The Open Construction and Building Technology Journal, 2010, 4, 105-112 105

Open Access

Critical Lateral-Torsional Buckling Moments of Steel Web-Tapered

I-beams
Ioannis G. Raftoyiannis* and Theodore Adamakos

Department of Civil Engineering, National Technical University of Athens, Athens 15780, Greece

Abstract: This paper deals with the stability of steel web-tapered I-beams subjected to bending loads. Tapered beams can
carry a maximum bending moment at a single location while in the rest of the member the moment carrying capacity is
considerably lower. This results in appreciable savings in materials as well as in construction. Numerous researchers have
focused on the investigation of the elastic behavior of tapered I-beams and many theoretical findings have been incorpo-
rated into the current specifications. According to Eurocode 3, the elastic critical moment is used for determining the de-
sign strength against lateral-torsional buckling (LTB) of I-beams with uniform cross-section and a number of coefficients
is employed accounting for the boundary conditions, the cross-sectional geometry and the type of transverse loading,
while no detailed information is given regarding non-uniform members. In this work a simple numerical approach is pre-
sented for determining the critical lateral-torsional buckling loads of web-tapered I-beams. Modification factors of the
elastic critical moment with reference to the mean cross-section are given for various taper ratios. The results presented in
graphical form are compared with those of previous investigations. The approach presented herein can be very easily ap-
plied for the design of tapered beams against lateral-torsional buckling.

Keywords: Elastic buckling, lateral-torsional buckling, steel profiles, web-tapered beams.

INTRODUCTION A stability analysis of the entire structure is required

when designing such members according to Timoshenko &
Steel members with non-uniform cross-section are com-
Gere [14], which leads to the determination of the exact
monly used as beams in the design of various structures such
buckling load of each member as well [15] according to the
as building frames, cranes, masts etc., due to the effort to
provisions of Eurocode 3. The main loading to the beams
minimize the total weight and subsequently the cost of the consists of bending moments and shear due to transverse
structure [1, 2]. The elastic buckling of beams with non-
loads or of end-moments. After determining the elastic criti-
uniform cross-section has been studied by numerous re-
cal moment of the beam, one may proceed to the design of
searchers. Kitipornchai & Trahair [3] presented analytical
the member against lateral-torsional buckling by following
and experimental results for tapered beams under transverse
the provisions of Eurocode 3 [15] and computing the LT
loads. Fraser [4] presented guidelines for the design of portal
buckling strength. This step involves the use of the reduction
frames with tapered members, while Mendera [5] proceeded factor
LT, which is out of the scope of this work.
to the buckling strength of such members. Ermopoulos [6-9]
analyzed buckling of tapered bars under bending and axial Tapered beams are manufactured either from hot-rolled
loading using the slope-deflection method, while Bradford profiles cut in the web lengthwise and re-welded in a special
[10] studied tapered beams subjected to unequal end- arrangement or from assembling steel plates with fillet
moments. Recently, Andrade et al. [11] presented a one- welds. As a result, the produced beam may also suffer from
dimensional model for studying the lateral-torsional buckling distortional buckling [3] and hence, the so-called taper ratio
behavior of singly symmetric web-tapered cantilevers and that is the ratio of the web heights at the two ends of the
simply supported thin-walled beams, while Zhang & Tong beam is of significant importance in our study. For the above
[12] studied the same problem based on a new theory and the reason and in order to avoid distortional buckling of the
variational principle for buckling analysis. Finally, Khelil & cross-section the taper ratio is kept within a reasonable range
Larue [13] presented simple solutions for the lateral buckling and thus, distortional buckling is not considered herein.
of beams with restrained flanges employing the Galerkin In the present study, the influence of the taper ratio and
method. Many other researchers not mentioned herein have cross-sectional geometry on the stability of non-uniform
studied the tosinal-flexural buckling problem of columns steel members that are subjected to bending loads is thor-
with variable cross-section under both axial and transverse oughly investigated. The effect of shear forces is considered
loads. negligible. The problem is studied by focusing on web-
tapered I-beams with build-up cross-sections that are usually
*Address correspondence to this author at the Department of Civil Engi- met in steel structures. The methodology is based on the
neering, National Technical University of Athens, Athens 15780, Greece; formulation of the total potential energy and the numerical
Tel: +30-210-7722454; Fax: +30-210-7722482; solution for buckling of beams with constant or variable
E-mail: [email protected] cross-section [16]. The beams are considered simply sup-

