task 2 testing to determine the effects of
torque-up on joint strength indicated that both strength and fail-
ure mode are altered at 0.0 in-lbs torque-up as compared to
typical 160 in-lb torque values for .375 inch diameter fasteners.
130
I
Joint
Hole Elongation (in.) After Fatigue Loading at joint

Test Spectrum TLL - as a Percent of Static Strength Ltimate
Conditions Hours Static
TLL @ 75 - 85% "1 - 96% 0 >96% (Ib)
50/40/10
RTD 16,000 0 0-0.035 0.018- 0.058 7,230
Residual Strength (Ib) 8 M80 8,340 8,960

19/76/5
RTD 16,000 0 0- 0.001 0.013 - 0.022 8,070
Residual Strength (Ib) 10,360 11 11,280 11,200
Figure 122. Summary of Effects of Stacking Sequbncs on Joint Spectrum Life
S-N data for both layups and at both R-ratios are presented
in Figures 123 and 124 along with curves obtained from testing of
baseline specimens with no fastener torque-up Significant
increases in static and fatigue strength are trvident. While
failure modes were the same for both torque-up levels, areas of
damage surrounding the fastener hole appear more pronounced for
specimens torqued to 160 inch pounds (Figure 125) and overall
fatigue failures occurred more abruptly (faster rates of hole
elongation). This is in contrast to the gradual hole elongation
observed in specimens having no fastener torque-up, which permit-
ted specimens to survive additional fatigue cycling while damage
progressed slowly.
Fastener torque-up, which constrains material through-the-

thickness locally about the fastener hole, inhibited local
failures and slowed the initiation of measurable fatigue damage.
This effect on strength and life was considerably more pronounced
in the 19/76/5 layup (60% increase) as compared to the 50/40/10
layup (20% increase).
Fastener torque-up produced no pronounced changes in joint
flexibility or local spring rate. For the 50/40/10 and 19/76/5
layups, no difference in double-shear spring rates for 0.0 in-lbs
torque-up and 160 in-lbs of torque-up was indicated. This may be
a result of diameter-to-thickness ratio (d/t) greater than one,
producing little bolt bending, especially for the double lap
configuration. Double-shear specimens exhibited a 13 percent
higher spring rate over single-shear specimens for the same
50/40/10 layup.
131
L _!
140
z / _ in To--.
-b -1.o
""16
Il
400
1 101 102
N - Cycles
Figure 123. Effect of Torque-Up on Joint Fatigue Life

50/40/10 Layup
Lobe-ha _mnfal~ 3d__
132
20
0 Specimen failure t•curred

10 in.-lb Torque-Up
ML
020
160 e0 - in.-Sb Torque-Up
__ -[__ - -- - _ __ -
0 pi ' R=+0.1
CL
80d 0.375
P 101 TN 102 103 _WINcle

=4105 10610
40 - Ri.
R-1.0 - (Baseline) -
[Doube.S
I(Baseline)
P 6d "80P1 R-=110-71
1 101 102 10314 051 o6 1

N - Cycles
Figure 124. Effect of Torque-Up on Joint Fatigue Life

19/76/5 Layup
133
(a) 0 in.-Ib Torque Up (b) 160 in.4b Torque-Up
50/40/10 Layup
(c) 0 in.-Ib Torque-Up (d) 160 in.-Ib Torque-Up

19/76/5 Layup
UP13-01oi5-88
Figure 125. Specimen Failures for Torque-Up Conditions
134• _
(4) Joint Geometry - The effects of specimen geometry on
ioint life were evaluated by testing specimens of the 50/40/10
and 19/76/5 layups with different edge distances and widths. In
static tests, these variations resulted in changes in joint
strength and failure mode. Associated nonlinear joint load-
deflection character and static failures in bearing or shearout,
as compared to net tension, might alter the fatigue performance
a- well.
Results for the baseline 50/40/10 layup, in which 4d and 3d

edge distances as well as 4d and Gd specimen widths were tested,
are summarized in Figure 126. These results indicate virtually
identical static and fatigue strength for this layup. Fatigue
and static failures were similar for this fiber-dominant layup
and did not change with geometry variations.
100
e/d = 4
80 ___
'•N~eld= 3-ý• a
U-) 60 -
,-Y
, 40 p- - 4d '"
20 Double-Shear
W• Specimen failure occurred

