Design of Monosymmetric Plate and Box Girders

Rapid Design of Monosymmetirc Plate and Box Girders
Design of Monosymmetric Plate and

Box Girders
By M. Khorasani & S. F. Stiemer
i
The spreadsheet checks for both strength and serviceability

Abstract requirements according to the applicable codes and standards.
This report deals with the design process for steel plate girders and Included with this project is a clear procedure manual so that the
box girders. The design of plate girders is quite prescriptive. A spreadsheet can be utilized for commercial design.
more fundamental approach is required for the design of box
girders. Equations explicitly for the design of plate and box girders
are heavily influenced by empirical data. This work considers pure List of Contents
steel box girders only, and does not include the design of Abstract ........................................................................................... ii
composite box girder sections. CAN/CSA-S6-00 “Canadian
Highway Bridge Design Code” provides detailed design List of Contents ............................................................................... ii
requirements for these composite girders. List of Tables ................................................................................. iv
The design of plate girders follows mostly the requirements List of Figures ................................................................................ iv
specified in the “Handbook of Steel Construction” CAN/CSA- 1.0 Introduction ............................................................................... 1
S16.1. However, the S16.1 clauses relating to bending capacity are 2.0 Buckling of Plates ..................................................................... 2
not well suited for the design of monosymmetric plate girders.
2.1 Buckling of Unstiffened plates ............................................. 2
Therefore, the code recommends a rational method of analysis
such as methods explained in the Structural Stability Research 2.1.1 Uniaxial Uniform Compression ..................................... 2
Council’s Guide to Stability Design Criteria for Metal Structures. 2.1.2 Pure Bending .................................................................. 3
In addition, “Canadian Highway Bridge Design Code” CAN/CSA- 2.1.3 Pure Shear ...................................................................... 4
S6-00 provides additional design information for monosymmetric 2.1.4 Combined Stresses ......................................................... 5
sections.
2.2 Buckling of stiffened Plates .................................................. 7
A steel box girder excluding composite design, hereon simply
referred to as a box girder, is a purely steel section that could be 3.0 Design of Plate Girders (CAN/CSA-S16-01) ........................... 8
designed in accordance with CAN/CSA-S16 “Limit States Design 3.1 Preliminary sizing ................................................................. 9
of Steel Structures”. However, this standard focusses on clauses 3.2 Design of cross section for flexure ..................................... 10
for plate girder design, with little specific reference to box girders. 3.2.1 Lateral Torsional Buckling .......................................... 11
Therefore, additional reference materials such as the Guide to
3.3 Design of cross section for shear ........................................ 13
Stability Design Criteria for Metal Structures, Crane
Manufacturer’s Association of America (CMAA 74-2) standards, 3.3.1 Unstiffened girder webs ............................................... 13
and the Canadian Highway Bridge Design Code CSA/CAN-S6 are 3.3.2 Transversely stiffened Girder webs ............................. 15
used for the design of monosymmetric box girders. 3.4 Design of Cross Section for Combined Flexure and Shear 17
An integrated design and analysis environment in a form of 3.5 Transverse stiffeners ........................................................... 17
formatted spreadsheet is implemented to ease the design process.
ii
3.6 Bearing stiffeners ................................................................ 18 7.2.2.3 Yielding failure ..................................................... 38
4.0 Design of Plate Girders (CAN/CSA-S6-06) ........................... 20 7.2.2.4 Buckling failure .................................................... 38
4.1 Moment Resistance ............................................................. 20 7.2.2.5 Stiffeners ............................................................... 38
4.1.1 Class 1 and 2 Sections.................................................. 20 8.0 Fabrication Considerations ..................................................... 39
4.1.2 Class 3 and 4 Sections.................................................. 21 8.1 Materials ............................................................................. 39
4.1.3 Stiffened Plate Girders ................................................. 22 8.2 Proportioning of spans ........................................................ 40
4.2 Shear Resistance ................................................................. 22 8.3 Selection of a girder cross section ...................................... 40
4.3 Combined Shear and Moment ............................................ 24 8.4 Webs ................................................................................... 40
4.4 Intermediate Transverse Stiffeners ..................................... 24 8.5 Stiffeners ............................................................................. 40
4.5 Longitudinal Web Stiffeners ............................................... 25 8.6 Flanges ................................................................................ 41
4.6 Bearing Stiffeners ............................................................... 26 8.7 Field splices ........................................................................ 42
5.0 Comparison of Codes for Plate Girder Design ....................... 27 8.8 Fatigue Details .................................................................... 42
5.1 Design Requirement for Plate Girders ................................ 27 9.0 Protecting Steel ....................................................................... 43
6.0 Box Girder Design .................................................................. 32 9.1 Weathering Steel: ................................................................ 43
6.1 Section classification .......................................................... 32 9.2 Protective Coatings: ............................................................ 44
Description of Element ..................................................... 32 9.2.1 Inhibitive Primers......................................................... 44
6.2 Shear strength................................................................... 32 9.2.2 Sacrificial Primers ........................................................ 44
6.2.1 Shear strength of box girders based on CAN/CSA-S16- 9.2.3 Barrier Coatings ........................................................... 44
01........................................................................................... 32 9.2.4 Surface Preparation ...................................................... 45
6.2.2 Shear strength of box sections according to SSRC ...... 33 10.0 Erection Considerations ........................................................ 46
6.3 Bending strength of box sections ........................................ 34 Conclusions ................................................................................... 48
7.0 Design using Formatted Spreadsheet ...................................... 36 References ..................................................................................... 48
7.1 Plate girder spreadsheet ...................................................... 36 Appendix A: Lateral-torsional buckling of monosymmetric beams
7.2 Box girder spreadsheet ........................................................ 36 ....................................................................................................... 50
7.2.1 Design using CSA standard and SSRC guidelines ...... 36 Monosymmetric I-girder: .......................................................... 50
7.2.2 Design using basic principles and CMAA ................... 37 Open top box girder with sloping or vertical webs: .................. 51
7.2.2.1 Sectional properties ............................................... 37
7.2.2.2 Stresses .................................................................. 37
iii
List of Tables Figure 21 – Plate girders in the Oak Street Bridge, Vancouver .... 27
Table 1 – Section Class ................................................................. 10 Figure 22 – Plate girders in a building floor system ..................... 27
Table 2 – Differences in the design of plate girders ..................... 31 Figure 23 – Transverse and longitudinal stiffeners in plate girders
............................................................................................... 30
Table 3 – Section Classification ................................................... 32
Figure 24 – Plate girders with openings ....................................... 30
Figure 25 – Common box sections ............................................... 32
List of Figures Figure 26 – Shear Buckling of web .............................................. 34
Figure 1 - Uniform compression coefficients, k, for equation 1 ..... 2 Figure 32 - Monosymmetric box girder designed in spreadsheet . 36
Figure 2 - Buckling of a plate with an aspect ratio of 3:1............... 3 Figure 33 - Shear flow around the section due to Vf .................... 37
Figure 3 - Definition of effective width .......................................... 3
Figure 4 - Plate subject to pure bending ......................................... 4
Figure 5 - Plate subject to pure shear .............................................. 4
Figure 6 - Buckling coefficients for combined bending and
compression ............................................................................ 5
Figure 7 - Stress states in a longitudinally stiffened plate .............. 7
Figure 8 – Unstiffened and stiffened plate girders.......................... 8
Figure 9 – Monosymmetric cross sections of plate girders ............ 9
Figure 10 – Effective distribution of bending stresses.................. 11
Figure 11 – Lateral-Torsional buckling motion ............................ 11
Figure 12 – M-Relationships for laterally unbraced beams ...... 12
Figure 13 – Beam Failure modes .................................................. 12
Figure 14 – Shear strength versus web slenderness ...................... 14
Figure 15 – Web shear strength – unstiffened web....................... 15
Figure 16 - Tension field in stiffened girder web ......................... 15
Figure 17 - Web shear strength – stiffened web ........................... 16
Figure 18 - Shear-moment interaction diagram ............................ 17
Figure 19 – Shear strength versus web slenderness ...................... 23
Figure 20 – Tension field action ................................................... 25
iv
Rapid Design of Monosymmetric Plate and Box Girders SFS & MK
aero-elastic problems with plate girders of I-shape was the Tacoma

1.0 Introduction Bridge.
Steel girders are typically rolled beams, or welded from steel rolled This report presents the design equations and their background for
plates resulting in plate or box girders. A rolled beam is a steel the design of monosymmetric plate girders according to
girder which has been formed by hot-rolling. The most common CAN/CSA-S16 and CSA/CAN-S6 standards. A fundamental first-
types of rolled beams used are the W-, H-, or I-shapes. These principle approach is chosen for the design of box girders along
shapes have the advantage of quick erection, straight forward with design equations based on CAN/CSA-S16, CMAA and
fabrication, and an overall simplicity of design. SSRC. Where appropriate the relevant clause numbers are listed
Rolled beams are sometimes reinforced with a flat plate, or cover next to any equations given in this text.
plate, at the flange. The advantage or cover plate is to increase
flexural resistance of the section without having to use a large size
rolled beam or plate girder.
A plate girder, like a rolled beam, has two flanges and a central
web section. Rather than being hot-rolled as a whole, the girder is
welded from steel plate elements. Older designs may be bolted or
rivetted. When the designer has the ability to specify the section
properties to accommodate the local forces it usually results in
greater economy of material usage. Variation in plate sizes may
represent a girder with the least weight; however, this may increase
the fabrication costs. Plate girders are generally used for larger
spans than rolled beams.
The box girder possess excellent torsional stiffness, thus they do
not usually require secondary members to provide bracing like
plate girders.
Bridge superstructures and stadium roofs are common applications
for steel plate and box girders. Box girders are more expensive to
fabricate than plate girders because of their complexity, however,
have a number of significant advantages, particularly for longer
spans. Firstly because of the shape of the box, the top flange itself
can act as the decking without the need for a concrete or timber
decking. They can also be designed with an aerodynamic shape,
again making them ideal for long spans. A famous example for
Plate_Box_Girder 2011-02-05 page 1 of 55

2.0 Buckling of Plates π2E

σc  k (1)
121  ν 2 b/t 
2
Webs and flanges of both steel plate and box girders are comprised
of flat steel plates with uniform thicknesses. The design of these 
girders consists of choosing appropriate cross sections and
boundary conditions to each of these plate elements, ensuring that Segment of long plate
having thickness t, width b,
each plate element does not fail due to local buckling, yielding or and various edge
an interaction of the two and that the girder does not fail due to conditions as tabulated
global buckling. below
It is desirable for global buckling to be the first mode of failure
reached as the loads on a girder are increased past the anticipated 
service loading. Global buckling is associated with large Description of
Case k b
deformations which give warning of failure, while local buckling is edge support
generally brittle and sudden. 1 Both edges simply supported 4.00
2.1 Buckling of Unstiffened plates 2 One edge simply supported, 5.42

The buckling stresses are obtained from the concept of bifurcation the other fixed
of an initially perfect structure. In practice, the response of the 3 Both edges fixed 6.97
structure is continuous, due to the inevitable presence of initial
imperfections. 4 One edge simply supported, 0.425
For a plate to be considered as slender, the in-plane dimensions, the other free
“a”, and, “b”, need to be significantly greater than the plate 5 One edge fixed, the other free 1.277
thickness, “t”. The dimension, “b”, is usually taken as the direction
transverse to the main direction of in-plane loading named width Figure 1 - Uniform compression coefficients, k, for equation 1
while “a” is taken as in line with the applied load, commonly
named length
The buckling coefficient, k, is determined by a theoretical critical
2.1.1 Uniaxial Uniform Compression load analysis, and is a function of plate geometry and boundary
conditions. The values given in figure 1 are the lower bounds, with
The elastic critical stress of a long plate segment, c, is determined the actual value depending on the plate aspect ratio, m (a/b). This
by the plate width-to-thickness ratio, b/t, by the restraint conditions is because a perfect plate under in-plane compression will buckle
along the longitudinal boundaries, and by the elastic material into m square half waves if the plate aspect ratio is an integer, as
properties. It is expressed as this corresponds to the lowest energy mode and for non integer
ratios the plate will in theory have a higher buckling stress.

