EJSE 200712 Published
EJSE 200712 Published
EJSE 200712 Published
net/publication/279554004
CITATIONS READS
81 4,266
1 author:
SEE PROFILE
All content following this page was uploaded by Ezzeldin Yazeed Sayed-Ahmed on 27 July 2015.
ABSTRACT: Corrugated web girders represent a new structural system emerged in the past two decades. The
girder’s flanges provide the flexural strength of the girder with no contribution from the corrugated web
which provides the girder’s shear capacity. Failure of the web occurs by steel yielding, web buckling or inter-
actively between them. Lateral torsion and local flange buckling of corrugated web girders represent another
two possible failure criteria. Here, the work previously performed by the author on corrugated web girders
was compiled and presented in a comprehensive format. The starting point is the shear behaviour of the cor-
rugated webs which is investigated focusing on the failure modes affecting the web design. An interaction
equation that considers web buckling and yielding is proposed. Numerical analyses are performed to investi-
gate the buckling modes of the corrugated web, verify the validity of the proposed equation and explore the
post-buckling strength of corrugated web girders. The numerical model is extended to determine the critical
moment causing lateral instability for corrugated web girders. The applicability of the critical moment design
equations, currently used for plane web girders, to corrugated web girders is examined. The numerical model
is then used to scrutinize the local buckling behaviour of the compression flange. The applicability of the cur-
rently used limiting values for the flange outstand-to-thickness ratios to corrugated web girders is investi-
gated.
27
Electronic Journal of Structural Engineering, 7(2007)
sibility of failure (El-Metwally 1999, Sayed-Ahmed fied using the numerical analysis. The validity of the
2001, 2005a). equivalent moment factor concept established earlier
Lateral torsion buckling is another possible fail- for plane web girders is verified for corrugated web
ure mode for steel I-girders with corrugated web. girders via the numerical analysis results. The nu-
Codes of practice commonly use the critical moment merical model is also used to investigate the local
of a simply supported I-girder with a plane web sub- buckling behaviour of the compression flange for
ject to a constant moment and relate it to the critical these girders. The applicability of the limits defining
moment of any other loading case via an “equivalent the section class for girders with plane web to corru-
moment factor”. Can the same concept still be used gated web girders is examined.
for corrugated web girders?
Local buckling of the compression flange of I-
girders with corrugated webs is another criterion 2 FAILURE MODES OF THE CORRUGATED
which affects design and strength of these girders. STEEL WEBS
Generally, local buckling of the compression flange
of an I-section mainly depends on the flange out- The corrugated steel web solely provides the shear
stand-to-thickness ratio. Limits are placed on this ra- capacity of the girder. The shear strength is thus,
tio such that the critical stress initiating local flange controlled by buckling and/or yielding of the web
buckling will not be reached before reaching the (Sayed-Ahmed 2001, Luo & Edlund 1994, Elgaaly
yield stress. Will the same limits adopted for steel et al. 1996, Bergfelt & Leiva-Aravena 1984, Leiva-
girders with plane webs still be applicable to steel Aravena 1987, El-Metwally & Loov 1999). The only
girders with corrugated webs? And what would be significant stress which appears in the corrugated
the value of the flange outstand for corrugated web web is pure shear stress.
girders?
This paper brings together the author previous in-
2.1 Steel Yielding of the Corrugated Web
vestigations [Sayed-Ahmed 2001, 2003b, 2005a] in
a comprehensive format. The different buckling The shear stress which causes an element of a corru-
modes of corrugated steel webs which may be en- gated web to yield is defined by:
countered for these plates are presented and the in-
Fy
teraction between these modes is inspected. The in- τy = (1)
teraction between the yield failure criterion and the 3
buckling modes is investigated. An interaction equa-
tion which considers the different corrugated web where Fy is the yield strength of the steel.
