adt-NU20040517 15270201whole

COMPRESSION STABILITY
OF HIGH STRENGTH STEEL SECTIONS WITH

LOW STRAIN-HARDENING
by
Demao YANG
Thesis presented for the Degree of

Doctor of Philosophy
School of Civil and Mining Engineering

University of Sydney
2003
SYNOPSIS
SYNOPSIS
Thin-walled steel sections made from high strength thin cold-reduced G550
steel to Australian Standard AS 1397-1993 under compression are investigated
experimentally and theoretically in this thesis.
This thesis describes three series of compression tests performed on box-section

stub columns, box-section long columns and lipped channel section columns
cold-formed from high strength steel plates in 0.42 mm or 0.60 mm thickness
with nominal yield stress of 550 MPa. The tests presented in this thesis formed
part of an Australian Research Council research project entitled "Compression
Stability of High Strength Steel Sections with Low Strain-Hardening".
For the fix-ended stub column tests, a total of 94 lipped-square and hexagonal
section stub columns were tested to study the influence of low strain hardening
of G550 steel on the compressive section capacities of the column members. For
the pin-ended long column tests, a total of 28 box-section columns were tested
to study the stability of members with sections which undergo local instability
at loads significantly less than the ultimate loads. For the fix-ended lipped
channel section columns, a total of 21 stub and long columns were tested to
study the failure resulting from local and distortional buckling with interaction
between the modes.
A numerical simulation on the three series of tests using the commercial finite
element computer program ABAQUS is also presented as part of this thesis.
The post-buckling behaviour of thin-walled compression members is
i
SYNOPSIS
investigated. The effect of changing variables, such as geometric imperfections

and end boundary conditions is also investigated. The ABAQUS analysis gives
accurate simulations of the tests and is in good agreement with the experimental
results.
Theoretical studies using finite strip methods are presented in this thesis to
investigate the buckling behaviour of cold-formed members in compression.
The theoretical studies provide valuable information on the local and
distortional buckling stresses for use in the interaction buckling studies. The
finite strip models used are the semi-analytical and spline models.
As expected from the stub columns tests, the greatest effect of low strain
hardening was for the stockier sections where material properties play an
important role. For the more slender sections where elastic local buckling and
post-local buckling are more important, the effect of low strain hardening does
not appear to be as significant. The pin-ended and fix-ended long column tests
show that interaction, which is between local and overall buckling in the box
sections, and between local and distortional buckling in the open channel
sections, has a significant effect on their member capacities.
The results of the successful column tests and ABAQUS simulation have been
compared with the design procedures in the Australian/New Zealand Standard
for Cold-Formed Steel Structures (AS/NZS 4600) and the North American
Specification for Cold-Formed Steel Structural Members prepared by the
American Iron and Steel Institute. The stub column tests show that the current
design rules give too conservative predictions on the compressive section
capacities of the column members; whereas the long column tests show that the
ii
SYNOPSIS
current column design rules are unconservative if used in their current form for
G550 steel.
Three design proposals are presented in this thesis to account for the effects of
high strength thin steels on the section and member capacities.
iii
PREFACE
PREFACE
This thesis is submitted to the University of Sydney, Australia, for the degree of
Doctor of Philosophy. The work described in this thesis was carried out by the
candidate during the years 1999 to 2003 in the Department of Civil Engineering
at the University of Sydney under the supervision of Professor G.J. Hancock,
BHP Steel Professor of Steel Structures.
In accordance with The Bye-Laws of the University of Sydney governing the

requirements for the degree of Doctor of Philosophy, the candidate submits that
the work presented in this thesis is original unless otherwise referenced within
the text. The complete experimental programs performed including the design
of the specimens and imperfection measurement rig are claimed as original. The
simulations of the complete experimental programs, in particular the analysis of
stress distribution on the lipped channel cross-section, are also claimed as
original. The design proposals for local buckling, interaction of local and overall
buckling, and interaction of local and distortional buckling are claimed as
original.
Twelve supporting papers and three research reports which are based on the
work presented in this thesis have been written. They are:
1. Compression Tests of Box-shaped Cold-Reduced High Strength Steel

Sections, Proceedings, 6th Pacific Structural Steel Conference, Beijing,
October, 2001, Seismological Press (Yang and Hancock)
2. Stability and Ductility of Thin-High Strength G550 Steel Members and

Connections, Keynote paper, Proceedings of the 3rd International Conference
iv
PREFACE
on Thin-Walled Structures, Kracow, Poland, (published as Thin-Walled

Structures-Advances and Developments-Elsevier 2001) (also published in
Thin-Walled Structures Vol. 41, 2003.) (Rogers, Yang, Hancock)
3. Compression Tests of Cold-Reduced High Strength Steel Stub Columns,

Proceedings, 16th International Specialty Conference on Cold-Formed Steel
Structures in Orlando, October, 2002 (Yang and Hancock)
4. Compression Tests of Cold-Reduced High Strength Steel Long Columns,.

Proceedings, 16th International Specialty Conference on Cold-Formed Steel
Structures in Orlando, October, 2002 (Yang, Hancock and Rasmussen)
5. The Behaviour of High Strength G550 Steel Sections as Used in Residential

Construction, Keynote paper, Proceedings, 2nd International Symposium on
Steel Structures, Seoul Korea, November, 2002 (Hancock, Rogers and Yang)
6. Stability of High Strength G550 Steel Compression Members, Keynote

paper, Proceedings of the 3rd International Conference on Advances in Steel
Structures, Hong Kong, December, 2002 (Yang and Hancock)
7. Compression Tests of Cold-Reduced High Strength Steel Channel Columns

Failing in the Distortional Mode, Proceedings of the 5th International
Conference on Steel and Aluminium Structures (ICSAS) & 7th International
Conference on Steel Concrete Composite Structures (ASSCCS03), Sydney,
June, 2003 (Yang and Hancock)

University of Sydney, Department of Civil Engineering, Research Report
R815, 2002 (Yang and Hancock)
v
PREFACE
9. Compression Tests of Cold-Reduced High Strength Steel Long Columns,

University of Sydney, Department of Civil Engineering, Research Report
R816, 2002 (Yang, Hancock and Rasmussen)

failing in the Distortional Mode, University of Sydney, Department of Civil
Engineering, Research Report R825, 2003 (Yang and Hancock)

(accepted for publication in ASCE) (Yang and Hancock)
12. Compression Tests of Cold-Reduced High Strength Steel Long Columns,

(accepted for publication in ASCE) (Yang, Hancock and Rasmussen)

failing in the Interaction between Local and Distortional Modes, (accepted
for publication in ASCE) (Yang and Hancock)
14. Developments in Design for Distortional Buckling of Thin-Walled

Members, Keynote paper, The Fourth International Conference on Thin-
Walled Structures, ICTWS 4, Loughborough, England, UK, June, 2004
(Yang and Hancock)
15. Numerical Simulations of High Strength Steel Box-Shaped & Channel

Columns, Proceedings, 17th International Specialty Conference on Cold-
Formed Steel Structures in Orlando, November, 2004 (will be submitted)
(Yang and Hancock)
vi
ACKNOWLEDGMENTS
ACKNOWLEDGEMENTS
I wish to express my sincerest thanks to my supervisor Professor Gregory J.

Hancock for his supervision, enthusiastic guidance and continual
encouragement throughout the course of my candidature. I am indebted to his
understanding, tolerance and precious support.
The scholarship provided by a joint Department of Civil Engineering and Centre

for Advanced Structural Engineering Scholarship during the course of this work
is gratefully acknowledged.
This thesis forms part of an ARC research project entitled “Compression

Stability of High Strength Steel Sections with Low Strain-Hardening” being
carried out in the Department of Civil Engineering at the University of Sydney.
I would like to thank the Australian Research Council and BHP Coated Steel
Australia for their financial support for these projects performed at the
University of Sydney.
The tensile specimens were milled in the William and Agnes Bennett
Supersonics Laboratory in the Department of Aeronautical Engineering. The
compression specimens were fabricated in the J.W. Roderick Laboratory for
Materials and Structures in the Department of Civil Engineering. I would like to
thank Mr. Todd Budrodeen for fabricating the specimens and designing the rig
for imperfection measurement. The finite element analyses were carried out on
UNIX terminal using ABAQUS in the Department of Civil Engineering at the
University of Sydney. I wish to thank Dr Tim Wilkinson for his advice and
assistance on using ABAQUS.
vii
ACKNOWLEDGMENTS
I wish to thank Associate Professor Kim J.R. Rasmussen for his advice and
suggestion on this work. My thanks go to Dr. Young Kwon, Dr. John
Papangelis, Dr. Lip Teh and Dr. Ben Young for their advice and suggestions.
My thanks go to Ms. Gwenda McJannet for her help on various aspects of the
presentation of this work.
My thanks go to my friends and colleagues in the Department of Civil

Engineering at the University of Sydney for their friendship and valuable
discussions, in particular Michael Bambach, Michael Dong, Elisha Harris,
Kelvin Ye and Wilson Yuan.
I wish to thank my parents and my brothers for their support, encouragement

and patience throughout the years of my postgraduate study.
I wish to thank my wife, Xiangyun Zhao, and my daughter, Zhaoyu Yang, for
their love, emotional support and encouragement as well as their appreciation of
my study. I am indebted to my wife so much for looking after our daughter for 4
years in China without my presence.
This thesis is dedicated to my parents Mr. Nianshen Yang and Mrs. Mingxiang
Han.
Demao Yang
viii
CONTENTS
CONTENTS
SYNOPSIS ................................................................................................................................ I
PREFACE.............................................................................................................................. IV
ACKNOWLEDGEMENTS ................................................................................................VII
CONTENTS........................................................................................................................... IX
NOTATION.........................................................................................................................XIV
CHAPTER 1
INTRODUCTION....................................................................................................................1
1.1 STATEMENT OF THE PROBLEM .........................................................................2
1.2 OBJECTIVES OF THIS THESIS..............................................................................4
1.3 OUTLINE OF THIS THESIS.....................................................................................5

1.3.1 General............................................................................................................................................... 5
1.3.2 Experimental investigation ................................................................................................................. 5
1.3.3 Numerical investigation...................................................................................................................... 6
1.3.4 Theoretical studies using finite strip methods .................................................................................... 7
1.3.5 Design recommendations using the effective width method and direct strength method. .................. 8
1.4 FIGURES....................................................................................................................10
CHAPTER 2
LITERATURE REVIEW .....................................................................................................12
2.1 RESEARCH ON HIGH STRENGTH STEEL .......................................................13
2.2 LOCAL BUCKLING, POST-BUCKLING OF PLATES AND INTERACTION

OF PLATE ASSEMBLIES...................................................................................19
2.2.1 General............................................................................................................................................. 19
2.2.2 Local buckling of plate ..................................................................................................................... 19
2.2.3 Post-buckling and interaction of plate assemblies ........................................................................... 21
2.3 BUCKLING OF COMPRESSION MEMBERS.....................................................25

2.3.1 General............................................................................................................................................. 25
2.3.2 Local buckling .................................................................................................................................. 25
2.3.3 Overall buckling ............................................................................................................................... 27
2.3.4 Distortional buckling........................................................................................................................ 30
2.4 INTERACTION OF MODES...................................................................................35

2.4.1 General............................................................................................................................................. 35
ix
CONTENTS
2.4.2 Interaction of local and overall buckling ......................................................................................... 35

2.4.3 Interaction of local and distortional buckling .................................................................................. 38
2.4.3.1. Linear interaction buckling................................................................................................................. 39
2.4.3.2. Non-linear interaction buckling.......................................................................................................... 39
2.4.4 General studies on mode interaction................................................................................................ 40
2.5 NUMERICAL METHODS .......................................................................................44

2.5.1 Finite element method ...................................................................................................................... 44
2.5.2 Finite strip method ........................................................................................................................... 47
2.6 SUMMARY ................................................................................................................53
2.7 FIGURE ......................................................................................................................56
CHAPTER 3
STRENGTH TESTS OF COLD-FORMED BOX-SHAPED-SECTION STUB

COLUMNS .................................................................................................................57
3.1 INTRODUCTION......................................................................................................58
3.2 TEST SPECIMENS ...................................................................................................58

3.2.1 General............................................................................................................................................. 58
3.2.2 Labelling........................................................................................................................................... 60
3.2.3 Geometric Imperfection Measurements............................................................................................ 61
3.3 MATERIAL PROPERTIES .....................................................................................61

3.3.1 Manufacturing process of G550 sheet steel...................................................................................... 61
3.3.2 Steel sheet thickness ......................................................................................................................... 63
3.3.3 Coupon Test specimens and procedures........................................................................................... 64
3.3.4 Coupon test results ........................................................................................................................... 64
3.4 STUB COLUMN TESTS ..........................................................................................65

3.4.1 Testing .............................................................................................................................................. 65
3.4.2 Test specimen behaviour, ultimate load-capacity and experimental local buckling loads............... 65
3.5 ANALYSES ................................................................................................................67

3.5.1 Elastic local buckling analyses......................................................................................................... 67
3.5.2 Test results and comparisons with design standards........................................................................ 68
3.5.2.1 General...................................................................................................................................................... 68
3.5.2.2 B-sections.................................................................................................................................................. 70
3.5.2.3 LB-sections ............................................................................................................................................... 71
3.5.2.4 HB-sections............................................................................................................................................... 71
3.5.2.5 Analyses of effective area ......................................................................................................................... 72
3.5.2.6 Explanation of some results ...................................................................................................................... 72
3.6 SUMMARY ................................................................................................................75
3.7 FIGURES....................................................................................................................77
3.8 TABLES......................................................................................................................90
CHAPTER 4
STRENGTH TESTS OF COLD-FORMED BOX-SHAPED-SECTION LONG

COLUMNS ...............................................................................................................106
x
CONTENTS
4.1 INTRODUCTION....................................................................................................107
4.2 TEST SPECIMENS .................................................................................................107

4.2.1 General........................................................................................................................................... 107
4.2.2 Labelling......................................................................................................................................... 109
4.2.3 Geometric Imperfection Measurements.......................................................................................... 109
4.2.4 Material Properties ........................................................................................................................ 110
4.3 LONG COLUMN TESTS .......................................................................................111

4.3.1 Testing ............................................................................................................................................ 111
4.3.2 Measurement of flexural rigidity (E*I)........................................................................................... 112
4.3.3 Test specimen behaviour and ultimate load-capacity..................................................................... 112
4.4 ANALYSES ..............................................................................................................114

4.4.1 Elastic local buckling analyses....................................................................................................... 114
4.4.2 Test Results compared with the Design Standards ......................................................................... 115
4.4.2.1 General.................................................................................................................................................... 115
4.4.2.2 Test results comparisons with design standards ...................................................................................... 116
4.5 SUMMARY ..............................................................................................................116
4.6 FIGURES..................................................................................................................118
4.7 TABLES....................................................................................................................134
APPENDIX 4.1 ................................................................................................................139
CHAPTER 5
STRENGTH TESTS OF COLD-FORMED LIPPED CHANNEL SECTION

COLUMNS ...............................................................................................................141
5.1 INTRODUCTION....................................................................................................142
5.2 TEST SPECIMENS .................................................................................................142

5.2.1 General........................................................................................................................................... 142
5.2.2 The determination of the intermediate stiffeners in the flanges...................................................... 144
5.2.3 The determination of the length of the flange stiffeners ................................................................. 144
5.2.4 The determination of the length of the long specimens................................................................... 145
5.2.5 Labelling......................................................................................................................................... 146
5.3 COLUMN TESTS....................................................................................................147

5.3.1 General........................................................................................................................................... 147
5.3.2 Stub column tests ............................................................................................................................ 148
5.3.3 Long column tests ........................................................................................................................... 149
5.3.4 Test specimen behaviour ................................................................................................................ 150
5.3.5 The effect of buckling modes on ultimate load-capacity................................................................. 151
5.3.6 Failure modes analysis................................................................................................................... 153
5.4 TEST RESULTS COMPARED WITH DESIGN METHODS ...........................153

5.4.1 General........................................................................................................................................... 153
5.4.2 The Effective Width Method (EWM)............................................................................................... 154
5.4.3 AS/NZS 4600 Clause 3.4.6(b) and Kwon & Hancock Equation ..................................................... 155
5.4.4 The Direct Strength Method (DSM) ............................................................................................... 157
xi
CONTENTS
5.5 SUMMARY ..............................................................................................................160
5.6 FIGURES..................................................................................................................162
5.7 TABLES....................................................................................................................187
CHAPTER 6
NUMERICAL SIMULATIONS .........................................................................................192
6.1 GENERAL DESCRIPTION OF ABAQUS...........................................................193

6.1.1 Element type ................................................................................................................................... 193
6.1.2 Material behaviour......................................................................................................................... 194
6.1.3 Boundary conditions....................................................................................................................... 195
6.1.4 Geometrical imperfection ............................................................................................................... 197
6.1.5 Eigenvalue buckling prediction and post-buckling analysis........................................................... 199
6.1.6 Solution Control Parameters.......................................................................................................... 200
6.2 SIMULATION OF STUB COLUMNS..................................................................201

6.2.1 General........................................................................................................................................... 201
6.2.2 Element type ................................................................................................................................... 201
6.2.3 ABAQUS Results ............................................................................................................................ 202
6.2.4 ABAQUS results compared with Test results and Design Standards ............................................. 203
6.2.4.1 General.................................................................................................................................................... 203
6.2.4.2 B-sections................................................................................................................................................ 204
6.2.4.3 LB-sections ............................................................................................................................................. 204
6.2.4.4 HB-sections............................................................................................................................................. 205
6.2.4.5 Explanation of some results .................................................................................................................... 205
6.3 SIMULATION OF LONG COLUMNS.................................................................206

6.3.1 General........................................................................................................................................... 206
6.3.2 Element type and Eccentricity Introduced...................................................................................... 207
6.3.3 ABAQUS Results ............................................................................................................................ 208
6.3.4 ABAQUS results compared with Test Results and Design Standards ............................................ 208
6.4 SIMULATION OF LIPPED CHANNEL SECTIONS.........................................209

6.4.1 General........................................................................................................................................... 209
6.4.2 Element type ................................................................................................................................... 209
6.4.3 ABAQUS results ............................................................................................................................. 210
6.4.4 Load-axial shortening curves ......................................................................................................... 210
6.4.5 Deformed shape and stress distribution ......................................................................................... 211
6.5 SUMMARY ..............................................................................................................213
6.6 FIGURES..................................................................................................................216
6.7 TABLES....................................................................................................................246
CHAPTER 7
DESIGN PROPOSALS .......................................................................................................251
7.1 INTRODUCTION....................................................................................................252
xii
CONTENTS
7.2 REDUCTION FACTOR MODIFIED TO 0.9FY ..................................................252

7.2.1 Limitations on the use of high strength steels................................................................................. 252
7.2.2 Brief description of the comparison of the test and ABAQUS results with Design Standards ....... 254
7.2.3 Modified reduction factor............................................................................................................... 256
7.2.4 Reliability study .............................................................................................................................. 257
7.3 INTERACTION OF LOCAL AND OVERALL BUCKLING............................261

7.3.1 The strength prediction of members subjected to flexural buckling ............................................... 261
7.3.3 Modified design curve .................................................................................................................... 265
7.4 INTERACTION OF LOCAL AND DISTORTIONAL BUCKLING.................266

7.4.1 The strength prediction of singly-symmetric sections subjected to distortional buckling .............. 267
7.4.3 Proposed design methods ............................................................................................................... 273
7.4.3.1 Method 1: based on Kwon & Hancock Equation .................................................................................... 274
7.4.3.2 Method 2: based on AS/NZS 4600 Clause 3.4.6(b) ................................................................................ 275
7.4.3.3 Comparison with the experimental results .............................................................................................. 275
7.5 SUMMARY ..............................................................................................................276
7.6 FIGURES..................................................................................................................280
7.7 TABLES....................................................................................................................287
CHAPTER 8
CONCLUSIONS ..................................................................................................................289
8.1 GENERAL................................................................................................................290
8.2 EXPERIMENTAL INVESTIGATIONS ...............................................................291

8.2.1 General........................................................................................................................................... 291
8.2.2 Box-shaped-section stub columns tests........................................................................................... 291
8.2.3 Box-shaped-section long columns tests .......................................................................................... 292
8.2.4 Lipped channel section columns tests............................................................................................. 293
8.2 NUMERICAL SIMULATIONS .............................................................................295
8.3 THEORETICAL STUDIES USING FINITE STRIP METHODS .....................297
8.4 DESIGN RECOMMENDATIONS ........................................................................298
8.5 FUTURE RESEARCH............................................................................................299
CHAPTER 9
REFERENCE.......................................................................................................................300
xiii
NOTATION
NOTATION
LATIN LETTERS
A cross-sectional area (mm2)

AB ultimate load of ABAQUS (kN)
Ae effective of area (mm2)
Aeq equivalent effective of area (mm2)
b web width (mm)
beq equivalent effective of width of web (mm)
CP correction factor
d lip width (mm)
D dead load
Dm mean dead load
Dn nominal dead load
e eccentricity (mm)
E Young's modulus of elasticity (MPa)
fcrl elastic local buckling stress (MPa)
fd elastic distortional buckling stress (MPa)
fn design stress (MPa)
fol elastic buckling stress (MPa)
fy yield stress (MPa)
F stress function defining the median fiber stress of the plate
Fm fabrication factor mean value
h flange width (mm)
hl glued lips thickness (mm)
k plate buckling coefficient
le effective length of column (mm)
lx, lx0 length of column (mm)
xiv
NOTATION
L length of column (mm) or Live load

Lm mean live load
Ln nominal live load
Lt length of pin-ended column (mm)
Mm material factor mean value
n number of tests.
Nc nominal member compression capacity (kN)
Ncred reduced member capacity (kN)
Nl , Nd theoretical local and distortional buckling loads (kN)
Nol elastic buckling load (kN)
Ns nominal section compression capacity (kN)
Ns0.75 nominal section compression capacity based on 0.75 fy (kN)
NsRb nominal section compression capacity based on Rbfy (kN)
Pc , Pkh distortional buckling strengths (kN)
Pcr elastic buckling load of test (kN)
Pcrl, Pcrd, elastic local and distortional buckling load respectively (kN)
Pe Euler load (kN)
Pn nominal axial strength (kN)
Pne inelastic long column buckling load (kN)
Pnl , Pnd resulting limiting strengths (kN)
Pnld nominal interaction axial strength (kN) accounting for
interaction of local and distortional buckling
Pt ultimate test load (kN)
Py squash load (kN)
Pm professional factor mean value
Q total load
Qn nominal load effect
Qm mean load effect
xv
NOTATION
r radius of corner (90o) (mm) or radius of gyration of

cross-section
rx radius of gyration of the cross-section (mm)
R radius of corner (135o) (mm)
Rn nominal resistance.
t thickness (mm)
tb thickness of base metal (mm)
tc thickness of coated metal (mm)
VD dead load coefficient of variation
VF fabrication factor coefficient of variation
VL live load coefficient of variation
VM material factor coefficient of variation
VP professional factor coefficient of variation
VR mean resistance coefficient of variation
GREEK LETTERS
β0 current target reliability (safety) index

β0 calculated target reliability (safety) index
γ0, γ reduction factor of radius of gyration
γD, γL load factors for dead and live load respectively
φ resistance (capacity) factor
φ0 current resistance factor
φc calculated resistance factor
xvi
Chapter 1: INTRODUCTION
CHAPTER 1
INTRODUCTION
CONTENTS
1.1 STATEMENT OF THE PROBLEM .........................................................................2
1.2 OBJECTIVES OF THIS THESIS..............................................................................4
1.3 OUTLINE OF THIS THESIS.....................................................................................5

1.3.1 General............................................................................................................................................... 5
1.3.2 Experimental investigation ................................................................................................................. 5
1.3.3 Numerical investigation...................................................................................................................... 6
1.3.4 Theoretical studies using finite strip methods .................................................................................... 7
1.3.5 Design recommendations using the effective width method and direct strength method. .................. 8
1.4 FIGURES....................................................................................................................10
1
CHAPTER 1
INTRODUCTION
1.1 Statement of the problem
The use of high strength steels with yield stress values up to 550 MPa is
increasing rapidly, particularly for steel framed houses with sections as thin as
0.4 mm. Steels with high yield stress usually have little or no strain hardening in
the stress-strain curve, and low ductility unlike conventional structural steel that
is highly ductile and strain hardens as shown in Fig.1.1. Strain hardening is
important in the stability of thin-walled sections and so the high strength steels
are likely to have their stability significantly affected by the lack of strain
hardening.
For high strength steel sections made from thin zinc-coated or aluminium/zinc-
coated cold-reduced steel to Australian Standard AS 1397-1993, no specific
investigation has been performed. Mainly due to lack of knowledge on their
structural behaviour, the 1996 Australia/New Zealand Standard AS/NZS 4600
for Cold-Formed Steel Structures and the 2001 North American Specification
(NAS) for Cold-Formed Structural Members have generally limited the design
stress for high strength low ductility steels to 75 percent of their yield stress or
tensile strength as applicable. The NAS further restricts the use of this steel to
multiple web configurations such as sheeting and decking.
A research project on these steels in tension, which was carried out by Rogers
and Hancock (1996), has shown that they have substantially reduced ductility
but this may not affect the net section strength of perforated sections. Steels of
2
this type are similar to Structural Grade 80 steels in the USA according to the
ASTM A653 (1997) and ASTM A792 (1994) standards. A research project led
by Professor W-W Yu at the University of Missouri-Rolla to investigate the
strength of ASTM A653 steel when formed into decking sections and subjected
to bending has demonstrated that their local and post-local buckling capacities
may be significantly influenced by the lack of strain hardening. In particular, the
ultimate moments of panels with slender sections (b/t>100) were lower than the
design moments calculated based on a conventional effective section model.
However, no significant definitive testing has been performed for sections
composed of AS 1397 steel in compression. The AS 1397 steel may be zinc-
coated or aluminum-zinc coated. Those studied in this thesis were aluminum-
zinc coated similar to ASTM A792.
Normally, one of three basic types of buckling, local, overall, and distortional,
can occur in thin-walled steel sections as shown in Fig. 1.2. However, the basic
types may interact with each other. For doubly symmetric box sections, local
and overall buckling or interaction between them rather than distortional
buckling may occur. It is customary to consider that a column may buckle in
either one of two ways: (a) by plate buckling of its component webs and flanges
in shorter half-waves (local or plate buckling) or (b) by deflection of the entire
column in a half-wave of length equal to the effective column length (overall
buckling). For a given column, buckling is supposed to occur at the lower of the
two critical stresses, local or overall. In reality, however, there is an interaction
between these two modes of buckling, so that the failure stress will be smaller
than the overall buckling stresses even if the column has no imperfections.
Imperfections play a significant role in interaction buckling. Column strength is
characterized by the maximum axial force that can be supported without
excessive lateral deformations. In view of the fact that post-buckling strength of
a flat plate is available for structural members to carry additional load, cold-
3
formed steel sections are normally designed on the basis of the post-buckling
strength of the plate elements rather than based on the local buckling stress.
In practice, singly symmetric open sections, such as channel-sections, are

commonly used in cold-formed steel design. For sections of this type, especially
for sections fabricated from such thin (less than 1 mm) sheet steel, distortional
buckling can occur and becomes a significant failure mode. The wavelength of
distortional buckling is generally intermediate between that of local buckling
and overall buckling. Research into the distortional mode of buckling has
attracted considerable attention in recent years. However, as discussed in Kwon
and Hancock (1992), the design method in the AISI Specification is
unconservative for distortional buckling of channel sections composed of high
strength steel of yield stress 500MPa, and so alternative design methods are
necessary to design against distortional buckling in this case. Design rules for
distortional buckling of compression and flexural members were included in the
1996 Edition of AS/NZS 4600.
1.2 Objectives of this thesis
The main objective of this thesis is to investigate the stability of high strength
steel sections and to determine the influence of the lack of strain hardening on
their capacity to resist buckling, particularly in the inelastic range. Several
research projects have been conducted which clearly indicate that the strength of
high strength sections is reduced by the lack of strain hardening. For high
strength quench and tempered steels to ASTM A514 fabricated by welding to
form conventional box, cruciform and I-sections of moderate thickness, a
research program on local instability by Rasmussen and Hancock (1992) has
shown that the lack of strain hardening influences the strength of the sections,
4
particularly for stockier sections where the lack of strain hardening eliminates
the usual rise in strength above the squash load which is the product of the
section area and the yield stress. A related objective of the thesis is to determine
whether the 75 percent limit on yield stress as expressed in the NAS (2001) and
AS/NZS 4600 is valid for these thin AS 1397 steels, and if not, whether it can
be increased, and to what extent.
The G550 sheet steels are very thin 0.4 mm~1.0 mm and so the effect of
slenderness is important. A related objective is therefore to investigate the
influence of the slenderness of the G550 sheet steels on section stability.
1.3 Outline of this thesis
1.3.1 General
This thesis consists of four main parts: experimental investigations, numerical

investigations, theoretical investigations and the design recommendations.
1.3.2 Experimen tal investigation
A series of compression tests was performed on thin-walled box and channel

section columns. The tests were performed on pin-ended box shaped long
columns and on fixed-ended stub and channel section columns. The main
purpose of the tests was to investigate local, overall and distortional stability.
The main objective of the local stability investigation using stub columns was to
determine the adequacy of the design rules in Section 2 (Elements) of AS/NZS
4600 and Section B of the NAS (2001). The investigation of the overall stability
5
and the interaction of local and overall buckling was carried out using pin-ended
columns to improve the design rules in Section 3.4 of AS/NZS 4600 and
Section C4 of the NAS (2001). The tests performed on the lipped channel
sections were performed on fixed-ended sections mainly to investigate local
buckling, distortional buckling and the interaction between them to determine
the adequacy of Clause 3.4.6 (Distortional Buckling) of AS/NZS 4600 for high
strength steels.
An accurate instrument was designed and used to measure the geometric

imperfections of the specimens.
1.3.3 Numerical investigation
The finite element non-linear analysis program “ABAQUS” is used to simulate

the geometric and material nonlinear behaviour of the columns. Two different
models are established, one for simulating the pin-ended box section and one for
the fixed-ended channel sections. The experimental data is very important in
calibrating and implementing the finite element non-linear analysis. Once this
has been performed, the finite element non-linear analysis can be used to extend
the range of test data, and to investigate the effect of changing variables, such as
stress-strain characteristics, residual stresses, geometric imperfections and
section geometry.
The 4-node doubly curved thin or thick shell, reduced integration, hour-glass
control, finite membrane strain element, type S4R, is used. The ratio of the
length to width of each element is kept approximately 2:1. For the long
columns, different mesh densities are adopted. In the longitudinal direction of
the column, the nodes were concentrated towards the middle of the column so
that a finer mesh is obtained around the centre. In the transverse direction, the
6
finer mesh is used at the corners based on the concept of effective area. The
material behaviour data used in the ABAQUS model was obtained from the
stress-strain curves of coupon tests in tension. For the ends, two different types
of boundary conditions are used to simulate the test situation in the stub column
tests. The ends are divided into an immovable end and a movable end, which
was the loaded end. Two different ways were used to introduce geometrical
imperfections into the ABAQUS analyses. The first is to use the initial out-of-
plane deflection at mid-length of the column based on the imperfection
measurement and Walker’s (1975) suggested expression. The second is to
introduce the imperfection based on an eigenvalue buckling analysis again with
Walker's suggested expression for the amplitude.
1.3.4 Theoretica l studies using finite strip methods
The semi-analytical finite strip method of buckling analysis of thin-walled

sections is a very efficient tool for investigating the buckling behaviour of cold-
formed members in compression and bending. It assumes that thin-walled
sections buckle with simply supported ends free to warp but with section
distortion prevented at the ends. The program THIN-WALL (TW) (1998) was
developed to perform a semi-analytical finite strip buckling analysis of thin-
walled sections under axial compression and bending. It can be used to
understand the general local, distortional and overall buckling behaviour of thin-
walled sections. It is applicable to both open and closed sections and quantifies
the different buckling stresses.
In order to account for the fixed-ended boundary conditions in the tests, the
Spline Finite Strip Method (SFSM) was developed for buckling analysis by Lau
and Hancock (1986). The method uses spline functions in the longitudinal
direction and can account for a range of end conditions including fixed and free.
7
It is used in the thesis to accurately simulate the test boundary conditions.
The TW analysis gives the elastic local and elastic distortional buckling stresses
at given half-wavelengths, whereas the SFSM analysis gives the actual buckling
stress of a given section length between fixed ends.
The theoretical studies also provide valuable information on the local and
distortional buckling stresses for use in the interaction buckling studies.
1.3.5 Design rec ommendations using the effective width method and direct
strength method.
The effective width method is an elemental method since it looks at the

elements forming a cross section in isolation. It was first proposed by von
Karman (1932) and calibrated for cold-formed members by Winter (1947). It
accounts for post-buckling by using a reduced (effective) plate width at the
design stress. This method is used in the thesis for computing the predicted load
for all test sections.
For high strength steels, Clause 1.5.1.5 (b) of AS/NZS 4600 and Section A2.3.2
of the NAS Specification (2001), have a reduction to 75% of the yield stress. It
appears from the series of stub column tests of box shaped sections described in
the thesis that a modified reduction factor should be used to better utilize the
strength of material. So a modified reduction factor of 0.90 can be used in place
of the reduction factor 0.75, which is specified for G550 steel with the thickness
being less than 0.9 mm in AS/NZS 4600.
In order to take account of interaction of local and overall buckling of long

columns a modified design method based on the column design curve of
8
AS/NZS 4600 is proposed. A reduced radius of gyration is used in Clause 3.4.2

of AS/NZS 4600 to replace the normal radius of gyration (Section C4.1, Eq.
C4.1-1 of the NAS Specification) when the design stress exceeds the local
buckling stress.
The Direct Strength Method (DSM) was proposed by Schafer and Pekoz (1998)
and summarized by Hancock, Murray and Ellifritt (2001) who also
demonstrated its applicability. The DSM determines the strength for local and
overall (L+E) interaction and distortional and overall (D+E) interaction and
takes the lesser of the two as the strength. This method is used for computing
the test strength of the lipped channel sections. However, it does not account for
the interaction of local and distortional buckling and underestimates some of the
test strengths.
Based on AS/NZS 4600 Clause 3.4.6(b) and the Kwon & Hancock equation
(1992), two simple design methods are proposed to account for the interaction
of local and distortional buckling in thin sections of high strength G550 steel.
To determine the nominal axial strength (Pn) of the lipped channel section at
intermediate lengths, it is proposed that the interaction of local and distortional
buckling is taken into account with the distortional mode treated as an overall
mode in the DSM.
9
1.4 FIGURES
fy fu High strength cold-reduced steel G550
600
fu Fully annealed steel G300

Stress (MPa)
fy
300
Strain-hardening range
0.2% Strain
Fig.1.1 Stress-Strain Curve
10
400
350
300
Buckling Stress (MPa)
250
200
Local mode
Distortional mode
150 Torsional-Flexural
mode
100
●
50
● ●
0
10 100 1000 10000
Buckle Half-Wavelength (mm)
Fig.1.2 Buckling Modes
11
Chapter 2: LITERATURE REVIEW
CHAPTER 2
LITERATURE REVIEW
CONTENTS
2.1 RESEARCH ON HIGH STRENGTH STEEL .......................................................13
2.2 LOCAL BUCKLING, POST-BUCKLING OF PLATES AND INTERACTION

OF PLATE ASSEMBLIES...................................................................................19
2.2.1 General............................................................................................................................................. 19
2.2.2 Local buckling of plate ..................................................................................................................... 19
2.2.3 Post-buckling and interaction of plate assemblies ........................................................................... 21
2.3 BUCKLING OF COMPRESSION MEMBERS.....................................................25

2.3.1 General............................................................................................................................................. 25
2.3.2 Local buckling .................................................................................................................................. 25
2.3.3 Overall buckling ............................................................................................................................... 27
2.3.4 Distortional buckling........................................................................................................................ 30
2.4 INTERACTION OF MODES...................................................................................35

2.4.1 General............................................................................................................................................. 35
2.4.2 Interaction of local and overall buckling ......................................................................................... 35
2.4.3 Interaction of local and distortional buckling .................................................................................. 38
2.4.3.1. Linear interaction buckling................................................................................................................. 39
2.4.3.2. Non-linear interaction buckling.......................................................................................................... 39
2.4.4 General studies on mode interaction................................................................................................ 40
2.5 NUMERICAL METHODS .......................................................................................44

2.5.1 Finite element method ...................................................................................................................... 44
2.5.2 Finite strip method ........................................................................................................................... 47
2.6 SUMMARY ................................................................................................................53
2.7 FIGURE ......................................................................................................................56
12
CHAPTER 2
LITERATURE REVIEW
2.1 RESEARCH ON HIGH STRENGTH STEEL
The use of high strength steels is increasing rapidly. The combination of greater
strength and low cost leads to advantages in many fields of industry including
house construction. Cold-reduced steels with high yield stress usually have little
or no strain hardening in the stress-strain curve, and low ductility unlike
conventional structural steel that is highly ductile and strain hardens as shown in
Fig. 1.1. Strain hardening is important in the stability of thin-walled sections
and so the high strength cold-reduced steels are likely to have their stability
significantly affected by the lack of strain hardening.
In past decades, many papers & research reports have been published on high
strength steels. Although, in recent years high strength steels have become
available, some standards & specifications have limited the use of some high
strength steels due to their lack of ductility. For example, the 1996
Australia/New Zealand Standard AS/NZS 4600 for Cold-Formed Steel
Structures and the 1996 American Iron and Steel Institute (AISI) Specification
for Cold-Formed Structural Members have limited the design stress for high
strength cold-reduced steels to 75 percent of their yield stress or tensile strength
as applicable. The AISI Specification has recently been revised in Supplement
No.1 (1999) to allow values higher than 75 percent for multiple web
configurations, the value depending mainly on plate slenderness.
Priest and Gilligan (1954) re-examined the design specifications for structural
13
carbon steel, particularly in regard to buckling and elastic stability. The essential
principles of structural design were discussed and some formulas, charts and
tables were developed to assist engineers in designing for high strength steels.
Dhalla et al. (1971) investigated the influence of low-ductility on the tension

and connection behaviour of cold-formed members under static loading. A
series of coupon and connection tests were carried out. The concepts of local
ductility and uniform ductility were introduced. Most of the low ductility steels
investigated showed significant local ductility, but very limited uniform
ductility. Dhalla and Winter (1974) suggested alternate approaches for
measuring local and uniform ductility and further presented ductility criteria to
ensure satisfactory structural performance of thin steel members under
essentially static load.
For uniform elongation outside the fracture zone

εun ≥ 3.0%
For local elongation in a ½ in. (12.7 mm) gage length across the fracture
ε1/2 ≥ 20%
Tensile strength to yield stress ratio
σu / σy ≥ 1.05
Maricic (1979) presented briefly the analysis and examples of galvanized high
strength steels. The analysis showed that the thinner the steel, the lower the
elongation will become. Currie (1989) examined the developments in research
and design of light gauge cold-formed steelwork used in the construction
industry, and reviewed the applications of these new developments in design
standards around the world.
Hancock et al. (1987) presented in detail the tests of thin-walled high tensile
14
steel columns, fabricated from nominal 5 mm thick Grade 350 hot rolled steel
plate. Strength tests of thin-walled high tensile steel columns consisted of
welded I-sections, welded channel sections and cold-formed square hollow
section. The use of these test strengths for selection of column curves in the
Australian Steel Standard AS 4100 (1998) for columns where local and Euler
buckling interact was explained.
Macadam et al.(1988) conducted a series of coupon, beam and column tests to

characterise the material such as the determination of local and uniform
elongation and to determine the effect of the low strain hardening on the
behaviour of members. All the specimens were fabricated from the high yield
strength carbon steel with low strain hardening. The coupon tests showed that
O-steel (original) had the ratio (σu /σy) of about 1.4, an uniform elongation of 4
to 6 percent and a total elongation of 15 percent, for R-steel (further reduced)
had the ratio (σu /σy) of about 1.01 to 1.02, an uniform elongation of 0.8 to 2.7
percent and the total elongation of 9 or 10 percent. The column tests showed
that the results for the intermediate length columns were predicted
unconservatively. As a result of their work, it was concluded that the ductility
requirements in the specification should be amended to permit the use of the
low-strain-hardening ductile steel, but that its application be limited to use as
flexural members.
Pan and Yu (1988) presented tests of sections with unstiffened elements

fabricated from materials with yield strengths from 581 to 1057 MPa. Modified
effective width formulae were proposed to predict the post-buckling strengths of
unstiffened elements with high yield stresses. The test results showed that the
yield strengths in tension and compression were not significantly different.
Rasmussen and Hancock (1992) studied the local instability of high strength
15
quench and tempered steels to ASTM A514 fabricated by welding to form

conventional box, cruciform and I-sections of moderate thickness. The coupon
tests showed that the ratios of tension and compression 0.2% proof stress are
about 0.89 and 0.98 for 5 and 6 mm thickness steels respectively. The analysis
showed that the lack of strain hardening influences the strength of the sections,
particularly for stockier sections where the lack of strain hardening eliminates
the usual rise in strength above the squash load which is the product of the
section area and the yield stress.
Bernard et al. (1992) performed a series of deck tests with intermediate

stiffeners. The decks were fabricated from steel sheets conforming to AS 1397
Grade 550 steel sheets with 0.60 mm nominal thickness. These steels are similar
to ASTM A653 Structural Grade 80 steel. The effects of stiffener size on the
ultimate moment capacity and buckling modes on the loss of section stiffness
were studied. The local and distortional buckling modes were investigated and
design rules proposed.
Kwon and Hancock (1992) conducted a series of compression tests on lipped

channel sections. All specimens were formed from a cold-reduced zinc-coated
steel conforming to AS 1397 grade G500 with a total coated thickness of 1.2
mm. The tests were carried out between fixed ends and investigated post-
buckling in the distortional and mixed local-distortional modes. The tests by
Bernard et al, and Kwon and Hancock were performed on sections for which the
yield stress was significantly higher than the distortional buckling stress and so
a substantial post-buckling reserve of strength occurred.
Rasmussen and Hancock (1995) presented a test program on long columns

fabricated from high strength steel plates with nominal yield stress of 690 MPa.
13 box and I-section specimens were tested. The analysis showed that the effect
16
of residual stresses was less detrimental to the strength of high strength steel
columns than to the strength of ordinary steel columns.
A research project led by Professor W-W Yu at the University of Missouri-

Rolla (1995, 1996) investigated the strength of ASTM A653 Structural Grade
80 steels when used as decking. The tension coupon tests indicated that the
0.2% offset yield and tensile strengths of the steel both in the rolling direction
and perpendicular to the rolling direction increases with a decrease in the
thickness of steel sheets, while the ductility of the steel decreases with the
decrease in the thickness of steel sheets. When the steel sheets were formed into
decking sections and subject to bending, they demonstrated that their local and
post-local buckling capacities may be significantly influenced by the lack of
strain hardening. In particular, the ultimate moments of panels with slender
sections (b/t>100) were lower than the design moments calculated based on a
conventional effective section model.
Rogers and Hancock (1996, 1997a) conducted a series of tensile coupon tests on
perforated & unperforated specimens. High strength G550 sheet steels which
ranged in base metal thickness from 0.40 to 0.60 mm were tested as tensile
coupons with various size and shape of perforations. The test results indicated
that the G550 sheet steels do not meet the Dhalla and Winter (1971, 1974)
material requirements regardless of direction, except for uniform elongation in
the longitudinal test specimens but this may not affect the net section strength of
perforated sections and the net cross-section ultimate tensile strength can still be
adequately predicted using current design provisions. Rogers and Hancock
(1997b) also investigated bolted connections of thin G550 and G300 sheet steels
in 0.42 mm and 0.60 mm thicknesses. The test results indicated that the
connection provisions in the AISI Specification (1996), Eurcode and AS/AZS
4600 cannot be used to accurately predict the failure mode of bolted
17
connections constructed using G550 and G300 steels. The design rules cannot
be used to accurately determine the bearing resistance of bolted test specimens
based on a failure criterion for predicted loads.
Earls and Galambos (1998) presented numerical studies on HSLA80 (High

Strength Low Alloy) wide flange beams. Two distinct inelastic flexural modes
emerged as the dominant modes occurring at failure in HSLA80 beams under
moment gradient: little out-of-plane deformation accompanying localized
buckling which occurs close to the mid-span stiffener and a great deal of out-of-
plane deflection associated with the regions near localized buckling in the
compression flange. The study showed that the structural ductility of the beams,
as quantified by plastic hinge rotation capacity, is very much dependent upon
which of the two mode shapes governs at failure and concluded that the design
specifications which are based on the behaviour of mild carbon steel will not be
applicable to the design of HSLA80 flexural members.
Hancock (1997, 2003) reviewed the changes of the main Standards and
Specifications in the world. Major research developments around cold-formed
steel structures were summarized. Gresnigt and Steenhuit (1997) gave an
overview of the developments towards higher strength structural steels, which
focused mainly on the research done in Europe and the consequent standards
and codes of practice. Hancock and Rogers (1998) summarized the research
performed at the University of Sydney on ductility of cold rolled high strength
steel sections and distortional buckling and reviewed the material standards
specified in the AISI Specification 1997, Eurocode 3 Part 1.3 (1996) and
AS/NZS 4600:1996.
18
2.2 LOCAL BUCKLING, POST-BUCKLING OF PLATES AND

INTERACTION OF PLATE ASSEMBLIES
2.2.1 General
A plate element subjected to compression, bending, shear or a combination of

these stresses in its plane may buckle locally or distort at a low stress level
(local buckling stress). Although the local buckling load may not be the design
basis since the post-buckling load can be much greater, local buckling may have
a very significant effect on the member strength, especially for members which
suffer interaction of local and overall buckling or local and distortional
buckling. Therefore, the local buckling behaviour, post-buckling behaviour and
interaction between plate elements have to be considered in structural design.
2.2.2 Local buck ling of plate
Bryan (1891) first discussed the problem of the buckling of a simply supported
thin-rectangular plate uniformly compressed in the longitudinal direction and
obtained the solution from the fundamental differential equation for the
deflection of the plate. Timoshenko (1936) discussed the stability of plates
under various conditions of support at the two edges parallel to the longitudinal
compressive forces and showed the application of the theory to the investigation
of the plate elements of the steel columns. The general soution based on Bryan’s
equation for the elastic critical local buckling stress (fol) is given by:
π 2E t
f ol = k ( )2 2.1
12(1 − ν ) b
2
where k is a plate buckling coefficient which depends on the support conditions.
19
Lundquist and Stowell (1942) assumed that restraint among structural members
is provided by a specially defined elastic restraining medium and derived a
formula for the critical compressive stress at which buckling may be expected to
occur in outstanding flanges and presented a general chart for determining the
value of k.
This theory was applied to the member cross-sections which are composed of
various connected elements by many researchers (Stowell et al. 1952, Bleich,
1952). Stowell et al. (1952) presented the calculation of the buckling stresses of
flat unstiffened plates and integral flat-plate sections and gave the buckling
stresses in the form of theoretical charts. Bleich (1952) presented the
relationship between k and the plate aspect ratio in graphical form and gave a
generalized form of the theory for critical local buckling stress. The equation for
the elastic critical buckling stress was modified by the plasticity reduction factor
and varied with the type of loading and support condition.
A generalized approach to the local instability based on the small deflection

theory of plate bending was developed by Chilver (1953). Application of this
method to any particular plate resulted in the derivation of the exact elastic
instability stresses. This method was used to deal with the flexural and torsional
effects of reinforcing flanges of struts with open section.
Timoshenko and Gere (1961) presented the values of k for rectangular plates of
high aspect ratio with different types of boundary conditions and stresses. More
cases of buckling were considered and more tables for calculating critical
stresses were added.
Bulson (1970), in his well-known book, summarized the available work on plate
buckling and extended it further. The values of k and the critical stresses of
20
unstiffened and stiffened rectangular plates with various loading and boundary
conditions were given.
2.2.3 Post-buckl ing and interaction of plate assemblies
Local buckling causes a loss of stiffness and a redistribution of stresses.