Actual distribution
1 No. of half waves
of stress
k
2 e
Region assumed not

be to transmit stress be
3 because of buckling
4 2 2
a = 3b 4
be is plate
buckling
b effective
1 2 6 3 4 e
2 12 Displacement width
Aspect ratio m = a/b controlled
loading b
Figure 2 - Buckling of a plate with an aspect ratio of 3:1
t
When a plate is relatively short in the direction of the compressive

stress (i.e. a/b << 1), the critical stress may be conservatively Figure 3 - Definition of effective width
estimated by assuming that a unit width of plate behaves like a
column. It assumed that the maximum edge stress acts uniformly over two
Local buckling causes a loss of stiffness and a redistribution of ‘strips’ of plate leaving the central region unstressed as shown in
stresses. Membrane tensions are set up, which resist the growth of Figure 3. This width is evaluated in order that the total force
deflection and give the plate postbuckling strength. carried by the plate is equal to the actual response.
Uniform edge compression in the longitudinal direction results in a
non-uniform stress distribution after buckling, and the plate derives 2.1.2 Pure Bending
almost all of its stiffness from the longitudinal edge supports. Equation 1 is used to calculate the critical buckling load, with
Elastic postbuckling stiffness is measured in terms of the apparent substitution of the appropriate buckling coefficient, k. The
modulus of elasticity E* (the ratio of the average stress carried by buckling coefficient of a plate in bending is significantly
the plate to the average strain). For simply supported edges E* = influenced by the fact that half (in the linear response range) of the
0.5E. load is applied in tension.
There is a decrease in stress at the center of the panel because of
the reduction of in-plane stiffness along the center line of the plate
caused by the lateral deflection. This reduction of stress due to
buckling action gives rise to a semi-empirical method of estimating
the maximum strength of plates by the use of the effective width
concept.

  
b b a
=
b
 
a
Figure 4 - Plate subject to pure bending

a
Pure bending plate buckling coefficients: Figure 5 - Plate subject to pure shear
k = 23.9 edges simply supported
k = 39.6 unloaded edges fixed
k = 0.85 top edge free, bottom edge simply supported 2 E
c  k s
12 1   2   b / t 
2
k = 2.15 top edge free, bottom edge fixed (2)
2.1.3 Pure Shear
In a plate subject to pure shear, there exists tension and Shear buckling coefficients:
compression stresses equal in magnitude to the shear stress and
inclined at 45º. The destabilizing influence of compressive stresses 1. Plate simply supported on four edges:
is resisted by tensile stresses in the perpendicular direction. The
critical stress can be obtained by substituting c and ks for c and k
5.34
in equation 1.  ≤ 1: k s =4.00+ (3)
α2
Unlike the case of edge compression, the buckling mode is
composed of a combination of several waveforms, making the 4.00
  1: k s =5.34+ 2 (4)
buckling analysis of shear more complex. α

2. Plate clamped on four edges:

8.98
 ≤ 1: k s =5.60+ 2 (5) Case Loading
α
 
5.60 1 Compressive stresses varying as
  1: k s =8.98+ 2 (6) a straight line b
α 0  1
   
3. Plate clamped on two opposite edges and simply supported on a= b
the other two edges:

2 Compressive and tensile stresses;  
8.98 varying as a straight line and with
 ≤ 1: k s  5.61  2  1.99 (7) the compression predominating b
 -1< <0   a= b  
5.61 1.99
  1: k s  2  8.98  2 (8)
 
 
3 Compressive and tensile stresses;
2.1.4 Combined Stresses varying as a straight line with b
equal edge values  = -1 or with
The Crane Code (CMAA Specifications #70 & #74, revised 2000) predominantly tensile stresses
 
a= b  
gives equations for the buckling coefficient for simply supported  < -1  
plates subject to combined in-plane bending and compression b
according to the cases shown in Figure 6.
   
a= b
Figure 6 - Buckling coefficients for combined bending and compression
The critical stress is then calculated using the buckling coefficient,

k (obtained using equations 9 – 13), in equation 1.
Case 1:
8.4
 1 k  (9)
  1.1

 1   2.1 
2 limit, buckling is said to be inelastic. For inelastic buckling the
  1 k      (10) compression stress is reduced to 1kR.
      1.1 
 y 1k
2
Case 2: 1kR  (15)
0.1836 2y  1k
2
k  1    k '   k ''  10 1     (11)

k’ is the buckling coefficient for  = 0 (case 1)
p = y/1.32 (16)
k’’ is the buckling coefficient for  = -1 (case 3)
The comparison stress is then used to calculate a safety factor, B,
Case 3: which is then compared with allowable design factor values, DFB,
  2/3 k  23.9 (12) for each load combination.
1.87
  2/3 k  15.87   8.6 2 (13)
 2
Elastic buckling
With predominant tension replace the width of the plate, b, by 2 1k

B   DFB (17)
times the width of the compression zone for calculation of  and 2  32
c.
The crane code also gives interaction failure criteria for plates
Inelastic buckling
subject to in-plane bending, compression and shear. First the
comparison stress, 1k, is calculated. 1kR
B   DFB (18)
2  32
 2  3 2
1k  (14) Design factor DFB requirements:
2 2
1       3       
 4       4      
 c  c   c Case 1 DFB = 1.7 + 0.175( - 1) (19)
Case 2 DFB = 1.5 + 0.125( - 1) (20)
If the resulting critical stress is below the proportional limit, p, Case 3 DFB = 1.35 + 0.05( - 1) (21)
buckling is elastic. If the resulting value is above the proportional

2.2 Buckling of stiffened Plates

d1 
d2

d3
 
Figure 7 - Stress states in a longitudinally stiffened plate
Figure 7 shows the stress state in each sub panel for a stiffened
plate subject to bending and shear (compressive stresses can easily
be included by modifying the ratio of longitudinal stresses at the
edge of each panel). Each sub panel can be checked for local
buckling subject to these stresses using the appropriate buckling
coefficients given.
A second mode of failure needs to be checked, too, which is local
buckling of the plate as a whole including the stiffeners.
The third and final mode of failure involves the local buckling of
elements of the stiffeners. This is usually avoided by adhering to
slenderness limits, which is a simplified buckling check.

Rapid
d Design of Monosymmetric Plate and Box Girders SFS & MK
3.0
0 Design of Plate Girrders (CAN
N/CSA-
S16-01)
Platee girders are buiilt up flexural members
m with sleender webs that
are uusually used in long
l spans betwween 15 to 45 m,, exceeding the
usefful range of avaiilable rolled shaapes. The webs can c fail due to
buckkling, yielding or
o a combination n of the two in either
e shear or
bendding. Web dimensions and stifffener spacing aree chosen in ordeer
to ennsure that there is an adequate safety
s margin with
w respect to
thesee failure modes. Flanges are sizzed to prevent lo ocal buckling orr
yieldding.
A tyypical plate girdeer consists of twwo flanges and a web welded
togeether to form an I-section. Theree are several typ pes of plate
girdeers depending ono the stiffeners used. There aree unstiffened
platee girders with noo stiffeners, stifffened plate gird
ders with only
transsverse web stifffeners, and stiffeened plate girderrs with
transsverse and longiitudinal web stifffeners. Figure 8 shows these
typees of plate girderrs and various sttiffeners generaally used in the
platee girders. Figuree 8d shows a typ pical section of a plate girder.
The stiffeners are ussed to improve the t shear capaciity of the webs
insteead of increasingg the overall weeb thickness. Th he use of
stiffe
feners is more ecconomical in lon nger spans undeer greater loads.
Althhough the plate girder
g sections are
a bisymmetricc in Figure 8, thee
monnosymmetric plaate girders are allso common. Fig gure 9 shows
varioous monosymm metric cross sectiions of plate gird ders.
Figure 8 – U nstiffened and stiffened plate gird

ders
Plate_
_Box_Girder 2011-02-05 page 8 of 55
Rapid
3.1 Preliminary
P s
sizing
The seelection of most economical girrder dimensions is dependent
on a nuumber of variabbles which are pprimarily a functtion of the
weightt of steel used anand the amount oof fabrication. A good value to
start with
w for the goodd depth in regardd of moment reesistance
accordding to allowabl e stress design iis.
1/ 3
 Mf 
h  5440 
 F 
(22)
 y 
Fy is th
he yield strengthh of steel and Mf is the maximuum factored
momen nt along the spaan.
An app proximated flannge area (Af) cann be calculated bby assuming
that latteral torsional bbuckling will nott govern the dessign, and the
contribbution of the weeb to the bendinng resistance of tthe girder is
negligiible. This assum mption is valid aas long as laterall supports are
providded at intervals cclose enough to prevent lateral--torsional
bucklin ng.
M
Af  f (23)
Fy h
The web
w thickness, w w, can be calculaated by assuminng that the
entire shear is carried by the web. Thherefore,
V
A w  f  wh (24)
Fs
Vf is th
he maximum facctored shear aloong the span andd Φ is the
perform mance factor eqqual to 0.9 for sttructural steel annd Fs is the
ultimaate shear stress oof steel. Fs will bbe discussed latter. This
parameeter depends onn the web slendeerness ratio and the existence
of trannsverse stiffenerrs. However, foor preliminary deesign Fs is
given byb equation 25.
Figure 9 – Monosymmetric
M cross sections of pla
ate girders Fs  0..66Fy (25)
Plate_
Furthermore, the code imposes limits on the maximum and and compression zones about the neutral axis. 0.9is the
minimum web thickness allowed. performance factor of steel and Fy is the yield stress of steel.
Fh Class 4 sections are such that they buckle locally at a moment less
w min  y (26) than My and the moment resistance is a function of the width-to-
83000
thickness ratios of the component elements. This class is
Fy  h
wmax  subdivided into three categories. The first category, (i), contains
1900 (27) those sections having both flange and web plates falling within
Class 4. The second category, (ii), contains those sections having
3.2 Design of cross section for flexure flanges meeting the requirements of Class 3 but having Class 4
A plate girder subject primarily to bending moment usually fails by webs. The third category, (iii), contains sections having web plates
lateral-torsional buckling, local buckling of the compression meeting the Class 3 requirements, but with compression flanges
flange, or yielding of one or both flanges. The code divides the exceeding Class 3 limits.
section into 4 classes according to their width to thickness ratios. The code recommends to use CSA Standard S136 for calculation
These limits are summarized in the table 1. of the moment resistances of Class 4(i) and 4(iii) sections.
However, it allows the use of equation 30 shown below for Class
4(iii) sections as an alternative to CSA S136.
Table 1 – Section Class
Class Plate girder Flange under Web under flexural
compression compression Mr = SeFy (30)
Class 1 b0 /t ≤ 145/√Fy h/w ≤1100/√Fy

Where Se is the effective elastic section modulus determined using
Class 2 b0 /t ≤170/√Fy h/w ≤1700/√Fy an effective flange width. The effective width is 670t/√Fy for
Class 3 b0 /t ≤200/√Fy h/w ≤1900/√Fy flanges supported along two edges parallel to the direction of stress
and 200t/√Fy for flanges supported along one edge parallel to the
Class 4 b0 /t >200/√Fy h/w >1900/√Fy direction of stress. For flanges supported along one edge, in no
case shall b/t exceed 60.
The Standard specifies for laterally supported members the Plate girder sections that have flanges meeting the requirements of
moment resistance are: Class 3 but having Class 4 webs (Class 4(ii)) can not attain the full
moment resistance from equation 8 due to the local buckling
Class 1, 2: Mr = Mp = ZFy (28) (softening) of the slender web. Most plate girders fall into this
Class 3: Mr = My = SFy (29) class of section. This effect is accounted for in the code through the
Where S denotes the section modulus (S = I/ŷ). Z represents the use of an effective width, by only considering 1/6 of the web area
plastic modulus which is the first moment of area of the tension

in the compression zone to be effective in resisting lateral buckling 3.2.1 Lateral Torsional Buckling
as shown in Figure 10. Beams subjected to flexure have much greater strength and
Compression stiffness in the plane in which the loads are applied (major
flange principal axis) than in the plane of the minor axis. It has been
assumed thus far that the strength of the beam is determined by the
1 of Web capacity of its cross-section, and this, in turn, is dependent on the
6 Area local buckling capacity of its plate elements.
Theoretical
However if the beam is laterally unsupported, the strength may be
governed instead by lateral-torsional buckling of the complete
member, as shown in figure 11. Beams are especially prone to this
type of buckling during the construction phase, where lateral
Experimental bracing are either absent or different in type from their permanent
ones.
Figure 10 – Effective distribution of bending stresses Position

before
loading
In the code, it is assumed that the maximum moment that can be
carried by the sections is that which causes the extreme fiber in the
compression zone to reach yield stress, as the thin web will not 
permit attainment of the theoretical plastic moment of the section.
A linear reduction to this maximum attainable value is then
applied, which is a function of web slenderness, the relative
proportions of the flange and web, and the buckling load of the
web.
 A h 1900   Position Position
M r '  M r 1.0  0.0005 w    (31) before after
 A f  w M f / S   buckling buckling
Where Aw and Af are the web and flange area respectively. Mf is the Figure 11 – Lateral-Torsional buckling motion
factored moment in the girder due to factored dead and live loads.
At a given stage of loading, the cross-section may twist and bend
about its weak axis, reducing its ultimate moment capacity due to

large deflections and yielding. The main parameter affecting

lateral-torsional buckling strength is the distance between lateral Mcr Inelastic
Local
braces. Other influences are: the type and position of the loads, the Buckling Lateral
restraints at the ends and at intermediate locations, the type of Buckling
cross sections, continuity at supports, the presence or absence of
stiffening devices that restrain warping at critical locations, the
material properties, the magnitude and distribution of the residual
stresses, prestressing forces, initial imperfections of geometry and Elastic
loading, discontinuities in the cross section, cross-sectional Lateral
distortion, and interaction between local and overall buckling. Buckling
M Stocky Intermediate Slender
Mp L
My Figure 13 – Beam Failure modes
B A stocky beam is defined as a beam which is able to reach its local
C buckling capacity before lateral buckling occurs. The local
M M A
buckling capacity of Class 1 or 2 sections is Mp and for Class 3
sections, My. A slender beam buckles laterally before the member
D yields, and the resistance to lateral-torsion buckling is based on full