failure criteria is proposed. A linear finite element
model is adopted to investigate the buckling modes 2.2 Stability of the Corrugated Web
of the corrugated web plates. Then, a nonlinear
model which considers both the geometric and the Two buckling modes are associated with corrugated
material nonlinearities is employed to verify the va- steel webs: local buckling and overall (global) buck-
lidity of the proposed interaction equation and to in- ling. The local buckling mode corresponds to the in-
vestigate the post-buckling strength (if any) of the stability of a panel simply supported between two
corrugated web girders. folds. The corrugated web, in this mode of failure,
Lateral torsion and local flange buckling of I- acts as a series of flat panels that mutually support
girders with corrugated steel webs are also numeri- each other along their vertical (longer) edges. The
cally investigated. A finite element model is adopted panels are supported by the flanges along their hori-
to determine the critical moment initiating lateral zontal (shorter) edges. Local buckling is investigated
torsion buckling of corrugated web girders and com- using equations derived for isotropic plates; an esti-
pare it to that of girders with plane webs. Corrugated mate for the elastic critical shear stress τcr,l for the
web girders subjected to constant moments are first local buckling mode is given by (Galambos 1998):
scrutinized. Then, the analysis is extended to inves- 2
π 2 ⋅E ⎛t ⎞
tigate the critical moments for girders subjected to τ cr ,l = k s ⋅ 2
⋅⎜ w ⎟ (2)
end moments causing moment gradient through the 12 ⋅ ( 1 − υ ) ⎝ b ⎠
girder’s span. Other corrugated web girders sub-
2
jected to central concentrated loads at, and away ⎛ b ⎞
from the shear centre are also analyzed. In evaluat- k s = 5.34 + 4.0 ⎜⎜ ⎟⎟ (3)
ing the critical moment for a corrugated web girder, ⎝ hw ⎠
a section with a plane web having a proposed equiv- where tw is the corrugated web thickness, b is the
alent thickness is suggested to handle the critical panel width, E and v are Young’s modulus and Pois-
moment calculations: an assumption which is veri- son’s ratio for the steel respectively and ks is a shear
28
Electronic Journal of Structural Engineering, 7(2007)
buckling coefficient for the local buckling mode. for τ cr ,l > 0.8τ y : τ cr ,in ,l = 0.8τ cr ,lτ y where τ cr ,in ,l ≤ τ y
(7)
The shear buckling coefficient defined by Equation for τ cr , g > 0.8τ y : τ cr ,in , g = 0.8τ cr , gτ y where τ cr ,in , g ≤ τ y
3 is applicable when all the sides of the panels are
simply supported which simulates steel girders with To calculate the critical stress for an inelastic inter-
corrugated webs. In Equation 2, if the “inclined” active buckling mode, τcr,in,l and τcr,in,g are used in
panel width a is larger than the “horizontal” panel Equation 6 instead of τcr,l and τcr,g respectively.
width b, it should be considered as the critical width. Another interaction equation which includes all
Global (overall) buckling is characterized by di- the failure criteria has been proposed [Sayed-Ahmed
agonal buckling over several corrugation panels. 2001, 2005a, El- Metwally & Loov 1999). The new
The critical shear stress for this mode is estimated equation takes the form:
by considering the corrugated web as an orthotropic n n n n
plate. The critical shear stress of this mode τcr,g is ⎛ 1 ⎞ ⎛ ⎞ ⎛ 1 ⎞ ⎛ 1 ⎞
⎜ ⎟ =⎜ 1 ⎟ +⎜ ⎟ +⎜ ⎟ (8)
defined by (Galambos 1998): ⎜τ ⎟ ⎜τ ⎟ ⎜τ ⎟ ⎜τ y ⎟
⎝ cr ,i ⎠ ⎝ cr ,l ⎠ ⎝ cr ,g ⎠ ⎝ ⎠
τ cr ,g = k g ⋅
(D y ⋅ D x3 )
1/ 4
D y = ⎜ ⎟⎜⎜ w
⎟⎟ = ⎜ ⎟⎜ w
⎟⎟ τcr,i Fy = 350 MPa
⎝ s ⎠⎝ 12 ⎠ ⎝ b + d / cos α ⎠⎝ 12 ⎠ 0.8
hw = 2000 mm
tw = 4 mm
where Ix is the second moment of area of one “wave- 0.6
hw /tw = 500
length” of the web having a projected length c and
β =a/b =1.0
an actual length s, tw is the web thickness, b is the 0.4
α = 37o
panel width, d is the horizontal projection of the in-
clined panel width and α is the corrugation angle. 0.2
0
2.3 Interaction between Failure Modes 0 100 200 300 400 500 600 700 800
Panel width (mm)
The following equation is currently adopted to ac- Figure 2. Critical shear stresses for trapezoidal corrugated
count for the interaction between the two buckling web plates with hw/tw = 500 (hw =2000 mm and tw= 4 mm).
