Chen (2011) PDF
Chen (2011) PDF
Chen (2011) PDF
com
ABSTRACT
Thin walled steel members, such as cold-formed steel (CFS) members, are susceptible for local
buckling at low loads. However, initiation of local buckling of elements does not necessarily mean the
ultimate limit state of the member, and thin-walled plate elements can exhibit substantial post-buckling
strength. International cold-formed steel design specifications recognize the post-buckling strength of
uniformly compressed stiffened and unstiffened elements, and the post-buckling strength of stiffened and
unstiffened webs in flexure. However, potential post-buckling shear strength of webs is not considered in
CFS design specifications, though post-buckling shear strength of stiffened webs is considered in
structural steel plate girder design. The objective of this study is to quantify the post-buckling shear
strength of thin-walled cold-formed steel members. This numerical investigation is based on finite
element method.
This paper presents the finite element modelling details associated with the thick/thin plates
representing the web of a cold-formed steel member. The study considered simply-supported rectangular
plates subjected to in-plane shear loadings experiencing buckling, post-buckling, and yielding until failure.
The plate was modelled using geometrically non-linear quadrilateral shell elements, and non-linear steel
stress-strain relationship derived from experiments. Total Langrangian with large displacement/small
strain formulation was used for such analyses. The model also considered the initial geometric
imperfections. The first part of the investigation established the ultimate shear strength of plates having
different parametric dimensions. The next part compared such results with the strength values based on
the current code provisions. For slender plates, the current code uses the shear buckling load as the
strength, whereas the analyses indicated significant post-buckling strength. The third part of this paper
establishes the shear design equations incorporating the post-buckling strength of plates.
Keywords: Cold-formed steel, thin-walled, finite element analysis, post-buckling, shear strength
a
Corresponding author: Email: siva@mcmaster.ca
b
Presenter: Email: siva@mcmaster.ca
1877–7058 © 2011 Published by Elsevier Ltd. Open access under CC BY-NC-ND license.
doi:10.1016/j.proeng.2011.07.081
642 BO CHEN and K.S. SIVAKUMARAN / Procedia Engineering 14 (2011) 641–647
1. INTRODUCTION
Thin walled steel members, such as cold-formed steel (CFS) members, are susceptible for local
buckling at low loads. Figure 1 indicates a plate subjected to pure shear loads along the edges,
representing a web of a steel bending member. At the initial loading stage, prior to shear buckling when
the plate is perfectly flat, equal tensile and compressive principal stresses are developed. As the applied
shear loads increase, the compressive and tensile stresses within the plate increase equally until the shear
buckling load is reached. In plated structural elements, initiation of local buckling of elements does not
necessarily mean the ultimate state of member, and thin-walled plate elements can exhibit substantial
post-buckling strength. Beyond shear buckling, plates cannot take any additional compressive stresses.
However, any additional shear loads are then resisted by the tension field action (tensile membrane stress)
only. The edge zones of the plate serve as an anchor to the diagonal tension field. Therefore, the rigidity
of edge members greatly influences the magnitude of the post-buckling strength of the plate. If the edge
elements are flexible, the edge member will bend inward, plate failure will be initiated by forming plastic
hinges at the edge members. If the edge elements are rigid, plate failure will be governed by frame action.
The objective of this study is to quantify the post-buckling shear strength of thin steel plates.
The critical shear buckling stress of a rectangular plate with length a, width h and thickness t subjected
to uniform shear loads can be formulated as (Timoshenko and Gere 1961);
S 2E
W cr kv (1)
12( 1 Q 2 )( h / t )2
Where, kv is the non-dimensional shear-buckling coefficient, which depends on the aspect ratio and the
boundary conditions of plates. For plates with all edges simply supported;
4
kv 5.34 , for a/ ht1 (2)
( a / h )2
Obviously the critical buckling stress depends on the slenderness ratio h/t of the plate. As h/t decreases,
IJcr can be as high as the shear yield stress IJy (ignoring
Ekvstrain
/ Fy hardening). It can be shown that, in theory,
when the slenderness ratio h/t is greater than 1.25 , the plate will shear buckle before it yields.
