Inelastic lateral buckling of steel beam-columns

P. E. CUK
Wnr;qorl Chnl~rnrrrlPrlrrrlc>r.sPre. Lrrl., Syrlriey. N.S.W.. 2001, Ait.srr.crlirr

M. A. BRADFORD
Depnrrr,7erlr qf S t r ~ ~ c tErlgir~eerir~g.
~ ~ r ~ ~ l Ur~iver:sifyqJ'New Sollrh Wclles, Kerlsirrgtori. N.S.W., 2033. Alrsrrrrlic~
AND

N . S. TKAHAIR'
Departrner~rof Civil Er~gineerirlg,The University of Alherrcr, Erlrrzor~tor~.
Alter., Concrdtr T6G 2G7
Received January 8, 1986
Can. J. Civ. Eng. Downloaded from www.nrcresearchpress.com by University of Waterloo on 04/29/11

Rcviscd manuscript accepted July 24, 1986

A series of investigations of the inelastic lateral buckling behaviour of stccl bcarn-columns is described. 'These originate
from many previous studics of the elastic latcral buckling of bcams. It was found that prcscnt methods of prcdicting the effects
of moment gradient in elastic beam-colun~ns arc unncccssarily conservative, and it was concluded that many practical
continuous beam-columns will have significant warping restraints.
Fourteen inelastic lateral buckling tests were carried out on 9 continuous stccl beam-columns. The results of these tests were
compared with predictions madc by a new and irnproved finitc elcmcnt computer method of analyzing inelastic buckling, and
very good agreement was found. Thc analytical mcthod was then used to develop a sirnplc approximation for predicting the
inelastic buckling of isolated beam-columns with unequal end moments, and a design mcthod was proposed.
Key ~vord.s:bcams, buckling, columns, flexure, residual strcsscs, stccl, structural design, structural engineering, torsion.

L'article dkcrit diffdrentes recherchcs sur le dkvcrscmcnt dcs poteaux-poutres dans Ic domaine semi-plastique. Ccs
recherches sont issues de diffkrentcs Ctudcs sur Ic dkvcrsement dcs poutrcs dans Ic domaine Clastiquc. Ces rcchcrches montrcnt
que les mtthodes actucllcs pcrmettant de prendre en consideration I'influence du diagrarnnlc des moments de flcxion sont, sans
aucune raison, trks stcuritaircs ct, en conclusion, indiqucnt quc la plupart dcs poteaux-poutrcs continus posskdent unc
For personal use only.

rtsistancc au gauchisscment non nCgligcablc.

Quatorze essais de dkversement dans Ic domaine semi-plastique furcnt rCalisCs sur 9 potcaux-poutrcs continus. Lcs rksultats
de ces essais furcnt comparks, avcc succks, aux valcurs obtenucs d'un nouveau programme d'CItments finis analysant Ic
flambcment dans le domainc semi-plastiquc. Cette methodc d'analysc a cnsuite permi d'knoncer unc f a ~ o nsimple mais
approximativc dc prCdire le flambemcnt dans le domaine scmi-plastiquc dcs potcaux-poutrcs soumis h dcs momcnts d'ex-
trCmitCs quelconques et unc mCthodc dc calcul est proposCe.
Mots clCs: poutres, flambcment, potcaux, flcxion. contraintes rCsiduelles, acicr, analysc structurale, science dcs structures,
torsion.
[Traduit par la rcvuc]
Can. J. Civ. Eng. 13. 693-699 (1986)

