Minia Journal of Engineering and Technology

(MJET)
Vol. 35, No 2, July 2016

Interactive shear buckling of corrugated webs

Prof. Dr: S. A. Tohamy 1, Prof.Dr:A.M.Abdel-Halim2 Ass.Prof: A. B. Saddek3, Eng: Asmaa .Y. Hamed 4
1, Professor of steel structures and bridges, Civil Engineering Departments, EL-Mina University, Egypt.
2, Professor of steel structures and bridges, Civil Engineering Departments, EL-Mina University, Egypt.
3, Vice Dean of Civil Eng. Dept., Beni-Swef University, Egypt. Email: amrbkr93@gmail.com.
4, Civil Engineer, General Authority for Educational Buildings (GAEB), Aswan, Egypt, Email:
asmaa_yasseen@yahoo.com

Abstract

This paper focuses on behaviour of the interactive shear stress of corrugated steel webs
with trapezoidal corrugations experimentally and analytically Proposed interaction equation,
which based on local buckling of the corrugation fold as isotropic flat plates, global buckling
of the entire web panels as an orthotropic plate, and steel yielding of the web is presented.
Four beams with corrugated steel webs were tested to failure under shear; the failure was due
to buckling of the web. Finite element method (FEM) was used to perform nonlinear analysis
to the models of the test specimens to determine ultimate load of these girders. Comparisons
between the results from the proposed equation, the finite element analysis, and the tests are
satisfactory. A new formula by the author is developed, which considers interaction among
the various shear failure modes. The conditions of many of these tests are found to be
inconsistent with the theoretical conditions assumed in deriving the shear strength formulas.
The various formulas for predicting critical shear stress are then compared with selected test
results.

Keywords: Corrugated webs, Shear buckling, experimental test, ANSYS.

1 Introduction
Several previous studies had been focused on steel girders with corrugated webs. Most of
these were about the shear and bending behaviour of simply supported beams.

Easley and McFarland [1], [2] discussed three different formulae for elastic buckling loads of
light-gauge corrugated metal shear diaphragms subjected to in plane shear loads; Easley and
McFarland buckling formula, Bergman-Reissner [3] buckling formula and Hlavacek [4]
buckling formula. They analysed the three different buckling formulae, and noted that the
Bergman-Reissner buckling formula was the most rigorous of the three, and also Easley –
McFarland formula was in agreement with Bergman - Reissner buckling formula and both of
them give lower buckling load than Hlavacek formulas which had 20% difference from the
test result. Galambos [5] presented an estimated equation to calculate elastic critical shear.
Elgaaly et al. [6] had studied the shear strength of beams with corrugated steel webs. They
modelled the test specimens performed by Hamilton [7] using finite elements and they
performed nonlinear analysis using computer program ABAQUS to calculate shear buckling
load for these specimens. They noted that the average ratio between analytical and
experimental buckling loads was 1.15. The primary reason why the analytical results are

63
 Minia Journal of Engineering and Technology
(MJET)
Vol. 35, No 2, July 2016

higher than the experimental is the presence of unavoidable imperfections in the webs of the
test specimens. They noted from the experimental and analytical results that the shear is
carried by the web and controlled by buckling; which is local or global for the coarse or
dense corrugation, respectively .They suggested buckling formulae for corrugated steel webs
which are based on local buckling of the corrugation folds as isotropic flat plates or global
buckling of the entire web panel as an orthotropic plate. They had recommended that the
local and global buckling values are calculated and the smaller value controlled the failure.
Elgaaly et al. [6] presented numerical and experimental investigations on the flexural strength
of steel beams with corrugated webs. Simply supported beams with corrugated webs were
tested to failure under uniform bending. Failures were sudden and due to the vertical buckling
of the compression flange until reach to the yield stress in the flange then a vertical buckling
of the flange into the web. The stresses in the web due to bending are equal to zero except for
very close to flanges where the web is restrained. Sayed [8] investigated shear behaviour of
corrugated steel webs. He proposed an interaction equation which considered the different
failure criteria including steel yielding. He performed numerical analysis for simply
supported steel girders with corrugated steel webs under shear load using finite element
technique. He performed a linear elastic analysis to assess the theoretical buckling modes.
Also he performed a nonlinear finite element model to verify the proposed interaction
equation. The results obtained from the numerical analysis were found to be in a good
agreement with the theoretical prediction obtained using the critical stress equations and the
proposed interaction equations.

