Thesis

STRUCTURAL BEHAVIOUR AND DESIGN
OF COLD-FORMED STEEL BEAMS AT

ELEVATED TEMPERATURES
NIROSHA DOLAMUNE KANKANAMGE

Structural Behaviour and Design
of Cold-formed Steel Beams at
Elevated Temperatures
By
Nirosha Dolamune Kankanamge
School of Urban Developments

Queensland University of Technology
A THESIS SUBMITTED TO THE SCHOOL OF URBAN DEVELOPMENTS

QUEENSLAND UNIVERSITY OF TECHNOLOGY IN PARTIAL
FULFILLMENT OF REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY
MAY 2010
KEYWORDS
Light gauge cold-formed steel, lateral-torsional buckling, section moment capacity,

flexural members, ambient temperature, elevated temperatures, tensile coupon tests,
reduced yield strength, reduced elastic modulus, stress-strain model, fire safety
design, finite element analysis.
Structural Behaviour and Design of Cold-formed Steel Beams at Elevated Temperatures iii
Structural Behaviour and Design of Cold-formed Steel Beams at Elevated Temperatures iv
ABSTRACT
Cold-formed steel members are extensively used in the building construction
industry, especially in residential, commercial and industrial buildings. In recent
times, fire safety has become important in structural design due to increased fire
damage to properties and loss of lives. However, past research into the fire
performance of cold-formed steel members has been limited, and was confined to
compression members. Therefore a research project was undertaken to investigate
the structural behaviour of compact cold-formed steel lipped channel beams subject
to inelastic local buckling and yielding, and lateral-torsional buckling effects under
simulated fire conditions and associated section and member moment capacities.
In the first phase of this research, an experimental study based on tensile coupon
tests was undertaken to obtain the mechanical properties of elastic modulus and yield
strength and the stress-strain relationship of cold-formed steels at uniform ambient
and elevated temperatures up to 700oC. The mechanical properties deteriorated with
increasing temperature and are likely to reduce the strength of cold-formed beams
under fire conditions. Predictive equations were developed for yield strength and
elastic modulus reduction factors while a modification was proposed for the stress-
strain model at elevated temperatures. These results were used in the numerical
modelling phases investigating the section and member moment capacities.
The second phase of this research involved the development and validation of two
finite element models to simulate the behaviour of compact cold-formed steel lipped
channel beams subject to local buckling and yielding, and lateral-torsional buckling
effects. Both models were first validated for elastic buckling. Lateral-torsional
buckling tests of compact lipped channel beams were conducted at ambient
temperature in order to validate the finite element model in predicting the non-linear
ultimate strength behaviour. The results from this experimental study did not agree
well with those from the developed experimental finite element model due to some
unavoidable problems with testing. However, it highlighted the importance of
magnitude and direction of initial geometric imperfection as well as the failure
direction, and thus led to further enhancement of the finite element model. The finite
element model for lateral-torsional buckling was then validated using the available
experimental and numerical ultimate moment capacity results from past research.
Structural Behaviour and Design of Cold-formed Steel Beams at Elevated Temperatures v

The third phase based on the validated finite element models included detailed
parametric studies of section and member moment capacities of compact lipped
channel beams at ambient temperature, and provided the basis for similar studies at
elevated temperatures. The results showed the existence of inelastic reserve capacity
for compact cold-formed steel beams at ambient temperature. However, full plastic
capacity was not achieved by the mono-symmetric cold-formed steel beams. Suitable
recommendations were made in relation to the accuracy and suitability of current
design rules for section moment capacity. Comparison of member capacity results
from finite element analyses with current design rules showed that they do not give
accurate predictions of lateral-torsional buckling capacities at ambient temperature
and hence new design rules were developed.
The fourth phase of this research investigated the section and member moment
capacities of compact lipped channel beams at uniform elevated temperatures based
on detailed parametric studies using the validated finite element models. The results
showed the existence of inelastic reserve capacity at elevated temperatures. Suitable
recommendations were made in relation to the accuracy and suitability of current
design rules for section moment capacity in fire design codes, ambient temperature
design codes as well as those proposed by other researchers. The results showed that
lateral-torsional buckling capacities are dependent on the ratio of yield strength and
elasticity modulus reduction factors and the level of non-linearity in the stress-strain
curves at elevated temperatures in addition to the temperature. Current design rules
do not include the effects of non-linear stress-strain relationship and therefore their
predictions were found to be inaccurate. Therefore a new design rule that uses a non-
linearity factor, which is defined as the ratio of the limit of proportionality to the
yield stress at a given temperature, was developed for cold-formed steel beams
subject to lateral-torsional buckling at elevated temperatures.
This thesis presents the details and results of the experimental and numerical studies
conducted in this research including a comparison of results with predictions using
available design rules. It also presents the recommendations made regarding the
accuracy of current design rules as well as the new developed design rules for cold-
formed steel beams both at ambient and elevated temperatures.
Structural Behaviour and Design of Cold-formed Steel Beams at Elevated Temperatures vi

TABLE OF CONTENTS
Keywords ................................................................................................................ iii

Abstract .................................................................................................................... v
Table of Contents .................................................................................................. vii
List of Figures ........................................................................................................ xii
List of Tables ........................................................................................................ xix
List of Symbols and Abbreviations................................................................... xxiii
Statement of Original Authorship ................................................................... xxvii
Acknowledgments ............................................................................................ xxviii
Dedication ........................................................................................................... xxix
1 Introduction ............................................................................................ 1-1

1.1 General ................................................................................................... 1-1
1.2 Research Problem................................................................................... 1-5
1.3 Research Objectives and Scope ............................................................. 1-8
1.4 Scope ...................................................................................................... 1-9
1.5 Research Methodology......................................................................... 1-10
1.6 Contents of the Thesis .......................................................................... 1-11
2 Literature Review .................................................................................. 2-1

2.1 General ..................................................................................................... 2-1
2.2 Characteristics and Mechanical Properties of Cold-Formed Steel Sections at
Ambient Temperature ............................................................................ 2-1
2.3 Flexural Behaviour of Cold-Formed Steel Beams ................................... 2-4
2.3.1 General Buckling Modes ................................................................ 2-4
2.4 Previous Research on Cold-Formed Steel Flexural Members at Ambient
Temperature ........................................................................................... 2-9
2.5 Current Design Rules for Cold-Formed Steel Beams at Ambient
Temperature ......................................................................................... 2-21
2.5.1 General ......................................................................................... 2-21
2.5.2 AS/NZS 4600 (SA, 2005)............................................................. 2-22
2.5.3 Eurocode 3 Part 1-3 (ECS, 2006) ................................................. 2-24
2.5.4 BS 5950-Part 5 (BSI, 1998).......................................................... 2-26
2.5.5 Direct Strength Method ................................................................ 2-28
2.6 Fire Safety .............................................................................................. 2-30
2.6.1 General ......................................................................................... 2-30
2.6.2 Fire Behaviour .............................................................................. 2-30
Structural Behaviour and Design of Cold-formed Steel Beams at Elevated Temperatures vii
2.6.3 Design for Fire Safety .................................................................. 2-31
2.6.4 Effect of Fire on Steel Structures ................................................. 2-32
2.7 Mechanical Properties of Cold-Formed Steels at Elevated
Temperatures ......................................................................................... 2-32
2.8 Flexural Behaviour of Steel Beams at Elevated Temperatures ............. 2-42
2.9 Current Design Rules for Cold-Formed and Hot-Rolled Steel Beams at
Elevated Temperatures......................................................................... 2-48
2.9.1 BS 5950 Part 8 (BSI, 1990) .......................................................... 2-49
2.9.2 Eurocode 3 Part 1-2 (ECS, 2005) ................................................. 2-51
2.9.3 Alternative Approach Proposed by Dharma and Tan (2007) ....... 2-54
2.10 Analysis Methods .................................................................................. 2-55
2.10.1 CU-FSM Program ........................................................................ 2-55
2.10.2 Finite Element Analysis (ABAQUS) ........................................... 2-56
2.11 Literature Review Findings ................................................................... 2-72
3 Mechanical Properties of Cold-formed Steels at Elevated

Temperatures ......................................................................................... 3-1
3.1 Introduction.............................................................................................. 3-1
3.2 Experimental Investigation ...................................................................... 3-2
3.2.1 Test Method.................................................................................... 3-2
3.2.2 Test Specimen ................................................................................ 3-3
3.2.3 Test Rig and Procedure .................................................................. 3-6
3.2.4 Mechanical Properties at Ambient and Elevated Temperatures .. 3-15
3.2.5 Comparison of Measured Yield Strength and Elastic
Modulus with Those Available in Design Codes for Cold-formed
Steel Structures ............................................................................. 3-32
3.2.6 Comparison of Yield Strength and Elastic Modulus Results
with Available Research Results .................................................. 3-34
3.2.7 Comparison of Reduction Factors for Yield Strength and
Elastic Modulus for Hot-rolled and Cold-formed Steels ............. 3-39
3.2.8 Stress-Strain Model ...................................................................... 3-41
3.3 Predictive Equations for Mechanical Properties.................................... 3-45
3.3.1 Yield Strength .............................................................................. 3-45
3.3.2 Elastic modulus ............................................................................ 3-52
3.3.3 Stress-strain Curves ...................................................................... 3-54
3.4 Conclusion ............................................................................................. 3-57
4 Experimental Investigation of Cold-formed Steel

Flexural Members .................................................................................. 4-1
4.1 Section Geometry and Specimen Sizes ................................................... 4-1
Structural Behaviour and Design of Cold-formed Steel Beams at Elevated Temperatures viii
4.2 Measurement of Initial Geometric Imperfections .................................... 4-4
4.3 Lateral-torsional Buckling Tests .............................................................. 4-7
4.3.1 Support System ............................................................................... 4-9
4.3.2 Loading System ............................................................................ 4-10
4.3.3 Measuring system ......................................................................... 4-11
4.3.4 Test Procedure .............................................................................. 4-13
4.4 Experimental Results ............................................................................. 4-15
4.5 Discussion of Test Results ..................................................................... 4-19
4.6 Conclusions ............................................................................................ 4-23
5 Finite Element Modeling of Cold-formed Steel