BO CHEN and K.S. SIVAKUMARAN / Procedia Engineering 14 (2011) 641–647 643
The North American Specification (AISI 2007) provides a procedure to estimate the ultimate shear
capacity of solid web plates, which is as given below:
The nominal shear strength [resistance], Vn of the web element shall be calculated as follows:
Vn = Aw Fv (3)
0.60 Ekv Fy
(b)- MODERATE PLATE: For Ekv / Fy < h/t d 1.51 Ekv / Fy ; Fv (4b)
(h / t )
S 2 Ekv
(c)- THIN PLATE: For h/t > 1.51 Ekv / Fy ; Fv = 0.904E kv / (h / t ) 2 (4c)
12(1 Q 2 )(h / t ) 2
According to the AISI (2007) procedure the cold-formed steel web plates can be divided into three
cases, namely, (a) thick, (b) moderate thick and (c) thin. Plates falling in region (a) will fully yield first,
since the corresponding shear buckling stress is higher than the yield strength. Region (c) is for plates
having their critical shear buckling stress less than the proportional limit in shear (0.8 IJy = 0.8*Fy /¥3).
Such plates undergo elastic buckling behavior. For plates with h/t ratio falling in case (b), which is the
transition zone between elastic buckling and yielding, plates undergo inelastic buckling. The ultimate
strength of such plates were established as the geometric mean of their shear buckling stress and 0.8 times
the shear yield stress, which is, Fv W cr ( 0.8W y ) . It is evident that the procedure provided in AISI (2007)
does not incorporate the post-buckling behavior of web plates into its design shear capacity.
In this section, a general finite element model for analyzing the post-buckling behavior of a simply
supported rectangular cold-formed steel plate subjected to pure shear loads is presented. Figure 1 shows
the dimensions and the loading conditions of plates under consideration. The quadrilateral four-node shell
elements were used in this model. Unless otherwise specified, all nodes were fixed in the z-direction
rotation ( T 3 ) to restrain the rigid body rotation about z-axis and were free in all the other degree of
freedoms. In order to simulate a simply supported boundary condition, the nodes along the four edges of
the plate were fixed in z-direction translation. In addition to the z direction translation, one point was also
fixed in the x and y translations to eliminate the rigid body movement of the plate in x and y directions.
Another point was also fixed in x-translation to avoid the rigid body rotation of the plate about z axis. To
simulate a pure shear loading state, uniformly distributed line loads were applied along the four edges of
the plate. The shear loads were directly applied to the nodes as a system of conservative forces and were
kept tangential to the edges of the plate during the deformation process. The proposed finite element
model used an idealized multi-linear stress-strain relationship for cold-formed steel proposed by
Sivakumaran and Abdel-Rahman (1997). Since this study focuses only on the flat plates, the stress-strain
relationship corresponding to the flat area was used for all models. The analyses considered commonly
used steel having yield strength of 350MPa. The von Mises yield criterion was adopted as the yielding
criterion for steel. Isotropic hardening rule was used. The model has included the effects of the
geometrical initial imperfection, and has ignored the effects of the residual stress. Geometrical
imperfection is a function of plate width, thickness, forming process, installation etc. For this study, a
double sine function was used to represent the distribution of the geometrical imperfection in a web plate.
Without going into details the amplitude of the imperfection was taken as w0 = 0.003h, where h is the
width of the main plate. Residual stresses are stresses that exist in steel sections as a result of the
deformations during the cold-forming fabricating processes, and the thermal gradients that are induced in
644 BO CHEN and K.S. SIVAKUMARAN / Procedia Engineering 14 (2011) 641–647
the welding process. Rodal (1992), Abdel-Rahman and Sivakumaran(1997) and Schafer and Peköz (1998)
have experimentally and analytically studied the changing of the yield strength and the residual stresses in
cold-formed steel sections due to cold-working. It was found by these researches that cold-formed steel
sections have elevated yield strengths in the corner regions due to the cold-working process. The increase
of yield strength and the induced residual stresses tend to compensate each another. Also, due to
modeling complications and lack of systematic data, residual stresses are often ignored in numerical
analysis. Based on above reasons, the residual stresses were neglected in the current study. The
automatic-time-step (ATS) was used as the analysis method. At least 100 load steps were used for each
loading process to ensure sufficient numbers of data points were available for the load-displacement
diagram. Based on convergence studies a 24x24 mesh configuration with a total of 576 elements was used
to model plates with an aspect ratio of 1, and the mesh was gradually increased to 24x120 to model plates
with a/h = 5.