1874-8368/10 2010 Bentham Open

106 The Open Construction and Building Technology Journal, 2010, Volume 4 Raftoyiannis and Adamakos

ported in bending, while other boundary conditions can be

easily dealt with the approach proposed herein. Some charac-
teristic model cases are also analyzed via the finite element
method [17] and the corresponding results are presented for
comparison along with the analytical ones in the form of
diagrams.
More specifically, in this paper special emphasis is given
on analytical models which provide qualitative results of
buckling loads for tapered beams and establish the effect of
properties’ variation on their buckling behavior. This is very
helpful for the designer who can easily apply the proposed
technique to compute with reasonable accuracy the elastic
critical moments and hence, the load carrying capacity of
such types of beams. The effect of taper ratio on the buckling Fig. (2). Boundary conditions for end-rotations of the beam.
strength of such beams is not dealt with in detail in the bibli-
ography. Employment of a detailed finite element analysis The beam is subjected to a transversely applied loading
serves herewith for verification purposes only and both ana- in the form of uniform load q, concentrated load P or end-
lytical and numerical results correlate with reasonable accu- moments M1 and M2=M1. In Fig. (3), one can see a tapered
racy. Then, the analytical results for critical loads (obtained beam subjected to a uniform load q and the corresponding
via the Rayleigh-Ritz method) referring to perfect members moment diagram for simply supported boundary conditions
can be employed to determine the corresponding strength of regarding bending. The moment distribution along the length
the member by using the reduction factor
LT. of the beam from the above diagram can be expressed as
ANALYTICAL MODEL follows

Consider the web-tapered I-beam with length
shown in x x

M ( x ) = 4M 0 [
( ) 2 ] (1)
Fig. (1). The beam has a doubly symmetric cross-section


with variable height h(x), while the width b of the flanges where M0 is the maximum moment value given by.
and the thickness tw of the web and tf of the flanges are con-
stant along the length of the beam. The boundary conditions
2
regarding the end-rotations of the tapered beam due to tor- M0 = q (2)
8
sion, which have been considered herein, are shown in Fig.
(2). These boundary conditions apply simultaneously with
the ones corresponding in bending. More specifically, if (x) q
is the angle of twist along the length of the tapered beam due
to lateral-torsional buckling, the simply supported case cor-
responds to fixed rotation and free warping at the beam-end,
while the fixed case corresponds to fixed rotation and fixed
warping. The free condition corresponds to free rotation and M0
warping of the beam-end and applies only to the case of a
cantilever beam.

a Fig. (3). Moment diagram for uniform load q.

x
h1 h2
P
a
l
z l /2 l /2
P
b
M0
tf
zg
y y
h(x)
tw è
Fig. (4). Moment diagram for concentrated load P.
tf
z z The beam in Fig. (4) is subjected to a concentrated load P
applied at the midlength producing the following moment
Fig. (1). Geometrical characteristics of a web-tapered I-beam. diagram
In this case, the moment distribution is given by
Critical Lateral-Torsional Buckling Moments of Steel The Open Construction and Building Technology Journal, 2010, Volume 4 107

 x where GJ(x) is the torsional stiffness for uniform torsion (St.