modes described earlier (Elgaaly et al. 1996, John-
son & Cafolla 1997a):
τcr,g τcr,l
1 1 1 1 τy
= + (6)
Critical stress / Yield stress
τ cr ,i τ cr ,l τ cr ,g Fy = 350 MPa
0.8 τcr,i
hw = 1000 mm
where τcr,i is the critical stress due to the interaction tw = 2 mm
between local and global buckling modes. Equation 0.6
hw /tw = 500
6 does not consider the steel yielding criterion or its β =a/b =1.0
interaction with the buckling failure criteria. Fur- 0.4
α = 37o
thermore, Equations 2 and 4 defining τcr,l and τcr,g re-
0.2
spectively do not account for inelastic buckling
which may occur if the critical shear stress of any
mode exceeds 0.8·τy. To overcome these defects, the 0
0 100 200 300 400 500 600 700 800
following equation was proposed (Elgaaly et al. Panel width (mm)
1996, Galambos 1998) for the inelastic critical stress Figure 3. Critical shear stresses for trapezoidal corrugated
τcr,in in both the local and global buckling modes: web plates with hw/tw = 500 (hw=1000 mm and tw=2mm).
29
Electronic Journal of Structural Engineering, 7(2007)
Stress
geometric characteristics of the analyzed girders are Linear elastic finite
listed in Table 1. element model
Fy=350 Et=0.001 E
Table 1. Numerically analyzed corrugated web girders.
MPa
Girder G5
G1 G2 G3 G4
No.
b (mm) 400 300 200 100 20 Non linear finite
tw (mm) 2 2 2 2 2 element model
hw (mm) 500 500 500 500 500 E=200 GPa
b/ hw 0.8 0.6 0.4 0.2 0.04
hw /tw 250 250 250 250 250 Strain Strain
bfl (mm) 300 300 300 150 50
tfl (mm) 20 20 20 20 20 Stress-strain relations ν=0.3 & E=200 GPa
G=77 GPa
L (span) 10.48 7.86 5.24 4.06 2.252
Figure 4. Geometry and loading configuration of the analyzed
The geometry and loading configurations of these (above); and material model adopted in the analysis (below).
girders are shown in Figure 4 together with the steel different element sizes. The typical meshes shown in
uniaxial stress-strain relation adopted in the nonlin- Figure 5 proved to be mesh-independent.
ear analysis. Stiffener plates 20 mm thick were The critical load at which web buckling was first
added at the loading and the support locations for all encountered was determined in both the Eigenvalue
the analyzed girders. The girders were simply sup- and the nonlinear analyses. The results of the nu-
ported and subjected to mid-span loads (line loads merical analyses are plotted in Figure 6 where the
acting on the top flanges as shown in Figure 5). ratio between shear force Vcr at which web buckling
Thus, the shear force acting on the web V is: initiated (obtained numerically and analytically) to
V = τ ⋅ hw ⋅ t w (9) the shear force causing steel yielding Vy (τy hw tw) is
plotted versus the panel width b. Figure 6 shows that
where τ is the shear stress acting on the web. Due to the results of the numerical analyses are in a good
symmetry, only one half of the girder was analyzed agreement with the behaviour theoretically predicted
with a plane of symmetry located at the girder cen- by the proposed interaction equation (Eq. 8). Fur-
tre-line. To eliminate the mesh sensitivity effect, the thermore, the results of the numerical analyses re-
finite element analyses were first performed using veal that Equations 6 and 7 are conservative com-
pared with the proposed interaction equation. Local,
30
Electronic Journal of Structural Engineering, 7(2007)
Vcr / Vy
Plane of
Yield
Symme-
1
try Inter.
buckling Linear Elastic FE model
Line of support
Eqs. 1 to 8
Girder G2 Eqs 6&7
L=7.86 m 0.5
Plane of
Symme- 0
t 0 50 100 150 200 250 300 350 400 450
Panel width (mm)
Line of support
Fy = 350 MPa
2
Global hw = 500 mm
Girder G3 buckling tw = 2 mm
L=5.24 m hw /tw = 250
Local
β =a/b =1.0
buckling
1.5 α = 37o
Plane of
Symme-
Vcr / Vy
t Yield
1
31
Electronic Journal of Structural Engineering, 7(2007)
V / Vy
0.6
mm panel width analyzed using mesh 1
0.4
Vcr/Vy
Global buckling of Girder G5 0.4
having a 20 mm panel width.