Uniform edge compression in the longitudinal direction results in a nonuniform
stress distribution after local buckling. The buckled (called post-buckled) plate
has the maximum stress occurring at the longitudinal edge supports and the
minimum at the centre. The redistribution of stress normally continues until the
stress at the edge reaches the yield point of the steel and then the plate begins to
fail. The post-buckled plate can carry further load. In some cases, yielding may
occur at the centre of the plate as a result of combined bending & compression.
The post-buckling behaviour of a plate can be analyzed by using large
deflection theory. Schuman and Back (1930) presented the experimental proof
to show that the ultimate load was higher than the buckling load, which
previously assumed that the buckling load was the highest load a plate could
carry.
The differential equations for large deflection buckling of a plate were

introduced by von Karman in 1910.
∂ 4ω ∂ 4ω ∂ 4ω t ∂ 2 F ∂ 2ω ∂ 2 F ∂ 2ω ∂ 2 F ∂ 2ω
+ 2 + = ( − 2 + ) 2.2a
∂x 4 ∂x 2∂y 2 ∂y 4 D ∂y 2 ∂x 2 ∂x∂y ∂x∂y ∂x 2 ∂y 2
∂4F ∂4F ∂4F ∂ 2ω 2 ∂ 2ω ∂ 2ω

+ 2 + = E{( ) −( 2 )} 2.2b
∂x 4 ∂x 2 ∂y 2 ∂y 4 ∂x∂y ∂x ∂y 2
where F is a stress function defining the median fiber stress of the plate
21
The solution of the differential equation is too complicated to be used for

practical design. von Karman (1932) presented a theoretical study of a simply
supported rectangular plate. A method for determining the post-buckling
strength was developed based on the concept of ‘effective width’ which is
illustrated in Fig. 2.1.
be f cr
= 2.3
b fy
where be is the effective width, fcr is the buckling stress, fy is the yield stress; in this method, it
is assumed that the total load is carried by a fictitious effective width b.
Based on many tests and studies of post-buckling strength of the long plates that
are stiffened along both longitudinal edges, such as webs of channels and I-
beams, Winter (1948) presented a modified formula for calculating the effective
width be.
E t E
be = 1.9t [1 − 0.475( ) ] 2.4
f max b f max
Coan (1951) studied the post-buckling of a simply supported rectangular plate

with a small initially curvature. Boundary conditions were stress free supported
edges and uniformed displaced loaded edges. Levy’s series solution of von
Karman’s compatibility equation taken in conjunction with the energy method
replacing the von Karman’s equilibrium equation was used to calculate the post-
buckling response assuming the nonuniform edge displacements that are
characteristic of stress free edges. Yamaki (1959) extended Levy’s and Coan’s
work to the nonlinear problem which was solved extensively under various
boundary conditions combining two kinds of loading and four kinds of
22
boundary conditions. Numerical solutions were obtained for the deflection, edge
shortening and effective width of a square plate in compression. Formulae for
the ultimate load of a square plate were derived by using the maximum shear
theory for the beginning of yielding.
Walker (1968) used the von Karman large deflection equations to carry out the
analysis of buckling and post-buckling of single plates and assumed that a short
channel column may be treated as a collection of individual plates connected
appropriately along their common edges. The analysis provided an engineering
estimate of the maximum load capacity of a channel. Walker (1969) analyzed
the post-buckling behaviour of flat square plates loaded along two opposite
straight edges using the von Karman equations, trigonometric series and
Galerkin’s method. The effects of initial geometric imperfections were studied.
A perturbation method was used to solve approximately the non-linear algebraic
equilibrium equations occurring in the analysis of uniformly compressed square
plates.
Abdel-Sayed (1969) presented an approximate theoretical approach for the two

cases of plates where the longitudinal edges are restrained to remain straight or
are free to move in the plane of the plate. The effective width of a wide thin
plate, under compression in its plane, was examined by solving the von Karman
governing differential equations. Formulas showed the effective width of the
plate to be reduced when there is a small deviation from flatness or the edges
parallel to loading are free to move in the plane of the plate.
Dawson and Walker (1972) studied the post-buckled plate behaviour and the
effects of different generalized geometric imperfection parameters on the
expression for the ultimate load. The explicit expressions for the collapse, end
shortening and stiffness of simply supported plates with stress free edges were
23
derived.
Walker and Murray (1975) described the manner plates buckle and studied the
behaviour of analogous mechanisms consisting of rigid links and springs. Based
on these studies, a plate mechanism was derived. The analysis showed that
membrane elastic energy plays a significant role in determining the post-
buckling behaviour of a thin plate.
Kalyanaraman et al. (1977, 1978, 1979) presented the experimental and

analytical investigations on the local buckling of unstiffened compression
elements in the elastic range. The effects of initial imperfection and rotational
edge restraint on the local buckling of compression elements were considered.
Rhodes (1981) presented an approach to the analysis of elastic-plastic behaviour

in plates subjected to uniaxial compressive loading. The general approach was
to base the analysis on equations derived for perfect linear elastic plates and to
use simple routine modification procedures to take into account material
yielding.
Based on the studies of Pekoz (1986a), the 1986 Edition of the AISI
Specification uses the following equation in the calculations of uniformly
compressed stiffened elements.
b=w λ≤0.673 2.5
b=ρw λ≥0.673 2.6
where ρ=(1-0.22/λ)/λ 2.7
24
λ = (1.052 / K )( w / t )( f / E ) 2.8
An unstiffened element has one longitudinal edge free and one supported.
Despite behavioural differences, the same method for stiffened elements can be
adopted for unstiffened elements. The elastic local buckling and post-buckling
behaviour of unstiffened elements was studied by many investigators. Pekoz
(1986b) proposed that Winter’s effective width equation could be used for
unstiffened elements providing the appropriate value (k) for the buckling
coefficient was adopted.
2.3 BUCKLING OF COMPRESSION MEMBERS
2.3.1 General
Steel sections may be subject to one of three basic types of buckling mode:
local, overall (Euler or flexural-torsional) or distortional buckling. Local
buckling is particularly prevalent in cold-formed sections and is characterized
by relatively short wavelength buckling of individual plate elements. Overall
buckling is a long wavelength mode in which the plate elements forming cross
sections undergo significant translations without cross-seectional distortion.
Distortional buckling is buckling which takes place as a consequence of
distortion of the cross section. The wavelength of distortional buckling is
generally intermediate between that of local buckling and overall buckling as
shown in Fig. 1.2.
2.3.2 Local buck ling
25
Local buckling of plate elements of columns represents a particular case of plate

instability where the adjacent plates forming the section buckle at the same
stress. The column design has to take into account the stability of the plate
elements. Bleich (1952) presented a theory to demonstrate the fundamental laws
which control the behaviour of compressed plates under various conditions of
restraint to which the plate element of a column are subject. The theory provides
the basis for reliable rules to compute the required thickness of plates for
practical purposes rather than requiring a tedious investigation of the condition
for the occurrence of local buckling in each individual case. A coefficient of
restrain (ζ) was introduced to study the rectangular plate. Design formulae for
the required thickness of plate elements of columns were given in order to
prevent premature failure of compression members by local buckling.
Walker (1966) studied plates and channel sections subjected to eccentric

compression by means of an approximate solution using the Ritz-Galerkin
method. In order to determine the channel load-bearing characteristics, the
corresponding plate properties were determined and combined to obtain
compatibility conditions along the adjoining edges. An approximate series
method was used to calculate the buckling loads of flat plates which are
subjected to linearly-varying compressive-load actions in the plane of the plate
along two opposite edges.
Reiss and Chilver (1968) studied the local buckling of columns using some
researchers’ test results of different shape sections. The study showed that the
assumption that strength of a section is the sum of the separate strengths of the
component plates, assuming these are simply-supported or free on the
longitudinal edges, may be in error on the unsafe side by as much as 20% for
some sections.
26
Venkataramaiah and Roorda (1978) presented a generalized method of

theoretical analysis for the problem of local buckling of thin walled channel
sections under eccentric compression. The theoretical analysis finds critical
stresses of the web and flange elements at their common edge by using the
criterion of vanishing combined stiffness at this edge.
Parks et al. (1988) investigated the local buckling of both stiffened and
unstiffened curved elements through the use of short stub column tests. Three
different curvatures of stiffened and unstiffened curved elements were formed
from high strength steels. The empirical expressions were developed for
predicting the local buckling stress.
Landolfo et al. (1999) conducted a series of stub column tests to study the
ultimate strength of aluminum alloy channels. The local buckling of internal and
outstand plate elements constituting the member cross section was analyzed by
considering the restraining action derived from their interaction. The test results
were used to investigate the degree of accuracy of the effective thickness
approach.
2.3.3 Overall bu ckling
A slender axially loaded column will tend to fail by overall buckling with
different type of buckling for different cross-section: flexural buckling, torsional
buckling and torsional-flexural buckling. Doubly symmetric shapes and closed
shapes may fail by overall flexural buckling. Singly symmetric shape columns
may fail by either flexural or a combination of flexural and torsional buckling.
Point symmetric shape columns may fail by flexural or torsional buckling. In
the elastic range, the elastic critical buckling stress (foc) for a long column can
be determined by the Euler formula. Two methods have been used for
27
determining the inelastic critical buckling load: the tangent modulus method and
the reduced modulus method.
π 2E
f oc = 2.9
(l r ) 2
where r is the radius of gyration of cross-section
Bleich (1952) gave a good prediction of the inelastic critical flexural buckling
stress (fcr) using the tangent-modulus theory together with the Euler formula for
critical flexural buckling stress. Three simultaneous differential equations of
buckling by torsion and flexure in their most general form were given. These
equations may be applied to columns with pinned, fixed, or free ends as in the
usual theory of buckling.
Chilver (1961) discussed the basic material properties of cold-formed steel

sections and of the geometry of formed sections. The flexural and torsional
properties of open thin-walled sections were reviewed. The strengths of tension
members, columns and beams were studied. A torsional-flexural buckling
parameter, which defines the mode of buckling and the buckling load, was
suggested. Vlasov (1961) and Timoshenko (1961) developed a general theory of
flexural-torsional buckling. Trahair (1993) provided an up-to-date treatment of
modern methods of analysis of flexural-torsional buckling and a detailed
summary of knowledge on flexural-torsional buckling.
A general formula proposed by Chajes and Winter (1965) investigated the

torsional-flexural buckling of centrally loaded thin-walled members and
presented a simple method of accounting for torsional-flexural buckling. Based
on the use of an interaction type of equation, a relatively uncomplicated and
28
straightforward procedure for determining the torsional-flexural buckling load

was described. The basic theory was limited to hinged and fixed end columns
but can be extended to include other than fixed or hinged boundary conditions
by using an effective length. Hone (1967) carried out a detailed analysis of the
differential equations of torsional-flexural buckling and established the critical
solutions for three types of end conditions. The solutions were converted into a
suitable non-dimensional form and various parameters were used in a series of
design charts. Barta (1967) reviewed the development of fundamental research
on the elastic flexural-torsional buckling of thin-walled bars and discussed the
fundamental differential equations governing the problem. The Chwalla (1950)-
Witte (1957/59) formula was improved and simplified. Two methods, which are
the energy method and the simplified finite-difference approach, were used to
solve the simplified equation.
Pekoz and Winter (1969) studied the general behaviour of thin-walled singly
symmetric open sections under eccentric axial loading on the plane of
symmetry. The analysis showed that for a given shape, depending on
eccentricity and slenderness, a variety of behaviour modes is possible. The
effect of pre-critical deflections and other modifications were introduced to the
basic theory to analyze these modes.
Wang (1986) presented a rational method for determining the torsional-flexural

buckling load for thin-walled columns of open cross sections to deal with the
battened and unbattened members under concentric or eccentric load. The
influence of pre-buckling deflection was taken into consideration.
Yang (1986) derived the differential equations of equilibrium for the stability of
thin-walled beams of an arbitrary open cross section. Various nonlinear effects
due to the geometric change were considered. These equations can be used to
29
study various torsional-flexural buckling problems of a thin-walled bar.
Seah and Rhodes (1990), studied the behaviour and the strength of edge-
stiffened beam sections which were subjected to pure bending and bent in such
a way that the stiffeners were in compression. Relatively long edge stiffened
equal-flanged channel beams were tested under four point bending. The
behaviour for lateral torsional buckling was investigated in the elastic range.
Two series of tests on edge stiffened beam sections of various geometries were
described. Design procedures were also proposed based on the effective width
concept and theoretical findings for prediction of the ultimate strength of the
edge stiffener beam with a limiting plate width to thickness ratio of 60. Seah
and Khong (1990) further compared their results with those obtained by semi-
analytical, semi-numerical approach adopting the Rayleigh-Ritz method based
on the energy principle and found close agreement between them.
2.3.4 Distortiona l buckling
Thin-walled channel sections and other mono-symmetric sections may undergo

a distortional buckling mode which is the lip-stiffened flange of the section
rotating about the flange-web junction or the lip of the section rotating about the
lip-flange junction.
Sharp (1966) reported some experimental data on buckling of hat sections

formed from aluminum sheet and summarized available information on the
strength and behaviour of plates with stiffeners oriented parallel to the direction
of applied loads. Three buckling modes were identified. As stated in the paper,
when the rigidity of restraining elements to a flanges becomes large, a torsion
buckling analysis that does not include the effects of flange flexibility becomes
highly unconservative. Therefore, an approximate modification to the restraint
30
term, which is rotational restraint offered to flange by other elements of the

section, was given.
Williams and Wittrick (1972) presented the numerical results for the buckling,
under uniform longitudinal compression, of a series of panels with unflanged or
flanged integral stiffeners, or with Z-section stiffeners. The results showed that
torsional buckling modes contain a substantial amount of deformation of the
stiffener cross-section and the assumption, which the cross-section does not
distort, can lead to considerable over-estimation of the torsional buckling stress.
Thomasson (1978) conducted a series of tests on the lipped channel sections

with/without stiffeners in the web. The interaction between different modes was
studied. The torsional mode of the stiffener, which is characterized by the fact
that the flange of the primary stiffener is displaced sideways at the same time as
the wide flange is deformed, was investigated. The associated critical load was
designated a torsional buckling load.
Hancock (1978) gave a definition of distortional buckling mode, which was

considered as a mode occurring at a half-wavelength between local and lateral
buckling. The finite strip method was used to study the local, distortional and
flexural-torsional buckling of I-beams bent about their major axis. The estimate
of the distortional buckling load for the beam subjected to continuous lateral
and torsional restraint on the tension flange was compared with the values
produced using Bleich’s “pony truss chord model”.
Desmond et al. (1981a) tested a variety of edge stiffened channel sections with
edge stiffening lips turned inward or outward at an angle to the flange. Two
interrelated yet fundamentally different buckling modes characterized the
behaviour of edge-stiffened elements; namely the stiffener buckling mode (more
31
recently called distortional buckling) and the local plate buckling mode (local
buckling). The local buckling interaction between web and flange elements was
not fully investigated, both the analytical and experimental work were limited to
those assemblies in which local instability is initiated in either the flange or the
stiffener.
Sridharan (1982) developed the finite strip method to study post-buckling in the
local-torsional (now called distortional) mode. The examples illustrated that one
likely consequence of buckling of an edge stiffener in its own plane would be of
plastic yielding. The yielding of a member which has been the main source of
stiffness cannot but hasten the collapse of the structure.
Mulligan and Pekoz (1984) tested a series of channel sections in either

concentric or eccentric loading in order to study the local buckling on the
overall buckling behaviour. General failure modes for the test specimens were
gradual lateral deflection without twisting of the cross section. Several
specimens exhibited significant local-torsional (now called distortional)
buckling rather than the interaction mode between local and overall buckling. A
simple limiting stress approach was proposed. To prevent the local-torsional
buckling mode, Mulligan and Pekoz proposed an effective section method with
the stress at the flange-stiffener junction being limited to the local buckling
stress with k=4.0. The proposed method was shown to be conservative when
compared with their limited test data on the local-torsional mode.
Hancock (1985) carried out a series of tests to study the effect of the distortional
mode on cold-formed lipped channel sections with rear flanges using the
material with thicknesses 1.6 & 2.0 mm and the yield stresses 523 & 487 MPa.
The distortional stresses were from 375 to 180 MPa. The experimental tests
showed that there was very little post buckling strength available for distortional
32
buckling mode probably because the sections yielded.
Lau and Hancock (1987) derived explicit analytical expressions to predict the
distortional buckling stress of thin-walled channel sections columns with a
range of section geometries. The rigorous analytical expressions were first
developed for an approximate model using the flexural-torsional buckling
theory of undistorted thin-walled columns.
Lau and Hancock (1988, 1989) tested a series of channel columns of different
section geometries, thicknesses and steel strength grades under uniform
compression in a fixed-ended boundary conditions. The section geometries
consisted of lipped channels, hat sections and storage rack sections. The lengths
of the test columns ranged from short stub columns which failed mainly in
inelastic local buckling to long columns which failed in the elastic or inelastic
flexural-torsional buckling mode. Based on the test results, two different design
methods using the effective width equation were proposed for the distortional
buckling and yielding.
Kwon and Hancock (1992) performed a series of compression tests on lipped

channel sections with/without intermediate stiffeners in the web and
investigated post-buckling in the distortional and mixed local-distortional
modes. The section geometry and yield strength were chosen to ensure to ensure
that a substantial postbuckling strength reserve occurred in the distortional
mode. Material with 1.0 mm thickness and 590 MPa yield stresses was used.
The distortional buckling stress (from 88 to 45 MPa) was much less than the
yield stress. The experimental tests showed a significant post-buckling strength
reserve in the distortional mode, even when local and distortional buckling
occurred simultaneously.
33
Davies and Leach (1994) introduced the second-order terms associated with
geometric nonlinearity into the basic equation of Generalized Beam Theory.
Simple explicit equations for the load to cause buckling in individual modes
under either axial load or uniform bending moment were derived. The explicit
procedure showed how to consider the linear interaction between local,
distortional and global buckling modes.
Schafer (1997) used finite strip and finite element analysis to demonstrate that
the distortional mode has greater imperfection sensitivity than local modes. The
results showed that distortional failure has a lower post-buckling strength than
local failures. Schafer and Pekoz (1999) developed new hand methods to predict
the critical buckling stress in both the local and the distortional mode.
A new design procedure for cold-formed steel members called the ‘Direct
Strength Method’ (DSM) was proposed by Schafer and Pekoz (1998) and
summarized by Hancock, Murray and Ellifritt (2001) who also demonstrated its
applicability. The method employs elastic buckling solutions for the cross-
section, instead of the element-by-element plate buckling solutions used in
traditional design. The method uses the entire cross-section in elastic buckling
determination and incorporates local, distortional and global buckling into the
design process.
Prola and Camotim (2002a, 2002b) presented investigations on the elastic

distortional post-buckling behaviour cold-formed lipped channel steel columns
and beams using spline finite strip analyses. The analysis showed a post-
buckling asymmetry phenomenon with respect to the cross-section initial
distortions.
34
2.4 INTERACTION OF MODES
2.4.1 General
To determine the capacity of a cold-formed member, consideration must be

given to local, overall and distortional behaviour. However, interaction between
buckling modes may result from the occurrence of simultaneous or nearly
simultaneous buckling loads and generally produces adverse effects on the
strengths of members with slender plate elements. The interaction between
buckling modes has been studied extensively over decades.
2.4.2 Interaction of local and overall buckling
Research into the interaction of local and overall buckling has been conducted
by many researchers. Bijlaard and Fisher (1952, 1953) analysed thin plated
columns of square or H-type cross section and produced some closed form
solutions using their semi-intuitive method of ‘split rigidities’ in conjunction
with the principle of virtual work. The tests showed that these columns buckled
elastically in a flexural mode at a load higher than the local buckling load but
less than the Euler buckling load. Considerable interaction buckling effects were
demonstrated.
Graves-Smith (1967) investigated the behaviour of locally buckled box section

columns and examined their load carrying capacity using an elastic-plastic
analysis. The column was assumed initially straight and a short length was
analyzed using a Rayleigh-Ritz energy method. The interaction between local
plate deformations and the spread of plasticity was treated on an approximate
basis because of the lack of a general relationship between stress and strain in
the plastic plates. The reduction in strength of longer columns caused by overall
35
buckling was assessed by obtaining the apparent internal bending stiffness of

the locally buckled section.
van der Neut (1969) studied initially straight columns containing only local
imperfection and columns having both local and overall imperfections. The
study showed in the case of an idealized column severe imperfection-sensitivity
when the ratio of Euler load to local buckling load is about one but the
imperfection of the column axis appears to have a minor effect upon the load
carrying capacity. van der Neut (1973) gave a correction and re-evaluated in the
case of an idealized column the reduction of the failure load due to the
imperfection of the column axis. The analysis showed that whereas initial
waviness of the composing thin walls reduces the buckling load mainly when
the ratio (R) of Euler load to local buckling load is about one, the imperfection
of the column axis has its strength reducing effect over a much wider range of R,
more in particular in the range R>1.
Skaloud and Zornerova (1970) carried out a series of column test and concluded
that the buckling of the plate elements diminished the effective section of the
column and consequently accelerated the column’s flexure. Interaction between
local and overall buckling was observed.
DeWolf et al. (1974) studied the behaviour of cold-formed columns subject to

local buckling and presented an analytical approach to account for the combined
effects of local buckling, column buckling, and nonuniform material properties
in compression members which was based on the tangent modulus concept and
the effective width expression.
Gibert and Calladine (1974) used a graphical method to extend Calladine’s

work and to make an exhaustive study of the combined effects of both local and
36
overall imperfections. The effects of imperfections were well described by the

celebrated Perry formula in conjunction with a single imperfection parameter
compounding simply the local and overall imperfections. The reduction of the
peak load due to local and overall imperfections together was much less than the
sum of the separate effects.
Little (1979) discussed the problem of predicting theoretically the strength of

square thin-walled steel box columns which are pin-ended and loaded centrally
at the ends. A general method for generating design curves for the local-flexural
interaction situation were described. The analysis showed that in a local-flexural
interaction situation, the stocky column strength is not usually the same as that
stress-value which, if reached in the critical element, is sufficient to cause rapid
failure of more slender columns. This is in contrast to the behaviour of simple
columns, for which both these values are simply the yield stress.
Hancock (1981a, 1981b) proposed a simple design method for I-columns which
are subjected to both local and flexural buckling phenomena. The finite strip
method was extended to include the nonlinear response of imperfect plate strips
under longitudinal compression and proposed a method to calculate the
reduction in the flexural buckling load of imperfect box and I-section columns
in the region of the local buckling load. For the box columns rather than for I-
section columns, the reduction calculated using the effective width finite strip
analysis agreed closely with the calculated using the effective width formula.
Bradford and Hancock (1984) presented a nonlinear finite strip method for the
post-local buckling of geometrically imperfect plate assemblies. The method
was used to provide an accurate alternative to the Winter effective width
formula for obtaining the effective section of a simply supported I-beam in the
post-buckling range.
37
Mulligan and Pekoz (1984) presented an effective section method for analysing
the effects of local buckling on the overall buckling modes of behaviour of
singly symmetric thin-walled columns and beam-column. The method
recognized the post-local bucking of the component plate elements and the
associated shift of the centroid.
Toneff et al. (1987) studied the interaction of local and flexural or flexural-
torsional buckling modes by adopting a finite element for a thin-walled beam-
column of arbitrary cross sections to include local degrees of freedom. The
method allowed distortion of the cross section.
Sridharan and Zeggane (2000) studied the interaction of local and overall
buckling in plate structures and stiffened shells using a specially formulated
shell element. Amplitude modulation, a key feature of the interactive buckling,
was incorporated in the element formulation. A procedure was outlined for
incorporating the key secondary local mode in the interactive buckling model.
Young and Rasmussen (2000) described a technique for determining the overall
flexural and flexural-torsional bifurcation loads of locally buckled cold-formed
channel columns. The inelastic nonlinear finite strip buckling analysis was used
to determine tangent rigidities of a locally buckled section which were used in
the overall flexural and flexural-torsional equations.
2.4.3 Interaction of local and distortional buckling
The interaction between local and distortional bucking modes, and distortional
and flexural-torsional modes has not been widely investigated. Hancock (2002)
summarised the research on the mode interaction and categorized the mode
interaction into: Linear interaction and Non-linear interaction. Linear
38
interaction is the interaction of the two modes which occurs at the same half-
wavelength and Non-linear interaction is that where the short half-wavelength
mode (usually local) occurs over multiple half-wavelengths and interacts with a
long half-wavelength mode by reducing its effective stiffness against buckling.
Therefore, the following review of the researches on the mode interaction can
be classified according to this definition.
2.4.3.1. Linear interaction buckling
Williams and Wittrick (1972) showed the coupling between the different modes.
The coupling between different modes was weak if these predominantly
consisted of the local or overall mode. When the torsional mode was dominant,
the effect due to coupling between the modes can be considerable.
Davies and Jiang (1996) presented an analysis on thin-walled columns with the
properties typically used in practice using GBT and showed that the GBT can
provide a particularly appropriate tool to analyze distortional buckling in
isolation and in combination with other buckling modes. For a column of
symmetrical cross-section, there is no modal interaction between symmetric and
asymmetric buckling and the degree of interaction with other bucking modes
depends primarily on the buckling half-wavelength. For distortional buckling in
columns, it is generally sufficient to consider a single buckling mode involving
symmetric rotation of the flanges about the flange/web junctions.
2.4.3.2. Non-linear interaction buckling
Pekoz (1991) presented an overview of research on thin-walled steel and

aluminum structures conducted at Cornell University. The first part of the paper
was aimed at developing sufficiently accurate and simple design formulations
39
for laterally unsupported cold-formed steel compression flanges. Possible

failure modes were illustrated in the paper. The interaction of local and
distortional buckling was discussed.
Kwon and Hancock (1992) conducted tests on sections of high strength steel for
which the buckling stresses were significantly less than the yield stress of the
material. Although a significant post-buckling strength reserve was observed, no
adverse interaction between local and distortional buckling occurred for these
sections with thickness down to 1.0 mm.
Serrette and Pekoz (1995a) discussed an analytical method that was formulated
for estimating the elastic distortional buckling stress of flexural members with
laterally unsupported compression flanges. The elastic buckling finite strip
program was used to study the interaction of local and distortional buckling.
The analysis showed that for some specimens, the interaction resulted in a
significant decrease in buckling stress. However, the relationship between the
degree of the interaction of buckling modes and the specimen of geometric
properties could not be identified from this study. Serrette and Pekoz (1995b)
presented the results from two experimental studies and proposed two design
recommendations for the interaction between local and distortional buckling in
standing-seam panels. The methods differ only in the procedure by which the
elastic distortional buckling stress is computed.
2.4.4 General st udies on mode interaction
Two different approaches have been adopted to predict post-buckling

phenomena. One approach is to make use of the concept of effective width to
account for the post-buckling strength of the locally buckled component plates.
The representative papers are such as Loughlan and Rhodes(1979), Hancock
40
(1981a, b), and Bradford and Hancock (1984) which are based on the empirical
formula suggested by Winter and the nonlinear finite strip analysis. Another
approach is on the basis of the general Koiter theory (1945). Hereafter, a brief
review on the second approach is given.
The general post-buckling theory of elastic structures was described by Koiter

(1945). The concept of initial imperfections was introduced to explain the
significant reductions of the critical load and the discrepancy between linear
theories and experiments on shells. Graves-Smith (1967) and van der Neut
(1968) first studied the interaction between local and Euler buckling in thin-
walled compression members by using mathematical models which included the
nonlinear membrane stiffness of the plates forming the columns. After those
pioneering works, the phenomenon of post-buckling was extensively studied by
many researchers who include Tvergaard, Byskov and Hutchinson, Sridharan
and his co-workers, Pignataro and his co-workers, and Kolakowski and his co-
workers.
Tvergaard (1973) studied the initial post-buckling behaviour of a wide,

integrally stiffened panel subjected to compression. Two buckling modes
occurred simultaneously and were investigated. Byskov and Hutchinson (1977)
presented a preliminary investigation of mode interaction in axially stiffened
cylindrical shells under axial compression. An initial post-buckling analysis
method was proposed for simultaneous and nearly simultaneous buckling
modes. The influence of a given level of imperfection on the optimum was
explored.
Sridharan (1983) developed a semi-analytical approach for the study of doubly

symmetric interactive buckling of plate structures subject to axial compression
using the theory of mode interaction and the finite strip concept. Sriharan and
41
Ali (1985) developed a new analytical model for a study of the response of
thin-walled beam-columns having doubly symmetric cross sections with some
novel features which are: incorporating the interaction of overall buckling with
two companion local modes; accounting for the phenomenon of amplitude
modulation and modeling any set of realistic end conditions. The singularity
problem in interactive buckling, which was bypassed in the previous study, was
considered. Sriharan and Peng (1989) described a new analytical model for
stiffened panels which was developed for the study of nonlinear interaction of
local and overall instabilities of axially compressed stiffened plates. The
phenomenon of amplitude modulation and the triggering of the secondary mode
were shown to be the logical consequences of the interaction of overall bending
with the primary local mode.
Pignataro et al. (1985) performed a post-buckling analysis of simply supported

channels under uniform compression on the basis of the general theory of elastic
stability due to Koiter. Attention was essentially focused on local/Eulerian and
local/flexural-torsional simultaneous buckling modes interaction. The columns
were treated as a plate assemblage. The results showed that the local/Eulerian
interaction is more detrimental than the local/flexural-torsional one. Luongo and
Pignataro (1988) analyzed the effect of boundary conditions on multiple
interactive buckling in thin-walled members under compression by means of the
general theory of stability. The total potential energy was expanded to up to
third-order terms to investigate structures with asymmetric post-buckling
behaviour and initial geometric imperfections were taken into account. Under
simplifying assumptions, the nonlinear equations governing the post-critical
behaviour reduced to a linear eigenvalue problem that is independent of the
geometric and mechanical properties of the column and depends on the
boundary conditions only. Pignataro et al. (2000) investigated the effect on
post-buckling behaviour of thin trapezoidal profiles of the interaction of many
42
simultaneous or nearly simultaneous buckling modes. The post-buckling

analysis was performed on the basis of the Koiter general theory of elastic
stability. Results showed that the occurrence of several simultaneous buckling
modes makes the structures sensitive to initial imperfections.
Kolakowski (1987) studied the interaction of nearly simultaneous buckling

modes in the presence of imperfections of thin-walled trapezoidal columns
under uniform compression. The bifurcation stress was determined analytically
and the asymptotically exact expansion was obtained for the initial bifurcation
behaviour. Kolakowski (1993) studied the influence of the modification of
boundary conditions at both ends of structures regarding interactive buckling on
the post-buckling behaviour of thin-walled elastic columns with imperfections.
The investigation concerned thin-walled closed and open cross-section columns
under axial compression. The interaction of three buckling modes was
considered in terms of the second order approximation having continuity
conditions at the junctions of the plates. Peter and Kolakowski (1996)
investigated the interactive buckling analysis of thin-walled open beam-columns
with intermediate and edge stiffeners under axial compression and a constant
bending moment carried out by means of the transition matrix method. In the
solution obtained, the effects of interactive buckling of two local modes having
the same wavelength, the shear lag phenomenon and also the effect of cross-
sectional distortions were included. Kolakowski (1999) studied the influence of
modal interactive elastic buckling on the post-buckling behaviour of thin-walled
composite beam-columns with imperfections when the distortional deformations
were taken into account. An interaction of three or four buckling modes was
considered in terms of the second order approximation, satisfying continuity
conditions at the junctions of plates.
43
2.5 NUMERICAL METHODS
There are many numerical methods which can be used to account for important
features specific to thin-walled structures. Generally speaking, all of the
required phenomena can be modeled using appropriate numerical methods. The
finite element method is the most general numerical method available. The
commercial computer program ABAQUS (1997) is a suite of powerful
engineering simulation programs based on the finite element method. They can
solve problems ranging form relatively simple linear analyses to the most
challenging nonlinear simulation. With this program, local and overall
geometric imperfections, residual stresses, geometric and material non-
linearities can be taken into account. However, the finite element method has
some disadvantages, such as considerable computing expense and tedious data
preparation. The finite strip method has an advantage over the finite element
method when applied to thin-walled members with the greatly reduced number
of degrees of freedom for a given member.
2.5.1 Finite elem ent method
Research on the finite element method started from the 1940’s. Since then, the
finite element method has developed extensively and many papers and books
have been published. Zienkiewicz (1971) presented the basic concepts about the
finite element method and some concepts of fundamental mathematical
approximation in the finite element method in his famous book. The detailed
procedures were developed to form and solve a finite element problem. Cook et
al. (1974) described the basic concepts and applications of finite element
analysis. Crisfield (1991) developed a solution procedure involving a
combination of incremental load or displacement control coupled with full or
modified Newton-Raphson iterations. The techniques provided the basis for
44
most nonlinear finite element computer programs. Here a brief review of the
applications of the finite element method in thin-walled member analysis is
given.
Rajasekaran and Murray (1973a, 1973b) studied the effects of local and member
buckling and developed a formulation for elastic and inelastic buckling of thin-
walled beams with local and member buckling interaction. A finite element
model was developed from this formulation for elastic wide flange beams.
Finite element stiffnesses were derived from these equations for the particular
case of elastic wide-flange sections. An investigation of the coupling in this type
of section was carried out. Rajasekaran and Murray (1973b) presented details of
a procedure for analyzing inelastic beam and beam-column problems of
arbitrary thin-walled open cross-section, under the assumption of no strain
reversal. The finite element formulation associated with such a procedure, and
the expressions for programming the necessary computation of section
properties and element matrices were given.
Waszczyszyn et al. (1994) presented the application of the finite element

method to the structural stability analysis and described the formulation of finite
element method models associated with engineering applications and the
phenomena of instability of structures. The extended and comprehensive
algorithms which are suitable to the linear buckling analysis were also
discussed, as was buckling of thin-walled structures.
Chou and Chai (1997) presented an alternative design procedure suitable for
estimating the ultimate load of thin-walled stub-columns and developed a
standard design procedures for the post-buckling analysis of these structures
using the finite element method. The ultimate loads for steel columns under
compressive loads were obtained using a nonlinear finite element analysis.
45
Schafer (1997) used the finite element analysis program ABAQUS to establish
the numerical modeling of ultimate strength of cold-formed steel members. The
existing literature on the geometric imperfections, residual stresses and the
stress-strain curve was reviewed. The conservative bounds for imperfection
magnitudes were established. For normal use in modelling residual stresses, the
average flexural residual stresses were suggested. An elastic-plastic strain
hardening model was suggested if direct experimental data was not available.
Sivakumaran and Abdel-Rahman (1998) developed a finite element analysis

model to predict the post-local buckling behaviour of non-perforated and
perforated cold-formed steel members subjected to axial compression. The
finite element model also utilized experimentally derived stress-strain
relationships and residual stress idealizations, geometric imperfection
distribution, a special loading system, and a displacement solution algorithm.
Wilkinson and Hancock (1999) investigated the sensitivity of the rotation

capacity of cold-formed RHS beams using the finite element analysis program
ABAQUS. The essential stages in the development of the model to simulate the
bending tests on RHS were described. A wider parametric study was also
performed on other variables in the model. The maximum loads predicted were
slightly lower than those observed experimentally, since the numerical model
assumed the same material properties in the whole cross-section.
Chou et al. (2000) presented a finite element analysis on the post-buckling

behaviour of thin-walled cold-formed lipped channel and hat-section stub
columns under axial compression. A standard design procedure to predict the
ultimate local capacity of these structures was developed using the finite
element computer program ANSYS. The ultimate load obtained using the
design procedure consistently underestimated the experimental results and
46
analytical predictions using BS 5950.
Sarawit et al. (2001) studied the post-buckling behaviour of an isolated flange-

stiffener model using a nonlinear finite element analysis program ABAQUS.
Two types of imperfections, local and distortional modes obtained from the
eigenvalue, were superimposed to give the initial geometric imperfections.
Analysis types, material models, elements taken into account in the finite
element analysis of thin-walled structures were discussed. Bakker and Pekoz
(2001) discussed the possible sources of error in linear and non-linear finite
elements solutions and gave some suggestions on how to check and prevent
these errors.
Mateus and Witz (2001) investigated the sensitivity of the buckling and post-
buckling behaviour of imperfect steel plates used in ship and related marine
structures when subjected to variations in their initial conditions. The analyses
were carried out using the non-linear finite element program ABAQUS. The
results showed that the three most relevant parameters affecting the buckling
and post-buckling of uniaxially loaded plates were aspect ratio, unloaded edge
boundary conditions and initial out-of-plane imperfection amplitudes.
Dhanalakshmi and Shanmugam (2001) carried out an experimental study of the
ultimate load capacity of non-perforated and perforated equal-angle cold-
formed steel stub columns. A finite element model was proposed based on the
experimental results and used to carry out a parametric study. A simplified
design formula was proposed to determine the ultimate load-carrying capacity
using the data obtained from the parametric study. ABAQUS was also used to
analyze the test results.
2.5.2 Finite strip method
47
The finite strip method originally developed by Cheung (1968) has become a
useful method to analyze thin-walled sections in the area of static analysis,
stability analysis, post-buckling analysis and vibration. A typical finite strip
discretizations for a channel section is shown in Fig 5.4a. The finite strip
method of structural analysis is a special form of the finite element method
where a prismatic structure consisting of an assembly of thin plates is
subdivided into a number of longitudinal strips. In contrast to the finite element
method which uses polynomial functions in both directions, the semi-analytical
finite strip method (SAFSM) uses continuous harmonic series functions in the
longitudinal direction to describe the strip displacements. The series have to
satisfy a priori boundary conditions at the end of the strips. Different finite strip
methods have been developed using complex functions, spline functions,
polynomial functions or mixed functions to describe the displacement fields.
Cheung (1968a) originally developed the finite strip method for a wide range of
solutions for isotropic and orthotropic variable thickness plates in which two
opposite ends are simply supported. Cheung (1968b, 1971) extended the method
to plates with a clamped or free end boundary. Cheung (1976) further extended
the method to the stress analysis of folded plate structures by incorporating the
membrane displacements in addition to the flexural displacements. In the study,
only four degrees of freedom were required to describe the displacements along
each nodal line. The numerical work was reduced very considerably when
compared with the generalized finite element method.
Przemieniecki (1973) showed that the finite element method can be effectively
applied to the analysis of local instability of plates, stiffened panels, and thin-
walled columns through the use of special wavelength dependent stiffness
matrices. The stiffness matrices were derived from the exact sinusoidal
lengthwise variation of displacements. The resulting sinusoidal stiffness
48
matrices depend on the wavelength of the buckled pattern, but this dependence
is of a very simple form, since all stiffness coefficients contain the buckling
wavelength as a common factor.
Plank and Wittrick (1974) extended Przemieniecki’s analysis and developed a

complex finite strip buckling analysis method for determining the initial
buckling of any structure consisting of a series of thin flat isotropic plates
rigidity connected together at their longitudinal edges. Przemieniecki’s analysis
was extended in two ways, namely abandonment of the assumption of the line
junctions between adjoining plates remaining straight during buckling and
adoption of a more complicated state of stress rather than an uniform biaxial
compression in each plate.
Hancock (1978) studied the interaction between local, distortional and lateral
buckling of I-beams using the finite strip method. Hancock (1981a) studied the
interaction between local and overall buckling of I-section columns using a
nonlinear finite strip method. The finite strip results of the problem of
interaction of local and overall buckling of I-section columns in the elastic
range were found to agree closely with the existing test results. Hancock
(1981b) extended the finite strip method to include the nonlinear response of
imperfect plate strips under longitudinal compression. Studies of the nonlinear
response of box and I-sections under axial compression were presented.
This method was also applied by Bradford and Hancock (1984) to investigate
the post-local buckling of geometrically imperfect plate assemblies. The method
provided an accurate alternative to the Winter effective width formula for
obtaining the effective section of a simply supported I-beam in the post-local
buckling range of structural response. An improved tangent stiffness was used
to solve the nonlinear simultaneous equations, resulting in rapid convergence
49
for bending strains up to three times the eigenstrain. Hancock (1985a)

performed a study of the elastic buckling and nonlinear elastic response of thin-
walled I-sections in bending using the finite strip method. Davids and Hancock
(1987) presented a theory which combined the finite strip method of nonlinear
elastic analysis of locally buckled thin-walled sections with the influence
coefficient method of analysis of beam-columns. The theory combined the
advantages of both methods of analysis to produce a computationally efficient
procedure for the estimation of the overall nonlinear elastic response of locally
buckled thin-walled beam-columns.
Graves-Smith and Sridharan (1978a) presented a finite strip method for the
post-locally buckled analysis of plate structures. The method was restricted to
prismatic plate structures subjected to uniform compression at their ends. A
brief study of uniformly compressed channel section struts was presented.
Graves-Smith and Sridharan (1978b) also presented a finite strip buckling
analysis for the stability analysis of prismatic thin walled structures subjected to
arbitrary loading. The method was restricted to prismatic plate structures with
the end sections supported by rigid diaphragms. This method of analysis was
subsequently used by Graves-Smith and Gierlinski (1982) to analyze the
buckling of longitudinally stiffened web plates subjected to transverse local
edge loading. Simple rules were suggested for determining the minimum
bending rigidity of the stiffener required for fully restraining the plate, and then
for calculating the buckling strength as a function of the torsional rigidity.
Graves-Smith et al. (1985) developed a combined finite strip/finite element
method for the linear analysis of prismatic structures containing transverse
diaphragms, stiffeners and intermediate supports.
Sridharan and Graves-Smith (1981) applied two versions of the finite strip
method to study the post-buckling behaviour of plated structures which might
50
buckle in a local or a distortional mode. They investigated the effect of corner

displacements on the post-buckling behaviour using displacement functions
Sridharan (1982) presented a semi-analytical method for the study of the post-
local torsional buckling and overall buckling of lipped channel sections which
buckle in modes that involve significant in-plane movement of one or more of
the plate elements. The influence of the method of loading which may either be
one of prescribed end compression or load eccentricity was discussed. Since a
perturbation technique was employed in the analysis, the validity of the result
was restricted to the close vicinity of bifurcation.
Lengyel and Cusens (1983) presented a finite strip method for the geometrically
non-linear analysis of prismatic structures under arbitrary loading. The method
employed an iterative finite element approach but replaced numerical
integration by an explicit formulation. Structural imperfections were taken into
account. Cheung and Fan (1983) developed a spline finite strip method to
overcome the difficulties of the semi-analytical finite strip method in dealing
with concentrated forces, multiple spans, discrete supports at strip ends etc. The
spline finite strip uses a linear combination of local spline functions in the
longitudinal direction to describe the displacements of the strip rather than a
series of harmonic functions. As a result, the method can be used to account for
non-simple end boundary conditions. The spline strip method is more flexible
than the finite strip method in modelling different boundary conditions and
complex loading conditions.
Lau and Hancock (1985) developed the spline finite strip method for buckling
analysis of folded plate systems, stiffened plates and thin-walled structural
members. Stability matrices for a thin plate strip subjected to longitudinal
compression and bending, transverse compression as well as shear were
developed. The stability matrices can be used in conjunction with the stiffness
51
matrices previously developed by Cheung and Fan (1983) to perform analysis of

thin flat walled structures of finite length and with end boundary conditions
other than simple supports. The method was further extended by Lau and
Hancock (1989, 1990) to the inelastic buckling analysis.
Kwon and Hancock (1989) developed a nonlinear elastic analysis based on the
spline strip method for studying the post-buckling behaviour of thin-walled
sections. The method can be used to deal with local, distortional and overall
buckling modes in the post-buckling range and the interaction between them.
The method has no restrictions on the end boundary conditions and loading. The
buckling modes occurring simultaneously at different wavelengths were
automatically accounted for in the analysis. Kwon and Hancock (1993)
presented a geometric nonlinear elastic spline strip method to analyze the post-
buckling behaviour of thin-walled sections. The method can handle sections
buckling simultaneously in local and distortional buckles at different half-
wavelengths.
Guo and Lindner (1993) presented a theoretical study on the elastic-plastic

interaction buckling of imperfect longitudinally stiffened panels under axial
loads using spline finite strip method. Key and Hancock (1993) presented a
material and geometric non-linear finite strip analysis to study the non-linear
behaviour and ultimate local capacity of plates and plate assemblies subjected to
axial compression and bending. Using the appropriate displacement functions,
sections undergoing either inelastic local buckling deformations or overall
buckling deformations were analyzed. The effects of geometric imperfections,
material yield stress variation and stress-strain responses and complex patterns
of residual stress can be included in the analysis.
Bradford and Azhari (1995a) extended the finite strip method to analyse the
52
buckling of plates with different boundary conditions other than simple

supports. In the analysis, various analytic functions were used. Bradford and
Azhari (1995b, 1997) employed bubble functions to improve the inelastic semi-
analytical complex finite strip method. The convergence could be significantly
improved by adopting bubble functions. The method can be applied to the
buckling analyses of plates with various support conditions, plates under
different loading conditions and inelastic buckling of plates and plate
assemblies.
Hancock (1998) summarized the finite strip analysis and nonlinear analysis
methods at the University of Sydney since 1976. The substantial test programs
on a range of thin-walled section types including welded I-sections and channel
sections, and cold-formed channel and square hollow sections performed at the
University of Sydney since 1981 were also reviewed. Cheung and Tham (2000)
reviewed the post-1995 researches on the finite strip method. The review
focused on the developments of various types of strips, implementation of finite
strip in parallel programming environment and applications in vibration and
stability analyses.
2.6 SUMMARY
This chapter has reviewed the development in research and design of thin-
walled structures. The conclusions that are discussed throughout the chapter are
summarised below.
High strength thin steels may have insufficient ductility and may not meet the
Dhalla & Winter or current design requirements for overall, local or uniform
elongation. The thin steels which give the advantages of lightness and high
53
strength-to-weight ratio are also responsible for some adverse effects in

structural behaviour. Results based on current design standards may therefore
predict unconservatively for such steels.
In the design of thin-walled steel members, the complex section profile can be
represented as an assemblage of plate elements and assessment of the full
profile is based on these individual plates. Each plate element may be treated as
a slender plate, which is supported along one or both of its longitudinal edges by
the adjoining plate elements and is susceptible to local buckling when subjects
to compression. Although local buckling may not be the design basis since the
post-buckling load can be greater, local buckling may have a very significant
effect in the member strength, especially for members which suffer the
interaction of local and overall buckling or local and distortional buckling.
Therefore, the local buckling behaviour, post-buckling behaviour and
interaction between plate elements have to be considered in structural design.
The analysis of the problem of the post-local buckling behaviour of a plate
element are complicated and therefore alternative approaches have therefore
been developed for this purpose such as the Effective Width Method and the
Direct Strength Method.
In the design of compression members, the thin material and the section
configurations associated with thin-walled steel sections result in local and
torsional effects. For closed sections, the thin-walled members are susceptible to
local and overall buckling. The interaction of local and overall buckling, in
which the effects of local buckling often drastically reduce the overall buckling
strength of the compression member, has been extensively investigated. In a
singly-symmetric section, thin-walled members are particularly susceptible to
local and distortional buckling. The interaction of local and distortional
buckling has not been extensively studied. Therefore the effects of the
54
interaction on the strength of the thin-walled members needs to be clarified.
There are many numerical methods to be used to account for important features
specific to thin-walled structures. Generally speaking, all of the required
phenomena can be modeled using appropriate numerical methods. The finite
element method is the most general numerical method, such ABAQUS (1997).
It can solve problems ranging form relatively simple linear analyses to the most
challenging nonlinear simulation. With this program, local and overall
geometric imperfections, residual stresses, geometric and material non-
linearities can be taken into account. However, the finite element method has
some disadvantages, such as considerable computing expense and tedious data
preparation. The finite strip method has an advantage over the finite element
method when applied to thin-walled members with the greatly reduced number
of degrees of freedom for a given member. Some computer programs based on
this method has been developed for buckling analysis, such as THIN-WALL
(1998). Computer programs for post-buckling analysis based on this method
may be needed to predict the strength of the thin-walled members. Although
these numerical methods are becoming more and more important in the research
on thin-walled structures, it is still difficult to give a rational modelling of
geometric imperfections.
55
2.7 FIGURE
fmax
b
True area
fmax
be/2 be/2
Effective area Effective area
Fig. 2.1 True and Effective Stress Distributions
56
Chapter 3: STRENGTH TESTS OF COLD-FORMED BOX-SHAPED SECTION STUB COLUMNS
CHAPTER 3
STRENGTH TESTS OF COLD-FORMED BOX-

SHAPED-SECTION STUB COLUMNS
CONTENTS
3.1 INTRODUCTION......................................................................................................58
3.2 TEST SPECIMENS ...................................................................................................58

3.2.1 General............................................................................................................................................. 58
3.2.2 Labelling........................................................................................................................................... 60
3.2.3 Geometric Imperfection Measurements............................................................................................ 61
3.3 MATERIAL PROPERTIES .....................................................................................61

3.3.1 Manufacturing process of G550 sheet steel...................................................................................... 61
3.3.2 Steel sheet thickness ......................................................................................................................... 63
3.3.3 Coupon Test specimens and procedures........................................................................................... 64
3.3.4 Coupon test results ........................................................................................................................... 64
3.4 STUB COLUMN TESTS ..........................................................................................65

3.4.1 Testing .............................................................................................................................................. 65
3.4.2 Test specimen behaviour, ultimate load-capacity and experimental local buckling loads............... 65
3.5 ANALYSES ................................................................................................................67