Flexural Members .................................................................................. 5-1
5.1 Development of Finite Element Models .................................................. 5-2
5.1.1 Element Type.................................................................................. 5-4
5.1.2 Finite Element Mesh....................................................................... 5-5
5.1.3 Symmetry ....................................................................................... 5-7
5.1.4 Ideal Model for Member Moment Capacity
(Lateral-torsional Buckling) ........................................................... 5-8
5.1.5 Ideal Model for Section Moment Capacity (Local Buckling) ...... 5-10
5.1.6 Experimental Finite Element Model ............................................ 5-12
5.1.7 Material Model and Properties ..................................................... 5-14
5.1.8 Geometric Imperfections .............................................................. 5-18
5.1.9 Residual Stresses .......................................................................... 5-27
5.2 Validation of Ideal Finite Element Models ............................................ 5-31
5.2.1 Elastic Buckling Analysis............................................................. 5-32
5.2.2 Non-linear Analysis ...................................................................... 5-36
5.3 Conclusion ............................................................................................. 5-50
6 Behaviour and Design of Cold-formed Steel Beams Subject to

Local Buckling and Lateral-torsional Buckling at Ambient
Temperature ........................................................................................... 6-1
6.1 Details of Finite Element Model Used in the Parametric Study .............. 6-2
6.2 Selection of Member Sizes for the Parametric Study .............................. 6-4
6.3 Section Moment Capacity ........................................................................ 6-6
6.3.1 Comparison of Ultimate Capacities from Finite Element Analysis and
Current Design Rules ................................................................................... 6-12
6.4 Member Moment Capacity of Lipped Channel Beams Subject to Lateral-
torsional Buckling ................................................................................ 6-20
6.4.1 Discussion..................................................................................... 6-23
Structural Behaviour and Design of Cold-formed Steel Beams at Elevated Temperatures ix

6.4.2 Comparison of Ultimate Moment Capacities from FEA with
Predictions from the Current Design Rules at
Ambient Temperature .................................................................. 6-34
6.5 Development of Design Rules ............................................................... 6-49
6.5.1 AS/NZS 4600 (SA, 2005) Design Rules ............................................... 6-50
6.6 Capacity Reduction Factor Calculations ............................................... 6-59
6.7 Conclusion ............................................................................................. 6-62
7 Behaviour and Design of Cold-formed Steel Beams Subject to

Local Buckling and Lateral-Torsional Buckling at Elevated
Temperatures ......................................................................................... 7-1
7.1 Details of Finite Element model used in the Parametric Study ............... 7-2
7.2 Selection of Member Sizes for the Parametric Study .............................. 7-5
7.3 Section Moment Capacity of Cold-formed Lipped Channel Beams at
Elevated Temperatures........................................................................... 7-6
7.3.1 Comparison of Ultimate Capacities from Finite Element Analyses and
Current Design Rules ................................................................... 7-10
7.3.2 Compressive Strain ...................................................................... 7-14
7.3.3 Use of Higher Strain Based Yield Strength ................................. 7-15
7.4 Member Moment Capacity of Lipped Channel Beams Subject to
Lateral-torsional Buckling ..................................................................... 7-19
7.4.1 Effect of Elevated Temperatures .................................................. 7-29
7.4.2 Effect of Steel Grade .................................................................... 7-33
7.4.3 Effect of Non-linear Stress-strain curves ..................................... 7-34
7.4.4 Effect of ky,T/kE,T Ratio ................................................................. 7-40
7.4.5 Effect of Residual Stresses ........................................................... 7-43
7.5 Comparison of Ultimate Member Moment Capacities with Predictions
from the Current Design Rules ............................................................ 7-45
7.5.1 Eurocode 3 Part 1.3 / Eurocode 3 Part 1.1 ................................... 7-46
7.5.2 New Design Method Proposed in Chapter 6 ................................ 7-52
7.5.3 Eurocode 3 Part 1.2 ...................................................................... 7-54
7.5.4 Design Methods Proposed by Dharma and Tan (2007) ............... 7-58
7.5.5 New Proposal ............................................................................... 7-62
7.6 Conclusion ............................................................................................. 7-68
8 Conclusions and Recommendations ..................................................... 8-1

8.1 Mechanical Properties of Cold-formed Steels at
Elevated Temperatures ............................................................................ 8-3
8.2 Section and Member Moment Capacities of Cold-formed
Steel Lipped Channel Beams at Ambient Temperature .......................... 8-6
Structural Behaviour and Design of Cold-formed Steel Beams at Elevated Temperatures x

8.3 Section and Member Moment Capacities of Cold-formed
Steel Beams at Elevated Temperatures .................................................... 8-9
8.4 Future Research ..................................................................................... 8-13
Appendix A Tensile Coupon Tests Results at Ambient and

Elevated Temperatures .................................................................. A-1
Appendix B Sample Calculations for Section Moment Capacity ...................... B-1
Appendix C Sample Calculations for Lateral-torsional
Buckling Capacity ......................................................................... C-1
Appendix D Sample Calculations for the Plotting of Lateral-torsional
Buckling Capacity Curves in Non-dimensional Format .............. D-1
Appendix E Sample Calculations for Capacity Reduction Factor...................... E-1
Appendix F Finite Element Analysis Results of Section Moment Capacities
and Predictions using Available Design Methods .......................... F-1
Appendix G Finite Element Analysis Results of Member Moment
Capacities and Predictions using Available Design Methods ....... G-1
REFERENCES ................................................................................................... R-1
Structural Behaviour and Design of Cold-formed Steel Beams at Elevated Temperatures xi

LIST OF FIGURES
Figure 1.1: Building Composed of Cold-Formed Steel Sections ..................... 1-1

