Steel Member Strength by Inelastic Lateral Buckling

N. S. Trahair1 and G. J. Hancock2

Abstract: This paper develops a simple advanced method of designing steel members against out-of-plane failure, in which reduced
Downloaded from ascelibrary.org by UNIV OF CONNECTICUT LIBRARIES on 05/14/13. Copyright ASCE. For personal use only; all rights reserved.

elastic moduli are used in an out-of-plane buckling analysis to model the effects of high moment, residual stresses and geometrical
imperfections on yielding. The reduced moduli are derived from the basic beam and column strength curves of the Australian steel code
AS4100 in 1998. The strengths predicted for simply supported beams in uniform bending are exactly the same as those of AS4100, while
those for simply supported columns are extremely close. The strengths predicted for simply supported beam-columns with equal and
opposite end moments are a little higher than the less conservative predictions of AS4100, and are very close to the basic beam and
column strengths when these are plotted against a consistent generalized slenderness. The strengths predicted for simply supported beams
under double curvature bending are somewhat less than those of the AS4100 method of design by buckling analysis, while those for
beams with central concentrated loads acting at or away from the centroid are very close, and those for end restrained beams under
uniform bending and for sway columns are generally a little higher. While the method has been developed from and compared with the
Australian code AS4100, it may be modified for any other modern code for the design of steel structures. It may be more widely applied
to two-dimensional frames with in-plane loading, as part of a simple method of advanced analysis in which separate assessments are made
of the in-plane and out-of plane strengths.
DOI: 10.1061/共ASCE兲0733-9445共2004兲130:1共64兲
CE Database subject headings: Beams; Beam columns; Columns; Buckling; Steel; Residual stress.

to how to incorporate the effects of realistic residual stresses and

Introduction
geometric imperfections so that the buckling analyses will repro-
duce the basic beam and column strength design curves.
In current structural steel design standards such as the Australian
One way to achieve this is to perform inelastic buckling analy-
steel structures standard AS4100 共SA 1998兲, the design of mem-
bers against out-of-plane 共lateral兲 buckling is founded on basic ses in which the elastic moduli E, G are reduced to allow for
beam and column design curves which reduce the full section imperfections and residual stresses as well as yielding. There has
strength. This empirical approach is achieved by using slender- been some initial work in this area 共Trahair and Kitipornchai
ness reduction factors which were determined using the results of 1972; Trahair 1993兲 for the residual stresses in perfectly straight
tests on full-scale members with practical values of residual stress members, which follows from the earlier work of Galambos
and geometric imperfections 共Trahair and Bradford 1998兲. 共1968兲. However, these analyses do not take account of geometric
The allowances made in AS4100 for other effects such as mo- imperfections. More recently, similar reductions in the Young’s
ment gradient, transverse loads acting away from the centroid, modulus E have been used to reduce the in-plane stiffnesses used
and elastic end restraints in beams, and elastic end restraints in in simple advanced analyses of the second-order in-plane bending
columns require the use of moment modification factors and ef- of steel frames 共Kim and Chen 1996a,b兲. The suggestion to re-
fective length factors in the design equations. For beam-columns, duce the out-of-plane stiffnesses by reducing the elastic moduli
the beam and column strengths must be combined in empirical was made by Trahair and Chan 共2003兲, who proposed that in the
interaction equations which are of somewhat variable accuracy, simple advanced analysis of two-dimensional frames with in-
and are often unnecessarily conservative. plane loading only, the out-of-plane analysis should be separated
With the advent of computer elastic buckling analysis pro- from the in-plane analysis. Similar suggestions and proposals
grams for both in-plane and out-of-plane buckling which allow have been made by Wongkaew and Chen 共2002兲.
for these effects, it should not be necessary to use such artificial This paper provides a simple advanced out-of-plane analysis
devices in the design process. However, the problem remains as method in which the elastic moduli E, G are reduced within a
finite element elastic flexural-torsional buckling analysis as
1
Emeritus Professor of Civil Engineering, Univ. of Sydney, NSW, implemented in the computer program PRFELB 共Papangelis et al.
2006, Australia. E-mail: N.Trahair@civil.usyd.edu.au 1993, 1998兲. The method can easily account for all the effects
2
BHP Steel Professor of Steel Structures, Univ. of Sydney, NSW, cited above without the need for effective lengths, moment modi-
2006, Australia. fication factors, and interaction equations. The method is general
Note. Associate Editor: Sherif El-Tawil. Discussion open until June 1, and can reproduce the basic beam and column strength design
2004. Separate discussions must be submitted for individual papers. To curves of existing design standards. It can also be used to deter-
extend the closing date by one month, a written request must be filed with
mine rational out-of-plane strengths for beam-columns, as an al-
the ASCE Managing Editor. The manuscript for this paper was submitted
for review and possible publication on October 7, 2002; approved on ternative to the conventional use of empirical interaction equa-
March 13, 2003. This paper is part of the Journal of Structural Engi- tions. In this paper, the Australian Standard AS4100 design curves
neering, Vol. 130, No. 1, January 1, 2004. ©ASCE, ISSN 0733-9445/ for beams and columns have been used, although the curves of
2004/1-64 – 69/$18.00. any design standard could be substituted.