3.5.1 Elastic local buckling analyses......................................................................................................... 67
3.5.2 Test results and comparisons with design standards........................................................................ 68
3.5.2.1 General...................................................................................................................................................... 68
3.5.2.2 B-sections.................................................................................................................................................. 70
3.5.2.3 LB-sections ............................................................................................................................................... 71
3.5.2.4 HB-sections............................................................................................................................................... 71
3.5.2.5 Analyses of effective area ......................................................................................................................... 72
3.5.2.6 Explanation of some results ...................................................................................................................... 72
3.6 SUMMARY ................................................................................................................75
3.7 FIGURES....................................................................................................................77
3.8 TABLES......................................................................................................................90
57
CHAPTER 3

SHAPED-SECTION STUB COLUMNS
3.1 INTRODUCTION
The aim of this chapter is to present the first stage compression test results. The
tests were performed on box (B), lipped-box (LB) and hexagonal (HB) shaped
specimens of three different cross-sections fabricated from G550 steel sheets
with 0.6 mm and 0.42 mm thickness. The specimens included stub columns and
long columns. The stub columns were tested between fixed ends and the long
columns were tested between pinned ends. This chapter presents the stub
column results only.
3.2 TEST SPECIMENS
3.2.1 General
A stub column is a structural member sufficiently short so as to preclude

member buckling when compressed, but sufficiently long to contain the same
initial residual stress pattern as a much longer member cut from the same stock.
For cold-formed steel sections, which generally have thin-walled plate elements,
the stub-column test is aimed at determining the effect of local buckling as well
as the effect of cold-forming and residual stress on the section capacity in
compression.
For cold-formed shapes, the length of the stub column should not be less than
58
three times the largest dimension of the cross section and no more than 20 times
the least radius-of-gyration (Galambos, 1988). In the tests in this chapter, the
lengths of stub columns were three times the width. The ends of the columns
were milled flat and perpendicular to the longitudinal axis of the column.
The tests were performed on closed sections brake pressed from aluminum/zinc-
coated Grade G550 structural steel sheet to AS 1397. The sections tested are
shown in Fig. 3.1. Epoxy was used to close the B & LB-sections and a
Teflon/Steel Support was used to close the HB-sections. Bolts & clamps were
also used on the B & LB-sections as discussed later. The sections were
fabricated from 0.42 mm and 0.6 mm steel sheets. The widths of the B-sections
shown in Fig. 3.1 ranged from 20 to 100 mm for the 0.6 mm sheet steel and
from 14 to 70 mm for the 0.42 mm sheet steel. The flats of the HB-sections
ranged from 20 to 100 mm and the thickness was 0.6 mm. The dimensions of
the specimens are shown in Tables 3.1~3.3, for the nomenclature shown in Fig.
3.2. The cross-section dimensions were the average of the measured values. The
base metal thickness (tb) was measured by removing the zinc coating by acid-
etching.
The first few specimens were found to fracture in the epoxy at loads very close
to the ultimate load. It was therefore decided to drill 3 mm holes in the corner
lips and to place small diameter bolts and nuts as shown in Figs. 3.3b & 3.3c to
ensure that the corners did not come apart. Initially, these bolts were only
located at the ends, but subsequent testing resulted in increased numbers of bolts
with the bolt spacing at approximately 20 mm then 10 mm for two series of
specimens using the steel sheets with 0.60 mm thickness. Later tests used
clamps as shown in Fig. 3.3a. During the tests, the number of clamps was
changed for some sizes in order to determine the effect of clamps on the lips.
The test results were not significantly different (less than 1%). Hence the effect
59
of clamp spacing on the ultimate loads can be eliminated. The holes/clamping

configurations are summarized in Table 3.4 and shown in Fig. 3.3.
For the B-sections, tests were performed on 0.60 mm thickness material and
0.42 mm thickness material. All sections had the same size lips. For the LB-
sections, tests were performed on both 0.42 mm and 0.60 mm thickness
material. The lips of the LB-sections in 0.42 mm thickness were the same for all
series as shown in Table 3.2d. However, for the 0.60 mm thickness LB-sections,
the lips were different from section size to size for the two series shown in
Tables 3.2a & 3.2c. The material used for the LB-sections as detailed in Tables
3.2a & 3.2c had different yield stress values as given in Table 3.4. The lips for
the sections shown in Table 3.2b all had the same lip size of 6 mm. For the HB-
sections, tests were only performed on 0.60 mm thickness material. Two slotted
pieces of Teflon and two long steel strips were used to join the two edges of the
HB-sections together. The objective of the Teflon was to connect the
longitudinal free edges without resisting axial compressive load.
3.2.2 Labelling
The B & LB-sections were divided into two different series for 0.6 mm sheet
steel and three different series for 0.42 mm sheet steel. The HB-sections had one
series. The test specimens were labelled such that the thickness of steel sheet,
type of section, nominal width of specimen and specimen number were
expressed by the label.
For example, the label “060LB40ra” defines the following specimen:
• The first three numbers indicate that the specimen is fabricated from 0.60
mm steel sheet.
• The fourth and fifth letters (LB) indicate that the specimen is a lipped box
(alternatives B for box and HB for hexagon).
60
• The “40” indicates that the nominal width of specimen is 40 mm (flat width
for a hexagon).
• The “r” indicates that the specimen is a repeat test.
• The last letter “a” indicates that the specimen was the first tested
(alternatives b, c).
3.2.3 Geometric Imperfection Measurements
The geometric imperfections were measured for the test specimens. The
instrument used was designed to measure the geometric imperfections and is
shown in Fig. 3.4. The main components include a guide frame, a laser sensor
and a Datataker data logger. The laser sensor was attached to a movable seat,
which could move up and down along the upright poles and from left to right
using a screw adjuster. The specimen was positioned on the seat in front of the
frame using a clamp to make it firm and upright. Once the specimen was
properly positioned on the seat and fixed in position, the laser sensor was
moved along the specimen. The readings were taken at 15 mm or 20 mm
intervals. For each side of a specimen, the measurements were taken along three
lines in the longitudinal and transverse directions of the specimen, two along the
corner and one along the centre (approximately 5mm away from the corner or
edge). The results were used in the finite element modelling analysis. The
maximum imperfections are shown in Table 3.5 and were computed relative to a
straight line between the ends.
3.3 MATERIAL PROPERTIES
3.3.1 Manufactu ring process of G550 sheet steel
61
The usual process for the production of these steels is cold reduction. This
process can be used to increase the strength and hardness, as well as to form an
accurate thickness for sheet steels and other steel products. Initially the sheet
steels are rolled to size in a hot strip mill with finishing and coiling temperatures
of approximately 940ºC and 670ºC, respectively. The hot worked coil of steel,
typically 2.5 mm in thickness with a minimum specified 300 MPa yield stress, is
uncoiled and cleaned in an acid solution to remove surface oxides and scale. The
uncoiled strip is then trimmed to size and fed into a cold reduction mill, which
may contain any number of stands. High compressive force in the stands and strip
tension systematically reduce the thickness of the steel sheet until the desired
dimension is reached, e.g. by approximately 85% for the 0.42 mm sheet steels.
The milling process causes the grain structure of cold reduced steels to elongate
in the rolling direction, which produces a directional increase in the material
strength and a decrease in the material ductility. The effects of cold working are
cumulative, i.e. grain distortion increases with further cold working as a result
of an increase in total dislocation density, however, it is possible to change the
distorted grain structure and to control the steel properties through subsequent
heat treatment. Various types of heat treatment exist and are used for different
steel products. G550 sheet steels are stress relief annealed, although
recrystallisation does not occur. Stress relief annealing involves heating the steel
to below the recrystallisation temperature, holding the steel until the
temperature is constant throughout its thickness, then cooling slowly. Mild sheet
steels of similar thicknesses are annealed to a greater extent in comparison with
G550 sheet steels, and hence recover their ductile behaviour. Annealing is
carried out in a hot dip coating line prior to application of either a zinc or
aluminum/zinc coating. Upon final cooling, the sheet steel is further processed
through a tension levelling mill, e.g. 0.35% extension, to improve the finish
quality and the flatness of the coil. The cold-reduced G550 sheet must be
62
differentiated from other sheet steels whose high yield stress and ultimate
strength values are obtained by means of an alloying process, i.e. high strength
low alloy (HSLA) steels.
The material property requirements for G550 or Grade E sheet steels are
specified in Australia by AS 1397 (1993) and in North America by the
following ASTM Standards, A611, A653, A792 and A875. The steels used for
this project were from the same coils as used by Rogers and Hancock (1997).
3.3.2 Steel sheet thickness
The thickness of a section is an important factor in determining the buckling

stress of the section and so must be determined accurately. The steel sheets used
in this test program were finished with a thin layer of coating for corrosion
protection purposes. From a structural point view, the aluminum/zinc coating is
a softer material compared with the steel. Therefore, it is reasonable to assume
that the thin layer of zinc coating does not carry load although, in reality, it does
carry some load for very thin sheets. Hence, the thickness of steel sheet used in
calculations is the base metal thickness (BMT)(without the coating).
The base metal thickness of each series of test specimens was determined using
a micrometer. Many coupons were made from the steel sheets that were used to
fabricate the column. The total coated thickness and base metal thickness were
measured with coating and without the coating using a digital vernier caliper
and a micrometer. For the nominal thickness 0.60 mm steel sheet, the average
total coated thickness and the base metal thickness are 0.65 mm and 0.60 mm
respectively. For the nominal thickness 0.42 mm steel sheet, the average total
coated thickness and the base metal thickness are 0.45 mm and 0.41 mm
respectively.
63
3.3.3 Coupon Te st specimens and procedures
The material properties of each series of specimens were determined by tensile

coupon tests. Two longitudinal coupons were tested for each series of
specimens. The coupon dimensions conformed to AS 1391 (Standards
Australia, 1991) for the tensile testing of metals. In order to ensure that fracture
occurred within the middle portion of the constant gauge length, the test
coupons were dimensioned with a more gradual change in cross-section from
the constant gauge width to the grip. Specimens were fabricated with larger
radii (i.e. 55 mm), which required a shortening of the constant gauge length
(Rogers and Hancock, 1996). The tensile coupons were 12.5 mm wide with
gauge length 50 mm.
The longitudinal coupons were tested according to AS 1391 in the Sintech/MTS

300 kN testing machine. The coupons were tested with the aluminum/zinc
coating removed. In the tests the longitudinal strains were measured using two
strain gauges for every coupon, which were attached at the centre of each face.
A SPECTRA data acquisition system was used to record the load and readings
of strain during the test.
3.3.4 Coupon te st results
The stress-strain curves were obtained from the coupon tests using strain gauges
for 0.42 mm and 0.60 mm material respectively. The yield stress fy was
obtained using the nominal 0.2% proof stress. The stress was the measured load
divided by the initial cross-section area of the coupon and the strain is the
average of the two strain gauge readings. The measured 0.2% proof stress of the
steel is shown for each specimen type in Table 3.4 and ranged from 634 MPa to
711 MPa. Young’s modulus of elasticity (E) was also calculated from the elastic
64
part of the stress-strain curves using Mathcad. The calculated mean values of
Young’s modulus of elasticity were 2.20x105 MPa and 2.16x105 MPa for
thicknesses of 0.42 mm and 0.60 mm respectively as shown in Table 3.6. The
mean values are given in Table 3.6. The three values for two thicknesses were
based on different batches of sheets used in the study. Care was taken to match
the coupons with the particular test specimens. The proof stresses for each
specimen type are given in Table 3.4.
3.4 STUB COLUMN TESTS
3.4.1 Testing
The test rig is shown in Fig. 3.5 and the test arrangement is shown in Fig. 3.6.
The rig consisted of the Sintech/MTS-300kN testing machine with fix-ended
bearings. The bottom bearing was adjustable so that it could orient specimens
vertically. However, Pattenstone was needed at the top to ensure perfect contact.
The load and shortening were recorded using the Sintech data acquisition
system. The compressive deformation rate was 0.05 mm/min.
The central deflections were measured using six transducers for the HB-
sections. For the LB-sections, four transducers and two transducers were used
for the sides and the lips respectively. For the B-sections, four transducers were
used, two for the plates and two for the lips. The transducers were connected to
a SPECTRA data acquisition system.
3.4.2 Test specim en behaviour, ultimate load-capacity and experimental

local buckling loads
65
The typical figures as shown in Fig. 3.7 show that initially the columns
remained elastic with the slope associated with axial stiffness approximately
constant. Stiffening occurred at take-up initially in some specimens. At local
buckling, as demonstrated by the lateral deflections, the axial stiffness reduced
with the following stiffness getting smaller and smaller until the ultimate load
was reached. After that point the load decreased but during the unloading, the
load did not fall too sharply.
From observation of the surface of the specimen, the buckling behaviour can be
seen. After the first lateral deflections occurred, elastic local buckling can be
observed with three half-wavelengths occurring along the specimen length.
Eventually the column entered the elastic-plastic range and the ultimate load
was reached when the local plastic mechanism formed. The box-specimens with
low b/t ratios (from 33.3 to 66.6) developed roof-shaped mechanisms as shown
in Fig. 3.8 and those with high b/t ratios developed the so-called flip-disc
mechanisms as shown in Fig. 3.9 with alternating concavity and convexity at
the four sides. The plots of load-deflection also show that the lips did not stay
straight during the tests. For some specimens, the deflection of the lips occurred
before the elastic buckling load was reached.
The ultimate loads (Pt) for all specimens are given in Tables 3.7~3.10. Figs.
3.10 & 3.11 are based on the data from Tables 3.7~3.10. The results have been
non-dimensionalised with respect to the theoretical stub column strength (Ns)
computed to AS/NZS 4600 as discussed in Section 3.5.2 following. A
significant change occurs at b/t equal to about 50 (i.e. b=30 mm for t=0.60 mm
or b=21 mm for t=0.42 mm). At this b/t value, a minimum value of the non-
dimensional strength between 0.85 and 0.9 occurs. The results are discussed in
detail in Section 3.5.2.
66
The experimental local buckling loads (Pcr) in Tables 3.7~3.10 were evaluated
from the plots of load-deflection and load-shortening. For the plots of load-
deflection, the P-w2 method (Ventaramaiah and Roorda, 1982) was used to
obtain the experimental elastic local buckling load.
It can be seen that as the width of cross-section (b) of column became larger, the
theoretical buckling loads (Nol) became smaller with the ultimate load (Pt)
increasing. The ultimate load (Pt) was about 5 times the elastic buckling load
(Pcr) for the largest sections.
3.5 ANALYSES
3.5.1 Elastic loca l buckling analyses
The theoretical elastic local buckling loads (Nol) were obtained using the THIN-
WALL program (Papangelis and Hancock, 1998). The average measured cross-
section dimensions of the specimens for each series as well as the measured
values of base metal thickness and Young’s modulus taken from the coupon
tests were used to determine the theoretical local buckling loads.
The theoretical local buckling stresses (fol) varied from approximately 30 MPa
for the largest B-sections to 710 MPa for the smallest B-sections. The effect of
the double thickness corner (lip) elements was to increase the theoretical local
buckling coefficient (k) for the flat elements from the simply supported value of
4.0 to approximately 4.5 due to slight torsional restraint at each of two corners.
The theoretical local buckling mode is showed in Fig. 3.12a. The local buckling
stresses (fol) varied from 220 MPa for the large LB-sections to 960 MPa for the
small LB-sections in 0.60 mm thickness. For the LB sections in 0.42 mm
67
thickness, three sections were fabricated (30, 40 and 50 mm) with the same
length of lip (5.25 mm). The local buckling stresses ranged from 70 MPa to 185
MPa for LB-sections in 0.42 mm thickness. The effect of the two webs with the
double thickness lips was to increase the theoretical local buckling coefficient
(k) for the flanges to approximately 5.6. The theoretical local buckling mode is
showed in Fig. 3.12b. This value is between the values for the simply supported
edges (k=4) and the built-in edges (k=6.97). (Bulson, 1970). The lips attached to
the web elements can be regarded as an intermediate stiffener similar to those
discussed by Desmond, Pekoz and Winter (1981b). However, they are of
sufficient size to ensure that the web elements buckle in an anti-symmetric
mode with the stiffener preventing movement normal to the plane of the web
elements. The local buckling stresses (fol) varied from 32 MPa for the large HB-
sections to 760 MPa for the small HB-sections in 0.60 mm thickness. The
hexagonal sections were assumed to have a buckling coefficient of 4.0 for all
faces. The theoretical local buckling coefficient (k) for all faces is about 4.5 due
to slight restraint at the Teflon supports. It was similar to the B-sections. Some
restraint at the supports is visible. The theoretical local buckling mode is shown
in Fig. 3.12c.
The full set of ratios of the experimental local buckling loads (Pcr) to the
theoretical local buckling load (Nol) are given in Table 3.7~3.10. The mean
ratios are 0.90 with a buckling coefficient average value of 4.8. These results
indicate that the THIN-WALL program was in reasonably good agreement with
the experimental local buckling loads.
3.5.2 Test result s and comparisons with design standards
3.5.2.1 General
68
The test results (Pt) are given in Tables 3.7~3.10. The test results (Pt) non-
dimensionalised with respect to the theoretical stub column strength (Ns) are
plotted against the plate slenderness ratios (b/t) in Figs. 3.10 & 3.11. The
theoretical stub column strengths (Ns) were calculated according to AS/NZS
4600 (1996). The theoretical strengths were calculated using the average
measured cross-section dimensions and the measured material properties as
detailed in Tables 3.1~3.3 & 3.6 respectively.
Solid and dotted curves are also plotted in Figs. 3.10 & 3.11. The dotted curve is
the ratio of NsRb/Ns versus plate slenderness (b/t). NsRb has been calculated based
on Rbfy as included in Section A3.3.2 of AISI Specification Supplement
No.1(1999). The solid curve is the ratio of Ns0.75/Ns versus the plate slenderness
(b/t). Ns0.75 has been calculated based on 0.75fy as included in Clause 1.5.1.5(b)
of AS/NZS 4600:1996.
The holes in the lips of the sections were taken into account in the calculation of
the theoretical stub column strength (Ns) by simply removing the area from the
lips. The plate slenderness (λ) for the lips according to Eq. 2.2.1.2(4) of
AS/NZS 4600 was less than 0.673 when the width of lip was less than 8.7 mm.
Hence the lips were fully effective for all B-sections and most LB-sections. For
the sections with holes in the lips, the perforated part was treated as an
ineffective portion. So the effective width of the lip was the full length of the lip
with the diameter of the hole removed. This effective width was used in the
calculation of the theoretical stub column strength (Ns). For the others without
holes in the lips, the calculation of the theoretical stub column strength (Ns) was
still based on Clause 2.2.1.2 of AS/NZS 4600:1996 with an appropriate
slenderness (λ). The plate buckling coefficient (k) values were as specified in
Clause 2.2.1.2 of AS/NZS 4600 for stiffened sections and Clause 2.3.1 of
AS/NZS 4600 for the unstiffened sections.
69
In this chapter, the plots were based on the Pt/Ns ratio vs. b/t ratio rather than the
fmax/fy ratio vs. b/t ratio which was used by Wu, Yu and LaBoube (1996). The
reason may be explained as follows. In their tests, the ultimate stress fmax was
obtained from the average compressive edge strain and the average measured
tensile strain was recorded at the failure of each panel. The problem with this
method may be that the change of the strain or stress on the surface of the
specimens can be very large from point to point to produce scatter in the data.
This may have made it difficult to obtain an accurate curve with this method.
3.5.2.2 B-sections
The ratio Pt to Ns of for the B-sections in 0.42 mm and 0.60 mm thickness

plotted against the plate slenderness (b/t) are shown in Figs. 3.10a & 3.11a
identified using symbols (♦, ■, ▲). It can be seen that both thicknesses had a
similar tendency. The ratios of Pt/ Ns decreased when the plate slenderness (b/t)
varies from 33 to 50. In this range, the test results are close to the results based
on the AISI Specification Supplement No.1(1999) and also have the same
tendency. As the plate slenderness (b/t) varies from 50 to 116, the ratios (Pt/ Ns)
increase. However, for plate slenderness (b/t) greater than 116, the ratios (Pt/ Ns)
decrease slightly. It can be noted that for the B-section in 0.60 mm thickness the
test results show more scatter than those of the B-section in 0.42 mm thickness.
The “♦” and “■” points are for the B-sections in 0.60 mm thickness with holes
in the lips at 20 mm spacing, whereas the “▲” points are for the B-sections in
0.60 mm thickness with holes in the lips at 10 mm spacing. It is obvious that the
latter are lower than the former by about 5%. To clarify the effect of the holes
on the column strength, three different sizes (30,50 and 70 mm plate width) of
B-sections in 0.60 mm thickness were fabricated and tested with clamps on the
lips. The test results (BWH) are shown in Fig. 3.11 expressed by the “∗” points.
70
It can be seen that the results obtained by simply removing the holes from the
net area when calculating Ns may underestimate the column strengths by about
5%.
3.5.2.3 LB-sections
Only three different sizes were tested for the LB-sections in 0.42 mm and 0.60
mm thickness respectively. The ratios of Pt to Ns for the LB-sections in 0.42
mm and 0.60 mm thickness plotted against the plate slenderness (b/t) are shown
in Figs. 3.10 & 3.11 identified using symbols (─). It can be seen that for the LB-
sections in 0.42 mm thickness, the tendency is quite similar to that of the B-
sections but a little bit lower by about 2% for the smaller and bigger sizes and
5% for the medium size compared with values at the equivalent plate
slenderness (b/t) of the B-sections. The tendency is contrary to the curve based
on AISI Specification Supplement No.1(1999). The largest difference of 5%
occurs for the largest size column. For the LB-sections in 0.60 mm thickness, it
can be seen that the test results are more scattered, however the tendency of the
average value is still similar to that of the B-sections.
3.5.2.4 HB-sections
HB-sections were tested in the 0.60 mm thickness only. The ratios of the test
results (Pt) to the theoretical stub column strengths (Ns) of the HB-sections in
0.60 mm thickness plotted against the plate slenderness (b/t) are shown in Fig.
3.10 identified using symbols (●). The results show a similar trend to the B&LB
sections but are lower at high slenderness with the test value at b/t=167 quite
low. It appears that the Teflon may not have fully restrained the sheet at high
slenderness.
71
3.5.2.5 Analyses of effective area
The tests results also indicated that when the critical stress was less than the
yield stress, the plate was capable of sustaining a greater ultimate load than the
load at which local buckling occurred as detailed in Tables 3.7~3.10. The
effective areas based on AS/NZS 4600 were also obtained. The effective areas
of the B-sections in 0.42 mm and 0.60 mm thickness are shown in Fig.3.13
identified using symbols (♦, ■). It is known from analysis of effective widths of
simply supported flat plates in edge compression that after the local buckling
stress is reached, the portion of the plate furthest from the supports deforms
significantly while those portions adjacent to the supports continue to develop
stress. Therefore the larger the size the section becomes, the more the central
part of the centre of the plate withdraws from the load. This also means that the
load is mainly carried by the corners and lips for the sections in this chapter. It
may explain why the slender section was not as sensitive to imperfections as the
stockier sections. According to the analysis and the test results, the conclusion
can be drawn that for a given shaped section, the load carrying capacity can not
be increased by enlarging the section. It is proven from the results of tests that
even though the size of section became larger, the ultimate loads obtained from
the tests stopped increasing at about 20.5 kN and 41 kN for the B-sections in
0.42 mm and 0.60 mm thickness respectively. This matches with the effective
areas derived from AS/NZS 4600 in Fig. 3.13.
3.5.2.6 Explanation of some results
In Figs 3.10 & 3.11, it can be noted that for the B-sections, the test results
decreased slightly when the ratio of plate slenderness (b/t) reached about 133. In
Tables 3.7, 3.8 & 3.10, it can be seen that when the plate slenderness (b/t)
ranged from 133 to 166, the critical stress was less than 40 MPa which can be
72
compared with the maximum average stress in the whole cross-section (Pt/A)
which was greater than 160 MPa. The higher plate slenderness (b/t) and the
lower critical buckling stress resulted in larger deformation of plates especially
at the two ends of column. Slight premature failure may have resulted from the
increased deformation of the plate at the ends before the ultimate load was
reached. The deformed shapes can be seen in the tested sections (Fig. 3.14).
It is also interesting to note that for the more slender sections, the test results are
closer to the theoretical strengths assuming no reductions in the yield stress.
This is to be expected since the effect of no strain hardening is likely to have
less effect for sections that are slender and fail principally by post-local
instability and then yielding rather than inelastic local buckling. For the stockier
sections, the results are closer to the revised AISI Specification Supplement
No.1(1999) strengths. The strengths based on AS/NZS 4600 using 75%
reductions in the yield stress are too conservative.
As shown in Fig. 3.15, the HB-sections had two failure modes. One was failure
at the top edges and the other was failure at the juncture of the two longitudinal
edges. The failure of some of the larger sections resulted from a combination of
the above two failure modes. The first failure mode occurred also for the larger
B-sections. For the thin plate, the ultimate load can be increased further because
of the post-buckling effect. In the post-buckling region, the plate will avoid
further increase of stress in the central region by deepening of the buckles. For
the larger sections, the lateral deflection of the central region was larger.
However, the friction between the column ends and the bearings prevented the
top edges spreading freely and eventually resulted in the bulging out near the
top end, which is similar to that of a tube under compression (Alexander, 1912).
The second failure mode occurred along the two longitudinal edges. It seems
that the Teflon supports did not fully restrain the steel sheets because the Teflon
73
supports may not have provided enough friction to the steel sheet to prevent a
part of the sheet from being pulled out of the Teflon.
For the LB-sections in 0.60 mm thickness as shown in Fig. 3.10, the results of
the largest sections varied by about 10%. The reasons may be explained as
follows. The ultimate load of LB40rb was significantly lower than LB40ra by
about 6% as detailed in Table 3.9b. In the test the LB40rb test, a special local
plastic mechanism can be observed on one face. A roof shaped plastic hinge
formed starting from one corner at L/3 from the top end and at another corner at
mid-length as shown in Fig. 3.16. The roof shaped plastic hinge had a 45o angle
with the longitudinal direction. This abnormal buckling shape may have resulted
from uneven loading of the two ends. For LB40a&b and LB40na&nb, the
common characteristic was a larger lip of about 10 mm. From Table 3.9a, it can
be seen that even though LB40a and LB40b had different configurations on the
lips, the test results (Pt) were the same but the theoretical column strengths (Ns)
were significantly different. The differences in the calculation or simulation
were that LB40a used the full length of lip and LB40b adopted a length of 3 mm
removed from the length of lip to account for the hole. Also the same method
was used for the LB40na&nb sections. It means that the actual effective length
of the lip of the LB40 section with a 10 mm length of lip was smaller than the
effective length calculated based on AS/NZS 4600. The latter results in an
overestimate of the theoretical strength and may be the main reason for the
decrease in the ratio Pt/Ns. Research on the intermediate stiffener also indicates
that an increase in the stiffener cross-sectional area contributes to increase at
most of a few percent in the ultimate strength. The results of the LB sections in
0.42 mm thickness did not have a significantly different ultimate strength from
the B sections, which used the normal length of lips as shown in Fig. 3.8b.
The effect of holes on the ultimate load should be considered. Many researchers
74
have investigated perforated members under compression. Levy, Woolley and

Kroll (1947) studied circular holes in the simply supported square plate, and
found that if the diameter of a hole is half of the width of the plate, the buckling
load will drop by 14%. In the tests in this chapter, sections with holes in the lips,
for which the diameter of the holes was nearly 50% of the width of the lips, had
an effect on the ultimate load which made it decrease by about 13%. However, a
significant effect on the buckling load was not observed. So for the sections
with holes in the lips, the ultimate loads may be considered to be about 87% of
those of the same section without holes in the lips if the diameter of hole is near
50% of the lip width.
3.6 SUMMARY
A range of stub columns in G550 sheet steel to AS 1397 has been tested in
compression to obtain the strength characteristics of this steel with low strain-
hardening. The following detailed observations & conclusions can be made.
• As detailed in Tables 3.7~3.10, for the stockiest sections, the average

ultimate stress (Pt/A) was lower than the theoretical local buckling stress (fol)
due to the effect of yielding and initial imperfections. However for the
slender sections, the average ultimate stress (Pt/A) was higher than their
critical stress (for the biggest sections, about 5 times the local buckling
stress) as a result of the post-buckling effect. This occurred even though the
initial imperfection was bigger than that of the stockiest sections. As detailed
in Tables 3.7~3.10 and shown in Figs. 3.10 & 3.11, the maximum non-
dimensional section strength is reached when the plate slenderness (b/t) is
equal to 120 and the minimum when the plate slenderness (b/t) is equal to
approximately 50.
75
• The results of the successful stub column tests have been compared with the
design procedures in the Australian/New Zealand Standard for Cold-formed
Steel Structures and recent (1999) Amendments to the American Iron and
Steel Institute Specification. As expected, the greatest effect of the low strain
hardening was for the stockier sections where material properties play an
important role. For the more slender sections where elastic local buckling
and post-local buckling are more important, the effect of low strain
hardening does not appear to be as significant. This is contrary to recent
design proposals for decking in flexure in the USA (Wu, Yu and LaBoube,
1996) where it was believed that the more slender sections had been
influenced. This design proposal was based on the plots of the fmax/fy ratio vs.
b/t ratio of Deck Panel Tests.
76
3.7 FIGURES
Flange Epoxy
Epoxy
Web
Epoxy Plate
Epoxy
Bolts in some
( a) Lipped-Box Section (LB) (b) Box Section (B)
Steel strip
Teflon
Bolt
Flange
(c) Hexagonal Section (HB)
Fig. 3.1 Test Section

d
tc
t
h
R
d
b
r R
b r
hl
a) Lipped-Box Section (LB) (b) Box Section (B)

bw
bf
(c) Hexagonal Section (HB)
Fig. 3.2 Cross-Section Dimension Nomenclature
77
(a) Lipped-Box Section (LB) with Clamps (b) Lipped-Box Section (LB) with Bolts
(c) Box Section (B) with Bolts (d) Hexagonal Section (HB) with Teflon Support
Fig. 3.3 Bolting & Clamping Configurations of

Test Specimens
78
Locations of
measurement on
each side of
specimen
Fig. 3.4 Rig for Imperfection Measurement and

Location of the Measurement
79
Fig. 3.5 Test Specimen in Test Rig
80
Cross head
Pattern stone to ensure for

End plates uniform compression
Test
stub Location of transducers
column
Adjustable bearing used to

orient specimen and fix it
before testing
Fig. 3.6 Test Arrangement
81
45
40
35
30
Load (kN)
25
20
15
10
5
0
0 0.5 1 1.5 2 2.5
Axial
AxisShortening
Shortening(mm)
(mm)
(a) Load-shortening curve
-45
-40
-35
-30
Load (kN)
-25
-20
-15
-10
-5
0
-5.0 -4.0 -3.0 -2.0 -1.0 0.0 1.0 2.0 3.0 4.0 5.0
Deflection (mm)
(b) Load-deflection curve
Fig. 3.7 Typical Load-shortening & Local-deflection curves

of stub column
82
(a) Roof-shaped mechanisms (B section) (b) Roof-shaped mechanisms (B section)
(c) Roof-shaped mechanisms (LB section)
Fig. 3.8 Buckling Mode
83
(a) Flip-disc mechanisms (B section) (b) Flip-disc mechanisms (B section)
(c) Flip-disc mechanisms (HB section)
84
1.10
60Ba/Ns 60Bb/Ns 60Bra&rb/Ns 60HB/Ns
1.05 60BWH/Ns 60LB(all)/Ns AISI/Ns AS4600/Ns
1.00
0.95
Pt/Ns,
0.90
0.85
0.80
0.75
20 40 60 80 100 120 140 160 180
b/t
Fig. 3.10 Comparison of B&LB&HB-section Test Results with

Design Standard (0.60mm)
1.1
42Ba/Ns 42Bb/Ns 42Bc/Ns
1.05 42LB(all)/Ns AISI/Ns AS4600/Ns
0.95
Pt/Ns,
0.9
0.85
0.8
0.75
20 40 60 80 100 120 140 160 180
b/t
Fig. 3.11 Comparison of B & LB-section Test Results with

85
Local mode
Fig. 3.12a Buckling Stress vs. Half-Wavelength for B-section
Local mode
Fig. 3.12b Buckling Stress vs. Half-Wavelength for LB-section
86
Local mode
Fig. 3.12c Buckling Stress vs. Half-Wavelength for HB-section
65
60
55
Effective area Ae (mm )
2
50
Ae vs b (0.60mm)
45
Ae vs b (0.42mm)
40
35
30
25
20
0 20 40 60 80 100 120
Width b (mm)
Fig. 3.13 Effective area vs Width for B-sections

of 0.42 mm & 0.60 mm
87
B-section failure at top
Fig. 3.14 B-section failure mode
(a) HB-section failure at top (b) HB-section failure at longitudinal edge
Fig. 3.15 HB-section failure mode
88
Fig. 3.16 LB-section failure mode
89
3.8 TABLES
Thickness
Web Thickness Radius Lip Length
Specimen of Lip
b (mm) tb (mm) tc (mm) r (mm) R (mm) d (mm) hl (mm) L (mm)
060B20a 22.1 0.60 0.65 0.60 1.20 7.2 1.6 59.7
060B20b 22.0 0.60 0.65 0.60 1.20 7.3 1.6 58.5
060B30a 31.8 0.60 0.65 0.60 1.20 7.4 1.4 89.5
060B30b 31.8 0.60 0.65 0.60 1.20 7.3 1.4 90.2
060B40a 41.8 0.60 0.65 0.60 1.20 7.5 1.6 119.3
060B40b 42.1 0.60 0.65 0.60 1.20 7.6 1.6 119.6
060B50a 51.7 0.60 0.65 0.60 1.20 7.6 1.7 149.9
060B50b 51.9 0.60 0.65 0.60 1.20 7.6 1.7 150.0
060B60a 61.9 0.60 0.65 0.60 1.20 7.5 1.7 178.9
060B60b 61.9 0.60 0.65 0.60 1.20 7.4 1.6 180.2
060B70a 71.8 0.60 0.65 0.60 1.20 7.4 1.6 210.0
060B70b 71.7 0.60 0.65 0.60 1.20 7.4 1.6 209.3
060B80a 81.8 0.60 0.65 0.60 1.20 7.4 1.5 239.9
060B80b 81.7 0.60 0.65 0.60 1.20 7.4 1.6 240.4
060B90a 91.7 0.60 0.65 0.60 1.20 7.5 1.9 270.0
060B90b 91.6 0.60 0.65 0.60 1.20 7.4 2.0 269.7
060B100a 102.3 0.60 0.65 0.60 1.20 7.3 1.7 298.2
060B100b 102.1 0.60 0.65 0.60 1.20 7.5 1.7 295.4
Table 3.1a Measured Dimensions of Box Shaped (B)

Sections (0.60mm) (Holes at 20 mm spacing)
90
Thickness
Specimen of Lip
060B20ra 21.3 0.60 0.65 0.60 1.20 7.5 1.5 59.1
060B20rb 21.3 0.60 0.65 0.60 1.20 7.4 1.5 58.9
060B30ra 30.7 0.60 0.65 0.60 1.20 7.5 1.6 88.6
060B30rb 30.8 0.60 0.65 0.60 1.20 7.5 1.6 88.8
060B40ra 41.3 0.60 0.65 0.60 1.20 7.3 1.8 119.0
060B40rb 41.2 0.60 0.65 0.60 1.20 7.4 1.8 118.3
060B50ra 51.3 0.60 0.65 0.60 1.20 7.5 1.5 148.6
060B50rb 51.3 0.60 0.65 0.60 1.20 7.4 1.6 148.9
060B70ra 71.6 0.60 0.65 0.60 1.20 7.1 1.7 209.0
060B70rb 71.4 0.60 0.65 0.60 1.20 7.1 1.7 209.1
060B80ra 82.0 0.60 0.65 0.60 1.20 7.6 1.4 240.0
060B80rb 82.0 0.60 0.65 0.60 1.20 7.9 1.4 238.7
060B90ra 91.7 0.60 0.65 0.60 1.20 7.2 1.7 268.5
060B90rb 91.5 0.60 0.65 0.60 1.20 7.2 1.9 267.0
060B100ra 101.4 0.60 0.65 0.60 1.20 7.6 1.7 299.2
060B100rb 101.7 0.60 0.65 0.60 1.20 7.5 1.7 299.8
Table 3.1b Measured Dimensions of Box Shaped (B)

Thickness
Specimen of Lip
BWH30 31.5 0.60 0.65 0.60 1.20 7.2 1.4 90.4
BWH50 50.9 0.60 0.65 0.60 1.20 7.3 1.5 150.1
BWH70 71.1 0.60 0.65 0.60 1.20 7.1 1.7 210.1
Table 3.1c Measured Dimensions of Box Shaped (B)

Sections (0.60mm) (No holes)
91
Thickness
Specimen of Lip
042B14a 15.5 0.41 0.45 0.60 1.20 5.7 0.9 41.5
042B14b 15.4 0.41 0.45 0.60 1.20 5.7 1.0 41.7
042B14c 15.6 0.41 0.45 0.60 1.20 5.7 0.9 41.8
042B21a 22.2 0.41 0.45 0.60 1.20 5.7 0.9 62.5
042B21b 22.3 0.41 0.45 0.60 1.20 5.5 1.0 62.3
042B21c 22.2 0.41 0.45 0.60 1.20 5.6 0.9 62.5
042B28a 29.4 0.41 0.45 0.60 1.20 5.7 1.1 83.5
042B28b 29.4 0.41 0.45 0.60 1.20 5.5 0.9 83.4
042B28c 29.3 0.41 0.45 0.60 1.20 5.7 1.0 83.6
042B35a 36.1 0.41 0.45 0.60 1.20 5.5 1.0 104.6
042B35b 36.1 0.41 0.45 0.60 1.20 5.6 1.0 104.1
042B35c 36.3 0.41 0.45 0.60 1.20 5.5 1.1 104.3
042B42a 43.2 0.41 0.45 0.60 1.20 5.5 1.0 125.5
042B42b 43.2 0.41 0.45 0.60 1.20 5.4 1.0 125.0
042B42c 43.4 0.41 0.45 0.60 1.20 5.4 1.0 125.6
042B49a 50.2 0.41 0.45 0.60 1.20 5.6 1.0 146.0
042B49b 50.2 0.41 0.45 0.60 1.20 5.7 1.0 146.5
042B49c 50.2 0.41 0.45 0.60 1.20 5.7 1.0 146.2
042B56a 57.2 0.41 0.45 0.60 1.20 5.7 0.9 167.2
042B56b 57.1 0.41 0.45 0.60 1.20 5.7 1.0 165.6
042B56c 57.1 0.41 0.45 0.60 1.20 5.7 1.0 165.1
042B63a 64.4 0.41 0.45 0.60 1.20 5.7 1.1 188.5
042B63b 64.4 0.41 0.45 0.60 1.20 5.7 1.1 187.8
042B63c 64.6 0.41 0.45 0.60 1.20 5.5 1.1 187.5
042B70a 71.2 0.41 0.45 0.60 1.20 5.7 1.1 209.3
042B70b 71.4 0.41 0.45 0.60 1.20 5.7 1.1 208.5
042B70c 71.3 0.41 0.45 0.60 1.20 5.6 0.9 209.2
Table 3.1d Measured Dimensions of Box Shaped (B)

Sections (0.42mm) (Clamps)
92
Thickness
Flange Web Thickness Radius Lip Length
Specimen of Lip
h (mm) b (mm) tb (mm) tc (mm) r (mm) R (mm) d (mm) hl (mm) L (mm)
060LB20a 10.4 20.8 0.60 0.65 0.60 0.60 5.9 1.3 59.21
060LB20b 10.4 20.9 0.60 0.68 0.60 0.60 5.7 1.3 58.4
060LB30a 14.9 32.2 0.60 0.69 0.60 0.60 8.2 1.3 89.7
060LB30b 15.0 31.6 0.60 0.65 0.60 0.60 8.0 1.4 89.5
060LB40a 19.7 42.4 0.60 0.64 0.60 0.60 10.6 1.6 119.5
Table 3.2a Measured Dimensions of Lipped-Box Shaped (LB)

Sections (0.60mm) (No holes & Clamps)
Thickness
Specimen of Lip
060LB20ra 10.7 20.3 0.60 0.65 0.60 0.60 6.7 2.7 60.07
060LB20rb 10.7 20.7 0.60 0.65 0.60 0.60 6.7 2.1 60.5
060LB30ra 15.5 30.6 0.60 0.65 0.60 0.60 6.8 2.4 90.3
060LB30rb 15.6 30.5 0.60 0.65 0.60 0.60 6.9 2.3 89.9
060LB40ra 20.5 40.9 0.60 0.65 0.60 0.60 6.9 2.4 120.2
060LB40rb 20.6 40.8 0.60 0.65 0.60 0.60 6.6 2.3 120.3
060LB40b 20.0 41.7 0.60 0.65 0.60 0.60 10.6 1.5 119.4
Table 3.2b Measured Dimensions of Lipped-Box Shaped (LB)

93
Thickness
Specimen of Lip
060LB20na 10.9 20.8 0.60 0.65 0.60 0.60 5.6 1.6 59.55
060LB20nb 10.9 20.7 0.60 0.65 0.60 0.60 5.5 1.5 59.2
060LB30na 16.1 30.2 0.60 0.65 0.60 0.60 7.8 1.4 89.7
060LB30nb 16.0 30.4 0.60 0.65 0.60 0.60 7.9 1.4 89.7
060LB40na 20.9 40.7 0.60 0.65 0.60 0.60 10.4 1.4 119.6
060LB40nb 21.0 40.6 0.60 0.65 0.60 0.60 10.3 1.4 119.7
Table 3.2c Measured Dimensions of Lipped-Box Shaped (LB)

Thickness
Specimen of Lip
042LB30a 15.7 32.6 0.41 0.45 0.60 0.60 6.5 1.0 90.64
042LB30b 15.7 32.8 0.41 0.45 0.60 0.60 5.8 1.1 90.4
042LB40a 20.6 41.6 0.41 0.45 0.60 0.60 6.1 1.0 119.2
042LB40b 20.8 41.9 0.41 0.45 0.60 0.60 6.1 1.0 119.8
042LB50a 26.0 51.5 0.41 0.45 0.60 0.60 6.0 1.1 150.4
042LB50b 25.9 51.7 0.41 0.45 0.60 0.60 5.9 1.1 150.7
Table 3.2d Measured Dimensions of Lipped-Box Shaped (LB)

94
Flange Web Thickness Radius Length

Specimen
bf (mm) b (mm) tb (mm) tc (mm) r (mm) L (mm)
060HB20 19.3 20.5 0.60 0.65 1.10 60.1
060HB20ra 19.1 20.6 0.60 0.65 1.10 60.0
060HB30a 29.3 30.5 0.60 0.65 1.10 89.5
060HB30b 29.1 30.9 0.60 0.65 1.10 90.5
060HB30c 29.4 30.6 0.60 0.65 1.10 91.0
060HB40a 39.2 41.0 0.60 0.65 1.10 119.8
060HB40b 40.4 40.5 0.60 0.65 1.10 120.0
060HB60 59.6 61.0 0.60 0.65 1.10 179.8
060HB80 79.4 81.2 0.60 0.65 1.10 240.1
060HB100 99.5 100.0 0.60 0.65 1.10 300.0
Table 3.3 Measured Dimensions of Hexagonal Shaped (HB)

Sections (0.60mm) (Teflon)
95
Specimen Dimensions Table Hole/Clamping 0.2% Proof Stress
Holes at 20 mm
060B20a~~060B100b 1a 711
spacing
Holes at 10 mm
060B20ra~~060B100rb 1b 711
spacing
BWH30~~BWH70 1c No holes or clamping 711
042B14a~~042B70c 1d Clamps 634
060LB20a~~060LB40a 2a No holes or clamping 711
060LB20ra~~060LB40rb Holes at 10 mm
2b 711
060LB40b spacing
060LB20na~~060LB40nb 2c Clamps 670
042LB30a~~042LB50b 2d Clamps 634
060HB20~~060HB100 3 Teflon/Steel support 711
Table 3.4 Hole/Clamping Configuration and Proof Stresses
96
Specimen Side1 Max (mm) Side2 Max (mm) Side3 Max (mm) Side4 Max (mm)
Ed11 Ce10 Ed12 Ed21 Ce20 Ed22 Ed11 Ce10 Ed12 Ed21 Ce20 Ed22
060B20ra 0.26 0.16 0.22 0.16 0.18 0.25 0.30 0.20 0.16 0.10 0.21 0.11
060B20rb 0.30 0.28 0.24 0.35 0.33 0.33 0.17 0.10 0.20 0.21 0.27 0.33
060B30ra 0.24 0.28 0.15 0.12 -0.20 -0.21 0.15 0.16 0.33 0.25 0.22 0.13
060B30rb -0.22 0.10 0.30 -0.21 0.30 0.17 0.16 0.23 0.12 0.26 0.25 0.19
060B40ra -0.16 -0.21 -0.41 -0.29 -0.21 -0.26 -0.26 -0.20 -0.24 0.24 -0.28 0.19
060B40rb 0.36 -0.25 0.37 0.30 -0.61 -0.44 -0.31 -0.27 -0.43 0.46 -0.47 -0.32
060B50ra -0.16 -0.24 -0.16 0.16 -0.42 -0.23 -0.33 0.18 -0.18 0.18 -0.19 0.13
060B50rb -0.38 -0.27 -0.17 -0.28 -0.26 0.26 0.12 -0.27 -0.39 -0.17 -0.45 -0.20
060B70ra -0.18 0.18 -0.11 -0.32 0.25 0.20 -0.13 -0.28 0.20 -0.19 0.17 0.23
060B70rb -0.19 0.23 0.21 -0.18 0.28 -0.22 0.24 -0.17 -0.24 -0.21 0.20 0.24
060B80ra -0.40 -0.71 -0.40 -0.28 -0.52 -0.82 -0.27 0.24 0.42 -0.18 -0.40 0.27
060B80rb -0.50 -0.39 0.20 0.22 -0.64 -1.03 -0.61 -0.50 -0.11 0.29 -0.92 -0.76
060B90ra -0.24 -0.45 -0.28 0.29 -0.17 -0.21 0.32 -0.18 -0.16 -0.29 0.22 0.19
060B90rb -0.15 0.27 0.22 0.10 0.30 -0.23 0.29 0.21 -0.27 0.15 -0.27 -0.37
060B100ra -0.20 -0.24 -0.30 -0.17 -0.32 -0.36 -0.49 -0.49 -0.25 -0.17 -0.64 -0.61
060B100rb -0.71 -0.40 -0.14 -0.49 -0.66 -0.70 -0.77 -0.41 0.14 -0.44 -0.45 -0.52
060LB20ra 0.23 0.33 0.33 -0.11 0.05 0.2
060LB20rb 0.08 0.2 0.19 0.1 0.18 -0.18
060LB30ra 0.21 -0.18 0.3 -0.16 0.31 0.22
060LB30rb -0.25 -0.22 0.28 0.35 -0.37 0.24
060LB40ra -0.31 -0.14 -0.29 -0.12 -0.3 -0.23
060LB40rb -0.11 -0.36 0.22 -0.18 -0.32 -0.18
Table 3.5a Measured Imperfections of B & LB Sections (Longitudinal)
Specimen Side1 Max (mm) Side2 Max (mm) Side3 Max (mm) Side4 Max (mm)
Ed11 Ce10 Ed12 Ed21 Ce20 Ed22 Ed11 Ce10 Ed12 Ed21 Ce20 Ed22
060B20ra 0.02 -0.04 -0.11 -0.02 -0.04 0.09 -0.01 -0.33 0.11 -0.06 -0.14 -0.12
060B20rb -0.18 0.02 -0.22 -0.03 -0.02 -0.19 -0.06 -0.05 -0.08 0.04 -0.18 -0.05
060B30ra -0.11 0.05 0.09 -0.06 -0.06 -0.17 0.02 0.01 -0.19 -0.10 -0.03 0.02
060B30rb -0.03 -0.17 -0.31 -0.17 0.08 0.03 0.15 0.16 0.04 -0.23 -0.23 0.02
060B40ra -0.07 0.03 -0.18 0.09 0.15 0.02 -0.10 0.07 -0.10 0.01 -0.05 0.01
060B40rb -0.03 0.05 -0.24 0.08 0.01 -0.05 0.12 -0.27 -0.04 0.17 -0.10 -0.02
060B50ra 0.21 0.04 0.08 -0.11 -0.10 -0.04 0.02 -0.11 0.17 0.22 0.06 0.12
060B50rb -0.08 -0.05 -0.04 -0.17 0.09 -0.07 0.09 0.15 0.04 0.10 -0.08 -0.12
060B70ra 0.13 -0.02 -0.09 -0.03 0.05 0.13 0.19 -0.06 0.05 0.04 0.15 -0.08
060B70rb -0.01 -0.05 0.12 0.01 0.31 0.16 0.23 0.03 -0.26 -0.15 0.13 -0.06
060B80ra -0.16 -0.17 0.34 0.38 -0.41 0.26 0.15 -0.30 0.31 -0.27 0.25 -0.38
060B80rb 0.25 -0.14 0.35 -0.17 -0.15 0.15 0.39 -0.31 0.39 -0.13 0.75 0.33
060B90ra -0.33 0.20 -0.44 0.26 0.25 -0.21 -0.11 -0.07 0.37 0.16 0.34 -0.21
060B90rb -0.31 0.10 0.10 0.37 0.16 -0.22 -0.59 0.20 -0.21 0.05 0.29 0.19
060B100ra 0.12 0.23 0.28 -0.22 0.09 0.23 0.19 -0.23 0.16 0.38 0.28 0.19
060B100rb 0.38 0.31 0.31 0.55 0.27 0.60 0.33 0.18 0.15 -0.12 -0.27 0.11
060LB20ra 0.16 0.08 -0.07 0.05 0.17 0.15
060LB20rb 0.05 0.06 0.06 -0.02 0.01 0.05
060LB30ra 0.10 0.07 0.01 0.15 -0.05 0.22
060LB30rb 0.02 -0.04 0.08 -0.02 -0.02 0.15
060LB40ra 0.27 -0.02 0.09 0.13 0.08 0.64
060LB40rb -0.07 -0.01 0.15 0.01 0.15 -0.02
Table 3.5b Measured Imperfections of B & LB Sections (Transverse)
97
Specimen E (103 MPa) fy (MPa)
060Coupon 216 711
060Coupon 216 670

