HBRC Journal

ISSN: (Print) 1687-4048 (Online) Journal homepage: https://www.tandfonline.com/loi/thbr20

Experimental and finite element study on the

inelastic lateral buckling behavior of coped I-
beams

Abdelrahim K. Dessouki, Sherif A. Ibrahim & Seham A. El-Sa’eed

To cite this article: Abdelrahim K. Dessouki, Sherif A. Ibrahim & Seham A. El-Sa’eed (2015)
Experimental and finite element study on the inelastic lateral buckling behavior of coped I-
beams, HBRC Journal, 11:3, 339-352, DOI: 10.1016/j.hbrcj.2014.06.009

To link to this article: https://doi.org/10.1016/j.hbrcj.2014.06.009

© 2014 Housing and Building National

Research Center. Production and hosting by
Elsevier B.V.

Published online: 17 May 2019.

Submit your article to this journal

Article views: 290

View related articles

View Crossmark data

Full Terms & Conditions of access and use can be found at

https://www.tandfonline.com/action/journalInformation?journalCode=thbr20
HBRC Journal (2015) 11, 339–352

Housing and Building National Research Center

HBRC Journal

http://ees.elsevier.com/hbrcj

Experimental and finite element study

on the inelastic lateral buckling behavior of
coped I-beams
Abdelrahim K. Dessouki, Sherif A. Ibrahim, Seham A. El-Sa’eed *

Structural Engineering Department, Ain Shams University, Abbasia, Cairo, Egypt

Received 7 March 2014; revised 6 June 2014; accepted 16 June 2014

KEYWORDS Abstract An experimental study of inelastic lateral torsional buckling of coped beams with simply
Lateral torsional buckling; supported ends is presented in this paper. Six full scale coped steel I-beam tests were conducted. The
Coped I-beams; test parameters include the aspect ratio of cope length to beam depth at coped region as well as the
Finite element; ratio of cope depth to beam depth. The results of tests were compared with finite element model
Local buckling results. The test results showed that a reduction in the inelastic buckling load due to coping could
reach more than 60% of the uncoped buckling capacity. A group of twelve finite element models for
steel coped beams are investigated. The study takes into consideration variable parameters such as
cope depth and length. A comparison between uncoped models and models with different geomet-
rical parameters, is performed. The finite element results showed that both the cope length and cope
depth have a significant influence on the lateral torsional buckling capacity. A parametric study of
coped beams with stiffeners at coped region is reported in this paper. Based on the results of coped
beams strengthened with either horizontal or vertical stiffeners, it is found that for cope depth to
beam depth (dc/D) P 0.25; both horizontal and vertical stiffeners are required to prevent local
web buckling at the coped region.
ª 2014 Production and hosting by Elsevier B.V. on behalf of Housing and Building National Research
Center.

Introduction

In steel construction, when beams are connected to girders at

the same elevation, beam flanges must be coped to provide suf-
* Corresponding author. Tel.: +20 1006922483. ficient clearance for proper attachment as shown in Fig. 1.
E-mail address: [email protected] (S.A. El-Sa’eed). Beams can be coped at the top, bottom, or at both flanges.
Peer review under responsibility of Housing and Building National When a beam is coped, the lateral torsional buckling of the
Research Center. beam will be affected [1]. Cheng and Snell [2] carried out both
experimental and analytical studies on elastic lateral buckling
of coped beams. Very little experimental data are available
regarding in-elastic lateral buckling of coped beams. Michael
Production and hosting by Elsevier

1687-4048 ª 2014 Production and hosting by Elsevier B.V. on behalf of Housing and Building National Research Center.
http://dx.doi.org/10.1016/j.hbrcj.2014.06.009
340 A.K. Dessouki et al.

presented herein and is developed to investigate the influence

of stiffened copes on coped beam lateral buckling resistance.
An experimental study of coped I-beams under two sym-
metrical point loads is considered herein. The effect of coping
on the type of failure of coped beams is also studied. Six tests
are conducted for coped beams with different coping details.
Fig. 2 shows the coped beam details. Finite element analyses
of all test specimens are also presented in this paper.
In light of these investigations, a finite element analysis
methodology has been developed to investigate the influence
of cope depth and length ratios. The results of the parametric
study along with general recommendations are also presented.