0
3 LATERAL AND LOCAL FLANGE 0 50 100 150 200 250 300 350 400 450
BUCKLING Panel width (mm)
Figure 9. Post-buckling strength of the analyzed corrugated
3.1 Lateral Torsion Buckling of Steel Girders with web girders.
Plane Webs ing constant of the beam’s cross section respec-
The critical moment initiating lateral instability for a tively.
simply supported I-girder is: For beams subjected to unequal end moments
(MA and MB) with no other loads through the beam’s
Cb π
M cr = Cb ⋅ M ocr = E I y G J (1 + WR2 ) (10) span, the equivalent moment factor Cb may be given
L by (Galambos 1998):
MA M
π EC w Cb = 1.75 + 1.05 + 0 .3 ( A ) 2 ≤ 2 .3 (12)
WR = (11) MB MB
L GJ
where Mocr is the critical moment for a beam sub- where MA is the smaller end moment. MA/MB is posi-
jected to constant bending, Cb is the equivalent mo- tive for beams bent in double curvature and negative
ment factor, E is the Young’s modulus, G is the elas- for beams bent in single curvature. CAN/CSA-
tic shear modulus, and L is the beam span. Iy, J and S16.1-02 specifications (CSA 2002) adopted Equa-
Cw are the second moment of area about the weak tion 12 in design with a change in its limiting value
axis of inertia, the torsional constant, and the warp- from 2.3 to 2.5. The same specifications use Cb of
32
Electronic Journal of Structural Engineering, 7(2007)
1.0 when the moment between the end supports be- even develop the yield moment then it is classified
comes greater than the end moment or when there is as a Class 4 slender section. According to the AISC-
no lateral support for the compression flange at one LRFD (AISC 2003), these limits are defined by:
of the ends of the unsupported length. The AISC- Compact sections :
LRFD specifications (AISC 2003) define the follow-
ing equation for the equivalent moment factor: b fl E hw E
≤ 0.38 ≤ 3.76
2t fl Fy tw Fy
12.5 ⋅ M max
Cb = (13) (16)
3 ⋅ M 1 + 4 ⋅ M 2 + 3 ⋅ M 3 + 2.5 ⋅ M max Non - compact sections :
33
Electronic Journal of Structural Engineering, 7(2007)
buckling 300analysis. The buckling analysis was per- Cw (109) mm3 5625 5625
formed on girders having the b x twcross
200x2 400x2
sections shown 3
20
in Figure 10. The span of all the analyzed girders J (10 ) mm
3
1602 1614
*
was 11.52 m. Two panel widths b x tw for
200x4 400x4
the corrugated section is considered as a double symmetric I-section with an
web were adopted in the analysis: 200 mm and 400 equivalent thickness teq=s/c tw
mm, which correspond 520 to panel width a = to
b web height Elastic mod- Shear mod-
Poisson’s
Uniaxial yield
t ulus ulus strength
ratios of 0.42 and 0.83. Web thicknesses
w
β= a/b = of1.0 2 mm ratio ν
E (GPa) G (GPa) Fy (MPa)
and 4 mm were considered in thedanalysis, = a cos α which
200 76.9 0.3 350
correspond to web height-to-thickness tan α of 120
h = dratios
and 240. The analyzed corrugationc panel = b + dwidth-to-
1
thickness ratio ranges between 50 and
Sec. 1-1 s = b200.
+ d/cos α in-
The The theoretical critical moments of corrugated
clined panel of the corrugated web teqwas assumed
= (s/c) tw to web girders subjected to constant bending moments
have the same width as the horizontal panel (a=b): a were determined via Equation 10 to be 526.1 kN.m
typical practical geometric1 configurationStiff for Pl.
the tra- and 527.4 kN.m for corrugated web thicknesses of 2
pezoidal corrugation
L = 11.52 m profile of the web. Stiffener mm and 4 mm respectively. It is clear that, theoreti-
plates 20 mm thick were added at the support loca- cally, neither the web thickness nor the corrugation
tions for all the analyzed girders hto prevent bearing panel width of the web would have significant effect
failure of the webs at these α locations. The flange on the value of the critical moment. In these calcula-
width and thickness (bfl/2tfl = 15) as well as the web tions, the girders were assumed to have plane webs
height were kept bconstantd through the analysis. The with an equivalent thickness [13] which is given by:
Figure 10. Geometric
mechanical properties
and material of corrugated
properties of thewebgirders’
girders
numerically analyzed for lateral torsion buckling.
cross-section are listed in Table 2. t eq = t w ⋅ s / c (17)
where tw is the corrugated web thickness, s is the
Table 2. Mechanical and material properties of the analyzed
corrugated web girders*. corrugated wave length and c is the projected length
of one corrugation wave (Figure 1).