Figure 1.2: Manufacturing Process of Cold-Rolled Steel Sections .................. 1-2
Figure 1.3: Press Braking (Karren, 1967) ......................................................... 1-2
Figure 1.4: Various Shapes of Cold-Formed Steel Sections (Yu, 2000) .......... 1-3
Figure 1.5: Fire in a Steel Structure .................................................................. 1-4
Figure 2.1: Effect of Cold Work on the Mechanical Properties of
Cold-Formed Steel Unstiffened Channel Sections
(Karren and Winter, 1967) ............................................................. 2-2
Figure 2.2: Effects of Strain Hardening and Strain Aging on Stress-Strain
Characteristics (Chajes et al., 1963) ............................................... 2-3
Figure 2.3: Stress-Strain Curves (Pi and Trahair, 1995)................................... 2-3
Figure 2.4: Buckling Modes of Lipped Channel Beams in Bending ................ 2-4
Figure 2.5: Local Buckling of Stiffened Top Flange of Hat-Shaped Beams .... 2-5
Figure 2.6: Distortional Buckling ..................................................................... 2-7
Figure 2.7: Lateral Distortional Buckling ......................................................... 2-8
Figure 2.8: Lateral Torsional Buckling............................................................. 2-9
Figure 2.9: Eccentrically Loaded Channel Beam (Hancock, 2003) ............... 2-12
Figure 2.10: Effects of Initial Imperfection and Twist (Pi et al., 1998) ........... 2-14
Figure 2.11: Effects of Web Distortion (Pi et al., 1998) ................................... 2-15
Figure 2.12: Effect of Slenderness on the Buckling Resistance of Beams ....... 2-16
Figure 2.13: Lateral Buckling Strength (Trahair, 1998) ................................... 2-17
Figure 2.14: Comparison of Lateral-torsional buckling strengths of RHS
and I-section Beams (Pi and Trahair, 1995) ................................. 2-20
Figure 2.15: Stress Distribution in Stiffened Compression Elements .............. 2-22
Figure 2.16: Time-Temperature Curve for Typical Fire Development
Situations ...................................................................................... 2-30
Figure 2.17: Stress-Strain Relationship for Steel at Elevated Temperatures .... 2-33
Figure 2.18: Simply Supported Beam (Real et al., 2004b) ............................... 2-44
Figure 2.19: Buckling Load Factor Versus Half-Wave Length........................ 2-56
Figure 2.20: Finite Element Modelling of Local Buckling Tests ..................... 2-58
Figure 2.21: Details of Finite Element Modelling (Yu and Schafer, 2004) ..... 2-58
Figure 2.22: Boundary Conditions (Yin and Wang, 2003)............................... 2-59
Figure 2.23: Plan View of the Model (Mahaarachchi and Mahendran, 2005a) 2-60
Figure 2.24: End Support Boundary Conditions .............................................. 2-61
Structural Behaviour and Design of Cold-formed Steel Beams at Elevated Temperatures xii
Figure 2.25: Axis System and Different Types of End Nodes.......................... 2-62
Figure 2.26: Residual Stress Distributions (Pi et al., 1998) .............................. 2-64
Figure 2.27: Residual Stresses in Cold-Formed Steel C-Section...................... 2-64
Figure 2.28: Assumed Flexural Residual Stress Distribution
(Ranawaka and Mahendran, 2009b) ............................................. 2-65
Figure 2.29: Definition of Geometric Imperfection (Schafer and Pekoz, 1998)2-66
Figure 2.30: Simply Supported Beam with Forks Supports (Real et al., 2003) 2-68
Figure 2.31: Experimental Set up (Real et al., 2003)........................................ 2-68
Figure 2.32: Measured Mid Span Displacement (Real et al., 2003) ................. 2-68
Figure 2.33: Test Arrangement Used by Put et al. (1999a) .............................. 2-69
Figure 2.34: Test Arrangement (Yu et al, 2002) ............................................... 2-70
Figure 2.35: Support System Used by Mahaarachchi and
Mahendran (2005a) ...................................................................... 2-71
Figure 2.36: Loading System Used by Mahaarachchi and
Mahendran (2005a) ...................................................................... 2-71
Figure 3.1: Tensile Test Coupons and Dimensions .......................................... 3-4
Figure 3.2: Test Rig .......................................................................................... 3-6
Figure 3.3: Details of Test Arrangement .......................................................... 3-8
Figure 3.4: Laser Speckle Extensometer......................................................... 3-10
Figure 3.5: Strain Measurements .................................................................... 3-10
Figure 3.6: Speckle Output for Strain Measurements ..................................... 3-12
Figure 3.7: Overall Test Set-up ....................................................................... 3-13
Figure 3.8: Stress-strain Curve for 1.9 mm thick G450 Steel at 20oC ........... 3-14
Figure 3.9: Stress-strain Curves ...................................................................... 3-16
Figure 3.10: Stress-strain Curves of Low and High Strength Steels
at Ambient and Elevated Temperatures ....................................... 3-18
Figure 3.11: Stress-strain Curves at Different Temperatures ........................... 3-19
Figure 3.12: Yield Strength Reduction Factors versus Temperature for
Different Steel Grades and Thicknesses ....................................... 3-20
Figure 3.13: Yield Strength Reduction Factor versus Temperature for
Different Strain Levels ................................................................. 3-24
Figure 3.14: Elastic Modulus versus Temperature for Different Steel
Grades and Thicknesses ............................................................... 3-26
Figure 3.15: Ultimate Strength Reduction Factors versus Temperature
for Different Steel Grades and Thickness .................................... 3-27
Figure 3.16: Stress-strain Curves at Various Temperatures ............................. 3-29
Structural Behaviour and Design of Cold-formed Steel Beams at Elevated Temperatures xiii
Figure 3.17: Failure Modes of Tensile Specimens at Elevated Temperatures . 3-31
Figure 3.18: Comparison of Yield Strength Reduction Factors at
Elevated Temperatures with Current Design Standards .............. 3-33
Figure 3.19: Comparison of Elastic Modulus Reduction Factors
at Elevated Temperatures with Current Design Standards........... 3-34
Figure 3.20: Comparison of Yield Strength Reduction Factors with
Those Obtained by Other Researchers ......................................... 3-35
Figure 3.21: Comparison of Yield Strength Reduction Factors ....................... 3-37
Figure 3.22: Comparison of Elastic Modulus Reduction Factors with
Those Obtained by Other Researchers ......................................... 3-38
Figure 3.23: Comparison of Yield Strength Reduction Factors with
Those Given in Hot-rolled Steel Design Standards ..................... 3-40
Figure 3.24: Comparison of Elastic Modulus Reduction Factors with
Those Given in Hot-rolled Steel Design Standards ..................... 3-40
Figure 3.25: Comparison of Predicted Stress-strain Curves by
Ranawaka (2006) with Test Results ............................................. 3-43
Figure 3.26: Comparison of Predicted Stress-strain Curves by
Chen and Young (2007) with Test Results .................................. 3-44
Figure 3.27: Comparison of Predicted Yield Strength Reduction
Factors from Equation 3.2 with Test Results .............................. 3-47
Factors from Equation 3.3 with Test Results ............................... 3-49
Figure 3.31: Comparison of Predicted Elastic Modulus Reduction
Factors with Test Results ............................................................. 3-54
Figure 3.32: Stress-strain Curves from Equations 3.5 and 3.6 with ................. 3-56
Figure 4.2: Buckling Plot of Selected Channel Section from Thin-Wall ......... 4-2
Figure 4.3: Negative and Positive Geometric Imperfections ............................ 4-4
Figure 4.4: Imperfection Measuring Device ..................................................... 4-5
Figure 4.5: Measured Initial Geometric Imperfections .................................... 4-6
Figure 4.6: Schematic View of Test Arrangement and Bending
Moment Distribution in the Test Beam .......................................... 4-7
Figure 4.7: Test Rig .......................................................................................... 4-8
Structural Behaviour and Design of Cold-formed Steel Beams at Elevated Temperatures xiv
Figure 4.8: Support System ............................................................................... 4-9
Figure 4.9: Loading System and its Components ........................................... 4-11
Figure 4.10: EDCAR (Experimental Data Collection And Recorder).............. 4-12
Figure 4.11: Measuring System ........................................................................ 4-12
Figure 4.12: End Connection of Test Beam ...................................................... 4-13
Figure 4.13: Connection at Loading Point ........................................................ 4-14
Figure 4.14: Cold-formed Steel Lipped Channel Beam at Failure ................... 4-16
Figure 4.15: Possible Failure Directions of Lipped Channel Beam ................. 4-16
Figure 4.16: Moment-Deflection Curves .......................................................... 4-19
Figure 5.1: Physical Arrangements of Finite Element Models ......................... 5-3
Figure 5.2: Lipped Channel Beam Cross-section Used in Finite
Element Modelling ......................................................................... 5-4
Figure 5.3: Comparison of Elastic Lateral-torsional Buckling
Moments Obtained from FEA and Thin-Wall................................ 5-5
Figure 5.4: Typical Finite Element Mesh for Lipped Channel Beams ............. 5-6
Figure 5.5: Structural Arrangement and Applied Boundary Conditions .......... 5-8
Figure 5.6: Ideal Model for Member Moment Capacity
(lateral-torsional buckling) ............................................................. 5-9
Figure 5.7: Ideal Model for Section Moment Capacity (local buckling) ........ 5-12
Figure 5.8: Experimental Finite Element Model ............................................ 5-13
Figure 5.9: Material Models............................................................................ 5-15
Figure 5.10: Different Imperfection Types in a Lipped Channel Beam ........... 5-19
Figure 5.11: Positive and Negative Geometric Imperfections .......................... 5-19
Figure 5.12: Effect of Imperfection Magnitude ................................................ 5-23
Figure 5.13: Effects of Negative and Positive Geometric Imperfections
on the Ultimate Moment Capacity of Lipped Channel Beams..... 5-24
Figure 5.15: Effect of Imperfection Direction .................................................. 5-26
Figure 5.16: Flexural Residual Stress Distribution ........................................... 5-27
Figure 5.17: Residual Stress Contours for Lipped Channel Beams .................. 5-28
Figure 5.18: Effect of Residual Stresses on the Ultimate Moment
Capacity of Lipped Channel Beams ............................................. 5-29
Figure 5.19: Comparison of Moment-Deflection Curves ................................. 5-39
Figure 5.21: Dimensions of Lipped Channel Beams Used by .......................... 5-42
Figure 5.22: Stress-Strain Curves Adopted by Pi et al. (1997) ......................... 5-42
Figure 5.23: Residual Stress Distributions used by Pi et al. (1997).................. 5-43
Figure 5.24: Comparison of Nonlinear Finite Element Analysis Results
Structural Behaviour and Design of Cold-formed Steel Beams at Elevated Temperatures xv