for predicting the effects of moment gradient than those of

1. Introduction
When a steel beam-column that is bent about its major axis
is insufficiently braced laterally, it may fail by deflecting lat-
erally out of the plane of bending and twisting. For
beam-columns of intermediate slenderness, the in-plane ac-
tions cause yielding, which reduces the resistance to lateral
buckling. This paper describes a series of investigations of
the inelastic lateral buckling of steel beam-columns.
The investigations have their origin in many previous studies
of elastic lateral buckling of beams. In these studies, the effects
of cross section, slenderness, support, moment gradient, load
height, and restraint have been thoroughly researched. Exten-
sions to inelastic buckling of beams have shown the importance
of residual stresses, moment gradient, and the location of yield
regions in both simply supported and continuous beams, and
have led to methods of incorporating these effects into design
procedures.
NOTE: Written discussion of this oaoer is welconled and will be
received by the Editor until March '3i, 1987 (address inside front
cover).
. - .~,-
'At the time of writing, Visiting Professor of Civil Engineering,
University of Alberta, on leave from School of Civil and Mining
Engineering, University of Sydney, N.S.W., 2006, Australia.
present design procedures was required. Following this the
restraining effects caused by concentrated moments in con-
tinuous members were investigated.
The second phase of the investigations was experimental,
and 14 tests were conducted on 9 continuous steel
beam-columns, which buckled inelastically. At the same
time, a new and improved analytical method was developed for
predicting the inelastic lateral buckling of continuous
beam-columns, and tested against the results of previous ana-
lytical studies. The predictions of the new method were then
compared with the experimental results, and a very high degree
of correlation was obtained.
The next phase of the investigations involved the use of the
new analytical method to undertake systematic research into
inelastic lateral buckling. The effects of moment gradient on
isolated beam-columns have been studied, and a significantly
improved design method has been developed. ~ u t u r ework
planned includes the lateral buckling of beam-columns that
sway in the plane of loading, and the buckling interactions
between adjacent segments of continuous beam-columns.
2. Lateral buckling of beams
2.1 Elastic buckling
The elastic flexural-torsional buckling of beams has been
studied by many investigators, and there are a number of re-
 694 CAN.J . CIV. ENG. VOL. 13. 1986

I* L
I
(a) Elevation
I
S u p p o r t s Prevent E n d Rotation (4-0)
and Allow E n d Warping (4"=0)
4
(b) Section
(a) Elevation

(b) Plan on Shear Cenlre Axls Restraint Parameter,

(aL/EIy) / (I+aL/EIy)

(c) Plan on Longitudinal Axis

Can. J. Civ. Eng. Downloaded from www.nrcresearchpress.com by University of Waterloo on 04/29/11

(c) Ellecllve Length Factors

FIG. I . Monosymmetric beam in uniform bending.

FIG.3. Elastic buckling of end-restrained beams.

(a) Beam

(b) Bending Momenl, M, End M o m e n t Ratio, fl

(c) Moment Faclors, m
For personal use only.

FIG. 2. Doubly symn~etricbeams under moment gradient. (a1 Three-Span Beam

DimensionlessLoad,

Q,L:/-

search summaries (Column Research Committee of Japan (b) Crlllcal Load Combinations
1971; Structural Stability Research Council 1976; Trahair
1977a, 6 ; Nethercot 1983). For simply supported beams of FIG.4. Interaction buckling of symmetrical three-span beams of
length L bent in uniform bending in a plane of symmetry as narrow rectangular cross section.
shown in Fig. 1 , the elastic buckling moment M is given by buckling moment can be obtained from [2] and [3] by substi-
tuting the effective length.

where M,,, is the buckling moment for a doubly symmetric for the actual length L, in which the effective length factor k is
beam: approximated by
[21 My: = ~ { P , , ( G J+ T'EI,/L~))
P , is the flexural buckling load of a column:
[3] P, = .rr2E~,/L2 where a is the moment-rotation stiffness of each of the four
flange end restraints.
p, is a monosymmetry property of the cross section: The restraining actions between adjacent segments of braced
or continuous beams are more difficult to assess. since there are
a number of different restraining modes possible, as shown in
Fig. 4. These include the easily analyzed zero-interaction case,
y , is the shear centre coordinate, and El,., GJ, and EI, are the in which all segments buckle simultaneously and independently
flexural, torsional, and warping rigidities of the cross section. of the adjacent segments, so that there are no buckling re-
For doubly symmetric beams bent by unequal end moments straints. An approximate method has been developed for more
M and PM as shown in Fig. 2, the maximum moment at elastic general analysis (Nethercot and Trahair 1976a; Trahair 19770)
buckling can be expressed as in which a lower bound is first produced by assuming that each
[5] ME = rnM,, segment buckles independently and by determining the most
critical segment. The restraining actions of the adjacent seg-
where rn is approximated by ments are then approximated and used to obtain an improved
estimate of the buckling load of the critical segment.
A recent study (Cuk and Trahair 1983) has considered the
The elastic buckling resistance of a beam may be signifi- elastic buckling of continuous beams with concentrated
cantly increased by end restraints (Trahair and Nethercot moments acting at the support points, as shown in Fig. 5. It was
1984). For doubly symmetric beams in uniform bending with found that the jump discontinuities in the bending moment
equal flange end restraints as shown in Fig. 3, the elastic caused unexpected restraint effects, which might be approxi-
 CUK ET AL. 695
Ratio, Q,/Q,
vo
A 0.6
Bending Moment Dlstributlon