In this paper we study the shear behaviour of girders with corrugated webs
experimentally and analytically. Buckling formulae for different buckling modes which may
appear in corrugated plates are investigated here. The interaction between the yield failure
criterion and these buckling modes is presented. Proposed interaction equation which
considers different failure criteria including the shear yielding of corrugated steel webs is
presented. Also the comparisons between the results obtained from test, finite element, and
proposed interaction equation are presented.

2 Shear Failure Modes of Corrugated Steel Webs

Shear failure of corrugated steel webs may occur due to shear yielding, buckling or
interactive between yielding and buckling.

2-1 Steel yielding of web

The shear stress which causes an element of corrugated web to yield when it is subjected
to pure shear stress state can be determined using von Mises yield criterion with f y being the
yield strength of the steel as:

[1]
√

63
 Minia Journal of Engineering and Technology
(MJET)
Vol. 35, No 2, July 2016

2-2 Local shear buckling

Corrugated web in local buckling mode of failure acts as a series of flat plate sub panels,
Timoshenko [10]. These flat plate sub panels are subjected to shear, the elastic buckling
stress considering this plate as isotropic plates. In this research we choose that the boundary
conditions of the web is simply supported each other along shorter edges (X-axis) direction.
Whereas the other two edges (Y-axis) directions, at the flange junction tend to be fixed due to
the large torsional stiffness in the case of laterally supported compression flanges, which had
been proposed by author [9].

The method of the minimum potential energy depends on assuming a suitable deflection
form, which satisfies the assumed end conditions equation suggested by the author [11], the
final form of the critical local buckling is as follows:

[2]
( )
( )

[3]
√
[4a]
[ ( ) ( ) ( )]
[4b]
[ ( ) ( ) ( )]

2-3 Global shear buckling

The global shear buckling stress of the corrugated webs treated the corrugated web as
an orthotropic flat plate. We recall that the assumption of isotropy implies that material
properties at a point are the same in all directions. This means that if an isotropic material is
subjected to an axial stress in a principal direction, the major deformation occurs in the
direction of the applied load. Lateral deformations of smaller magnitude occur in the other
principal directions. Also, shear stress causes only shear deformation. So, normal strains and
stresses are not coupled to shear strains and stresses. The deformations are dependent on the
two independent elastic constants for instance, E and υ.

The final equation that relates shear stress for orthotropic corrugated web plate by the
author [11] is as follows:

[5]

63
 Minia Journal of Engineering and Technology
(MJET)
Vol. 35, No 2, July 2016

[6]
√ ( ) ( ) √ ( )
√
√( ) ( )
[ ]
[( ) ] [7a]

[7b]
( )

2-4 Interactive shear buckling

The interaction equation which includes all failure criteria (steel yielding, local and global
buckling stress) which had been proposed by El-Metwally and Loov [12, 13] as previously
given in equation:

[8]

Where are defined by Equations [2], [5] and [1] respectively. Eq. [8] gives
the least value of the limits of the right-hand side as the maximum limit of the resulting
in the lift-hand side, regardless the exponent "n" value. A low value for n (e.g. n=1) gives
less than the least of the three limits. On the other hand, higher values for n close to
the least of the three limits. In this investigation two values of "n" are considered in models
under study, n=3, and n=5, to determine the best value of them to be used in studies and
applications.