64 / JOURNAL OF STRUCTURAL ENGINEERING © ASCE / JANUARY 2004

J. Struct. Eng. 2004.130:64-69.

 The advantage of the method is that design software for steel
structures can use the finite element buckling analysis directly
without empirical adjustments of the results. The resulting inelas-
tic buckling strengths accurately reproduce the current code basic
design strengths for beams and columns, and closely approximate
code predictions for other more complex cases.
The inelastic buckling method of this paper also provides a
simple method of advanced analysis of two-dimensional steel
frames against out-of-plane buckling. The examples analyzed in
this paper demonstrate that such an advanced analysis method can
be consistent with the basic member design rules of current stan-
Downloaded from ascelibrary.org by UNIV OF CONNECTICUT LIBRARIES on 05/14/13. Copyright ASCE. For personal use only; all rights reserved.

dards, while providing a rational and consistent method of design-

ing a wide range of structures against out-of-plane failure.

Beam Strength

Nominal Strength
The AS4100 共SA 1998兲 nominal lateral buckling moment strength
M b of a compact I-section beam is given by
Fig. 1. Basic beam and column strengths of AS4100
M b ⫽␣ m ␣ s M px (1)
in which M px ⫽ f y S x ⫽major axis full plastic moment in which

冉 冊 冉 冊
f y ⫽yield stress and S x ⫽plastic section modulus; ␣ m ⫽moment 2
modification factor which allows for nonuniform moment distri- 1 M M
␥ IM ⫽0.9⫺ while ⭐1.0 (6)
butions (␣ m ⫽1.0 for uniform bending兲; and ␣ s ⫽slenderness re- 1.2 M px M px
duction factor which allows for the effects of elastic buckling, in which M⫽bending moment at the cross-section.
initial crookedness and twist, and residual stresses, and which is These values of ␥ IM are shown in Fig. 2 and compared with
given by 共Trahair 1993兲 simplified versions ␥ IE , ␥ IG 共Trahair 1993兲 of the approximations

␣ s ⫽0.6 再 冑冋冉 冊 册 冎
M px
M yz
2
⫹3 ⫺
M px
M yz
⭐1.0 (2)
used by Trahair and Kitipornchai 共1972兲 for the influence of re-
sidual stresses 共i.e., without initial crookedness and twist兲 on the
inelastic lateral buckling of I-beams in uniform bending. The dif-
in which M yz ⫽elastic buckling moment of a simply supported ferences between ␥ IM and ␥ IE and ␥ IG are due to the effects of
beam in uniform bending given by initial crookedness and twist. When the values of ␥ IM given by

冑 冉 冊
Eq. 共6兲 are used in Eq. 共5兲, then the AS4100 basic strengths
␲ 2 EI y ␲ 2 EI w shown in Fig. 1 are obtained. This suggests that nominal strengths
M yz ⫽ GJ⫹ (3)
L 2
L2 which are consistent with the design procedures of AS4100 can
be predicted for other beams with different loading and restraint
in which E and G⫽Young’s and shear moduli of elasticity; I y , J, conditions by using an inelastic buckling analysis with reduced
and I w ⫽minor axis second moment of area, the uniform torsion moduli ␥ IM E and ␥ IM G obtained using Eq. 共6兲.
section constant; and the warping section constant, respectively;
and L⫽length of the beam.
The variation of the dimensionless AS4100 basic beam
strength M b /M px with the modified slenderness 冑 (M px /M yz ) is
shown in Fig. 1.