Problem statement

The study of the elastic lateral torsional buckling of coped

beams by Cheng et al. [7] investigated only coped beams under
mid-span point load. Study of two symmetrical point loads is
taken into consideration in this study both experimentally and
numerically. It also discusses the effectiveness of using differ-
ent types of horizontal and vertical stiffeners for strengthening
coped beams.
Fig. 1 Typical coped beam-to-beam connection.

Test program
and Lam [3] studied experimentally the effect of inelastic
buckling of coped beams, and compared the results with the
Tested coped beams
theoretical results by Cheng and Snell [2], they found the max-
imum reduction in strength due to the effect of residual stresses
and initial imperfections to be 35% for short and braced spec- Six tests were conducted to study the inelastic lateral torsional
imens. Maljaarsa et al. [4] presented numerical models to study buckling strength of coped beams loaded at their top flanges of
the effect of lateral torsional buckling to (coped) beams with the beams. The copes and connection details are shown in
end plates and they recommended not to use stocky beams Fig. 3. Built-up sections with web 200 · 5 mm and flanges
with large copes in combination with short end plates, as this 125 · 8 mm of nominal yield strength of 345 MPa are investi-
gives the largest reduction of the ultimate buckling resistance gated. The beams’ cross sections are classified as compact sec-
of all studied connections. Yam et al. [5] presented an experi- tions to insure avoiding local buckling in the failure mode. The
mental study of the strength and behavior of reinforced coped nominal measured dimensions are given in Table 1. Test beams
beams, they recommended for a coped beam section with a lar- have designations as described below;
ger d/tw ratio, a stiffener arrangement consisting of longitudi-
nal and transverse stiffeners. Yam and Chung [6] proposed
reinforcement details accounted for the effects of various cope
details and the results show the reinforcements were able to
increase the capacity of the coped beam specimens. ‘‘Uncoped specimen’’ with no cope is used as a pilot test for
Cheng et al. [7] studied both lateral and local buckling of comparison. Specimens 120 and 360-B-0.25 have the same
coped beams, as well as possible strengthening of coped region. cope depth but with different cope lengths. Conversely, 180-
They recommended using stiffeners at the coped region in B-0.1 to 0.50 have the same cope length but with different cope
order to improve the buckling strength of coped beams. How- depths. All models have a span of 3000 mm. In order to min-
ever, no theoretical data are available for stiffening coped imize the end restraints, a conservative pined ended condition
I-beams for inelastic lateral torsional buckling. A parametric is assumed. A double clip angle connection is used in the tests
analysis of coped beams for inelastic buckling behavior is as shown in Fig. 3. The clip angle was bolted between the web

bf C
dc

D h tw
ho ho
tw
y
tf tf
bf

Fig. 2 Cope details designation **‘‘After Cheng and Yura’’ [5].

Lateral buckling behavior of coped I-beams 341

45
6
45
20 L80X80X6
PLAN

120,180,360-B-0.25
180-B-0.10
120, 180 OR 360
180

8
44 37 44 37

30
14

46
130
130

70
70 30

30
2-M16
2-M16 8

43 37
180 180-B-0.50 UNCOPED
8

30 70 30

7
100

57 43
44 44

2-M16

1-M24

(3-a) End conditions for test specimens

P P

D
dc
C
L=3000mm
Reaction Reaction

(3-b) Static system of the test

Fig. 3 Details of coping connections of lateral torsional buckling tests. Note: all dimensions in millimeter.

Table 1 Cope details of specimens (dimensions are in mms).