L elastic action. For the intermediate beam, the bending moment at
the instant before lateral buckling is sufficient to cause portions of
the section to yield, thus the resistance to both lateral and twisting
 motions is reduced.
Figure 12 – M-Relationships for laterally unbraced beams The Standard provides an equation for calculating the elastic lateral
buckling strength of doubly symmetric beams.
ω2 π
Beams can also be classified in terms of the effect of lateral-torsion Mu  EI y GJ  πE/L  I y C w
2
(32)
buckling on the ultimate moment capacity attainable, as shown in L
figure 13. 2  1.75  1.05  0.3 2  2.5 (33)

equations given in ‘Guide to stability Design Criteria for metal

where  is the ratio of the smaller bending moment to the larger structures’; however these are in the general form. CAN/CSA-S6-
bending moment at opposite ends of the unbraced length. 06 gives worked examples based on the same expressions listed in
‘Guide to Stability Design Criteria for metal structures’ for
This equation provides a reasonable estimate of the moment at
monosymmetric plate girders and open top box girders.
which lateral buckling will occur, provided that the strains in the
member are less than the yield strain at the instant before buckling. Lateral torsional buckling can be avoided by properly spaced and
Thus, equation 32 is accepted as the basis for the design of slender designed lateral bracing, or by using cross sections which are
members. torsionally stiff, such as box-shaped sections or open-section beam
groups connected intermittently by triangulated lacing or by
Due to relatively large residual stresses in the flange tips, yielding
diaphragms or by ensuring that the required design moment does
will occur when the applied moment reaches approximately two-
not exceed the lateral-torsional buckling capacity.
thirds of the buckling capacity of the member, Mp for Class 1 or 2
sections, and My for Class 3 or 4 sections. Equation 32 is thus valid
until Mu reaches two-thirds of Mp for Class 1 or 2 sections, and My
3.3 Design of cross section for shear
for Class 3 or 4 sections. CAN/CSA-S6-06 identifies 3 limiting states for determining the
shear capacity of the web; shear buckling, shear yielding or a
combination of both. The first mode of failure is dependent on the
Mu ≤ 0.67Mp (Slender members): web slenderness and stiffener spacing. In the following section the
shear resistance of the plate girders with and without transverse
Mr = Mu for Classes 1- 4 (33) stiffeners will be considered.
3.3.1 Unstiffened girder webs

Mu ≥ 0.67Mp (Stocky members):
When steel is subject to a combined stress condition, the yield
stress in shear, Fy, is normally approximated by the Von-Misses
 0.28M p  value, which is increased to allow for the strengthening effects of
M r  1.15M p  1-   M p Class 1 or 2 (34) strain hardening. The shear yielding strength is given by:
 Mu 
F
 0.28M y  Fs   y  0.66Fy (35)
M r  1.15M y  1    M y Class 3 or 4 (35) 3
 Mu 
The general equation for buckling of a plate subject to pure shear
(equation 2) is used to calculate the resistance to shear buckling
CAN/CSA-S16 does not give equations for the calculation of after substituting the correct notation for dimensions of plate girder
mono-symmetric sections such as box girders or plate girders with webs.
flanges of differing width. The code recommends the use of

k 2 E A third mode of failure due to combined shear yielding and

cr  (36) buckling creates a transition curve between the curves given by

12 1   2 h / w  2
equation 40 shown in Figure 14. The equation for this curve is
For a/b ≥ 1.0, for simply supported edges, it is found that: given in the code and was chosen mainly on the basis of
experimental evidence.
4.0
k  5.34  (37)
 a/h 
2
1.  h / w   439 k v / Fy :
The Standard assumes representative values for the terms in
equation 36 (E = 200 GPa,  = 0.3, k = kv, cr = Fs). With these Fs  0.66Fy (39)
values equation 4.9 reduces to:
180000k v
Fs  (38) 2. 439 k v / Fy   h / w   621 k v / Fy :
h / w 
2
Figure 14 below shows equations 35 and 38 plotted on the same 290 Fy k v

Fs  Fcri  (40)
chart, and clearly demonstrates that the failure mode is dependent  h/w 
on the slenderness, (h/w), of the web.
3.  h / w   621 k v / Fy :
180000k v
Fs  Fcre  (41)
 h/w 
2
Fs = 180000 kv
Fs
MPa (h/w)2
The Standard also imposes a limit on slenderness:

Fs = 0.66Fy
 h  83000
  (42)
w Fy
These equations are presented graphically in figure 15.
Web slenderness h/w
Figure 14 – Shear strength versus web slenderness

the plate girder with transverse stiffeners. This tension field is

Fs = 0.66Fy
Fs anchored by the flanges and stiffeners.
MPa Fs = 290 Fykv
(h/w)
a a
Fs = 180000 kv
s
(h/w)2
kv
83000 h
439 621 kv Web slenderness h/w Fy
Fy Fy

Figure 15 – Web shear strength – unstiffened web
The capacity of the section is calculated by multiplying the

ultimate shear stress for the web multiplied by a performance a a
F
factor  and the web area Aw. 2 2
taw
Vr  φA w Fs (43) N.A.
V
Tw
V 2 Tw
3.3.2 Transversely stiffened Girder webs 2
The upper limit of the strength of a girder web stiffened by 2 2
transverse stiffeners is similar to that of the unstiffened girder T + T T
which the strength is corresponding to shear yielding. Stiffened
webs may fail due to shear buckling before shear yielding occurs, Figure 16 - Tension field in stiffened girder web
however, subsequent to buckling the stress distribution in the web
changes and significant amount of postbuckling strength may be
developed because of the diagonal tension that develops in web The code specifies several equations for the shear resistance of the
panels and compressive forces in the transverse stiffeners that web based on the web slenderness and stiffener spacing.
border those panels. This is called tension field action. Figure 16
shows the general distribution of the tension field that develops in 1.  h / w   439 k v / Fy :

Fs  0.66Fy (44) 4.0

a / h  1: k v  5.34  (51)
a / h 
2
2. 439 k v / Fy   h / w   502 k v / Fy : a / h  1: k v  4.0 

5.34
(52)
a / h 
2
290 Fy k v
Fs  Fcri  (45) Where w, h and a are web thickness, web height and stiffeners
 h/w  spacing respectively. Fcre and Fcri are critical elastic and inelastic
buckling stress in shear. Ft is the contribution due to tension field
action.
3. 502 k v / Fy   h / w   621 k v / Fy :
The Standard also imposes the same limit on slenderness as for
Fs  Fcri  Ft (46) unstiffened webs:
 h  83000
  (53)
0.50Fy  0.866Fcri w Fy
Ft  (47)
1  (a / h) 2 These equations are shown diagrammatically in Figure 17. The
dotted line shows the shear resistance of the unstiffened web which
is shifted upwards by the development of the tension field Ft in the
4.  h / w   621 k v / Fy : stiffened web.
Fs  Fcre  Ft (48) Fs = 0.66Fy
Fs Fs = 290 Fykv
MPa
(h/w) Fs = 290 Fykv + Ft
180000k v
Fcre  (49) (h/w)
 h/w 
2 Ft Fs = 180000 kv + Ft
(h/w)2
Ft
0.50Fy  0.866Fcre
Ft  (50)
1  (a / h) 2 kv 621 kv Web slenderness h/w
83000
439 Fy
Fy Fy
502 kv
where Fy
Figure 17 - Web shear strength – stiffened web

The capacity of the section is calculated by multiplying the

ultimate shear stress for the web by a performance factor and the Mf V
web area Aw. 0.727 + 0.455 f = 1.0
Mr Vr
1.0
Vr  φA w Fs (54)
0.75
3.4 Design of Cross Section for Combined
Flexure and Shear Mf
The presence of significant shear and moment together can occur Mr
at certain locations along the girder span such as at the interior
supports of a continuous beam. In such cases, the effect of the
interaction between these two forces upon girder strength must be
considered. This is shown diagrammatically in figure 18. The Vf 0.6 1.0
diagram shows that if the factored moment is less than 75% of the Vr
moment resistance, the full shear shear resistance may be used.
Figure 18 - Shear-moment interaction diagram
Similarly, if the factored shear is less than 60% of the shear
resistance, the full moment resistance is used.
The code applies a straight line to the interaction curve in order to
simplify the equations.
Mf V
0.727  0.455 f  1.0 (55)
Mr Vr
3.5 Transverse stiffeners

Transverse stiffeners are used throughout the web to provide
tension field action. The S16-01 Standard does place limits on
maximum stiffener spacing. The limits for stiffener spacing are as
follows:

67500h D = 2.4 stiffeners composed of plates placed on one side of

a when h / w  150 or; (56) web only
(h / w) 2
Furthermore, the code imposes a limit on the moment of inertia of
the stiffener to prevent lateral displacement of the web. This is
a  3h when h / w  150 (57) given as follows:
4
 h 
Is    (61)
Each stiffener should resist the summation of the vertical  50 
components of the tension field action (F) over one panel width.
To prevent local buckling of the stiffener under the compressive
force, F, the slenderness ratio (b/t) should not exceed:
 hw  a a / h  
2
F t    (58)
2  h 1  (a / h) 2 b 200
  (62)
t Fy
Based on the assumption that the stiffener will yield before
buckling, the Standard provides an equation for the stiffener area 3.6 Bearing stiffeners
required. The application of a concentrated load to the flange of a girder can
result in local failure. This can happen either by localized buckling
of the web in the region where it connects to the flange or by
aw  a/h  Fy overall buckling of the web. In a stockier web, the web will fail
As  1  C D (59) due to yielding. In a more slender web, crippling, or localized
2  1  (a / h) 2  Fys buckling, will govern. Bearing resistance must be calculated for
both possible modes of failure, and the smaller value will govern
 310000K v  as the bearing resistance. The equations for bearing for an interior
C  1  2 
 0.1 (60) location of the girder are given in clause 14.3.2(a):
 Fy (h / w) 
Br  bi w( N  10t ) Fy (63)
The stiffener factor, D, can be found as follows
D = 1.0 stiffeners furnished in pairs
Br  1.45bi w2 Fy E (64)
D = 1.8 stiffeners composed of angles placed on one side of
web only

Where N= length of bearing

w= web thickness
t= flange thickness
bi = 0.8
While the equations for bearing for the end reactions are given by:
Br  be w( N  4t ) Fy (65)
Br  0.60be w2 Fy E (66)
Where be = 0.75