Forty simply supported plate models covering five different aspect ratios (a/h =1, 2, 3, 4, 5), and eight
slenderness values were studied. The AISI (2007) limits the slenderness value h/t for an unreinforced
flexural web panel of a cold-formed steel member to a maximum value of 200, thus the h/t ratios selected
in this study started from h/t =50, and increased at every 25 interval until h/t =200. However, one extra h/t
value, namely h/t =250, was also studied in order to obtain a better view on how the ultimate shear
strength of a plate changes as the h/t ratio is changed. Thus, this study considered eight slenderness values,
using a fixed width of the plates of h =100mm. Figure 2 shows the relationship between the average
applied shear stress and the average shear strain for plates with aspect ratios of 1 and 5. It can be seen
from these figures that the ultimate shear strength of solid plates decreases at a decreasing rate as the h/t
values increase. It can also be found that the ultimate shear strength of solid plates decreases as the a/h
ratios increase. However, as the aspect ratio becomes large, the effect of the a/h value on the ultimate
shear capacity of plates fades out. In practice, cold-formed steel members are normally used without
intermediate stiffener, which means that the a/h for the web panels tend to be very large (a/h >>5). Thus,
the effect on the ultimate shear strength of web panels caused from the aspect ratio a/h can be normally
neglected.
Figure 2: Shear stress Vs Shear strain for plates with aspect ratio a/h = 1 and 5
Figure 3 shows the ultimate shear strength of plates as a function of slenderness ratio obtained from
both the finite element analysis and the AISI (2007) method for plates with a/h =1 and 5. Regions (a), (b)
and (c) in this figure correspond to the thick, moderately thick, and thin web panels defined earlier. In
regions (a) and (b), the ultimate shear capacities obtained from the finite element analysis are somewhat
BO CHEN and K.S. SIVAKUMARAN / Procedia Engineering 14 (2011) 641–647 645
lower than those calculated following the AISI (2007) specification. For plates in region (a), the AISI
(2007) uses IJy = 0.60 Fy instead of IJy =Fy/¥3 | 0.577 Fy when calculating the ultimate shear capacity of
thick webs. Thus the nominal ultimate shear strength for plates using the AISI (2007) method in region (a)
tends to be over-estimated. The major difference in the ultimate shear strengths between the finite
element analysis results and the AISI (2007) results occurs for case (c) plates (thin plates). The shear
capacities obtained from the finite element study for thin plates are higher than those calculated from the
AISI (2007) equations. This is because that the AISI (2007) calculates the ultimate shear strength of
plates based on either the shear yield stress or the shear-buckling stress of web plates. Any additional
post-buckling strength due to tension field action that may exist after the web has buckled is neglected.
However, appreciable amounts of post-buckling strengths can build up for thin plates. Thus the current
AISI specification has conservatively estimated the ultimate shear strength of thin shear webs. Thus,
modifications may be made to the AISI equations to include the post-buckling strength of thin web plates.
250
250
FEM
AISI (2007) Region(b) Ekv / Fy d h/ t d1.51 Ekv / Fy
150 150
FEM Region(c) h / t d1.51 Ekv / Fy
AISI (2007)
100 100
Region (a) 0 d h / t d Ekv / Fy
Region (b) Ekv / Fy d h / t d1.51 Ekv / Fy
50 50
Region (c) h / t d 1.51 Ekv / Fy
a/h = 1.0 a/h = 5.0
0 0
0 50 100 150 200 250 0 50 100 150 200 250
h/t h/t
Figure 3: Ultimate strength Vs Slenderness for plates with aspect ratio a/h = 1 and 5
Parametric studies indicated considerable post-buckling shear capacity in thin plates. The post-
buckling strength of plates IJp may be quantified by subtracting the elastic buckling stress from the
ultimate shear strength of plates. Elastic bucking stress of plates IJcr was given in equation (1). In order to
establish the impact of various parameters on the post-buckling capacity, first we established a
dimensionless parameter IJp /IJcr, which is essentially the post-buckling strength divided by the shear
buckling stress of the corresponding plates. Figure 4[A] shows the relationship between the post-buckling
strength normalized by the shear buckling strength and the slenderness ratios up to h/t =200 for plates
with varying aspect ratios. Figure 4[B] shows the relationship between the post-buckling strength
normalized by the shear buckling strength and the aspect ratios up to a/h =5 for plates with varying
slenderness ratios.