Specimen D (mm) bf (mm) tw (mm) tf (mm) h/tw C (mm) dc/D
Uncoped 216 125 5 8 40 – –
180-B-0.10 180 0.10
180-B-0.25 180 0.25
180-B-0.50 180 0.50
120-B-0.25 120 0.25
360-B-0.25 360 0.25

of the coped beam and that of the main girder. The size of the Test setup
clip is 80 · 80 · 6 mm. It should be noted that the clips were
bolted on both sides of the web. Two bolts of grade 8.8 with The test setup is shown, schematically, in Fig. 4 and in test lab
nominal diameter 16 and 70 mm spacing in between were used in Fig. 5. The main girder that is supporting the coped beams
for connection of the coped beams. The average tensile yield has a web of 300 · 5 mm and flanges 150 · 10 mm. This girder
strength of the clips was 345 MPa and the average elastic mod- is 1000 mm in length. Vertical stiffeners with thickness 12 mm
ulus of clips and beams is 210,000 MPa. were placed at 1/3 and 2/3 of the main girder span. A
342 A.K. Dessouki et al.

Fig. 4 Schematic drawing of test setup.

Fig. 5 Coped beam experimental test setup.

distributer beam, (web 250 · 5 mm and flanges 150 · 10 mm), applied directly to the distributer beam by connecting the jack
is used to apply the two concentrated loads on the tested beam. head and the distributer beam using a supporting system as
Transverse web stiffeners were used to strengthen the distrib- shown in Fig. 6. For loading cell, 4-M24 anchors of grade
uter beam at the loading positions. 10.9 were used as shown in the figure. Other two supporting
The load is applied to the bottom flange of coped beam ver- plates were used with dimensions 200 · 400 · 25 mm and of
tically upward, where the beam is flipped upside down. It was nominal yield strength 240 MPa. A steel rod with diameter
meant to have the beam in this position to avoid direct loading 30 mm is used to connect the distributer beam with the coped
on the compression flange. A 200 kN tension loading cell was beam, see Fig. 6. The load was applied quasi-statically at
Lateral buckling behavior of coped I-beams 343

Fig. 6 Loading cell and test setup details.

Fig. 7 Typical location of dial and strain gauges in test beams.

Table 2 Comparison between test results and finite element

results.
Specimen PTest (kN) PTest/Puncoped PF.E. (kN) PTest/PF.E.
Uncoped 75.75 1.00 84.00 0.90
180-B-0.10 62.50 0.82 76.30 0.82
180-B-0.25 59.00 0.78 51.66 1.14
180-B-0.50 46.90 0.62 37.25 1.26
120-B-0.25 73.50 0.97 65.14 1.13
360-B-0.25 36.00 0.47 32.00 1.13
Average 1.07

intervals of 2.5 kN each and the load cell was used to measure
the applied loads.

Instrumentation

The in-plane deflection and lateral displacement of the tested Fig. 8 Load versus vertical deflection curves of the tests with
beams were measured using six dial gauges as shown in varied cope length.
344 A.K. Dessouki et al.

Fig. 11 Test result of strain profile for specimen 180-B-0.50 near

the cope.

Fig. 9 Load versus vertical deflection curves of the tests with

varied cope depth.

Fig. 12 Test results of load versus out-of-plane deflection at

point 1 and 5 of specimen (360-B-0.25).
Fig. 10 Test result of strain profile for specimen 180-B-0.25 near
the cope.