2M 0
, 0< x <
/2 Venant) and ECw(x) is the warping stiffness for non-uniform
M(x ) =  (3) torsion. Note that the above terms are not constant since the
2M (1
x ),
/2< x <
cross-sectional height is assumed to vary linearly along the
 0

length of the beam as given in the following relation
where the maximum moment M0 is given by x x
h ( x ) = h 1 (1
) + h 2 (10)




M0 = P (4)
4 The cross-sectional properties needed for lateral-torsional
buckling of the I-section shown in Fig. (1) can be determined
In Fig. (5), one can see a beam subjected to end moments from the following relations
M1 and M2 producing the corresponding moment diagram
In this case, the moment distribution is given by J ( x ) = 13 [ b t 3f + h ( x ) t 3w ] (11)
x
M ( x ) = M 1 [1 + (
1) ] (5) C w (x) = 1 b 3 h 2 (x) t f (12)

24

where in Fig. (5), M2=M1. The load potential
can be determined with the assumption
of incompressible axis [16] from the following relation
As a final case study, we consider the cantilever shown in
Fig. (6) that is subjected to a tip load P. The corresponding
M2 2
 =
 [ 
qz g  2 ] dx
Pz g  02 (13)
moment distribution can be expressed as follows 0 EI z
x
M ( x ) = M 0 (1
) (6) where EIz is the flexural stiffness about the weak axis, zg is

the distance between the point of application of the trans-
where verse load q or P and the centroid and 0 is the angle of twist
in the position of the concentrated load P, i.e. for x=
/2.
M 0 = P
(7) Thus, after substitution of eq(9) and eq(13) into eq(8), we
arrive at the expression of the total potential energy VT, in
which the only unknown is the angle of twist (x). Note also
that secondary effects such as the Wagner effect are negligi-
ble for this type of cross-section.
We next approximate the buckled shape of the beam tak-
ing into account the boundary conditions for rotation with
reference to Fig. (2). Hence, in the simply supported case a
suitable approximation can be taken as follows

x 2
x 3
x
( x ) = C1 sin + C 2 sin + C 3 sin (14)




Fig. (5). Moment diagram for end-moments M1, M2.
Note that the three-term approximation for (x) is proven
to be sufficient since the accuracy of the results is less than
0.1% when keeping more terms. Introducing next the expres-
sion for (x) from eq(14) into the final form of the total po-
tential energy eq(8) and after integrating over the length, we
obtain an expression of VT with respect to the unknown con-
stants Ci (i=1,2,3). Finally, we formulate the following ei-
genvalue problem

2 VT
det =0 (i, j = 1,2,3) (15)
Fig. (6). Moment diagram for tip load P.
C i
C j

from which the buckling load and the corresponding mode is

The total potential energy V T of the above system is derived.
generally given by Similarly, in the fixed-fixed case we take
VT = U +
(8) 2ix
( x ) =  C i (1
cos ) (16)
where U is the strain energy and
is the load potential. The i =1, 2,3

strain energy for twisting (uniform and non-uniform torsion)
can be written as and for the fixed-free case
x x x




( x ) = C1 ( ) 2 + C 2 ( ) 3 + C 3 ( ) 4 (17)
U=  GJ( x )
 dx + 2  EC w ( x )
 dx
1 2 1 2
2
(9)



0 0
108 The Open Construction and Building Technology Journal, 2010, Volume 4 Raftoyiannis and Adamakos

For the fixed-simply supported case we choose a two- load is computed for a beam with h1=450 mm, h2=150 mm
mode approximation consisting from the corresponding flex- and h0=300 mm.
ural buckling problem, where the buckled shape of the beam
In Fig. (8), one can see the corresponding diagram for the
is
same beam, which is now considered simply supported for
x x x bending and fixed-fixed for twisting. In this case one can see
( x ) = C1[( ) 4
2.5( ) 3 + 1.5( ) 2 ] + that as the taper ratio increases from h2/h1=1.0 (uniform




(18) beam) to h2/h1=3.0 (tapered beam), there is an increase of the
kx kx kx
+C 2 [sin
+ k (cos
1)] critical load Pcr up to 6% for
=10h, while for longer beams



the increase drops to 3.5%.
while, in eq(18), it is k=4.493.
fixed-fixed , zg=h/2
NUMERICAL RESULTS
1.06
Solving the eigenvalue problem in eq(15), we obtain the
1.05
buckling load and the corresponding mode. Our main con- length 10h
1.04
cern in this study lies in the cases outlined in the previous length 15h