(New Sheet)
042Coupon 220 634
Table 3.6 Mean Values of Tensile Coupon Tests
98
Specimen B20a B20b B30a B30b B40a B40b B50a B50b B60a
A (mm2) 59.4 59.8 83.9 83.5 107.6 108.5 131.5 131.9 155.8
fy (MPa) 711 711 711 711 711 711 711 711 711
Py (kN) 42.3 42.5 59.7 59.4 76.5 77.2 93.5 93.8 110.8
Ns (kN) 35.9 36.0 38.8 38.6 40.3 40.4 41.2 41.1 41.6
Pt (kN) 35.2 30.4 35.8 35.7 38.1 37.8 39.1 37.8 39.0
Pt//Ns 0.98 0.84 0.92 0.93 0.95 0.94 0.95 0.92 0.94
fol (MPa) 712.5 720.2 328.0 328.4 200.8 188.0 123.4 122.1 85.0
Pt/A 592.2 508.8 426.7 427.6 354.2 348.4 297.3 286.5 250.4
b/t 34.7 36.7 53.0 53.0 69.6 70.2 86.1 86.5 103.2
Nol (kN) 42.4 43.0 27.5 27.4 21.6 20.4 16.2 16.1 13.2
Pcr (kN) \ \ 29.9 31.3 22.7 22.4 17.3 17.0 13.4
Pcr/Nol \ \ 1.09 1.14 1.05 1.1 1.07 1.06 1.01
Note: Holes at 20 mm spacing; 0.60 mm thickness
Table 3.7a Summary of Results for B-Sections
Specimen B60b B70a B70b B80a B80b B90a B90b B100a B100b
A (mm2) 155.6 179.3 179.1 203.4 203.3 226.9 226.3 252.2 252.1
fy (MPa) 711 711 711 711 711 711 711 711 711
Py (kN) 110.7 127.5 127.4 144.6 144.6 161.3 160.9 179.3 179.3
Ns (kN) 41.3 41.7 41.6 41.9 42.0 42.3 42.2 42.2 42.6
Pt (kN) 39.4 \ 41.8 42 42 \ 41.5 39.2 41.1
Pt//Ns 0.96 \ 1 1 1 \ 0.98 0.93 0.97
fol (MPa) 84.4 62.2 65.0 47.2 47.5 38.3 38.3 29.9 30.1
Pt/A 253.2 \ 233.3 206.5 206.6 \ 183.4 155.5 163
b/t 103.2 119.7 119.6 136.3 136.2 152.9 152.7 170.4 170.1
Nol (kN) 13.1 11.2 11.6 9.6 9.7 8.7 8.7 7.5 7.6
Pcr (kN) 13.7 \ 10.6 11.0 9.5 \ 10.3 8.6 8.4
Pcr/Nol 1.04 \ 0.91 1.15 0.99 \ 1.19 1.14 1.11
Table 3.7a Summary of Results for B-Sections (continued)
99
Specimen B20ra B20rb B30ra B30rb B40ra B40rb B50ra B50rb

A (mm2) 58.6 58.4 80.9 81.2 105.5 105.7 130.6 130.2
fy (MPa) 711 711 711 711 711 711 711 711
Py (kN) 41.6 41.6 57.6 57.7 75.0 75.2 92.9 92.6
Ns (kN) 36.0 35.9 38.7 38.6 39.8 40.0 40.9 40.8
Pt (kN) 33.9 32.6 34.7 34.1 37.1 37.1 38.1 37.9
Pt//Ns 0.94 0.91 0.9 0.88 0.93 0.93 0.93 0.93
fol (MPa) 803.8 800.5 382.2 378.0 208.6 208.8 129.5 130.4
Pt/A 578.8 557.5 428.4 420.5 352 351 291.9 291.3
b/t 35.5 35.5 51.2 51.3 68.8 68.7 85.5 85.5
Nol (kN) 47.1 46.8 30.9 30.7 22.0 22.1 16.9 17.0
Pcr (kN) \ \ 32.9 32.8 23.6 23.8 16.3 15.6
Pcr/Nol \ \ 1.06 1.07 1.07 1.08 0.96 0.92
Table 3.7b Summary of Results for B-Sections

2
A (mm ) 178.2 177.4 204.8 205.4 226.5 225.8 250.8 251.1
fy (MPa) 711 711 711 711 711 711 711 711
Py (kN) 126.8 126.1 145.6 137.7 161.1 160.6 178.4 178.6
Ns (kN) 41.3 41.2 42.3 41.2 41.9 41.9 42.6 42.5
Pt (kN) 39.4 39.2 40.6 38.6 39.5 38.7 36.8 39.3
Pt//Ns 0.96 0.95 0.96 0.94 0.94 0.92 0.86 0.93
fol (MPa) 65.5 66.2 48.3 48.8 39.4 39.9 32.0 31.8
Pt/A 221.3 221 198.3 188 174.2 171.3 146.6 156.5
b/t 119.4 118.9 136.7 136.7 152.8 152.5 169.1 169.5
Nol (kN) 11.7 11.7 10.1 10.2 8.9 9.0 8.0 8.0
Pcr (kN) 11.9 10.6 10.0 9.5 10.5 8.8 6.7 6.9
Pcr/Nol 1.02 0.9 0.99 0.93 1.18 0.97 0.84 0.86
Table 3.7b Summary of Results for B-Sections (continued)
100
Specimen BWH30 BWH50 BWH70
A (mm2) 93.0 140.0 188.4
Fy (MPa) 711 711 711
Py (kN) 66.2 99.6 133.9
Ns (kN) 43.4 45.6 46.2
Pt (kN) 42.1 45.7 45.5
Pt//Ns 0.97 1.00 0.98
fol (MPa) 327.1 124.8 63.8
Pt/A 452.6 326.4 241.5
b/t 50.0 83.3 116.7
Nol (kN) 30.4 17.5 12.0
Pcr (kN) 31.6 18.0 12.2
Pcr/Nol 1.04 1.03 1.02
Note: No holes; 0.60 mm thickness
Table 3.7c Summary of Results for B-Sections
Specimen B14a B14b B14c B21a B21b B21c B28a B28b B28c
2
A (mm ) 33.2 33.1 33.4 44.3 44.0 44.2 56.0 55.6 55.9
fy (MPa) 634 634 634 634 634 634 634 634 634
Py (kN) 21.1 21.0 21.2 28.1 27.9 28.1 35.5 35.3 35.5
Ns (kN) 18.7 18.7 18.7 20.0 19.7 19.8 20.6 20.3 20.6
Pt (kN) 17.1 17.3 17.1 17.7 16.8 17.4 19.1 19.1 19.3
Pt//Ns 0.91 0.92 0.91 0.89 0.85 0.88 0.93 0.94 0.94
fol (MPa) 721.1 737.9 708.8 337.8 337.9 337.7 195.8 191.2 194.3
Pt/A 514.6 523.4 512.1 400.0 382.1 393.3 341.3 343.3 345.3
b/t 35.0 35.0 35.0 52.5 52.5 52.5 70.0 70.0 70.0
Nol (kN) 24.0 24.4 23.7 14.9 14.9 14.9 11.0 10.6 10.9
Pcr (kN) \ \ \ 13.7 14.5 13.0 8.9 10.3 10.3
Pcr/Nol \ \ \ 0.91 0.98 0.87 0.81 0.97 0.94
Note: Clamps; 0.42 mm thickness
Table 3.8 Summary of Results for B-Sections
101
A (mm2) 66.6 66.7 66.9 78.2 78.7 78.4 89.8 90.1 90.2
fy (MPa) 634 634 634 634 634 634 634 634 634
Py (kN) 42.2 42.3 42.4 49.6 49.9 49.7 57.0 57.2 57.2
Ns (kN) 20.7 20.8 20.7 20.9 20.8 20.9 21.2 21.4 21.4
Pt (kN) 20.0 20.0 20.1 20.5 20.5 20.5 19.6 20.8 21
Pt//Ns 0.97 0.96 0.97 0.98 0.99 0.98 0.92 0.97 0.98
fol (MPa) 125.2 125.2 124.9 85.2 85.9 85.6 62.8 63.0 63.1
Pt/A 300.4 299.9 300.7 262.1 260.6 261.4 218.2 230.8 232.9
b/t 87.5 87.5 87.5 105.0 105.0 105.0 122.5 122.5 122.5
Nol (kN) 8.3 8.4 8.4 6.7 6.8 6.7 5.6 5.7 5.7
Pcr (kN) 7.0 8.1 7.7 6.1 5.7 7.2 5.3 6.3 6.3
Pcr/Nol 0.84 0.97 0.92 0.92 0.84 1.08 0.94 1.10 1.10
Table 3.8 Summary of Results for B-Sections (continued)
A (mm2) 101.6 101.5 101.4 113.2 113.4 113.2 124.4 124.6 124.5
fy (MPa) 634 634 634 634 634 634 634 634 634
Py (kN) 64.5 64.4 64.3 71.8 71.9 71.8 78.9 79.0 79.0
Ns (kN) 21.5 21.5 21.5 21.6 21.6 21.3 21.7 21.7 21.5
Pt (kN) 20.9 20.7 21.1 20.9 20.3 20.3 20.9 20.4 20.5
Pt//Ns 0.97 0.96 0.98 0.97 0.94 0.95 0.96 0.94 0.95
fol (MPa) 47.4 48.2 48.1 37.9 37.8 37.6 30.7 31.1 30.0
Pt/A 205.7 204 208.1 184.6 179.1 179.3 168 163.7 164.3
b/t 140.0 140.0 140.0 157.5 157.5 157.5 175.0 175.0 175.0
Nol (kN) 4.8 4.9 4.9 4.3 4.3 4.3 3.8 3.9 3.7
Pcr (kN) 4.7 5.5 4.7 3.9 4.0 3.9 3.7 3.5 3.5
Pcr/Nol 0.98 1.12 0.96 0.92 0.93 0.92 0.95 0.91 0.94
Note:) Clamps; 0.42 mm thickness
Table 3.8 Summary of Results for B-Sections(continued)
102
Specimen LB20a LB20b LB30a LB30b LB40a LB40b

A (mm2) 62.8 63.1 94.6 95.3 132.0 128.3
fy (MPa) 711 711 711 711 711 711
Py (kN) 44.6 45.3 67.3 67.7 93.9 91.2
Ns (kN) 39.4 39.8 52.8 \ 60.9 56.6
Pt (kN) 37.0 38.0 48.7 \ 53.0 53.3
Pt//Ns 0.94 0.96 0.92 \ 0.87 0.94
fol (MPa) 902.3 905.1 390.0 \ 214.5 220.4
Pt/A 589.5 602.5 515 \ 401.6 415.4
b/t 33.3 33.3 50.0 \ 66.7 66.7
Nol (kN) 56.6 57.1 36.9 \ 28.3 28.3
Pcr (kN) \ \ 35.7 \ \ 29.0
Pcr/Nol \ \ 0.97 \ \ 1.02
Note: No holes & Clamps; 0.60 mm thickness
Table 3.9a Summary of Results for LB-Sections
Specimen LB20ra LB20rb LB30ra LB30rb LB40ra LB40rb

A (mm2) 67.1 67.1 90.9 91.1 115.5 115.0
fy (MPa) 711 711 711 711 711 711
Py (kN) 47.7 47.7 64.6 64.8 82.1 81.8
Ns (kN) 38.6 38.5 48.0 48.2 51.8 51.5
Pt (kN) 37.3 35.6 46.0 45.8 47.3 44.6
Pt//Ns 0.97 0.93 0.96 0.95 0.91 0.87
fol (MPa) 978.4 945.2 440.2 442.7 238.3 238.6
Pt/A 555.4 531.1 506.1 502 409.9 387.7
b/t 33.3 33.3 50.0 50.0 66.7 66.7
Nol (kN) 65.7 63.4 40.0 40.3 27.5 27.4
Pcr (kN) \ \ 37.5 40.1 24.3 \
Pcr/Nol \ \ 0.94 0.99 0.88 \
Table 3.9b Summary of Results for LB-Sections
103
Specimen LB20na LB20nb LB30na LB30nb LB40na LB40nb

2
A (mm ) 61.9 61.9 91.8 91.9 121.8 121.4
fy (MPa) 670 670 670 670 670 670
Py (kN) 41.5 41.5 61.5 61.6 81.6 81.4
Ns (kN) 37.6 37.7 51.2 51.3 58.8 58.7
Pt (kN) 36.4 35.4 45.8 45.8 50.6 51.3
Pt//Ns 0.97 0.94 0.89 0.89 0.86 0.87
fol (MPa) 954.6 953.2 411.7 412.6 225.6 224.2
Pt/A 587.8 571.9 498.9 498.5 415.4 422.6
b/t 33.3 33.3 50.0 50.0 66.7 66.7
Nol (kN) 59.1 59.0 37.8 37.9 27.5 27.2
Pcr (kN) \ \ 37.2 36.2 28.1 25.2
Pcr/Nol \ \ 0.98 0.95 1.02 0.93
Table 3.9c Summary of Results for LB-Sections

A (mm2) 62.8 62.9 78.3 78.7 95.4 95.2
fy (MPa) 634 634 634 634 634 634
Py (kN) 39.9 39.9 49.7 49.9 60.5 60.4
Ns (kN) 26.9 26.9 28.0 28.1 28.2 28.1
Pt (kN) 24.8 23.5 25.1 25.0 26.1 26.1
Pt//Ns 0.92 0.87 0.90 0.89 0.93 0.93
fol (MPa) 183.7 185.2 108.1 109.3 69.7 69.0
Pt/A 394.7 373.5 320.4 317.9 273.6 274.1
b/t 73.0 73.0 98.0 98.0 122.0 122.0
Nol (kN) 11.5 11.7 8.5 8.5 6.7 6.6
Pcr (kN) 12.6 11.3 8.9 8.5 6.2 7.1
Pcr/Nol 1.09 0.97 1.05 0.99 0.93 1.08
Table 3.9d Summary of Results for LB-Sections
104
Specimen HB20a HB20b HB30a HB30b HB30c HB40a HB40b HB60 HB80 HB100
A (mm2) 71.5 71.6 107.5 108.2 107.9 145.0 144.2 217.1 289.3 358.6
fy (MPa) 711\ 711 711 711 711 711 711 711 711 711
Py (kN) 50.8 50.9 76.4 77.0 76.7 103.1 102.6 154.4 205.7 255.0
Ns (kN) 42.1 42.1 46.2 46.3 46.3 48.3 48.3 50.2 51.9 51.8
Pt (kN) 37.3 37.4 39.9 43.6 41.1 44.4 43.6 45.4 47.7 45.5
Pt//Ns 0.89 0.89 0.86 0.94 0.89 0.92 0.90 0.90 0.92 0.88
fol (MPa) 764.3 766.2 340.3 345.9 341.2 200.5 197.1 86.5 48.6 31.9
Pt/A 521.8 522.3 371.3 403 380.9 306 302.4 208.9 164.9 126.9
b/t 33.3 33.3 50.0 50.0 50.0 66.7 66.7 100.0 133.3 166.7
Nol (kN) 54.6 54.9 36.6 37.4 36.8 29.1 28.4 18.8 14.0 11.4
Pcr (kN) \ \ 34.5 37.2 35.4 27.8 29.4 18.9 14.9 12.8
Pcr/Nol \ \ 0.94 0.99 0.96 0.96 1.04 1.01 1.06 1.12
Note: Teflon; 0.60 mm thickness
Table 3.10 Summary of Results for HB-Sections
105
Chapter 4: STRENGTH TESTS OF COLD-FORMED BOX-SHAPED SECTION LONG COLUMNS
CHAPTER 4

SHAPED-SECTION LONG COLUMNS
CONTENTS
4.1 INTRODUCTION....................................................................................................107
4.2 TEST SPECIMENS .................................................................................................107

4.2.1 General........................................................................................................................................... 107
4.2.2 Labelling......................................................................................................................................... 109
4.3 LONG COLUMN TESTS .......................................................................................111

4.3.1 Testing ............................................................................................................................................ 111
4.3.2 Measurement of flexural rigidity (E*I)........................................................................................... 112
4.3.3 Test specimen behaviour and ultimate load-capacity..................................................................... 112
4.4 ANALYSES ..............................................................................................................114

4.4.1 Elastic local buckling analyses....................................................................................................... 114
4.4.2 Test Results compared with the Design Standards ......................................................................... 115
4.4.2.1 General.................................................................................................................................................... 115
4.4.2.2 Test results comparisons with design standards ...................................................................................... 116
4.5 SUMMARY ..............................................................................................................116
4.6 FIGURES..................................................................................................................118
4.7 TABLES....................................................................................................................118
APPENDIX 4.1 ................................................................................................................139
106
CHAPTER 4

SHAPED-SECTION LONG COLUMNS
4.1 INTRODUCTION
Research on these steels as stub columns in compression, which is detailed in

Chapter 3, has shown that the greatest effect of the low strain hardening was for
the stockier sections where strain hardening plays an important role. For the
more slender sections where elastic local buckling and post-local buckling are
more important, the effect of low strain hardening does not appear to be as
significant which is contrary to recent design proposals in the USA where it was
believed that the more slender sections had been also influenced. The aim of
this Chapter is to present the second stage test results for long columns. The
tests were performed on lipped-box shaped specimens (LB-section) fabricated
from G550 steel sheets to AS 1397 with 0.6 mm and 0.42 mm thickness similar
to those in Chapter 3. In this Chapter, two series of long columns tests (LB-
section) with pin-ends are described. The purpose of the tests was to investigate
whether high strength steel members with low strain hardening can be designed
according to the existing standard AS/NZS 4600 or the AISI Specification or
whether these standards need to be modified for high strength steel.
4.2 TEST SPECIMENS
4.2.1 General
107
In Chapter 3, a series of stub column tests was presented on sections fabricated

from G550 steel sheets in 0.60 mm and 0.42 mm thickness. For cold-formed
steel sections, which generally have thin-walled plate elements, the stub-column
test is aimed at determining the effect of local buckling as well as the effect of
cold-forming, residual stress and yielding on the section capacity in
compression.
In order to determine the capacity of long columns in compression, local and

overall buckling and the interaction between them should be considered. For the
high strength steel columns assembled from thin plates, the section
configuration is an important factor affecting their behaviour. Thin plate
elements will generally continue to carry load after local buckling into the post-
buckling range so that local buckling does not mean failure of the whole
column. However, a singly-symmetric section may have a neutral axis shift after
local buckling occurs resulting in an additional moment. To eliminate this
problem, in this test program, doubly-symmetric sections were chosen. These
sections are the LB-sections previously described in Chapter 3 for stub
columns.
In this Chapter, a long column is defined as a member whose length is

considerably larger than any of its cross-sectional dimensions and which is
subjected to compression in the longitudinal direction. A centrally loaded
column means that the resultant compressive force is approximately coincident
with its longitudinal centroidal axis (Galambos, 1988).
The tests were performed on closed sections brake pressed from

aluminium/zinc-coated Grade G550 structural steel sheet to AS 1397 in 0.42
mm and 0.6 mm thickness. The sections tested are shown in Fig. 4.1. Epoxy was
used to close the LB-sections. The lengths of the LB sections ranged from 450
108
to 1100 mm for the 0.6 mm sheet steel and 550 to 1700 mm for the 0.42 mm
sheet steel. The cross-section dimensions of the specimens are shown in Tables
4.1, for the nomenclature shown in Fig. 4.2. The cross-section dimensions are
the average of the measured values. The base metal thickness (tb) was measured
by removing the zinc coating by acid-etching. The ends of each specimen were
milled flat and parallel to ensure full contact between specimens and end
bearing. Bolts & clamps were also used on the LB-sections as specified in Table
4.2 and shown in Fig. 4.3.
4.2.2 Labelling
The LB-sections were divided into two different series: one for the 0.6 mm
sheet steel and one for the 0.42 mm sheet steel. The test specimens were
labelled such that the thickness of steel sheet, type of section, nominal width of
specimen and specimen number were expressed by the label.
For example, the label “060LB20P450a” defines the following specimen:

• The first three numbers indicate that the specimen is fabricated from 0.60
mm steel sheet.
• Th fourth and fifth letters (LB) indicate that the specimen is a lipped box
• The “20” indicates that the nominal width of specimen is 20 mm
• The “P” indicates that the specimen is pin-ended
• The "450" indicates that the nominal length of specimen is 450 mm
• The last letter “a” indicates that the specimen was the first tested (alternative
b)
109
Geometric imperfections were measured for all of the specimens. Only minor
axis imperfections were measured and two methods were used. For the 0.60 mm
specimens, a small vertical scale was designed and moved along the specimens
and the scale was read optically. Readings were obtained using an optical level
at 20 mm intervals from end to end. For the 0.42 mm specimens, a more
accurate instrument was designed to measure the geometric imperfections as
shown in Fig. 4.4. The main components included a guide frame, a laser sensor,
a Datataker data logger and a synchronous motor. The laser sensor was attached
to a movable seat, which could move along two accurately levelled channels.
The seat was drawn by a synchronous motor in order to keep the laser sensor
moving smoothly. The laser sensor can move up and down along the upright
poles and from left to right using a screw adjuster. The specimen was positioned
on the seat in front of the frame using a clamp to make it firm and level. The
data was recorded using the Datataker data logger. Once the specimen was
properly positioned on the seat and fixed in position, the laser sensor was
moved along the specimen. The readings were taken at 15 mm or 20 mm
intervals. For each side of a specimen, the measurements were taken along three
lines in the longitudinal direction of the specimen, two along the corner
(approximately 5mm away from the corner or edge) and one along the centre.
The results were used in the finite element modelling analysis. The maximum
imperfections are shown in Table 4.3 and were computed relative to a straight
line between the ends.
4.2.4 Material P roperties
The material and its properties are fully described in Chapter 3. The proof
stresses are given in Table 4.2.
110
4.3 LONG COLUMN TESTS
4.3.1 Testing
The test rig is shown in Fig. 4.5 and the test arrangement is shown in Fig. 4.6.
The rig consisted of the Sintech/MTS-300kN testing machine with pin-ended
bearings. The load and shortening were recorded using the Sintech data
acquisition system. The compressive deformation rate was 0.05 mm/min.
The measured specimen lengths are given in Table 4.1. The measured specimen
lengths (L) were different from the lengths of the pin-ended column (Lt) which
were obtained as the sum of the specimen length and the distance from the
platen face to the centre of the end bearings (Lt=L+150 mm for the 0.60 mm
series and Lt=L+108 mm for the 0.42 mm series). The first column was found to
fracture in the epoxy at loads very close to the ultimate load. It was therefore
decided to drill 3 mm holes in the corner lips and to place small diameter bolts
and nuts to ensure that the corners did not come apart for the first series. These
bolts were only located at the ends (four for each) and were only used for the
0.60 mm sections. Clamps were used on the lips of 0.42 mm sections as shown
in Fig. 4.3.
The central deflections were measured using two transducers on two opposite
sides of the LB-sections. Two strain gauges were attached at the centre of each
side of each column. The transducers and strain gauges were connected to the
SPECTRA data acquisition system.
During the test, the trial axial load which was 1/15 or 1/10 of the estimated
ultimate load was applied and readings of the strain gauges were taken. Based
on a calculation of the location of the initial central eccentricity (e) of the action
111
line as shown in Fig. 4.6, the location of the column was adjusted at each end.
This procedure was repeated until the value of the initial eccentricity (e) was
approximately equal to the nominal value of L/1000 which is the maximum
normally specified in structural design standards. The values are shown in Table
4.4 for the method presented in Appendix 4.1.
4.3.2 Measurem ent of flexural rigidity (E*I)
For the LB sections in 0.60 mm thickness, the flexural rigidities (E*I) (minor
axis) were measured. The ends of each specimen were simply supported and sag
of the specimen between supports under self-weight was permitted. A point load
was then applied at mid-length. The mid-length deflections were measured with
increasing load. The measured flexural rigidities and theoretical values obtained
from a finite strip analysis program (THIN-WALL) (Papangelis and Hancock,
1998) using the measured dimensions are shown in Table 4.5.
It can be seen that the experimental results are in good agreement with the
theoretical results except for the 060LB40P450 specimen as detailed in Table
4.5. This short wide section is very sensitive to a deviation of measured
deflection, which may be the main reason for the large difference between the
experimental results and theoretical results. The accurate rigidity measurements
mean that the epoxy was effective in causing the two hat-shaped halves to act in
a fully integral manner.
4.3.3 Test specim en behaviour and ultimate load-capacity
The graphs in Fig. 4.7 & 4.8 show that initially the columns remained elastic
with slope of the load-shortening diagram approximately constant after some
initial take-up. The ultimate loads (Pt) are given in Table 4.7 for each specimen.
112
For the 060LB20 and 060LB30 series, the ultimate load (Pt) was less than the
theoretical local buckling load (Nol) so that the slope kept approximately
constant until the ultimate load was reached. The 450 mm specimens in 0.60
mm material and the 550 mm specimens in 0.42 mm material buckled in-
elastically soon after the ultimate load with a sudden drop in load. The 900 mm
specimens in 0.60 mm material and the 1100 mm and 1700 mm specimens in
0.42 mm material generally had a more gradual decrease in load until inelastic
local buckling occurred at which point there was a sudden drop in load as for
the 450 mm specimens. Only the 042LB30P1100 specimens appeared to
inelastically locally buckle before the ultimate load caused by overall
deformation was reached. For the intermediate length columns in 0.42 mm
thickness and the 060LB40P450 column, the slope reduced continuously from
the local buckling load to reaching the ultimate load.
By observation of the surface of the specimens, the local buckling behaviour

can be observed. Initially, elastic local buckling occurred with many half-
wavelengths occurring along the length of the specimens. Eventually the
columns entered the elastic-plastic state and a local plastic mechanism formed
accompanied by a sudden drop in load. The specimens (in 0.60 mm thickness)
with lower web slenderness (b/t) (from 33 to 50) developed roof-shaped
mechanisms. However, all the others developed the so-called flip-disc
mechanisms at about mid-length on one side as shown in Fig.4.9a. All of the
test columns are shown in Fig. 4.10.
The experimental local buckling loads were evaluated from the plots of load-
deflection and load-shortening. For the plots of load-deflection, the P-w2
method (Ventaramaiah and Roorda, 1982) was used to obtain the elastic local
buckling load. For the two series of pin-ended tests, the experimental local
buckling loads were obtained except for the 060LB20 and 060LB30 specimens
113
whose ultimate loads were lower than the elastic local buckling loads of the
section.
As shown in Figs. 4.9 (c) & (d) taken during the test, after the local buckling
load was reached, elastic local buckling can be observed on the concave side of
the columns. Due to the small eccentricity of the load, one side of the column
was in greater compression than the other and modulation of the local buckling
occurred on this side. Finally, interaction of local and overall buckling resulted
in failure of column.
It is noticeable that the column strength increases as the plate slenderness (b/t)
becomes larger and the overall slenderness becomes smaller for a given column
length as shown in Figs. 4.7 & 4.8. The column and plate slenderness values are
given in Table 4.6.
4.4 ANALYSES
4.4.1 Elastic loca l buckling analyses
The theoretical elastic local buckling loads (Nol) were obtained using the THIN-
WALL (Papangelis and Hancock, 1998). The average measured cross-section
dimensions of the specimens for each series as well as the measured values of
base metal thickness and Young’s modulus taken from the coupon tests were
used to determine the theoretical local buckling loads. The theoretical local
buckling mode is shown in Fig.4.11. The theoretical local buckling loads (Nol)
are given in Table 4.7 where the ratios of the experimental to theoretical local
buckling loads (Pcr/Nol) are computed.
114
The mean ratio of the experimental local buckling loads (Pcr) to the theoretical
local buckling loads (Nol) was approximately 1.06 and 0.95 respectively for the
0.60 m and 0.42 mm sections respectively. These results indicate that the THIN-
WALL program was in good agreement with the experimental local buckling
load as detailed in Table 4.7.
4.4.2 Test Resul ts compared with the Design Standards
4.4.2.1 General
The test results (Pt) are given in Tables 4.7. Figs. 4.12a, 12b & 12c show the test
strength (Pt) compared with a range of design curves for the 30 mm, 40 mm and
50 mm sections in 0.42 mm material respectively. Figs. 4.13a, 13b & 13c show
same for the 20 mm, 30 mm and 40 mm sections in 0.60 mm material
respectively. In Figs. 4.12 & 4.13, the test strengths (Pt) have been non-
dimensionalised with respect to the squash load (Py = Afy) as computed for the
measured yield stress and dimensions in Table 4.1.
Dashed and dash-dotted curves are plotted in Figs. 4.12 & 4.13. The dashed
curve is the ratio of NcRb/Py against the column length (lx). NcRb was calculated
using the AISI Specification based on a yield stress Rbfy as included in Section
A3.3.2 of the AISI Specification Supplement No.1(1999). The dash-dotted
curve is the ratio of Nc0.75/Py against column length (lx). Nc0.75 was calculated
based on AS/NZS 4600 with a yield stress of 0.75fy as included in Clause
1.5.1.5(b) of AS/NZS 4600.
The horizontal dashed line represents the ratio of the theoretical local buckling
load to the squash load (Nol /Py) against the column length (lx). The solid curve
is the ratio of Euler load to the squash load against the column length (lx).
115
Although each column in 0.60 mm thickness had 2 holes in the lips at both ends,
the effect of the holes on the ultimate load was not taken into account in the
calculation of the section or the member capacity.
4.4.2.2 Test results comparisons with design standards
It can be seen for all sections and lengths that when the elastic local buckling
loads (Nol) are generally lower than the test results (Pt), the test results are lower
than the design curves at intermediate columns lengths where local and Euler
buckling interact.
As shown in Figs. 4.12a&12b&12c, for the columns in 0.42 mm thickness, the

results are lower than the curves based on AS/NZS 4600 and the AISI
Specification. The low values for the 060LB20P450 specimen is probably due
to a large load eccentricity. The test was repeated and gave a higher result.
For the sections of stockier cross-section (060LB20 and 060LB30), the AS/NZS
4600 design curve based on 0.75fy and the AISI design curve based on a stress
Rbfy are generally slightly conservative.
The conclusion is that a modified design method is required for sections with
low local buckling loads as may occur in high strength steel columns with
slender plate elements.
4.5 SUMMARY
Pin-ended column tests with box-sections and constructed from high strength
G550 steel have been successfully performed. The plate slenderness (b/t) ranged
116
from 33 to 119 and the column slenderness (L/rx) ranged from 27 to 148. A load
eccentricity which produced a column response equivalent to L/1000 was used
for all tests.
• The columns with stockier plate elements, which had high local buckling
stresses (fol), failed by overall buckling. The test results were close to the
curves based on AS/NZS 4600 and the AISI Specification. For very long
columns with slender plate elements which had lower local buckling stresses
(fol), failure was still governed by overall buckling although local buckling
occurred.
• For the intermediate length columns, the failure mode was governed by the
interaction of local and overall buckling. The interaction of local and overall
buckling reduced the column strength and made the test results lower than
the design curves. The worst case had a difference between test results and
the results based on AS/NZS 4600 and the AISI Specification of about 14%,
which means that for the slender sections AS/NZS 4600 and the AISI
methods are unconservative.
117
4.6 FIGURES
Bolt in some sections,

clamps in others
Epoxy
Epoxy
Lipped-Box Section (LB)
Fig. 4.1 Tested Sections
t
h
r
b
Lipped-Box Section
Fig. 4.2 Cross-Section Dimensions
118
(a) Lipped-Box Section (LB) with Clamps (b) Lipped-Box Section (LB) with Bolts at ends
Fig. 4.3 Bolting & Clamping Configurations of

Test Specimens of LB section
119
Fig. 4.4 Rig for Imperfection Measurement
120
Fig. 4.5 Test Specimen in Test Rig
121
Cross head
P
Pin-ended bearing
End plate
Test machine centre line
Test long column

e
Location of transducers
& strain gauges at column mid-
Lt L height
Specimen centre line
End plate
Pin-ended bearing
Fixed base
Fig. 4.6 Test Arrangement
122
50
45 060LB40P450
40
35
30
Load (kN)
060LB30P450
25
20
15
10
5 060LB20P450
0
0 0.5 1 1.5 2
Axial Shortening (mm)
Axis Shortening (mm)
Fig. 4.7a. Load- Shortening of 060LB450 Series
35
30 060LB40P900
25
Load (kN)
20
060LB30P900
15
10
5
060LB20P900
0
0 0.5 1 1.5 2 2.5
Axis Shortening
Axial Shortening(mm)
(mm)
Fig. 4.7b. Load- Shortening of 060LB900 Series
123
25
042LB50P550
20
042LB40P550
15
Load (kN)
10
060LB40P450
042LB30P550
0 060LB30P450
0 0.5 1 1.5 2 2.5

AxisShortening
(mm)
Fig. 4.8a. Load- Shortening of 042LB550 Series
25
042LB40P1100
042LB50P1100
20
042LB50P1700
15
Load (kN)
042LB30P1100
10
5
042LB40P1700
0
0 0.5 1 1.5 2 2.5 3 3.5 4
Axis Shortening (mm)
Fig. 4.8b. Load- Shortening of 042LB1100&1700 Series
124
(a) Local buckle at failure (b) Overall buckle
125
(c) Local & Overall buckling (Bolts at ends) (d) Local & Overall buckling (Clamps)
Fig. 4.9 Buckling Mode (continued)
126
Fig. 4.10a Tested specimens of 060LB450&900 series
127
Fig. 4.10b Tested specimens of 042LB550&1100 series
128
Fig. 4.10c Tested specimens of 042LB1700 series
129
Fig. 4.11 Buckling Stress vs. Half-Wavelength for LB section
130
1 AS/NZS 4600 (Nc)
AISI, NcRb
0.75 Euler Curve (Noc)

Dimsionless Load
Local buckling Load (Nol)
0.5 T est Results (Pt)
0.25
0
0 500 1000 1500 2000 2500 3000
Length (mm)
Fig. 4.12a Comparison of Test Results with Design Standard

(0.42mm), 30x30 mm Sections
1 AS/NZS 4600 (Nc)
AISI, NcRb

Dimsionless Load
0.25
0
0 500 1000 1500 2000 2500 3000
Length (mm)
Fig. 4.12b Comparison of Test Results with Design Standard

131
1 AS/NZS 4600 (Nc)
AISI, NcRb
Euler Curve (Noc)

0.75
Dimsionless Load
T est Results (Pt)

0.5
0.25
0
0 500 1000 1500 2000 2500 3000
Length (mm)
Fig. 4.12c Comparison of Test Results with Design Standard

1.5
AS/NZS 4600 (Nc)

1.25
AISI, NcRb
1 Euler Curve (Noc)

Dimsionless Load
0.5
0.25
0
0 500 1000 1500 2000 2500 3000
Length (mm)
Fig. 4.13a Comparison of Test Results with Design Standard

132
1
AS/NZS 4600 (Nc)
AISI, NcRb
0.75
Euler Curve (Noc)
Dimsionless Load
0.25
0
0 500 1000 1500 2000 2500 3000
Length (mm)
Fig. 4.13b Comparison of Test Results with Design Standard

1 AS/NZS 4600 Nc)
AISI, NcRb

Dimsionless Load
0.25
0
0 500 1000 1500 2000 2500 3000
Length (mm)
Fig. 4.13c Comparison of Test Results with Design Standard

133
4.7 TABLES
Thickness
Specimen of Lip
060LB20P450a 10.5 22.3 0.60 0.65 0.60 0.60 6.0 1.5 448.5
060LB20P450b 10.3 20.6 0.60 0.65 0.60 0.60 5.8 1.6 449.0
060LB20P900a 10.7 21.4 0.60 0.65 0.60 0.60 5.2 1.9 896.5
060LB20P900b 11.5 20.5 0.60 0.65 0.60 0.60 5.9 1.6 896.0
060LB30P450a 15.5 32.0 0.60 0.65 0.60 0.60 8.2 1.7 448.8
060LB30P450b 15.1 31.4 0.60 0.65 0.60 0.60 8.1 1.4 449.0
060LB30P900a 15.8 30.8 0.60 0.65 0.60 0.60 8.7 1.5 898.3
060LB30P900b 15.2 30.8 0.60 0.65 0.60 0.60 8.5 1.6 897.5
060LB40P450a 19.8 41.9 0.60 0.65 0.60 0.60 10.8 1.6 448.8
060LB40P450b 20.0 41.9 0.60 0.65 0.60 0.60 10.7 1.5 448.6
060LB40P900a 19.8 41.5 0.60 0.65 0.60 0.60 11.4 1.7 897.5
060LB40P900b 19.8 41.2 0.60 0.65 0.60 0.60 11.1 1.6 898.5
Table 4.1a Measured Dimensions of Long Columns (0.60mm)
Thickness
Specimen of Lip
042LB30P550a 15.4 32.5 0.41 0.45 0.60 0.60 5.8 1.1 550.1
042LB30P550b 15.9 32.4 0.41 0.45 0.60 0.60 5.8 1.2 550.0
042LB30P1100a 15.7 32.7 0.41 0.45 0.60 0.60 6.1 1.1 1099.8
042LB30P1100b 15.8 32.6 0.41 0.45 0.60 0.60 5.9 1.2 1100.1
042LB40P550a 21.2 41.5 0.41 0.45 0.60 0.60 5.9 1.0 550.0
042LB40P550b 21.3 41.2 0.41 0.45 0.60 0.60 5.9 1.1 549.5
042LB40P1100a 21.3 41.4 0.41 0.45 0.60 0.60 6.1 1.4 1099.8
042LB40P1100b 21.2 41.6 0.41 0.45 0.60 0.60 6.0 1.2 1099.5
042LB40P1700a 20.4 41.2 0.41 0.45 0.60 0.60 6.0 1.3 1698.0
042LB40P1700b 20.5 41.4 0.41 0.45 0.60 0.60 5.9 1.1 1697.0
042LB50P550a 26.4 51.0 0.41 0.45 0.60 0.60 5.9 1.2 550.0
042LB50P550b 26.5 50.8 0.41 0.45 0.60 0.60 6.2 1.2 550.0
042LB50P1100a 26.5 50.3 0.41 0.45 0.60 0.60 6.6 1.5 1099.8
042LB50P1100b 26.5 50.7 0.41 0.45 0.60 0.60 6.2 1.3 1100.1
042LB50P1700a 25.3 52.0 0.41 0.45 0.60 0.60 6.2 1.2 1697.0
042LB50P1700b 25.4 51.7 0.41 0.45 0.60 0.60 6.0 1.2 1698.0
Table 4.1b Measured Dimensions of Long Columns (0.42mm)
134
Dimensions
Specimen Hole/Clamping 0.2% Proof Stress
Table
060LB20P450a~~060LB40P900b 1a 4 Holes at ends for each 711
042LB30P550a~~042LB50P1100b
1b Clamps 690
042LB40P1700a~~042LB50P1700b
Table 4.2 Hole/Clamping Configuration and Proof Stresses
Side1 Max Side2 Max

Specimen
(mm) (mm)
060LB20P450a -0.18 0.14
060LB20P450b 0.12 -0.15
060LB20P900a -0.51 -0.33
060LB20P900b -0.97 0.19
060LB30P450a 0.19 -0.25
060LB30P450b 0.06 -0.12
060LB30P900a -0.75 0.79
060LB30P900b -0.33 -0.46
060LB40P450a 0.19 -0.21
060LB40P450b 0.13 0.19
060LB40P900a 1.27 0.99
060LB40P900b 0.85 0.81
Table 4.3a Measured Imperfections of LB sections (Longitudinal)
Side1 Max (mm) Side2 Max (mm)

Specimen
Ed11 Ce10 Ed12 Ed21 Ce20 Ed22
042LB30P550a 0.448 0.19 -0.556 -0.435 -0.427 0.452
042LB30P550b -0.538 -0.376 -0.894 -0.4 -0.59 -0.399
042LB30P1100a 0.496 -0.75 -0.812 -0.75 -1.071 -1.55
042LB30P1100b 0.317 -1.04 -1.26 -0.441 -0.678 -0.37
042LB40P550a -0.589 -0.397 -0.367 -0.81 0.674 0.458
042LB40P550b 0.708 -0.482 -0.385 0.413 -0.596 -0.571
042LB40P1100a -0.768 -0.803 -0.35 -0.642 -0.498 0.53
042LB40P1100b 0.419 -0.47 0.606 -0.64 -0.91 -1.01
042LB40P1700a -0.38 -0.65 -0.98 -0.78 -1.41 -1.39
042LB40P1700b -1.59 -1.63 -1.36 -1.25 -1.58 -1.25
042LB50P550a 0.568 1.35 -0.543 0.562 -0.666 -0.329
042LB50P550b 0.453 0.414 -0.373 0.378 -1.052 -0.365
042LB50P1100a -0.443 -0.424 -0.723 -0.352 -0.555 0.54
042LB50P1100b -0.682 -0.443 -0.455 0.674 -0.907 0.408
042LB50P1700a -2.19 -2.5 -1.8 -1.17 -1.32 -0.82
042LB50P1700b -1.55 -1.63 -1.32 -1.25 -1.59 -1.56
Table 4.3b Measured Imperfections of LB sections (Longitudinal)
135
e (mm)
Specimen
Average
060LB20P450a 0.277
060LB20P450b 0.046
060LB20P900a 0.828
060LB20P900b 0.072
060LB30P450a 0.3960
060LB30P450b 0.049
060LB30P900a 0.143
060LB30P900b 0.773
060LB40P450a 0.486
060LB40P450b 0.071
060LB40P900a 0.632
060LB40P900b 0.062
Table 4.4a Measured Eccentricity of LB sections (0.60 mm)
e (mm)
Specimen
Average
042LB30P550a 0.122
042LB30P550b 0.072
042LB30P1100a 0.166
042LB30P1100b 0.099
042LB40P550a 0.609
042LB40P550b 0.429
042LB40P1100a 0.879
042LB40P1100b 0.764
042LB40P1700a 1.165
042LB40P1700b 0.9325
042LB50P550a 0.481
042LB50P550b 0.346
042LB50P1100a 0.263
042LB50P1100b 0.573
042LB50P1700a N/A
042LB50P1700b 1.450
Table 4.4b Measured Eccentricity of LB sections (0.42 mm)
136
E*I (m2N) E*I (m2N)

Specimen (E*I)m/(E*I)tw
from beam tests from Thin-Wall
060LB20P450a 587.4 552.6 1.06
060LB20P450b 663.0 665.6 1.00
060LB20P900a 847.4 895.5 0.95
060LB20P900b 810.2 762.5 1.06
060LB30P450a 2164 1960 1.10
060LB30P450b 2152 1963 1.10
060LB30P900a 2178 2155 1.01
060LB30P900b 2178 2180 1.00
060LB40P450a 5048 3791 1.33
060LB40P450b 5054 3744 1.35
060LB40P900a 4874 4607 1.06
060LB40P900b 4976 4568 1.09
Table 4.5 Measurement of E*I of LB Sections (0.60mm)
L/rx
Specimen b/t
Overall Slenderness
Plate Slenderness
L(mm) 550 1100 1700
042LB30 48 95 148 71
042LB40 35 70 108 95
042LB50 27 55 85 119
Table 4.6a Slenderness of Sections in 0.42 mm Thickness
L/rx
Specimen b/t
Overall Slenderness
Plate Slenderness
L(mm) 450 900
060LB20 63 126 33
060LB30 43 85 50
060LB40 32 63 67
Table 4.6b Slenderness of Sections in 0.60 mm Thickness
137
A fy Py Pt fol Pcr Nol

Specimen b/t Pcr/Nol
(mm2) (MPa) (kN) (kN) (MPa) (MPa) (kN)
060LB20P450a 61.0 33.3 711.0 43.4 16.4 954.6 N/A 58.2 N/A
060LB20P450b 60.4 33.3 711.0 43.0 21.3 954.6 N/A 57.7 N/A
060LB20P900a 61.9 33.3 711.0 44.0 8.1 954.6 N/A 59.1 N/A
060LB20P900b 62.1 33.3 711.0 44.2 N/A 954.6 N/A 59.3 N/A
060LB30P450a 90.8 50.0 711.0 64.5 N/A 411.7 N/A 37.4 N/A
060LB30P450b 90.4 50.0 711.0 64.3 35.5 411.7 N/A 37.2 N/A
060LB30P900a 91.4 50.0 711.0 65.0 21.9 411.7 N/A 37.6 N/A
060LB30P900b 90.3 50.0 711.0 64.2 20.3 411.7 N/A 37.2 N/A
060LB40P450a 120.8 66.7 711.0 85.9 47.6 225.6 31.6 27.3 1.16
060LB40P450b 120.7 66.7 711.0 85.8 45.5 225.6 32.6 27.2 1.20
060LB40P900a 120.9 66.7 711.0 86.0 28.7 225.6 24.4 27.3 0.89
060LB40P900b 120.5 66.7 711.0 85.7 30.1 225.6 27.3 27.2 1.00
Table 4.7a Summary of Results for LB Sections (0.60mm)
A fy Py Pt fol Pcr Nol

Specimen b/t Pcr/Nol
(mm2) (MPa) (kN) (kN) (MPa) (MPa) (kN)
042LB30P550a 60.1 71.4 690 41.5 17.5 183.6 10.1 11.0 0.91
042LB30P550b 60.5 71.4 690 41.8 17.4 183.3 9.7 11.1 0.87
042LB30P1100a 60.5 71.4 690 41.7 10.8 192.8 9.9 11.7 0.85
042LB30P1100b 60.4 71.4 690 41.7 11.7 182.3 10.4 11.0 0.94
042LB40P550a 79.9 95.2 690 55.1 20.9 108.0 9.2 8.6 1.07
042LB40P550b 77. 95.2 690 53.1 23.0 109.0 9.6 8.4 1.14
042LB40P1100a 76.9 95.2 690 53.1 15.0 108.7 8.8 8.4 1.05
042LB40P1100b 77.1 95.2 690 53.2 14.2 107.4 8.5 8.3 1.03
042LB40P1700a 75.8 95.2 690 52.3 8.4 116.3 6.2 8.8 0.70
042LB40P1700b 75.7 95.2 690 52.3 8.8 116.3 7.6 8.8 0.86
042LB50P550a 93.3 119.0 690 64.3 23.4 70.6 6.6 6.6 1.00
042LB50P550b 93.5 119.0 690 64.5 23.6 71.2 6.8 6.7 1.02
042LB50P1100a 93.8 119.0 690 64.7 19.7 72.2 7.2 6.8 1.06
042LB50P1100b 93.3 119.0 690 64.4 17.5 86.0 6.5 8.0 0.81
042LB50P1700a 92.1 119.0 690 63.6 11.2 74.0 6.5 6.8 0.96
042LB50P1700b 91.9 119.0 690 63.4 10.7 74.0 6.2 6.8 0.91
Table 4.7b Summary of Results for LB Sections (0.42mm)
Specimen 042LB30 042LB40 042LB50 060LB20 060LB30 060LB40

fy (MPa) 634 634 634 711 711 711
Ns(kN) 26.9 28.1 28.1 39.6 52.8 60.9
Pt (kN) 24.9 25.1 26.1 37.5 48.7 53.0
Table 4.7c Average Results for Stub Column of LB Sections