Fig. 7. Two dial gauges were placed near the coped end to buckling. The ultimate failure loads, PTest, of all the specimens
record the lateral and vertical movements of the beam. are given in Table 2. The test results show that when the cope
Another two dial gauges were placed at the loading position depth to beam depth ratio ‘‘dc/D’’ increases, the failure load of
as well as two dial gauges at mid span of the tested beam to the specimens decreases. The failure load decreased up to 47%
measure lateral and vertical displacements. Five longitudinal of the un-coped beam failure load (More results can be found
strain gauges mounted on the beam web near the end of the in Ahmed thesis [8]). The load versus in-plane vertical deflec-
cope and at mid span of the beam are installed to record the tion curves, in Figs. 8 and 9, respectively, show that a linear
strain distribution across the beam depth as shown before in load–deflection behavior exists before inelastic lateral buckling
Fig. 4. starts to occur.
Strain readings at the junction of the I-section and the
T-section (cope regions) of specimens 180-B-0.25 and 180-B-
General test procedure
0.50 are shown in Figs. 10 and 11, respectively. The neutral
axes of the I-section and the T-section are also included in
The test procedure was the same for all coped beams. The load the figures. Strain readings at the cope corner are much higher
was applied in increments of 2.5 kN each so that the points than the calculated values using the T-section geometrical
along the load deflection curve would nearly be equally spaced. properties. This is mainly caused by the stress concentration
The tests were terminated when the maximum load was at the cope corner.
reached and then unloading occurred and the deformation A typical out-of-plane deformation plot of the web at the
under the loading became uncontrollable. cope corner versus the applied load of specimen 360-B-0.25
is shown in Fig. 12. It can be seen that insignificant out-
Test results of-plane web deformation at the cope corner was observed
once the load deformation curve starts to be nonlinear. When
The general failure mode of all specimens was inelastic lateral the failure load was reached, significant lateral deflection was
torsional buckling at the cope region accompanied with local observed with a value of 25 mm at the cope corner.
Lateral buckling behavior of coped I-beams 345

Fig. 13 Finite element model of tested coped beam.

Fig. 14 Comparison of load vs. lateral deflection between finite Fig. 15 Comparison of load vs. lateral deflection between finite
element and test results for specimen (120-B-0.25). element and test results for specimen (180-B-0.25).

Finite element model of test specimens restrained in UY, and UZ. The load is represented by symmet-
rical point loads applied at the bottom flange of the coped
The test specimens were analyzed using the commercial finite beam. Since the out-of plane imperfection was not measured
element program, ANSYS V.11 [9]. A ‘‘thin shell 93’’ element during the test, suitable values of imperfections were assigned
is used in the models which has six degrees of freedom at each in the finite element analysis in order to capture the load–
node; three translations in the nodal x, y, and z directions and deflection behavior of the specimens. The maximum initial
three rotations about the nodal x, y, and z-axes. The deforma- imperfection used, was 1/1000 of the span at the load applica-
tion shapes are of quadratic function in both in-plane and out- tion point in out-of-plane direction.
of-plane directions. The element used has plasticity, stress stiff-
ening, large deflection, and large strain capabilities. The finite Comparison between finite element and test results
element models of each beam consisted of 7188 elements and
21,989 nodes, 8 elements per flange width and 30 elements The comparison between the inelastic buckling loads obtained
per web height. Typical model of the test beam is shown in by finite element analysis and those obtained from the test
Fig. 13. All nodes in web and flanges at the end of the beam results is given in Table 2. The ratio of test to finite element
are restrained in UX, UY, and UZ while at the other end results (PTest/PFEM) has a mean value of 1.07. The load versus
346 A.K. Dessouki et al.

Fig. 16 Comparison between typical failure mode of test specimens and finite element models.

lateral deflection curves of the specimens were also compared angle connection and to identify important parameters that
with the test results, see Figs. 14 and 15, respectively. Typical may influence the connection strength, a parametric study is
buckled shapes of all specimens of the test and the finite ele- conducted. Based on the test results and the finite element
ment models are shown in Fig. 16. For all specimens, it is results, two parameters were chosen in this study, the cope
noticed from the deformed shape of the coped beams, that length and the cope depth. The variable parameters considered
the failure is due to inelastic lateral torsional buckling accom- in this study are summarized in Table 3. Seven beam sections,
panied by localized buckling effect and a relatively high stress 600-B-0.25, 120-B-0.10, 360-B-0.10, 600-B-0.10, 120-B-0.50,
level was observed at the coped location. 360-B-0.50 and 600-B-0.50, in addition to the previous coped
beams were chosen for the finite element analysis. All models
have a beam length of 3000 mm. The end condition is kept
Finite element parametric study the same for all beams by using two clip angles L80x6 back-
to-back. Initial imperfection at the load application points is
To provide a better understanding of the inelastic lateral considered as 1/1000 of the span in X-direction, i.e. in the
torsional buckling behavior of coped beams with bolted clip out-of-plane direction. The yield strength and ultimate tensile
Lateral buckling behavior of coped I-beams 347