from the Developed Ideal Model with those of Pi et al. (1997) ... 5-44
Figure 5.25: Dimensions of Hot-rolled Channel Beam Used by
Poutre et al. (2000) and Snijder et al. (2002) ............................... 5-45
Figure 5.26: Stress-Strain Curves Used by Snijder et al. (2002) ...................... 5-45
Figure 5.27: Shape of Initial Imperfection in Test Beams ................................ 5-46
Figure 5.28: Finite Element Models used by Snijder et al. (2002) and
Developed Model in this Research............................................... 5-47
Figure 5.29: Load versus Deflection Plots........................................................ 5-49
Figure 6.1: Elastic Buckling and Ultimate Failure Modes ............................... 6-9
Figure 6.2: Von Mises Stress Distribution over the Cross-section ................. 6-12
Figure 6.3: Strain Distribution ........................................................................ 6-15
Figure 6.4: Von Mises Strain Variation along the Compression Flange – ..... 6-16
Figure 6.5: Longitudinal Membrane Strain Variation with the Applied
Moment at the Point of Maximum Strain –
G250-1.95-125-50-15 ................................................................... 6-16
Figure 6.6: Lateral-Torsional Buckling Failures of Cold-formed Steel
Lipped Channel Beams (G250-1.95-100-60-15-5000) ................ 6-20
Figure 6.7: Moment-Deflection Curves .......................................................... 6-22
Figure 6.8: Ultimate Moment Capacity Curves Based on Finite Element
Analyses ....................................................................................... 6-23
Figure 6.9: Effect of d/b ratio on Ultimate Moment Capacity........................ 6-27
Figure 6.10: Effect of Lip Size on Ultimate Moment Capacity........................ 6-28
Figure 6.11: Effect of Steel Thickness.............................................................. 6-29
Figure 6.12: Effect of Mechanical Properties on Ultimate Moment Capacity . 6-30
Figure 6.13: Effect of Imperfection Magnitude on Ultimate Moment
Capacity ........................................................................................ 6-32
Figure 6.14: Ultimate Moment Capacity Increase as a Function of
Initial Geometric Imperfection Magnitude .................................. 6-33
Figure 6.15: Comparison of Moment Capacity Design Curves in
Cold-formed Steel Design Codes ................................................. 6-35
Figure 6.16: Comparison of FEA Results with AS/NZS 4600 Moment
Capacity Curve for G250 and G450 Steel Beams ........................ 6-42
Figure 6.17: Comparison of FEA Results for Beams with Varying Initial
Imperfections with AS/NZS 4600 Moment Capacity Curve ....... 6-44
Figure 6.19: Comparison of FEA Results with Different Buckling
Curves of Eurocode 3 Part 1.1 (ECS, 2005) ................................ 6-47
Figure 6.20: Comparison of Buckling Curve ‘a’ of Eurocode 3 Part 1.3
Structural Behaviour and Design of Cold-formed Steel Beams at Elevated Temperatures xvi
(ECS, 2006) with FEA Results..................................................... 6-47
Figure 6.21: AS/NZS 4600 Moment Capacity Curve for Lateral-torsional
Buckling ....................................................................................... 6-51
Figure 6.22: Comparison of New Design Equations with FEA Results –
Option 1 ........................................................................................ 6-54
Option 2 ........................................................................................ 6-55
Figure 6.24: Comparison of FEA Results with New Design Beam Curve –
Option 3 ........................................................................................ 6-57
Figure 7.1: Elastic Buckling and Ultimate Failure Modes at Elevated
Temperatures – G250-1.55-120-45-15 ........................................... 7-9
Figure 7.2: Longitudinal Membrane Strain Variation with Applied
Moment at the Point of Maximum Strain for Different
Temperatures G250-1.55-120-45-15 Beam.................................. 7-14
Figure 7.3: Elastic Lateral-torsional Buckling and Ultimate Failure
Modes of Cold-formed Steel Lipped Channel Beams at
Elevated Temperatures ................................................................. 7-19
Figure 7.4: Finite Element Analysis Results at Elevated Temperatures –
For All the Cross-sections ............................................................ 7-20
Figure 7.5: Finite Element Analysis Results at Elevated Temperatures ......... 7-26
Figure 7.6: Effect of Elevated Temperatures on the Ultimate Moment
Capacity ........................................................................................ 7-31
Figure 7.7: Reduction in Moment Capacities at Elevated Temperatures
Relative to Ambient Temperature Moment Capacity .................. 7-32
Figure 7.8: Effect of Steel Grade .................................................................... 7-34
Figure 7.9: Non-linearity in Stress-strain Graphs ........................................... 7-35
Figure 7.10: Determining the fp,T Value at 300oC............................................. 7-35
Figure 7.11: Variation of fp,T/fy,T Ratio with Elevated Temperatures .............. 7-36
Figure 7.12: Effect of Stress-strain Relationship .............................................. 7-38
Figure 7.13: Strain Hardening and Elastic-perfect Plastic Material Models .... 7-39
Figure 7.14: Variation of ky,T/kE,T Ratio at Elevated Temperatures ................. 7-40
Figure 7.15: Variation of Non-Dimensional Member Slenderness with
Temperature .................................................................................. 7-42
Figure 7.16: Effect of ky,T/kE,T Ratio on the Non-Dimensional Member
Moment Capacity ......................................................................... 7-44
Figure 7.17: Comparison of FEA Results with Buckling Curves ‘a’ and ‘b’
of Eurocode 3 Part 1.3 / Eurocode 3 Part 1.1 ............................... 7-47
Structural Behaviour and Design of Cold-formed Steel Beams at Elevated Temperatures xvii
Figure 7.18: Comparison of FEA Results with Different Buckling
Curves of Eurocode 3 Part 1.1 (ECS, 2005) ................................ 7-49
Figure 7.19: Comparison of Ultimate Member Moment Capacities
with Predictions based on Selected Buckling Curves in
Different Temperature Ranges – G450 Steel Beams ................... 7-50
Figure 7.20: Comparison of Ultimate Member Moment Capacities
with Predictions based on Selected Buckling Curves in
Different Temperature Ranges – G250 Steel Beams ................... 7-51
Figure 7.21: Comparison of FEA results with the New Design Method
Proposed in Chapter 6 .................................................................. 7-53
Figure 7.22: Comparison of FEA results with Eurocode 3 Part 1.2 ................. 7-55
Figure 7.23: Comparison of Experimental Results of Real et al. (2003)
with their Finite Element Analysis Results .................................. 7-57
Figure 7.24: Comparison of FEA Results at Elevated Temperatures
with the Alternative Approach Proposed by
Dharma and Tan (2007) ............................................................... 7-60
Figure 7.25: Comparison of FEA Results at Elevated Temperatures
with the Rankine Approach Proposed by
Dharma and Tan (2007) ............................................................... 7-61
Figure 7.26: Comparison of the New Proposal with FEA Results ................... 7-65
Figure 7.27: Comparison of the New Proposal with FEA Results ................... 7-67
Structural Behaviour and Design of Cold-formed Steel Beams at Elevated Temperatures xviii
LIST OF TABLES
Table 2.1: Values of for Determining the Critical Buckling Stress ............... 2-6
Table 2.2: Recommended Values for Imperfection Factors for
Lateral-torsional Buckling Curves ............................................... 2-26
Table 2.3: Strength Reduction Factors for Cold-Formed Steels.................... 2-33
Table 2.4: Reduction Factors for the Mechanical Properties of Steels at
Elevated Temperatures (Eurocode 3 Part 1-2 (ECS, 2005)) ........ 2-34
Table 2.5: Reduction Factors for Mechanical Properties of S350GD+Z
Steel Proposed by Outinen et al. (2002a,b) .................................. 2-36
Table 2.6: Reduction Factors for Mechanical Properties of Flat Regions
of Cold-Formed Steel at Elevated Temperatures
(Chen and Young, 2004) .............................................................. 2-37
Table 2.7: Reduction Factors for Mechanical Properties of Corner
Regions of Cold-Formed Steel at Elevated Temperatures
(Chen and Young, 2006) .............................................................. 2-38
Table 2.8: Moment Correction Factors for the New Proposal..................... 2-45
Table 2.9: Values of the Severity Factor, (Real et al., 2006) ...................... 2-46
Table 2.10: Limiting Temperatures for the Design of Protected and
Unprotected Beams (BS 5950-Part 8 (BSI, 1990)) ...................... 2-50
Table 2.11: Design Temperatures for Beams (BS 5950-Part 8 (BSI, 1990)) .. 2-51
Table 2.12: Applied Boundary Conditions ...................................................... 2-62
Table 3.1: Dimensions of Tensile Test Coupons ............................................. 3-5
Table 3.2: Yield Strength and Elastic Modulus at Ambient Temperature .... 3-15
Table 3.3: Yield Strength Values (σy,T) for Cold-formed Steels
at Various Strain Levels ............................................................... 3-21
Table 3.4: Yield Strength Reduction Factors (fy,T/fy,20) for
Cold-formed Steels at Various Strain Levels ............................... 3-22
Table 3.5: Elastic Modulus and Reduction Factors of Cold-formed
Steel at Elevated Temperatures .................................................... 3-25
Table 3.6: Ultimate Strengths (fu,T) and Reduction Factors (fu,T/fu,20) of
Cold-formed Steel at Elevated Temperatures............................... 3-27
Table 3.7: Typical chemical composition of tested steel ............................... 3-28
Table 4.1: Test Program .................................................................................. 4-3
Table 4.2: Measured Section Dimensions ....................................................... 4-3
Structural Behaviour and Design of Cold-formed Steel Beams at Elevated Temperatures xix
Table 4.3: Measured Initial Geometric Imperfections of Test Beams and
Fabrication Tolerance Limit (AS4100, 1998) ................................ 4-6
Table 4.4: Ultimate Lateral-torsional Buckling Moment Capacities ............ 4-15
Table 4.5: Comparison of Experimental Moment Capacity Results
with Code Predictions .................................................................. 4-22
Table 5.1: Comparison of Elastic Lateral-torsional Buckling Moments for
Different Element Sizes ................................................................. 5-6
Table 5.2: Ambient Temperature Mechanical Properties used in the
Parametric Study .......................................................................... 5-16
Table 5.3: Effect of Imperfection Magnitude ................................................ 5-22
Table 5.4: Comparison of Elastic Buckling Moments of Lipped
Channel Beams (a) 20oC .............................................................. 5-33
Table 5.5: Comparison of Elastic Local Buckling Moments of Lipped
Channel Beams ............................................................................. 5-36
Table 5.6: Comparison of Experimental and Finite Element Analysis
Results .......................................................................................... 5-37
Table 5.7: Magnitude and Direction of Initial Geometric Imperfections
and Experimental Results ............................................................. 5-46
Table 5.8: Comparison of Finite Element Analysis Results with
Experimental Results from Snijder et al. (2002) .......................... 5-48
Table 6.1: Selected Lipped Channel Beam Sections and their Dimensions.... 6-5
Table 6.2: Section Moment Capacity Results from FEA ................................ 6-7
Table 6.3: Maximum Strains Observed in Finite Element Analyses ............ 6-14
Table 6.4: Comparison of Section Moment Capacities from FEA and
Design Methods Given in Shifferaw and Schafer (2008) and
EC3 Part 1.3 ................................................................................. 6-18
Table 6.5: Finite Element Analysis Results .................................................. 6-21
Table 6.6: Comparison of FEA Ultimate Moment Capacities with Code
Predictions for G250-1.55 Series at Ambient Temperature ......... 6-36
Predictions for G250-1.95 Series at Ambient Temperature ......... 6-37
Predictions for G450-1.5 Series at Ambient Temperature ........... 6-39
Predictions for G450-1.9 Series at Ambient Temperature ........... 6-40
Table 6.10: Mean and COV Values and Associated Capacity Reduction
Factors of the Ratio of FEA to Predicted Moment Capacities
Structural Behaviour and Design of Cold-formed Steel Beams at Elevated Temperatures xx