Reduced Elastic Moduli

The nominal beam moment strength M b may also be obtained by
making an inelastic lateral buckling analysis using inelastic
moduli E I ⫽␥ IM E and G I ⫽␥ IM G which are reduced below their
elastic values E and G to allow for the effects of initial crooked-
ness and twist and residual stresses in reducing the elastic lateral
buckling moments M yz to the nominal strengths M b . The reduc-
tion factor ␥ IM is then obtained by setting the uniform bending
inelastic buckling moment

M I⫽ 冑 ␲ 2 ␥ IM EI y
L2
冉 ␥ IM GJ⫹
␲ 2 ␥ IM EI w
L2
冊 (4)

equal to M b so that
M b ⫽M I ⫽␥ IM M yz (5)
Fig. 2. Reduced elastic moduli factors for beams and columns
when

JOURNAL OF STRUCTURAL ENGINEERING © ASCE / JANUARY 2004 / 65

J. Struct. Eng. 2004.130:64-69.

Downloaded from ascelibrary.org by UNIV OF CONNECTICUT LIBRARIES on 05/14/13. Copyright ASCE. For personal use only; all rights reserved.

Fig. 4. Effect of load height on strengths of beams with central loads

Fig. 3. Strengths of beams in double curvature bending

with a central concentrated load Q, for which the maximum mo-

Moment Gradient
ment is M m ⫽QL/4. The inelastic buckling of such a beam may
The effectiveness of the inelastic buckling method of predicting also be analyzed by using the finite element computer program
the nominal strengths of beams under moment gradient has been PRFELB 共Papangelis et al. 1993, 1998兲 with a sufficiently large
investigated for the extreme example of a simply supported beam number of elements to approximate the variations of ␥ IM with the
with equal end moments M which cause double curvature bend- moment along the beam. Inelastic buckling strength predictions
ing. The inelastic buckling of such a beam may be analyzed by for 250UB37.3 section beams 共BHP 1998兲 with ␥ IM ⫽0.6 ob-
using the finite element computer program PRFELB 共Papangelis tained using 20 elements are plotted in Fig. 4 against the dimen-
et al. 1993, 1998兲 with a sufficiently large number of uniform sionless load height 2y Q /h in which h⫽2 冑 (I w /I y ).
property elements to approximate the variations of ␥ IM with the Also shown in Fig. 4 are the dimensionless nominal strengths
moment along the beam. Inelastic buckling predictions for M b /M px of AS4100 obtained by using the method of design by
250UB37.3 section beams 共BHP 1998兲 obtained using 20 ele- buckling analysis. It can be seen that the inelastic predictions of
ments are shown in Fig. 3. The properties of the 250UB37.3 are the beam strengths are very close to the values of AS4100.
A⫽4,750 mm2 , I x ⫽55.7E6 mm4 , I y ⫽5.66E6 mm4 , J
⫽158E3 mm , I w ⫽85.2E9 mm and S x ⫽486E3 mm3 , and the
4 6
Elastic End Restraints
material properties are E⫽2E5 MPa, G⫽8E4 MPa, and f y
⫽250 MPa, in which A is the area of the cross-section and I x is The effectiveness of the inelastic buckling method of predicting
the major axis second moment of area. the nominal strengths of beams with elastic end restraints has
Also shown in Fig. 3 are the nominal strengths of AS4100 been investigated for a simply supported beam in uniform bend-
obtained by using the method of design by buckling analysis, in ing with equal elastic restraints against minor axis end rotation
which the nominal strength is given by Eqs. 共1兲 and 共2兲 with M yz acting at each end of each flange.
replaced by The elastic buckling moment of such a beam is given by 共Tra-
hair 1993兲
M oa ⫽M ob /␣ m (7)
in which M ob ⫽maximum moment at elastic buckling and the
moment modification factor ␣ m is obtained from
M yzr ⫽ 冑 冉␲ 2 EI y
共 kL 兲 2
GJ⫹
␲ 2 EI w
共 kL 兲 2
冊 (9)