Centroidal Deflection

Angle of Twist
Can. J. Civ. Eng. Downloaded from www.nrcresearchpress.com by University of Waterloo on 04/29/11

TOD

Jm
0.3 0.4 0.5 0.6 0.7 0.8 09 1.0 1.1 1.2

Modified Slenderness,

Flange Flange Deflection FIG.7. Inelastic buckling predictions for continuous beams.
FIG.5. Buckled shapes of two-span beams with concentrated rno-
ments.
For personal use only.

0
Computer Solution
C Approximate Eq. 191 -1.0 -0.5 0 0.5 1.0 -1.0 -0.5 0 0.5 1.0

Ratio of End Moments, P Ratio of End Moments, P

mE
g 0 . 0.2 5 04 b 06
1 10.8 1 10 1 121
Modified Slenderness,
FIG.6. Hot-rolled beams with unequal end moments.
mated as equivalent end warping restraints. Recognition of the
fact that concentrated moments will often require significant
web stiffening of the beam (Cuk et al. 1986), which will
produce further warping restraints, suggests that it is not unrea-
sonable to assume that end warping is effectively prevented in
continuous members with concentrated moments.
2.2 Inelastic buckling
The buckling resistance of an intermediate length steel beam
is reduced by yielding caused by a combination of the effects
of the applied loads and the residual stresses left in the beam
after manufacture. A tangent modulus theory of inelastic buck-
ling has been developed (Trahair and Kitipornchai 1972;
Trahair 1983), in which the initial elastic moduli E and G are
replaced by the strain-hardened values E, and G, for all yielded
and strain-hardened regions of the beam, and which accounts
for the nonuniform, monosymmetric nature of the beam after
partial yielding.
For a simply supported steel beam under moment gradient,
the inelastic buckling resistance MI may be approximated
(Nethercot and Trahair 1976b) as shown in Fig. 6 by
(a) \I(TZEI JGJL') = 0.3 (b) J ( r r 2JGJL~)
~~ = 1.0
:1.4
FIG.8. Beam-column factors for unequal end moments.
that the moment distribution is very important, as there are very
substantial reductions in buckling resistance for uniform
bending ( P = - 1) when all the beam is yielded. On the other
hand, the reductions are quite small for double curvature
bending ( P = l), for which yielding is concentrated near the
supports.
For simply supported beams with central concentrated loads
(Fig. 7), the resistance is a little higher than for uniform
bending, because while yielding occurs in the mid-span region
of the beam, it is limited in its extent. A similar conclusion can
be drawn for continuous beams (Yoshida et al. 1977) except in
the special cases where yielding first occurs at the supports, in
which case the inelastic buckling resistance is much higher, as
indicated by the results for Q , / Q 2 = 1.56 shown in Fig. 7.
3. Elastic buckling of beam-columns
The elastic flexural-torsional buckling of a simply sup-
ported beam-column in uniform bending (P = - 1) is approx-
imated by
[lo] (M + Pya)' = (P, - P ) {(ri + yi) (P; - P ) + MP,)
in which Pi is the torsional buckling load of a column