using Buckling Curve ‘a’ of Eurocode 3 Part 1.1 ........................ 6-48
Factors of the Ratio of FEA to Predicted Moment
Capacities using Option 1 ............................................................. 6-54
using Option 2 .............................................................................. 6-56
using Option 3 .............................................................................. 6-58
Table 6.14: Summary of Mean and COV Values and Associated Capacity
Reduction Factors of the Ratio of FEA to Predicted Moment
Capacities ..................................................................................... 6-61
Factors for the New Equations and EC3 Part 1.1 – Buckling
Curve ‘a’ ....................................................................................... 6-62
Table 7.1: Ambient and Elevated Temperature Mechanical Properties
Used in the Parametric Study ......................................................... 7-3
Table 7.2: Dimensions of Selected Cross-sections for the Parametric Study
at Elevated Temperatures ............................................................... 7-6
Table 7.3: Section Moment Capacities at Elevated Temperatures .................. 7-7
Table 7.4: Comparison of Section Moment Capacities Obtained from
FEA with Predicted Capacities using Eurocode 3 Part 1.3
(ECS, 2006) .................................................................................. 7-11
FEA with Predicted Capacities using Eurocode 3 Part 1.2
(ECS, 2005) .................................................................................. 7-12
FEA with Predicted Capacities using the Design Method
Proposed by Shifferaw and Schafer (2008) .................................. 7-13
Table 7.7: Maximum Longitudinal Membrane Strains Observed in the
Finite Element Analyses of G250-1.55-120-45-15 Beam at
Elevated Temperatures ................................................................. 7-15
Table 7.8: Ratios of Yield Strength Based on 1.5% Total Strain and 0.2%
Proof Stress................................................................................... 7-17
Table 7.9: Comparison of Section Moment Capacity Predictions using
0.2% Proof Stress and 1.5% Total Strain Based Yield Strength .. 7-18
Structural Behaviour and Design of Cold-formed Steel Beams at Elevated Temperatures xxi
Table 7.10: Elastic Lateral-torsional Buckling and Ultimate Moment
Capacities at Elevated Temperatures from FEA .......................... 7-21
Table 7.11: fp,T/fy,T Ratios of G250 and G450 steels at Elevated
Temperatures ................................................................................ 7-36
Table 7.12: Yield Strength and Elastic Modulus Reduction Factors and their
Ratios ............................................................................................ 7-41
Table 7.13: Proposed Buckling Curves from Eurocode 3 Part 1.1 ................. 7-46
Table 7.14: Mean, COV and Capacity Reduction Factors of the
Ratio of FEA to Predicted Member Moment Capacities using
the Proposed Buckling Curves ..................................................... 7-52
Table 7.15: Mean, COV and Capacity Reduction Factors of FEA to
Predicted Moment Capacities....................................................... 7-68
Structural Behaviour and Design of Cold-formed Steel Beams at Elevated Temperatures xxii
LIST OF SYMBOLS AND ABBREVIATIONS
A Cross-sectional area
AISI American Iron and Steel Institute
AS Australian Standard
AS/NZS Australian Standard/New Zealand Standard
b Flange width
be Effective plate width
BMT Base Metal Thickness
BSI British Standard Institute
Cb Moment distribution coefficient
Cp Correction factor depending on the number of tests
Cy Average membrane compressive strains
COV Coefficient of Variation
d Web depth
D Overall depth of the cross-section
Dc Twice the depth of the compression zone
Dw Section depth or twice the depth of the compression zone (Dc )
E , E20 Elastic modulus at ambient temperature
ET Elastic modulus at temperature T o C
E y ,T Elastic modulus at yield strength at temperature T o C
E fi ,d Design effect of actions for the fire design situation
ECS European Committee for Standardization
f 0.2 0.2% yield stress
f* Design stress in the compression element
f cr Elastic critical stress
f ol Elastic local buckling stress of the cross section
f od Elastic distortional buckling stress of the cross section
f pT , f p ,T Elevated temperature limit of proportionality
fT Stress at temperature T o C
f y , f y , 20 Yield strength at ambient temperature
f y ,T Yield strength at temperature T o C
f u ,T Ultimate strength at temperature T o C
Fm Mean value of the fabrication factor (1, 0.05)
FEA Finite Element Analysis
G Shear modulus
GT Shear modulus at temperature T.
Ix Second moment of area about x axis
Structural Behaviour and Design of Cold-formed Steel Beams at Elevated Temperatures xxiii
Iy Second moment of area about y axis
Iw Warping constant
J Torsion constant
k Plate local buckling coefficient
k E ,θ , k E ,T Elastic modulus reduction factor at elevated temperatures
kp Proof stress reduction factor at elevated temperatures
k y ,θ , k y ,T , k y The yield strength reduction factor at elevated temperatures
l Lip length
L Beam span
Le Effective length
m Degree of freedom (n-1)
M* Design bending moment
Ms Nominal section moment capacity
Mb Nominal member moment capacity
M b,T Nominal member moment capacity at temperature ToC
M be Nominal member moment capacity for lateral-torsional buckling
M bd Nominal member moment capacity for distortional buckling
M bl Nominal member moment capacity for local buckling
Mc Critical moment capacity
M c ,T Critical moment capacity at elevated temperature
My Yield moment capacity
M y ,T Yield moment capacity at temperature ToC
M o , M o, 20 Elastic lateral-torsional buckling moment at ambient temperature
M o ,T Elastic lateral-torsional buckling moment at temperature ToC
M od Elastic distortional buckling moment
M ol Elastic local buckling moment
Mf Applied moment at the fire limit state
M Rd Design resistance moment for normal temperature design
M fi ,θ , Rd Design resistance moment at temperature θ a
M pl , Rd Plastic moment resistance of the gross section for normal temperature
design
M el , Rd Elastic moment resistance of the gross section for normal temperature
design
M b , fi Ultimate buckling moment capacity at elevated temperatures
M u , M u , 20 Ultimate moment capacity at ambient temperature
M u ,T Ultimate moment capacity at temperature ToC
ME Elastic buckling moment capacity considering major axis curvature
Mm Mean of material factor
Structural Behaviour and Design of Cold-formed Steel Beams at Elevated Temperatures xxiv
n Number of tests
NAS North American Standard
Pm Mean value of the tested to predicted load ratio
po Limiting compressive stress
r Root radius
ry Radius of gyration of the section about the y axis
R fi ,d ,0 Design resistance of the steel member for the fire design situation for
time t = 0
t Thickness
T ,θa Uniform temperature
Vf Coefficient of variation of the fabrication factor
Vm Coefficient of variation of the material factor
Vp Coefficient of variation of the tested to predicted load ratio
Vq Coefficient of variation of load effect
w Plate width
Wel , Z f Full unreduced elastic section modulus
W pl , y , S x , S p , x Plastic section modulus of the gross section
Z c , Z eff , Weff Effective section modulus
σp Proof stress
σ nom Nominal stress
σ true True stress
ε nom Nominal strain
ε true True strain
εT Strain at temperature T o C
ε u ,T Strain corresponding to yield strength at temperature T o C
αT Temperature dependent thermal expansion coefficient
ν Poisons ratio
χ LT Reduction factor for lateral-torsional buckling
χ LT , fi Reduction factor for lateral-torsional buckling in fire design situation
α , α LT Imperfection factor
φb Capacity reduction factor for bending
φ Capacity reduction factor
γ M1 Partial factors for buckling
γ Mo Partial factor for yielding
γ M , fi Partial factor for the fire situation (usually γ M , fi =1).
η Perry coefficient
kc Moment correction factor
β Severity factor
Structural Behaviour and Design of Cold-formed Steel Beams at Elevated Temperatures xxv
β0 Target reliability index = 2.5
ρ Effective width factor
λ Slenderness ratio
λb Non-dimensional slenderness ratio used to determine Mc for members
subjected to lateral buckling
λb ,T Non-dimensional slenderness ratio used to determine Mc for members
subjected to lateral buckling at temperature ToC
λl Non dimensional slenderness used to determine M bl
λd Non dimensional slenderness used to determine M bd
_
λ LT Non dimensional beam slenderness
λ LT ,θ ,com Non-dimensional slenderness at elevated temperature, θ a,com
λ_ LT Beam slenderness at ambient temperature
λp Plate element slenderness
θ a,com Maximum temperature in the compression flange reached at time t
θ a,cr Critical temperature of carbon steel at time t
μo Degree of utilization at time t = 0
η LT
T
Initial curvature and twist term at elevated temperatures
Structural Behaviour and Design of Cold-formed Steel Beams at Elevated Temperatures xxvi
STATEMENT OF ORIGINAL AUTHORSHIP
The work contained in this thesis has not been previously submitted for a degree or
diploma at any other higher education institution. To the best of my knowledge and
belief, the thesis contains no material previously published or written by another
person except where due reference is made.
Signed:
Date:
Structural Behaviour and Design of Cold-formed Steel Beams at Elevated Temperatures xxvii
ACKNOWLEDGEMENTS
First and foremost I would like to express my deep and sincere gratitude to my
supervisor, Prof. Mahen Mahendran who supported me throughout this research with
his encouragement and guidance. His patience and understanding at difficult times
helped me to finish this thesis once I thought would not be possible. He always
guided me in the right direction providing constructive suggestions and ways of
improving the thesis further. Without his help provided throughout this three and
half years, I would not be able to finish my thesis successfully.
My sincere thanks goes to Queensland University of Technology (QUT) for giving

me this opportunity of doing a PhD and for providing financial support of my
research project through the International Postgraduate Research Scholarship (IPRS).
I would also like to thank the School of Urban Development of QUT for providing
necessary facilities and technical support during this research. I would also like to
express my appreciation to the Structural Laboratory staff members for their
assistance in my experimental studies. I especially appreciate my fellow post-
graduate students for their friendship and support.
I am grateful to my husband, Yasintha, for his love and care provided me from the
beginning we first met and throughout these three and half years. My deepest love
goes to my precious son, Ramindu, for bringing happiness to our lives. Even if he is
very little with just one year old, it seems he understood that I am busy with works.
My deep love and appreciation goes to my brother, Prasanna, and sister, Ruvini.
They always gave me lots of love and encouragement throughout my life. I am very
grateful to my mother, Chandra, for giving me lots of love and encouragement
throughout my life. She always let me know that she is proud of me which
encouraged me in finishing this PhD.
Finally, I am very grateful to my father, D.K. Ariyadasa, who is the reason I came
this far. He is in my mind always giving me lots of love and encouragement. Even if
it is painful to remember he is not with me at this juncture, one thing I know for sure
that he will be proud of me for this achievement.
Structural Behaviour and Design of Cold-formed Steel Beams at Elevated Temperatures xxviii
DEDICATION
I dedicate this thesis to my beloved father, D.K. Ariyadasa,

and mother, Chandra Fernando.
Structural Behaviour and Design of Cold-formed Steel Beams at Elevated Temperatures xxix
1 Introduction
1.1 General
In steel construction, two main types of structural steel are used. They are hot-rolled
and cold-formed steels. Hot-rolled steel is more popular and has many applications
in the construction industry. However, demand for cold-formed steel structures
especially for residential, industrial and commercial buildings has increased
significantly during the last decade. Figure 1.1 shows the use of cold-formed steel
members in a building. There are two major types of cold-formed structural
members. They are primary load bearing members such as beams in floor
assemblies, columns in wall assemblies, individual beams and columns, and truss
members, and non-load bearing members such as members in partition walls.
Figure 1.1: Building Composed of Cold-Formed Steel Sections

(Yu, 2000)
Three methods are generally used in the manufacturing of cold-formed steel

sections: cold roll-forming, press brake operation and bending brake operation. The
most widely used method is cold roll-forming. In the cold roll-forming method, cold-
formed steel sections are manufactured by sending a steel strip through a series of
rolls without the application of heat. Each pair of rolls progressively forms the steel
strip until the required cross-section is produced (see Figure 1.2). Press braking is
Structural Behaviour and Design of Cold-formed Steel Beams at Elevated Temperatures 1-1
normally used for production of small quantity of simple shapes such as angles,
channels and Z sections. The equipment used for press breaking consists of a moving
top beam and a stationary bottom bed on which dies are mounted (Figure 1.3).
Unlike hot-rolled steel sections, the manufacturing process of cold-formed steel
sections is entirely carried out at room temperature. A variety of cold-formed steel
sections with different thicknesses is available to meet the requirements of a wide
range of structural and non-structural applications. Figure 1.4 shows the different
shapes of cold-formed steel sections available today.
This Figure is not available online.

Please consult hardcopy thesis available
from QUT library
(a) Cold-roll forming sequence for a C-section (b) Cold roll-forming

(http://www.vatsalametsec.com/index.htm) (http://www.indiamart.com/valson-
fabricators/)
Figure 1.2: Manufacturing Process of Cold-Rolled Steel Sections
Figure 1.3: Press Braking (Karren, 1967)
Figure 1.4: Various Shapes of Cold-Formed Steel Sections (Yu, 2000)
Due to the strain hardening effect resulting from the cold-forming process, cold-
formed steel sections have a higher yield strength compared to the parent steel strip.
Therefore, cold-rolled steel members have a very high strength to weight ratio
compared to thicker hot-rolled steel members.
The interest of engineers and architects in using cold-formed steel members in

construction is increasing because of its qualities over other construction materials
such as hot-rolled steel, wood, concrete, etc. Cold-formed steel members have the
following unique qualities;
• Consistent quality – They have uniform properties throughout the member

and also have a uniform cross-section.
• Light weight – They are easy to handle contributing to reduced labour and
transportation costs. Reduced foundation loads due to the lightweight of cold-
formed steel elements leads to more economical designs.
• Fast and easy installation.
• Allows easy modifications at a later date.
• Excellent surface appearance – Cold-formed steels have manufacturing
controls assuring consistent surface quality.
• Design flexibility – Cold-formed steel members are available in a variety of
shapes and sizes to accommodate flexible cost effective designs.
• Resistance to termites and pests – cold-formed steel members are impervious
to termites and other wood destroying insects.
• Non-combustible.
Fire safety of cold-formed steel structures need greater attention since fire in a
building can cause not only loss of property but also loss of lives (see Figure 1.5).
Even though cold-formed steel is a non-combustible material, it does not guarantee
the fire safety of structures. To acquire the required fire resistance, the common
practice is to provide external fire protection to steel members to reduce the
temperature increase when exposed to fire. These protected cold-formed steel
members are used in wall and floor assemblies which are covered with gypsum
board with or without insulation material in the cavity. However, this type of fire
protection increases the cost of cold-formed steel buildings, which would affect the
demand for such structures. Also in recent years, the demand for using unprotected
cold-formed steel members in steel structures has increased significantly because
they improve the aesthetic appearance.
This Figure is not available online.