(0.42mm & 0.60mm)
138
Appendix 4.1
The calculation of eccentricity-load curves for LB-sections in 0.60 mm

thickness
P
A--A
P
M=P*
A A
ε1
ε2
Test long h/2
column
Fig App-4.1
Fig 4.6 shows the loading condition of an eccentrically loaded column. The Fig
App-4.1 shows the initial behaviour of the fibers of the cross-section of the
column at trial load P. ε1 and ε2 are strain obtained from two attached strain-
gauges.
The axial force (P) loaded eccentrically results in the extra bending moment
(M). So the stress in the column is the sum of axial and bending stress.
σ=P/A+M/(2*I/h) 4.1
139
ε=σ/E 4.2
From the strain diagram and Eqs. 4.1 & 4.2, the variation of the strain in two
sides of the cross-section (ε1-ε2) can be expressed by Eq. 4.3
ε1-ε2=P*h*e/(E*I) 4.3
Hence the distance between the loading and centroidal axes of the column (e)
can be expressed by Eq. 4.4
e=(ε1-ε2)*(E*I)/(P*h) 4.4
The value of the initial eccentricity (e) was used in the tests to adjust the
location of columns. The load-shortening (e - P) curves shown in this chapter
are based on data from well-adjusted tests which ensured the values of e within
the ranges of nominal values (L/1000).
140
Chapter 5: STRENGTH TESTS OF COLD-FORMED LIPPED CHANNEL SECTION COLUMNS
CHAPTER 5
STRENGTH TESTS OF COLD-FORMED LIPPED

CHANNEL SECTION COLUMNS
CONTENTS
5.1 INTRODUCTION....................................................................................................142
5.2 TEST SPECIMENS .................................................................................................142

5.2.1 General........................................................................................................................................... 142
5.2.2 The determination of the intermediate stiffeners in the flanges...................................................... 144
5.2.3 The determination of the length of the flange stiffeners ................................................................. 144
5.2.4 The determination of the length of the long specimens................................................................... 145
5.2.5 Labelling......................................................................................................................................... 146
5.3 COLUMN TESTS....................................................................................................147

5.3.1 General........................................................................................................................................... 147
5.3.2 Stub column tests ............................................................................................................................ 148
5.3.3 Long column tests ........................................................................................................................... 149
5.3.4 Test specimen behaviour ................................................................................................................ 150
5.3.5 The effect of buckling modes on ultimate load-capacity................................................................. 151
5.3.6 Failure modes analysis................................................................................................................... 153
5.4 TEST RESULTS COMPARED WITH DESIGN METHODS ...........................153

5.4.1 General........................................................................................................................................... 153
5.4.2 The Effective Width Method (EWM)............................................................................................... 154
5.4.3 AS/NZS 4600 Clause 3.4.6(b) and Kwon & Hancock Equation ..................................................... 155
5.4.4 The Direct Strength Method (DSM) ............................................................................................... 157
5.5 SUMMARY ..............................................................................................................160
5.6 FIGURES..................................................................................................................162
5.7 TABLES....................................................................................................................187
141
CHAPTER 5
STRENGTH TESTS OF COLD-FORMED LIPPED

CHANNEL SECTION COLUMNS
5.1 INTRODUCTION
Normally, one of three basic types of buckling, local, overall, and distortional,
can occur in thin-walled steel sections. However, the basic types may interact
with each other. For double symmetric box sections as shown in the previous
two chapters (3 & 4), local and overall buckling or interaction between them
rather than distortional buckling occurs. However, in practice, singly symmetric
open sections, such as C-sections, are commonly used in cold-formed design.
For sections of this type, especially for sections fabricated from such thin (less
than 1 mm) sheet steel, distortional buckling can occur and becomes a
significantly failure mode. The wavelength of distortional buckling is generally
intermediate between that of local buckling and overall buckling.
The aim of this chapter is to present the test results and structural behaviour of
0.42 mm G550 steel C-sections and to investigate local and distortional
buckling and the interaction between them.
5.2 TEST SPECIMENS
5.2.1 General
Compression tests were performed on channel sections with intermediate web
142
and flange stiffeners brake-pressed from aluminium/zinc-coated Grade G550

structural steel sheet to AS 1397 (1993) in 0.42 mm thickness. The section type
tested is shown in Fig. 5.1a. The lengths of the sections (designated SLC and
LC) ranged from 360 to 2000 mm. The cross-section dimensions of the
specimens are shown in Table 5.1, for the nomenclature shown in Fig. 5.1b. The
cross-section dimensions are the average of the measured values. The base
metal thickness (tb) was measured by removing the aluminium/zinc coating by
acid-etching. The ends of each specimen were milled flat and parallel to ensure
full contact between specimens and end bearing.
The length of the shortest column (360 mm), which was treated as a stub
column (SLC), was decided based on the SSRC guidelines (Galambos, 1988).
For cold-formed steel sections, which generally have thin-walled plate elements,
the stub column test is aimed at determining the effect of local buckling as well
as the effect of cold-forming, residual stress and yielding on the section capacity
in compression.
In order to determine the capacity of longer columns in compression, local and

distortional buckling and the interaction between them should be considered.
For the high strength steel columns assembled from thin plates, the section
configuration is an important factor affecting their behaviour. Thin plate
elements will generally continue to carry load after local buckling into the post-
buckling range so that local buckling does not mean failure of the whole
column. However, a singly-symmetric section may have a shift in the line of the
axial force after local buckling occurs resulting in an additional moment if it is
loaded between pinned ends. Moreover, long columns are prone to overall
buckling, which may complicate the problem. To eliminate these problems, in
this test program, fixed ends were used.
143
5.2.2 The determ ination of the intermediate stiffeners in the flanges
A finite strip analysis program (THIN-WALL) (Papangelis and Hancock, 1998)

was used to optimize the geometric shape to achieve local and distortional
buckling at similar stresses. Generally speaking, for most channel sections, the
sections’ performance may be enhanced significantly with a stiffener located in
the web. However, for the sections fabricated from very thin sheet steel (less
than 1 mm thick), a stiffener may be needed in each flange as well as the
stiffener in the web. Two graphs of buckling stress versus buckle half-
wavelength for two cross-sections in compression are shown in Figs. 5.2a &
5.2b, where the section in Fig. 5.2a is without stiffeners in the flanges whereas
the section in Fig. 5.2b has a stiffener in each flange. The local buckling minima
occur in the curves at half-wavelengths of 80 and 40 mm and the buckling
stresses of 20 and 73 MPa for the cross-sections without/with stiffeners in the
flanges respectively. The distortional buckling minima occur in the curves at
half-wavelengths of 1200 and 1300 mm and buckling stresses of 34 and 38 MPa
for the sections without/with stiffeners in the flanges respectively. For the
sections without stiffeners in the flanges, the local buckling stress is lower than
the distortional buckling stress. On the other hand, for the cross-section with
stiffeners in the flanges, the local buckling stress is higher than the distortional
buckling stress. To ensure that the local buckling stress is higher than the
distortional buckling stress, a stiffener was used in each flange in the tests, as
well as the one in the web.
5.2.3 The determ ination of the length of the flange stiffeners
As for the determination of the need for the flange stiffeners, a similar method
was used to determine the length of the lip. A series of sections with different
lip length of 3, 6, 9, 12, 15, and 18 mm, was analyzed using THIN-WALL.
144
Although for these sections the local buckling stresses were approximately 73
MPa, the distortional buckling stress increased with lip size. Lip lengths of 12
mm were adopted for this project and corresponded with a distortional buckling
stress of 38 MPa as shown in Table 5.2a. As demonstrated later, the effect of
fixed ends increased this value in the tests.
5.2.4 The determ ination of the length of the long specimens
The long columns tested mainly focused on the interaction of local and
distortional buckling. To determine the optimum length for testing, the accurate
determination of the distortional buckling stresses is a very important step. In
this chapter, two different methods of finite strip buckling analysis were used to
analyze the test sections so that the lengths of the long specimen could be
determined properly including end boundary conditions.
The solid line shown in Fig. 5.3 obtained using the semi-analytical finite strip
(SAFSM as described in Chapter 2 and implemented in THIN-WALL) presents
the buckling stress versus half-wavelength curve for the chosen test sections.
Two minimum points occur in the curve at half-wavelengths equal to 40 and
1300 mm corresponding to the local and distortional buckling mode with the
buckling stresses approximately 73 and 38 MPa respectively.
A spline finite strip method (SFSM) analysis was performed using BFPLATE
(as described in Chapter 2) on the stiffened C-section. The section was
discretises with 20 strips by 40 spline sections longitudinally as shown in Fig.
5.4a. The SFSM values and modes are given in Table 5.2b and the modes are
shown in Fig. 5.4b. As shown in Fig. 5.3, the local buckling stress (73 MPa)
computed using SAFSM for the test section was higher than the distortional
buckling stress (38 MPa) whereas the actual buckling stresses computed using
145
the SFSM taking account of the fixed-end boundary conditions were the same
as the local buckling stresses and higher than the distortional buckling stresses
computed using the SAFSM, for the lengths up to 1500 mm. However, as the
length further increases, the buckling stresses obtained using the SFSM become
closer to the distortional buckling stress obtained using the SAFSM. It can be
seen clearly that the fixed ends enhanced the buckling stress for the test section
particularly the distortional buckling stress. The length chosen for the longest
columns aimed to minimise the effect of the end boundary conditions on the
distortional buckling stress. A length of 2000 mm was adopted to fit the
maximum range of the test machine.
5.2.5 Labelling
The test specimens were labelled such that the type of section, nominal length
of specimen and specimen number were expressed by the label.
For example, the label “SLC360a” defines the following specimen:

• The first three numbers indicate that the specimen is a stub column lipped
section (alternative LC for long column lipped section)
• The “360” indicates that the nominal length of specimen was 360 mm
• The last letter “a” indicates that the specimen was the first tested (alternative
b, ca, cb, ra, rb and rc.)
• All specimens were 0.42 mm total coated thickness in G550 steel.
Geometric imperfections were measured for all of the specimens. The readings
were taken at 15 mm or 20 mm intervals along the specimen lengths. For each
side of a specimen, the measurements were taken along two lines in the
146
longitudinal direction of the specimen, along the centreline between the corner
and the stiffener as shown in Fig. 5.5a. The results are used later in the finite
element modelling analyses. The recording procedure and the description of the
measurement rig are detailed in Chapters 3 and 4. The maximum imperfections
are shown in Table 5.3 and were computed relative to a straight line between the
ends.
5.2.7 Material P roperties
The same material was used as in the stub column tests described in Chapter 3.
The 0.2% proof stress values are 690 MPa for all specimens.
5.3 COLUMN TESTS
5.3.1 General
A total of 9 stub columns and 12 long columns were tested. The specimen
lengths were 360 mm, 800 mm, 1300 mm, and 2000 mm. Test specimens in the
rig are shown in Figs. 5.6~5.8. The rig consisted of the Sintech/MTS-300kN
testing machine with fix-ended bearings. The bottom bearing was adjustable so
that it could orient specimens vertically before the bearing was locked for
testing. Specially designed end plates were filled with pattern-stone as shown in
Fig. 5.9 and were designed to guarantee a fixed end boundary condition. They
also ensured uniform compression and prevented the top and bottom edges from
localized damage which occurred in some of the stub column tests of the B and
HB sections detailed in Chapter 3. Load and shortening were recorded using the
Sintech data acquisition system. The compressive deformation rate was 0.05
mm/min.
147
The central deflections were measured using six transducers around the
specimen located at mid-height of the specimen, two for the web and one for lip
on each flange as shown in Fig. 5.5b. The transducers were connected to a
SPECTRA data acquisition system.
5.3.2 Stub colum n tests
The test results obtained for the 9 stub column specimens are given in Table
5.4a. The average ultimate test strength (Pt) was 37kN. The average load at
ultimate of the specimens was approximately 37% of the theoretical squash load
(99 kN) but approximately three times the theoretical local buckling load (10.5
kN). They indicate that these specimens had a very high post-local buckling
reserve strength. The failure modes are shown in Figs 5.10a & b and the full set
of load-shortening graphs in Fig. 5.11a.
The load-deflection graphs presented in Fig. 5.12a show that the deflection
behaviour of the flanges and web in the cross section were quite different. For
the web, lateral deflections started from a load equal to approximately 11 kN
which was slightly lower than the theoretical local buckling load obtained using
the SFSM. However, the tip deflection of the flanges did not change until the
ultimate load was approached. After that, the deflection in the web and flanges
increased quickly accompanied by a sudden drop in load (see Fig 5.11a) when
inelastic local buckling occurred in the lips as shown in Figs 5.10a & b.
Inelastic local buckles also occurred soon after in the web. This behaviour
indicates that the failure of these stub columns resulted mainly from local
buckling and significant distortional buckling only occurred when the ultimate
load for local buckling was approached.
148
5.3.3 Long colum n tests
In order to investigate the interaction between local and distortional buckling

and the member strength reduction resulting from this interaction, a series of
long specimens was tested.
The average ultimate test strengths were approximately 22, 19, and 18 kN
respectively for the 800, 1300, and 2000 mm length specimens as shown in
Table 5.4b. The average ultimate load of the specimens was approximately
22%, 19%, and 18% of the theoretical squash load for the 800, 1300 and 2000
mm lengths respectively but significantly higher than the theoretical local
buckling load (10.5 kN). These columns still have a very high post-local-
buckling reserve strength. It is interesting to note that the results of the 1300 and
2000 mm length columns are the same even though the lengths differ by 700
mm. The test strengths are also well above the theoretical distortional buckling
loads of 10.7 kN, 10.6 kN and 9.1 kN for the 800 mm, 1300 mm and 2000 mm
sections respectively.
A difference between tests SLC360ca&cb, LC1300a&b and all of the other

columns was the larger radius value of the corners at the junction of the web and
the flange as shown in Table 5.1. A comparison of the test results with the same
length specimens appears to indicate no significant difference.
The buckling modes for the 800 mm, 1300 mm and 2000 mm specimens are
shown in Figs. 5.6 & 5.7 & 5.8 respectively and the failure modes in Fig. 5.10c
for the 800 mm specimens, Fig. 5.10d for the 1300 mm specimens and Fig.
5.10e for the 2000 mm specimens. The load-shortening graphs are shown in Fig.
5.11b.
149
5.3.4 Test specim en behaviour
For the 360 mm columns, significant deflections occurred in the flanges as

shown in Fig. 5.12a after the ultimate load was approached. As shown in Figs.
5.10a & b, three types of flange buckling modes can be identified, which are
one flange moving inwards and another moving outwards (O-I mode); both
flanges moving inwards (I-I mode); and both flanges moving outwards (O-O
mode).
For the 800 mm columns, significant lateral deflections occurred in the flanges
at approximately 16 kN as shown in Fig. 5.12b(a). This is higher than the
theoretical elastic distortional buckling load of 10.7 kN based on the SFSM.
These deflections occurred when the lips developed inelastic local buckling as
shown in Fig. 5.6 or Fig. 5.10c for the flanges moving outwards (O-O mode) or
inwards (I-I mode) in the distortional mode.
For the 1300 mm columns, significant lateral deflections occurred in the flanges
from a low load for specimens 1300a&b&c&d as shown in Fig. 5.12b(b).
Specimens appeared to bifurcate in the distortional mode at approximately 12
kN as shown in Fig. 5.12b. For all specimens, there was a significant change in
axial stiffness at approximately 12 kN as shown in Fig. 5.11b. The theoretical
local and distortional buckling loads are quite close at approximately 10.6 kN
respectively for these specimens The eventual failure mode of all specimens
was local buckling in the web as shown in Fig. 5.11b. These deflections
occurred when the lips developed inelastic local buckling as shown in Fig. 5.7
or Fig. 5.10d for the flanges moving outwards (O-O mode) or inwards (I-I
mode) in the distortional mode. Significant interaction of local and distortional
buckling occurred.
150
As shown in Fig. 5.12b(c), flanges deflection of the 2000 mm long columns

occurred as soon as the columns were subjected to load. The load shortening
graphs of both specimens show a significant reduction in axial stiffness at about
10 kN. This can be compared with the theoretical distortional and local buckling
loads by the SFSM of 9.1 kN and 10.5 kN respectively. For the 2000a
specimen, the flanges lips moved inwards (I-I mode) as shown in Fig. 5.8a so
that final local buckling failure occurred in the web as shown in Fig. 5.10e. By
comparison, for the 2000b specimen, the flange lips moved outwards (O-O
mode) as shown in Fig. 5.8b with local buckling occurring in the flange lips
earlier than that in the web. The specimens clearly failed with interaction of
local and distortional buckling well into the post-buckling range of both modes.
The buckling modes are also shown in Fig. 5.4b which are 3D drawings based
on the data obtained using the SFSM. For the short column of 360 mm length
and the long column 2000 mm length, only the local or distortional buckling
mode appeared. However, for the 800 mm and 1300 mm lengths, the local and
distortional buckling modes appeared simultaneously. The buckling in the local
and distortional modes occurs for very close stress values, so that the
eigenvactor routine in the program cannot separate the modes and both appear
in the eigenmode at the same time.
5.3.5 The effect of buckling modes on ultimate load-capacity
Three kinds of buckling shape were observed during the stub column tests and
can also be identified from Figs. 5.10a & 5.10b. In some cases, the two flanges
both move outwards (O-O mode), both move inwards (I-I mode) and one moves
inwards while the other moves outwards (O-I mode). However, no clear
relationship can be established between the member capacity and the buckling
shape (or initial imperfection) for the stub columns. The main reason is that for
151
stub columns, the failure modes are dominated by the local buckling mode.
For the long column tests, the test results and buckling characteristics are
summarized in Table 5.5. For the 800 mm specimens, the LC800a&b sections
are in the O-O and O-I modes, whereas the LC800sa&sb sections are both in the
I-I mode. The peak loads of the LC800sa&sb specimens are significantly lower
than those of the LC800a&b specimens which failed in the O-O and O-I modes.
For the 1300 mm specimens, the failure modes of the LC1300sa&sb specimens
are the O-O mode. The peak loads are significantly higher than those of the
LC1300a&b&c&d specimens which failed in the I-I mode. For the 2000 mm
specimens, the influence of the buckling shape on the member capacity clearly
exists. The buckling mode of the LC2000b specimen is the O-O mode and had a
higher load than that of the LC2000a which failed in the I-I mode. Clearly, the
outward modes had a higher failure load than the corresponding inward modes.
As shown in the load-shortening graphs in Fig. 5.11b, the difference between

the O-O modes and I-I modes can also be observed for the 1300 and 2000 mm
lengths. For the 800 mm specimens, no special relationship can be established
between the buckling modes and stiffnesses using the slopes of the load-
shortening-deflection curves. However, for the 1300 mm long specimens, a
significant loss of the stiffness occurred after the elastic local buckling load was
approached even though in the beginning all specimens had a similar stiffness.
The LC1300a~d specimens (I-I mode) were considerably less stiff in axial
compression than the LC1300sa&sb specimens (O-O mode) and carried a lower
load (17.8 kN c.f. 21.3 kN). For the 2000 mm specimens, Fig. 5.11b shows that
the LC2000a specimen was considerably less stiff initially than the LC2000b
specimen and carried a lower load (16.8 kN c.f. 19.4 kN). Significant interaction
of local and distortional buckling occurred in the post-buckling range of both
modes.
152
It was also observed that the inward or outward initial imperfections may not
have a significant effect on inducing the specific buckling shape unless the
initial imperfections are very large such as for the LC1300sb or LC2000b
specimens (see Table 5.3b).
5.3.6 Failure mo des analysis
Two kinds of lip buckling mode shape were observed during the tests. For the
specimens which buckled inwards (I-I mode), large deformation occurred at the
top-and-bottom of the specimens. However, for specimens which buckled
outwards (O-O mode), large deformation occurred at the mid-height of the
specimens. For most specimens, which failed in the O-O mode, as shown in Fig.
5.11b, the load-shortening curves had small sudden drops before the ultimate
loads were approached as a result of the large deformations of the lips at mid-
height. For the 800 mm specimens which failed in the O-O modes, the first
large drop in the load resulted from failure of the stiffener in the flange but the
final big drop in the load was a result of the failure at the joint of the flange and
web. For the 800 mm specimens which failed in the I-I mode, the large drop in
the load resulted from the failure at the joint of the flange and web only. For the
1300 and 2000 mm specimens, the failure of the joint of flange and web led a
large drop in the load corresponding of the buckling modes.
5.4 TEST RESULTS COMPARED WITH DESIGN METHODS
5.4.1 General
Two methods can be used for predicting the member capacity: the Effective
Width Method (EWM) and the Direct Strength Method (DSM). The effective
153
width method is an elemental method since it looks at the elements forming a

cross section in isolation. It was first proposed by von Karman (1932) and
calibrated for cold-formed members by Winter (1947). It accounts for post-
buckling by using a reduced (effective) plate width at the design stress.
As sections become more complex with additional edge and intermediate

stiffeners, the computation of the effective widths becomes more complex.
Interaction between the elements also occurs so the method may be less
accurate for complex sections. To overcome these problems, a new method has
been developed by Schafer and Pekoz (1998), called the Direct Strength
Method. It uses elastic buckling solutions for the whole section rather than the
individual elements, and strength curves for the whole section.
In this Chapter, both methods are used to predict the member axial capacity
(Nc).
5.4.2 The Effect ive Width Method (EWM)
The new North America Specification (NAS) (AISI, 2001) was recently revised
to provide EWM design provisions for the determination of the effective widths
of uniformly compressed stiffened elements with multiple intermediate
stiffeners or edge stiffened elements with intermediate stiffeners. In this thesis,
the test sections adopted were C-sections with intermediate stiffeners in the web
and flanges and a simple edge stiffener on each flange. Therefore, the new NAS
(AISI, 2001) was used to determine the effective widths of the sections.
The section axial capacity (Ns) (Pn in NAS) was obtained using the EWM, and
gave a result of 34 kN which was slightly lower than the average ultimate test
strength 37 kN for the stub column tests. As shown in Fig. 5.13 the solid
154
horizontal line is Ns, the dash-dotted curve is the member capacity (Nc) (Pn in
NAS) plotted against the effective length (Le). The effective length Le has been
taken as half the column length to allow for the fixed ends. Nc was calculated
using the new NAS (AISI, 2001) based on the clauses for determination of the
long column strength of a locally buckled section. The test results (Pt) plotted
against the effective length (Le) are also shown in Fig. 5.13 identified using
symbols (■), (▲) and (♦) for buckling in O-O, O-I and I-I modes respectively.
As can be seen, for the long specimens, the test results were much lower than
the strengths based on the new NAS (AISI, 2001). There are two reasons for
these unconservative results. Firstly, distortional buckling is not adequately
accounted for by the NAS. Secondly, the interaction of local and distortional
buckling, which causes the additional dip in the test strength curves for the 800
mm and 1300 mm specimens, is also not adequately accounted for.
5.4.3 AS/NZS 46 00 Clause 3.4.6(b) and Kwon & Hancock Equation
Two alternative methods were used to calculated the member ultimate strength
as shown in Fig. 5.13.
The Winter formula based on the distortional buckling stress (fd) was modified
by Kwon and Hancock (1991) as given in the following formula to predict the
distortional buckling strength.
for λ ≤0.561
Pkh =Afy 5.1-1
for λ >0.561
155
0.6 
0.6
  f d    f d 
Pkh = Af y * 1−0.25  5.1-2
  f y   f y 
     
fy
where λ =
fd
As shown in Fig. 5.13, the line-dotted curve ( • ) is the member ultimate

strength (Pkh) against the effective length (Le) and is assumed invariant with Le
to produce a horizontal line. fd was obtained as 63 MPa (see Fig. 5.3) using the
SFSM at a length of 2000 mm and accounts for the fixed ends.
Clause 3.4.6 (b) in AS/NZS 4600 was also used to calculate the member
distortional strength. These equations are:
for fd > fy /2
Pc = Afy(1- fy /4 fd) 5.2-1
for fy /13 ≤ fd ≤ fy /2
 2 
 fy  
Pc = Af y  0.055 − 3.6  + 0.237  5.2-2
  fd  
   
The line-cross curve ( ∗ ) is the member ultimate strength (Pc) plotted against
the effective length (Le) and is assumed invariant with Le to produce a
horizontal line.
156
The distortional buckling strength (fd) used in these calculations was based on
the SFSM and therefore allows for the effect of the warping restrained end
conditions. As can be seen, the results based on AS/NZS 4600 Clause 3.4.6(b)
and Kwon & Hancock’s equation are approximately 21% and 12% higher than
the test results of 1300 mm section which buckled in the I-I mode and
approximately 32% and 22% higher than the test results of 2000 mm long
columns. The test results of the 1300 mm specimens which buckled in O-O
mode are close to Kwon & Hancock’s prediction. These results indicate that for
the long columns, the interaction of local and distortional buckling has a
significant effect on the member strength since it is not accounted for in the
Kwon & Hancock and AS/NZS 4600 equations. Hence, for the monosymmetric
sections of the type tested subject to distortional buckling, the interaction of
local and distortional buckling has to be taken into account when determining
the design member strength.
5.4.4 The Direct Strength Method (DSM)
The direct strength method (DSM) was proposed by Schafer and Pekoz (1998)
and summarized by Hancock, Murray and Ellifritt (2001) who also
demonstrated its applicability.
The DSM determines the strength for local and overall (L+E) interaction and
distortional and overall (D+E) interaction and takes the lesser of the two as the
strength. For the sections subjected to compression, these are the axial strengths
Pnl and Pnd. The equations for Pnl considering the L+E interaction are given by:
for λl ≤ 0.776
Pnl = Pne 5.3-1
157
for λl > 0.776
 0 .4  
0.4
P    Pcrl 
Pnl = Pne 1− 0.15 crl   P 
5.3-2
  Pne    ne 

Pne
where λl =
Pcrl
Pcrl is the elastic local buckling load; Pne is the compression member design strength Afn,,
where fn is based on the overall failure mode
The equations for Pnd considering the D+E interaction are given by:
for λd ≤ 0.561
Pnd = Pne 5.4-1
for λd > 0.561
 0.6  
0.6
P    Pcrd 
Pnd = Pne 1− 0.25  crd   P  5.4-2
  Pne    ne 

Pne
where λd =
Pcrd
Pcrd is the elastic distortional buckling load
A series of results was obtained using this method (DSM) based on the elastic
158
bucking stresses which was computed using the SFSM. A comparison of the
results with the test results is presented in Fig. 5.14.
In Fig. 5.14, the Pnd (+) and Pnl ( • ) represent the results using the DSM
based on the elastic buckling stresses from the SFSM. As can be seen, the
results of the DSM are in good agreement with the results of the tests which
buckle in the O-O mode (■) and O-I mode. (▲). The results of the tests which
buckle in I-I mode (♦) are clearly lower than Pnd. The results of Pnd and Pnl,
which were obtained using the elastic buckling stresses based on the SFSM
taking into account the fixed ends boundary conditions, conform with the test
results at 360 and 2000 mm lengths rather than at the intermediate lengths. As in
the preceding discussion, the failure modes at these lengths (360 and 2000 mm)
were mainly dominated by local buckling and distortional buckling respectively.
Hence, the lower test results for the 1300 mm length may be a result of the
significant interaction of L+D buckling. It can also be noted that even though
the length variation from 1300 to 2000 mm was large, the test results of 1300
mm length which buckle in I-I modes are quite close to the results of 2000 mm
length. This may be a consequence of the different failure mode. The failure
mode of the column at 2000 mm length was mainly distortional buckling,
whereas the failure of the column at the 1300 mm length was severe interaction
of local and distortional buckling. This interaction induced the premature failure
of the column at the 1300 mm length.
As shown in Fig. 5.14, the results based on the DSM are unconservative for the
columns at intermediate lengths, especially, for the members which buckle in I-I
mode.
159
5.5 SUMMARY
A series of channel section column tests with intermediate stiffeners in the web
and flanges and constructed from high strength G550 steel has been successfully
performed. The tests were carried out to investigate the effects of local and
distortional bucking as well as the interaction between them.
For the stub columns, local buckling is the dominant failure mode. However, for
the long specimens, interaction between local and distortional buckling led to
failure.
• The new NAS (AISI 2001) gives an accurate prediction of the stub column
strength using the new rules for an edge stiffened element with an
intermediate stiffener.
• The adverse interaction between local and distortional buckling reduced

significantly the member ultimate strength for the intermediate length
specimens fabricated from thin sheet steel.
• The effect of the buckling shape on the member capacity clearly exists. The
outwards buckling shape mode (O-O mode) has a much higher member
strength than the inwards buckling shape mode (I-I mode)
• The NAS (AISI, 2001) gives an unconservative prediction since it generally

ignores distortional bucking.
• The AS/NZS 4600 gives an unconservative prediction at intermediate

lengths because it ignores interaction of local and distortional buckling.
160
• More general theoretical models and design methods are required for
sections with adverse interaction of local and distortional buckling.
161
5.6 FIGURES
Fig. 5.1a End view
h/2 h/2
r
b/2
sw R
sd
b/2
t sd d
sw
Fig. 5.1b Nomenclature
Fig. 5.1 Test SLC & LC-Section
162
360
Stress (MPa) 240
160
Distortional Buckling
Local Buckling
80
0
10 100 1000 10000
Buckle Half-Wavelength
Fig. 5.2a Cross-section without stiffeners in the flanges (12 mm lips)
360
240 Local Buckling

Stress (MPa)
160 Distortional Buckling
80
10 100 1000 10000

Buckle Half-Wavelength
Fig 5.2b Cross-section with stiffeners in the flanges (adopted) (12 mm lips)
Fig. 5.2 Stiffeners adopted in the flanges
163
400
T W
Buckling Stress (MPa)
300 SFSM
200 L(1) D(1)

L(5) D(1)
100 D(1)
Local Minimum (73 MPa)

L(2) L(9)
0
1 10 100 1000 10000
Distortional Minimum (38 MPa)
Fig. 5.3 Finite Strip buckling Analysis (TW and SFSM)
164
Fixed end
Spline section h
h
h
Spline Strip
40 Sections
h
.
.
.
Nodal lines .
.
h
h
h
h
Fixed end
Spline Strip
Fig. 5.4a Spline Finite Strip analysis of Fixed end LC-section
165
Local mode Local & distortional modes

L=360 mm (fcr=73.0 MPa) L=800mm (fcr=73.3 MPa)
Local & distortional modes Mainly distortional mode

L=1300 mm (fcr=74.2 MPa) L=2000 mm (fcr=63.2 MPa)
Fig.5.4b 3D Drawing based on Spline Finite Strip Method
166
G2 G1
G3
G6
G4 G5
Fig. 5.5a Geometric Imperfection Measurement Positions
T3 T4
T5 T6
T1 T2
Fig. 5.5b Transducer Arrangement at Mid-Height
167
Fig. 5.6a 800 mm Long Test Specimen Under Load

in Test Rig (LC800a) (O--O mode)
168
Fig. 5.6b 800 mm Long Test Specimen Under Load

in Test Rig (LC800sa) (I--I mode)
169

in Test Rig (LC1300sa) (O—O mode)
170

in Test Rig (LC1300c) (I—I mode)
171

in Test Rig (LC2000b) (O—O mode)
172

in Test Rig (LC2000a) (I—I mode)
173
Fig. 5.9a End plate before insertion of specimen
Fig. 5.9b Specimen sitting in end plate

filled with pattern-stone
Fig. 5.9 End Plate
174
Fig. 5.10a Failure modes of SLC360 (a,b,c,d)
Fig. 5.10b Failure modes of SLC360 (ca,cb,ra,rb,rc)
175
Fig. 5.10c Failure modes of LC800 (a,b)
176
Fig. 5.10c Failure modes of LC800 (sa,sb) (continued)
177
Fig. 5.10d Failure modes of LC1300 (a,b,c,d)
178
Fig. 5.10d Failure modes of LC1300 (sa,sb) (continued)
179
Fig. 5.10e Failure modes of LC2000 (a,b)
180
45
40
35
30
Load (kN)
25
20
SLC360 Series
15
10
0
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0
AxisAxial Shortening (mm)

Shortening (mm)
Fig. 5.11a Load- Shortening of SLC360 Series
181
30
LC800a
O-O mode LC800sa&sb LC1300sa&sb
I-I mode O-O mode
25
LC800b
O-I mode
20
Load (kN)
15
10 LC2000b LC2000a
O-O mode I-I mode
5 LC1300a&b&c&d
I-I mode
0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5

Axis S hortening (mm)
Fig. 5.11b Buckling shape modes of LC Series
182
-45
-40
-35
-30
Load (kN)
-25
-20
-15
-10
-5
0
-1.0 0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0
Deflection (mm) (Flange)
(a) Load-deflection curve SLC360 (T1,T2)
-45
-40
-35
-30
Load (kN)
-25
-20
-15
-10
-5
0
1.0 0.0 -1.0 -2.0 -3.0 -4.0 -5.0 -6.0 -7.0
Deflection (mm) (Web)
(b) Load-deflection curve SLC360 (T3,T4)
Fig. 5.12a Typical Local-deflection curves of stub column
183
-25
-20
Load (kN) -15
-10
-5
0
0.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0

(a) Load-deflection curve LC800 (T1,T2)
-20
-18
-16
-14
Load (kN)
-12
-10
-8
-6
-4
-2
0
2 -8 -18 -28 -38 -48 -58 -68
Deflection (m m ) (Flange)
(b) Load-deflection curve LC1300 (T1,T2)
-25
-20
Load (kN)
-15
-10
-5
0
0 10 20 30 40 50 60

(c) Load-deflection curve LC2000 (T1,T2)
Fig. 5.12b Typical Local-deflection curves of long column
184
55
Member capacity (NAS, 2001) ( Nc) Test (O-O) (Pt)
50 Section Capacity (NAS, 2001) (Ns) AS/NZS 4600 Clause3.4.6(b) (Nc)
Kwon & Hancock Equation Test I-I (Pt)
45 Test (O-I) (Pt) Trendline (Test I-I (Pt))
Trendline (Test (O-O) (Pt))
40
35
Load (kN)
30
25
20
15
10
0
0 200 400 600 800 1000 1200 1400 1600 1800 2000
Effective Length (mm)
Fig. 5.13 Comparison of Test, NAS 2001 and AS/NZS 4600
185
45
Pcrd (2000mm-SFSM) Pcrl (360mm-SFSM)
Test (O-O) (Pt) Pnd
40
Pnl Test (I-I) (Pt)
Test (O-I) (Pt) Trendline (Test (I-I) (Pt))
Trendline (Test (O-O) (Pt))
35
30
Load (kN)
25
20
15
10
0
0 200 400 600 800 1000 1200 1400 1600 1800 2000
Fig. 5.14 Comparison of Test results and the results based on SFSM
186
5.7 TABLES
Web Flange Stiffener(w) Stiffener(F) Thickness Radius Lip Length

Specimen
b h sw sd sw sd tb tc r R d L
SLC360a 122.5 90.9 20.5 10.5 20.1 10.1 0.41 0.45 0.60 0.60 12.4 359.9
SLC360b 123.0 90.8 21.1 10.0 20.1 9.5 0.41 0.45 0.60 0.60 12.5 360.0
SLC360c 122.1 91.4 20.5 10.4 20.2 9.9 0.41 0.45 0.60 0.60 12.4 360.1
SLC360d 122.8 91.5 20.7 10.3 20.5 10.0 0.41 0.45 0.60 0.60 12.4 360.0
SLC360ca 123.2 91.0 20.8 10.2 19.6 10.3 0.41 0.45 0.60 1.05 12.8 361.0
SLC360cb 122.0 92.0 20.5 10.3 20.5 10.1 0.41 0.45 0.60 1.05 12.8 361.0
SLC360ra 122.2 91.2 20.3 10.3 20.5 10.3 0.41 0.45 0.60 0.60 12.4 360.1
SLC360rb 122.5 90.8 20.2 10.6 20.2 10.5 0.41 0.45 0.60 0.60 12.4 360.1
SLC360rc 122.7 91.4 20.3 10.2 20.1 10.2 0.41 0.45 0.60 0.60 12.3 360.1
Note: All dimensions in mm
Table 5.1a Measured Dimensions For SLC-Sections
Web Flange Stiffener(w) Stiffener(F) Thickness Radius Lip Length

Specimen
b h sw sd sw sd tb tc r R d L
LC800a 122.7 91.1 21.1 10.5 20.2 10.8 0.41 0.45 0.60 0.60 12.6 801.0
LC800b 122.2 91.6 21.1 11.3 20.5 10.0 0.41 0.45 0.60 0.60 12.5 801.0
LC800sa 122.6 91.0 20.9 10.0 21.1 10.1 0.41 0.45 0.60 0.60 12.6 800.3
LC800sb 122.0 91.3 20.8 9.8 21.1 10.1 0.41 0.45 0.60 0.60 12.7 800.0
LC1300a 121.4 92.5 21.1 10.5 21.2 9.5 0.41 0.45 0.60 1.05 12.6 1299.6
LC1300b 122.8 90.6 21.6 9.9 20.5 10.3 0.41 0.45 0.60 1.05 12.8 1297.5
LC1300c 120.4 91.5 20.3 10.3 21.1 10.6 0.41 0.45 0.60 0.60 12.7 1300.0
LC1300d 121.2 91.7 20.8 10.1 21.1 9.8 0.41 0.45 0.60 0.60 12.1 1300.0
LC1300sa 122.5 91.1 21.1 10.1 20.9 9.4 0.41 0.45 0.60 0.60 12.8 1300.5
LC1300sb 122.8 91.4 19.9 9.9 20.0 9.6 0.41 0.45 0.60 0.60 12.9 1300.0
LC2000a 121.1 90.7 21.6 9.9 21.2 10.1 0.41 0.45 0.60 0.60 12.9 2000.0
LC2000b 121.5 92.0 21.1 9.3 22.0 11.8 0.41 0.45 0.60 0.60 13.0 2001.0
Note: All dimensions in mm
Table 5.1b Measured Dimensions For LC-Sections
187
Lip length Local buckling stress Distortional buckling stress

d fl fd
mm MPa MPa
3 73.2 29.2
6 73.2 29.4
9 73.2 32.9
12 73.2 38.0
15 73.2 44.3
18 73.2 50.6
Table 5.2a Determination of Lip Length Using THIN-WALL
Test SAFSM SFSM

(Average) (THIN-WALL) (BFPLATE)
Distortion
Elastic Local Elastic
Specimen Ultimate al
Buckling Failure Buckling Buckling Buckling
Stress Buckling
Stress Mode Stress Stress Mode
(MPa) Stress
(MPa) (MPa) (MPa)
(MPa)
SLC360 258.7 81.1 L(9) 73.0 L(9)
D(1)+L(20
LC800 166.1 74.4 D(1)+L(20) 73.3
)
73.2 38.1
LC1300 126.2 / D(1)+L(32) 74.2 D(1)
LC2000 128.2 / D(1)+L(46) 63.2 D(1)
L--- Local buckling mode

D--- Distortional buckling mode
( )--- Number of Buckling half-waves
Table 5.2b Buckling Stresses and Buckling Modes
188
Web Max
Flange1 Max (mm) Flange2 Max (mm)
(mm)
Specimen
G1 G2 G3 G4 G5 G6
SLC360a 0.37 -0.45 -0.60 -0.34 -0.22 0.30

SLC360b 0.22 -0.25 0.29 0.25 0.37 0.28
SLC360c 0.22 0.27 -0.21 0.20 -0.42 0.37
SLC360d -0.6 -0.34 0.29 0.25 0.27 0.37
SLC360ca -0.34 -0.19 0.24 -0.55 -0.76 -0.52
SLC360cb -0.45 -0.86 -0.32 0.33 -0.42 -0.49
SLC360ra 0.22 -0.28 -0.56 -0.33 -0.32 0.53
SLC360rb -0.29 0.35 -0.32 0.22 -0.74 -0.33
SLC360rc -0.36 0.40 -0.48 0.43 -0.42 -.23
Note: Sign (-) indicates the inward initial imperfection
Table 5.3a Measured Imperfections of SLC-Sections(Longitudinal)
Web Max
Flange1 Max (mm) Flange2 Max (mm)
(mm)
Specimen
G1 G2 G3 G4 G5 G6
LC800a -0.16 -0.28 -0.21 0.50 -0.39 0.27

LC800b 0.80 0.64 0.30 0.25 -0.69 -0.34
LC800sa 0.67 -0.69 0.61 -0.47 -1.36 -0.03
LC800sb 0.35 0.08 0.46 -0.67 -0.51 0.39
LC1300a 0.54 -0.30 -1.84 -0.32 1.22 -0.45

LC1300b -0.93 -0.93 -0.75 -0.75 -0.88 -1.25
LC1300c -0.43 -0.28 -1.61 -3.90 -2.18 -0.65
LC1300sa 0.36 -0.48 0.40 3.40 0.66 -1.09
LC1300sb -0.48 -0.13 2.14 -1.02 5.02 2.25
LC1300d -1.21 -1.21 -0.53 -3.95 -2.30 -0.60
LC2000a 1.73 1.84 -3.48 0.93 -5.62 -3.00

LC2000b -1.65 1.56 -1.96 5.72 8.58 3.24
Note: Sign (-) indicates the inward initial imperfection
Table 5.3b Measured Imperfections of LC-Sections(Longitudinal)
189
Specimen SLC360a SLC360b SLC360c SLC360d SLC360ca SLC360cb SLC360ra SLC360rb SLC360rc
2
A(mm ) 144.6 142.5 144.3 144.3 145.2 145.0 144.6 145.5 144.8
fy (MPa) 690 690 690 690 690 690 690 690 690
Pt (kN) 40.2 37.5 35.6 33.9 36.5 36.6 40.7 37.0 38.5
Pt /A(MPa) 284.7 265.6 252.1 240.1 258.5 259.2 288.2 262.0 272.7
Ns (kN) 34.1 33.7 34.0 33.9 34.1 34.0 34.1 34.3 34.0
Pt / Ns 1.2 1.1 1.0 1.0 1.1 1.1 1.2 1.1 1.1
Nl (kN) 10.6 10.4 10.5 10.5 10.6 10.6 10.6 10.6 10.6
Pcr (kN) 12.5 11.5 10.0 11.5 11.3 11.9 11.6 12.2 13.0
Table 5.4a Summary of Results for SLC-Sections
Specimen LC800a LC800b LC800sa LC800sb LC1300a LC1300b LC1300c LC1300d

A(mm2) 145.8 145.7 143.2 143.0 143.5 143.2 144.5 142.5
fy (MPa) 690 690 690 690 690 690 690 690
Pt (kN) 22.8 24.1 21.5 21.5 18.4 17.3 18.4 17.2
Pt /A(MPa) 161.5 170.7 150.1 150.3 130.3 122.5 130.3 121.8
Nc (kN) 33.0 32.9 32.5 32.4 30.5 30.6 30.9 30.5
Pt / Nc 0.69 0.73 0.66 0.66 0.60 0.57 0.60 0.56
Nl (kN) 10.6 10.6 10.6 10.6 10.5 10.5 10.5 10.4
Nd (kN) 10.7 10.7 10.7 10.7 10.6 10.6 10.7 10.6
Specimen LC1300sa LC1300sb LC2000a LC2000b
A(mm2) 142.3 144.1 141.9 144.9
fy (MPa) 690 690 690 690
Pt (kN) 21.7 20.9 16.8 19.4
Pt /A(MPa) 152.4 145.0 119.0 137.4
Nc (kN) 30.2 30.3 26.7 26.8
Pt / Nc 0.72 0.69 0.63 0.72
Nl (kN) 10.5 10.5 10.4 10.6
Nd (kN) 10.6 10.6 9.0 9.2
Table 5.4b Summary of Results for LC-Sections
190
Specimen LC800a LC800b LC800sa LC800sb LC1300a LC1300b LC1300c LC1300d
Pt (kN) 22.8 24.1 21.5 21.5 18.4 17.3 18.4 17.2
Buckling Outwards Outwards Inwards Inwards Inwards Inwards Inwards Inwards

Characteristic (O-O) (O-I) (I-I) (I-I) (I-I) (I-I) (I-I) (I-I)
Specimen LC1300sa LC1300sb LC2000a LC2000b
Pt (kN) 21.7 20.9 16.8 19.4
Buckling Outwards Outwards Inwards Outwards

Characteristic (O-O) (O-O) (I-I) (O-O)
Table 5.5 Summary of Results for LC-Sections (Long)
191
Chapter 6: NUMERICAL SIMULATIONS
CHAPTER 6
NUMERICAL SIMULATIONS
CONTENTS
6.1 GENERAL DESCRIPTION OF ABAQUS...........................................................193

6.1.1 Element type ................................................................................................................................... 193
6.1.2 Material behaviour......................................................................................................................... 194
6.1.3 Boundary conditions....................................................................................................................... 195
6.1.4 Geometrical imperfection ............................................................................................................... 197
6.1.5 Eigenvalue buckling prediction and post-buckling analysis........................................................... 199
6.1.6 Solution Control Parameters.......................................................................................................... 200
6.2 SIMULATION OF STUB COLUMNS..................................................................201

6.2.1 General........................................................................................................................................... 201
6.2.2 Element type ................................................................................................................................... 201
6.2.3 ABAQUS Results ............................................................................................................................ 202
6.2.4 ABAQUS results compared with Test results and Design Standards ............................................. 203
6.2.4.1 General.................................................................................................................................................... 203
6.2.4.2 B-sections................................................................................................................................................ 204
6.2.4.3 LB-sections ............................................................................................................................................. 204
6.2.4.4 HB-sections............................................................................................................................................. 205
6.2.4.5 Explanation of some results .................................................................................................................... 205
6.3 SIMULATION OF LONG COLUMNS.................................................................206

6.3.1 General........................................................................................................................................... 206
6.3.2 Element type and Eccentricity Introduced...................................................................................... 207
6.3.3 ABAQUS Results ............................................................................................................................ 208
6.4 SIMULATION OF LIPPED CHANNEL SECTIONS.........................................209