Pcr/Pcr0
1.03 length 20h
section, while results for more complex loading cases and
1.02 length 25h
boundary conditions can be similarly derived. In the numeri- length 30h
1.01
cal results presented herein in forms of diagrams, the influ-
1.00
ence of the web-tapering ratio on the critical load and mode 1.00 1.50 2.00 2.50 3.00
of buckling in steel web-tapered I-beams is thoroughly in- h2/h1
vestigated.
In Fig. (7), one can see the critical load ratio Pcr/Pcr0 ver- Fig. (8). Critical load ratio Pcr/Pcr0 versus taper ratio h2/h1 for vari-
sus the taper ratio h2/h1 for various lengths of a tapered beam ous slenderness ratios of a fixed-fixed beam with concentrated load
made from an IPE-300 profile. The loading consists from a P applied at the top flange.
concentrated load P applied at the midlength with point of
application the top flange, i.e. zg=h0/2. The index 0 refers to
properties and loads at the midlength of the beam corre- Similarly in Fig. (9), one can see the corresponding to the
sponding to the original uniform beam, i.e. IPE-300. The above diagram, where the beam is considered simply sup-
beam is considered simply supported for both bending and ported for bending and fixed-simply supported for twisting.
twisting.
fixed - simply supported , zg=h/2
simply supported , zg=h/2
1.00
1.00
1.00 0.99 length 10h
0.99
length 10h length 15h
Pcr/Pcr0

0.99 0.98
length 15h 0.98 length 20h
Pcr/Pcr0

0.97
0.98 length 20h length 25h
0.97
length 25h 0.96 length 30h
0.97 length 30h 0.96
0.95
0.96 1.00 1.50 2.00 2.50 3.00
1.00 1.50 2.00 2.50 3.00 h2/h1
h2/h1

Fig. (9). Critical load ratio Pcr/Pcr0 versus taper ratio h2/h1 for vari-
Fig. (7). Critical load ratio Pcr/Pcr0 versus taper ratio h2/h1 for vari- ous slenderness ratios of a fixed-simply supported beam with con-
ous slenderness ratios of a simply supported beam with concen- centrated load P applied at the top flange.
trated load P applied at the top flange.
In this case one can see that as the taper ratio increases
It becomes clear that as the taper ratio increases from from h2/h1=1.0 (uniform beam) to h2/h1=3.0 (tapered beam),
h2/h1=1.0 (uniform beam) to h2/h1=3.0 (tapered beam), there there is a drop of the critical load Pcr up to 4.5% for
=10h.
is a drop of the critical load Pcr, that becomes even more The maximum drop occurs for taper ratios in the range of
pronounced in the cases of short beams. This drop is about h2/h1=2.0 and h2/h1=2.5, while for higher taper ratios this
4% for
=10h, while for long beams the drop is less than effect is reversed.
1%. It must be noted at this point that although the differ-
ences in the critical loads seem to be small, the correspond- Next, we examine the same simply supported tapered
ing differences in the critical load values may be significant beam where the concentrated load P is now applied at the
depending on the both the taper ratio h2/h1 and the length
. centroid, i.e. zg=0. From Fig. (10), it becomes clear that as
For example, the load Pcr0 which used as a reference load, the taper ratio increases from h2/h1=1.0 to h2/h1=3.0, there is
corresponds to a uniform IPE-300 beam with h1=h2=h0=300 a drop of the critical load Pcr, that becomes even more pro-
mm, while if the taper ratio increases to h2/h1=3, the critical nounced in the cases of short beams. This drop is now 3%
for
=10h, while for long beams the drop is less than 1%. In
Critical Lateral-Torsional Buckling Moments of Steel The Open Construction and Building Technology Journal, 2010, Volume 4 109