␣ m ⫽M os /M yz (8) in which the elastic effective length factor k satisfies

in which M os ⫽maximum moment at elastic buckling of a beam ␣ RL ␲ ␲
without lateral rotation end restraints and with any transverse ⫽⫺ cot (10)
EI y 2k 2k
loads applied at the centroid 共the values of M ob and M os may be
obtained by using PRFELB, for example兲. It can be seen in Fig. 3 in which ␣ R ⫽stiffness of each elastic restraint.
that the inelastic buckling predictions of the beam strengths are a The inelastic buckling of such a beam may also be analyzed by
little lower than the somewhat empirical values of AS4100. using the finite element computer program PRFELB 共Papangelis
et al. 1993, 1998兲. Inelastic buckling predictions for 250UB37.3
section beams 共BHP 1998兲 with ␥ IM ⫽0.6 obtained using 20 ele-
Transverse Loads Acting At or Away
ments are shown in Fig. 5.
From the Centroid
Also shown in Fig. 5 are the dimensionless nominal strengths
The effectiveness of the inelastic buckling method of predicting M b /M px of AS4100 obtained by using the method of design by
the strengths of beams with loads at different distances y Q below buckling analysis. It can be seen that the inelastic predictions of
the centroid has been investigated for a simply supported beam the column strengths are a little higher than the values of AS4100.

66 / JOURNAL OF STRUCTURAL ENGINEERING © ASCE / JANUARY 2004

J. Struct. Eng. 2004.130:64-69.

Downloaded from ascelibrary.org by UNIV OF CONNECTICUT LIBRARIES on 05/14/13. Copyright ASCE. For personal use only; all rights reserved.

Fig. 5. Effect of end restraints on uniform bending beam strength Fig. 6. Effect of end restraint on unbraced column strength

This is because elastic restraints are more effective for inelastic equal to N cy so that
beams with reduced values of E I than for elastic beams. It may be
concluded that the AS4100 method of design by buckling analysis N cy ⫽N I ⫽␥ IN N y (15)
is a little conservative for beams with elastic end restraints. The variation of ␥ IN obtained by this method with the value of
N cy /N Y is shown in Fig. 2 and compared with the corresponding
variation of ␥ IM with M b /M px for beams. It can be seen that
Column Strength these variations are quite similar. The close approximation for ␥ IN
of
Nominal Strength ␥ IN ⫽0.98⫺0.57共 N/N Y 兲 ⫹0.59共 N/N Y 兲 2 ⫺0.99共 N/N Y 兲 3
The AS4100 共SA 1998兲 minor axis flexural buckling nominal (16)
compression strength N cy of a fully effective 共against local buck- in which N⫽axial compression at the section is also shown in Fig.
ling兲 I-section column is given by 2.
N cy ⫽␣ c N Y (11) When the approximate values of ␥ IN given by Eq. 共16兲 are
used in Eq. 共15兲, then the close approximations to the AS4100
in which N Y ⫽A f y ⫽squash load in which A⫽cross-sectional area; basic strengths shown in Fig. 1 are obtained.
and ␣ c ⫽slenderness reduction factor which allows for the effects
of elastic buckling, initial crookedness, and residual stresses, and
which depends on the type of cross-section 共there are five groups Elastic End Restraints
of these in AS4100兲 and on the modified slenderness The effectiveness of the inelastic buckling method of predicting
␭ c ⫽ 冑 共 N Y /N y 兲 (12) the nominal strengths of columns with elastic end restraints has
been investigated for an unbraced column with an elastic restraint
in which against minor axis end rotation acting at the free end.
N y ⫽␲ 2 EI y /L 2 (13) The elastic buckling load of such a column is given by
is the elastic minor axis flexural buckling load of the column. The ␲ 2 EI y
variation of the dimensionless AS4100 basic column strength N yr ⫽ (17)
共 kL 兲 2
N cy /N Y with the modified slenderness ␭ c for the central cross-
section group is shown in Fig. 1. in which the elastic effective length factor k satisfies
␣ RL ␲ ␲
Reduced Elastic Moduli ⫽ tan (18)
EI y k k
The nominal column compression strength N cy may also be ob- in which ␣ R ⫽stiffness of the elastic restraint.
tained by making an inelastic lateral buckling analysis using an The inelastic buckling of such a column may be analyzed by
inelastic modulus E IN ⫽␥ IN E which is reduced below its elastic using the finite element computer program PRFELB 共Papangelis
value E to allow for the effects of initial crookedness and residual et al. 1993, 1998兲. Inelastic buckling predictions for 250UB37.3
stresses in reducing the elastic flexural buckling load N y to the section columns 共BHP 1998兲 with ␥ IN ⫽0.6 obtained using 20
nominal strength N cy . The reduction factor ␥ IN may be obtained elements are shown in Fig. 6.
by setting the inelastic buckling load Also shown in Fig. 6 are the dimensionless nominal strengths
N cy /N Y of AS4100 obtained by using a method of design by
␲ 2 ␣ IN EI y buckling analysis. It can be seen that the inelastic predictions of
N I⫽ (14)
L2 the column strengths are a little higher than the values of AS4100.