4.0 Design of Plate Girders (CAN/CSA-S6- Laterally supported members are expected to attain their full
plastic moment strength, and thus the moment resistance varies
06) only with the plastic section modulus and the yield strength of the
This section thoroughly covers the design of plate girders steel. A laterally unsupported member may fail by lateral torsional
according to Canadian Highway Bridge Design Code CAN/CSA- buckling or a combination of weak axis buckling and lateral
S6-06. buckling. In order to account for this reduction in bending
resistance, clause 10.10.2.3 (a) introduces equations that reduces
4.1 Moment Resistance the moment resistance of the section based on its geometric
The moment resistance calculations for plate girders are divided properties. These equations are as follows:
into two categories: those for Class 1 and 2 sections in 10.10.2 and
those for Class 3 and 4 sections in 10.10.3. The fundamental 0.28M p
distinction is that Class 1 and 2 sections use the plastic section M r  1.15 s M p [1  ]   s M p when M u  0.67 M p (68)
modulus to calculate the moment resistance, while Class 3 and 4 Mu
sections use the elastic modulus. The reasoning for this is
described in the previous section of this report.
Both 10.10.2 and 10.10.3 include four main subsections. These M r   s M u when M u  0.67 M p (69)
describe (i) the limiting width-to-thickness ratios for steel sections
of the corresponding classes, (ii) moment resistance for laterally
supported members, (iii) moment resistance for laterally
unsupported members and (iv) moment resistance for bending The Mu term in these equations is the critical elastic moment,
about the minor axis. Each of these conditions makes use of which can be described as the moment that will cause buckling in
different geometric properties of the plate girder to calculate the the unbraced beam. This moment is defined in clause 10.10.2.3 (b)
moment resistance. by the equations:
4.1.1 Class 1 and 2 Sections
The basic equation for the moment resistance of Class 1 and 2  2 2
Mu  [ E s I y G s J ( B1  1  B2  B1 )] (70)
sections is the factored plastic moment resistance for laterally L
supported members, defined in clause 10.10.2.2 as:
 2  1.75  1.05  0.3 2  2.5 (71)
M r   s Z x Fy   s M px (67)

x Es I y The last clause in section 10.10.2 identifies the moment resistance

B1   (72) for bending about the minor axis of the box girder. This equation
2L Gs J is:
 2 Es Cw M r  s Z y Fy  s M py (75)
B2  2
(73)
L Gs J
4.1.2 Class 3 and 4 Sections
The coefficient 2 accounts for the increased moment resistance of Class 3 sections will not attain the plastic moment capacity, so the
the beam when subjected to a moment gradient. This coefficient moment resistance is based on the yield moment. The equation for
depends on , the ratio of the smaller factored moment to the larger laterally supported Class 3 sections in bending is defined in clause
factored moment at opposite ends of the beam. The coefficients B1 10.10.3.2 as:
and B2 are included to account for the monosymmetric nature of
the plate girder. These coefficients vary with x, the coefficient of M r  s S x Fy  s M y (76)
monosymmetry, as well as J, the St. Venant torsional constant and
Cw, the warping torsional constant.
These values all take into account the complex torsional buckling Laterally unsupported plate girder sections may be subject to
tendencies of the unusual box girder cross-section. Equations for lateral-torsional buckling, thus beams in these conditions have
these terms are given in the S6 commentary, in section C10.10.2.3, reduced moment resistance equations listed in clause 10.10.3.3:
Laterally Unsupported Members. The coefficient of
monosymmetry is defined by the equation: 0.28My
M r  1.15s M y [1  ]  s M y when M u  0.67M y (77)
Mu
1
x   y( x  y 2 )dA  2e
2
(74)
I xx A M r   s M u when M u  0.67 M y (78)
The closed-form solution to the integral is also provided in the

These equations are very similar to the laterally unbraced bending
commentary, but is not shown here. Also not shown are the
equations from clause 10.10.2.3, except that the plastic moment
equations for J and Cw. These formulas are complex, and would
capacity has been replace with the elastic moment capacity.
be beyond the scope of this report to derive from the geometry of
the cross-section.

Similar to Class 1 and 2 sections, the moment resistance for Class 1 2d 1900
3 sections is based solely on the section properties about the weak [1.0  ( c )] (80)
1200 Acf w M f / s S
axis. The plastic section modulus has once again been replaced 300 
Aw
with the elastic section modulus, leading to the following equation
in 10.10.2.5:
4.2 Shear Resistance

M r  s S y Fy  s M y (79)
The shear resistance of the plate girder is calculated using the area
of the webs, neglecting the contribution of the flanges. The
Section 10.10.3 also provides for the calculation of the moment ultimate shear stress Ft, depends on the slenderness of the web.
resistance for certain Class 4 sections. Plate girders must meet The ultimate shear stress includes two components, the shear
certain additional conditions, such as that the compression flange buckling stress Fcr, and the tension field component of the post-
must have continuous lateral support, and the web must still meet buckling stress Ft. These components are added together to
Class 3 requirements. In such cases, the moment resistance of the provide the ultimate shear stress.
section can be calculated using the equations for Class 3 sections,
except that the elastic section modulus is replaced with the Vr  s Aw Fs (81)
effective section modulus, Se. The section modulus has been
reduced by using only a portion of the actual flange width.
Fs  Fcr  Ft (82)
4.1.3 Stiffened Plate Girders
The clauses in section 10.10.4 reduce the moment capacity of the
plate girders when the webs are slender. This is due to the
susceptibility of slender webs to buckle during flexure, which
would seem to be equally applicable to single webbed plate girders
or double webbed box girders. This section provides additional
limits the width-to-thickness ratios of webs that have transverse
stiffeners, which will be further discussed later in this report. For
girders that have webs without longitudinal stiffeners and that are
more slender than the limit of 2d c / w  1900 / Fy , the moment
resistance will be reduced by the following factor:

Rapid
Fcr  0.577 Fy (83)
Ft  0 (84)
As thee slenderness of the web increasses, the failure m

mode shifts
fu yielding to innelastic bucklinng. This occurs when
from full
k h k
502 v   621 v , correspondding to another sset of
Fy w Fy
equatio
ons for the ultim
mate shear stresss in 10.10.5.1(b)). These
ons now includee a component ffor the post-buckkling strength
equatio
due to tension field acction. This tensiion field action is only
relevan
nt if the web off the girder is stiffened.
290 Fy kv
Figure 19
9 – Shear strength
h versus web slend
derness Fcr  (85)
h/w
The buckling stress and post-buckliing stress vary depending
d on th
he
slendderness of the web.
w The variatiion of these streesses with the 1
Ft  (0.5 Fy  0.866 Fcr )( ) (86)
web slenderness is shown
s in the graaph above, from
m the S16.1 1  ( a / h) 2
Commmentary. Threee sets of equatio ons are given to represent the
diffeerent modes of behavior
b for the webs in shear. The first set off
equaations, in clausee 10.10.5.1(a), describes the critteria for stockierr quations for the most slender caategory of webs, when
The eq
h k h kv
webs, where  502 v , and thee web will fail in full yielding.  621
6 , are deffined in clause 110.10.5.1(c). Foor this case, the
w Fy w Fy
The post-buckling stress
s is neglecteed, and the buck
kling stress is web will
w fail due to ellastic buckling. The post-bucklling stress,
derivved from the vo
onMises-Hencky y-Huber yield crriterion. This assumiing the web is sstiffened, remainns the same as fo
for the inelastic
valuue is more conseervative than thaat used in S16.1.. bucklinng case. Howevver, the equationn describing thee buckling
stress has
h been modifi fied.
Plate_
h k
tension field action to carry shear, that is with  502 v . The
180000kv w Fy
Fcr  (87)
(h / w) 2 girder does not lose a significant amount of shear strength if
Mf/Mr<0.75 because the moment will be carried by the flanges of
1 the girder, and the web will still be able to carry the shear. The
Ft  (0.5 Fy  0.866 Fcr )( ) (88) combined shear and moment equation is shown below:
1  ( a / h) 2
Mf Vf
The shear buckling stress in both elastic and inelastic buckling 0.727  0.455  1.0 (91)
Mr Vr
varies with the shear buckling coefficient kv. This coefficient
depends on the ratio of the stiffener spacing to the height of the
web, and is calculated from the following equations.
4.4 Intermediate Transverse Stiffeners
5.34 The first portion of section 10.10.6, relating to intermediate
kv  4  when a / h  1 (89) transverse stiffeners, deals with the conditions that requires these
( a / h) 2
stiffeners. Web stiffeners in a plate girder will not be required if
the factored shear load is less than the unstiffened shear resistance,
4 and the slenderness of the web does not exceed the limit of
kv  5.34  when a / h  1 (90) h / w  150 . In such instances, the web does not need tension field
( a / h) 2
action to resist the shear loads on the girder. If these restrictions
are not met, then web stiffeners are required.
Once all of the parameters that define the ultimate shear stress of The following paragraph in section 10.10.6 identifies the limits for
the web have been evaluated, the factored shear resistance of the spacing between web stiffeners. The limits for stiffener spacing
plate girder can be determined. This value is compared to the are as follows:
factored shear force in the girder to determine if the shear capacity
is adequate.
67500h
a when h / w  150 or; (92)
4.3 Combined Shear and Moment (h / w) 2
Clause 10.10.5.2 provides the check for sections subject to
combined shear and moment. The equation involves linear a  3h when h / w  150 (93)
interaction with coefficients for shear and moment based on
Basler. The clause only applies to webs of girders that depend on

The limit of 3h for stockier webs is necessary to ensure that tension

field action is properly developed. Interestingly, the limit for the
case of slender webs is provided only to ensure the ease in
handling and fabrication the girders. This limit is not related to the
strength of the girder, and is based practical limits and experience
in the industry.
Once the maximum spacing of the stiffeners has been determined,
the stiffeners must be sized. Lower limits have been placed on the
moment of inertia and the cross-sectional area of the stiffeners.
These limits are necessary to ensure that the stiffeners can
withstand the compression loads from the tension field action in
the web. The limits for moment of inertia and area of stiffeners are Figure 20 – Tension field action
given below:
Additionally, limits are placed on the width-to-thickness ratios of
I  aw 3 j where j  2.5( h / a ) 2  2 but not less than 0.5 (94) the stiffeners to prevent local buckling
4.5 Longitudinal Web Stiffeners

aw a/h Vf
As  ( [1  ] CD  18w2 )Y  0 (95) The main effect of installing longitudinal web stiffeners is the
2 1  (a / h) 2 Vr change in the allowable spacing between intermediate transverse
stiffeners. These new spacing requirements are defined in section
10.10.7. This modified spacing may provide an overall reduction
310000kv
C  1 but not less than 0.10 (96) in material if the transverse stiffener spacing is increased
Fy (h / w) 2 sufficiently. The maximum transverse stiffener spacing is
modified to 1.5hp, where hp is the maximum subpanel depth.
However, this spacing limit is not necessarily an increase, since
The tension field action in the web develops like a truss, where the webs without longitudinal stiffeners have a maximum spacing of
segments of web between the stiffeners will transfer the shear
3h, if h / w  150 . The creation of subpanels due the longitudinal
loads in tension to the adjacent stiffeners. These stiffeners will
stiffener reduces the slenderness of the web, and increases the web
then carry the shear in compression, completing the truss.
stiffness. Therefore, slender webs with longitudinal stiffeners are
not subjected to the stiffener spacing limit of

67500h and the smaller value is used as the overall bearing resistance.
a when h / w  150 . Two sets of equations are given for bearing at the end of the beam
(h / w) 2
and along the span of the beam. The equations are similar in form,
Clause 10.10.7.2 provides limits for the size and thickness of but have slightly different coefficients. Notably, these clauses
longitudinal stiffeners. The maximum width-to-thickness ratio is include a distinct resistance factor for bearing, being smaller than
200 / Fy , and the maximum width of the stiffener is 30t. This the resistance factor for flexural compression. The equations for
clause also specifies minimum values for the moment of inertia bearing within the span of the beam are givin in clause
and the radius of gyration for the longitudinal stiffeners: 10.10.8.1(a):
I  hw3 [ 2.4( a / h) 2  0.13] (97) Br  bi w( N  10t ) Fy (99)
Br  1.45bi w2 Fy E
Fy
ra (98)
1900 While the equations for bearing at the end of the beam are given in
These limits ensure that the longitudinal stiffeners will not undergo clause 10.10.8.1(b):
local buckling. Clause 10.10.7.3 provides a number of adjustments
to the parameters used to calculate the properties of transverse Br  be w( N  4t ) Fy
stiffeners when longitudinal stiffeners are present. The main
difference is that typically the subpanel height is used instead of Br  0.60be w2 Fy E (100)
the full web height when calculating slenderness ratios. In this
way, longitudinal stiffeners can significantly improve the
effectiveness of transverse stiffeners. These equations are not the same as those given in S16. The S6
equations do not consider the distance from the flange to the toe of
4.6 Bearing Stiffeners the web fillet weld, nor the depth of the web. The equations in S6
Section 10.10.8 provides equations to determine the strength of the are derived from finite element analysis, while those from S16.1
web in bearing. Webs can fail in bearing either due to crippling or are based on empirical analysis.
yielding. The failure mode depends on the slenderness of the web. If the bearing resistance of the web is less than the factored
In a stockier web, the web will fail due to yielding. In a more concentrated loads at the point of application, bearing stiffeners are
slender web, crippling, or localized buckling, will govern. Unlike required. S6 also includes equations that determine the bearing
similar clauses, a limiting width-to-thickness ratio is not provided resistance of the bearing stiffeners.
to determine which condition will govern. Instead, a bearing
resistance must be calculated for both possible modes of failure,