Property tw =2 mm tw =4 mm
Figure 11 show two typical meshes used to model
3 2
A (10 ) mm 13.07 14.13 two girders having corrugated panel widths of 400
6
Ix (10 ) mm 4
770.5 791
mm and 200 mm, respectively. The girders were
6 4
simply supported in flexure and in torsion: Uy, Uz
Iy (10 ) mm 90 90 and Rx at both ends and Ux at one of the two ends
20 mm end
rx mm 243 236 were restrained. Figure 12 shows the loading con-
plate
figuration adopted in the No of elements: 5261
analyses. For girders sub-
ry mm 83 79.8 No. of nodes: 16308
jected to end moments, the moments were applied as
Sx (103) mm3 DOFs: 97848
2963 3042 pressure and tension on the top and bottom flanges
Sy (10 ) mm
3 3
600 600
respectively at the two ends of the analyzed girder.
For girders subjected to concentrated loads, a point
3
Zx (10 ) mm
3
3127 3255 load was
y applied at the mid-span of the girder either
Zy (103) mm3 900 900
on the top flange, from the bottom flange or in the
mid-height of the corrugated web.
z x
z x
Figure 11. FE mesh for the 400 mm (above) and the 200
mm (below) corrugated web panel width girder.
34
Electronic Journal of Structural Engineering, 7(2007)
Mcr-FE /
1.202 1.19 1.12
Mcr-Eq 11*
3.3.2 Results of the Numerical Analysis *
Based on Cb from Equation 14.
Corrugated web girders subjected to end mo-
ments or central concentrated loads (Tables 3 and 4)
The numerically obtained critical moments Mcr-FE
were numerically analyzed (Sayed-Ahmed 2003a,
were also related to the critical moment of Girder G6
2005b). Results of the numerical analysis are listed
subjected to constant flexure through the span. Thus,
in Tables 3 and 4 and plotted in Figures 13 and 14.
equivalent moment factors for all the numerically
M M M analyzed corrugated web girders were evaluated:
those are listed in Tables 3 and 4 and plotted in Fig-
Girder G6 Girder G8 ures 13 and 14 versus the equivalent moment factors
M M/2 M M/2 analytically calculated using the currently adopted
equations (Equations 12 to 14).
Girder G7 Girder G9 It is evident from Tables 3 and 4 and Figures 13
M M and 14 that resistance to lateral torsion buckling of
girders with corrugated webs is different from that
Girder G10 of plate girders with traditional plane webs. The
P P P P critical moment initiating lateral buckling for corru-
gated web girders is larger than that of traditional
girders. The numerical analysis reveals that Mcr for
Girder G11 Girder G12
corrugated web girders subject to end moments is
P 15% to 37% higher than Mcr for plane web girders.
This ratio becomes 12% to 20% for corrugated web
Girder G13 girders subject to central concentrated loads.
P The numerical analysis also reveal that the con-
Figure 12. Loading configuration adopted in the FE analysis. cept of the equivalent moment factor which is cur-
rently used for I-girders with plane web to account
The “analytical” values of the critical moments Mcr- for the moment gradient is applicable to girders with
Eq.11 were calculated using Equation 11 with the corrugated webs. Furthermore, the analysis results
equivalent moment factor Cb evaluated using Equa- show that Equations 12 and 13 which are currently
tion 12 for girders subjected to end moments or used to calculate the equivalent moment factor are
Equation 14 for girders subjected to concentrated also applicable to steel girders with corrugated webs.
loads. The theoretically calculated critical moments Figure 13 shows that in case of girders subjected
Mcr-Eq.11 were compared to the critical moment re- to end moments, Equation 12 with the CAN/CSA-
sulting from the numerical analysis Mcr-FE. S16.1-02 limiting value provides a good match to
Table 3. Results of the finite element buckling analysis for cor- the results of the numerical analysis.
rugated steel web girders subjected to end moments.