Table 3 Comparison of ratio (PF.E./Py) and (PF.E./PPILOT) for different cope details.
Models C (mm) dc/D PF.E.(kN) PF.E./Pplastic PF.E./Py PF.E./PPILOT
SP1 120 0.10 81.60 0.91 1.02 0.97
SP2 180 76.30 0.85 0.95 0.91
SP3 360 42.00 0.47 0.52 0.50
SP4 600 28.00 0.31 0.35 0.33
SP5 120 0.25 65.14 0.72 0.81 0.78
SP6 180 51.66 0.57 0.64 0.62
SP7 360 32.00 0.36 0.40 0.38
SP8 600 23.54 0.26 0.29 0.28
SP9 120 0.50 40.00 0.44 0.50 0.48
SP10 180 37.25 0.41 0.46 0.44
SP11 360 25.80 0.29 0.32 0.31
SP12 600 15.40 0.17 0.19 0.18

Fig. 17 Stress in z-direction for SP3 model at dc = 0.10D.

Fig. 18 Stress in z-direction for SP11 model at dc = 0.50D.

348 A.K. Dessouki et al.

beam. The figure shows clearly the existence of high stress con-
centration at the coped region. The effect of residual stresses
was not taken into consideration, may influence the capacity
of coped beams.

Effect of cope length on lateral torsional buckling strength of

beams

The cope length, in this study varies from 120 to 600 mm with
the change of the cope depth from 0.10D to 0.50D. The beam
span is chosen as 3000 mm.
Curves representing the load–displacement relation are
shown in Fig. 19. In this figure, the effect of increasing coping
length on the lateral torsional buckling capacity is shown. As
an example, for beam at (dc/D = 0.25), by increasing the cope
length from 120 to 180 mm, the buckling capacity decreased by
Fig. 19 Effect of cope length at dc/D = 0.25 on lateral buckling 20% and by increasing cope length from 180 to 360 mm, the
capacity for parametric study models. buckling capacity decreased by 60% at the same cope depth.

Effect of cope depth on lateral torsional buckling strength of

beams

Fig. 20 shows the ratio of (PF.EM/Pplsatic) (where Pplastic is the

corresponding plastic load capacity of uncoped beam), versus
(C/ho) (where ho = (D dc) = depth of beam at coped region)
for different cope depths. This implies that the cope depth will
have a greater influence than for the copes that are short. The
reduction in buckling capacity due to the increase in cope
depth is greater for short cope lengths (C/ho = 1.0) than for
long cope lengths (C/ho = 2.0).
The buckling capacity of the coped beam is reduced by
about 10% for dc/D = 0.10 at C/ho 6 1.0 and for deeper
copes, the buckling capacity decreases by about 38% when
dc/D = 0.50 and C/ho 6 1.0 as given in Table 3.

Strengthening techniques of coped beams

Possible strengthening methods were studied by Yam et al. [5].