Rapid
5.0
0 Comparis
son of Cod
des for Pla
ate Girder
Deesign
In thhis section the reequirements for the design of pllate girder
desiggn will be comp pared between th he “Handbook of o Steel
Connstruction CAN/CSA-S16” and the “Canadian Highway H Bridgee
Desiign Code CAN//CSA-S6-00”. Plate girders can be used in
bridgges, as crane girrders in industriial buildings, an
nd for long floor
spanns in other builddings. As their crross section are chosen
accoording to the loaad demand of thee individual stru ucture, they are
most economical an nd efficient. Figu ure 21 shows pllate girders
empployed in bridge span. Figure 22 2 shows plate girders as a part ofo
buildding floor system m.
Figure 22 – Plate girders in a building floor sysstem
5.1 Design
D Requ
uirement forr Plate Gird
ders
The bu uilding code andd bridge code arre similar in the major aspects
of desiign of the plate girders. The prooportioning of fl
flanges, web
transveerse stiffeners, aand bearing stiff
ffeners, design fo
for resistance of
combin ned shear and mmoment, and ressistance to web ccrippling and
Figure 21 – Pla
ate girders in the Oak
O Street Bridgee, Vancouver yieldinng are similar inn most parts in bboth the codes. M
Minor
differeences exist betw
ween the codes inn the proportionning.
Plate_
The bridge code limits the minimum web thickness to 10mm while Furthermore, the bridge code replaces the web height h, with 2dc
the building code has no such restriction (S6.06 Clause 10.7.2). where dc is the depth of compression portion of web, for the
This restriction in bridge code is to reduce the susceptibility of web computation of web slenderness under flexural compression (S6.06
to fatigue and brittle fracture. On the other hand, the building code Clause 10.10.3.1). This implies that the depth of compression in
restricts the web slenderness ratio, h/w, to less than 83000/Fy, the web must be equal to half of web height in order to be
(S16.01 Clause 14.3.1) where comparable to similar restrictions in the building code. Therefore,
h = height of the web and clear distance between the flanges the bridge code considers the web as slender when (2dc/w)
w = thickness of the web >1900/√Fy, whereas the building code considers it for h/w >1900/√
Fy.
Fy = yield strength of the steel in MPa.
The reduction in moment resistance due to slender web, in building
For Fy =350 MPa, the slenderness ratio is limited to 237. This limit code is computed as, (S16.01 Clause 14.3.4),
is waived if it is proven that the compression flange does not
buckle under factored loads. In the case of bridge code, the
maximum slenderness ratio is 6000/√Fy, for webs stiffened with  A h 
both longitudinal and transverse stiffeners (S6.06 Clause M r '  M r 1  0.0005 w   1900  (101)
 Af w M f / S 
10.10.4.2). This ratio is equal to 321. A comparison of the ratios at   
first indicates that the bridge code is more lenient with respect to
web slenderness. However, it must be noted that the use of
longitudinal web stiffener would reduce to the effective height of where
web and provides increased resistance to web buckling. In the case
of building code, although the use of transverse stiffeners is Mr = factored moment resistance of the plate girder
implicit, the limit on slenderness ratio is for webs without Aw = web area
longitudinal stiffeners. Hence, the building code is more lenient in
limiting the web slenderness. However, the building code specifies Af = flange area
that the structures subjected to fatigue must have web slenderness Mf = maximum bending moment
ratio less than 3150/√Fy, for webs with transverse stiffeners = resistance factor of steel = 0.90
(S16.01 Clause 26.4.2). This is similar to the restriction in the S = elastic section modulus
bridge code for webs with transverse stiffeners (S6.06 Clause
10.17.2.5). Therefore, the bridge code is more conservative with
proportioning of webs due to fatigue considerations. Similarly, for In bridge code, it is computed as (S6.06 Clause 10.10.4.3),
flanges the building code limits the maximum width-to-thickness   2d 
Aw  c  1900
ratio to 60 (S16.01 Clause 13.5), whereas the bridge code restricts M r '  M r 1   (102)
 300 A  1200 A  w M f / S 
it to 30 (S6.06 Clause 10.10.3.4).  w cf
 

In designing the transverse stiffeners, the building codes specifies

where the minimum moment of inertia of the stiffener about the web, I, as
(h/50)4, whereas the bridge code specifies I ≥ aw3j where,
= resistance factor of steel = 0.95
Acf = area of the compression flange
a = spacing between the stiffeners
The comparison of Eq.s (25) and (26) yields two main differences.
The web height in building code is replaced by 2dc in the bridge j = 2.5(h/a)2 – 2 ≥ 0.5
code and the factor of 0.0005 applied to the ratio of web area to
flange area. The factor of 0.0005 Aw/Af is equivalent to Furthermore, the area of the transverse stiffeners is specified in the
Aw/2000Af. If compression flange area and web area are equivalent building code as (S16.01 Clause 14.5.3),
in Eq. (26), ratio of areas becomes Aw/1500Acf. This implies that
the bridge code specifies a greater reduction in the moment
resistance for plate girders with slender webs compared to the  aw  a/h  
building code with a similar girder design. As   1  CYD  (103)
 2  1  ( a / h) 2  
Moreover, for the web in flexural and axial compression, the    
slenderness ratio limit in building code includes  factor applied to Where
Cy, axial compression load at yield stress (S16.01 Table 2, Clause
11.2). The bridge code follows the previous edition of the building 310000k v
code, which does not include the  factor (S6.06 Table 10.3). Both C  1  0.10 (104)
codes are similar in all other aspects of width-to-thickness ratios. Fy (h / w) 2
Further major differences between the codes are in two main areas:
stiffeners and openings. The building code is more lenient with D = stiffener factor
respect to the stiffeners than the bridge code. In building code, the Y = ratio of specified minimum yield point of web steel to
intermediate transverse stiffeners are waived if the factored shear specified minimum yield point of the stiffener steel
resistance, i, is greater than the shear force under factored load, Vf.
In bridge code, the intermediate transverse stiffeners must be
designed unless h/w ≤ 150, and Vr > Vf (S6.06 Clause 10.10.6.1). The area in Eq. (27) can be further decreased by a ratio of Vf / Vr.
Another minor difference between the codes is in the computation In the case of bridge code, the area of transverse stiffener is
of factored shear resistance. The building code specifies greater proportioned as (S6.06 Clause 10.10.6.2(b)),
shear resistance for h/w ≤ 439√kv/Fy, where kv is the shear  aw  a/h Vf 
buckling coefficient (S16.01 Clause 13.4.1.1(a)). In contrast, the As   1   CD  18w 2 Y  0 (105)
 2  1  ( a / h) 2  Vr 
bridge code maintains that all webs with slenderness ratio less than    
502√kv/ Fy, have same shear resistance (S6.06 Clause 10.10.5.1).

Rapid
The comparison of thet proportionin ng of As reveals that the bridge

codee allows lesser area
a of stiffenerss than the buildiing code.
Morreover, the bridgge code specifies the width of thhe plate used as
fener to be greater than a quarter of the flange width.
stiffe w It should
also be greater than (50+h/30) (S6.0 06 Clause 10.100.6.2). The
buildding code does not
n have these restrictions.
r
Figuree 24 – Plate girderrs with openings
In addition to the abovve differences, tthe building codde addresses

the connnections of stifffeners, proportiioning of the ennd panel and
cover plates
p to the flannges. These details are not addrressed in the
bridge code. Instead, tthe proportioninng of the longituudinal
stiffen
ners and the desiign of transversee stiffeners in thhe presence of
longituudinal stiffenerss are detailed. Inn contrast, the buuilding code
does not
n stipulate desiign procedures ffor longitudinall stiffeners. The
plate girders
g in the buuildings do not uusually need the longitudinal
Figure 23 – Tran
nsverse and longitu
udinal stiffeners in
n plate girders
stiffen more economicall in buildings to employ other
ners since it is m
ural systems, succh as trusses, if greater shear reesistance is
structu
warrannted. Figure 15 sshows the transvverse and longittudinal
ners in a plate girrder for bridge. Furthermore, openings in the
stiffen
plate girders
g are expliicitly addressed in the building code. For the
plate girders
g utilized iin the buildings,, the openings are essential to
run thee building utilitiies. The openinggs are not a conccern in case of
bridges and hence, theese are not addressed in the briddge code.
Figuree 16 shows the pplate girders witth openings. Tabble 2
summaarizes the differrences between tthe building andd bridge codes
Plate_
in the design of plate girders, along with the relevant clauses. The Proportioning of end panel 14.4.1 Not
design examples for composite decks and plate girders according addressed
both the codes are presented in the following section.
Cover plates of flanges 14.2.4 Not
addressed
Table 2 – Differences in the design of plate girders
Longitudinal stiffeners Not addressed 10.10.7
Openings 14.3.3 Not
Aspect of design Building Bridge Code
addressed
Code Clause Clause
Minimum web thickness Not addressed 10.7.2
Web slenderness ratio 14.3.1 10.10.4.2
Width-to-thickness ratio Table 2 Table 10.3
Class 4 sections 13.5(c) 10.10.3.4
Reduction in moment 14.3.4 10.10.4.3
resistance for slender webs
Monosymmetric sections Not 10.10.2.3
Addressed
Waiver of transverse stiffeners 13.4.1.1, 14.5 10.10.6.1
Factored shear resistance 13.4.1.1(a) Not
addressed
Proportioning of moment of 14.5.3 10.10.6.2(a)
inertia of transverse stiffeners
Proportioning of area of 14.5.3 10.10.6.2(b)
transverse stiffeners
Proportioning of width of Not addressed 10.10.6.2(b)
transverse stiffeners
Connections of stiffeners 14.5.4 10.10.6.4

6.0 Box Girder Design 6.1 Section classification

Steel box girders are used in bridges due to their torsional capacity The maximum width-to- thickness ratios are presented as a
and good stability during construction. Box girders are more stable constant divided by the square root of the specified yield strength
and able to span greater distances than plate girders. However the of the steel. The limiting values are given in the table below:
design and construction of box girders are more difficult than plate
girders. The shapes of box girders are usually rectangular or Table 3 – Section Classification
trapezoidal in and can be either open or closed, as shown in below.
Class 1 Class 2 Class 3
Description of Element
Flanges of box girders b 525 b 525 b 670
  
Multi-Spine Multi-Cell t Fy t Fy t Fy
Webs in flexural h 1100 h 1700 h 1900

  
compression t Fy t Fy t Fy
"Bathtub" Trapezoidal Rectangular
Open top 6.2 Shear strength

Closed top
Figure 25 – Common box sections 6.2.1 Shear strength of box girders based on CAN/CSA-
One component of the overall design procedure is to determine the S16-01
capacity of the box girder prior to attaining composite action with The shear design of box girders based on CSA-S16-01 is presented
the concrete deck. This situation arises during construction, when in this section. The design is applicable to design of both
the concrete deck has not yet been poured, and after the concrete symmetric and mono-symmetric sections. The factored shear
has been poured but has not yet hardened. In such cases, the box resistance, Vr, developed by the web of the flexural member shall
girder is designed as a non-composite steel section. This report be taken as
will discuss this non-composite aspect of steel box girder design. Vr   Aw Fs (106)
The Standard guides the designer to consult alternative reference
material such as “Guide to stability design criteria for metal
structures” for the design of box girders. where
Aw = Shear Area
Fs = as follows

a = distance between the stiffeners

h k h = web depth
when  439 v (107)
w Fy
Fy K v
Fs  0.66 Fy Fcri  290 (113)
( h / w)
Kv h k
(b) when 439   502 v (108) K a = aspect coefficient (114)
Fy w Fy
1
Fs  Fcri =
1  ( a / h) 2
Kv h k
(c) when 502   621 v (109)
Fy w Fy 180000k v
Fcre  (115)
Fs  Fcri  k a (0.50Fy  0.866Fcri ) (h / w) 2
6.2.2 Shear strength of box sections according to SSRC

k h To find the shear strength of box section, the approach defined by
(d) when 621 v  (110)
Fy w SSRC’s (Structural Stability Research Council) “Guide to
Stability design criteria for metal structures” is also applicable
Fs  Fcre  ka (0.50Fy  0.866Fcri ) because in such sections the shear strength is the combination of
strength provided by both, the web before buckling and the
diagonal tension after the buckling of the web.
kv = shear buckling coefficient
The shear strength of the box-section can thus be found by the
when a / h < 1 (111) following formula:
5.34
kv  4 
( a / h) 2 Vu  VB  VT
(ii) when a / h  1 VB  Dt w Fvcr (116)
(112)
Dt w FT
VT 
2( 1   2   )
4
kv  5.34  Where
( a / h) 2