The importance of the location of the applied load
Panel Girder G6 G7 G8 G9 G10
with respect to the shear centre of the girder’s cross-
dim. Cb-Eq.12 1 1.3 1.75 2.3 2.3 section is emphasized through the results plotted in
Mcr-FE
1523* 1661* Figure 14 and listed Table 4. It is clear from these
b = 200 mm
(kN.m)
Cb-FE 1 1.31 1.8 2.45 2.62
Mcr-FE /
Mcr-Eq 11
1.204 1.219 1.24 1.285 1.372 3.4 Local Buckling of Corrugated Web girders’
Compression Flanges
*Lateralbuckling is preceded by flange and/or web local buck-
ling mode.
The flange outstand for girders with plane webs was
Table 4. Results of the finite element buckling analysis for cor- commonly approximately as bfl/2. For corrugated
rugated steel web girders subjected to central loads. web girders at a section where the web is parallel to
Girder G11 G12 G13 the axis of the girder, there is a large outstand on one
Panel
dim. Cb-Eq.13 1.31 1.31 1.31 side and a small outstand on the other (Figure 15).
Cb-Eq.14 0.96 1.35 1.9 Thus, the flange outstand which should be used to
Mcr-FE 609 848 1122 classify the cross section class of corrugated web
tw =
mm
35
Electronic Journal of Structural Engineering, 7(2007)
2.75 1.70
Cb Eq. 12 Cb-F.E.
2.50 CAN/CSA 1.60
MA MB
S16.1
2.25 Cb-Eq. 12 1.50
Mcr-FE / Mcr-Eq. 11
bxt w =400 mm x 4mm
Cb 2.00 1.40
1.75 1.30
Mcr-FE / Mcr-Eq. 11
1.50 1.20
1.25 1.10
1.00 1.00
-1.00 -0.50 0.00 0.50 1.00
MA / MB
2.75 1.70
Cb-F.E.
Cb Eq. 12
2.50 MA MB 1.60
CAN/CSA
2.25 S16.1 Cb-Eq. 12 1.50
Mcr-FE / Mcr-Eq. 11
bxt w =200 mm x 2 mm
2.00 1.40
Cb
1.75 1.30
Mcr-FE / Mcr-Eq. 11
1.50 1.20
1.25 1.10
1.00 1.00
-1.00 -0.50 0.00 0.50 1.00
MA / MB
Figure 13. Equivalent moment factor and the critical moments for girders with corrugated webs subjected to end moments.
2.00 1.25
P
Cb-Eq. 14
1.80
bxt w = 200 mm x 4 mm
1.60 1.20
Mcr-FE / Mcr-Eq. 11
Cb - F.E.
Mcr-FE/Mcr-Eq. 11
1.40
Cb
1.20 1.15
Cb - CAN/CSA-S16.1
1.00
Below the shear centre Above the shear centre
0.80 1.10
-0.50 -0.25 0.00 0.25 0.50
Load distance from the shear centre / cross section height
Figure 14. Equivalent moment factor for girders with corrugated webs subjected to central concentrated loads.
the average flange outstand Lav (Fig. 15). The aver- greater than 0.14, which is more practical, it is rec-
age flange outstand corresponds to bfl/2 which is tra- ommended to be conservative using the large flange
ditionally used for girders with plane webs. outstand Ll. Thus, a considerable uncertainty still ex-
Based on a previous investigation (Johnson & ists regarding the correct value which should be used
Cafolla 1997b), it was argued that the average flange for the flange outstand of corrugated web girders.
outstand Lav may only be used as bfl/2 if a ratio R is
less that 0.14 where R is the ratio between area
EFGH and area ABCD defined in Figure 15. For R
36
Electronic Journal of Structural Engineering, 7(2007)
B b d C
300 Figure 15. Flange outstands and geometric properties of the corrugated web
girders analyzed for local flange buckling.
CAN/CSA-S16.1 limit
3.00 Flange outstand based on
Compact
*
AISC limit average outstand
Non-
2.50
Non-Compact
2.00
Ncr / Ny
1.50
Ncr/Ny = 1.0
1.00
b/t w = 50
0.50
b/tw = 100
0.00
0.00 10.00 20.00 30.00 40.00
bfl / 2 t fl (based on Lav)
CAN/CSA-S16.1 limit
3.00 *
AISC limit
Non-Compact
2.50
Flange outstand based on
Non-Compact
1.50
N cr/N y = 1.0 *
1.00 The AISC-LRFD limit
b/t w = 50 on bfl/2tfl is calculated
0.50 without considering the
b/t w = 100
residual stresses
0.00
0.00 10.00 20.00 30.00 40.00 50.00
bfl / 2t fl (based on Ll)
Figure 16. Local buckling of corrugated web girders’ flanges and current compact section limits.