They tested 10 specimens to investigate the strength and behav-
ior of reinforced coped steel I-beams and the results show that
the length of longitudinal stiffeners, length of transverse stiffen-
Fig. 20 Effect of cope depth at different cope lengths on
ers, combined longitudinal and transverse stiffeners, double
parametric study models.
transverse stiffeners were able to increase the capacity of the
coped beam specimens. Yam and Chung [6] examined the effect
strength are 345 and 500 MPa, respectively. The output results of various parameters on the strength of reinforced coped
of the finite element models are given in Table 3. beams and found that the strength of the reinforced coped
Table 3 illustrates a comparison of (PF.E./Py) for different beams decreases with increasing cope depth to beam depth
coped beams where PF.E. and Py are the failure load and the ratio (dc/D), irrespective of the cope length to beam depth ratio
corresponding yield load, respectively. Table 3 also illustrates (C/D) and types of stiffener. In this research the study is
a comparison of (PF.E./PPILOT) for different coped beams extended to cover the effect of dc/D = 0.50 and also to study
where PF.E. and PPILOT are the failure load of the finite ele- the effect of cope depth during strengthening even short or long
ment model for coped beams and the failure load for the unc- cope lengths. Cheng et al. [7] and Yam and Lam [10] recom-
oped beam respectively. The ratio of the maximum failure load mended using stiffeners at the coped regions to improve the lat-
to the yield load ranged from 1.02 to 0.19. eral torsional buckling strength. Two reinforcing details as
Figs. 17 and 18, respectively, show the stress distribution in shown in Fig. 21 were recommended. Reinforcing details of
z-direction for coped beams SP3 at dc = 0.10D and SP11 at type B was more recommended than type A in case of large
dc = 0.50D where the maximum values of stresses reached cope depth. However, no theoretical data are available for stiff-
393 and 363 MPa, respectively. For both beams, it is noticed ening coped I-beams in case of inelastic lateral torsional buck-
that the failure mode is due to localized buckling effect at ling of compression flange. Therefore, in order to verify the
the connection between cope region and the uncoped part edge strengthening details recommended by Cheng et al. [7], in case
accompanied with the overall lateral torsional buckling of the of inelastic lateral torsional buckling, an analytical study of
Lateral buckling behavior of coped I-beams 349

Fig. 21 Types of coped beam reinforcement suggested by Yam et al. [5].

Fig. 22 Different types of proposed web reinforcement.

Table 4 Comparison between test results and finite element results for strengthened beams.
Specimen PF.E. STIFF (kN) PF.E. (kN) PF.E (Type 4)/PF.E.

Type 1 Type 2 Type 3 Type 4

180-B-0.10 77.29 81.36 77.00 82.90 76.30 1.07
180-B-0.25 71.10 74.25 68.75 81.20 51.66 1.57
180-B-0.50 55.60 62.56 58.10 75.75 37.25 2.03
120-B-0.25 70.30 76.44 68.05 78.25 65.14 1.20
360-B-0.25 71.00 74.16 68.67 79.32 32.00 2.40
PF.E STIFF, the failure loads of the stiffened coped beams. PF.E., finite element load of un- stiffened test specimen.

coped I-beams with horizontal stiffeners and vertical stiffener is beams with both horizontal and vertical stiffeners or each type
conducted in this paper. separately. In the analytical study, horizontal stiffener, with
extended length, ‘‘LHR’’, not larger than C (LHR 6 C) was pro-
Finite element modeling of strengthened models and results vided at the cope position as shown in Fig. 22. Hence, two
types of stiffeners with extensions equal to zero (Type 1) and
The test specimens were analyzed by using the finite element equal to C (Type 2), were investigated. It has been noted that
methodology, to study the lateral torsional buckling of coped the stiffeners are installed on both sides of the web. The
350 A.K. Dessouki et al.

Fig. 23 Stress in z-direction for coped beam stiffened by LHR = C at dc = 0.25D.