Rapid
D  depth of the web

w between flan nges
d 0  transverse stifffener separation
n
  do / D
t w  web thickness
Fvcr  critical buckliing shear stress
FT  tension-field stress
s
Therre are various models

m to calculaate the post-buckkling strength of
o
the wweb of the box oro plate girder. Basler
B was first to model the
tensiion field action for the plate girrder (1953). Wo oulchuk and
Mayyrbourl (1980) suggest the appliication of the Baasler’s model,
whicch is based on th he assumption ofo negligible ben nding rigidity off
the fflanges, for the box
b girders.
Accoording to Baslerr tension field model
m (1953) tennsion field stresss
and critical buckling g shear stress arre additive, assu
uming that Figuure 26 – Shear Bu
uckling of web
tensiion field stress acts
a at 45 degreee and the modell uses the
resuulting combinatio on of the princippal stresses in th
he linear 6.3 Bending
B stre
ength of box
x sections
apprroximation of th he Mises yield condition. This results in the SSRC guidelines are ffollowed to estim mate the flexuraal capacity of
folloowing: the boxx-section.
Accord ding to SSRC guguide lines, “Thee flexural strenggth of box
 section
ns is rarely goveerned by flexuraal torsional buckkling. Instead
F 
FT  Fyw 1  vcr  (117) the govverning criteria is buckling of ccompression flannge or yielding
 F 
 vyw  of the tension flange, which ever occuurs first”.
The buucking of the coompression flangge can calculateed using the
where basic plate
p buckling eequation:
Fyw  yield stress fo
or web in tension
n
 2E
Fvyw  yield stress fo
or web in shear c  k (118)
12(1  2 )(b / t ) 2
Fvcr  critical buckliing shear stress
Plate_
where
E = modulus of elasticity
 = Poisson ratio
k= buckling coefficient
t = thickness of the compression flange
b = width of the plate (distance between the webs for this case)
The top flange is in the state of uniform compression. It can be

assumed as a plate with simply supported edges. In other words,
buckling coefficient can be assumed as k=4.0.
If buckling stress is less than the yield stress then
M r  S c c (119)
Otherwise
M r   S t t (120)
where
 t = yield stress in tension
 c = compression flange buckling stress
S t = section modulus for extreme tension fiber
S c = section modulus for extreme compression fiber

7.0 Design using Formatted Spreadsheet ITFS,ATFS

LTF tTF,ITF,ATF
Formatted spreadsheets are developed for ease of design. The
spreadsheets follow the same approach given by the Standard and XTFS2 hws
presented in the preceding sections. XTFS1
7.1 Plate girder spreadsheet

The plate girder spreadsheet consists of three sub sheets. The first yTFS
sheet is the design based on CAN/CSA-S16 standard. The second
sheet is the design based on CAN/CSA-S6-00 standard. The third yWS2 LW
IWS,AWS
sheet is the design of plate girder based on the combination of both
codes. yWS1
yBFS XBFS1
7.2 Box girder spreadsheet
XBFS2
For the design of closed box girders, a formatted spreadsheet was
developed which checks for local buckling or yielding of a LBF
monosymmetric box girder stiffened longitudinally and IBFS,ABFS
transversely, with webs and flanges each stiffened longitudinally tBF,IBF,ABF
with two stiffeners. The spreadsheet contains of two sub Figure 27 - Monosymmetric box girder designed in spreadsheet
spreadsheets. The first spreadsheet is based on CAN/CSA-S16
standard and SSRC guidelines. The second spreadsheet is based on
The spreadsheet is designed in a way that it calculates the flexural
basic principles and CMAA standard. The spreadsheets require the
resistance and shear resistance of the section with and without
user to enter the complete dimensions of the girder according to the
longitudinal stiffeners.
figure provided below.
7.2.1 Design using CSA standard and SSRC guidelines
The analysis checks for the web-crippling and compression flange
buckling and calculates the ultimate moment resistance of the box
girder using SSRC guidelines. Furthermore, the shear resistance of
the girder is calculated using both the CSA-S16 code and SSRC
guidelines.

7.2.2 Design using basic principles and CMAA

The analysis uses solid mechanics principles to calculate stresses at M y
 f (125)
each panel in the box girder and then uses the equations in CMAA
I
standard to combine the stresses as previously described in section xtot
2. The equations used in this analysis are described in this section
The factored shear (assumed to be applied along the line of
7.2.2.1 Sectional properties
symmetry) gives rise to a shear flow around the section, q, which is
The neutral axis height, ў, and moment of inertia of the box girder calculated using equation 126. The general distribution of shear
section, Ixtot, are easily calculated using equations 121 and 122. around the section is given in Figure below:
 yi A i Vf
y (121) q Dx  qo (126)
A tot I xtot
VF
 I  A (y  y) 2
xi  i i
I (122)
xtot
For the calculation of effective thickness of flanges or webs with

the longitudinal stiffeners smeared, equations 123 and 124 are
used.
N. A.
I xstiffenedplate  I xplate   I xstiffeners   A(yi  ystiffenedplate )2 (123)
1/ 3
 12I xstiffenedplate 
t eff   (124)
 L
 plate 
7.2.2.2 Stresses
The factored moment and previously calculated sectional
Figure 28 - Shear flow around the section due to Vf
properties allow the calculation of the longitudinal stresses at any
point in the section (assuming that plane sections remain plane).

By taking a cut along the vertical line of symmetry of the bottom occurs when the energy of distortion reaches the same energy for
flange, the constant shear flow term, qo, is zero and the shear flow, yield/failure in uniaxial tension).
q, can be evaluated traveling anti-clockwise around the section by
evaluating Dx.
12  1 2   22  Fy2 (130)
D x   os tyds   ysi Asi (127)

The von Mises yield criterion was chosen over other yield failure
criteria such as the Tresca criterion, as it is more conservative.
q
 xy  (128) 7.2.2.4 Buckling failure
t
Equations presented in section 2 are used to check for local
buckling of the sub panels between longitudinal stiffeners and also
Once the longitudinal (normal) and shear stresses are known, the
the wider panels with the longitudinal stiffeners smeared to create
principal stresses, 1 and 2, can be calculated (the maximum and a plate with increased thickness.
minimum normal stresses in a plane, always perpendicular to each
other and oriented in directions for which the shear stresses are Lateral-torsional buckling is not evaluated for the closed cell box
zero). girder, as it is unlikely that this failure mode will dominate for
typical box girders. However if the height to width ratio of the box
is relatively large, lateral-torsional buckling may dominate and
x  y  x  y 
2 needs to be evaluated.
2
1 ,  2       xy (129)
2  2  7.2.2.5 Stiffeners
No provision for design of stiffeners has been made in the box
girder formatted spreadsheet, buckling of the longitudinal
stiffeners can be avoided through conformance with slenderness
The principal stresses are calculated at several discrete points; the limits given in the Standard.
intersection of the flanges and web, the web stiffener locations, and
at the neutral axis.
7.2.2.3 Yielding failure

These principal stresses are used to check for local yielding failure
of the cross section using the Von Mises failure criterion for plane
stress, which is given in equation 130 (this assumes that failure

8.0 Fabrication Considerations The allowable fabrication tolerances are defined in W59, clauses
5.8 and 5.9. The tolerances in the individual pieces that make up a
Evaluation of the economics of design often includes a perception
continuous span will be additive.
that least weight and least cost are synonymous. Although cost of a
structure is related to the weight of steel material, there are
8.1 Materials
numerous other considerations in purchasing, fabricating, shipping,
and erection and effective use of material locally which may Weathering steel is now the norm for bridges in Canada. Painted
override the decision to aim for a least weight structure. Some of steels are used in environments not considered acceptable for the
these considerations require familiarity with purchasing, weathering process, such as continued wetness due to climate and
fabricating and erecting processes. precipitation, proximity to airborne chlorides e.g. near the sea coast
or above a high traffic volume expressway, and exposure to harsh
Various fabricators have their own processes, and it is difficult for
industrial environments. In many cases weathering steel is selected
a designer to produce a design to satisfy everyone. Fabricators
even when a paint system is to be applied due its lower strength to
should be allowed flexibility in detailing, with designer approval,
cost ratio and its ability to form a superior base for paint systems.
to make adjustments to the number and location of splices.
A designer needs to be aware of the plate sizes available so that
Material content is only one element in the cost equation and will
spices in webs and flanges are kept to a minimum, particularly
represent about 20% to 30% of the total ‘in place’ cost in fairly
longitudinal splices which should be avoided. The maximum
standard bridges.
length of plate that may be supplied is dependent on the thickness
The total rate per tonne (metric = 1000 kg) of steel depends on of the plate and the material type, and will vary from mill to mill
several factors, including: and hence local fabricators should be consulted.
 Complexity of details The designer should also be aware of other factors which influence
 Quality control requirements material cost such as:
 Amount of welding, including grinding, type and amount of  There is a small premium on plates longer than 18 meters
inspection etc… (about 4%)
 The amount of repetition and reuse of assembly jigs  Plates less than 9 mm thick and more than 25mm thick
 Size and number of individual pieces to be fabricated attract a premium of from 4% - 8%
 Other demands on shop space, particularly when large box  Small orders also incur mill extras and small quantities of
girders are involved any one plate thickness should be avoided.
 The access for erection As a general guide the maximum piece-weight for handling in the
shop is of the order of 50 tonnes, and the optimum length for the
 Number of girder field splices shop is about 27m although these values are increasing.

8.2 Proportioning of spans throughout the girder, varying the spacing of intermediate
When there is choice in the positioning of piers for a continuous transverse stiffeners according to the shear diagram and possible
bridge. The end spans should be approximately 75% of the length eliminating transverse stiffeners in the areas of low shear.
of the main span, this will permit balancing of dead and live load The recommended minimum web thickness is ½ inch as thinner
moments, reduce the potential for uplift at the abutments, and plate is subject to excessive distortion from welding. Web
permit the most economical design when proportioning the girder. thickness increments should be 1/16 inch up to a plate thickness of
¾ inch, use 1/8 increments up to 1 inch, if the web plate needs to
8.3 Selection of a girder cross section be thicker than 1 inch, use ¼ inch increments.
For compositely designed continuous spans, the designer should For web splices use the submerged arc process and avoid grinding
start with a main span to girder depth ration of approximately 28 if possible, or use only nominal grinding to touch up the profile of
for box girders and 26 for I girders. On bridges where there are no full penetration butt welds in the web. Web shop splices should be
pedestrians the bridge may be made more slender due to the at least 10 feet apart and at least 6 inches away from a flange splice
reduced deflection requirements, ratios of 30 to 34 may be used or transverse stiffener, in order to facilitate testing of the weld.
successfully.
8.5 Stiffeners
8.4 Webs Welding of bearing stiffeners to the bottom flange should be
The optimum web thickness and subsequent number of transverse specified as fillet welds, use of full penetration welds is costly and
stiffeners depends on the depth of the web and should be can cause distortion of the bottom flange, thus making it difficult
considered. For example, it is economically advantageous to have to achieve the desired flatness for the sole plate or bearing.
an unstiffened web if the girder is 1200 mm deep or less. The For composite bridge girders, stiffeners welded to the top flange
economics of unstiffened webs decreases as web depth increases. throughout do not alter the fatigue category of the flange (already
Sometimes the minimum web thickness is dictated by the method Class ‘C’ because of the studs).
of construction such as launching, in this case thicker webs will The use of both transverse and longitudinal stiffeners is difficult to
usually be the economical solution because of local bearing, avoid on deeper girders. However every effort should be made to
buckling, and crippling considerations, as well as overall stability. place longitudinal stiffeners on one side of the web, with transverse
Changes in web thickness should coincide with either a field splice stiffeners on the other so that interferences occur only where the
or a maximum length of mill material available for the thickness longitudinal stiffener meets the double sided web stiffeners used on
and depth of web being considered. In addition, it will usually be I girders for connection of cross frames.
found satisfactory to avoid grinding or to use only nominal Fabricators have indicated that flat bars are typically more
grinding to touch up the profile of full penetration butt welds in the economical than plates for stiffeners.
web when using the submerged arc process. In many cases it will
The clear distance between longitudinal stiffeners should be no less
be found economical to maintain a constant thickness of web
than 24 inches, to accommodate automated welding equipment.