Fig. (11), one can see the corresponding diagram for the the centroid, i.e. zg=0. From Fig. (16), we can see that as the
same beam, which is considered fixed-fixed for twisting. As taper ratio increases from h2/h1=1.0 to 3.0, there is a drop of
the taper ratio increases from h2/h1=1.0 to h2/h1=3.0, the cor- the critical load qcr, which becomes more pronounced in the
responding increase of the critical load Pcr is up to 4.5% for cases of short beams. This drop is 3.3% for
=10h, while for

=10h, while for longer beams the increase drops to 2.8%. long beams the drop is less than 1%.

simply supported , zg=0 fixed - simply supported , zg=0

1.00 1.00

0.99 length 10h length 10h

0.99
length 15h length 15h
Pcr/Pcr0

Pcr/Pcr0
0.98 length 20h 0.98 length 20h
length 25h length 25h
0.97 0.97 length 30h
length 30h

0.96 0.96
1.00 1.50 2.00 2.50 3.00 1.00 1.50 2.00 2.50 3.00
h2/h1
h2/h1

Fig. (10). Critical load ratio Pcr/Pcr0 versus taper ratio h2/h1 for vari- Fig. (12). Critical load ratio Pcr/Pcr0 versus taper ratio h2/h1 for vari-
ous slenderness ratios of a simply supported beam with concen- ous slenderness ratios of a fixed-simply supported beam with con-
trated load P applied at the centroid. centrated load P applied at the centroid.

In Fig. (12), we show the corresponding to the above simply supported , zg=-h/2
diagram, where the beam is considered fixed-simply sup-
ported for twisting. 1.00

0.99 length 10h

fixed-fixed , zg=0 length 15h
Pcr/Pcr0

0.98 length 20h

1.05 length 25h
0.97 length 30h
1.04 length 10h
length 15h 0.96
Pcr/Pcr0

1.03
length 20h 1.00 1.50 2.00 2.50 3.00
1.02 length 25h h2/h1
FIG.10
1.01 length 30h

1.00 Fig. (13). Critical load ratio Pcr/Pcr0 versus taper ratio h2/h1 for vari-
1.00 1.50 2.00 2.50 3.00 ous slenderness ratios of a simply supported beam with concen-
h2/h1 trated load P applied at the bottom flange.

Fig. (11). Critical load ratio Pcr/Pcr0 versus taper ratio h2/h1 for vari-
fixed-fixed , zg=-h/2
ous slenderness ratios of a fixed-fixed beam with concentrated load
P applied at the centroid. 1.04

length 10h
λόγος Pcr/Pcr0

1.03
In this case one can see that as the taper ratio increases length 15h
from h2/h1=1.0 to h2/h1=3.0, there is a drop of the critical 1.02 length 20h

load Pcr up to 3% for
=10h. The maximum drop occurs length 25h
1.01 length 30h
again for taper ratios h2/h1=2.0 to 2.5, while for higher taper
ratios this effect is reversed, as expected. In Figs (13 to 15) 1.00
are shown the corresponding critical load versus taper ratio 1.00 1.50 2.00 2.50 3.00
diagrams for various lengths and boundary conditions of the λόγος h2/h1

beam with concentrated load P applied at the bottom flange,

i.e. zg=-h/2. The buckling behavior observed in these dia- Fig. (14). Critical load ratio Pcr/Pcr0 versus taper ratio h2/h1 for vari-
grams follows the pattern established previously with corre- ous slenderness ratios of a fixed-fixed beam with concentrated load
sponding differences in the critical loads less than 3%. From P applied at the bottom flange.
a comparison between Figs (7 to 15) the stabilizing effect of
the load position zg is verified and the effect of the taper ratio
on the critical load is well established for various boundary Comparing Figs (10 and 16) we see that there is a slight
conditions. increase in the drop of the critical load and hence, the dis-
tributed load affects more the buckling behavior compared to
As a second case study, we examine the simply sup- the concentrated load.
ported tapered beam subjected to a uniform load q applied at
110 The Open Construction and Building Technology Journal, 2010, Volume 4 Raftoyiannis and Adamakos