0 1 1 ,J I . i I 1 [1 1 - I1 1 . 1 1 11 . 1 1 1 1. 1 1 L6 7
S1' 2 04 10 in6 10
N - Cycles
kIOPIS-O115-12
Figure 126. Effect of Geometry on Joint Fatigue Life

50140/10 Layup
Results for the matrix-dominant 19/76/5 layup, in which

var-iations of specimen widths at 4d and 3d were tested (Figure
127), indicate a pronounced effect on fatigue life. Further, a
change of failure fro-a bearing to net section occurred as the w/d
ratio was reduced as shown in Figure 128.
135
I
F..
100
T____
d4
80
(n 60
0 in.-lb Tocique-Up / 3
G Double-Shear
LI g
40 0.375~d~ .3d
10 12 13 14 1 06 17
Np- Cycles
Figure 127. Effect of Geometry on Joint Fatigu. Life

19/76/5 Layup
136
w/d =6 w/d 4
dl w
d 0.37E in All Specimens
w/d -3
GP13-O1IIgo
Figure 128. Changes In Failure Due to Geometry
137
A'
(5) Interference Fit - This series of fatigue tcsts were
conducted to evaluate the effect of interference fits of .005
inch on the fatigue life of no-torque and torqued joints. In
Task 3 testing, interference fits of .003 and .008 inch produced
an increase in static strength (up to 14%) over neat-fit speci-
mens; however, evidence of localized cracking and delamination
was noted as well. Task 4 tests were to evaluate the effects of
these characteristics on fatigue life.
Static strengths of specimens without torque-up were 7.5%

higher than baseline (neat-fit) specimens. These data were con-
sistent with earlier Task 3 data.
Fatigue tests, R = +0.1, of no torque-up specimens resulted

in lower fatigue life relative to neat-fit baseline data. Data
are summarized in Figure 129. Associated with lower life was
large test data scatter. It was suspected that minor installa-
tion damage caused by the interference fit produced the wide
scatter in data when no through-the-thickness constraint existed
due to zero fastener torque-up. Limited fatigue testing of
torqued-up specimens (160 in-lb) produced higher fatigue life
relative to neat-fit specimens with the same installation torque,
Two specimens were tested at 85% and 94% of the neat-fit static
strength, respectively. Both showed fatigue improvements rela-
tive to baseline data (Figure 129) with no hole elongation
recorded at the 85% load level after 106 cycles.
(6) Single-Shear Loading - The effects of joint eccentrici-

ties on fatigue life were evaluated using a single-shear specimen
configuration (Reference Figure 23). Two types of fastener head
configurations were tested; protruding head and countersunk
tension-head. As-manufactured specimens were tested under
tension-tension, R = +0.1, fatigue loading at room temperature.
All specimens were torqued to 160 in-lbs for comparison to data
from torqued-up double-shear specimens.
Effects on joint life are illustrated in Figure 130. For

the protruding head single-lap configuration, static strengths
were identical to the double shear configuration. However, the
added effect of bending stresses due to the single shear confi-
guration indicates a significant reduction in fatigue life.
Further- fatigue life reductions were exhibited by the countersunk
fastener with its associated loss of direct bearing material and
added fastener head flexibility. Joint spring rates for the
double-shear, noncountersunk single-shear, and countersunk single--
shear specimens were respectively 396,000 lbs/in, 347,000 lbs/in,
and 250,000 lbs/in. Representative failures for the three
specimen configurations are shown in Figure 131.
138
p4
* 160 mn.-Ib torque-up
o No torque-up
120-
Baseline 160 in.-tb Torque-Up (Neat F it)

100 • , • /-Baseline 0 in.-Ib Torque-Up
(Nat Fit)
A!
m"•.u:• • •' "-Id =0.375-\ )•• -- • 0( "
40 Doube-Shear --
0.
SG 3d 1301 s7
110 1 102 103 104 105 106 107

N - C y c les O I 1 57
Figure 129. Effect of 0.005 In. Interference Fit on Joint Fatigue Life
50440/10 Lay0 p
C--3
__ " m •0
3d.39
140
Protruding HID. Fouble-Sihear