It is highly preferable not to have several stiffener sizes for a although this is not always possible e.g. a constant width may
girder. Bearing stiffener thickness that matches the flange require a flange which is locally beyond a practical thickness. If
thickness is suggested. Bearing stiffeners should be thick enough flange widths are varied, it is best to change the width at field
to preclude the need for multiple bearing stiffeners at any given splices only
bearing, as multiple stiffeners present fabrication difficulties and Top and bottom flanges should be the same width. Girders in
usually are not needed. positive bending that are composite with a slab can have a top
It is very important that the width be sufficient to provide clearance flange narrower than the bottom flange, but the weight savings
for field welding of diaphragm members to the stiffener. Four achieved are typically not worth the reduced lateral stability prior
inches or more of clearance between the web face and a vertical to hardening of the deck. Also, if continuous construction is used,
weld on a gusset plate/diaphragm member is required for good the top flange width would normally have to be increased for the
welding access. Three inches or more of clearance is needed negative moment sections, which creates slab-forming difficulties.
between a gusset plate/diaphragm member and a flange. The desirable maximum flange thickness is 3 inches. Grade 50 and
For box girders, the current trend is to longitudinally stiffen the HPS70W steels are not available in thicknesses greater than 4
webs and flanges and use internal diaphragms, without the inches. Weld time is disproportionately increased when splicing
requirement for transverse stiffeners. plates thicker than 3 inches.
A 10-foot minimum length should be used for any given flange
8.6 Flanges segment on a girder. It is only economical to introduce a flange
When deciding how to fit the flange sizes of a girder to the splice if it is possible to save about 800 – 1000 pounds, these
moment envelope, the designer must consider the cost implications numbers are approximate and are a function of the current cost of
as well as technical factors. The trade-off to be considered here is steel plate.
the cost of the material saved by reducing the flange size versus the Flange thickness increments should be 1/8 inch for thicknesses
cost of the full penetration but welded splice in the material from ¾ to 1 inch, ¼ inch from 1 to 3 inches, and ½ inch from 3 to
including; material preparation, fitting up, welding, gouging, 4 inches. A change in thickness should be made at a slope of 1 in
grinding, inspection and possibly repairs and so can involve a 2½.
considerable number of man-hours. Flange thicknesses should be sufficient to preclude the need for
In I girders and open top box girders, the designer may change the lateral bracing. Lateral bracing is to be avoided because it creates
flange width or thickness or both. It is usually more economical to fatigue-sensitive details and is costly to fabricate and install.
produce several flange splices simultaneously, this process Flange splices should be located at least 6 inches away from a web
involves butting two thicknesses of plate, wide enough to produce splice or transverse stiffener, in order to facilitate testing of the
2,4 or 6 flanges, producing one butt weld across them and then weld. Splices should be at least 10 feet apart. Field splices are good
flame cutting (stripping or ripping) the flanges longitudinally. Thus locations to change flange sizes.
by making the flanges a constant width between field splices, the
costly procedure of butt welding individual plates is avoided,

Top flanges for open box girders should follow the suggestions for welding, and the accuracy of fabrication and fit up is more critical
plate girder flanges, except for the stability criteria. Top and than with a bolted splice.
bottom flanges of closed box girders and bottom flanges of open
box girders should extend past the centerline of each web a 8.8 Fatigue Details
minimum of 2 inches to allow for automated welding equipment. Flanges with welded shear studs and a web with welded transverse
Flange width is somewhat dependent on the need for enough room stiffeners both fall into Category ‘C’. Grinding is expensive and if
inside the box girder to allow the passage of inspection personnel. carried out improperly can be detrimental to the fatigue life of the
Provision must be made for entrance to the box girder by structure. Each tension flange butt weld should be radiographed,
inspection personnel, typically a hatch-type, lockable door at each compression flange butt splices should be radio-graphed randomly
end of the box is sufficient. (form 10% to 25%) and only butt splices in webs in critical tensile
For wide bottom flanges of box girders, plate distortion during areas (e.g. 20% of the web adjacent to a tension flange) should be
fabrication and erection can be a problem. Designer should check radio-graphed. A radius should be provided at the end of the gusset
with fabricators when using bottom tension flange plates of less to eliminate a sharp notch, reduce the stiffness at the tip and
than 1” thickness in order to determine whether practical stiffness minimize longitudinal stresses at the tip of the attachment.
needs are met. In no case should bottom tension flanges be less
than ½” thick.
8.7 Field splices

The gap between girder ends should be made large enough to
accommodate normal shop tolerances. A dimension of 10mm is
commonly used, smaller values would be difficult to work to, and
unnecessarily expensive.
Designer should use one bolt diameter throughout a structure, if
practical, and ensure that it is physically possible to install bolts in
their specified locations.
When welded field splices are specified, usually it is because
aesthetics are paramount and a bolted splice is deemed
unattractive. They have several disadvantages compared with
bolted and are rarely seen. It can be difficult to detail an all welded
splice to have acceptable fatigue and fracture performance, not to
mention the problems of welding (including possible repairs),
grinding and inspecting the welds in the field. Temporary
connections are required to hold the parts in alignment during

9.0 Protecting Steel airborne chlorides e.g. near the sea coast or above a high traffic
volume expressway, and exposure to harsh industrial
Corrosion of steel is an electrochemical process. When two
environments. Weathering steel’s strength to cost ratio and ability
different metals are placed in an electrolyte, an electrical current is
to form a superior base for a paint system means that it may still be
created between the two metals. An example of an electrolyte
the chosen material even when a coating is specified.
would be salt water. In essence, one metal acts as the anode and the
other as the cathode with deterioration being the end result. This If weathering steel is used, protective measures such as drip tabs,
process of deterioration is known as galvanic corrosion. drip pans, etc… must be taken to prevent staining of the
substructure.
The corrosion of a steel structure is accelerated by:
To prevent galvanic corrosion, avoid contact between weathering
 The presence of an electrolyte like salt water, due to deicing
steel and galvanized steel, such as galvanized anchor bolts, rigid
chemicals or a marine environment, metal conduit, etc… Zinc or cadmium coated bolts should not
 The presence of carbon dioxide, come into contact with weathering steel, the thin sacrificial
 The presence of hydrogen sulfide. coatings reportedly corrode quickly when in contact with
weathering steel.
 Extremely high temperatures.
An adequate protective coating should be applied to weathering
This corrosion can cause significant loss of section to structural steel that will be embedded in soil or gravel pockets or subject to
members and impact the integrity of the structure as a whole. standing water.
There are several measures that can be taken to limit corrosion of Weathering steel is not recommended if:
steel:  The atmosphere contains concentrated corrosive industrial or
 Reduce the contact between dissimilar metals.
chemical fumes
 Galvanize the surface of the metal.
 The steel is subject to heavy salt-water spray or salt-laden
 Separate surfaces by a different material such as rubber or
fog.
paint.
 The steel is in direct contact with timber, because timber
 Use of weathering steel.
retains moisture and may have been treated with corrosive
preservatives.
9.1 Weathering Steel:
 The location has very high rainfall and humidity or there is
Atmospheric Corrosion steel (CSA G40.21 Grades 350A and
350AT) commonly referred to as weathering steel, is now the norm constant wetness.
for bridges in Canada. Painted steels are used in environments not  There is low clearance (less than 8 to 10 feet) over stagnant
considered acceptable for the weathering process, such as or slow moving water.
continued wetness due to climate and precipitation, proximity to

 The steel is used for a low urban-area bridge/overpass that process of chemical or mechanical inhibition. This inhibition is
will create a tunnel like configuration over a road on which designed to prevent deterioration caused by moisture and oxygen.
deicing salt is used. In these situations, road spray from The pigment of inhibitive primers gives coatings their color,
hardness, and corrosion resisting properties and can be either
traffic under the bridge causes salt to accumulate on the steel.
organic or inorganic.
9.2 Protective Coatings: 9.2.2 Sacrificial Primers

When using a protective coating system, corrosion of steel member Like an anode in a cathodic protection system, a sacrificial or
is controlled through use of one of three products: galvanic primer protects the underlying steel surface by creating a
 Inhibitive primers surface which is electrochemically negative in relation to the steel.
 Sacrificial primers Zinc is the most common material used to make the primer act as
an anode. The zinc is dispersed through the paint film as a pigment
 Barrier coatings
and is applied directly to the surface.
The coatings used can come in a variety of forms, the major types A limitation of inorganic, sacrificial primers is that the surface
being metallic, organic, and inorganic in composition. must receive a near white blast cleaning prior to application, and
It should be noted that while the coating material is itself the difficulty in determining which areas of the structure are the
expensive, the surface preparation and eventual application of the base metal and which have been newly primed.
protective coating is often several times the cost of that for the
basic materials. 9.2.3 Barrier Coatings
Another important aspect of applying protective coatings, is the A barrier system is designed to prevent water, oxygen, and ionic
containment and disposal of paint waste. The preparation of steel material from coming in contact with the underlying steel surface.
and the subsequent coating methods used can lead to the A barrier system is typically composed of multiple layers of
production of toxic by-products. This waste must be collected and essentially the same substance.
disposed of to the satisfaction of environmental regulations. This Types of barrier system coatings are:
can lead to severe increases in cost and in extreme cases may make  Coal-tar enamels
replacement a viable alternative to recoating.
 Low-build vinyl lacquers
9.2.1 Inhibitive Primers  Epoxy and aliphatic urethanes
A primer is the initial coat of paint which is applied to the virgin  Coal-tar epoxies
surface of a steel member. The quality of a primer is indicated by
its ability to adhere to the surface of the steel. An inhibitive primer
functions through use of a coating which stops corrosion through a

The thickness of barrier coatings is usually insufficient in White blast Excellent Excellent Excellent Very good
preventing moisture and oxygen from precipitating the cathodic
Water Good Good Poor Fair
reaction which causes deterioration. However the ionic
impermeability of the barrier systems usually offers high enough blasting
electrical resistance that the cathodic reaction is minimal. High pressure Very good Good-very Poor good
water blast good
9.2.4 Surface Preparation
The bond between coating and the base steel is of paramount Wet abrasive Very good - Very good - Excellent excellent
importance and adequate surface preparation is required in order to blasting excellent excellent
create a sound and secure bond. Proper surface preparation
increases the bonding action and also creates a uniform surface The use of power or hand tools by themselves is generally
upon which to place the coating system and thereby minimizes the insufficient and does not satisfy any of the major requirements of
amount of foreign particles which could potentially accelerate the surface preparation. The use of hand and power tools is typically
cathodic reaction such as chloride or sulfate ions. confined to the removal of heavy rust scale, loose paint etc… after
The type of surface preparation required depends on the type of which a more robust method of surface preparation is used. An
protective coating used. Inhibitive primers generally require less adverse effect of power and hand tools is their tendency to force
surface preparation than sacrificial and barrier systems. corroding back into the steel surface itself.
Dry abrasive blasting cleans the steel surface by blasting small
Table 6.1 Surface preparation methods abrasive particles at the steel surface which strips off the layers of
Cleaning Normalization of Removal of Increase in Removal
paint, rust etc... above the bare steel. Zinc based particles can also
methodology substrate interface surface area of soluble
be used which offers cathodic protection to the steel.
material salts The main advantages of water blasting, is that it is good at
removing chloride contaminants from the steel surface and that the
Hand tool Poor Poor-fair Poor-fair Poor water does not scarify (scratch) the surface of the steel as
Power tool Fair Fair Fair Poor-fair mechanical methods are prone to do.
Due to find limitations it is not always possible to completely clean
Brush-blast Fair Fair Good Poor-fair
and recoat a structure. In these cases spot cleaning is used where
Commercial Good Good Excellent Good only the deteriorated areas are prepared and then coated.
blast
Near white Very good Very good Excellent Very good