As a third case study, we examine the simply supported

fixed - simply supported , zg=-h/2 tapered beam subjected to end moments M1 and M2. In Fig.
(18), one can see the critical load values versus taper ratio
1.00
for a beam with length
=10h and various values of the end-
length 10h moment ratio
. It is clear that as the taper ratio increases
0.99 length 15h from h2/h1=1.0 to 3.0, there is a significant drop of the criti-
Pcr/Pcr0

length 20h
cal moment Mcr with maximum value 12% for
=-0.5, while
for
=1 the drop is only 4%.
0.98 length 25h
length 30h

0.97 In Fig. (19), one can see the same as above diagram for
1.00 1.50 2.00 2.50 3.00 the fixed-fixed case. We can see that as the taper ratio in-
h2/h1
creases from h2/h1=1.0 to 3.0, the critical moment Mcr also
increases with maximum value 6.5% for
=-1, while for
other values of
the increase in almost 5.2%.
Fig. (15). Critical load ratio Pcr/Pcr0 versus taper ratio h2/h1 for vari-
ous slenderness ratios of a fixed-simply supported beam with con- As a final case study, we examine the tapered cantilever
centrated load P applied at the bottom flange. subjected to a tip load P applied at the centroid. In Fig. (20),
one can see the critical load values versus taper ratio for a
cantilever with various length values. In this case, as the ta-
simply supported , zg=0
per ratio increases from h2/h1=1.0 to 3.0, there is a significant
increase of the critical load Pcr with maximum value 34% for
1.00
=3h, while for longer lengths the drop decreases to 22%.

0.99 length 10h

simply supported , L= 10h
length 15h
qcr/qcr0

0.98 length 20h

1.00
length 25h
0.98
0.97 length 30h ψ=1
0.96
ψ=0,5
Mcr/Mcr0

0.96 0.94
ψ=0
0.92
1.00 1.50 2.00 2.50 3.00 ψ=-0,5
0.90
h2/h1 ψ=-1
0.88
0.86
1.00 1.50 2.00 2.50 3.00
Fig. (16). Critical load ratio qcr/qcr0 versus taper ratio h2/h1 for vari- h2/h1
ous slenderness ratios of a fixed-simply supported beam with uni-
form load q applied at the centroid. Fig. (18). Critical moment ratio Mcr/Mcr0 versus taper ratio h2/h1 for
various end-moment ratios of a simply-supported tapered beam.
In Fig. (17) we show the effect of the twisting boundary
conditions on the critical load for a beam simply supported
fixed-fixed , L= 10h
for bending with zg=0,
=10h0 and various values of the
taper ratio. 1.07
1.06
ψ=1
1.05
ψ=0,5
Mcr/Mcr0

zg=0 , L=10h 1.04

ψ=0
1.06 1.03
ψ=-0,5
1.04 1.02
ψ=-1
1.01
1.02
Pcr/Pcr0

1.00
1.00 SS - SS 1.00 1.50 2.00 2.50 3.00
0.98 F - SS h2/h1

0.96 F-F

0.94 Fig. (19). Critical moment ratio Mcr/Mcr0 versus taper ratio h2/h1 for
1.00 1.50 2.00 2.50 3.00 various end-moment ratios of a fixed-fixed tapered beam.
h2/h1