Single-Shear R=+0.1 /R=+0.1
1 T peFSingle-Shear
600 180 R=+1.1 '
-u 60 d =0.375
100 in.-Ib Torque-Up 3d [ < Specimem Failure Occured

PN -A Ccle
t
Figre13. Efe ofSnl-harLaigo JonaigeLf
1 40401Lau
10 10 10 10 10 10 10i
N - Cycles
Figure 130. Effect of Single-Shear Loading on Joint Fatigue Lifg

50/40/10 Layup
214
140
(a) Single-Shear (b) Single-Shear
Protruding Head Countersunk
I I
I I
4 -J
160 in.-lb Torque-Up
(c) Double-Shear
GP13.0115-91
Figure 131. Failures of SingleShear Specimens
141
__- _ _ _ _ _ _ _ _ ~ -
(7) Porosity - Tests of specimens with moderate porosity
were conducted to evaluate the effects of this anomaly on joint
durability. Testing, in Task 3, indicated that moderate levels
of porosity had a minor effect on static joint strength at room
temperature. Specimens with moderate porosity in regions of
fastener holes were tested under R = +C.1 and R = -1.0 fatigue
loadings at room temperature dry conditions. Life data is com-
pared in Figure 132 with baseline data. No reduction of stat'c
strength or joint fatigue life was indicated.
120
2:00
with Porosity -
240 __'_-'
0 in.-lb Torque-Up
Double-Shear 3d
0 .JI lI I I II I ! II I I 1.L..J.1.....
L LLI I l 1 1 1 I
N - Cycles GPIS.O11B.4
Figure 132. Effect of Porosity on Joint Fatigue Life

50/40/10 Layup
142
I. 60
SECTION V
RECOMMENDATIONS
Based on the results of this program, which were summarized

in Section II, several areas of future methodology development
and experimental verin (•axion are recommended.
The analysis procedure developed in this program s

extended to analyze the general class of stress risers in compo-
site structures. Overall geometry and stress concentration
interactions (e.g. multiple fastener holes) should be included.
Experimental evaluations should be performed to virify developed
methodology, with emphasis placed on multifastener j-iints.
The analysis procedure developed in this program should also

be extended to predict the effects of through-the-thickness
variables, such as stacking sequence, fastener torque-up, and
joint eccentricity. These variables were shown by tests in this
program to have a pronounced effect on static and fatigue
performance. The improved methodology should be verified by
test. The extended analysis technique would permit application
of the methodology to a broader range of joint designs.
Hole elongation mechanisms and their relation to joint
fatigue life should be further studied. Analytic techniques
should be developed to identify critical plies, failure loca-
tions, and failure modes for both hole elongation and overall
fatigue life. Developed methodology should incorporate the
capability to predict effects of finite joint geometry and
through-the-thickness variables. An experimental verification
progra-m should be conducted to verify the developed methodology
for fatigue and hole elongation predictions.
143
S -
r!
SECTION VI
REFERENCES
I. S. P. Garbo, and J. M. Ogonowski, "Effect o4 Variances and

M.-anufacturing Tolerances on the Design Strength and Life of
ý-chanically Fastened Composite Joints," Interim Report
. 'FDL-TR-78-179, 1978.
2. B. Burroughs, et al, "Advanced Composite Serviceability",

Rockwell International, F33615-76-C-5344, Quarterlies, 1976
to Present.
3. S. G. Lekhnitskii, "Anisotropic Plates", Gordan and Breach

Science Publishers, 1968.
4. J. P. Waszczak and T. A. Cruse, "A Synthesis Procedure for

Mechanically Fastened Joints in Advanced Composite Mater-
ials", AFML-TR-73-145, Volume II, 1973.
5. J. M. Whitney and R. J. Nuismer, "Stress Fracture Criteria

for Laminated Composites Containing Stress Concentrations,"
J. of Composite Materials, Volume 8, July 1974.
6. T. DeJong, "Stresses Around Pin-Loaded Holes in Elastically

Orthotropic or Isotropic Plates", J. of Composite Materials,
Volhune II, July 1977.
144
11U.S.Government Printing Office: 1981 757h-002/591