in which Mp is the fully plastic moment capacity. It can be seen and r, is the polar radius of gyration obtained from
 696 CAN. J . CIV. ENG. VOL. 13. 1986
For doubly symmetric sections, yo = 0 and P, = 0, and a more
accurate solution (Trahair 1977u) is obtained from
[13] ( M I M Y , ) ' = (1 - P I P , ) (1 -PIP,.) (1 - P I P T )
in which P, is the in-plane column buckling load
The term (1 - P I P , ) in [13] is often close to unity.
For beam-columns with end moments M and PM, the elas-
tic buckling resistance may be approximated by
Can. J. Civ. Eng. Downloaded from www.nrcresearchpress.com by University of Waterloo on 04/29/11
[I51 (MImM,,)' = (1 - P I P , ) (1 - PIP,)
in which the moment distribution factor m is given by [6].
However, this approximation is often conservative (Cuk and
Trahair 198 I), as indicated in Fig. 8 , and more accurate predic-
tions may be obtained (Cuk and Trahair 198 1; Bradford and
Trahair 1985) using
for I l m in [15].
4. Analysis of inelastic buckling
4.1 Prebucklirzg crrzalysis of in-plcme bending
Before a prediction can be made of inelastic lateral buckling,
For personal use only.
the in-plane bending must be analyzed so that the distributions
of the elastic, yielded, and strain-hardened regions throughout
the member can be determined. The effective out-of-plane
rigidities that contribute to the inelastic buckling resistance can
be evaluated using these distributions.
When the member is statically determinate, the in-plane
analysis can be made in two separate stages. First, the variation
of the axial force and bending moment along the member can
be determined from statics. Following this, the locations of the
boundaries of the elastic, yielded, and strain-hardened regions
within selected cross sections can be determined using the
cross-section geometry, material properties, residual stresses,
and the axial force and bending moment.
When an elastic member is siatically indeterminate, the two
stages cannot be separated, because the material nonlinearity
closes the chain of dependence of yielding on stress resultants,
on redundant actions, on deflections, on stiffnesses, on yield-
ing. In addition, it may be necessary to consider the effects of
geometric nonlinearity, as for example when the term (1 -
P I P , ) is not close to unity.
A finite element computer method of analyzing the in-plane
behaviour of steel frames is discussed by El-Zanaty and Murray
(1983). This method, which allows for the effects of residual
stresses, yielding, strain-hardening, and finite deflections, is
used to determine the yielded and strain-hardened boundaries.
4.2 Analysis of out-of-plarze buckling
Finite element methods of analyzing elastic flexural-
torsional buckling (Hancock and Trahair 1978) may be simply
adapted for the analysis of inelastic buckling. For members
with equal flanges, it is easiest to specify the buckling displace-
ments in terms of the lateral displacement u and twist of the + shown in Fig. 9. The hot-rolled I-section members were loaded
elastic centroidal axis, in which case the strain energy stored in
an element can be expressed as (Bradford et al. 1984)
where the generalized strain vector is
[I81 { E , , } ~ = { u ' ~ , + ~ , + ' ~ }
and the subscript t denotes the tangent modulus values of G J
and the top and bottom flange rigidities El.[. and El,, and the
primes indicates differentiation with respect to z.
The work done by the forces acting on the element during
buckling can be expressed as
[20] V = lL{~,.}'
0
[D,.] { E , } dz
where the generalized stability strain vector is
[211 { E , , } ~= br,+, + I 1
and f is the total normal longitudinal stress and y , is the distance
below the centroidal axis at which the distributed load S; acts.
The element stiffness and stability matrices may be formed
from [17] and [20], and these may be transformed and assem-
bled into the global matrices [K] and [GI in
1
[24] 3 {ulT [K + G] {u} = 0
in which {u} is the vector of the global nodal deformations. In
inelastic buckling problems, [K] and [ G I must be recalculated
for each load level on the structure, and so the usual eigenvalue
methods used for elastic buckling problems lose their effi-
ciency. Instead, a series of calculations are made at increasing
load levels until an approximately zero determinant is obtained
+
from [K GI, which determines the buckling load. Some care
must be taken to ensure that the lowest buckling load is not
missed.
5. Inelastic buckling predictions
5.1 Tests on continuous beam - columns
Cuk et al. (1986) describe a series of 14 tests on 9
beam-columns, which were continuous over three spans, as
by end forces P , and P,, and concentrated in-plane moments
developed by the forces P 4 and P 5 , and were restrained against
in-plane sway by the bracing force P,. These forces caused
significant yielding of the beam-columns, reducing their re-
sistances to o u t - ~ f - ~ l a nbuckling.
e Because of this, the re-
 CUK ET AL. 697