blast

10.0 Erection Considerations Plate girders can often be nested during shipping whereas the
internal diaphragms present in common box girders prevent
A well-conceived economical steel bridge requires consideration of
nesting.
its erection at two stages in the design process. Firstly, erection
must be considered at the concept stage because it typically The particular configuration of box selected has a very significant
represents about 30% of the superstructure cost and therefore the effect on the erection cost of the bridge. Unless circumstances
most economical arrangement cannot evolve without its dictate boxes larger than about 3.5 m in width should be avoided
consideration. Truss versus girder, curved versus parallel chords because they will cause excessive shipping and handling problems
and flanges, continuity, main member dimensions, drop in spans, and, in the limit, will require a longitudinal splice.
pier arrangements, etc… all have significance at this stage. Box girders having more than two webs should be avoided except
Secondly, erection must be considered at the detail stage. Details for special situations such as an axial girder cable stayed bridge.
of splices, diaphragms, bracing and pier members are very Flange width has an impact on the stability of the girder during
significant contributors to erection cost. handling and erection. According to an industry rule of thumb, I-
Those elements which are in the control of the designer should be girders will be stable if their length is less than or equal to 60 times
designed to facilitate construction wherever possible. Field labor is the flange width. If this is exceeded the erector and fabricator may
very expensive, therefore keep things simple. Realistic tolerances need to use temporary bracing to handle and erect the girder.
must be built into the system wherever shop fabricated elements
meet field construction. Access to splices, anchor bolts and Splicing:
bearings and adequate space to install jacks is very necessary for The maximum economy will result if the fabricator/erector is
proper installation, inspection and future maintenance. permitted freedom to choose the splice locations that best suit his
Constant depth or Curved chords: Strictly from an erection point of equipment. If the strength requirements of the splices are spelled
view, constant depth girders have the advantage. Pier sections of out in general terms in the drawings and specifications, then the
haunched girders frequently require extraordinary effort in Contractor can detail the bridge with his preferred splice locations
shipping, handling and turning because of their increased bulk. for the Engineer’s approval.
Constant depth girders are much easier to ship, to turn and to lift It is common to have all holes drilled or punched sub-size and then
and block. reamed to full size in full girder assembly of not fewer than three
Plate girder or box girder: From an erection point of view, the box girder sections, laid on blocking corresponding to the cambered
girders are usually preferable to the plate girders because there are shape. If this method is performed accurately, all components
fewer pieces of girder and less bracing. Box girders are reasonably should fit precisely in the field and the required bridge geometry
stable in shipping, handling and free cantilever, whereas plate will be attained. This method has the advantage of minimum time
girders, particularly if slender in flange width, can pose stability spent on field fitting and rework, as well as optimum quality in the
problems in shipping, and handling and frequently require top connection. However it should be noted that the large assemblies
chord stiffening trusses in cantilever erection. tie up a lot of shop space and reaming is very time consuming.

Diaphragms and bracing:

While angle bracing and diaphragms are very cheap to fabricate,
they are generally very expensive to erect, due to the cost of the
crane and labor for erection. In order to minimize these costs, the
designer should not use bracing and diaphragms indiscriminately,
but only were strictly necessary.

Conclusions next and also with time. In any case it is of paramount importance
to collaborate with and receive input from local fabricators when
This report has outlined the CAN/CSA S16.1 approach to the
designing plate and box girders, in order to ensure a practical and
design of doubly symmetric plate girders. The Standard gives a
economical design.
prescriptive method which was easily transferable to formatted
spreadsheets, this report expands on the clause equations to give
some explanation of their derivation and why they are applied. In References
order to expand on the applicability of the spreadsheet, equations
for analyzing the lateral-torsional buckling resistance of [1] Bridge Design Specifications, American Association of
monosymmetric plate girders was obtained from CAN/CSA-S6-06 Highway and Transportation Officials (AASHTO), 1997,
(Canadian Highway Bridge Design Code). There is a high level of 2nd Edition Washington D.C.
confidence in using the plate girder spreadsheet, as it is based
entirely on CSA standards equations and clauses.
[2] Bridge Engineering, D.E. Tonias, 1995
Design of steel box girders is based on a more fundamental
approach was required. This approach is influenced by equations
given in the Crane Code (CMAA Specifications #70 & #74), which [3] Bridge Erection – The designer’s Role’, P.R. Taylor,
allows plate elements of box girders subject to combined linearly Canadian Structural Engineering Conference, 1982
varying normal edge stress and shear stress to be checked for
buckling. A formatted spreadsheet was created for the design of [4] Bridge Design Specifications, American Association of
closed cell box girders, however unlike the spreadsheets for the Highway and Transportation Officials (AASHTO), 2004,
design of plate girders which are based on limit states design, this 3rd Edition, Washington D.C.
spreadsheet is based on allowable or working stress design. The
second spreadsheet is also developed for the design of box girders
based on CAN/CSA S16.1 Standard and the equations provided in [5] Canadian Highway Bridge Design Code (CHBDC) –
SSRC’s Guide to Stability Design Criteria for Metal Structures. To CAN/CSA-S6-00, December 2000, CSA International
ease the use of spreadsheet, a complete manual is provided in the
report. [6] Canadian Highway Bridge Design Code (CHBDC) –
This report also includes a discussion of economical and practical CAN/CSA-S6-88, June 1988, CSA International
aspects associated with the design, fabrication and erection of steel
plate and box girders. The general considerations given are true [7] Concrete Design Handbook, 3rd Edition, January 2006,
across North America, and will continue to be true for the Cement Association of Canada, Ottawa, Ontario, Canada
foreseeable future. However the reader should bear in mind that the
exact values given will vary from one geographical location to the

[8] Commentary to the Canadian Highway Bridge Design

Code – CAN/CSA-S6-00, December 2000, CSA
International
[9] Commentary to the Ontario Highway Bridge Design Code

– OHBDC-91-01, 3rd Edition, 1992, Ontario Ministry of
Transportation, Downsview, Ontario
[10] Handbook of Steel Construction, December 2005,

8th Edition, third printing, Canadian Institute of Steel
Construction, Toronto, Ontario, Canada
[11] Crane Code (CMAA Specifications #70 & #74,

revised 2000, Material Handling Industry)
[12] Limit states design in structural steel’, G.L. Kulak,

P.F. Adams, M.I. Gilmor, CISC, 4th Edition, 1990
[13] Preferred Practices for Steel Bridge Design,

Fabrication and Erection’, November, 2000, Texas Steel
Quality Council, Texas, Texas Department of
Transportation (TxDOT)
[14] Steel Bridges, Design, Fabrication, Construction,

‘Notes and References’, Canadian Institute of Steel
Construction

Appendix A: Lateral-torsional buckling of loading). g is positive (corresponding to an increase in Mu) when

the load is below the shear center.
monosymmetric beams
The general expression given in the CHBDC CAN/CSA-S6-00 for Monosymmetric I-girder:
Mu: ($10.10.2.3)
2 EI y  Cw  y
 C 2 g  C3 k 
2
M u  C1 C g  C3 k  
2  2
  Compression flange
 KL   Iy  t1 b1
GJ(KL) 2
  1  C1 S
2 EC w C e
x x d1
k =x/2
y2
C1, C2, C3 and K depend on conditions of loading and support of

t2
the beam. Two common conditions are:
b2 Tension flange
Uniform moment: C1 = 1.0 C2 = 0.0 C3 = 1.0 K = 1.0 y
Uniform load: C1 = 1.13 C2 = 0.45 C3 = 0.5 K = 1.0 Figure 1 Monosymmetric I section plate-girder
These values assume unrestrained weak axis bending and b12 t1

unrestrained warping at the ends of the member. For the values of e  d1 3  y2
b1 t1  b32 t 2
the coefficients for other conditions of loading and support, refer to
the paper by Clark and Hill (1960). Note that C1 is equivalent to
the moment factor 2. d12 b13 t1b32 t 2
Cw 
12  b13 t1  b32 t 2 
x is the coefficient of monosymmetry
e is the distance from the shear center to the centroid of the section
1
g is the distance from the shear center to the point of application of J
3
 b1t13  b 2 t 32  dw 3 
transverse load (i.e. top flange, mid-height loading, bottom flange

e  (e ' y 2 )  t 2
 I yc    I y  
2
x  0.9d1  2  1 1    
1
 I y    I x   J
3
 2b1t13  2d w w 3  b 2 t 32 
Open top box girder with sloping or vertical 2 A A 

C w   k12 f 2   k12  k 22  k1k 2  A w   4k 22  (k 2  k 3   (k 2  k 3 ) 2 fl 
2
webs: 3 2 2 
y e ' bb
k1 
bt Compression flange 2
t1
db
k 2  t  s(e ' d  t 2 )  k1
2
y1 b
d d1 k 3  (e ' d  t 2 ) 1
C
2
x x

y2 e Tension flange k x
2
t2
e' s bb s
1   b 22 2  bt 2 b12 
S x  y A
 2 f2   y 2  y A
1 fl    2y12   
y I xx   12   2 6 
Figure 2 Monosymmetric open top box girder   4cs 3 1  s   4
2

 w c 2 (y 22  y12 ) 
  3d
 2 1  2   d    2 1  
y  y 3
  1   y  y 4
     
A f 1  b1t1
b 2 ys2
Af 2  b2 t 2 C 
2 d
Aw  dw w
  b A  bb   d 2 w 
e '   b 2 d  A fl t  w  bt     1 A fl  / I yy
  2 6  2  6 

Design of Monosymmetric Plate and Box Girders

Uploaded by

Copyright:

Available Formats

Design of Monosymmetric Plate and Box Girders

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Design of Monosymmetric Plate and Box Girders

Uploaded by

Copyright:

Available Formats

What are the main differences between plate girders and box girders?

What are the main differences between plate girders and box girders?

How does the code recommend designing monosymmetric plate girders?

How does the code recommend designing monosymmetric plate girders?

Rapid Design of Monosymmetirc Plate and Box Girders

Design of Monosymmetric Plate and

By M. Khorasani & S. F. Stiemer

The spreadsheet checks for both strength and serviceability

aero-elastic problems with plate girders of I-shape was the Tacoma

Plate_Box_Girder 2011-02-05 page 1 of 55

2.0 Buckling of Plates π2E

2.1 Buckling of Unstiffened plates 2 One edge simply supported, 5.42

Plate_Box_Girder 2011-02-05 page 2 of 55

Region assumed not

When a plate is relatively short in the direction of the compressive

Plate_Box_Girder 2011-02-05 page 3 of 55

Plate_Box_Girder 2011-02-05 page 4 of 55

2. Plate clamped on four edges:

the other two edges:

Figure 6 - Buckling coefficients for combined bending and compression

The critical stress is then calculated using the buckling coefficient,

Plate_Box_Girder 2011-02-05 page 5 of 55

k  1    k '   k ''  10 1     (11)

With predominant tension replace the width of the plate, b, by 2 1k

Plate_Box_Girder 2011-02-05 page 6 of 55

2.2 Buckling of stiffened Plates

Figure 7 - Stress states in a longitudinally stiffened plate

Plate_Box_Girder 2011-02-05 page 7 of 55

Figure 8 – U nstiffened and stiffened plate gird

Class 1 b0 /t ≤ 145/√Fy h/w ≤1100/√Fy

Plate_Box_Girder 2011-02-05 page 10 of 55

Figure 10 – Effective distribution of bending stresses Position

Plate_Box_Girder 2011-02-05 page 11 of 55

large deflections and yielding. The main parameter affecting

M Stocky Intermediate Slender

Plate_Box_Girder 2011-02-05 page 12 of 55

equations given in ‘Guide to stability Design Criteria for metal

3.3.1 Unstiffened girder webs

Plate_Box_Girder 2011-02-05 page 13 of 55

k 2 E A third mode of failure due to combined shear yielding and

Figure 14 below shows equations 35 and 38 plotted on the same 290 Fy k v

The Standard also imposes a limit on slenderness:

Web slenderness h/w

Figure 14 – Shear strength versus web slenderness

Plate_Box_Girder 2011-02-05 page 14 of 55

the plate girder with transverse stiffeners. This tension field is

The capacity of the section is calculated by multiplying the

Plate_Box_Girder 2011-02-05 page 15 of 55

Fs  0.66Fy (44) 4.0

2. 439 k v / Fy   h / w   502 k v / Fy : a / h  1: k v  4.0 

Figure 17 - Web shear strength – stiffened web

Plate_Box_Girder 2011-02-05 page 16 of 55

The capacity of the section is calculated by multiplying the

3.5 Transverse stiffeners

Plate_Box_Girder 2011-02-05 page 17 of 55

67500h D = 2.4 stiffeners composed of plates placed on one side of

Plate_Box_Girder 2011-02-05 page 18 of 55

Where N= length of bearing

Plate_Box_Girder 2011-02-05 page 19 of 55

Plate_Box_Girder 2011-02-05 page 20 of 55