Fig. (17). Critical load ratio Pcr/Pcr0 versus taper ratio h2/h1 for vari- In Fig. (21), one can see a comparison between the criti-
ous support conditions of a tapered beam with concentrated load P cal loads (moments) for lateral-torsional buckling of an IPE
applied at the centroid. and a wide-flange HEB beam. The wide flanges have a stabi-
lizing effect on the critical load since the warping stiffness of
It becomes clear that a fixed condition tends to stabilize the cross-section increases significantly. This effect seems
the tapered beam and the critical load increases with the in- not to be influenced by the web-tapering since the drop of
crease of the taper ratio (see fixed-fixed case) whereas in the the critical load is almost the same in both cases.
cases with one or two simple supports the critical load drops.
Critical Lateral-Torsional Buckling Moments of Steel The Open Construction and Building Technology Journal, 2010, Volume 4 111

cantilever , zg=0

1.40

1.30
length 3h
Pcr/Pcr0

length 5h
1.20
length 7h
length 10h
1.10

1.00
1.00 1.50 2.00 2.50 3.00
h2/h1

Fig. (20). Critical load ratio Pcr/Pcr0 versus taper ratio h2/h1 for vari-
ous slenderness ratios of a cantilever with a tip load P applied at the
centroid.
Fig. (22). FE model of a tapered beam made from an IPE-300 pro-
file.
simply supported, ψ=1 , ℓ=10h

1.00

0.99
Mcr/Mcr0

HEB400
0.98
IPE300

0.97

0.96
1.00 1.50 2.00 2.50 3.00
h2/h1

Fig. (21). Critical moment ratio Mcr/Mcr0 versus taper ratio h2/h1 for
tapered beams made from IPE and HEB profiles.

The above results obtained via the energy method corre-

late well, where applicable, with the analytical and experi- Fig. (23). Lateral-torsional buckling of a simply sup-ported
mental results presented by Kitipornchai & Trahair [3] with tapered beam under uniform load q.
a maximum deviation 1.2% regarding the analytical results
and 2.3% regarding the experimental ones. Fine agreement is Note at this point that employment of the FE Method for
also observed with the results in form of graphs given by lateral-torsional buckling always requires a detailed model of
Bradford [10] for beams with unequal end-moments M at the the beam with special care paid to boundary conditions in
larger and
M at the smaller end. The maximum deviation order to represent correctly bending, twisting and warping at
with the results of this study was 1.9%. the supported ends. On the other hand, the technique pre-
FINITE ELEMENT ANALYSIS sented herein is much simpler regarding computational ef-
fort, while the boundary conditions are employed by a proper
Some characteristic cases of the above studies have also selection of the moment and buckling mode shape functions.
been validated via the Finite Element Method. For this pur-
pose, the Algor FEA commercial code [17] has been em-
ployed for linear buckling analysis. Such a detailed FE
model of an I-beam with taper ratio h2/h1=2.0 build up from
an original profile IPE-300 is shown in Fig. (22). The model
consists of 1200 3-D oriented quadrilateral plate elements
that are employed to model the flanges and the web of the
tapered beam. More specifically, the lateral-torsional buck-
ling mode of the above model subjected to a uniform load q
is shown in Fig. (23) for simply supported boundary condi-
tions. In Fig. (24), one can see the buckling mode for a ta-
pered cantilever model with taper ratio h2/h1=1.5 consisting
from 600 elements which is subjected to a tip load P applied
at the top flange. It is found that the analytical results pre-
sented herein correlate well with the corresponding FEA
results with a maximum error less than 0.2%. Fig. (24). Lateral-torsional buckling of a tapered IPE cantilever
under tip load P.
112 The Open Construction and Building Technology Journal, 2010, Volume 4 Raftoyiannis and Adamakos