Continuous Beam - Column ation 1984), except that P I ,is the inelastic out-of-plane flexural
I
buckling load of a simply supported column and M l o is the
uniform buckling moment of a simply supported inelastic
beam. Approximations for P l y and M l o were developed (Brad-
ford and Trahair 1985) from inelastic buckling analyses of a
wide range of hot-rolled 1-section members as
[27] PI\./PY = 1.035 - 0.181

FIG.9. Arrangement of three-span beam-column in which P y = A F y is the squash load.

Can. J. Civ. Eng. Downloaded from www.nrcresearchpress.com by University of Waterloo on 04/29/11

It was found that the approximations calculated from [25]

I I
Load
set
Nominal Load Configuralion
1 Specimen
Number
1 (P51P~) I (PIIP~ 1 were generally conservative, and especially so for high mo-
ment gradients (P 2 0.5). This conservatism was attributed to
the use of a linear interaction equation, instead of a parabolic
one similar to [13] and [15], and to the use of a C,,, factor for
nonuniform bending which was independent of the axial load
P , instead of varying with P as in [16].
Because of this it was decided to modify the elastic parabolic
interaction equation and the nonuniform bending factor to
For personal use only.

These equations proved to be of high accuracy, as is demon-

strated in Fig. 1 1.

6. Application to design
Beams and beam-columns that fail by flexural-torsional
buckling must be almost perfectly straight and untwisted before
loading, and the applied loads must initially cause deflections
only in the plane of loading. Real members have initial curva-
tures and twists, and their loads are applied eccentrically and
with components that cause out-of-plane bending and torsion
immediately, as shown in Fig. 12.
Note: 'A' indicates specimen previously tested to failure
Because of this, the strengths of real members are reduced
FIG. 10. Experimental failure conditions. below their buckling resistances. Design rules usually allow for
this by modifying the buckling predictions. For example, for
straining out-of-plane actions developed by weak axis beams simply supported columns (Rotter 1982) the column strength
played important roles in increasing the member strengths. P o may be approximated by
The purpose of the tests was to obtain experimental data that
could be used to evaluate inelastic buckling theories. A com-
parison of the experimental failure loads P , with the predic-
tions P I (Bradford and Trahair 1986) obtained from the theory
developed by Bradford et al. (1984) is shown in Fig. 10, which
indicates extremely close agreement.
in which q is an imperfection parameter given by
5.2 lsolated beam - columns under moment gradient
Inelastic buckling predictions (Bradford and Trahair 1985) [32] q = 0.293 {m)
0.15) - 30
of isolated hot-rolled beam-columns with end moments M and while the strength M o of a beam in uniform bending may be
PM have been compared with approximations obtained from approximated by (Nethercot and Trahair 1983; Trahair 1984)
the linear interaction equation
[33]
Mo
=
M l
0.6 ([(
My,
+l '
31')-
M
My:
L} 1

At present, the flexural-torsional design strengths of

in which beam-columns are approximated by using equations similar to
[25] and [26], but with P I , and M l o replaced by equations
[26] C,, = 0.6 - 0.4P 2 0.4
similar to [31] and [32]. The unsatisfactory nature of [25] and
These equations are similar to those used in present design [26] for inelastic flexural-torsional buckling has been noted
codes, such as CAN3-S 16.1-M84 (Canadian Standards Associ- above, as has the marked improvement provided by [28] and
 C A N . 1. CIV. ENG. VOL. 13, 1986

Computer Solution
- - - - - - Eq. [291
13 M
Can. J. Civ. Eng. Downloaded from www.nrcresearchpress.com by University of Waterloo on 04/29/11

FIG. I I. Improved buckling interaction equation

inelastic buckling of beams has also been studied, including the

For personal use only.

effects of residual stresses and the yield distribution.