Ada 101657

Uploaded by

Ada 101657

Uploaded by

A101657

EFFECT OF VARIANCES AND

S.P. Garbo and J.M. O4onowski J

McDonnell Aircraft Cormnvey

Final Rnport for Period 15 February 1978- 15 April 1981

-pwnftr URbIftISS idt'IbUwid UUUI*

-__ T DYNAMICS LABGS : ?ORY

When Government drawings, specifications, or other data are

Thie report has been reviewed by the Office of Public

ROBERT L. GALLO, Capt, USAF DAVEY L/ SAITH, Chief

FOR THE COMMANDER

RALPH L. KUSTER, JR., Col, USAF

If your address "is changed, if you wish to be removed front

AFAL R-8--3,4 1' V I h V±

16 DISTRBUTION STATEMENT (of th

Approved for public release; distribution unlimited

IS SUPPLEMENTARY IES " '

Bolted Joints Orthotropic Ltad Distributions

The subject of this program was structural evaluation of mechanically

,'Program activities to accomplish these objectives were organized under

This report documents all program activitics performed under Tasks 2, 3,

Tbis final report is organized in the •oll]owin$ three volumes:

Vlolume 1 - M1ethodology Development ard asta Eval.uation

SECURITY LASSIFICATION OF THIS PArkir-'When Date Entered)

The work reported herein was performed by the McDonnell

Program Manager was Mr. Raman A. Garrett, Branch Chief

I1 SUMMARY AND COMCLUSIONS . . . . . . . . . . . . . . . . 2

a. Effects of Anisotropy on Stress

IV TEST DATA EVALUATION ..... .................. 26

a. Task 2 Test Plan . ....... .......... 26

(i) Strength of Laminates With Unloaded

(a) Layup Variation .. .. ........ . 50

2. EV\LIUATION OF MANUFACTURING AN.) SERVICE

a. Task 3 Test Plan .......... ............... 93

(1) Out of Round Holes .. .......... 94

3. EVALUATION OF CRITICAL JOINT DESIGN VARIABLES

a. Task 4 Test Plan . . . . ............ . . 110

(1) Layup Variation ...... .............. .. 117

V RECOMM>,NDATIONS ....................... 143

VI REFERENCES .............. ....................... 144

1 Superposition of Linear-Elastic Stress Solutions . 6

2 Uniaxially Loaded Infinite Plate ........ .......... 9

3 Assumed Cosine Bolt-Load Distribution ....... ...... 10

4 Characteristic Dimension Failure Hypothesis . . .. 12

5 Failure Criteria Comparison ....... ........... .. 13

6 Superposition of Solutions to Account

7 CircuMferential Stress Solutions at

8 Circ mferential Stress Solutions at

9 Effect of Hole Size on Stress Distributions .... 17

10 Correlation of Loaded Hole Analyses ..... ........ 18

11 Effects of Width on Stress Ccncentrations ....... 18

12 Correlation of Loaded Hole Analysis for Small

13 Effect of Characteristic Dimension on

15 Effect of Biaxial Loading on Failure Location . . . 22

16 effect of Biaxial Loading on Ply Strain

19 Task 2 - Joint Design Variables - Test Matrix . . 27

20 MCAIR Evaluation of Unloaded Hole Specimens -

22 Unloaded Hole Specimen Geometry ... .......... 30

23 Loaded Hole Specimen Configuration ... ......... .. 31

24 Specimen Loading Configurations ... .......... 32

25 Effect of Laminate Stiffness on Specimen

26 Task 2 - Layup Number and Stacking Sequence ... 34

27 AS/3501-6 Lamina Mechanical Properties ....... .. 35

2R Effect of Layup Variation on Unloaded

29 Effect of Hole Size on Unloaded Hole

Tensile Strength ............ .................. 38

32 Effect Off-Axis Loading on Failure Stress ...... 40

33 Efrect of Variation of R, on Predicted

34 Effect of Hole Size on Unloaded Hole

35 Effect of Layup Variation on Unloaded

37 Two-Bolt-In-Tandem Spezimen Raseline

38 Variations of Single 7nstener Specimen

40 Variations of Load-Interaction Specimen

41 Fastener Pattern Specimen Geometry ... ......... .. 49

42 Effect of Layup Variation on Joint

43 Effect of Layup Variation on Joint