SUMMARY AND CONCLUSIONS REFERENCES

In this study, the elastic buckling behavior of steel web- [1] T. Galambos, Ed., Guide to Stability Criteria for Metal Structures,
tapered I-beams by means of determining the corresponding Wiley: New York, 1988.
[2] G. Ballio, and F. M. Mazzolani, Ed., Theory and Design of Steel
elastic critical moment for lateral-torsional buckling. This Structures, Chapman and Hall: London, 1983.
critical moment can be used for the determination of the cor- [3] S. Kitipornchai, and N. S. Trahair, “Elastic stability of tapered I-
responding member strength. More specifically, the influ- beams”, J Struct Div ASCE, vol. 98, no. ST3, pp. 713-728, 1972.
ence of the taper ratio and cross-sectional geometry, the [4] D. J. Frazer, “Design of tapered member portal frames”, Construct.
Steel. Res., vol. 3, no. 3, pp. 20-26, 1983.
boundary conditions and the loading type on the stability of [5] Z. Mendera, Buckling Strength of Tapered Steel Beams, Stahlbau,
non-uniform steel members that are subjected to bending Ernst & Sohn, Berlin: 1995, vol. 64, no. 4.
loads is thoroughly investigated. [6] J. Ermopoulos, “Buckling of tapered bars under stepped axial
loads”, J. Struct. Eng. ASCE, vol. 112, no. 6, pp. 1346-1354, 1986.
The most important conclusions that can be drawn from [7] J. Ermopoulos, “Slope-deflection method and bending of tapered
this study are: bars under stepped loads”, Construct. Steel Res., vol. 11, no. 2, pp.
121-141, 1988.
• The lateral-torsional buckling of web-tapered I-beams is [8] J. Ermopoulos, “Equivalent buckling length of non-uniform mem-
always affected by the web-tapering ratio and the slen- bers”, Construct. Steel Res., vol. 42, no. 2, pp. 141-158, 1997.
derness ratio. [9] J. Ermopoulos, “Buckling length of nonuniform members under
stepped axial loads”, Int. J. Comput. Struct., vol. 73, pp. 573-582,
• Depending on the taper ratio and the slenderness ratio, 1999.
the critical buckling load can be determined with refer- [10] M. A. Bradford, “Stability of tapered I-beams”, Construct. Steel
ence to the original profile from which the tapered beam Res., vol. 9, no. 3, pp. 195-206, 1988.
[11] A. Andrade, D. Camotim, and P. B. Dinis, “Lateral-torsional buck-
is manufactured. ling of singly symmetric web-tapered thin-walled I-beams”, Com-
• As the taper ratio increases, the torsional buckling load put. Struct., vol. 85, no. 17-18, pp. 1343-1359, 2007.
[12] L. Zhang, and G. S. Tong, “Lateral-torsional buckling of web-
decreases slightly for simply supported beams, while for tapered I-beams: A new theory”, Construct. Steel Res., vol. 64, no.
fixed boundary conditions this effect is strongly reversed. 12, pp. 1379-1393, 2008.
[13] A. Khelil, and B. Larue, “Simple solution for the flexural-torsional
• The influence of the taper ratio on the critical load of buckling of laterally restrained I-beams”, Eng. Struct., vol. 30, no.
web-tapered I-beams is proven to be significant espe- 10, pp. 2923-2934, 2008.
cially in the case of cantilevers and must be taken into [14] S. P. Timoshenko, and J. M. Gere, Ed., Theory of Elastic Stability,
when designing such members against buckling. McGraw-Hill: New York, 1961.
[15] Eurocode 3, Design of steel structures. Part 1.1: General rules and
• The technique presented herein is much simpler regard- rules for buildings. EN 1993-1-1. European Committee for Stan-
ing computational effort, while the boundary conditions dardization: Brussels, 2004.
[16] A. Chajes, Ed., Principles of Structural Stability, Englewood
are employed by a proper selection of the moment and Cliffs: Prentice-Hall Inc, New Jersey, 1974.
buckling mode shape functions. [17] C. C. Spyrakos, and I. G. Raftoyiannis, Ed., Linear and Non-linear
Finite Element Analysis in Engineering Practice: Algor Publishing
Division: Pittsburgh, PA, 1997.

Received: December 09, 2009 Revised: January 20, 2010 Accepted: January 21, 2010
© Raftoyiannis and Adamakos ; Licensee Bentham Open.
This is an open access article licensed under the terms of the Creative Commons Attribution Non-Commercial License (http://creativecommons.org/-
licenses/by-nc/3.0/) which permits unrestricted, non-commercial use, distribution and reproduction in any medium, provided the work is properly cited.