JOURNAL OF STRUCTURAL ENGINEERING © ASCE / JANUARY 2004 / 67

J. Struct. Eng. 2004.130:64-69.

 This is because elastic restraints are more effective for inelastic
columns with reduced values of E I than for elastic columns. It
may be concluded that the method of design by buckling analysis
when used with AS4100 is a little conservative for columns with
elastic end restraints, just as it was for beams with elastic end
restraints.

Beam-Column Strength
Downloaded from ascelibrary.org by UNIV OF CONNECTICUT LIBRARIES on 05/14/13. Copyright ASCE. For personal use only; all rights reserved.

Nominal Strength
The general design rules of AS4100 共SA 1998兲 for a beam-
column under axial compression and in-plane bending can be
used to formulate a lesser nominal out-of-plane moment strength
M on of
M on ⫽M b 共 1⫺N/N cy 兲 (19)
although a more accurate and economical expression can be de-
rived from a higher tier design rule for beam-columns which are Fig. 7. Strengths of beam-columns in uniform bending
without transverse loads and which are fully effective against
local buckling.

is the torsional buckling load of the column. The term (1

Reduced Elastic Moduli
⫺N/N x ) in Eq. 共24兲 allows for increases caused by the prebuck-
The values of ␥ IM for bending moment are close to those of ␥ IN ling in-plane deflections as well as the decreases caused by the
for axial compression, and so a reasonably simple compromise amplification of the moment from M to M m . If the effects of the
for a moment reduction factor ␥ IM N for beam-columns can be prebuckling deflections are ignored, as they usually are, then Eq.
obtained by using 共24兲 may be replaced by

␥ IM N ⫽
␥ IN ⫹␥ IM 共 N Y /N 兲共 M /M px 兲
1⫹ 共 N Y /N 兲共 M /M px 兲
(20) 冉 冊 冉
Mm
M yz
2
⫽ 1⫺
N
Ny 冊冉 1⫺
N
Nz 冊 (26)

for the reduced modulus E I ⫽␥ IM N E, while the value of ␥ IM G

may be used for G I . In this, ␥ IM is the value determined from Eq. Inelastic Buckling Strength Approximations
共6兲 using M /M pxr for M /M px and ␥ IN is the value determined The inelastic buckling of a simply supported beam-column may
from Eq. 共16兲 using N/N Y r for N/N Y , in which M pxr and N Y r are be analyzed by using the finite element computer program
the reduced values of M and N at full plasticity. AS4100 uses the PRFELB 共Papangelis et al. 1993, 1998兲. Approximate inelastic
full plastic approximation buckling strength predictions for 250UB37.3 beam-columns with
equal and opposite end moments are shown in Fig. 7 for the case
M pxr /M px ⫽1.18共 1⫺N Y r /N Y 兲 ⭐1 (21) where M m /M px ⫽N/N Y so that M pxr /M px ⫽N Y r /N Y ⫽0.541
关after using Eq. 共21兲兴. For these approximations, it was conserva-
tively assumed that the in-plane bending moment was constant
Uniform Bending
and equal to the maximum amplified value M m given by Eq. 共22兲.
It can be seen that the inelastic buckling strength approxima-
Elastic Behavior
tions are significantly higher that the very conservative values
For a simply supported beam-column with axial compression
obtained from Eq. 共19兲, but only a little higher than the more
loads N and equal and opposite end moments M, the first-order
accurate higher tier approximations derived from AS4100. It is of
bending moments are amplified by second-order effects to a
interest that the inelastic buckling beam-column strength approxi-
maximum moment approximated by
mations are very close to the AS4100 basic beam lateral buckling
M m ⫽M / 共 1⫺N/N x 兲 (22) strength curve shown in Fig. 7 and therefore also to the AS4100
in which basic column flexural buckling strength curve 共see Fig. 1兲.