6.4.1 General........................................................................................................................................... 209
6.4.2 Element type ................................................................................................................................... 209
6.4.3 ABAQUS results ............................................................................................................................. 210
6.4.4 Load-axial shortening curves ......................................................................................................... 210
6.4.5 Deformed shape and stress distribution ......................................................................................... 211
6.5 SUMMARY ..............................................................................................................213
6.6 FIGURES..................................................................................................................216
6.7 TABLES....................................................................................................................216
192
CHAPTER 6
6.1 GENERAL DESCRIPTION OF ABAQUS
The finite element method has recently been used successfully to predict the
behaviour of cold-formed members. In general, it is not a simple exercise to
produce a reliable finite element model since geometric imperfections, residual
stresses and end boundary conditions play a large role in the structural
behaviour. In fact, it can be argued that an accurate finite element simulation
may take longer to set up and execute than a carefully controlled test. Hence the
experimental data becomes very important in calibrating and implementing the
finite element analysis. Once this has been performed, the finite element
analysis can be used to extend the range of test data, and to investigate the effect
of changing variables, such as stress-strain characteristics, residual stresses,
geometric imperfections and section geometry. This chapter presents a finite
element analysis of the post-buckling behaviour of thin-walled compression
members. A commercial finite element program ABAQUS 5.7 is used to
simulate the behaviour of columns in this project.
6.1.1 Element ty pe
ABAQUS includes general-purpose shell elements as well as elements that are

valid for thick and thin shell problems. The general-purpose shell elements
provide robust and accurate solutions to most applications and will be used for
most applications. Element types S3/S3R, S4R, S4 SAX1 and SAX2 are general-
purpose shells. Thick shells are needed in cases where transverse shear
flexibility is important and second-order interpolation is desired. When a shell is
193
made of the same material throughout its thickness, this occurs when the
thickness is more than about 1/15 of a characteristic length on the surface of the
shell. Thin shells are needed in cases where transverse shear flexibility is
negligible and the Kirchhoff constraint must be satisfied accurately. For
homogeneous shells this occurs when the thickness is less than about 1/15 of a
characteristic length on the surface of the shell.
Element type S4R is a general-purpose, finite-membrane-strain, reduced

integration shell element. Reduced integration usually provides more accurate
results and significantly reduces running time, especially in three dimensions. In
this thesis, the element type S4R is used.
6.1.2 Material b ehaviour
Most materials of engineering interest initially respond elastically. Elastic

behaviour means that the deformation is fully recoverable. If the load exceeds
some limit, the deformation is no longer fully recoverable. Some part of the
deformation will remain when the load is removed. Plasticity theories model the
material’s mechanical response as it undergoes such nonrecoverable
deformation in a ductile fashion. Most of the plasticity models in ABAQUS are
“incremental” theories in which the mechanical strain rate is decomposed into
an elastic part and a plastic part. ABAQUS also has a “deformation” plasticity
model, in which the stress is defined from the total mechanical strain.
Incremental plasticity models are usually formulated in terms of a yield surface,
a flow rule, and evolution laws that define the hardening.
Most materials exhibit ductile behaviour and yield at stress levels that are orders
of magnitude less than the elastic modulus of the material, which implies that
the relevant stress and strain measures are “true” stress (Cauchy stress) and
194
logarithmic strain. Material data for all of these models should, therefore, be
given in these measures.
If the nominal stress-strain data for a uniaxial test is obtained and the material is
isotropic, a simple conversion to true stress and logarithmic plastic strain as
shown in Fig. 6.1 is
σ true = σ nom (1 + ε nom ) 6.2.1
ε pl
= ln(1 + ε nom ) − σ true E 6.2.2
where E is the Young’s modulus; εpl is true plastic strain; σnom , εnom are nominal stress and
strain.
When performing an elastic-plastic analysis at finite strains, ABAQUS assumes

that the plastic strains dominate the deformation and that the elastic strain is
small. From the load-deflection graphs of column tests such as shown in Fig.
5.11a, the “incremental” assumption can be adopted for the G550 sheet steel.
The elastic response can be modeled accurately as being linear. The plastic
strain values were used for the post-buckling analysis. The data used in the
ABAQUS model was obtained from the stress-strain curves of the coupon tests
in tension. The second point on the stress-strain curve corresponded with the
onset of plasticity. A large number of points was used around the corner on the
stress-strain curve in order to simulate the true behaviour of the material. The
data used is shown in Table 6.1.
6.1.3 Boundary conditions
A rigid body is created by defining rigid elements that all share a common rigid
195
body reference node. The reference node has both translational and rotational
degrees of freedom and acts as a master node to the nodes on the rigid elements.
The rigid element nodes are slave nodes: loads can be applied to the nodes of
rigid elements and the nodes can be connected to other elements (rigid or
deformable). Deformable elements can be connected to a rigid body by using
the same nodes to define both the deformable and rigid elements. This
capability provides a convenient alternative to certain kinds of multi-point
constraints.
For each of the two ends, two different types of boundary conditions were used
to simulate the test situation in the stub column tests. The ends were divided
into an immovable end and a movable end, which was the loaded end.
Initially, load was exerted directly on the movable end with an even stress
distribution. Secondly, for the purpose of comparing with the first method and
to eliminate the concern of the top edge being damaged for such thin sheet steel,
a displaced rigid body was used. The top end of the column was linked to the
rigid body. The results obtained from the two methods were not found to be
significantly different, so to simulate the behaviour of the stub columns either
method can be used.
Multi-point constraints (MPCs) specify linear or nonlinear constraints between

nodes. These relations between nodes can either be the default types that are
provided or can be coded in the form of a user subroutine. MPC type PIN
provides a pinned rigid link between two nodes to keep the distance between the
two nodes constant. The displacements of the first node are modified to enforce
this constraint. The rotations at the nodes, if they exist, are not involved in this
constraint. MPC type LINK provides a pinned joint between two nodes. This
MPC makes the displacements equal but leaves the rotations, if they exist,
196
independent of each other.
In order to join the two parts in the B & LB tests, lips had to be used. To
simulate the connection between the lips, three different ways were tried.
Firstly, the lips were treated entirely as a double thickness. The second and third
methods were LINK and PIN in the Multi-Point-Constraints method of
ABAQUS. The difference between the results from the different methods was
found to be less than 2%. However for the small size columns, some
convergence problems resulted from the Link method.
6.1.4 Geometric al imperfection
Imperfections are usually introduced by perturbations in the geometry. Chou,

etc. (2000) presented a buckling analysis using finite element method and
pointed out that to obtain the ultimate loads of the section which undergo
buckling, the structure must have initial geometric imperfections to trigger
deformation which can be done by either modelling the structure with an initial
out-of-plane deflection or by using small transverse forces. ABAQUS offers
three ways to define an imperfection: as a linear superposition of buckling
eigenmodes, from the displacements of a *STATIC analysis, or by specifying
the node number and imperfection values directly on the data lines. Only the
translational degree of freedom are modified. The usual approach involves two
analysis runs with the same model definition: (a) In the first analysis run,
perform an eigenvalue buckling analysis on the “perfect” structure to establish
probable collapse modes and to verify that the mesh discretizes those modes
accurately. (b) In the second analysis run, introduce an imperfection in the
geometry by adding these buckling modes to the “perfect” geometry. The
lowest buckling modes are assumed to provide the most critical imperfections,
so usually these are scaled and added to the perfect geometry to create the
197
perturbed mesh. (c) Finally, perform a geometrically nonlinear load-

displacement analysis of the structure containing the imperfection using the
Riks method. In this way the Riks method can be used to perform post-buckling
analyses of “stiff” structures that show linear behaviour prior to buckling, if
perfect. By performing a load-displacement analysis, other important nonlinear
effects, such as material inelasticity or contact, can be included.
The response of some structures depends strongly on the imperfections in the

original geometry, particularly if the buckling modes interact after buckling
occurs. Hence, imperfections based on a single buckling mode may yield
nonconservative results. By adjusting the magnitude of the scaling factors of the
various buckling modes, the imperfection sensitivity of the structure can be
assessed. Normally, a number of analyses should be conducted to investigate the
sensitivity of a structure to imperfections. Mateus and Witz (2001) presented
the investigation of the sensitivity of the buckling and post-buckling behaviour
of imperfect steel plates using ABAQUS.
In this thesis, the degree of initial imperfection was specified as the maximum
amplitude of the buckling mode shape and was usually prescribed as a
percentage of the thickness of sheet steel. The expression suggested by Walker
(1975) for the degree of imperfection is 0.3(Py /Pcr)1/2t, where Py is the yield
load and Pcr is the critical buckling load, and t is the thickness of sheet steel.
Comparing the values measured with results based on the expression suggested
by Walker (1975), the difference was very small. So the results based on the
expression suggested by Walker (1975) were used in the finite element
analyses.
In this chapter, two different ways were used to introduce the geometrical
imperfection into the post-buckling analysis. The first was to use the initial out-
198
of-plane deflection at mid-length of the column based on the imperfection

measurement, which can be found in Chapters 3&4&5. The second was to
introduce the imperfection based on an eigenvalue buckling analysis with the
expression suggested by Walker (1975) for the amplitude. The two ways were
not significantly different for the stub columns except for the stockiest B-
sections.
6.1.5 Eigenvalue buckling prediction and post-buckling analysis
Eigenvalue buckling is generally used to estimate the critical bucking loads of

stiff structures. Stiff structures carry their design loads primarily by axial or
membrane action, rather than by bending action. Their response usually
involves very little deformation prior to buckling. However, even when the
response of a structure is nonlinear before collapse, a general eigenvalue
buckling analysis can provide useful estimates of collapse mode shapes. An
incremental loading pattern is defined in *BUCKLE step. The important thing is
to find the load multipliers in the eigenvalue problem. Normally, the lowest
value of those multipliers is of interest.
In some cases linear eigenvalue analysis may be sufficient for design

evaluation; but if there is concern about material nonlinearity, geometric
nonlinearity prior to buckling, or unstable post-buckling response, a load-
deflection analysis must be performed to investigate the problem further. As
discussed in previous chapters, the sections made of the high strength thin steel
have very high post-buckling reserve. The Riks method uses the load magnitude
as an additional unknown; it solves simultaneously for loads and displacements.
Therefore, another quantity must be used to measure the progress of the
solution; ABAQUS uses the “arc length” along the static equilibrium path in
load-displacement space. This approach provides solutions regardless of
199
whether the response is stable or unstable.
Riks (1972, 1979) proposed an incremental approach to deal with the buckling
and snapping problems. The Riks method works well in snap-though
problems—those in which the equilibrium path in load-displacement space is
smooth and does not branch. The Riks method can also be used to solve post-
buckling problems, both with stable and unstable post-buckling behaviour such
as that shown in Fig. 6.2. However, the exact post-buckling problem can not be
analyzed directly due to the discontnuous response at the point of buckling. To
analyse a post-buckling problem, it must be turned into a problem with
continuous response instead of bifurcation. This effect can be accomplished by
introducing an initial imperfection into a “perfect” geometry so that there is
some response in the buckling mode before the critical load is reached. Herein,
the *STATIC, RIKS procedure was used to perform the collapse or post-
buckling analysis.
6.1.6 Solution C ontrol Parameters
The convergence and integration accuracy algorithms are very important factors
in the finite element analysis, especially for simulation of sections made of thin
steel with high post-buckling reserve. There are many control parameters
associated with them in ABAQUS. Normally, these parameters are assigned
default values which are chosen to optimize the accuracy an efficiency of the
solution for a wide spectrum of nonlinear problems. However, large sections
and sections with large longitudinal dimensions have a convergence problem in
the analysis. To solve this problem, the control parameters have to be adjusted
in the *CONTROL option within a step definition. The main parameters which
are adjusted in using ABAQUS to analyze the test sections are I0 for specifying
the equilibrium iteration for a residual check and IR for specifying the
200
equilibrium iteration for a logarithmic rate of convergence check.
Both I0 and IR are adjusted to 7 and 10 respectively in the analyses in this thesis.
With these values, convergent solutions can be obtained. In some cases, the
adjustment of the parameters in the *STATIC, RIKS option can avoid
divergence.
6.2 SIMULATION OF STUB COLUMNS
6.2.1 General
The aim of the simulation of the stub columns is to predict the ultimate strength.
Different contact conditions, geometric and material nonlinearities are
considered in the finite element modeling adopted in this section.
6.2.2 Element ty pe
The shell element, type S4R, was used. The ratio of the length to width of the
element is about 2:1.
For the whole column, different mesh densities were adopted. In the
longitudinal direction of the column, the nodes were concentrated towards the
middle of the column so that a fine mesh was obtained around the center. In the
transverse direction, the finer mesh is used at the corners based on the concept
of effective area as shown in Fig.6.3. The mesh density for ABAQUS models of
B, LB and HB sections is shown in Figs. 6.4a & 6.5a & 6.6a.
The model consisted of 18960 nodes and 18720 elements. The mean mesh size
201
varied from 0.35 mm by 0.19 mm for the stockiest section to 2.5 mm by 1.4 mm
for the largest section.
6.2.3 ABAQUS Results
The three tested sections, which are the B, LB, and HB-sections, are simulated
using ABAQUS. The ABAQUS results are given in Tables 6.2~6.5. The results
are also plotted in Fig. 6.7 along with the test results.
As shown in Figs. 6.7a & 6.7b and detailed in Tables 6.2~6.5, the differences
between the results of ABAQUS and tests for the stockiest sections were larger
compared with those of the other sizes. From Tables 6.2~6.5, it can be seen that
the yield stress of the material (fy) and the local buckling stress (fol) were about
equal for these sections. The analysis of equilibrium indicates that the
bifurcation of equilibrium will occur when the critical load is reached. It means
that when the yield stress is about equal to the critical buckling stress, the
structure will be very sensitive to initial imperfections. Although the values
adopted in ABAQUS were based on the measurements and Walker’s suggested
expression, the differences between the results of ABAQUS and the tests were
still quite large for the stocky sections. The main reason for the differences may
be very difficult to determine accurately due to the difficulty in determining the
true initial imperfection in the sections. However, it has been proven by further
study using ABAQUS modelling that the larger the initial imperfection, the
smaller the results become.
For the stockiest B-sections, the ABAQUS results were very sensitive to the
imperfections as shown in Fig. 6.7a for the 30 mm sections.
In order to clarify the degree of imperfection sensitivity, a further investigation
202
is performed. Two steps are involved in setting up the analyses. The first step
was an eigenvalue buckling analysis performed on the "perfect" column to
establish probable collapse modes. The second step was creating an
imperfection in the geometry by adding these modes to the "perfect" geometry.
The lowest buckling modes are assumed to provide the most critical
imperfections, so usually these are scaled and added to the perfect geometry to
create the perturbed mesh. The magnitudes of the perturbations used are a
percentage of the sheet thickness.
The input scale factor varied from 0.05, 0.2 to 0.5 for the stockiest B-sections in
0.42 mm thickness. The differences in the computed strengths were
approximately 10% and 20% respectively. For the stockiest B-section in 0.60
mm thickness, the differences were approximately 3% corresponding to the
input scale factor 0.05 and 0.5. For other sections in both thicknesses, the
differences were less than 1%. So for the stockiest B-section, the thinner the
sheet used, the more sensitive they became to the imperfections.
6.2.4 ABAQUS results compared with Test results and Design Standards
6.2.4.1 General
The buckling shapes using ABAQUS of B, LB & HB section are shown in Figs
6.4b, 6.5b and 6.6b respectively.
The test results (Pt) and ABAQUS (AB) results are given in Tables 6.2~6.5. The
test results (Pt) and ABAQUS (AB) results non-dimensionalised with respect to
the theoretical stub column strength (Ns) are plotted against the plate
slenderness ratios (b/t) in Figs. 6.7a & 6.7b. The theoretical stub column
strengths (Ns) were calculated according to AS/NZS 4600 (1996). The
203
theoretical strengths were calculated using the average measured cross-section

dimensions and the measured material properties as detailed in Chapter 3.
6.2.4.2 B-sections
The ratios of the ABAQUS strengths (AB to Ns) for the B-sections in 0.42 mm
and 0.60 mm plotted against the plate slenderness (b/t) are also shown in Figs.
6.7a & 6.7b expressed using symbols (∆). For the B-sections in 0.42 mm
thickness, the ABAQUS results are in good agreement with the test results.
However, for the B-sections in 0.60 mm thickness, the ABAQUS results were
about 5% higher than the test results obtained from sections with holes in their
lips and slightly lower than the test results obtained for sections without holes.
The ABAQUS analysis did not model the holes but the area of the holes was
deducted.
6.2.4.3 LB-sections
Only three different sizes were tested for the LB-sections in 0.42 mm and 0.60
mm thickness respectively.
The ratios of the ABAQUS strengths (AB to Ns) for the LB-sections in 0.42 mm
and 0.60 mm plotted against the plate slenderness (b/t) are shown in Figs. 6.7a
& 6.7b expressed using symbols (□). The ABAQUS results for the LB-sections
in 0.42 mm thickness are a little bit higher than the test results by about 2% and
5% for the big and small sizes respectively. For the LB-sections in 0.60 mm
thickness, the ABAQUS results were close to the test results except for some of
the results of the small sizes for which ABAQUS was somewhat higher as
discussed earlier in Section 5.2.5.
204
6.2.4.4 HB-sections
HB-sections were tested in the 0.60 mm thickness only. The ratios of the test
results (Pt) to the theoretical stub column strengths (Ns) of the HB-sections in
0.60 mm thickness plotted against the plate slenderness (b/t) are shown in Fig.
6.7a identified using symbols (●). The results show a similar trend to the B&LB
sections but are lower at high slenderness with the last test value quite low. It
appears that the Teflon may not have fully restrained the sheet at high
slenderness. The ratios of the ABAQUS strengths (AB to Ns) of the HB-sections
0.60 mm plotted against the plate slenderness (b/t) are also shown in Fig. 6.7a
identified using the symbols (◊). As can be seen, the larger the section, the
higher the ultimate load became except for the largest section. For the last three
sections, the differences of results between ABAQUS and tests were 5%, 5%
and 12% respectively. The high slenderness test results seem to be faulty as
discussed above.
6.2.4.5 Explanation of some results
In Figs 6.7a & 6.7b, it can be noted that for the B-sections, the test results
decreased slightly when the ratio of plate slenderness (b/t) reached about 133. It
can also be seen that for the B-sections in 0.60 mm thickness, the results of
ABAQUS deviate from the test results by less than 2% for the slender sections
and 5% for the stockier sections. The predicted results overestimate the strength
of B-sections by 3% except for the stockier sections, whereas for the B-sections
in 0.42 mm thickness, the ABAQUS results are in good agreement with the test
results.
For the LB40a&b and LB40na&nb sections in 0.60 mm thickness, the common
characteristic was a larger lip of about 10 mm. From Table 6.4a, it can be seen
205
that even though LB40a and LB40b had different configurations on the lips, the
test results (Pt) and the ABAQUS results (AB) were the same respectively but
the theoretical column strengths (Ns) were significantly different. The
differences in the calculation or simulation were that LB40a used the full length
of lip and LB40b adopted a length of 3 mm removed from the length of lips to
account for the holes. Also the same method was used for the LB40na&nb
sections. A further comparison of ABAQUS results (AB) with theoretical
results (Ns) indicates that the results based on AS/NZS 4600 were significantly
larger than those of ABAQUS. It is interesting to note that the ABAQUS results
with the section using the full length of lip and with the section using a length of
3 mm removed from the lip are the same. It means that the actual effective
length of the lip of the LB40 section with a 10 mm length of lip was smaller
than the effective length calculated based on AS/NZS 4600. The latter results in
an overestimate of the theoretical strength and may be the main reason for the
decrease in the ratio Pt/Ns. Research on the intermediate stiffener also indicates
that an increase in the stiffener cross-sectional area contributes to increase at
most of a few percent in the ultimate strength. The results of the LB sections in
0.42 mm thickness did not have a significantly different ultimate strength from
the B sections, which used the normal length of lips.
6.3 SIMULATION OF LONG COLUMNS
6.3.1 General
The aim of the simulation of the long column is to predict their capacity in
compression and to determine the effects of imperfection and eccentricity. The
ABAQUS model adopted is similar to the model used in Section 6.2 since it has
the same materials (in 0.60 and 0.42 mm thicknesses) and the same (LB)
206
section. Therefore, here, only the differences from the model used in Section 6.2
are presented.
6.3.2 Element ty pe and Eccentricity Introduced
The shell element, type S4R, was used. The ratio of length to width of element
varied from about 2:1 to 6:1.
For the whole column, different mesh densities were adopted. In the
longitudinal direction of each column, the nodes are concentrated towards the
middle of the column so that the fine mesh can be obtained around the centre. In
the transverse direction, the finer mesh was used at the corners on the basis of
the concept of effective area as shown in Fig.6.8a.
The model consisted of 12848 nodes and 12212 elements. The mean mesh size
varied from 4.6 mm by 2.3 mm for the stockiest section to 14.2. mm by 2.3 mm
for the largest section.
In order to simulate the pin-ended bearings, rigid bodies were used. The rigid
bodies were created by defining rigid elements (R3D4) that all share two rigid
body reference nodes. The top rigid body had both longitudinal translational and
rotational degrees of freedom and the bottom had only a rotational degree of
freedom. The lengths of the rigid bodies were 75 mm and 54 mm for the
columns in 0.60 mm thickness and in 0.42 mm thickness respectively. The rigid
surfaces were also defined to connect with the column. The reference nodes
associating with each rigid body were adopted to realize the rotation in the weak
axis direction of the cross-section as shown in Fig. 6.8b. Eccentric loading was
introduced. The values of eccentricity used are the experimental values as
shown in Table 4.4. The load was applied through the top reference point.
207
6.3.3 ABAQUS Results
The ABAQUS results (AB) ultimate loads are given in Table 6.6 where the
ratios AB/Pt are computed. The average difference between the ABAQUS
results and test results is approximately 6%. It can be seen in Table 6.6 that for
most specimens the ABAQUS results are slightly higher than the test results.
The differences including variability are most likely related to assumed
imperfection in the ABAQUS model. Further investigation of the imperfection
to use in ABAQUS is required.
6.3.4 ABAQUS results compared with Test Results and Design Standards
Figs. 6.9a & 6.9b & 6.9c show the ABAQUS values (AB) compared with a
range of design curves for the 30 mm, 40 mm and 50 mm sections in 0.42 mm
material respectively. Figs. 6.10a & 6.10b & 6.10c show the same for the 20
mm, 30 mm and 40 mm sections in 0.60 mm material respectively. In Figs. 6.9
& 6.10, the ABAQUS values (AB) have been non-dimensionalised with respect
to the squash load (Py) as computed for the measured yield stress and
dimensions in Chapter 4.
The ABAQUS results show a similar trend to the test results when compared
with the design curves although, as detailed above, they are slightly higher
probably due to assumptions for the imperfections.
Fig. 6.11 obtained using ABAQUS/post shows the buckled shape. Fig 6.11a
shows the local buckling on compression side. Fig. 6.11b shows the
combination of local and overall buckling. All the buckling shapes are the same
as those of tests which are shown in Chapter 4. The interaction of local and
overall buckling is clearly shown in Fig. 6.11.
208
For the largest 50 mm section in 0.42 mm material, two more long columns
were simulated using ABAQUS in order to obtain a complete graph of variation
with length and to make up the shortage where no test could be performed on
such long columns. The lengths were 2400 mm and 3000 mm. The dimensions
of cross-section used are the same as 042LB50P1700 as detailed in Chapter 4.
The ABAQUS results are 7.27 kN and 5.42 kN for 2400 mm and 3000 mm
respectively. It can be seen in Fig. 6.9c that the results are 7% and 4% lower
than the curves based on AS/NZS 4600 and AISI Specification. The longer the
column, the less the difference becomes for a given size of cross-section.
6.4 SIMULATION OF LIPPED CHANNEL SECTIONS
6.4.1 General
Much effort has been made to simulate the lipped channel section because of its
complicated buckling modes. The material used, the way to introduce geometric
imperfections into the ABAQUS model and the boundary conditions adopted
are the same as that of previous chapters.
6.4.2 Element ty pe
The shell element, type S4R, was used. The ratio of width to length of element
varied from about 0.9:1 to 1.6:1.
The model consisted of elements varying from 5500 to 11000 . The mean mesh
size varied from 3.86 mm by 3 mm for the stub column to 7.4. mm by 8.3 mm
for the largest column. The model is shown in Fig. 6.12a. The element division
and elements number over the cross-section is shown in Fig, 6.12b
209
6.4.3 ABAQUS results
The ABAQUS ultimate loads are given in Table 6.7 where the ratios AB/Pt are
computed. The average difference between the ABAQUS results and test results
are less than 4% for both the stub and long columns.
6.4.4 Load-axial shortening curves
The load-shortening curves from ABAQUS for the stub and long columns are
shown in Fig. 6.13.
For the 360 mm stub columns as shown in Fig. 6.13, the axial stiffness does not
change until the load approaches 20 kN which is much higher than the
theoretical elastic local buckling load. It indicates that the loss of axial stiffness
results from local buckling increasing as shown in Fig. 5.11a.
For the 800 mm columns, before the peak load is reached a small drop occurs at
about 18 kN which corresponds to the lip failure in O-I mode as shown in Fig.
5.11b. After that, the load keeps increasing until a significant sharp drop occurs
as shown in Fig. 6.13 which is similar to the test curve. For the 1300 mm
columns, the significant turning point of the slope is at about 10 kN as shown in
Fig. 6.13. However, for the 2000 mm columns, the test curve as shown in Fig.
5.11b is without an obvious turning point while the ABAQUS curve occurs at
about 8 kN.
The difference between the O-O mode and I-I mode can be seen clearly in the
load-shortening curves for the 1300 and 2000 mm columns as shown in Fig.
6.13. For the 1300 and 2000 mm columns in O-O mode, the axial stiffness is
much higher than that in I-I mode for the whole loading process, which is the
210
same as the test curves of the 2000 mm columns as shown in Fig. 5.11b but
different from the test curves of the 1300 mm columns. For the test curves of
the 1300 mm columns, only after the theoretical local buckling load was
reached, the axial stiffness of the columns in the O-O mode became
significantly higher than that in the I-I mode.
6.4.5 Deformed shape and stress distribution
Fig. 6.14 shows the deformed shapes and stress (von Mises stress) distributions
of stub and long columns. The buckling shapes are the same as those of the test.
The local buckling mode, the distortional buckling mode and the combination of
local and distortional bucking mode can be identified from Fig. 6.14. Also it can
be seen that high stress concentration occurs in the lips and the joints of web-
flanges at mid-length of the columns. This is in good agreement with the failure
mode observed during the tests.
In order to get an overview of the stress distribution on the whole cross-section,

several graphs are plotted at different loading steps for the different length
locations. The data obtained from the ABAQUS output file is used for this
purpose. For each length column, four graphs are plotted as shown in Fig. 6.15,
two reflect the stress distribution before and after the elastic local buckling load
whereas another two reflect the stress distribution before and after the peak
load.
The stress distribution changing on whole cross-section can be seen clearly from
Figs. 6.15a-1&2, b-1&2 for 360 and 800 mm length columns. Before the elastic
local buckling load is reached, the stress is higher at the central area of each
portion; however after the local buckling load is reached, the stress concentrates
at the corners on each portions. However, this phenomenon is not clearly shown
211
on Fig. 6.15c-1&2, d-1&2 for 1300 and 2000 mm length columns, it may result
from the distortional buckling occurring early on the flanges.
Figs. 6.15a,b,c,d-3&4 show the stress distributions changing before and after the
peak load. The highest stress concentration areas shown clearly on all figures
are the mid-height lips and the joints of the web and flanges. The areas with the
stress concentrations will yield earlier than other area. With the load increasing,
the yielding areas are getting larger. Especially, the corner areas of the joints of
flanges and web at the mid-height will lose totally their stiffness as the yielding
area becomes large enough. Those stress concentrations will finally result in the
failure of columns in those areas. It may give an explanation of the phenomenon
(failure mode) observed during the tests. Also it may explain the sharp drop on
the load-deflection curves in Chapter 5.
6.4.6 ABAQUS results compared with Test Results and Design Standards
The ABAQUS results (◊, ∆) are plotted in Figs. 6.16 & 6.17 and compared with
test results and the design standards.
From the figures, it can be seen, the ABAQUS results are in good agreement to
the test results for all columns. Interestingly, as shown in Figs. 6.16 or 6.17 the
ABAQUS results in the O-O mode and I-I mode are lower than the test results
in the O-O mode and I-I mode. This may be due to the reduced integration
which lowers the shear stiffness of this shell element and which may affect the
predictions of the distortional buckling mode. The ABAQUS results are
consistent with the test results, which are that the results in the O-O buckling
mode are higher than the results in the I-I buckling mode as shown in Table
6.7b.
212
The description of the comparison of the ABAQUS results and the design
standards and other methods is similar to the description of the test results
detailed in Chapter 5.
6.5 SUMMARY
A series of ABAQUS simulations was carried out on Box-Shaped Stub

Columns, Box-Shaped Long Columns, and Lipped-Channel Section Columns.
1. Fix-ended stub columns with box-shaped sections are simulated, the

following conclusions can be made.
• The expression suggested by Walker (1975) used to scale initial

imperfections in an ABAQUS model is generally successful. Based on Py
and Pcr of such thin sheet steel sections, this expression can provide an initial
imperfection magnitude suitable for the simulation of the test results.
• The use of the finite element program ABAQUS for simulating the
behaviour of the stub columns is successful since the ABAQUS results were
generally in good agreement with experimental values. The results of
ABAQUS are sensitive to the initial imperfections for the stockier sections
but not for the slender sections. ABAQUS can be used for further work on
such thin sheet steel sections.
2. Pin-ended long column tests with box-shaped sections have been

successfully simulated. The ABAQUS results are compared with the test
results and the design standards. ABAQUS simulations of the test results
with local imperfections and overall eccentricity are made.
213
• The ABAQUS results (AB) are generally in good agreement with the test
results (Pt). The difference is on average less than 6% for all columns
although the ABAQUS results were higher. More detailed investigation of
imperfections to use in ABAQUS is required.
stresses (fol), failed by overall buckling. The test results and the ABAQUS
results are close to the curves based on AS/NZS 4600 and the AISI
Specification.
• Since ABAQUS results are in reasonably good agreement with the test
results, ABAQUS can be used to simulate long column tests which could not
be performed in the available machine.
3. Fix-ended stub and long column tests with lipped channel sections are
simulated. The buckling modes and stress distribution are discussed.
• The ABAQUS results (AB) are generally in good agreement with the test
results (Pt). The difference is for the stub and long columns being less than
4%
• The comparison of the failure modes based on ABAQUS plots and the test
photos shows the reliability of the ABAQUS results. The stress distribution
shows the changing of stress over the whole cross-section during the loading
and gives an explanation of the failure modes which are identical with the
observation during the tests.
• For all columns, the simulation gives a reasonable prediction results. The
214
differences between ABAQUS results and Test results are less than 4%.
• In the future, based on the foregoing study of ABAQUS model of lipped

channel section, more work will focus on the difference between different
flange distortional buckling modes and its effect on the member capacity
because of the findings in the experimental tests. The effect of imperfection
introduced on the deformed shape requires further study.
4. General Summary
• A series of closed and open section columns is simulated using ABAQUS

5.7. The main attention has been paid to the choice of the mesh density, the
adoption of the initial imperfections and the determination of the plastic
stress-strain data.
• The ABAQUS analyses give reasonable results. In Section 6.3, ABAQUS

gives a good prediction for the 2400 and 3000 mm B-section columns which
shows the expected trend in Fig. 6.9c.
• The buckling shape and the stress distributions are in good agreement to
those observed experimentally.
• Finite element analysis can be used to predict the ultimate load of thin-
walled members, analyze the post-buckling behaviour of thin-walled
sections, and also helps to design and optimize shapes of thin-walled section.
215
6.6 FIGURES
True stress
εpl εel True strain
where: εel is true elastic strain; εpl is true plastic strain; ε is total true strain
Fig. 6.1 ABAQUS Stress-Strain Data Input Approach

(ABAQUS 5.7)
Load
Displacement
Fig. 6.2 Unstable Response (ABAQUS 5.7)
216
be/2
be/2
where be is effective width
Fig. 6.3a Effective Width of B-Section
be/2 be/2
h’e/2
h’
h’e/2
where be ,h’e are effective width
Fig. 6.3b Effective Width of LB-Section
217
Fig. 6.4a ABAQUS Mesh Density of B- Stub Columns
Fig. 6.4b ABAQUS Deformed Shape of B- Stub Columns
218
Fig. 6.5a ABAQUS Mesh Density of LB- Stub Columns
Fig. 6.5b ABAQUS Deformed Shape of LB- Stub columns
219
Fig. 6.6a ABAQUS Mesh Density of HB- Stub Columns
Fig. 6.6b ABAQUS Deformed Shape of HB- Stub Columns
220
1.10
60Ba/Ns 60Bb/Ns 60Bra&rb/Ns
60HB/Ns 60BWH/Ns 60LB(all)/Ns
60AB(B)/Ns 60AB(LB)/Ns 60AB(HB)/Ns
1.05 AISI/Ns AS4600/Ns
1.00
0.95
Pt/Ns, AB/Ns
0.90
0.85
0.80
0.75
20 40 60 80 100 120 140 160 180
b/t
Fig. 6.7a Comparison of B&LB&HB-section Test and ABAQUS

Results with Design Standard (0.60mm)
1.1
42Ba/Ns 42Bb/Ns 42Bc/Ns

1.05 42AB(B)/Ns 42LB/Ns 42AB(LB)/Ns
AISI/Ns AS4600/Ns
0.95
Pt/Ns, AB/Ns
0.9
0.85
0.8
0.75
20 40 60 80 100 120 140 160 180
b/t
Fig. 6.7b Comparison of B & LB-section Test and ABAQUS

Results with Design Standard (0.42mm)
221
Fig. 6.8a ABAQUS Mesh Density of LB- Long Columns (mid-length portion)
Fig. 6.8b ABAQUS Rigid Body and Reference Point of LB- Long Columns
(top portion)
222
1 AS/NZS 4600 (Nc)
AISI, NcRb
Euler Curve (Noc)

0.75
Dimsionless Load
T est Results (Pt)
0.5 ABAQUS Results (AB)
0.25
0
0 500 1000 1500 2000 2500 3000
Length (mm)
Fig. 6.9a Comparison of Test and ABAQUS Results with Design Standard
1 AS/NZS 4600 (Nc)
AISI, NcRb
Euler Curve (Noc)

0.75
Dimsionless Load
T est Results (Pt)
0.25
0
0 500 1000 1500 2000 2500 3000
Length (mm)
Fig. 6.9b Comparison of Test and ABAQUS Results with Design Standard
223
1 AS/NZS 4600 (Nc)
AISI, NcRb
Euler Curve (Noc)

0.75
Dimsionless Load
T est Results (Pt)
ABAQUS Results (AB)

0.5
0.25
0
0 500 1000 1500 2000 2500 3000
Length (mm)
Fig. 6.9c Comparison of Test and ABAQUS Results with Design Standard
1.5
AS/NZS 4600 (Nc)

1.25
AISI, NcRb
Euler Curve (Noc)

1
Dimsionless Load
T est Results (Pt)

0.75
ABAQUS Results (AB)
0.5
0.25
0
0 500 1000 1500 2000 2500 3000
Length (mm)
Fig. 6.10a Comparison of Test and ABAQUS Results with Design Standard
224
1
AS/NZS 4600 (Nc)
AISI, NcRb
Euler Curve (Noc)

0.75
Dimsionless Load
T est Results (Pt)
0.25
0
0 500 1000 1500 2000 2500 3000
Length (mm)
Fig. 6.10b Comparison of Test and ABAQUS Results with Design Standard
1 AS/NZS 4600 Nc)
AISI, NcRb
Euler Curve (Noc)

0.75
Dimsionless Load
T est Results (Pt)
0.25
0
0 500 1000 1500 2000 2500 3000
Length (mm)
Fig. 6.10c Comparison of Test and ABAQUS Results with Design Standard
225
Fig. 6.11a ABAQUS Local Deformed Shape of LB-Long column
226
Fig. 6.11b ABAQUS Local and Overall Deformed Shape

of LB-Long column
227
Fig. 6.12a Mesh Density of SLC&LC Sections
Fig. 6.12b Element Number over the Cross-Section
228
40
35
SLC360
30
LC800 (O—O)
25
Load (kN)
LC800 (I—I)
20
15
10
LC1300 (I--I)
LC2000 (I--I)
5 LC2000 (O--O)
LC1300 (O--O)
0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5
Axis Shortening
Axial Shortening(mm)
(mm)
Fig. 6.13 Load vs Shortening Curve of ABAQUS
229
Chapter 6:
230
Fig. 6.14a ABAQUS Deformed Shape and Stress Distribution of SLC360

Chapter 6:
231
Fig. 6.14b ABAQUS Deformed Shape and Stress Distribution of LC800

(O—O mode)
Chapter 6:
SECTION POINT 1
MISES VALUE
+1.93E+01
+7.21E+01
+1.25E+02
+1.78E+02
+2.30E+02
+2.83E+02
+3.36E+02
+3.89E+02
+4.41E+02
+4.94E+02
+5.47E+02
+6.00E+02
+6.52E+02
+7.05E+02
232
3
2 DISPLACEMENT MAGNIFICATION FACTOR = 1.00
RESTART FILE = lc800-22i-i STEP 1 INCREMENT 130
1 TIME COMPLETED IN THIS STEP 11.6 TOTAL ACCUMULATED TIME 11.6
ABAQUS VERSION: 5.7-1 DATE: 20-MAY-2003 TIME: 15:46:17
Fig. 6.14b ABAQUS Deformed Shape and Stress Distribution of LC800

(I—I mode)
Chapter 6:
233
Fig. 6.14c ABAQUS Deformed Shape and Stress Distribution of LC1300

(O—O mode)
Chapter 6:
SECTION POINT 1
MISES VALUE
+1.07E+01
+4.14E+01
+7.21E+01
+1.03E+02
+1.34E+02
+1.64E+02
+1.95E+02
+2.26E+02
+2.57E+02
+2.87E+02
+3.18E+02
+3.49E+02
+3.79E+02
+4.10E+02
234
3
2 DISPLACEMENT MAGNIFICATION FACTOR = 1.00
RESTART FILE = lc1300-22i-i32 STEP 1 INCREMENT 2040
1
TIME COMPLETED IN THIS STEP 202. TOTAL ACCUMULATED TIME 202.
ABAQUS VERSION: 5.7-1 DATE: 23-MAY-2003 TIME: 12:18:42
Fig. 6.14c ABAQUS Deformed Shape and Stress Distribution of LC1300

(I—I mode)
Fig. 6.14d ABAQUS Deformed Shape and Stress Distribution of LC2000

(O—O mode)
235
Fig. 6.14d ABAQUS Deformed Shape and Stress Distribution of LC2000

(I—I mode)
236
LC360 P=8.9kN
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
1
2
3
4
5
6
7
8
9
-61.8
-62
-62.2
-62.4
-62.6
-62.8
L/4
-63 L/2
Fig. 6.15a-1 Stress Distribution over the whole cross-section of SLC360

based on ABAQUS (note small stress range)
LC360 P=19.3kN
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
1
2
3
4
5
6
7
8
9
-40
-60
-80
-100
-120
-140
-160
L/4
-180 L/2

based on ABAQUS
237
LC360 P=32.5kN Before Peak-Load
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
1
2
3
4
5
6
7
8
9
0
-50
-100
-150
-200
-250
-300
-350
L/4
-400 L/2

based on ABAQUS
LC360 P=32.7kN After Peak-Load

10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
1
2
3
4
5
6
7
8
9
50
0
-50
-100
-150
-200
-250
-300
-350
-400
-450 L/4
-500 L/2

based on ABAQUS
238
LC800 P=10.7kN
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
1
2
3
4
5
6
7
8
9
-74.4
-74.5
-74.6
-74.7
-74.8
-74.9
-75
-75.1
L/8
-75.2 L/4
-75.3 L/2
Fig. 6.15b-1 Stress Distribution over the whole cross-section of LC800

LC800 P=17.7kN
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
1
2
3
4
5
6
7
8
9
-60
-80
-100
-120
-140
-160 L/8
L/4
-180 L/2

based on ABAQUS
239
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
1
2
3
4
5
6
7
8
9
400
300
200
100
-100
-200
-300
L/8
-400
L/4
-500
L/2

based on ABAQUS

10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
1
2
3
4
5
6
7
8
9
400
300
200
100
0
-100
-200
-300
-400
L/8
-500
L/4
-600 L/2

based on ABAQUS
240
LC1300 P=10.9kN
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
1
2
3
4
5
6
7
8
9
-60
-65
-70
-75
-80
-85
-90
L/12
-95 L/4
L/2
Fig. 6.15c-1 Stress Distribution over the whole cross-section of LC1300

LC1300 P=16.3kN
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
1
2
3
4
5
6
7
8
9
300
200
100
-100
-200
L/12
-300 L/4
L/2

based on ABAQUS
241
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
1
2
3
4
5
6
7
8
9
400
300
200
100
0
-100
-200
-300
-400 L/12
L/4
-500
L/2

based on ABAQUS

10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
1
2
3
4
5
6
7
8
9
400
300
200
100
0
-100
-200
-300
-400
-500 L/12
-600 L/4
L/2

based on ABAQUS
242
LC2000 P=8.4kN
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
1
2
3
4
5
6
7
8
9
-30
-40
-50
-60
-70
-80
-90 L/16
L/4
-100
L/2
Fig. 6.15d-1 Stress Distribution over the whole cross-section of LC2000

LC2000 P=11.60kN
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
1
2
3
4
5
6
7
8
9
100
50
-50
-100
-150
-200 L/16
L/4
-250
L/2

based on ABAQUS
243
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
1
2
3
4
5
6
7
8
9
150
100
50
0
-50
-100
-150
-200
-250
-300 L/16
-350 L/4
L/2

based on ABAQUS

10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
1
2
3
4
5
6
7
8
9
150
100
50
0
-50
-100
-150
-200
-250
-300
L/16
-350
L/4
L/2

based on ABAQUS
244
55
50
AS/NZS 4600 Clause3.4.6(b) (Nc) Test I-I (Pt)
45
Test (O-I) (Pt) ABAQUS (O-O) (AB)
40 ABAQUS (I-I) (AB)
35
Load (kN)
30
25
20
15
10
0
0 200 400 600 800 1000 1200 1400 1600 1800 2000
Fig. 6.16 Comparison of ABAQUS, Test, NAS 2001 and AS/NZS 4600
45
Test (O-O) (Pt) Pnd
Pnl Test (I-I) (Pt)
40 Test (O-I) (Pt) ABAQUS (O-O) (AB)
ABAQUS (I-I) (AB)
35
30
Load (kN)
25
20
15
10
0
0 200 400 600 800 1000 1200 1400 1600 1800 2000
Fig. 6.17 Comparison of ABAQUS, Test and the results based on SFSM
245
6.7 TABLES
For sections in 0.60 mm For sections in 0.60 mm

For sections in 0.42 mm
(060LB20na~40nb) (others )
Stress Strain Stress Strain Stress Strain
393.75, 0 620 0 434.923 0
514.065 6.25E-05 633.33 6.44E-04 475.243 2.16E-05
568.75 1.53E-04 640 4.94E-03 525.424 5.27E-05
590.63 2.13E-04 643.33 9.87E-03 561.415 1.24E-04
605.21 2.88E-04 640 1.20E-02 597.431 2.11E-04
634.38 5.00E-04 636.67 1.70E-02 633.497 3.49E-04
645.315 6.63E-04 640 2.20E-02 653.321 4.62E-04
650.78 8.55E-04 643.33 2.70E-02 683.109 8.04E-04
663.54 1.02E-03 640 3.70E-02 703.135 1.32E-03
667.19 1.18E-03 643.33 4.70E-02 709.968 1.77E-03
668.19 1.38E-03 711.15 2.33E-03
665.2 1.59E-03 707.602 3.10E-03
665.1 2.25E-03 705.365 3.34E-03
704.929 3.57E-03
704.723 3.92E-03
703.568 4.41E-03
702.957 5.05E-03
702.669 5.92E-03
696.743 6.30E-03
Table 6.1 Plastic data using for ABAQUS model
Specimen B20a B20b B30a B30b B40a B40b B50a B50b B60a
fy (MPa) 711 711 711 711 711 711 711 711 711
Ns (kN) 35.9 36.0 38.8 38.6 40.3 40.4 41.2 41.2 41.6
Pt (kN) 35.2 30.4 35.8 35.7 38.1 37.8 39.1 37.8 39.0
Pt /Ns 0.98 0.84 0.92 0.93 0.95 0.94 0.95 0.92 0.94
AB (kN) 38.5 38.5 36.8 36.8 38.6 38.6 40.0 40.0 40.9
AB/Ns 1.08 1.07 0.95 0.95 0.96 0.95 0.97 0.97 0.98
Specimen B60b B70a B70b B80a B80b B90a B90b B100a B100b
fy (MPa) 711 711 711 711 711 711 711 711 711
Ns (kN) 41.3 41.7 41.6 41.9 42.0 42.3 42.2 42.2 42.6
Pt (kN) 39.4 \ 41.8 42.0 42.0 \ 41.5 39.2 41.1
Pt /Ns 0.96 \ 1.00 1.00 1.00 \ 0.98 0.93 0.97
AB (kN) 40.9 40.5 40.5 41.6 41.6 41.2 41.2 41.7 41.7
AB/Ns 0.99 \ 0.97 0.99 0.99 \ 0.98 0.99 0.98
Table 6.2a Summary of Results for B-Sections
246

fy (MPa) 711 711 711 711 711 711 711 711
Ns (kN) 36.0 35.9 38.7 38.6 39.8 40.0 40.9 40.8
Pt (kN) 33.9 32.6 34.7 34.1 37.1 37.1 38.1 37.9
Pt /Ns 0.94 0.91 0.90 0.88 0.93 0.93 0.93 0.93
AB (kN) 38.5 38.5 36.8 36.8 38.6 38.6 40.0 40.0
AB/Ns 1.07 1.07 0.95 0.95 0.97 0.96 0.98 0.98
fy (MPa) 711 711 711 711 711 711 711 711
Ns (kN) 41.3 41.2 42.3 41.2 41.9 41.9 42.6 42.5
Pt (kN) 39.4 39.2 40.6 38.6 39.5 38.7 36.8 39.3
Pt /Ns 0.96 0.95 0.96 0.94 0.94 0.92 0.86 0.93
AB (kN) 40.5 40.5 41.6 41.6 41.2 41.2 41.7 41.7
AB/Ns 0.98 0.98 0.98 1.01 0.98 0.98 0.98 0.98
Table 6.2b Summary of Results for B-Sections
fy (MPa) 634 634 634 634 634 634 634 634 634
Ns (kN) 18.7 18.7 18.7 20.0 19.8 19.8 20.6 20.3 20.6
Pt (kN) 17.1 17.3 17.1 17.7 16.8 17.4 19.1 19.1 19.3
Pt /Ns 0.91 0.92 0.91 0.89 0.85 0.88 0.93 0.94 0.94
AB (kN) 19.9 19.9 19.9 18.3 18.3 18.3 19.4 19.4 19.4
AB/Ns 1.06 1.06 1.06 0.92 0.93 0.92 0.94 0.95 0.94
fy (MPa) 634 634 634 634 634 634 634 634 634
Ns (kN) 20.7 20.8 20.7 20.9 20.8 20.9 21.2 21.4 21.4
Pt (kN) 20.0 20.0 20.1 20.5 20.5 20.5 19.6 20.8 21.0
Pt /Ns 0.97 0.96 0.97 0.98 0.99 0.98 0.92 0.97 0.98
AB (kN) 19.8 19.8 19.8 20.1 20.1 20.1 20.4 20.4 20.4
AB/Ns 0.96 0.95 0.96 0.96 0.97 0.96 0.96 0.95 0.95
fy (MPa) 634 634 634 634 634 634 634 634 634
Ns (kN) 21.5 21.5 21.5 21.6 21.6 21.3 21.7 21.7 21.5
Pt (kN) 20.9 20.7 21.1 20.9 20.3 20.3 20.9 20.4 20.5
Pt /Ns 0.97 0.96 0.98 0.97 0.94 0.95 0.96 0.94 0.95
AB (kN) 20.5 20.5 20.5 20.3 20.3 20.3 20.7 20.7 20.7
AB/Ns 0.95 0.95 0.95 0.94 0.94 0.95 0.95 0.95 0.96
Table 6.3 Summary of Results for B-Sections
247

fy (MPa) 711 711 711 711 711 711
Ns (kN) 39.4 39.8 52.8 \ 60.9 56.6
Pt (kN) 37.0 38.0 48.7 \ 53.0 53.3
Pt /Ns 0.94 0.96 0.92 \ 0.87 0.94
AB (kN) 42.5 42.5 50.0 \ 55.3 55.3
AB/Ns 1.08 1.07 0.95 \ 0.91 0.98
Note: No holes & Clamps; 0.60 mm thickness
Specimen LB20ra LB20rb LB30ra LB30rb LB40ra LB40rb
fy (MPa) 711 711 711 711 711 711
Ns (kN) 38.6 38.5 48.0 48.2 51.8 51.5
Pt (kN) 37.3 35.6 46.0 45.8 47.3 44.6
Pt /Ns 0.97 0.93 0.96 0.95 0.91 0.87
AB (kN) 37.6 37.6 44.5 44.5 49.7 49.7
AB/Ns 0.98 0.98 0.93 0.92 0.96 0.97
Specimen LB20na LB20nb LB30na LB30nb LB40na LB40nb
fy (MPa) 670 670 670 670 670 670
Ns (kN) 37.6 37.7 51.3 51.3 58.8 58.7
Pt (kN) 36.4 35.4 45.8 45.8 50.6 51.3
Pt /Ns 0.97 0.94 0.89 0.89 0.86 0.87
AB (kN) 40.6 40.6 48.8 48.8 53.7 53.7
AB/Ns 1.08 1.08 0.95 0.95 0.91 0.91
Table 6.4a Summary of Results for LB-Sections

fy (MPa) 634 634 634 634 634 634
Ns (kN) 26.7 26.9 28.0 28.1 28.2 28.1
Pt (kN) 24.8 23.5 25.1 25.0 26.1 26.1
Pt /Ns 0.92 0.87 0.90 0.89 0.93 0.93
AB (kN) 26.2 26.2 26.4 26.4 26.6 26.6
AB/Ns 0.98 0.97 0.94 0.94 0.95 0.95
Table 6.4b Summary of Results for LB-Sections
Specimen HB20a HB20b HB30a HB30b HB30c HB40a HB40b HB60 HB80 HB100
fy (MPa) \ 711 711 711 711 711 711 711 711 711
Ns (kN) 42.1 42.1 46.2 46.3 46.3 48.3 48.3 50.2 51.9 51.8
Pt (kN) 37.3 37.4 39.9 43.6 41.1 44.4 43.6 45.4 47.7 45.5
Pt /Ns 0.89 0.89 0.86 0.94 0.89 0.92 0.90 0.90 0.92 0.88
AB (kN) 41.1 41.1 42.1 42.1 42.1 45.3 45.3 50.3 52.7 54.8
AB/Ns 0.98 0.98 0.91 0.91 0.91 0.94 0.94 1.00 1.01 1.06
Note: Teflon; 0.60 mm thickness
Table 6.5 Summary of Results for HB-Sections
248
Specimen fy(MPa) Pt(kN) AB(kN) AB/Pt