Elastic Buckling of
Recent research studies have extended this work to the in-
M~"--------
/- Perfect Member
.......'.. . ................
, , ......;;F elastic flexural-torsional buckling of steel beam-columns
+
C
.... and have led to the development of a general computer method
,. ,,.:."" of predicting inelastic buckling, and this has received experi-
E
0
M~
/Elastic Bending of
Imperfect Member mental confirmation. The computer method has been used to
r . ---- study the inelastic buckling of beam-columns with unequal
end moments and to develop comparatively simple equations
Perfect Member for predicting their inelastic buckling resistances. This has al-
lowed the formulation of an improved method of estimating
their design out-of-plane member strengths. Thus a

0
I h e a l Behaviour
beam-column bent in-plane would be checked for cross-
section capacity and in-plane member strength as at present,
Lateral Deflection, Twist and for out-of-plane strength using the new formulation.
FIG. 12. Behaviour of real members. Acknowledgements
[291 It seems logica1 therefore to propose (Trahair 1986) that The work described in this paper has been supported by
this should be extended to estimate the out-of-plane design research grants made under the Australian Research Grants
strengths of beam-columns from Scheme, by the Natural Sciences and Engineering Research
Council of Canada. and bv the Universitv of Alberta. and has
been assisted by ~ r b f e s s o r ' ~W.
. ~ u r r a i o the
f ~ e ~ a i t m eof
nt
Civil Engineering of the University of Alberta.
in which Cbc,M o , and Po are given by [30], [31], and [33]. BRADFORD, M. A., and TRAHAIR, N. S. 1985. Inelastic buckling of
Thus the design of beam-columns will require three condi- beam-columns with unequal end moments. Journal of Construc-
tions to be satisfied: tional Steel Research, 5(3), pp. 195-212.
(1) cross-section capacity, 1986. Inelastic buckling tests on beam-columns. ASCE
(2) in-plane member strength, and Journal of Structural Engineering, 112(3), pp. 538-549.
(3) out-of-plane member strength (eq. [34]) BRADFORD, M. A,, CUK,P. E., GIZEJOWSKI, M. A., and TRAHAIR,
for which the present methods may be retained for assessing the N. S. 1984. Inelastic lateral buckling of beam-columns. School of
first two conditions of cross-section capacity and in-plane Civil and Mining Engineering, University of Sydney, Sydney,
strength. N.S.W., Australia, Research Report No. R474, pp. 1-34.
CANADIAN STANDARDS ASSOCIATION. 1984. Steel structures for build-
7. Conclusions ings (limit states design). CAN3-S 16. I-M84, CSA, Rexdale, Ont.
pp. 1-163.
Previous studies of the elastic flexural-torsional buckling of COLUMN RESEARCH COMMITTEE OF JAPAN.197 1. Handbook of struc-
beams and beam-columns have demonstrated the importance tural stability. Corona Publishing Company, Tokyo, Japan.
of the bending moment distribution and of end restraints. The CUK, P.E., and TRAHAIR,N.S. 1981. Elastic buckling of
 CUK ET AL

beam-colunins with unequal end moments. Civil Engineering beam-column factors for unequal end moments
Transactions, lnstitution of Engineers. Austrz~lia,CE23(3), pp.
166-171. element matrices (eqs. [I91 and [22])
1983. Buckling of bcarns with concentrated moments. ASCE Young's modulus of elasticity
Journal of Structural Engineering, 109(6), pp. 1387- 1401. strain-hardening modulus
CUK, P. E., ROGERS,D. F., and TRAHAIR, N. S. 1986. Inelastic total longitudinal stress
buckling of continuous stccl beam-columns. Journal of Construc- yield stress
tional Stcel Research. 6(1), pp. 2 1 -52. global stability matrix
EL-ZANATY,M. H., and MURRAY, D. W. 1983. Non-linear finitc shear modulus of elasticity
elcmcnt analysis of stcel framcs. ASCE Journal of Structural strain-hardening shear modulus
Enginecring, 109(2), pp. 353-368. distance between flange centroids
HANCOCK, G. J., i~ndTRAHAIR, N. S. 1978. Finite clcmcnt analysis second moments of area of bottom and top flanges
of thc lateral buckling of continuously rcstraincd bcam-columns.
warping section constant
Can. J. Civ. Eng. Downloaded from www.nrcresearchpress.com by University of Waterloo on 04/29/11