37
Electronic Journal of Structural Engineering, 7(2007)
average flange outstand or 10.5 and 52.5 based on tio is more appropriate for classifying a corrugated
the large flange outstand. The finite element meshes web girder’s cross section according to the AISC-
used to model the girders is similar to the one shown LRFD provisions: Ncr/Ny > 1.0 for bfl/2tfl < 19.8 only
in Figure 12 for the 200 mm panel width girder. when bfl/2 is based on the large flange outstand.
The AISC-LRFD assumes that the web of a plane
3.4.2 Results of the Numerical Analysis web I-girder provide a support condition for the
The critical compressive force Ncr acting on the flange outstand (Fig. 17) which lies somewhere be-
upper flange at which local flange buckling occurred tween a fixed support (k=1.277) and a simple sup-
was numerically determined. This force was normal- port (k=0.425). The buckling coefficient k was as-
ized with respect to the yield strength of the flange sumed for this hypothetical case of the AISC-LRFD
Ny where Ny = bfl x tfl x Fy and plotted versus the av- to be 0.7 hence, the limit on bfl/2tfl given in Equation
erage flange outstand-to-thickness and the large 16. The critical loads causing local flange buckling
flange outstand-to-thickness ratios in Figure 16. The for a corrugated web girder flange, obtained from
non-compact section limits according to the AISC- the numerical analysis, were plotted in Figure 18
LRFD and the CAN/CSA-S16.1-02 were also plot- versus the flange large outstand-to-thickness ratio.
ted in Figure 16. The AISC-LRFD limit on bfl/2tfl in The basic plate buckling equation with the three
this figure does not consider the residual stress (69 buckling coefficients described above were also
MPa for rolled section and 114 MPa for welded sec- plotted in Figure 18. The figure confirms that the
tions) as the yield strength of the compression flange flange outstand-to-thickness ratio should be based
Ny was based on the full yield stress Fy. on the large outstand for girders with corrugated
Figure 16 reveals that using the average flange webs (Sayed-Ahmed 2005b). Figure 18 also reveals
outstand-to-thickness ratio to classify the section of that the assumed value of k=0.7 which is adopted by
a corrugated web girder’ is acceptable according to the AISC-LRFD specifications for plane web girders
the restricted limits of the CAN/CSA-S16.1-02 is equally applicable to girders with corrugated web
specifications: Ncr/Ny > 1.0 for bfl/2tfl < 10.7 with provided that the large outstand is used for bfl/2.
bfl/2 based on the average flange outstand. On the
other hand, the large flange outstand-to-thickness ra-
4 CONCLUSIONS
S.S. edge
Free edge
Free edge
Free edge
38
Electronic Journal of Structural Engineering, 7(2007)
4.00
N cr /N y (k=0.7)
Non-Compact
3.50
AISC limit
3.00
N cr /N y (k=0.425) b/t w = 50
2.50 b/t w = 100
Ncr / Ny
2.00
0.50
0.00
0.00 10.00 20.00 30.00 40.00
bfl / 2 t fl (based on Ll)
Figure 18. Local buckling of the corrugated web girders’ flanges and the assumed buckling coefficients of the AISC-LRFD.
Combault, J. 1988. The Maupré Viaduct near Charolles,
of such girders is 12% to 37% higher than that of France, Proceedings of the NEC/COP National Steel Con-
plate girders with plane webs. Thus, the equations struction Conference, Miami Beach, USA, pp.12.1-12.22.
used to calculate the critical moment of girders with Combault, J, Lebon, J.D. and Pei, G. 1993. Box girders using
plane webs would underestimate the capacity of Corrugated Steel Webs and Balanced Cantilever Construc-
plate girders with corrugated web to resist lateral tion. Proc. of the FIP Symposium, Kyoto, pp. 417-424.