060LB20P450a 711 16.4 22.9 1.40
060LB20P450b 711 21.3 22.9 1.07
060LB20P900a 711 8.1 7.4 0.92
060LB20P900b 711 N/A 7.4 N/A
060LB30P450a 711 N/A 37.9 N/A
060LB30P450b 711 35.5 37.9 1.07
060LB30P900a 711 21.9 19.5 0.89
060LB30P900b 711 20.3 19.5 0.96
060LB40P450a 711 47.6 50.2 1.05
060LB40P450b 711 45.5 50.2 1.10
060LB40P900a 711 28.7 31.8 1.11
060LB40P900b 711 30.1 31.8 1.06
Specimen fy(MPa) Pt(kN) AB(kN) AB/Pt

042LB30P550a 690 17.5 18.7 1.07
042LB30P550b 690 17.4 18.7 1.07
042LB30P1100a 690 10.8 10.6 0.98
042LB30P1100b 690 11.7 10.6 0.90
042LB40P550a 690 20.9 23.3 1.11
042LB40P550b 690 23.0 23.3 1.01
042LB40P1100a 690 15.0 16.8 1.12
042LB40P1100b 690 14.2 16.8 1.18
042LB40P1700a 690 8.4 8.1 0.96
042LB40P1700b 690 8.8 8.1 0.92
042LB50P550a 690 23.4 24.5 1.05
042LB50P550b 690 23.6 24.5 1.04
042LB50P1100a 690 19.7 20.2 1.03
042LB50P1100b 690 17.5 20.2 1.15
042LB50P1700a 690 11.2 10.5 0.94
042LB50P1700b 690 10.7 10.5 0.98
249
Specimen 042LB30 042LB40 042LB50 060LB20 060LB30 060LB40
fy (MPa) 634.0 634.0 634.0 711.0 711.0 711.0
Pt (kN) 24.2 25.1 26.1 37.5 48.7 53.0
AB (kN) 26.2 26.4 26.6 42.5 50.0 55.3
Table 6.6c Average Results for Stub Column of LB Sections

(0.42mm & 0.60mm)
Specimen SLC360a SLC360b SLC360c SLC360d SLC360ca SLC360cb SLC360ra SLC360rb SLC360rc
fy (MPa) 690 690 690 690 690 690 690 690 690
Pt (kN) 40.2 37.5 35.6 33.9 36.5 36.6 40.7 37.0 38.5
Ns (kN) 34.1 33.7 34.0 33.9 34.1 34.0 34.1 34.3 34.0
Pt / Ns 1.2 1.1 1.0 1.0 1.1 1.1 1.2 1.1 1.1
AB (kN) 38.8 38.8 38.8 38.8 38.8 38.8 38.8 38.8 38.8
AB/Pt 0.97 1.03 1.09 1.14 1.06 1.06 0.95 1.05 1.01
Table 6.7a Summary of Test and ABAQUS Results for SLC-Sections
Specimen LC800a LC800b LC800sa LC800sb LC1300a LC1300b

fy (MPa) 690 690 690 690 690 690
Pt (kN) 22.8 (O-O) 24.1 (O-I) 21.5 (I-I) 21.5 (I-I) 18.4 (I-I) 17.3 (I-I)
Nc (kN) 33.0 32.9 32.5 32.4 30.5 30.6
Pt / Nc 0.69 0.73 0.66 0.66 0.60 0.57
AB (kN) 23.0 (O-O) 23.0 (O-O) 19.1 (I-I) 19.1 (I-I) 16.6 (I-I) 16.6 (I-I)
AB/Pt 1.01 0.95 0.89 0.89 0.90 0.96
Specimen LC1300c LC1300d LC1300sa LC1300sb LC2000a LC2000b
fy (MPa) 690 690 690 690 690 690
Pt (kN) 18.4 (I-I) 17.2 (I-I) 21.7 (O-O) 20.9 (O-O) 16.8 (I-I) 19.4 (O-O)
Nc (kN) 30.9 30.5 30.2 30.3 26.7 26.8
Pt / Nc 0.60 0.56 0.72 0.69 0.63 0.72
AB (kN) 16.6 (I-I) 16.6 (I-I) 18.9 (O-O) 18.9 (O-O) 15.1 (I-I) 17.4 (O-O)
AB/Pt 0.90 0.97 0.87 0.90 0.90 0.90
Note: Buckling Characteristic O-O; O-I, I-I represent Outwards and Inwards flanges buckling modes
Table 6.7b Summary of Test and ABAQUS Results for LC-Sections
250
Chapter 7: DESIGN PROPOSALS
CHAPTER 7
DESIGN PROPOSALS
CONTENTS
7.1 INTRODUCTION....................................................................................................252
7.2 REDUCTION FACTOR MODIFIED TO 0.9FY ..................................................252

7.2.1 Limitations on the use of high strength steels................................................................................. 252
7.2.3 Modified reduction factor............................................................................................................... 256
7.2.4 Reliability study .............................................................................................................................. 257
7.3 INTERACTION OF LOCAL AND OVERALL BUCKLING............................261

7.3.1 The strength prediction of members subjected to flexural buckling ............................................... 261
7.3.3 Modified design curve .................................................................................................................... 265
7.4 INTERACTION OF LOCAL AND DISTORTIONAL BUCKLING.................266

7.4.1 The strength prediction of singly-symmetric sections subjected to distortional buckling .............. 267
7.4.3 Proposed design methods ............................................................................................................... 273
7.4.3.1 Method 1: based on Kwon & Hancock Equation .................................................................................... 274
7.4.3.2 Method 2: based on AS/NZS 4600 Clause 3.4.6(b) ................................................................................ 275
7.4.3.3 Comparison with the experimental results .............................................................................................. 275
7.5 SUMMARY ..............................................................................................................276
7.6 FIGURES..................................................................................................................280
7.7 TABLES....................................................................................................................287
251
CHAPTER 7
DESIGN PROPOSALS
7.1 INTRODUCTION
The test programs described in Chapter 3, Chapter 4, and Chapter 5 provide

reliable and accurate test data for calibrating the design rules. Based on this
data, the finite element method as described in Chapter 6 has been used
successfully to predict the behaviour of thin-walled sections fabricated from
G550 thin sheet steel.
The main aim of this chapter is to further compare with the design standards and
propose design rules for high strength steels with limited ductility, for the
strength prediction of members subjected to the interaction of local and overall
buckling, and for the strength prediction of singly-symmetric sections subjected
to the interaction of local and distortional buckling.
Design proposals are made for the reduction factor for the design stress of high
strength steels; for the design of the thin-walled columns failing by local and
overall buckling; and for the design of the thin-walled columns with steel (0.42
~ 0.60 mm) failing by local and distortional buckling including interaction.
7.2 REDUCTION FACTOR MODIFIED TO 0.9fy
7.2.1 Limitation s on the use of high strength steels
252
The Australian/New Zealand Standard AS/NZS 4600, Clause 1.5.1.5(b):

restricts the design stress for AS 1397, Grade 550 steel, less than 0.9 mm in
thickness to 75 percent of the yield/tensile strength. In the North American
Specification (NAS, 2001), all structural Grade 80 steels are limited 75 percent
of the yield and tensile strengths and can only be used for multiple-web
configurations such as roofing, siding and floor decking. As an exception: for
multiple-web configurations, a reduced yield point, Rbfy, shall be used for
determining the nominal flexural strength. The reduction factor, Rb, shall be
determined as follows:
(a) Stiffened and Partially Stiffened Compression Flanges
for w/t≤ 0.067E/fy
Rb = 1.0
for 0.067E/fy <w/t<0.974E/fy
wf y
Rb = 1 − 0.26[ − 0.067]0.4 7.1-1
tE
for 0.974E/fy≤ w/t≤ 500
Rb = 0.75
(b) Unstiffened Compression Flanges
for w/t≤ 0.0173E/fy
Rb = 1.0
253
for 0.0173E/fy<w/t≤60
wf y
Rb = 1.079 − 0.6
(tE )
7.1-2
where E = modulus of elasticity; fy = yield point; t = thickness of section; w = flat width of

compression flange
Therefore, the design stress for the high strength steels has normally been
limited to 75 percent of their yield stress in both Standards. The main reason is
the lower ratio (<1.08) of the tensile strength to yield stress (fu/fy) and the lack
of knowledge on their structural behaviour.
In order to determine whether the 75 percent limit on yield stress is valid for
these thin AS 1397 steels, a series of carefully designed stub column tests were
carried out and detailed in Chapter 3.
7.2.2 Brief descr iption of the comparison of the test and ABAQUS results
with Design Standards
Here, only a brief description of the comparison of the tests and the ABAQUS
results with the design standards is given in order to emphasize the difference
between the test & ABAQUS results and design standards. More details were
given in Chapter 3 and Chapter 6. The test results (Pt) and ABAQUS (AB)
results non-dimensionalised with respect to the theoretical stub column strength
(Ns) are plotted against the plate slenderness ratios (b/t) in Figs. 7.1 & 7.2. The
theoretical stub column strengths (Ns) were calculated according to AS/NZS
4600 (1996) using the full yield stress fy without a reduction factor.
254
Solid and dotted curves are also plotted in Figs. 7.1 & 7.2. The dotted curve is
the ratio of NsRb/Ns versus plate slenderness (b/t). NsRb has been calculated based
on Rbfy as included in Section A3.3.2 of NAS (AISI, 2001). The solid curve is
the ratio of Ns0.75/Ns versus the plate slenderness (b/t). Ns0.75 has been calculated
based on 0.75fy as included in Clause 1.5.1.5(b) of AS/NZS 4600:1996.
Effective widths were used in the calculation of the theoretical stub column
strength (Ns). The calculation of the theoretical stub column strength (Ns) was
based on Clause 2.2.1.2 of AS/NZS 4600:1996 with an appropriate slenderness
(λ). The plate buckling coefficient (k) values were as specified in Clause 2.2.1.2
of AS/NZS 4600 for stiffened sections and Clause 2.3.1 of AS/NZS 4600 for the
unstiffened sections.
For determining the section or member capacity, the effective widths (be) of
uniformly compressed stiffened elements can be determined from the following
equations.
for λ ≤ 0.673:
be=b 7.2-1
for λ > 0.673:
be=ρb 7.2-2
where b = flat width of the element excluding the corner radii; ρ = effective width factor; λ =
slenderness ratio
255
0.22
ρ = (1 − )
λ
7.2-3
λ
1.052 b f*
λ=( )( )( ) 7.2-4
k t E
where, k = plate buckling coefficient; t = thickness of the element; E = Young’s modulus of

elasticity; f* = design stress in the compression element.
For the uniformly compressed unstiffened elements, the effective widths are
also based on the previous equations except that k is taken as 0.43 rather than 4
for the stiffened elements with simple supported longitudinal edges. The design
stress (f*) is taken as fy.
It can be seen that the results of the ABAQUS analyses deviate less for the
slender sections from the test results and slightly more for the stockier sections
from the test results. The ABAQUS results are in good agreement with the test
results. For the stockier sections, the results are closer to the NAS (AISI, 2001)
strengths. The strengths based on AS/NZS 4600 using a 75% reduction factor
for the yield stress are too conservative. These strengths were about 15% lower
than the results of tests and ABAQUS simulations. More details were given in
Chapter 6.
7.2.3 Modified r eduction factor
In Figs. 7.1 & 7.2, it can be seen that the ratios of the test results and theoretical
values become higher as the sections became more slender up to a plate
slenderness of about 133. The tendency of the test results was contrary to the
results based on the NAS Section A2.3.2 Exception Clause which decreases
with plate slenderness increases as shown in Figs. 7.1 & 7.2. Although the test
256
results and the results based on AS/NZS 4600 using a 75% reduction factor
applied to the yield stress have the same tendency, the design predictions are too
conservative for efficient design.
From the test data, it appears that a modified reduction factor applied to the
yield stress should be used. Three trial reduction factors (0.85, 0.90 and 0.95)
were chosen to be used to calculate the section capacity. The results (Ns0.85,
Ns0.90 and Ns0.95) based on those reduction factors were obtained and are plotted
in Figs. 7.1 & 7.2. The dash double-dot, dash and dash-dot curves are the ratios
of Ns0.85/Ns, Ns0.90/Ns and Ns0.95/Ns respectively versus the plate slenderness (b/t).
As can be seen in Figs. 7.1 & 7.2, the dash curves based on a reduction factor of
0.90 fit the mean test results well. A modified reduction factor of 0.90 can
therefore be used for compression to replace the reduction factor 0.75, which is
specified for G550 steel with a thickness less than 0.9 mm in AS/NZS 4600 or
in the NAS (AISI, 2001).
7.2.4 Reliability study
The reliability of a structure at various limit states can be estimated by means of

a first order second moment (FOSM). The analytical model has been described
by Ellingwood et al. (1980) to calculate the reliability index. Here, this method
is used to calculate the resistance (capacity) factors ( φ ) based on target
reliability (safety) indices (β0 ) for the stub column tests in which the specimens
were fabricated from G550 sheet steels. The target reliability indices used here
are 2.5 for cold-formed members as specified in AISI (1996).
Limit state design philosophy requires that the sum of the factored loads applied
to a member be less than or equal to the factored resistance of the member, as
257
follows;
∑γ Qi ni ≤φ ⋅ Rn 7.2-5
where Qni represents the general nominal load effect and Rn represents the nominal
resistance.
Standards which incorporate limit state design principles require that the total
load, Q, applied to a member be less than or equal to the resistance of the
member,
Q ≤ φ ⋅ Rn 7.2-6
The total load values used in this reliability study are based on combinations of
dead and live loads specified in the applicable design standards. In AISI-1996
requires the following combinations of dead and live loads.
Q = 1.2 D + 1.6 L 7.2-7
The mean load effect can be expressed as a function of the mean dead and live
loads, Dm and Lm
Qm = Dm + Lm 7.2-8
In an investigation of load statistics Ellingwood et al. (1980) concluded that

dead loads are typically underestimated while the mean live load is equal to the
code specified live load "if the tributary area is small enough so that no live load
reduction is included (AISI Commentary, 1996), hence Dm=1.05D and Lm=L.
258
The AISI Commentary (1996), as well as Elligwood et al. specify the ratio of
dead to live found in common cold formed steel structures as D/L=0.20. With
this information, as well as the above equations, it is possible to express the
mean load effect as a ratio of the factored tensile resistance, as follows:
Qm = 1.05 ⋅ (0.20 L) + L = 1.21L 7.2-9
Hence, Equation Q≤φ ⋅ Rn becomes Qm ≤ 0.657⋅φ ⋅ Rn (AISI).
Load statistics have been analyzed in a study of the National Bureau of

Standards (NBS) (Ellingwood et al. (1980), where it was shown that
Dm=1.05Dn, VD=0.1, Lm=Ln, VL=0.25. Dm and Lm are the mean dead and live
load intensities, respectively. Dn and Ln are the nominal dead and live load
intensities, respectively. VD and VL are the corresponding coefficient of
variation. (Commentary AISI 1996, VI-19)
The mean resistance can be defined by Rm/Rn,
Rm Rn = M m ⋅ Fm ⋅ Pm 7.2-10
where Mm, Fm and Pm are the mean values of the material factor, the fabrication factor and
the professional factor respectively.
The coefficient of variation can be defined by VR,
VR = VM2 + VF2 + VP2 7.2-11
where VM, VF and VP represent the coefficient of variations of the material factor, the
259
fabrication factor and the professional factor respectively.
The values of Pm and VP are determined from comparison of test results and
predictions of theoretical models based on measured values of material
properties and dimensions. The values of Mm and VM , Fm and VF are based on
statistical readings.
According to Ravindra and Galambos (1978), the coefficient of variation of the

load effect, VQ, can be expressed as,
2 2 2
 Dm  D  L 
  ⋅ VD 2 ⋅  n  +  m  ⋅ VL2
 Dn   Ln   Ln 
VQ = 7.2-12
 Dm   Dn   Lm 
  ⋅   +  
 Dn   Ln   Ln 
The resistance factor ( φ ) can be expressed as,
− β 0 ⋅ VM2 +VF2 + C P ⋅VP 2 +VQ2

φ c = 1.5 ⋅ ( M m ⋅ Fm ⋅ Pm ) ⋅ e 7.2-13
1 m
where CP is the correction factor, C P = (1 + ) ⋅ ; m=n-1, n is the number of tests.
n m−2
If the current resistance factor ( φ 0 ) is used, then the reliability index can be
calculated as follows,
Dn
γ D ⋅(
)+γ L
Rm Ln R 1
= ⋅( m )⋅ 7.2-14
Qm D D L Rn φ 0
( m )⋅( n ) + ( m )
Dn Ln Ln
260
where γ D and γ L are the load factors for dead and live load respectively.
For the FOSM method when the resistance (R) and the load (Q) both have
lognormal distributions, the reliability index ( β c ) can be expressed (Ravindra
and Galambos (1978)) as follows,
ln (Rm Qm )
βc = 7.2-15
VR2 + VQ2
The resistance factors calculated for given values of β 0 and the statistical data
used in the calculations are given in Table 7.2.
For the stub column tests fabricated from 0.42 mm and 0.60 mm thicknesses
sheet steels, the resistance factors are about 1.21 and 1.24 respectively.
Therefore, the resistance factors determined using AISI dead and live load
combinations are greater than the current specified value, φ 0 =0.85, which
indicates that the load carrying capacity of G550 stub column tests can be
adequately predicted using existing procedures. The main reason for the high
resistance factors is the Mm ratio of 1.308 & 1.278 for 0.42 mm & 0.60 mm
respectively. This ratio commonly occurs for high strength cold-formed steel.
7.3 INTERACTION OF LOCAL AND OVERALL BUCKLING
7.3.1 The streng th prediction of members subjected to flexural buckling
The design rules for the members subject to concentrically loaded compression
are given in Clause 3.4 in AS/NZS 4600 and the NAS (AISI, 2001).
261
For members subjected to concentric compression, the nominal section capacity

(Ns) and the nominal member capacity (Nc) of the member can be computed
using the following equations:
Ns = nominal section capacity of the member in compression
= Ae fy 7.3-1
where Ae = effective area at yield stress (fy)
Nc = nominal member capacity of the member in compression
= Ae fn . 7.3-2
where Ae = effective area at the critical stress (fn).
fn = critical stress and is determined from the following equations:
for λc ≤ 1.5:
2
f n = (0.658λc ) f y 7.3-3
for λc > 1.5:
0.877
fn = 2
fy 7.3-4
λc
where λc = non-dimensional slenderness
262
fy
λc = 7.3-5
f oc
where foc = least of the elastic flexural, torsional and torsional-flexural buckling stress
A slender, axially loaded column will fail by overall flexural buckling if the
cross-section of the column is a doubly-symmetric closed shape. For the
columns with those section shapes that can be shown not to be subject to
torsional or flexural-torsional buckling, the elastic flexural buckling stress (foc)
can determined from the following equation:
π 2E
f oc = 7.3-6
(le r ) 2
where le = effective length of member; r = radius of gyration of the full, unreduced cross-
section.
Section C4 of the NAS has the same formulae specification for concentrically
loaded compression members.
The test results (Pt), ABAQUS results (AB) and theoretical proposed column
strength Ncred are given in Tables 7.1 where the ratios Pt/Ncred and AB/Ncred show
the comparisons. The Ncred values are the proposed column design strengths as
detailed later.
Figs. 7.3a &7.3b & 7.3c shows the test strength (Pt) ABAQUS values (AB)
compared with a range of design curves for the 30 mm, 40 mm and 50 mm
263
sections in 0.42 mm material respectively. Figs. 7.4a & 7.4b & 7.4c show the
same for the 20 mm, 30 mm and 40 mm sections in 0.60 mm material
respectively. In Figs. 7.3 & 7.4, the test strengths (Pt) and ABAQUS values
(AB) have been non-dimensionalised with respect to the theoretical squash load
(Ny).
Dashed and dash-dotted curves are plotted in Figs. 7.3 & 7.4. The dashed curve
is the ratio of NcRb/Ny against the column length (lx). NcRb was calculated using
the AISI Specification based on a yield stress Rbfy as included in Section A3.3.2
of the NAS (AISI, 2001). The dash-dotted curve is the ratio of Nc0.75/Ny against
column length (lx). Nc0.75 was calculated based on AS/NZS 4600 with a yield
stress of 0.75fy as included in Clause 1.5.1.5(b) of AS/NZS 4600.
The horizontal dashed line represents the ratio of the theoretical local buckling
load to the squash load (Nol /Ny) against the column length (lx). The horizontal
solid line represents the ratio of the section capacity to the squash load
(Ns0.90/Ny) against the column length (lx). Ns0.90 is calculated based on AS/NZS
4600 with a yield stress of 0.90fy as proposed in Section 7.3.1.
The curved solid curve is the ratio of the Euler load (Noc=A foc) to the squash
load (Ny) against the column length (lx). The heavy dash-dot curve is the
proposed design curve, which is the ratio of the nominal reduced member
capacity to the squash load (Ncred/Ny) against the column length (lx). The
proposed design method is described in the following Section 7.3.2.2.
than the existing design curves at intermediate columns lengths where local and
Euler buckling interact.
264
As shown in Figs. 7.3a & 7.3b& 7.3c, for the columns in 0.42 mm thickness, the
results are lower than the curves based on AS/NZS 4600 and the AISI
Specification. For the sections of stockier cross-section, the AS/NZS 4600
design curve based on 0.75fy and the AISI design curve based on a stress Rbfy
are generally slightly conservative. More details are given in Chapter 4.
The ABAQUS results show a similar trend to the test results when compared
with the design curves although, as detailed above, they are slightly higher
probably due to assumptions associated with the assumed imperfections. More
details are given in Chapter 6.
The conclusion is that a modified design method is required for sections with
low local buckling loads as may occur in high strength steel columns with
slender plate elements.
7.3.3 Modified d esign curve
The effect of local buckling on overall buckling behaviour has been studied in
several research projects. On the basis of tests and analytical studies, DeWolf,
Pekoz, Winter (1974) and Kalyanaraman, Pekoz, Winter (1977) concluded that
a satisfactory approach is to calculate the overall buckling load using the
effective radius of gyration and the effective area, both calculated at the overall
buckling stress. The unified approach proposed by Pekoz (1986) was adopted
by the AISI Specification (1986, 1991, 1996) and AS/NZS 4600 (1996) to
determine the strength of columns where local and Euler buckling interact. This
proposed approach does not take account of the reduction of the radius of
gyration (flexural rigidity) resulting from local buckling which was proposed by
the earlier researchers. However, for the high strength steel sections, the effect
of the reduction in the radius of gyration on the interaction of local buckling and
265
overall buckling is significant, especially buckling of these very slender sections

at loads above the local buckling load.
To fit the test data, a new design approach is proposed. The proposed approach
consists of two steps. Firstly, the reduced yield point 0.90fy, which was
determined in Section 7.3 for stub columns, is used to replace the yield stress in
Clause 3.4 of AS/NZS 4600 and Section C4 of the AISI specification. Secondly,
a reduction factor γ is applied to the radius of gyration as defined in Equation
7.3-7. It is a function of the length varying from some limit γ0 at length lx=0
(here γ0 taken as 0.65) to 1.0 at length lx=1.10*lx0. Here, lx0 is the length where
the local buckling load equals the Euler buckling load as defined by Equation
7.3-8. The reduction factor γ accounts for the loss of flexural rigidity due to
local buckling. The value of reduced radius of gyration γ*rx is used in Clause
3.4.2 of AS/NZS 4600 to replace the normal radius of gyration rx (Section C4.1,
Eq. C4.1-1 of the NAS (AISI, 2001). It can be seen in Figs 7.3 & 7.4 that the
proposed design curves based on the reduced the radius of gyration γ fit the test
data well.
(1 − γ 0 ) * l x
γ =γ0 + 7.3-7
1.1 * l x 0
E
l x 0 = π * rx * 7.3-8
f ol
These design rules has recently been proposed to the AISI Specification
committee as Ballot S03-197A.
7.4 INTERACTION OF LOCAL AND DISTORTIONAL BUCKLING
266
7.4.1 The streng th prediction of singly-symmetric sections subjected to

distortional buckling
As detailed in Chapter 5, many methods, such as : the Effective Width Method

(EWM) and the Direct Strength Method (DSM) and AS/NZS 4600 Clause
3.4.6(b) and the Kwon & Hancock Equation, are used for computing the
member axial capacity (Nc) of the LC-section columns. In AS/NZS 4600, the
formulae are given in Clause 3.4.6 for singly-symmetric sections subject to
distortional buckling as detailed in Chapter 5.
In this section, only the method based on the new North American Specification
(NAS) (AISI, 2001) which was recently revised to provide EWM design
provisions for the determination of the effective widths of uniformly
compressed stiffened elements with multiple intermediate stiffeners or edge
stiffened elements with intermediate stiffeners are presented in more detail for
determining the effective widths of the sections.
The main difference between the 1996 AISI Specification and the North
American Specification (2001) for determining the effective widths is mainly in
determining the plate buckling coefficient. The NAS (AISI, 2001) Section B5.1
was revised to take account of the two buckling modes (local buckling and
distortional buckling) for members with multiple intermediate stiffeners in the
compression flange. The reduction factor (ρ) is applied to entire element instead
of only the flat portions. The following steps present the basic formulae for
obtaining the plate buckling coefficient (k) for the web element with one
intermediate stiffener and the flange element with an edge stiffener and one
intermediate stiffener.
• Uniformly compressed web with one intermediate stiffener (NAS Section
267
B4.1)
k = 3( RI ) n + 1 7.4-1
b0 / t 1
n = (0.583 − )≥ 7.4-2
12 S 3
Is
RI = ≤1 7.4-3
Ia
where: for S < b0/t < 3S
50b0 / t
Ia = t 4 ( − 50) 7.4-4
S
for b0/t > 3S
128b0 / t
Ia = t 4 ( − 285) 7.4-5
S
S = 1.28 E f 7.4-6
where Is = moment of inertia of full section of stiffener about its own centroidal axis parallel
to element to be stiffened.
• Uniformly compressed edge stiffened flange with intermediate stiffeners

(NAS Section B5.2)
Section B5.2 uses the provisions of Section B4.2 when b0 / t > 0.328S but with
b0 replacing w in all expressions.
268
if b0 / t > 0.328S
then, from B4.2 for a simple lip edge stiffener,
for d / b0 ≤ 0.25
k = 3.57( RI ) n + 0.43 ≤ 4 7.5-1
for 0.25 < d / b0 ≤ 0.8
k = (4.82 − 5d / b0 )( RI ) n + 0.43 ≤ 4 7.5-2
RI = Is / Ia ≤1 7.5-3
b0 / t b /t
I a = 399t 4 ( − 0.328)3 ≤ t 4 [115( 0 ) + 5] 7.5-4
S S
b0 / t
n = [0.582 − ] ≥ 1/ 3 7.5-5
4S
If k calculated from the above formulae is less than 4.0, the intermediate
stiffener is ignored and the provisions of Section B4.2 described above are used.
If k is equal to 4.0, the following formulae from Section B5.2 are used to
determine the k for calculating the effective widths.
k = min{kloc , Rkd } 7.6-1
kloc = 4(n+1)2 7.6-2
269
(1 − β 2 ) 2 + γ (1 + n)
kd = 7.6-3
β 2 [1 + δ (n + 1)]
β = [1 + γ (n + 1)]1 / 4 7.6-4
10.92 I sp
γ = 7.6-5
b0 t 3
As
δ = 7.6-7
b0 t
for b0/h < 1
R=2
for b0/h ≥ 1
11 − b0 / h 1
R= ≥ 7.6-8
5 2
It should be noted that a recent ballot C/S03-196A of the AISI Specification

committee confirms that Section B5.2 is applicable for edge stiffened element
with only one intermediate stiffener as used in this thesis.
The section axial capacity (Ns) (Pn in NAS) was obtained using the EWM, and
gave a result of 34 kN which was slightly lower than the average ultimate test
270
strength 37 kN for the stub column tests. As shown in Figs. 7.5a&b&c, the
dash-dotted curve is the member capacity (Nc) (Pn in NAS) plotted against the
effective length (Le). The effective length Le has been taken as half the column
length to allow for the fixed ends. Nc was calculated using the new NAS (AISI,
2001) based on the clauses for determination of long column strength of a
locally buckled section. The test results (Pt) plotted against the effective length
(Le) are also shown in Figs. 7.5a&b&c identified using symbols (■,▲,◆). As can
be seen, for the long specimens, the test results were much lower than the
strengths based on the new NAS (AISI, 2001). There are two reasons for these
unconservative results. Firstly, distortional buckling is not adequately accounted
for by the NAS. Secondly, the interaction of local and distortional buckling,
which causes the dip in the test strength curve for the 800 mm and 1300 mm
specimens, is also not adequately accounted for.
Two alternative methods were used to calculate the member ultimate strength as
shown in Fig. 7.5a.
1. The Winter formula based on the distortional buckling stress was modified
by Kwon and Hancock (1991) as given Chapter 5. As shown in Fig. 7.5a, the
line-dotted curve ( • ) is the member ultimate strength (Pkh) against the
effective length (Le). (fd) was obtained as 63 MPa using the SFSM at a
length of 2000 mm.
2. Clause 3.4.6 (b) in AS/NZS 4600 (2002) was also used to calculate the
member ultimate strength as given in Chapter 5. As shown in Fig. 7.5a, the
line-cross curve ( ∗ ) is the member ultimate strength (Pc) plotted against the
effective length (Le).
271
The distortional buckling strength (fd) used in these calculations was based on
the SFSM and therefore allows for the effect of the warping restrained end
conditions. As can be seen, the results based on AS/NZS 4600 Clause 3.4.6(b)
and Kwon & Hancock’s equation are approximately 32% and 22% higher on
average than the test results of the 1300 mm and 2000 mm long columns. These
results indicate that for the long columns, the interaction of local and
distortional buckling has a significant effect on the member strength. Hence, for
the monosymmetric sections of the type tested subject to distortional buckling,
the interaction of local and distortional buckling has to be taken into account
when determining the design member strength.
The ABAQUS results (◊, ∆) are plotted in Fig. 7.5b and compared with test
results and the design standards. From the figures, it can be seen that the
ABAQUS results are in good agreement with the test results for all columns. As
shown in Fig. 7.5b the ABAQUS results in the O-O mode and I-I mode are
lower than the test results in O-O mode and I-I mode. The ABAQUS results are
consistent with the test results, since the results in the O-O buckling mode are
higher than the results in the I-I buckling mode as shown in Table 6.7b.
A series of results was obtained using a modification of the Direct Strength

Method (DSM) based on the elastic bucking stresses which were computed
using the SFSM. A comparison of the results with test results is presented in
Figs. 7.6a&b&c.
The Pcrl (‫־־‬+‫ )־־‬and Pcrd (‫● ־‬ ) as shown in Fig. 7.6a represent the elastic local
‫־‬
buckling load and elastic distortional buckling load based on the elastic local
buckling stress at 360 mm half-wavelength and the elastic distortional buckling
stress at 2000 mm half-wavelength using SFSM respectively.
272
In Figs. 7.6a&b&c, the Pnl (+) and Pnd ( • ) represent the results using the
DSM based on the elastic buckling stresses from the SFSM. As can be seen, the
results of the DSM are in good agreement with the results of the tests (■) except
the results at 1300 mm length which are clearly lower than Pnd. The results of
Pnd and Pnl, which were obtained using the elastic buckling stresses based on the
SFSM taking into account the fixed ends boundary conditions, conform with the
test results at 360, 800 and 2000 mm lengths but not at 1300 mm. As in the
preceding discussion, the failure modes at these lengths (360, 800 and 2000
mm) were mainly dominated by local buckling and distortional buckling
respectively. Hence, the lower test results for the 1300 mm length may be a
result of the significant interaction of L+D buckling. It can also be noted that
even though the length variation from 1300 to 2000 mm was large, the test
results increase slightly. This may be a consequence of the different failure
mode. The failure mode of the column at the 2000 mm length was mainly
distortional buckling, whereas the failure of the column at the 1300 mm length
was interaction of local and distortional buckling. This interaction induced the
premature failure of the column at the 1300 mm length.
The ABAQUS results (◊, ∆) are plotted in Fig. 7.6b and compared with test
results and the design standards. The ABAQUS results are lower than the test
results and much lower than the results based on the DSM.
As shown in Figs 7.6a&b, the results based on the DSM are unconservative for
the columns at intermediate lengths.
7.4.3 Proposed d esign methods
As shown in Figs 7.5a&b and 7.6a&b, design procedures accounting for the
interaction of local and distortional buckling in thin sections of high strength
273
G550 steel are required. Based on AS/NZS 4600 Clause 3.4.6(b) and the Kwon
& Hancock Equation, two simple design methods are proposed.
intermediate lengths, the interaction of local and distortional buckling is taken
into account with the distortional mode treated as an overall mode in the DSM.
This nominal axial strength designated Pnld is used here.
7.4.3.1 Method 1: based on Kwon & Hancock Equation
Using the preceding assumptions, the distortional buckling strength (Pkh) based
on Equation 5.1-2 is adopted as the compression member design strength (Pne)
in Equations 5.3-1, 5.3-2.
From Equations 5.1-2, 5.3-1 and 5.3-2, the equations for Pnld considering the
L+D interaction are obtained.
For λl ≤0.776
Pnld =Pkh 7.7-1
For λl >0.776
 0.4  
0.4
P    Pcrl 
Pnld = Pkh 1− 0.15 crl   P  7.7-2
  Pkh 
   kh 
Pkh
where λl =
Pcrl
Pcrl is the elastic local buckling load; Pnld is the nominal axial strength accounting for
274
7.4.3.2 Method 2: based on AS/NZS 4600 Clause 3.4.6(b)
With the preceding assumption, the distortional buckling strength (Pc) based on
Equation. 5.2-2 is adopted as the compression member design strength (Pne) in
Equations 5.3-1, 5.3-2.
From Equations. 5.2-2, 5.3-1 and 5.3-2, the equations for Pint considering the
L+D interaction are obtained.
For λl ≤0.776
Pnld =Pc 7.8-1
For λl >0.776
 0.4  
0.4
P    Pcrl 
Pnld = Pc 1− 0.15 crl   P  7.8-2
  Pc    c 

Pc
where λl =
Pcrl
Pcrl is the elastic local buckling load; Pnld is the nominal axial strength accounting for
7.4.3.3 Comparison with the experimental results
The calculations were carried out using the two proposed design methods. The
stresses (fcrl and fd) and load (Pcrl) used in Equations 7.7-1~7.8-2 were obtained
using the SFSM. In Figs. 7.5c & 7.6c, the nominal interaction axial strength Pint
275
(▬) and (●) represent the results using Method 1 and Method 2 respectively.
The results based on Method 1 are 13% and 19% lower than the average test
results while the results based on Method 2 are 11% and 15% lower than the
average test results for the 1300 mm and 2000 mm lengths respectively. On the
other hand, the results are 10%, 13% and 8% , 9% lower than the lowest test
results (the worst cases, 17.2 kN and 16.8 kN for 1300 mm and 2000 mm
lengths) respectively.
Obviously, Method 1 is more conservative than the Method 2 but the difference
is less than 4%. Taking the worst case into consideration, Method 2 is in good
agreement with the test results. The differences are within 9%.
Therefore, to account for the adverse interaction of local and distortional

buckling when it occurs, the proposed design methods have to be used for
intermediate length sections when the local buckling and distortional buckling
occur simultaneously or nearly simultaneously.
7.5 SUMMARY
From Stub Column Tests:
A range of stub columns in G550 sheet steel to AS1397 have been tested in
compression to obtain the strength characteristics of this steel with low strain-
hardening. The following detailed conclusions can be made:
• The results of the successful stub column tests and ABAQUS simulation
have been compared with the design procedures in the Australian/New
276
Zealand Standard for Cold-Formed Steel Structures and North American

Specificatio0n. As expected, the greatest effect of low strain hardening was
for the stockier sections where material properties play an important role.
For the more slender sections where elastic local buckling and post-local
buckling are more important, the effect of low strain hardening does not
appear to be as significant. A reduction factor for design stress needs to be
proposed.
• As shown in Figs. 7.1 & 7.2, the ratios of AB/Ns and Pt/Ns were mostly
larger than 0.90 and higher than the results based on the NAS (AISI, 2001)
(Section A02.3.2, Exception) and AS/NZS 4600 when the plate slenderness
(b/t) was greater than 88. When the plate slenderness (b/t) was less than 50,
the formulae used in NAS (AISI, 2001) was slightly unconservative due to
the imperfection sensitivity. The conclusion drawn from the test results is
that for stub compression members, 0.90fy may be used as the reduced yield
point to determine the nominal strength of the stub columns.
From Closed-Section Long Column Tests:
Long column tests with box-sections and constructed from high strength G550
steel have been successfully performed. The plate slenderness (b/t) ranged from
33 to 119 and the column slenderness (L/rx) ranged from 27 to 148. A
comparison is made between design strength curve and the design standards.
stresses (fol), failed by overall buckling. The test results and the ABAQUS
results were close to the curves based on AS/NZS 4600 and the AISI
Specification. For very long columns with slender plate elements which had
lower local buckling stresses (fol), failure was still governed by overall
277
buckling although local buckling occurred. However, for the intermediate

length columns, the failure mode was governed by the interaction of local
and overall buckling. The interaction of local and overall buckling reduced
the column strength and made the test results lower than the design curves.
The worst case had a difference between test results and the results based on
AS/NZS 4600 and the AISI Specification of about 14%, which means that
for the slender sections AS/NZS 4600 and the AISI method are
unconservative.
• To account for the loss of flexural rigidity due to local buckling for the
slender sections, a reduction of the radius of gyration is needed to take
account of interaction buckling in the design curve. The proposed design
curve based on a reduction factor ( γ ) fits the test data well. So the reduced
radius of gyration γ*rx may be used in the design curve for the slender
sections.
From Open-Section Long Column Tests:
A series of channel section column tests with intermediate stiffeners in the web
and flanges and constructed from high strength G550 steel has been successfully
performed. The tests were carried out to investigate the effects of local and
distortional bucking as well as the interaction between them.
For the stub columns, local buckling is the dominant failure mode. However, for
the long specimens, interaction between local and distortional buckling led to
failure.
• The new NAS (AISI 2001) (Clause B5.2) gives an accurate prediction of the
stub column strength using the new rules for an edge stiffened element with
278
an intermediate stiffener but gives an unconservative prediction of the long

column strength since it ignores distortional bucking.
• The adverse interaction between local and distortional buckling reduced

significantly the member ultimate strength for the intermediate length
specimens fabricated from thin sheet steel.
• AS/NZS 4600 gives an unconservative prediction at intermediate lengths

because it ignores interaction of local and distortional buckling.
• Two simple design methods based on the Direct Strength Method are
proposed for intermediate length lipped channel section fabricated from
G550 thin sheet steel to account for the adverse interaction of local and
distortional modes. Those two method give a lower bound of the test results.
279
7.6 FIGURES
1.10
60Ba/Ns 60Bb/Ns 60Bra&rb/Ns
60HB/Ns 60BWH/Ns 60LB(all)/Ns
60AB(B)/Ns 60AB(LB)/Ns 60AB(HB)/Ns
1.05 AISI/Ns AS4600/Ns 085fy/Ns
090fy/Ns 095fy/Ns
1.00
Pt/Ns, AB/N s
0.95
0.90
0.85
0.80
0.75
20 40 60 80 100 120 140 160 180
b/t
Fig. 7.1 Comparison of Proposed Method with Test, ABAQUS Results, and
1.1
42Ba/Ns 42Bb/Ns 42Bc/Ns 42LB/Ns
42AB(B)/Ns 42AB(LB)/Ns AISI/Ns AS4600/Ns
1.05 085fy/Ns 090fy/Ns 095fy/Ns
1
Pt/Ns, AB/N s
0.95
0.9
0.85
0.8
0.75
20 40 60 80 100 120 140 160 180
b/t
Fig. 7.2 Comparison of Proposed Method with Test, ABAQUS Results, and
280
1 AS/NZS 4600 (Nc)

AISI, NcRb
Euler Curve (Noc)
Section Capacity, Ns0.90
0.75
Dimsionless Load

Proposed Design Curve, Ncred
T est Results (Pt)
0.25
0
0 500 1000 1500 2000 2500 3000
Length (mm)
Fig. 7.3a Comparison of Design Method with Test, ABAQUS Results ,and
Design Standard (0.42mm), 30x30 mm Sections
AS/NZS 4600 (Nc)

1
AISI, NcRb
Euler Curve (Noc)
0.75
Dimsionless Load

T est Results (Pt)
0.5
ABAQUS Results (AB)
0.25
0
0 500 1000 1500 2000 2500 3000
Length (mm)
Fig. 7.3b Comparison of Design Method with Test, ABAQUS Results ,and
281
1 AS/NZS 4600 (Nc)

AISI, NcRb
Euler Curve (Noc)
0.75 Local buckling Load (Nol)
Dimsionless Load

T est Results (Pt)
ABAQUS Results (AB)
0.5
0.25
0
0 500 1000 1500 2000 2500 3000
Length (mm)
Fig. 7.3c Comparison of Design Method with Test, ABAQUS Results ,and
1.5
AS/NZS 4600 (Nc)

1.25 AISI, NcRb
Euler Curve (Noc)
1
Dimsionless Load

ABAQUS Results (AB)
0.5
0.25
0
0 500 1000 1500 2000 2500 3000
Length (mm)
Fig. 7.4a Comparison of Design Method with Test, ABAQUS Results ,and
282
1
AS/NZS 4600 (Nc)
AISI, NcRb
Euler Curve (Noc)
0.75 Section Capacity, Ns0.90
Dimsionless Load

T est Results (Pt)
0.25
0
0 500 1000 1500 2000 2500 3000
Length (mm)
Fig. 7.4b Comparison of Design Method with Test, ABAQUS Results ,and
1 AS/NZS 4600 Nc)

AISI, NcRb
Euler Curve (Noc)
0.75
Dimsionless Load

T est Results (Pt)
0.5
ABAQUS Results (AB)
0.25
0
0 500 1000 1500 2000 2500 3000
Length (mm)
Fig. 7.4c Comparison of Design Method with Test, ABAQUS Results ,and
283
55
50 Section Capacity (NAS, 2001) (Ns) AS/NZS 4600 Clause3.4.6(b) (Nc)
Kwon & Hancock Equation Test I-I (Pt)
45
Test (O-I) (Pt)
40
35
Load (kN)
30
25
20
15
10
0
0 200 400 600 800 1000 1200 1400 1600 1800 2000
Fig. 7.5a Comparison of Test, NAS 2001 and AS/NZS 4600
55
50
AS/NZS 4600 Clause3.4.6(b) (Nc) Test I-I (Pt)
40
ABAQUS (I-I) (AB)
35
Load (kN)
30
25
20
15
10
0
0 200 400 600 800 1000 1200 1400 1600 1800 2000
Fig. 7.5b Comparison of ABAQUS, Test, NAS 2001 and AS/NZS 4600
284
55
50 AS/NZS 4600 Clause3.4.6(b) (Nc) Test I-I (Pt)
Test (O-I) (Pt) Method-1
45
Method-2
40
35
Load (kN)
30
25
20
15
10
0
0 200 400 600 800 1000 1200 1400 1600 1800 2000
Fig. 7.5c Comparison of Test, NAS 2001, AS/NZS 4600 with

the proposed design methods
45
Pcrd (2000mm-SFSM) Pcrl (360mm-SFSM)
40 Test (O-O) (Pt) Pnd
Pnl Test (I-I) (Pt)
Test (O-I) (Pt)
35
30
Load (kN)
25
20
15
10
0
0 200 400 600 800 1000 1200 1400 1600 1800 2000
Effe ctive Le ngth (mm)
Fig. 7.6a Comparison of Test and the DSM
285
45
Test (O-O) (Pt) Pnd
Pnl Test (I-I) (Pt)
ABAQUS (I-I) (AB)
35
30
Load (kN)
25
20
15
10
0
0 200 400 600 800 1000 1200 1400 1600 1800 2000
Fig. 7.6b Comparison of ABAQUS, Test and the DSM

45
Test (O-O) (Pt) Pnd Pnl

40
Test (I-I) (Pt) Test (O-I) (Pt) Method-1
Method-2
35
30
Load (kN)
25
20
15
10
0
0 200 400 600 800 1000 1200 1400 1600 1800 2000
Effe ctive Le ngth (mm)
Fig. 7.6c Comparison of Test, the DSM with the proposed design methods
286
7.7 TABLES
fy Ny Ncred Pt AB fol
Specimen b/t Pt//Ncred AB/Ncred AB/Pt
(MPa) (kN) (kN) (kN) (kN) (MPa)
060LB20P450a 33.3 711 43.35 15.7 16.4 1.04 22.88 1.46 1.40 954.6
060LB20P450b 33.3 711 42.95 15.7 21.3 1.36 22.88 1.46 1.07 954.6
060LB20P900a 33.3 711 44.03 5.2 8.1 1.56 7.43 1.43 0.92 954.6
060LB20P900b 33.3 711 44.17 5.2 N/A N/A 7.43 1.43 0.49 954.6
060LB30P450a 50.0 711 64.54 29.0 N/A N/A 37.86 1.31 N/A 411.7
060LB30P450b 50.0 711 64.25 29.0 35.5 1.22 37.86 1.31 1.07 411.7
060LB30P900a 50.0 711 64.98 16.4 21.9 1.34 19.50 1.19 0.89 411.7
060LB30P900b 50.0 711 64.22 16.4 20.3 1.24 19.50 1.19 0.96 411.7
060LB40P450a 66.7 711 85.87 39.3 47.6 1.21 50.17 1.28 1.05 225.6
060LB40P450b 66.7 711 85.79 39.3 45.5 1.16 50.17 1.28 1.10 225.6
060LB40P900a 66.7 711 85.97 25.3 28.7 1.13 31.76 1.26 1.11 225.6
060LB40P900b 66.7 711 85.68 25.3 30.1 0.12 31.76 0.13 1.06 225.6
fy Ny Ncred Pt AB fol
Specimen b/t Pt//Ncred AB/Ncred AB/Pt
(MPa) (kN) (kN) (kN) (kN) (MPa)
042LB30P550a 71.4 690 41.50 16.1 17.50 1.09 18.66 1.16 1.07 183.6
042LB30P550b 71.4 690 41.75 16.1 17.38 1.08 18.66 1.16 1.07 183.3
042LB30P1100a 71.4 690 41.71 8.1 10.78 1.33 10.58 1.31 0.98 192.8
042LB30P1100b 71.4 690 41.66 8.1 11.70 1.44 10.58 1.31 0.90 182.3
042LB40P550a 95.2 690 55.13 20.1 20.90 1.04 23.30 1.16 1.11 108.0
042LB40P550b 95.2 690 53.10 20.1 23.00 1.14 23.30 1.16 1.01 109.0
042LB40P1100a 95.2 690 53.08 13.9 15.00 1.08 16.81 1.21 1.12 108.7
042LB40P1100b 95.2 690 53.21 13.9 14.20 1.02 16.81 1.21 1.18 107.4
042LB40P1700a 95.2 690 52.30 8.0 8.40 1.05 8.09 1.01 0.96 116.3
042LB40P1700b 95.2 690 52.26 8.0 8.79 1.10 8.09 1.01 0.92 116.3
042LB50P550a 119.0 690 64.34 22.5 23.40 1.04 24.50 1.09 1.05 70.57
042LB50P550b 119.0 690 64.54 22.5 23.60 1.05 24.50 1.09 1.04 71.18
042LB50P1100a 119.0 690 64.72 17.5 19.70 1.13 20.21 1.15 1.03 72.24
042LB50P1100b 119.0 690 64.35 17.5 17.50 1.00 20.21 1.15 1.15 85.96
042LB50P1700a 119.0 690 63.58 10.4 11.2 1.08 10.53 1.01 0.94 73.98
042LB50P1700b 119.0 690 63.40 10.4 10.7 1.03 10.53 1.01 0.98 73.98
287
Number βc φc
Thickness Pm VP Mm VM Fm VF
of Tests
0.42 mm 0.941 0.037 1.308 0.0578 0.962 0.0161 33 4.26 1.21
0.60 mm 0.929 0.041 1.278 0.01 0.994 0.012 62 4.55 1.24
Note: 0.42 mm and 0.60 mm are the nominal base metal thicknesses; Target reliability index
β 0 = 2.5 ; Current resistance factor φ 0 = 0.85 .
Table 7.2 Summary of Resistance factors and Statistical data
288
Chapter 8: CONCLUSIONS
CHAPTER 8
CONCLUSIONS
CONTENTS
8.1 GENERAL................................................................................................................290
8.2 EXPERIMENTAL INVESTIGATIONS ...............................................................291