It can be seen that the ratio IJp/IJcr increases approximately linearly as the h/t ratio increases. Thus the
new proposed equation could use a linear equation to represent the relationship between the ratio IJp/IJcr
and their slenderness ratios h/t. It can also be seen that when a/h d 2 , the ratio IJp/IJcr increases for plates
with a/h =1 to plates with a/h =2. However, the real function between the ratio IJp/IJcr and the aspect ratio
a/h is unknown for plates with a/h d 2 since only two data points (a/h =1 and a/h =2) are available. In
this study, a linear relationship is assumed as an approximate estimation of the relationship between the
ratio IJp/IJcr and the aspect ratio a/h for plates with a/h d 2 . When a/h>2, the ratio IJp/IJcr decreases
approximately linearly as the a/h ratio increases. As a result, one linear equation could be used in each a/h
range to include the effects on the post-buckling strength of plates caused by changing of the aspect ratio.
646 BO CHEN and K.S. SIVAKUMARAN / Procedia Engineering 14 (2011) 641–647
Based on above discussion, this paper proposes a modified shear strength equation for thin plates (Region
c).
1.2 1.8
a/b=1 h/t=250
a/b=2 1.6
1.0
a/b=3 1.4
0.8 a/b=4
1.2
a/b=5 h/t=200
ʏ p /ʏ cr
IJ p/IJcr
1.0
0.6
0.8 h/t=175
0.4 0.6 h/t=150
0.4
0.2
[A] 0.2 [B] h/t=125
0.0 0.0
0 50 100 150 200 250 0 1 2 a/h 3 4 5 6
h/t
Figure 4: Effect of h/t and a/h on the post-buckling shear strength of plates
Accordingly, for plates having h/t > 1.51 Ekv / Fy the design shear stress could be taken as;
Fv=IJcr+ IJp (5)
where
S 2 Ekv
W cr
12( 1 Q 2 )( h / t )2
and W p [0.048(h/t - 1.51 Ekv / Fy ) . f ( a / h )] W cr
and, where f(a/h) = 0.01*(a/h) + 0.2, for 1 d a/h d 2 and f(a/h) = -0.02*(a/h) + 0.26, for a/h > 2
6. CONCLUDING REMARKS
This paper studied the ultimate shear capacity of simply supported plates subjected to pure edge shear
loading, through finite element analysis. The relationships between the shear capacity, the slenderness
ratio and the aspect ratio of plates were obtained. It was shown that the ultimate shear capacity of a plate
decreases as the slenderness ratio h/t of the plate increases. The ultimate shear capacity also decreases,
but to a lesser extent, as the aspect ratio a/h of the plate increases. By comparing the shear strength
obtained from the finite element analyses with that obtained from the AISI (2007) equations, the results
indicated that by ignoring the post-buckling strength of slender plates, the AISI (2007) has
underestimated the shear strength of thin plates, since the post-buckling strength can be many times larger
than the shear-buckling strength of thin plates. This paper quantified the post-buckling strength and had
proposed a modified design equation for thin plates which incorporates the beneficial effects of post-
buckling shear strength of such plates. The material included in this paper is part of the thesis by Chen
(2009).
References
[1] Abdel-Rahman, N. and Sivakumaran, K. S. (1997), Material Properties Models for Analysis of Cold-Formed Steel Members,
Journal of Structural Engineering, 123 (9), pp. 1135-1143.
[2] AISI (2007), North American Specification for the Design of Cold-Formed Steel Structural Members, American Iron and
BO CHEN and K.S. SIVAKUMARAN / Procedia Engineering 14 (2011) 641–647 647