Fig. 24 Stress in z-direction for coped beam stiffened by LHR = C at dc = 0.50D.

thickness of stiffeners was equal to the thickness of the flange The deformed shapes, distribution of stresses along the con-
while the width of the stiffeners on each side was equal to half nection of coped beam stiffened with both horizontal and ver-
the width of the flange. In addition to the two coped beams tical stiffeners from the finite element analysis are shown in
with horizontal stiffeners, one coped beam with vertical stiff- Figs. 23–25.
ener of length, ‘‘LVR’’ (Type 3), with full depth is analyzed. The finite element results showed that the maximum ulti-
Another type having both vertical and horizontal stiffeners mate loads of the coped beams with stiffeners were larger than
(Type 4), vertical stiffener with full depth and horizontal stiff- those of the coped beams without stiffeners by 7–50%.
ener with length equal to cope length is also verified, as shown
in Fig. 22. The output results of the finite element models of Effect of horizontal stiffeners
stiffened coped I-beams considered in the study are given in
Table 4. Table 4 shows comparison between strengthened and
The failure mode of the coped beams without stiffeners was unstrengthened models with respect to ultimate capacity.
identified to lateral torsional buckling accompanied by high Using horizontal stiffeners in coped parts, Type (1) and Type
local stresses at the cope region. The failure mode of the coped (2), leads to a large increase in buckling capacity by up to
beams strengthened with horizontal stiffeners was identified to 137%. The local buckling in the coped region, which contrib-
be lateral torsional buckling of beams as shown in Figs. 23–25. utes strongly in reducing the capacity of the coped beam, is
Lateral buckling behavior of coped I-beams 351

Fig. 25 Stress in z-direction for coped beam stiffened by LHR = C and LVR = H at dc = 0.50D.

prevented herein and the failure is mostly due to pure lateral

torsional buckling only.
The effect of increasing horizontal stiffener length is also
studied. It is found that when LHR increased, the web crippling
is terminated. As illustrated in deformed shapes, the increase in
horizontal stiffener length from LHR = 0 to C leads to an
increase in buckling capacity. For example, for model 180-B-
0.50, the increase in buckling capacity by changing horizontal
stiffener length from zero to C is 49% and 68%, respectively.
Figs. 23 and 24, respectively, show the deformed shape of
models (LHR = C at dc equal to 0.25 and 0.5) with failure
due to inelastic lateral torsional buckling at the compression
flange. The maximum stresses for beams with dc equal to
0.25 and 0.50 are 403 and 397 Mpa, respectively. The beams
failed by inelastic lateral torsional buckling.

Effect of adjoined horizontal and vertical stiffeners

Fig. 26 Comparison between different stiffening methods on the
To decide which of the horizontal or vertical stiffeners is the coped beam 180-B-0.25.
best improvement technique with respect to lateral buckling
capacity of coped beams, two additional different types of
The buckling capacity increased when using both vertical
stiffeners were examined. The first type included only vertical
and horizontal stiffeners by 10–140% compared with unstiff-
stiffeners, Type 3, whereas the second type included both hor-
ened beams as given in Table 4.
izontal and vertical stiffeners, Type 4, (Fig. 22).
The load versus lateral deflection curves for different stiff-
The use of vertical stiffeners was believed to be able to pre-
ening types show the effect of stiffening on the lateral buckling
vent the sideway movement of the web, (Cheng et al. [7]).
capacity of coped beam 180-B-0.25, see Fig. 26.
From Table 4, it can be seen that when vertical stiffeners with
length equal to beam depth are installed, the buckling capacity
of the stiffened coped beam increased from 33% to 114% Summary and conclusions
when compared with the un-stiffened coped beam.
Since it was found that using vertical stiffeners only did not The lateral torsional buckling behavior of coped beams sub-
increase the lateral buckling capacity of coped beams effi- jected to two symmetrical point loads is investigated both
ciently, it was decided to provide vertical stiffeners together experimentally and using finite element analysis. In the
with horizontal stiffeners at the coped section. It was found experimental program, six full scale tests of coped I-beams
that the failure load of the coped beam with both horizontal are carried out for the lateral torsional buckling strength. It
and vertical stiffeners ranged from 10% to 25% higher than was shown from the test results that web distortion occurred
that of the beam with horizontal stiffener only of length equal at the coped region, and high stress concentration zone is
to cope length. developed.
352 A.K. Dessouki et al.