N x ⫽␲ 2 EI x /L 2 (23)
is the in-plane column elastic buckling load. Conclusions
Elastic lateral buckling occurs when M and N satisfy 共Vachara-
jittiphan et al. 1974; Trahair 1993兲 This paper develops a simple advanced method of designing steel

冉 冊 冉 冊冉 冊冉 冊
2
members against out-of-plane failure, in which reduced elastic
M N N N moduli are used in an out-of-plane buckling analysis to simulate
⫽ 1⫺ 1⫺ 1⫺ (24)
M yz Nx Ny Nz the effects of high moment, residual stresses and geometrical im-
in which perfections on yielding. The reduced moduli are derived from the
basic beam and column strength curves of the Australian steel
共 GJ⫹␲ 2 EI w /L 2 兲 code AS4100 共SA 1998兲. The method is verified by comparing its
N z⫽ (25)
共 I x ⫹I y 兲 /A predictions of member strength with those of AS4100.

68 / JOURNAL OF STRUCTURAL ENGINEERING © ASCE / JANUARY 2004

J. Struct. Eng. 2004.130:64-69.

 The strengths predicted for simply supported beams in uni- M px ⫽ full plastic moment about x axis;
form bending are exactly the same as those of AS4100, while M pxr ⫽ reduced full plastic moment about x axis;
those for simply supported columns are extremely close. M yz ⫽ elastic buckling moment of beam in uniform
The strengths predicted for simply supported beams under the bending;
extreme moment gradient of double curvature bending are some- M yzr ⫽ value of M yz for beam with end restraints;
what less than those of the AS4100 method of design by buckling N ⫽ axial compression;
analysis, which themselves are somewhat higher than those of the N cy ⫽ column minor axis nominal strength;
simpler AS4100 method of allowing for nonuniform bending. The N I ⫽ column inelastic buckling load;
predictions for beams with central concentrated loads acting at or N x , N y ⫽ column elastic flexural buckling loads;
away from the centroid are very close to those of the AS4100 N yr ⫽ value of N y for column with end restraints;
method of design by buckling analysis, which itself is more ra- N Y ⫽ column squash load;
Downloaded from ascelibrary.org by UNIV OF CONNECTICUT LIBRARIES on 05/14/13. Copyright ASCE. For personal use only; all rights reserved.