8.2.1 General........................................................................................................................................... 291
8.2.2 Box-shaped-section stub columns tests........................................................................................... 291
8.2.3 Box-shaped-section long columns tests .......................................................................................... 292
8.2.4 Lipped channel section columns tests............................................................................................. 293
8.2 NUMERICAL SIMULATIONS .............................................................................295
8.3 THEORETICAL STUDIES USING FINITE STRIP METHODS .....................297
8.4 DESIGN RECOMMENDATIONS ........................................................................298
8.5 FUTURE RESEARCH............................................................................................299
289
CHAPTER 8
CONCLUSIONS
8.1 GENERAL
High Strength Low-Alloy Steels, commonly designated “High Strength Steels” ,

are a specific class of steels with enhanced mechanical properties. These steels
are generally intended for applications where savings in weight can be effected
by reason of their greater strength and characterized particularly by a
considerably higher yield point than that of conventional structural steels as
shown in Fig 1.1. Steels with high yield stress usually have little or no strain
hardening in the stress-strain curve. For High Strength Steels, the fu/fy ratio is
often close to 1.0, especially for thin-cold-reduced steels with yield stress equal
to 550 MPa. Strain hardening is important in the stability of thin-walled sections
and so the High Strength Steels may have their stability significantly affected by
the lack of strain hardening. Therefore, the investigation of the stability of High
Strength Steel sections to determine the influence of the lack of strain hardening
on their capacity to resist buckling, particularly in the inelastic range, becomes
an important task for the widespread use of those steels. To achieve this goal,
three series of tests were carefully designed using High Strength Steels with
0.60 mm or 0.42 mm thicknesses. One series of tests was intended to study the
section capacity using stub columns. The other two series of tests were intended
to study the member capacity using long columns with closed or open cross-
sections. Also for this purpose, numerical simulation was carried out on the test
sections. The finite strip method, as a theoretical method, was also used to
further study the test sections. Finally, based on those investigations, some
design recommendations are proposed in this thesis for High Strength Steels
290
cold-reduced which are less than 0.9 mm thick.
8.2 EXPERIMENTAL INVESTIGATIONS
8.2.1 General
A series of compression tests was performed on thin-walled box and channel

section columns. The tests were performed on pin ended box shaped long
columns and on fixed ended stub and channel section columns. The main
purpose of the tests was to investigate local, overall and distortional stability.
8.2.2 Box-shape d-section stub columns tests
The main objective of the local stability investigation using stub columns was to
determine the adequacy of the design rules in Section 2, Elements of AS/NZS
4600 and Section B of the NAS (2001).
The tests included lipped-square and hexagonal sections, including 94 box-

shaped fix-ended stub columns. The purpose of these tests was to determine the
influence of low strain hardening of G550 steel on the compressive section
capacities of the column members.
The results of the successful stub column tests have been compared with the
design procedures in the Australian/New Zealand Standard for Cold-Formed
Steel Structures and recent (1999) Amendments to the American Iron and Steel
Institute Specification. As expected, the greatest effect of low strain hardening
was for the stockier sections where material properties play an important role.
For the more slender sections where elastic local buckling and post-local
buckling are more important, the effect of low strain hardening does not appear
291
to be as significant. This is contrary to recent design proposals in the USA

where it was proposed that the more slender sections would be more greatly
influenced by low strain hardening.
The ratios of Pt/Ns were mostly larger than 0.90 and higher than the results
based on the AISI Specification Supplement No.1(1999) and AS/NZS 4600
when the plate slenderness (b/t) was greater than 88. When the plate slenderness
(b/t) was less than 50, the formulae used in AISI Specification Supplement
No.1(1999) was slightly unconservative due to imperfection sensitivity.
8.2.3 Box-shape d-section long columns tests
The investigation of the overall stability and the interaction of local and overall
buckling was carried out using pin-ended columns to improve the design rules
in Section 3.4 of AS/NZS 4600 and Section C4 of the NAS (2001).
A series of compression tests was performed on long columns fabricated from

cold-formed high strength steel plates with nominal yield stress of 550 MPa. A
total of 28 long columns, which were made from two thicknesses of sheet steel
(0.42 mm and 0.6 mm), were tested. Three different cross-sections and shapes
were tested at a range of lengths between pinned ends.
The comparison of loads obtained experimentally with those predicted on the

basis of the AS/NZS4600 and the AISI specification including Supplement
No.1, 1999 is shown.
For columns of high slenderness, the test results agree well with curves based
on 0.75fy and Rbfy yield stress reduction. For low slenderness, the results are
close to the 0.75fy curve, but lower than the Rbfy curve by about 7%. At
292
intermediate slenderness, the test results are lower than both curves. The
difference ranges from 7% to 14%, which may be due to the interaction of local
and overall buckling.
than the design curves at intermediate columns lengths where local and overall
buckling interact.
For the sections of stockier cross-section, the AS/NZS 4600 design curve based
on 0.75fy and the AISI design curve based on a stress Rbfy are generally slightly
conservative.
Conclusions related to improvements in these design capacities are given later

in Section 8.4.
8.2.4 Lipped cha nnel section columns tests
The tests performed on lipped channel sections were performed on fixed-ended

sections mainly to investigate local buckling, distortional buckling and the
interaction between them to determine the adequacy of Clause 3.4.6 Distortional
Buckling of AS/NZS 4600 for high strength steels.
A series of lipped channel (LC) section columns fabricated from cold-reduced

high strength steel of thickness 0.42 mm with nominal yield stress 550 MPa. A
range of lengths of lipped channels with intermediate stiffeners in the web and
the flanges were tested between fixed ends to determine the strength of the
sections.
293
For the lipped channel sections, the failure resulted from local and distortional
buckling with interaction between the modes. The tests indicated that
distortional buckling and the interaction of local and distortional buckling may
have a significant effect on the strength of the sections formed from such thin
steel. The results obtained experimentally and theoretically using the effective
width method and the direct strength method are presented.
For the long column tests, the peak loads were affected significantly by the
different flange failure modes which are both flanges buckling inward (I-I
mode) and both flanges buckling outward (O-O mode). For the long columns,
the outward modes had a higher failure load than the corresponding inward
modes.
The new North American Specification (NAS) (AISI, 2001) was recently
revised to provide EWM design provisions for the determination of the effective
widths of uniformly compressed stiffened elements with multiple intermediate
stiffeners or edge stiffened elements with intermediate stiffeners. In this thesis,
the test sections adopted were C-sections with intermediate stiffeners in the web
and flanges and a simple edge stiffener on each flange. It should be noted that a
recent ballot C/S03-196A of the AISI Specification committee confirms that
Section B5.2 is applicable for edge stiffened element with only one intermediate
stiffener as used in this thesis. Therefore, the new NAS (AISI, 2001) was used
to determine the effective widths of the sections. For the long specimens, the
test results were much lower than the strengths based on the new NAS (AISI,
2001). There are two reasons for these unconservative results. Firstly,
distortional buckling is not adequately accounted for by the NAS. Secondly, the
interaction of local and distortional buckling, which causes the dip in the test
strength curve for the 800 mm and 1300 mm specimens, is also not adequately
accounted for.
294
The results based on AS/NZS 4600 Clause 3.4.6(b) and Kwon & Hancock’s
equation are approximately 32% and 22% higher than the test results of 1300
mm and 2000 mm long columns. These results indicate that for the long
columns, the interaction of local and distortional buckling has a significant
effect on the member strength. Hence, for the monosymmetric sections of the
type tested subject to distortional buckling, the interaction of local and
distortional buckling has to be taken into account when determining the design
member strength.
The results of the Direct Strength Method (DSM) are in good agreement with
the results of the tests except the results at 1300 mm length which are clearly
lower than Pnd. The results of Pnd and Pnl, which were obtained using the elastic
buckling stresses based on the Spline Finite Strip Method (SFSM) taking into
account the fixed ends boundary conditions, conform with the test results at 360
and 2000 mm lengths rather than at the intermediate lengths. As in the
preceding discussion, the failure modes at these lengths (360 and 2000 mm)
were mainly dominated by local buckling and distortional buckling respectively.
Hence, the lower test results for the 1300 mm length may be a result of the
significant interaction of L+D buckling. Therefore, the results based on the
Direct Strength Method (DSM) is unconservative for the columns at
intermediate lengths.
8.2 NUMERICAL SIMULATIONS
The finite element non-linear analysis program “ABAQUS” was used to

simulate the geometric and material nonlinear behaviour of the columns. Two
different models were established, one for simulating the pin-ended box sections
and one for the fixed-ended channel sections. The experimental data is very
295
important for two reasons. Firstly the data can be used for calibration and
implementation of the finite element non-linear analysis and, secondly, the
finite element non-linear analysis can be used to extend the range of test data,
and to investigate the effect of changing variables, such as stress-strain
characteristics, residual stresses, geometric imperfections and section geometry.
Walker’s suggested expression used to scale initial imperfections in an

ABAQUS model was generally successful. Based on Py and Pcr of such thin
sheet steel sections, this expression can provide an initial imperfection
magnitude suitable for the simulation of the test results.
The use of the finite element program ABAQUS for simulating the behaviour of
the stub columns was successful since the ABAQUS results were generally in
good agreement with experimental values. The results of ABAQUS were
sensitive to the initial imperfections for the stockier sections but not for the
slender sections. ABAQUS can be used for further work on such thin sheet steel
sections.
For the pin-ended long column, the ABAQUS results (AB) were generally in
good agreement with the test results (Pt). The difference was on average less
than 6% for all columns although the ABAQUS results were higher. Since
ABAQUS results were in reasonably good agreement with the test results,
ABAQUS can be used to simulate long column tests which could not be
performed in the available machine.
For the lipped channel (LC) sections, the simulation gives a reasonable
prediction results. The differences between ABAQUS results and Test results
are less than 4%. The comparison of failure modes based on ABAQUS plots
and test photos shows the reliability of the ABAQUS results. The stress
296
distribution shows the change of stress over the whole cross-section during the
loading and gives an explanation of the failure modes which are identical with
the observation during the tests.
8.3 THEORETICAL STUDIES USING FINITE STRIP METHODS
The semi-analytical finite strip method of buckling analysis of thin-walled

sections is a very efficient tool for investigating the buckling behaviour of cold-
formed members in compression and bending. It assumes that thin-walled
sections buckle with simply supported ends free to warp but with section
distortion prevented at the ends. The program THIN-WALL (TW) (1998) was
developed to perform a semi-analytical finite strip buckling analysis of thin-
walled sections under axial compression and bending. It can be used to
understand the general local, distortional and overall buckling behaviour of thin-
walled section. It is applied to both the open and closed sections to quantify the
different buckling stresses.
In order to account for the fixed ended boundary conditions in the tests, the
Spline Finite Strip Method (SFSM) was developed for buckling analysis by Lau
and Hancock (1986). The method uses spline functions in the longitudinal
direction and can account for a range of end conditions including fixed and free.
It is used in the thesis to accurately simulate the test boundary conditions.
The TW analysis gives the elastic local and elastic distortional buckling stresses
at given half-wavelengths, whereas the Spline Finite Strip Method (SFSM)
analysis gives the actual buckling stress of a given length of section between
fixed ends.
297
The theoretical studies also provide valuable information on the local and
distortional buckling stresses for use in the interaction buckling studies.
8.4 DESIGN RECOMMENDATIONS
For high strength steels, Clause 1.5.1.5 (b) of AS/NZS 4600 and Section A3.3.1
of the AISI Specification (1999), have a reduction to 75% of the yield stress.
From the series of stub column tests of box shaped sections, it appears from
work described in the thesis that a modified reduction factor should be used to
fully utilize the strength of material. So a modified reduction factor of 0.90 can
be used in place of the reduction factor 0.75, which is specified for G550 steel
with the thickness being less than 0.9 mm in AS/NZS 4600.
In order to take account of the interaction of local and overall buckling of long
columns a modified design method based on the column design curve of
AS/NZS 4600 is proposed. A reduced radius of gyration is used in Clause
3.3.3.2(a) of AS/NZS 4600 to replace the normal radius of gyration (Section
C3.1.2, Eq. C3.1.2-8 of the AISI Specification) when the design stress exceeds
the local buckling stress.
Based on AS/NZS 4600 Clause 3.4.6(b) and the Kwon & Hancock equation
(1992), two simple design methods are proposed to account for the interaction
of local and distortional buckling in thin sections of high strength G550 steel.
intermediate lengths, it is proposed that the interaction of local and distortional
buckling is taken into account with the distortional mode treated as an overall
mode in the Direct Strength Method (DSM).
298
8.5 FUTURE RESEARCH
• The stub column tests have been carried out in compression, the other
loading conditions such as bending may need to be studied on the section
fabricated from such thin high strength steels.
• The closed-section long column tests have indicated that the adverse
interaction may significantly reduce the member stiffness for slender
sections. Other sections, which may fail mainly by local and overall buckling
including flexural-torsional buckling, needs to studied.
• The lipped channel sections have been studied experimentally and

theoretically. The interaction of local and distortional buckling for the long
columns has been demonstrated. Therefore, further theoretical investigation
will be very helpful for understanding its mechanism and extending the use
of such thin steels.
• The spline finite strip method of buckling analysis developed previously is

applicable only to a bifurcation analysis. The extension of this method is
needed to the inelastic post-buckling range, especially to obtain the ultimate
strength of member due to the considerable post-buckling reserve of
strength.
299
Chapter 9: REFERENCE
CHAPTER 9
REFERENCE
Abdel-Sayed G. (1969). “Effective width of thin plates in compression”, Journal

of the Structural Division, Vol. 95 No. ST10, 2183-2203.
Alexander W. (1912). “Columns and struts”, E. & F. Spon, Ltd., New York.
American Iron and Steel Institute. (1997). “1996 Edition of the Specification for
the Design of Cold- Formed Steel Structural Members”, Washington, DC, USA.
American Iron and Steel Institute. (2001). “North American Specification for the
Design of Cold-Formed Steel Structural Members, Washington”, DC, USA.
American Iron and Steel Institute. (2000). "1996 Edition of the Specification for
the Design of Cold-Formed Steel Structural Members, Supplement 1, July 1999”,
Washington, DC, USA.
American Society for Testing and Materials A611 (1997). “Standard Specification
for Steel Sheet, Carbon, Cold-Rolled, Structural Quality”, Philadelphia, PA, USA.
for Steel Sheet, Zinc-Coated (Galvanized) or Zinc-Iron Alloy-Coated
(Galvannealed) by the Hot-Dip Process”, Philadelphia, PA, USA.
for Steel Sheet, 55% Aluminium-Zinc Ally-Coated by the Hot-Dip Process,
Philadelphia”, PA, USA.
Bakker M.C.M. and Pekoz T. (2001). “The finite element method for thin-
walled members—basic principles”, 3rd Conference on Thin-Walled Structures,
Ed. Zaras et al. Elsevier Science Ltd, 417-425.
Barta T.A. (1967). “On the tosional-flexural buckling of thin-walled elastic bars
wuth monosymmetric open cross-section”, Thin-Walled Structures, Ed Chilver
A.H., Chatto & Windus Ltd., 60-86
300
Bernard E.S., Bridge R.Q. and Hancock G.J. (1992). “Intermediate stiffeners in
cold-formed prefiled steel decks part 1-‘V’ shaped stiffeners”, Research Report,
R653, Department of Civil Engineering, University of Sydney, Australia
Bijlaard P.P. and Fisher G.P. (1952). “Interaction of column and local buckling
in compression members”, NACA T.N. 2640.
Bijlaard P.P. and Fisher G.P. (1953). “Column strength of H-sections and
square tubes in the post buckling range of the component plates”, NACA T.N.
2994.
Bleich F. (1952). “Buckling strength of metal structures”, McGraw Hill, New

York.
Bradford M.A. and Azhari M. (1995a). “Buckling of plates with different end
conditions using the finite strip method”, Computers & Structures, Vol. 56,
No.1. 75-83.
Bradford M.A. and Azhari M.(1995b). “Inelastic local buckling of plates and
plate assemblies using bubble functions”, Engineering Structures, Vol. 17, No.2,
95-103.
Bradford M.A. and Azhari M. (1997). “Use of bubble functions for the stability
of plates with different end conditions”, Engineering Structures, Vol. 19, No.2,
151-161.
Bradford M.A. and Hancock G.J. (1984). “Elastic interaction of local and lateral
buckling in beams, Thin-Walled Structures, Vol.2, No. 1, 1-25.
Bulson P.S. (1970). "The stability of flat plates", Chatto & Windus Ltd,.
Byskov E. and Hutchinson J.W. (1977). “Mode interaction in axially stiffened

cylindrical shells, AIAA, Vol. 15, No. 7, 941-948.
Chajes A. and Winter G. (1965). Torsional-flexural buckling of thin-walled

members, Journal of the Structural Division, ASCE, Vol. 91 No. ST4, 103-124.
Cheung Y.K. (1968a). “The finite strip method in the analysis of elastic plates
with two opposite simply supported ends”, Proceedings of the Institution of
Civil Engineers, Vol. 40 (5), 1-7.
Cheung Y.K. (1968b). “Finite strip method analysis of elastic slabs”, Journal of
the Engineering Mechanics Division, Vol. 94, No. EM6, 1365-1378.
301
Cheung Y.K and Cheung M.S. (1971). “Flexural vibrations of rectangular and
other polygonal plates”, Journal of the Engineering Mechanics Division, Vol.
97, No. EM2, 391-411.
Cheung Y.K. (1976). “Finite strip method in structural analysis”, Pergamon

Press, Inc., New York, N.Y.
Cheung Y.K. and Fan S.C. (1983). “Static analysis of right box girder bridges
by spline finite strip method”, Proceedings the Institution of Civil Engineers,
Part 2, Vol. 75, 323-331.
Cheung Y.K. and Tham L.G. (2000). “A review of the finite strip method”,
Prog. Struct. Engng. Master., Vol. 2, 369-375.
Chilver A.H. (1953). “The stability and strength of thin-walled steel struts”, The
Engineer, N5089, 196, 180-183.
Chilver A.H. (1961). “Structural problems in the use of cold-formed steel

sections”, Proceedings Institution of Civil Engineers, Vol. 20, 233-258.
Chou S.M. and Chai G.B. (1997). “Ultimate design load of thin-walled stub
columns”, Int. Journal of Computer Applications in Technology, Vol. 10 No. ½,
27-33.
Chou S.M., Chai G.B and Ling L. (2000). “Finite element technique for design
of stub columns”, Thin-Walled Structures Vol. 37(2), 97-112.
Cook R.D., Malkus D.S. and Plesha M.E. (1974). “Concepts and applications of
finite element analysis”, John Wiley & Sons Ltd.
Crisfield M.A. (1991). “Non-linear finite element analysis of solids and

structures, John Wiley & Sons Ltd.
Currie D.M. (1989). “The use of light-gauge cold-formed steelwork in

construction”, Building Research Establishment, Garston, Watford, WD2 7JR.
Davids A.J. and Hancock G.J. (1987). “Nonlinear elastic response of locally
buckled thin-walled beam-columns, Thin-Walled Structures Vol. 5(3), 211-226.
Davies J.M. and Jiang C. (1996). “Design of thin-walled columns for

distortional buckling”, Proceedings of the 2nd International Conference on
Coupled Instabilities in Metal Structures, Imperial College Press, Liege,
Belgium, 165-172.
302
Davies J.M. and Leach P. (1994). “Second order Generalised Beam Theory”, J.
Constructional Steel Research, Vol. 31, 221-241.
Dawson, R.G. and Walker A.C., (1972), “Post-buckling of geometrically

imperfect plates”, Journal of Structural Engineering, Vol.98, No. ST1, 75-94
Desmond T.P., Pekoz T. and Winter G. (1981a). “Edge stiffeners for tin-walled
members”, Journal of Structural Engineering, Vol.107, No.ST2, 329-353.
Desmond T.P., Pekoz T., and Winter G. (1981b). “Intermediate stiffeners for
thin-walled members”, Journal of Structural Engineering, Vol.107, No. ST4, 627-
648.
DeWolf J.T., Pekoz T. and Winter G. (1974). “Local and overall buckling of
cold-formed members”, Journal of the Structural Division, Vol. 100 No. ST10,
2017-2036.
Dhalla A.K., Errera S.J. and Winter G. (1971). “Connections in thin low-
ductility steels, Journal of the Structural Division, ASCE Vol. 97 No.ST 10,
2549-2566.
Dhalla A.K. and Winter G. (1974). “Steel ductility measurements”, Journal of

the Structural Division, ASCE, Vol. 100 No.ST 2, 427-444.
Dhalla A.K. and Winter G. (1974). “Suggested steel ductility requirements”,

Journal of the Structural Division, ASCE, Vol. 100 No.ST 2, 445-462.
Dhanalakshmi M. and Shanmugam N.E. (2001). “Design for openings in equal-

angle cold-formed steel stub columns”, Thin-Walled Structures, Vol. 39(2),
167-187.
Earls C.J and Galambos T.V. (1998). “Inelastic failure of high strength steel
wide flange beams under moment gradient and constant moment loading”,
Proceedings, the 1998 Annual Technical Session and Meeting of SSRC,
Altanta, USA, 219-246.
Ellingwood, B., Galambos, T.V., MacGregor, J.G. and Cornell, C.A. (1980).
“Development of a probability based load criterion for American National
Standard A58: Building code requirements for minimum design loads in
buildings and other structures”, NBS special pubilivation No. 577, National
Bureau of Standards, Washington, D.C., USA.
Galambos, T.V. (1998). “Guide to stability design criteria for metal structures”,
303
John Wiley & Sons, Inc.
Gilbert R.B. and Calladine C.R. (1974). “Interaction between the effects of local
and overall imperfections on the buckling of elastic columns”, J. Mech. Phys.
Solids, Vol. 22, 519-540.
Graves-Smith T.R. (1967). “The ultimate strength of locally buckled columns of

arbitrary length”, Symposium on Thin-Walled Steel Structures, Swansea, 35-60.
Graves-Smith T.R. and Sridharan S. (1978a). “A finite strip methods for the
post-locally buckled analysis of plate structures”, Int. J. Mech. Sci., Vol. 20(12),
833-842.
Graves-Smith T.R. and Sridharan S. (1978b). “A finite strip method for the
buckling of plate structures under arbitrary loading”, Int. J. Mech. Sci., Vol.
20(10), 685-693.
Graves-Smith T.R. and Gierlinski J.T. (1982). “Buckling of stiffened webs by

local edge loads”, Journal of the Structural Division, Vol. 108, No. ST6, 1357-
1366.
Graves-Smith T.R., Gierlinski J.T. and Walker B. (1985). “A combined finite

strip/finite elements method for analysis thin-walled structures”, Thin-Walled
Structures, Vol. 3(2), 163-180.
Gresnigt A.M. and Steenhuis C.M. (1997). “Light gauge construction”, Progress
in Structural Engineering and Materials, Vol 1, No. 1, 31-41.
Guo Y. and Linder J. (1993). “Analysis of elastic-plastic interaction bucking of

stiffened panels by spline finite strip method”, Computer & Structures, Vol. 46,
No. 3, 529-536.
Hancock G.J. (1978). “Local, distortional and later buckling of I-beam”, Journal
of the Structural Division, Vol. 104, No. ST11, 1787-1798.
Hancock G.J. (1981a). “Interaction buckling in I-section columns”, Journal of

the Structural Division, Vol. 107, No. ST1, 165-179.
Hancock G.J. (1981b). “Nonlinear analysis of thin sections in compression”,

Journal of the Structural Division, Vol. 107, No. ST3, 455-471.
Hancock G.J. (1985). “Distortional buckling of steel storage rack columns,

Journal of Structural Engineering, Vol. 111, No. 12, 2770-2783.
304
Hancock G.J., Davids A.J. and Rasmussen K.J.R. (1987). “Strength tests of
thin-walled high tensile steel columns”, Int. Conference on Steel and
Aluminium Structures, Cardiff, UK, 475-487.
Hancock G.J. (1997). “Light gauge construction”, Progress in Structural

Engineering and Materials, Vol 1, No. 1, 25-30.
Hancock G.J. and Rogers C.A. (1998). “Design of cold-formed steel structures
of high strength steel”, Journal of Constructional Steel Research, Vol. 46:1-3,
167-168 (Full paper on CD-ROM).
Hancock G.J. (1998). “Finite strip buckling and nonlinear analyses and
distortional bucking analysis of thin-walled structural members”, Coupled
Instabilities in Metal Structures, CISM379, Ed. Rondal J., Springer Wien New
York, USA, 225-289.
Hancock G.J., Murray T.M. and Ellifritt D.S. (2001). “Cold-formed steel
structures to the AISI Specification”, Marcell-Dekker, New York, USA, .
Hancock G.J. (2002). “The influence of sectional stability in engineering design

of thin-walled members”, Proceedings of the 3rd Australian Congress on
Applied Mechinics, Applied Mechanics Progress and Applications, Sydney,
Australia, 13-24
Hancock G.J. (2003). “Cold-formed steel structures”, Journal of Constructional

Steel Research, Vol. 59(4), 473-487.
Hibbitt, Karlsson & Sorensen, Inc. (1997). “ABAQUS/Standard User’s

Manual”, Ver.5.7.
Hone C.P. (1967). “Torsional-flexural buckling of axially-loaded, thin-walled,

elastic struts of open cross-section”, Thin-Walled Structures, Ed Chilver A.H.,
Chatto & Windus Ltd.
Kalyanaraman V., Pekoz T. and Winter G. (1977). “Unstiffened compression

elements”, Journal of the Structural Division, ASCE Proceedings, Vol. 103,
No.ST9, 1833-1848.
Kalyanaraman V. and Pekoz T. (1978). “Analytical study of unstiffened

compression elements”, Journal of the Structural Division, ASCE, Vol. 104,
No.ST9, 1507-1524.
Kalyanaraman V. (1979). “Local buckling of cold-formed steel members”,
305
Journal of the Structural Division, ASCE, Vol. 105, No.ST5, 813-828.
Key P.W. and Hancock G.J. (1993). “A finite strip method for the elastic-plastic
large displacement analysi of thin-walled and cold-formed steel sections”, Thin-
Walled Structures Vol. 16(1), 3-29.
Koiter W.T. (1970). “On the stability of elastic equilibrium”, PhD thesis , Delft,
1945; English translation as AFFDL Report-TR-70-25.
Kolakowski Z. (1987). “Mode interaction in thin-walled trapezoidal column

under uniform compression”, Thin-Walled Structures, Vol. 5(4), 329-342.
Kolakowski Z. (1993). “Influence of modification of boundary conditions on

load carrying capacity in thin-walled columns in the second order
approximation”, Int. J. Solids Structures, Vol. 30 No. 19, 2597-2609.
Kolakowski Z., Krolak M. and Kowal-Michalska K. (1999). “Modal interactive

buckling of thin-walled composite beam-columns regarding distorional
deformations”, Int. J., Engng. Sci., Vol. 37, 1577-1596.
Kwon Y.B. and Hancock G.J. (1989). A nonlinear elastic spline finite strip
analysis for thin-walled sections, Research Report, R607, Department of Civil
Engineering, University of Sydney, Australia.
Kwon Y.B. and Hancock G.J. (1992). “Tests of cold-formed channels with local
and distortional buckling”, Journal of the Structural Division, ASCE, Vol. 118,
No. 7, 1786-1803.
Kwon Y.B. and Hancock G.J. (1993). “Post-buckling analysis of thin-walled

channel sections undergoing local and distortional buckling”, Computers &
Structures, Vol. 49, No. 3, 507-516.
Landolfo, Mazzolani and Piluso (1999). “Local buckling of aluminium

channels: test and simulation”, Stability and ductility of Steel Structures,
SDSS’99, Ed Dubina and Ivanyi, 525-532
Lau S.C.W and Hancock G.J. (1985). “Buckling of thin walled structures by a
spline finite strip method”, Research Report, R478, Department of Civil
Lau, S.C.W and Hancock, G.J. (1986). “Buckling of thin flat-walled structures
by a spline finite strip method”, Thin-Walled Structures, Vol. 4(4), 269-294
306
Lau S.C.W and Hancock G.J. (1987). “Distortional buckling formulas for
channel columns”, Journal of Structural Engineering, Vol. 113, No. 5, 1063-
1078
Lau, S.C.W and Hancock, G.J. (1988). “Distortional buckling tests of cold-
formed channel sections”, Ninth International Specialty Conference on Cold-
formed Steel Structures”, St. Louis, Missouri, USA, 45-73.
Lau S.C.W. (1988). “Distortional buckling of thin-walled columns”, Ph.D

Thesis, University of Sydney, Australia.
Lau S.C.W and Hancock G.J. (1989). “Inelastic buckling analysis of beams,
columns and plates using the spline finite strip method”, Thin-Walled
Structures, Vol. 7(3-4), 213-238.
Lau S.C.W and Hancock G.J. (1990). “Inelastic buckling analysis of channel
columns in the distortional mode”, Thin-Walled Structures, Vol. 10(1), 59-84.
Lengyel P. and Cusens A.R. (1983). “A finite strip method for the geometrically
nonlinear analysis of plate structures”, Int. J. for Numerical methods in
Engineering, Vol. 19, 331-340.
Levy, S. Wooley, R.M., and Kroll, W.D. (1947). “Instability of simply

supported square plate with reinforced circular hole in edge compression”,
Research Paper RP1849, Vol.39, 571-577.
Little G.L. (1979). “The strenght of square steel box columns—design curves
and their theoretical basis”, The Structural Engineer, Vol. 52, No. 2, 49-61.
Loughlan J. and Rhodes J. (1979). “The interactive buckling of lipped channel

columns under concentric of eccentric loading”, Thin-Walled Structures, Int.
Conf. at the University of Strathclyde, Glasgow, 40-55.
Lundquist E.E. and Stowell E.Z. (1942). “Critical compressive stress for
outstanding flanges”, NACA R734.
Luongo A. and Pignataro M. (1988). “Multiple interaction and localization

phenomena in the post buckling of compressed thin-walled members”, AIAA
journal, Vol. 26, No.11, 1395-1402.
Macadam J.N., Brockenbrough R.L., Laboube R.A. and Pekoz T. and Schneider
E.J. (1988), Low strain hardening ductile steel cold formed members,
Proceedings of the 9th International Specialty Conference on Cold-Formed Steel
307
Structures, Missouri-Rolla, USA, 459-487
Maricic A. (1979). “Cold-formed structures of high strength steel”, Thin-Walled

Structures, International Conference at the University of Strathclyde, Glasgow,
386-397.
Mateus A.F. and Witz J.A. (2001). “A parametric study of the post-buckling
behaviour of steel plates”, Engineering Structures, Vol. 23, 172-185.
Mathcad 2000 Professional, (1999), ©1989-1999 Mathsoft Inc..
Mulligan G.P. and Pekoz T. (1984). “Locally buckled thin-walled columns”,

Journal of Structural Engineering, Vol. 110, No. 11, 2635-2654.
Pan L.C. and Yu W.W. (1988). “High strength steel members with unstiffened
compression elements”, Proceedings of the 9th Int. Specialty Conference on
Cold-Formed Steel Structures, Missouri-Rolla, Nov. 8-9, USA, 489-511.
Papangelis, J.P. and Hancock, G.J. THIN-WALL 2.0 (1998), Centre for
Advanced Structural Engineering, Department of Civil Engineering, University
of Sydney, Australia.
Parks M.B., Santaputra C. and Yu W.W. (1986), Local buckling of curved

elements, Proceedings of the 8th Int. Specialty Conference on Cold-Formed
Steel Structures, Missouri-Rolla, USA, 277-294.
Pekoz T.B. and Winter G. (1969). “Torsional-flexural buckling of thin-walled

sections under eccentric load”, Journal of the Structural Division, ASCE, Vol.
95, No. ST5, 941-963.
Pekoz T. (1986a). “Development of a unified approach to the design of cold-

formed steel members”, AISI Report SG-86-4.
Pekoz T. (1986b). “Development of a unified approach to the design of cold-

formed steel members”, Proceedings of the 8th Int. Specialty Conference on
Cold-Formed Steel Structures, Missouri-Rolla, Nov. 11-12.
Pekoz T. (1991). “Current research oh thin-walled steel and aluminum

structures at Cornell University”, Proceedings, Int. Conference on steel and
Aluminum Structures, ICSAS 91, Singapore, 25-34.
Peter A. and Kolakowski Z. (1996). “Interactive buckling of thin-walled open

elastic beam-columns with intermediate stiffeners”, Int. J. Solids Structures,
308
Vol. 33 No. 3, 315-330.
Pignataro M., Luongo A. and Rizzi N. (1985). “On the effect of the local overall
interaction on the postbuckling of uniformly compressed channels”, Thin-
walled Structures, Vol. 3(4), 293-231.
Pignataro M., Pasca M. and Franchin P. (2000). “Post buckling analysis of

corrugated panels in the presence of multiple interacting modes”, Thin-Walled
Plank R.J. and Wittrick W.H. (1974). “Buckling under combined loading of thin
flat walled structures by a complex finite strip method”, Int. Journal for
numerical Methods in Engineering, Vol. 8 No. 2, 323-339.
Priest H.M. and Gilligan J.A. (1954). “Desigh manual for high strneght steels”,
United States Steel Corporation, USA.
Prola, L.C. and Camotim D. (2002a). “On the distortional post-buckling

behaviour of cold-formed lipped channel steel columns”, Proceedings of SSRC
Annual Stability Conference, Seattle, USA, 571-590.
Prola, L.C. and Camotim D. (2002b). “On the distortional post-buckling

behaviour of cold-formed lipped channel steel beams”, The 3rd International
Conference on Advances in Steel Structures, Hong Kong, Dec. China, 331-339.
Przemieniecki J.S. (1973). “Finite element structural analysis of local

instability”, AIAA Journal, Vol. 11, No. 1, 33-39.
Rajasekaran S. and Murray D.W. (1973a). “Coupled local buckling in wide-

flange beam-column”, Journal of Structural Engineering, Vol. 99, No. ST6,
103-1023.
Rajasekaran S. and Murray D.W. (1973b). “Finite element solution of inelastic

beam equations”, Journal of Structural Engineering, Vol. 99, No. ST6, 1025-
1041.
Rasmussen K.J.R. and Hancock G.J. (1992). “Plate Slenderness limits for high
strength steel sections”, Journal of Constructional Steel Research, Vol. 23, 73-
96.
Rasmussen K.J.R. and Hancock G.J. (1995). “Tests of high strength steel
columns”, Journal of Constructional Steel Research, Vol. 34, 27-51.
309
Ravindra, M.K. and Galambos, T.V. (1978). “Load and resistance factor design
for steel”, Journal of Structural Division, Vol. 104, No. ST9, 1337-1353.
Reiss M. and Chilver A.H. (1968). “Computation of the postbuckling of Thin-

Walled sections”, Int. Ass. For Bridge and Structural Eng. Final Report, Sept. 9-
14, New York, USA, 299-309.
Rhodes J. (1981). “On the approximate prediction of elasto-plastic plate

behaviour”, Proc. Instn. Civ. Rngre., Vol. 71, Part 2, 165-183.
Riks E.(1972). “The application of Newton’s method to the problem of elastic

stability”, J. Applied Mechanics, Vol. 39, 1060-1066.
Riks E.(1979). “An incremental approach to the solution of snapping and

buckling problems”, Int. J. Solids Structures, Vol. 15, 529-551.
Rogers C.A. and Hancock G.J. (1996). “Ductility of G550 sheet steels in
tension-elongation measurements and perforated tests”, Research Report, R735,
Department of Civil Engineering, University of Sydney, Australia.
Rogers, C.A. and Hancock, G.J. (1997a). "Ductility of G550 sheet steels in
tension", Journal of Structural Engineering, ASCE, Vol. 123, No. 12, 1586-1594.
Rogers, C.A. and Hancock, G.J. (1997b). "Bolted connection tests of thin G550
and G300 sheet steels” Research Report, R749, Department of Civil
Sarawit A.T., Kim Y., Bakker M.C.M. and Pkoz T. (2001). “The finite element
method for thin-walled members—applocations”, 3rd Coference on Thin-Walled
Structures, Ed. Zaras et al, 437-448.
Schafer B.W. (1997). “Cold-formed steel behaviour and design: analytical and
numerical modeling of elements and members with longitudinal stiffeners”,
Ph.D. Thesis, Cornell University, Ithaca, New York, USA.
Schafer B.W. and Pekoz T. (1998). “Direct strength prediction of cold-formed

steel members using numerical elastic buckling solutions”, 14th International
Specialty Conference on Cold-Formed Steel Structures, St. Louis, MO., USA,
69-76.
Schafer B.W. and Pekoz T. (1999). “Laterally braced cold-formed steel flexural
members with edge stiffened flanges”, Journal of structural Engineering, Vol.
125, No. 2, 118-127.
310
Schuman L. and Back G. (1930). “Strength of rectangular flat plates under edge
compression”, Annual Report NACA, No. 356.
Seah L.K. and Khong K.H. (1990). “Lateral-torsional buckling of channel

beams”, Journal of Constructional Steel Research, Vol. 17, 265-282.
Seah L.K. and Rhodes J. (1990). “Behaviour of channel beams with unbraced
compression flange”, Proceedings of the 10th International Specialty Conference
on Cold-Formed Steel Structures, St. Louis, Missouri, USA, 335-361.
Serrette R.L and Pekoz T. (1995a). “Distortional buckling of thin-walled

beams/panels I: Theory”, Journal of Structural Engineering, Vol. 121, No. 4,
757-766.
Serrette R.L and Pekoz T. (1995b). “Distortional buckling of thin-walled

beams/panels II: Design methods”, Journal of Structural Engineering, Vol. 121,
No. 4, 767-776.
Sharp, M.L. (1966). “Longitudinal stiffeners for compression members”,

Journal of the Structural Division, Vol. 92, No. ST5, 187-211
Sivakumaran K.S. and Abdel-Rahman N. (1998). “A finite element analysis

modal for the behaviour of cold-formed steel members”, Thin-Walled
Skaloud M. and Zornerova M. (1970). “Experimental investigation into the

interaction of the buckling of compressed thin-walled columns with the
buckling of their plate elements”, ACTA Technica CASV No. 4, 389-424.
Srirdharan S. and Graves-Smith T.R. (1981). “Postbuckling analysis with finite

strip strips”, Journal of the Engineering Mechanics, Vol. 107, No. EM5, 869-
888.
Sridharan S. (1982). “A semi-analytical method for the post local torsional

buckling analysis of prismatic plate structures”, International Journal for
Numerical Methods in Engineering, Vol. 18, 1685-1697.
Sridharan S. (1983). “Doubly symmetric interactive buckling of plates

structures”, Int. J. Solids Structures, Vol. 19 No. 7, 625-641.
Sridharan S. and Ali M.A. (1985). “Interactive buckling in thin-walled beam-

columns”, Journal of Engineering Mechanics, Vol. 111, No. 12, 1470-1486.
311
Sridharan S. and Peng M. (1989). “Performance of axially compressed stiffened

panels”, Int. J. solids Structures, Vol. 25, No. 8, 879-899.
Sridharan S. and Zeggane M. (2000). “Modeling interactive buckling of plate

structures using special elements”, Journal of structural Mechanics, Vol. 126,
No. 12, 1247-1256.
Standards Australia / Standards New Zealand. (1996). "Cold-formed steel

structures - AS/NZS 4600", Sydney, NSW, Australia
Standards Australia. (1991). "Methods for tensile testing of metals-AS 1391-

1991", Sydney, NSW, Australia
Standards Australia. (1993). “Steel sheet and strip - Hot-dipped zinc-coated or

aluminium/zinc coated - AS 1397”, Sydney, NSW, Australia
Standards Australia. (1993). “Steel storage racking - AS 4084”, Sydney, NSW,

Australia
Stowell E.Z., Heimerl G.J. and Lundquist E.E. (1952). “Buckling stresses for
flat plates and sections”, Trans. Am. Soc. Civ. Eng., Vol. 117, 545-578.
Teter A. and Kolakowski Z. (1996), Interactive buckling of thin-walled open

elastic beam-columns with intermediate stiffeners, Int. J. Solids Structures, Vol.
33, No. 3, 315-330.
Thomasson, P. (1978). “Thin-walled C-shaped panels in axial compression”,

Swedish Council for Building Research, D1:1978
Timoshenko S.P. and Gere J.M. (1961). “Theory of elastic stability”, 2nd,
McGraw-Hill, New York, USA.
Toneff J.D., Stiemer S.F. and Osterrieder P. (1987). “Local and overall buckling
in thin-walled beams and columns”, Journal of Structural Engineering, Vol.
113, No. 4, 769-786.
Trahair N.S. (1993). “Flexural-torsional buckling of structures”, Chapman &

Hall, London.
Tvergaard V. (1973). “Imperfection-sensitivity of a wide integrally stiffened

panel under compression”, Int. J. Solids Structures, Vol. 9, 177-192.
van der Neut A. (1969). “The interaction of local buckling and column failure of
312
thin-walled compression members”, Proceedings, 12th International congress

applied mechanics, Springerverlag, Germany, 389-399.
van der Neut A. (1973). “The sensitivity of thin-walled compression members

to column axis imperfection”, Int. J. Solids Structures, Vol. 9, 999-1011.
Venkataramaiah K.R. and Roorda J. (1978). “Local buckling of thin walled

channels”, Proceedings of the 4th International Specialty Conference on Cold-
Formed Steel Structures, Missouri-Rolla, USA, 363-408.
Venkataramaiah, K.R. and Roorda, J. (1982). “Analysis of local plate buckling

experimental data”, 6th International Specialty Conference. on Cold-Formed
Steel Structures, St. Louis, Missouri, USA, 45-74.
Vlasov V.Z. (1961). “Thin-walled elastic beams”, 2nd Israel Program for
Scientic Translation, Jerusalem.
von Karman, T., Sechler, E.E., and Donnel, L.H. (1932). “The strength of thin
plates in compression”, Transactions ASME, Vol. 54, MP 54-5.
Walker A.C. (1966). “Local instability of thin plates and channel structs”,
Journal of the Structural Division, Vol. 92, No. ST3, 2193-2203.
Walker A.C. (1968). “Maximum loads for eccentrically loaded thin-walled

channel struts”., Int. Assoc. Brid. Struct. Eng., IABSE, Piblications 28-II, 169-
181.
Walker A.C. (1969). “The post-buckling behaviour of simply supported square

plates”, The Aero. Qusrterly, Vol. XX, 203-222.
Walker A.C. and Murray N.W. (1975). “A plastic collapse mechanism for
compressed plates”, Int. Assoc. Brid. Struct. Eng., IABSE, Piblications 35-I,
217-236.
Walker, A. C. (1975). “Design and analysis of cold-formed sections”,

International Textbook Company Limited.
Wang S. (1986). “Torsional-flexural buckling of partially closed thin-walled

columns”, Int. Assoc. Brid. Struct. Eng., IABSE, 91-98.
Waszczyszyn Z., Cichon C. and Radwanska M. (1994). “Stability of structures

by finite element methods”, Elsevier Science B.V.
313
Wilkinson T. and Hancock G.J. (1999). “Finite element analysis of plastic

bending of cold formed rectangular hollow section beams”, Research Report,
R792, Department of Civil engineering, University of Sydney, Australia (also
published as “Predicting the rotation capacity of cold-formed RHS beams using
finite element analysis" in Journal of Constructional Steel Research, Vol 58,
1455 1471, 2002).
Williams, D.G. and Wittrick W.H., (1972). “Numerical results for the initial
buckling of some stiffened panels in compression”, Aeronautical Quarterly,
Feb., 24-40
Winter, G. (1947). “Strength of thin steel compression flanges”, Transactions,

ASCE, Vol. 112, 527-554.
Wu S., Yu W.W. and LaBoube R. A. (1995). “Strength of flexural members

using structural Grade 80 of A653 and Grade E of A611 steels, First Progress
Report, Department of Civil Engineering, University of Missouri-Rolla, USA.
Wu, S., Yu, W.W, and LaBoube, R.A. (1996). “Strength of flexural members
using structural Grade 80 of A653 steel (deck panel tests)”, Second Progress
Report, Department of Civil Engineering, University of Missouri-Rolla, USA.
Yamaki N. (1959). “Post-buckling behaviour of rectangular plates with small

initial curvature loaded in edge compression”, Journal of Applied Mechanics,
ASME, 407-414
Yang Y.B. (1986). “Stability of thin-walled beams- A general theory”, Shell and
Spatial Structures”: Computational Aspects, Proceedings of the International
Symposium, Leuven, Belgium, 280-289.
Young B. and Rasmussen K.J.R. (2000). “Inelastic bifurcation of cold-formed

singly symmetric columns”, Thin-Walled Structures, Vol. 36, 213-230.
Zienkiewicz O.C. (1971). “The finite element method in engineering science”,

McGraw-Hill Ltd.
314

adt-NU20040517 15270201whole

Uploaded by

Copyright:

Available Formats

adt-NU20040517 15270201whole

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

adt-NU20040517 15270201whole

Uploaded by

Copyright:

Available Formats

COMPRESSION STABILITY

OF HIGH STRENGTH STEEL SECTIONS WITH

Thesis presented for the Degree of

School of Civil and Mining Engineering

This thesis describes three series of compression tests performed on box-section

investigated. The effect of changing variables, such as geometric imperfections

In accordance with The Bye-Laws of the University of Sydney governing the

1. Compression Tests of Box-shaped Cold-Reduced High Strength Steel

2. Stability and Ductility of Thin-High Strength G550 Steel Members and

on Thin-Walled Structures, Kracow, Poland, (published as Thin-Walled

3. Compression Tests of Cold-Reduced High Strength Steel Stub Columns,

4. Compression Tests of Cold-Reduced High Strength Steel Long Columns,.

5. The Behaviour of High Strength G550 Steel Sections as Used in Residential

6. Stability of High Strength G550 Steel Compression Members, Keynote

7. Compression Tests of Cold-Reduced High Strength Steel Channel Columns

8. Compression Tests of Cold-Reduced High Strength Steel Stub Columns,

9. Compression Tests of Cold-Reduced High Strength Steel Long Columns,

10. Compression Tests of Cold-Reduced High Strength Steel Channel Columns

11. Compression Tests of Cold-Reduced High Strength Steel Stub Columns,

12. Compression Tests of Cold-Reduced High Strength Steel Long Columns,

13. Compression Tests of Cold-Reduced High Strength Steel Channel Columns

14. Developments in Design for Distortional Buckling of Thin-Walled

15. Numerical Simulations of High Strength Steel Box-Shaped & Channel

I wish to express my sincerest thanks to my supervisor Professor Gregory J.

The scholarship provided by a joint Department of Civil Engineering and Centre

This thesis forms part of an ARC research project entitled “Compression

My thanks go to my friends and colleagues in the Department of Civil

I wish to thank my parents and my brothers for their support, encouragement

1.1 STATEMENT OF THE PROBLEM .........................................................................2

1.2 OBJECTIVES OF THIS THESIS..............................................................................4

1.3 OUTLINE OF THIS THESIS.....................................................................................5

LITERATURE REVIEW .....................................................................................................12

2.1 RESEARCH ON HIGH STRENGTH STEEL .......................................................13

2.2 LOCAL BUCKLING, POST-BUCKLING OF PLATES AND INTERACTION

2.3 BUCKLING OF COMPRESSION MEMBERS.....................................................25

2.4 INTERACTION OF MODES...................................................................................35

2.4.2 Interaction of local and overall buckling ......................................................................................... 35

2.5 NUMERICAL METHODS .......................................................................................44

2.6 SUMMARY ................................................................................................................53

2.7 FIGURE ......................................................................................................................56

STRENGTH TESTS OF COLD-FORMED BOX-SHAPED-SECTION STUB

3.2 TEST SPECIMENS ...................................................................................................58

3.3 MATERIAL PROPERTIES .....................................................................................61

3.4 STUB COLUMN TESTS ..........................................................................................65

3.5 ANALYSES ................................................................................................................67

3.6 SUMMARY ................................................................................................................75

STRENGTH TESTS OF COLD-FORMED BOX-SHAPED-SECTION LONG

4.2 TEST SPECIMENS .................................................................................................107

4.3 LONG COLUMN TESTS .......................................................................................111

4.4 ANALYSES ..............................................................................................................114

4.5 SUMMARY ..............................................................................................................116

APPENDIX 4.1 ................................................................................................................139

STRENGTH TESTS OF COLD-FORMED LIPPED CHANNEL SECTION

5.2 TEST SPECIMENS .................................................................................................142

5.3 COLUMN TESTS....................................................................................................147

5.4 TEST RESULTS COMPARED WITH DESIGN METHODS ...........................153

5.5 SUMMARY ..............................................................................................................160

NUMERICAL SIMULATIONS .........................................................................................192

6.1 GENERAL DESCRIPTION OF ABAQUS...........................................................193