Please consult hardcopy thesis available
from QUT library
Figure 1.5: Fire in a Steel Structure

http://www.whatreallyhappened.com/spain_fire_2005.html
It has been found that the rate of temperature rise in steel members depends on its
section factor. Because cold-formed steel sections are thinner, it has a high section
factor, P/A where P is the section perimeter and A is the cross-sectional area. The
temperature of cold-formed steel members increases rapidly when exposed to fire,
because its thermal conductivity and section factor are high. The mechanical
properties of cold-formed steel such as the yield strength and the modulus of
elasticity deteriorate with increasing temperature. Therefore it is important to have
proper and accurate design methods for the cold-formed steel structural members at
elevated temperatures so that the structure as a whole does not fail within a time
period ensuring safe evacuation of people.
To date, more attention has been given to the studies into the behaviour of hot-rolled
steel structural members at elevated temperatures. Engineers and architects face
problems in using cold-formed steel members in structures because of the lack of
knowledge about the physical, mechanical and fire resistance properties at elevated
temperatures and the lack of design specifications for fire conditions. The current
design allows the use of the same design guidelines for cold-formed steel members
as for hot-rolled steel members under fire conditions, despite the significant
differences between them. Therefore it is important to develop more accurate design
methods to assess the stability of unprotected cold-formed steel members when
subjected to fire events.
A good understanding of the structural behaviour of cold-formed steel members

under fire is essential and needs greater attention. But very few researchers have
focused on this area. In particular, the flexural behaviour of cold-formed steel beams
has not been adequately investigated under fire conditions.
1.2 Research Problem
Cold-formed steel structures are becoming more popular in building construction for
up to five storey buildings. These buildings fail early when exposed to fire events
because cold-formed steel members have less fire resistance. Therefore cold-formed
steel members should be properly designed so that they can withstand unexpected
fire situations. Engineers and architects face problems in using cold-formed steel
members in structures because of the lack of knowledge about the mechanical
properties at elevated temperatures and the lack of design specifications for fire
conditions. Although some researchers focused on the behaviour of cold-formed
steel columns at elevated temperatures, no study has been undertaken on the flexural
behaviour of individual cold-formed steel beams at elevated temperatures. Therefore
this research is focused on the study of flexural behaviour of cold-formed steel
beams at elevated temperatures.
Eurocode 3 Part 1.3 (ECS, 2006): “supplementary rules for cold-formed thin gauge
members and sheathing” gives design rules for cold-formed steel structural elements
at ambient temperature. Eurocode 3 Part 1.2 (ECS, 2005) “General rules – Structural
fire design” gives design rules for hot-rolled steel members subjected to fire.
Eurocode 3 Part 1.2 recommends the same guidelines for cold-formed steel thin
gauge steel members and sheathing although the recent studies have shown that even
the variations of mechanical properties of cold-formed steels with increasing
temperature are clearly different from that of hot-rolled steels (Outinen and
Makelainen, 2001, Ranawaka and Mahendran, 2009).
Both buckling behaviour and mechanical properties of cold-formed steel members

are different from those of hot-rolled steel members. Cold-formed steel sections are
mono-symmetric. Therefore their shear centre does not coincide with the centroid
which results in major changes to their structural properties compared to that of
symmetric sections. In addition, differences in material properties, residual stresses,
and initial imperfections also exist between hot-rolled and cold-formed steel
members due to the use of different manufacturing methods. Unlike hot-rolled steel,
the yield strength and the modulus of elasticity of cold-formed steel may vary
significantly depending on the location in the cross-section. The yield strength and
the ultimate tensile strength are higher in the rounded corners than in the flats of
cold-formed steel sections. The stress-strain relationship of cold-formed steel is also
different from that of hot-rolled steel. The initial imperfections of cold-formed steel
members may also be different because of the different manufacturing methods used.
Because of all these differences, it is obvious that independent design guidelines are
necessary for cold-formed steel members under fire conditions.
The Australian cold-formed steel structures design code, AS/NZS 4600 (SA, 2005),
gives design guidelines for cold-formed steel structural members at ambient
temperatures, but does not cover fire design.
The British fire design code, BS 5950 Part 8 (BSI, 1990), proposes two methods to
determine the fire behaviour of hot-rolled steel members: the limiting temperature
method and the moment capacity method. But it is limited to hot-rolled steel
members and does not cover cold-formed steel members. BS 5950 Part 5 (1998):
“Code of practice for design of cold formed thin gauge sections” only provides
design guidelines at ambient temperatures.
No research has been conducted to study the flexural behaviour of individual cold-
formed beams at elevated temperatures although some research has been conducted
on cold-formed steel beams incorporated in floor assemblies.
Kaitila (2002) presented numerical modelling of simply supported cold-formed steel

floor beams under fire conditions. The beams were considered as restrained. Because
the floor assemblies are subject to fire from one side, a temperature gradient
develops across the cross-section of the steel joists. In the numerical simulations, the
temperature was taken as linearly varying from bottom flange to mid-height of web
and from mid height of web to top flange.
Alfawakhiri et al. (2001) presented a comprehensive thermal-structural model for the

assessment of fire resistance of unrestrained cold-formed steel floor assemblies. The
steel beams inside the floor assemblies were found to have varying temperatures
across the cross-section.
Thus, it appears that no research has been conducted into the flexural behaviour of
individual cold-formed steel beams subjected to uniform elevated temperatures and
also there are no independent design guidelines for cold-formed steel members under
fire conditions. Hence it is important to study the flexural behaviour of cold-formed
steel beams in relation to their member moment capacity of long unrestrained beams
subject to lateral-torsional buckling and the section moment capacity of short or fully
restrained long beams subject to local buckling and yielding, and to develop accurate
design rules for fire conditions.
1.3 Research Objectives and Scope
Overall Objective
The main objectives of this research are to investigate the behaviour of cold-formed
steel beams under simulated fire conditions and to develop safe and economical
design rules for their section and member moment capacities relating to their local
buckling and lateral-torsional buckling effects, respectively.
Specific Objectives and Tasks
Experimental Study
1. To investigate the mechanical properties of cold-formed steels at

elevated temperatures.
2. To conduct an experimental study of simply supported cold-formed steel
lipped channel beams loaded at quarter points to investigate the lateral-
torsional buckling behaviour at ambient temperature.
Numerical Study
1. To develop advanced finite element models of cold-formed steel lipped

channel beams capable of simulating local and lateral torsional buckling
effects based on the cross-section geometry, section and member
imperfections, residual stresses, material properties, load type, position
and magnitude at ambient and elevated temperatures using finite element
program ABAQUS and validate them using experimental results.
2. To investigate the lateral torsional buckling failure of long unrestrained

lipped channel beams and the section bending and local buckling failure
of short beams and fully laterally braced long beams at ambient and
elevated temperatures using extensive finite element analyses based on
validated models.
3. To undertake a detailed numerical parametric study using the validated

finite element model in order to investigate the effects of a number of
influential parameters on the section and member moment capacities of
cold-formed steel lipped channel beams both at ambient and elevated
temperatures.
Development of Design Rules
1. To assess the accuracy of the current design rules for cold-formed steel
beams at ambient and elevated temperatures using the results of the
numerical simulations.
2. To develop accurate design rules for the member moment capacity of

cold-formed steel beams subject to lateral-torsional buckling at ambient
and elevated temperatures.
3. To develop accurate design rules for the section moment capacity of cold-
formed steel beams subject to local buckling and yielding.
1.4 Scope
This research concentrates on improving the understanding of the flexural behaviour

of cold-formed steel lipped channel beams including their buckling characteristics at
ambient and elevated temperatures, and developing suitable design guidelines for
their fire safety. This research has the following limitations.
• The member moment capacity of beams subject to lateral torsional buckling

and the section moment capacity of beams subject to local buckling and
yielding were investigated in this research. Distortional buckling or the
interaction effects of local, distortional and lateral-torsional buckling modes
were not investigated.
• Beam testing was limited to cold-formed steel beams loaded at the shear
centre at ambient temperature. The behaviour of eccentrically loaded beams
was not investigated in this research.
• Experimental testing and finite element analyses were limited to one of the
commonly used cold-formed section, the lipped channel beam. Other cross
section types were not investigated.
1.5 Research Methodology
This research on cold-formed steel beams at ambient and elevated temperatures was
based on the following.
• Detailed literature review
• Tensile coupon tests to obtain the mechanical properties of cold-formed
steels at ambient and elevated temperatures.
• Development of suitable finite element models of cold-formed steel beams to
compare with experimental results for the purpose of validation, and to
develop suitable design rules.
• Detailed parametric studies using the validated finite element models at
ambient and elevated temperatures.
There are three different methods of assessing the fire safety of structures: time domain,
temperature domain, and strength domain. This research is based on the strength
domain used in fire safety design. Fire safety in the strength domain is verified by
comparing the applied load at the time of the fire with the minimum load capacity of
structural members during the design fire event. The development of accurate design
moment capacity equations at elevated temperatures in this research will allow
engineers to calculate the minimum moment capacities of cold-formed steel lipped
channel beams at the expected maximum elevated temperature during the design fire
event. These moment capacities can then be easily checked against the design fire
loads, thus enabling a fire safe design for cold-formed steel beams.
1.6 Contents of the Thesis
The flexural behaviour of cold-formed steel beams at ambient and elevated

temperatures was investigated based on experimental tests and extensive finite
element analyses, and new design rules for cold-formed steel lipped channel beams
both at ambient and elevated temperatures have been developed in this thesis. This
thesis contains eight chapters and the contents of each chapter except this chapter are
briefly described as follows.
Chapter 2:
This chapter presents a literature review covering various aspects of the behaviour of
flexural members at ambient and elevated temperatures, effects of temperature
increase on the mechanical properties of cold-formed steel, different design methods
available at ambient and elevated temperatures, experimental and analytical
investigations conducted by previous researchers.
Chapter 3:
This chapter presents the experimental investigation of mechanical properties of