Elgaaly, M., Hamilton, R.W. and Seshadri, A. 1996. Shear
buckling but they are conservative for design pur- strength of beams with corrugated webs, Journal of Struc-
poses. Furthermore, it was concluded that the equiv- tural Engineering, ASCE, Vol. 122, No. 4, pp. 390-398.
alent moment factor concept which is used for the El-Metwally, A.S. 1999. Prestressed Composite Girders with
plate girders with plane web is equally applicable to Corrugated Webs. MSc Thesis, Department of Civil Eng.
plate girders with corrugated webs. Thus, all the eq- The University of Calgary, Calgary, Alberta, Canada, 189p.
uations and tables which are currently used to de- El-Metwally, A.S. and Loov, R.E. 1999. Composite
prestressed concrete beams with corrugated webs, Proc. of
termine the equivalent moment factor for plane web the Annual Conference of the Canadian Society for Civil
girders may be used for corrugated web girders. Engineering, Regina, Saskatchewan, Canada, Vol. 1, pp.
The numerical model was used to investigate the 305-314.
local buckling of the compression flange of corru- Galambos, T.V. 1998. Guide to stability design criteria for
gated web girders. It is concluded that the flange metal structures, 5th ed., John Wiley and Sons, NY, USA.
Lebon, J. 1998. Steel corrugated web bridges - first achieve-
outstand-to-thickness ratio, which is currently used ments, Proc. of the 5th Int. Conf. on Short and Medium
by codes of practice as one of the criteria classifying Span Bridges, CSCE, Calgary, Canada, CD-Proceedings.
the section compactness, should be based on the Leiva-Aravena, L. 1987. Trapezoidally corrugated panels -
large outstand of the corrugated web girder’s flange. buckling behaviour under axial compression and shear, Di-
vision of steel and Timber Structures, Chalmers University
of Technology, Gothenburg, Report S84:2, Sweden, Publi-
cation S 87:1.
5 REFERENCES Luo, R. and Edlund, B. 1994. Buckling of trapezoidally corru-
gated panels using spline finite strip method, Thin Walled
American Institute of Steel Construction (AISC). 2003. Man- Structures, Elsevier Science Limited, Vol . 18, pp. 209-240.
ual of steel construction: Load and Resistance Factor De- Johnson, R.P. and Cafolla, J. 1997a. Corrugated webs in plate
sign (LRFD). 3rd edition, Illinois, USA. girders for bridges, Structures and Buildings, ICE, Vol 123,
Bergfelt, A. and Leiva-Aravena, L. 1984. Shear buckling of pp.157-164.
trapezoidal corrugated girder webs, Division of steel and Johnson, R.P. and Cafolla, J. 1997b. Local flange buckling in
Timber Structures, Chalmers University of Technology, plate girders with corrugated webs, Structures and Build-
Gothenburg, Publication S 84:2, Sweden, 64p. ings, ICE, Vol 123, pp. 148-156.
CAN/CSA-S16.1-02. 2002. Limit states design of steel struc- Reinhard, J.M. 1994. Pont de la Corniche, Ouvrages d’Art No.
tures. Canadian Standard Association, Ontario, Canada. 19, pp. 14-19 (in-French).
Capra, A. and Leville, A. 1996. The bridge at Dole, Proceed- Sayed-Ahmed, E.Y. 2001. Behaviour of steel and/or composite
ings of the FIP Symposium, Post-Tensioned Concrete girders with corrugated steel webs, Canadian Journal of
Structures. London, Vol. 1, pp. 135-141. Civil Engineering, vol. 28, No. 4, pp. 656-672.
Chen, W., and Lui, E. 1987. Structural stability: theory and Sayed-Ahmed, E.Y. 2003a. Lateral Stability of Plate Girders
implementation. Elsevier Science Publishing Co., NY, with Corrugated Steel Webs. Proceedings, 31st Canadian
USA. Society for Civil Engineering Conference: Building our Ci-
Cheyrezy, M. and Combault, J. 1990. Composite bridges with vilization, Moncton, New-Brunswick, Canada. June 2003.
corrugated steel webs - achievement and prospects, IABSE GCF231-1 to GCF231-10.
Symposium, Mixed Structures: Including New Materials, Sayed-Ahmed, E.Y. 2003b. Girders with Corrugated Steel
IABSE Reports, Brussels, pp.479-484. Webs: Buckling Modes and Numerical Modeling. Proceed-
ings, Advances in Structures: Steel, Concrete, Composite
and Aluminum, Sydney, Australia, June 2003, pp. 807-812.
39
Electronic Journal of Structural Engineering, 7(2007)
40
View publication stats