3 Experimental Work
3-1 Laboratory Specimens
Four tests were performed on four hanged steel girders with corrugated webs in the
steel construction laboratory, Faculty of Engineering, EL-Minia University. Girders were
manufactured by local steel fabricator in welding work shop; in two girders the web depth is
300 mm and in the other two girders it is 400 mm, the thickness of web for all girder is1.5mm
(web height to thickness ratio equal 200 and 266.67 respectively). The flange width in all
girders is 100 mm and thickness is 9 mm in all girders. Four corrugation profiles are shown in
table (2) and Fig.(1). Table (2) lists the dimensions of the tested girders. Beams MC1 and
MC3 have the same width 45 mm of horizontal and inclined fold; and beam MC2 and MC4
have the same width 60mm of horizontal and inclined fold, with angles of corrugation equal
37 for all girders. Three steel plate stiffeners (100x9) mm were used in every girder; two over
the supports and one under the load. Webs in all girders were welded continuously to flanges
and vertical stiffeners using fillet welds. The stiffener plates have had fillet weld to the flange
plates from two sides. The size of the weld for connecting built up section and end
connecting plates is taken according to the Egyptian code of practice for steel constructions

63
 Minia Journal of Engineering and Technology
(MJET)
Vol. 35, No 2, July 2016

and bridges [14]. Careful procedures of welding were followed to avoid the initial
imperfection of the web as can as possible. All girders were over hanged and one
concentrated loads was applied at the mid span of each girders; as shown in Fig. [3].

Table (1): Dimensions of tested girders see Fig [1] and Fig [2].

Models L bf tf
Span mm mm
(mm)
MC1 1296 100 9 200 2.16 0.15 3.0
MC2 1296 100 9 200 2.16 0.2 4.0
MC3 1296 100 9 266.7 1.62 0.11 3.0
MC4 1296 100 9 266.7 1.62 0.15 4.0

Fig.[1]: Dimension of test specimens.

Fig.[2]: Dimension of corrugation profiles used in test.

Table (2): Corrugation dimensions for tested girders see Fig. (2).

Models hw tw b d i hr θ
mm mm mm mm mm mm degree

MC1 300 9 45 36 45 27 37
MC2 300 9 60 48 60 36 37

04
 Minia Journal of Engineering and Technology
(MJET)
Vol. 35, No 2, July 2016

MC3 400 9 45 36 45 27 37
MC4 400 9 60 48 60 36 37

Table (3): Mechanical properties of web specimens.

stress of Fy Fu E Elongation
web (t/cm2) (t/cm2) (t/cm2) %
3.1 3.9 2050 24

3-2 Test load

In the steel construction laboratory, Faculty of Engineering, El-Minia University. One
concentrated loads was applied at the mid span of each girders across the top flange over
mid stiffener, as shown in fig. [3].The rate of increasing the load [1t], was carefully
applied till failure.

Fig. [3]: Test setup and instrumentation

3-3 Test Results

Each test started by adjusting the loading piston so that contact with the top flange of the
distributor beam by suitable number of steel plates. Initial readings of all used
instrumentation were recorded. The specimens were loaded using an increment load of 1t to
allow for readings of the dial gage to be recorded. Reading was recorded at two seconds
interval time. The readings of the dials gages were recorded manually. Visual observations
were also made during the test. Loading of the girders was terminated after the maximum
load was achieved and the recorded load was 10% below the maximum load achieved. Once
the recorded load started to drop, it was difficult to maintain load control. This was due to the
failure of the specimen.

04
 Minia Journal of Engineering and Technology
(MJET)
Vol. 35, No 2, July 2016

All girders tested failed due to the buckling of the web as shown in figs. [5 a, b, c and d]. The
test results are given in Tables [4] and [5]. The load increased gradually to the max capacity
at failure then dropped suddenly.

Fig. [4]: The beam MC1 during testing.

(a) (b)

(c) (d)
Fig. [5]: Failure mode of specimen (MC1, MC2, MC3 and MC4) after test

For all tested specimens, the vertical deflection, the lateral displacement at mid height of the
web of the load application points as a function of the applied load are plotted in figs.[6a, b, c
and d] and figs.[11a, b, c and d].

04