Civil Engineering Transactions, Institution of Enginecrs, Australia,

CE20(2), pp. 120- 127. second moments of area about x- and y-axes
NETHERCOT, D. A. 1983. Elastic latcral buckling of beams. 111 Beams torsion section constant
and beam columns. Editerl by R. Narayanan. Applicd Scicncc Pub- effective length factor
lishers, Barking, England, pp. 1-33. global stiffness matrix
NETHERCOT, D. A,. and TRAHAIR, N. S . 19760. Latcral buckling effective length
approximations for elastic beams. The Structural Enginccr, 54(6), length of member or element
pp. 197-204. beam factor for moment gradient
197617. Inelastic lateral buckling of dcterminatc beams. ASCE moment
Journal of the Structural Division, 102(ST4), pp. 701 -717. elastic buckling moment
1983. Design of laterally unsupportcd beams. In Bcams and
beam columns. Edited by R. Narayanan. Applicd Science Publish-
inelastic buckling moment
ers, Barking, England, pp. 71 -94. value of M ,for uniform bending
ROTTER,J. M. 1982. Multiple column curves by modifying factors. uniform bending strength
ASCE Journal of the Structural Division, 108(ST7), pp. full plastic moment
1665-1669. value of M E for uniform bending
axial load
For personal use only.

STRUCTURAL STABILITY RESEARCH COUNCIL. 1976. Guide to stability

design criteria for metal structures. 3rd ed. Edited by B. G. John- forces on beam-column
ston. John Wiley and Sons, New York, NY. force at failure
TRAHAIR, N.S. 1977a. The behaviour and design of steel structures. inelastic buckling load
Chapman and Hall, London, England. value of PI for flexural buckling
19776. Lateral buckling of bcarns and beam-columns. In
Thcory of beam-columns. Vol. 2. Edited t>y W. F. Chen and T.
out-of-plane column strength
Atsuta. McGraw-Hill Book Co., New York. NY, Chap. 3. elastic buckling loads for flexure about x- and
1983. Inelastic latcral buckling of beams. 111 Beams and beam y-axes
columns. Erlited by R. Narayanan. Applicd Science Publishers, squash load
Barking. England. pp. 35-69. elastic torsional buckling load
1984. Lateral buckling design strength of stccl beams. Civil transverse loads
Engineering Transactions, Institution of Engineers, Australia, value of Q at elastic buckling
CE26(4), pp. 319-326. value of Q at nominal first yield
1986. Design strengths of steel bcani-columns. Canadian polar radius of gyration
Journal of Civil Engineering, 13, this issue. stress resultants (eq. [23])
TRAHAIR, N. S., and KITIPORNCHAI, S. 1972. Buckling of inelastic
lateral deflection of shear centre
I-beams under uniform moment. ASCE Journal of the Structural
Division, 98(STI I), pp. 2551 -2566. vector of global nodal displacements
TRAHAIR, N. S., and NETHERCOT, D. A. 1984. Bracing requirements strain energy stored in element
in thin-wallcd structures. In Developments in thin-walled stiuc- work done on element
tures. Edited by J. Rhodes and A. C. Walker. Applied Science principal axes of cross section
Publishers, Barking, England, pp. 93- 130. shear centre coordinate
YOSHIDA, H., NETHERCOT, D. A,, and TRAHAIR, N. S. 1977. Anal- distance below centroid of distributed load S''
ysis of inelastic buckling of continuous beams. Proceedings of the longitudinal axis through centroid
International Association for Bridge and Structural Enginecring No. stiffness of flange end restraints
P-3/77, pp. I - 14. end moment ratio
monosymmetry section constant
strain vectors
List of symbols angle of twist rotation
cross-sectional area imperfection parameter (eq. [32])