cold-formed steels at ambient and elevated temperatures using tensile coupon tests
and the new equations developed for the reduction factors of yield strength and
elastic modulus. Comparison of the yield strength and elastic modulus reduction
factors obtained from tensile coupon tests with values obtained from available
equations in other sources are also presented.
Chapter 4:
This chapter presents the experimental investigation carried out on the lateral-
torsional buckling behaviour of simply supported cold-formed steel lipped channel
beams at ambient temperature.
Chapter 5:
This chapter presents the details of the finite element models for section and member
moment capacities of cold-formed steel lipped channel beams subject to local
buckling and lateral-torsional buckling, respectively. The validation of developed
finite element models is also presented.
Chapter 6:
The detailed parametric study carried out on the section and member moment
capacities of cold-formed steel lipped channel beams at ambient temperature using
finite element analyses is presented in this chapter. The effects of various parameters
on the lateral-torsional buckling behaviour were investigated. The new design
equations developed for cold-formed steel lipped channel beams at ambient
temperature are also presented.
Chapter 7:
This chapter presents the detailed parametric study carried out on the section and
member moment capacities of cold-formed steel lipped channel beams at elevated
temperatures. The effects of various parameters on the lateral-torsional buckling
behaviour at elevated temperatures were investigated. The new design equations for
lateral-torsional buckling were developed to satisfy the elevated temperature
conditions and presented in this chapter.
Chapter 8:
The summary of the most significant findings of this research and the
recommendations made for further research are presented in this chapter.
2 Literature Review
2.1 General
Cold-formed steel structures are becoming more popular throughout the world due to
their good characteristics such as high strength to weight ratio, availability in various
sections etc. Extensive research has been carried out during the past decades on cold-
formed steel members and there are many research publications dealing with cold-
formed steel members at ambient temperature. Findings of these research works have
contributed towards the increasing usage of cold-formed steel members for structural
purposes while improving the quality and safety of construction. In recent times, the
fire safety of structures is also considered very critical and many researchers have
paid their attention to fire safety. Fire performance of cold-formed steel floor and
wall assemblies that are widely used in residential buildings has been recently
studied by several researchers (Alfawakhiri et al., 2001, Sultan, 2004, Kaitila, 2002).
Considerable research has also been undertaken on individual cold-formed steel
columns at elevated temperatures. However, flexural behaviour of individual cold-
formed steel beams has not been well studied yet, and no research publications could
be found on this topic. Therefore research carried out on hot-rolled steel beams at
elevated temperature is considered as the basis for this research including the
research on cold-formed steel beams at ambient temperature. In this research, our
main attention is on the lateral torsional buckling behaviour of cold-formed steel
beams, but also including their section moment capacities. This chapter provides a
review of past research carried out on cold-formed and hot-rolled steel beams that is
relevant to this research.
2.2 Characteristics and Mechanical Properties of Cold-Formed

Steel Sections at Ambient Temperature
Generally, the depth of cold-formed steel sections ranges from 50 mm to 300 mm

and the thickness of light-gauge cold-formed steel sections varies from 0.42 mm to
2.0 mm. However, recent developments have allowed the production of cold-formed
steel sections up to 25 mm thick. The nominal yield strength of cold-formed steels
generally ranges from 250 to 550 MPa while their modulus of elasticity is
approximately 200,000 MPa.
The cold-forming process can alter the mechanical properties of cold-formed steel
from the parent steel sheet. Since the mechanical properties of cold-formed steel play
an important role in the structural performance, these changes should be considered
in the designs. The cold-forming process increases both the yield strength and the
ultimate tensile strength while reducing the ductility (Yu, 2000, Chen and Young,
2006). However, according to Yu (2000) and Chen and Young (2006), the
percentage increase in tensile strength is much smaller than that of yield strength
with a consequent marked reduction in the difference between yield point and tensile
strength. Due to the differences in the degree of cold working in the flat parts and the
corner regions, the mechanical properties vary over the cross-section as shown in
Figure 2.1. The studies conducted by Karren and Winter (1967), Karren (1967) and
Chajes et al. (1963) on the influence of cold work showed that the changes of
mechanical properties due to cold-working are mainly due to strain hardening and
strain aging (Yu, 2000), as shown in Figure 2.2. Abdel-Rahman et al. (1997) stated
that the induced deformations due to cold forming in the flat parts of the section may
be elastic deformations while the deformations expected in the corner parts are
plastic deformations.
Figure 2.1: Effect of Cold Work on the Mechanical Properties of Cold-Formed

Steel Unstiffened Channel Sections (Karren and Winter, 1967)
The stress-strain relationship for the corners and the flat parts is also different and
two types of idealized stress-strain curves, tri-linear curves for flat parts and
Ramberg-Osgood type curves for the corners have been widely used by researchers
(Key and Hancock, 1993, Pi and Trahair, 1995, Pi et al., 1998). These two types of
stress-strain curves are shown in the Figure 2.3.
Figure 2.2: Effects of Strain Hardening and Strain Aging on Stress-Strain

Characteristics (Chajes et al., 1963)
Stress, Stress,
Strain, Strain,
(a) Tri-Linear (Flats) (b) Ramberg-Osgood Type Curves (Corners)
Figure 2.3: Stress-Strain Curves (Pi and Trahair, 1995)
The design standards such as AS/NZS 4600 (SA, 2005), American Iron and Steel
Institute Specification (AISI, 2007), BS5950-Part 5 (BSI, 1998) and Eurocode 3-Part
1-3 (ECS, 2006) allow using the increased yield strength that resulted from cold-
forming process and also give recommendations on how to compute the increased
yield strength.
In the next section, a brief description of buckling modes that are associated with the
failures of cold-formed steel members and a review of literature on the flexural
behaviour of steel beams at ambient temperature are presented.
2.3 Flexural Behaviour of Cold-Formed Steel Beams
The use of thinner cold-formed steel sections as flexural members leads inevitably to
complex problems in stability. The problems associated with cold-formed steel
members are discussed in the following sections.
2.3.1 General Buckling Modes
Cold-formed steel beams generally fail due to material yielding, local buckling,
distortional buckling, and lateral torsional buckling. In addition, local buckling and
distortional buckling may interact with lateral torsional buckling because cold-
formed steel sections are often thin and flexible. Figure 2.4 presents the result of a
finite strip buckling analysis which illustrates different buckling modes that can
occur with a commonly used cold-formed steel C-section at different half wave
lengths.
Figure 2.4: Buckling Modes of Lipped Channel Beams in Bending

(Hancock, 2003)
2.3.1.1 Local Buckling
Local buckling is the buckling of individual plate elements which are under
compression without changing the fold lines of the cross-section as shown in Figure
2.5. In the case of beams, local buckling may occur in the compression flange or part
of the web which is under compression. Local buckling is characterized by short half
wave lengths compared to the other buckling modes. Further, the half wave length of
local buckling is in the order of the width of individual plate elements.
Figure 2.5: Local Buckling of Stiffened Top Flange of Hat-Shaped Beams

(Winter, 1970)
Earlier researchers, Timoshenko and Gere (1961), Bleich (1952), Bulson (1970) and
Allen and Bulson (1980), extensively investigated and summarised the elastic critical
stresses for local buckling (Hancock, 1998). The elastic critical stress for local
buckling of a plate element in compression, bending or shear is given by (Hancock,
1998),
2
kπ 2 E ⎛ t ⎞
f cr = ⎜ ⎟
( )
12 1 − ν 2 ⎝ b ⎠ (2.1)
Where k is called the plate local buckling coefficient which depends on the support
conditions, (b / t ) is the plate slenderness in which b is the plate width and t is the
plate thickness. A summary of plate local buckling coefficients ( k ) for a long
rectangular plate subjected to different types of stress (compression, shear or
bending) and boundary conditions (simply supported, fixed, or free edge) are given
in Table 2.1.
A plate element is said to be slender if the elastic critical local buckling stress ( f cr )
calculated using the above equation is less than the material yield strength. Therefore
if a particular section is slender, the local buckling takes place prior to stresses
reaching the yield strength. This leads to a large reduction of section moment
capacity of beams because the local buckling and the material yielding are the two
failure criteria that govern the section moment capacity. If the elastic critical
buckling stress ( f cr ) exceeds the yield stress, the plate element under compression
buckles in the inelastic range (Yu, 2000).
Table 2.1: Values of k for Determining the Critical Buckling Stress

(Timoshenko and Gere, 1961)
If the beam carries increasing load after the local buckling failure, the beam is said to
have post-local buckling capacity. Therefore even if the local buckling occurs in a
plate element prior to yielding it does not necessarily mean that the failure of the
section has occurred. Normally the postbukling reserve is allowed for in the design
to achieve an economic solution (Hancock, 1998).
2.3.1.2 Distortional Buckling
Cold-formed steel flexural members may undergo a mode of buckling called

distortional, unless the compression flange is adequately restrained. In the flange-
distortional buckling, flange and edge stiffener rotate about the flange-web junction
with some rotational resistance provided by the web (see Figure 2.6). This type of
buckling mode is most likely to occur in the high strength open thin-walled sections
such as C and Z sections.
Buckled shape
Original shape
Figure 2.6: Distortional Buckling
This type of buckling may occur in thin sections in compression or bending at

stresses significantly below the yield stress, especially for high strength steels
(Hancock and Rogers, 1998). The half wave length of distortional buckling is
generally in between that of local buckling and lateral torsional buckling and
typically several times larger than the largest characteristic dimension of the section.
If the sections are composed of high-strength steel, then there may be a significant
postbuckling reserve of strength beyond the elastic distortional buckling stress in a
similar manner to that which normally occurs for local buckling (Kwon and
Hancock, 1992). Distortional buckling has less post-buckling capacity than local
buckling mode (Schafer and Pekoz, 1999). This is also observed experimentally by
Hancock et al. (1994). Distortional buckling failures occur even when local buckling
is at a lower critical elastic moment than distortional buckling because of reduced
post buckling strength in distortional failure mode (Schafer and Pekoz, 1999).
2.3.1.3 Lateral Distortional Buckling
The lateral distortional buckling involves transverse bending of vertical web and is
most likely to occur if the tension flange is restrained (Figure 2.7). This type of
buckling often occurs in hollow flange beams because the tubular flanges of hollow
flange beams are very stiff torsionally while their webs are comparatively slender
and easily undergo web distortion. Eventhough, when the tension flange is not
restrained laterally and flanges are not torsionally stiff, cold-formed C-section beams
may still fail by lateral distortional buckling if the web is particularly slender (Pi et
al., 1998).
Buckled shape
Original shape
Figure 2.7: Lateral Distortional Buckling
2.3.1.4 Lateral Torsional Buckling
Lateral torsional buckling failure is the most complex failure criterion of steel
beams. This type of failure is identified by the simultaneous bending and twisting of
the entire cross-section without cross-sectional deformation (see Figure 2.8). If a
beam is not restrained laterally, it tends to fail by lateral torsional buckling in cases
where lateral stiffness and torsional stiffness are low. It occurs generally if the beam
has a higher bending stiffness in the vertical plane compared to the horizontal plane.
Because of the geometry of the open cross-section, which gives great flexural
rigidity about one axis at the expense of low torsional rigidity and low flexural
rigidity about a perpendicular axis, cold-formed steel members are particularly
susceptible to lateral torsional buckling (Chu et al., 2004). The parameters affecting
lateral torsional buckling resistance are the length between lateral restraints, the type
and the positions of load, the geometry of cross-sections, the type of end supports,
the presence or absence of stiffening devices that restrain warping at critical
locations, the material properties, the magnitude and distribution of residual stresses,
initial imperfections of geometry and loading, and cross-sectional distortion
(Galambos, 1998).
Buckled shape
Original shape
(a) Sectional View (b) Experimental Results (Trahair, 1993)
Figure 2.8: Lateral Torsional Buckling
A review of the research conducted on cold-formed steel beams at ambient