A finite element study on tested coped beams is carried out 8- It is recommended that using horizontal stiffeners is
using the commercial program ANSYS. A non-linear finite more effective for small cope depths due to elimination
element analysis, in which the material non-linearity, geomet- of coped region local buckling. By increasing the cope
ric non-linearity, and initial imperfection at the load applica- depth (dc > 0.25D), it is preferable to use both vertical
tion points as 1/1000 of the span in X-direction, i.e. in the and horizontal stiffeners. The finite element results
out-of-plane direction were considered, was conducted to showed that the maximum ultimate loads of the coped
study the behavior of coped beams with different geometrical beams with stiffeners were larger than those of coped
configurations. The effect of residual stresses was not taken beams without stiffeners by 7–50%.
into consideration. The finite-element analysis results were in
acceptable agreement with the test results.
The calibrated finite element model was then used to con- Conflict of interest
duct a parametric study. Two parameters were chosen in this
study, the cope length and the cope depth. None.
Based on both test and analytical results, the following con-
clusions are drawn:
References
1- The reduction of the ultimate buckling capacity due to
[1] J.J.R. Cheng, Design of steel beams with end copes, J. Constr.
coping is very significant depending upon the cope Steel Res. ASCE 25 (1993) 3–22.
length and cope depth. [2] J.J.R Cheng, W.J. Snell, Experimental study of lateral buckling
2- The results showed that the inelastic buckling load was of coped beams. Proceedings, Structural Stability Research
reduced due to coping by more than 60% compared to Council, Illinois,1991, pp. 181–192.
un-coped buckling lateral torsional buckling capacity [3] C.H. Michael, C.C. Lam, In elastic lateral torsional buckling of
in case of C/ho P 1. coped beams. Sixth EP MESC Conference, China, 1997.
3- Effective local buckling can be avoided by keeping the [4] J. Maljaars, J.W.B. Starkc, H.M.G.M. Steenbergenb, R.
limits of coped beam as dc/D 6 0.25. Abspoel, Lateral–torsional buckling resistance of coped beams,
4- It is found that when dc > 0.25D and h/tw < 60, the J. Constr. Steel Res. 61 (2005) 1559–1575.
[5] Michael C.H. Yam, Hongwei Ma, Angus C.C. Lamc, K.F.
local buckling is effective and will reduce dramatically
Chungd, Experimental study of the strength and behaviour of
the buckling capacity of the tee-section, hence decreases reinforced coped beams, J. Constr. Steel Res. 67 (2011) 1749–
the overall lateral torsional buckling capacity of the 1759.
beam. [6] Michael C.H. Yam, K.F. Chung, A numerical study of the
5- The reduction in the lateral buckling capacity due to the strength and behavior of reinforced coped beams, J. Constr.
increase in cope depth is greater for short cope lengths Steel Res. 80 (2013) 224–234.
than for long cope lengths. [7] J.J.R. Cheng, J.A. Yura, C.P. Johnson, Design and behavior of
6- It is recommended to have a minimum reduction coped beams. Ferguson Structural Engineering Laboratory
capacity of coped beams by keeping the coping limits Report No. 84–1, University of Texas, 1984.
as C/ho 6 1 and dc/D 6 0.10. [8] S. Ahmed, Lateral torsional and distortional buckling of coped
beams with different end conditions (Ph.D. thesis), Egypt, 2014.
7- The lateral buckling capacity of coped beams
[9] ANSYS User’s Manual (version 11). Houston: Swanson
stiffened using vertical and horizontal stiffeners in Analysis Systems Inc., 2006.
coped region is investigated. Finite element modeling is [10] Michael C.H. Yam, Angus C.C. Lam, The local web buckling
employed to study the effects of using horizontal stiffen- strength of stiffened coped steel I-beams, J. Steel Struct. 7 (2007)
ers only or both horizontal and vertical stiffeners on the 129–138.
inelastic buckling strengths of coped beams.