tional and consistent than the simpler AS4100 method of allowing N Y r ⫽ reduced column squash load;
for load height by using empirical effective length factors. N z ⫽ column elastic torsional buckling load;
The strengths predicted for end-restrained beams under uni- Q ⫽ beam transverse load;
form bending are generally higher than those of the AS4100 S x ⫽ major axis plastic section modulus;
method of design by buckling analysis, because the AS4100 pro- x, y ⫽ cross-section principal axes;
cedure is based on the relative stiffness of the elastic restraints y Q ⫽ distance of load below centroid;
compared with that of the elastic beam, and so under-estimates z ⫽ distance along a member;
the relative stiffness compared with that of the inelastic beam. ␣ c ⫽ column slenderness reduction factor;
Similar differences were found for end-restrained unbraced col- ␣ m ⫽ moment modification factor;
umns. ␣ R ⫽ restraint stiffness;
The strengths predicted for simply supported beam-columns ␣ s ⫽ beam slenderness reduction factor;
with equal and opposite end moments are significantly higher ␥ IE , ␥ IG ⫽ inelastic modulus reduction factors;
than the more conservative predictions of AS4100, and a little ␥ IM , ␥ IN ⫽ beam and column inelastic modulus reduction
higher than the less conservative predictions. The strengths pre- factors;
dicted are very close to the basic beam and column strengths ␥ IM N ⫽ beam-column inelastic modulus reduction
when these are plotted against a consistent generalized slender- factor; and
ness. ␭ c ⫽ column modified slenderness.
While the method has been developed from and compared
with the Australian code AS4100, it may be modified for any
other modern code for the design of steel structures.
References
The method may be more widely applied to two-dimensional
frames with in-plane loading, as part of a simple method of ad-
Broken Hill Proprietary 共BHP兲. 共1998兲. Hot rolled and structural steel
vanced analysis in which separate assessments are made of the
products, Broken Hill Proprietary Co. Ltd, Melbourne, Australia.
in-plane and out-of plane strengths.
Galambos, T. V. 共1968兲. Structural members and frames, Prentice-Hall,
Englewood Cliffs, N.J.
Kim, S. E., and Chen, W. F. 共1996a兲. ‘‘Practical advanced analysis for
Notation unbraced steel frame analysis.’’ J. Struct. Eng., 122共11兲, 1259–1265.
Kim, S. E., and Chen, W. F. 共1996b兲. ‘‘Practical advanced analysis for
The following symbols are used in this paper: braced steel frame analysis.’’ J. Struct. Eng., 122共11兲, 1266 –1274.
A ⫽ cross-sectional area; Papangelis, J. P., Trahair, N. S., and Hancock, G. J. 共1993兲. ‘‘Computer
E ⫽ Young’s modulus of elasticity; analysis of elastic flexural-torsional buckling.’’ J. Singapore Struct.
E I ⫽ inelastic Young’s modulus; Steel Soc., 4共1兲, 59– 67.
E IN ⫽ column inelastic Young’s modulus; Papangelis, J. P., Trahair, N. S., and Hancock, G. J. 共1998兲. ‘‘Elastic
flexural-torsional buckling of structures by computer.’’ Comput.
f y ⫽ yield stress;
Struct., 68, 125–137.
G ⫽ shear modulus of elasticity;
Standards Australia 共SA兲. 共1998兲. ‘‘AS4100, Steel structures.’’ Standards
G I ⫽ inelastic shear modulus; Australia, Sydney, Australia.
h ⫽ 冑 (I w /I y ); Trahair, N. S. 共1993兲. Flexural-torsional buckling of structures, E & FN
I w ⫽ warping section constant; Spon, London.
I x , I y ⫽ second moments of area about x, y axes; Trahair, N. S., and Bradford, M. A. 共1998兲. The behaviour and design of
J ⫽ torsion section constant; steel structures to AS4100, 3rd Australian Ed., E & FN Spon, London.
k ⫽ effective length factor; Trahair, N. S., and Chan, S. L. 共2003兲. ‘‘Out-of-plane advanced analysis
L ⫽ member length; of steel structures.’’ Eng. Struct. 25, 1627–1637.
M ⫽ applied moment; Trahair, N. S., and Kitipornchai, S. 共1972兲. ‘‘Buckling of inelastic
M b ⫽ nominal member moment capacity; I-beams under uniform moment.’’ J. Struct. Div. ASCE, 98共11兲, 2551–
2566.
M I ⫽ inelastic buckling moment;
Vacharajittiphan, P., Woolcock, S. T., and Trahair, N. S. 共1974兲. ‘‘Effect
M m ⫽ maximum bending moment; of in-plane deformation on lateral buckling.’’ J. Struct. Mech., 3共1兲,
M oa ⫽ M ob /␣ m ; 29– 60.
M ob ⫽ maximum moment at elastic buckling; Wongkaew, K., and Chen, W. F. 共2002兲. ‘‘Consideration of out-of-plane
M on ⫽ nominal out-of-plane moment capacity; buckling in advanced analysis for planar steel frame design.’’ J. Con-
M os ⫽ modified value of M ob ; str. Steel Res., 58, 943–965.

JOURNAL OF STRUCTURAL ENGINEERING © ASCE / JANUARY 2004 / 69

J. Struct. Eng. 2004.130:64-69.