temperature is summarised in the next section.
2.4 Previous Research on Cold-Formed Steel Flexural Members

at Ambient Temperature
Rondal (2000) addressed stability problems of cold-formed steel members and gave
a review of recent progress in the field of cold-formed steel members. Davies (2000)
outlined a review of developments in cold-formed steel section technology,
applications and design procedures for cold-formed sections, the application of
generalized beam theory (GBT) to buckling problems, current design models and
their deficiencies, and design using whole section modulus. Hancock (2003)
presented an updated review of references on cold-formed steel research in leading
journals in 1999-2001. He further described the developments in the North American
Specification (NAS, 2001) for the design of cold-formed steel structural members.
To account for the local buckling in the slender steel members, design codes such as
AS/NZS 4600 (SA, 2005), AISI (1996), BS5950 Part 5 (BSI, 1998) and Eurocode 3-
Part 1-3 (ECS, 2006) suggest using the effective section properties which are
calculated by considering the effective width of stiffened or unstiffened individual
slender plate elements. This method is known as the effective width method. This
was originally proposed by von Karman (1932) and calibrated for cold-formed steel
members by Winter (1947) from the post-buckling analysis of plates under
compression. However, this method becomes more complicated for the sections
having complex shapes and additional edge and intermediate stiffeners. It is also
argued by several researchers that the consideration of isolated plate elements in the
calculation of effective section is less accurate since the interaction between
elements may also occur. To overcome these problems, Schafer and Pekoz (1998)
developed a new method called ‘Direct Strength Method’. This method uses elastic
buckling solutions for the entire cross-section rather than for the individual elements,
and strength curves for the entire member. The direct strength method was approved
by the American Iron and Steel institute in 2003, and it has been adopted in
Appendix 1 of the North American Specification for the design of cold-formed steel
structural members (Yu and Schafer, 2004).
Two primary modes of instabilities observed in the laterally braced cold-formed steel
beams are local and distortional buckling. More recent experiments on laterally
braced flexural members with edge stiffened flanges by Willis and Wallace (1990),
Schuster (1992), Moreyra (1993) and Ellifritt et al. (1997) demonstrate the un-
conservative nature of strength predictions using AISI specification. Schafer and
Pekoz (1999) investigated the buckling behaviour of the laterally braced cold-formed
steel flexural members with edge stiffened flanges, and presented a new procedure
for hand prediction of the critical buckling stress in the local and distortional modes.
A new design method for strength prediction based on unified effective width
approach was also presented including a new approach for determining the effective
width of the web. Comparison with experimental tests revealed that the new
approach is more consistent and reliable than the existing design methods.
Two series of tests were conducted on C and Z section cold-formed steel beams to
study two modes of failure: local buckling (Yu and Schafer, 2002) and distortional
buckling (Yu and Schafer, 2004, Yu and Schafer, 2006). The local buckling test
results indicated that AISI (1996), CSA (1994) and the new NAS (2001) design
methods provide adequate strength predictions in non-slender members that had
inelastic reserve capacity ( M test M y > 1 ). Only the direct strength method provides
the best test to predicted ratio for both slender and non-slender specimens in the case
of local buckling. The comparison of distortional buckling test results with the
design specifications revealed that the American specification (AISI, 2007),
Canadian standard (CSA, 1994) and the newly developed North American
Specification (NAS, 2001) provide a poor prediction of the strength for members
with failures initiated by distortional buckling. Further, AS/NZS 4600 (SA, 1996)
and the direct strength method provide reliable predictions of the capacity in
distortional buckling with conservative errors. Yu and Schafer (2007) developed a
finite element model and verified it using the experimental results of cold-formed
steel C and Z sections. The results showed that the direct strength method provides
reasonable strength predictions for both local and distortional buckling failures of Z
and C section beams with a wide range of geometries and yield stresses of steel.
They also showed that the direct strength method predicts conservatively the
increased strength due to moment gradient in distortional buckling by using
appropriate elastic buckling moments.
There are many investigations carried out on inelastic reserve capacity of cold-
formed steel flexural members at ambient temperature. Different types of sections
have been tested including hats and deck sections (Acharya and Schuster, 1997,
Desmond, 1997, Hoglund, 1980, Papazian et al., 1994, and Winter, 1996) and C and
Z sections (Cohen, 1987, LaBoube and Yu, 1978, Rogers, 1995, Shan et al., 1994,
Yu and Schafer, 2003, 2006) in which inelastic reserve capacity has observed as
reported by Shifferaw and Schafer (2008). Shifferaw and Schafer (2008) investigated
the inelastic reserve capacity of cold-formed steel flexural members. Even though
there exists inelastic reserve capacity of cold-formed steel beams, full plastic
moment capacity cannot be achieved. Shifferaw and Schafer (2008) developed new
design method to calculate inelastic reserve capacity of cold-formed steel beams
based on direct strength method.
Most common cold-formed steel sections that are used as beams are the C and Z-
sections which are geometrically mono-symmetric. Cold-formed open sections are
thin-walled and hence have very low torsional rigidity and are therefore more likely
to undergo torsional deformations. Because of the open nature of the cold-formed
steel beams, the shear centre does not coincide with the centroid. These mono-
symmetric cold-formed steel beams are often loaded eccentrically to the shear centre
causing failure by combined torsion and bending (Figure 2.9). Several research
projects have been carried out to test cold-formed steel beams loaded at the shear
centre and also eccentrically (Celebi et al., 1972, Ellifrit et al., 1991 & 1992, Put et
al., 1999a, Gotluru et al., 2000).
Figure 2.9: Eccentrically Loaded Channel Beam (Hancock, 2003)
Celebi et al. (1972) conducted tests on two unlipped cold-formed C-section beams
loaded eccentrically through the web. Lindner and Kurth (1979) conducted
approximately 160 tests on lipped and unlipped simply supported cold-formed steel
beams loaded above the top flange, either through the web or centroid at the mid
span. The test results showed that the strengths of beams loaded through the centroid
were significantly lower than those loaded through the web. Kavanagh and Ellifrit
(1994) reported ten tests of discretely braced cold-formed steel C-section beams
loaded at the web centreline and these tests indicated that the strengths generally
increased as the amount of bracing increased. Ellifrit et al. (1991, 1992) and Winter
et al. (1949) have observed the same in their tests conducted on C-section beams.
Put et al. (1999a) performed bending and torsion tests on unbraced simply supported
cold-formed steel channel beams loaded eccentrically at the mid span. The test
results showed that the beam strengths decrease as the load eccentricity increases and
that the strength is higher when the load acts on the centroid side of the shear centre
than when it acts on the side away from the shear centre. They developed simple
interaction equations that can be used in the design of eccentrically loaded cold-
formed channel beams.
Gotluru et al. (2000) investigated the behaviour of cold-formed steel beams

subjected to torsion and bending because the transverse load was not applied at the
shear centre. They performed simple geometric nonlinear analyses, finite element
analyses and finite strip analyses and compared their results with experimental
results. The influence of typical support conditions was studied and they were found
to produce partial warping restraints at the ends. Yu (2000) provides numerical
methods for calculating torsional properties of thin-walled sections.
Pi et al. (1997, 1998) developed a non-linear inelastic finite-element model for

analysing the lateral distortional buckling behaviour of cold-formed steel lipped
channel section beams. The results were used to develop improved design rules for
cold-formed steel channel section beams. Pi et al. (1998) also stated that the effects
of residual stresses, direction and the magnitude of the initial crookedness and twist,
bending moment distribution and the load height are very significant on the design
capacity of beams. They observed the inelastic strengths of cold-formed steel lipped
channel beams with positive imperfections and twist rotations are higher than those
of the beams with negative imperfections and twist rotations (Figure 2.10). They
gave a theoretical explanation to this behaviour. Figure 2.10 (a) shows the effects of
initial negative imperfections on the capacity. It can be seen that when imperfection
is –L/1000 beam capacities of very slender beams are lower than the elastic buckling
capacities and as the imperfection reduces the beam capacities going up. The effects
of web distortion were found to be small due to low torsional rigidities of cold-
formed steel lipped channel beams (Figure 2.11).
(a) Negative Imperfection and Twist
(b) Positive Initial Imperfection and Twist
Figure 2.10: Effects of Initial Imperfection and Twist (Pi et al., 1998)
Figure 2.11: Effects of Web Distortion (Pi et al., 1998)
Because of the low torsional stiffness of cold-formed steel beams as discussed

before, lateral torsional buckling is the predominant mode of buckling of unbraced
cold-formed steel beams loaded through the shear centre. According to Pi et al.
(1998), the lateral torsional buckling of cold-formed steel beams under shear centre
loading have rarely been studied and there is lack of well documented test data. This
was also supported by Pekoz (1987) and Trahair (1993). Even today, compared to
local and distortional buckling, research conducted on lateral torsional buckling
behaviour of laterally unbraced cold-formed steel members is very little at ambient
temperature.
The lateral torsional buckling behaviour is illustrated in Figure 2.12, relating to

buckling resistance and slenderness. There are three different ranges of behaviour
namely, elastic buckling, inelastic buckling and plastic behaviour. The elastic lateral
torsional buckling occurs in slender beams with low resistance to lateral bending and
twisting. As the slenderness decreases, the resistance of a beam to undergo elastic
buckling increases and the beam may yield before its elastic buckling moment is
reached. Yielding reduces the effective out-of-plane rigidities, and hence, lateral
torsional buckling occurs before reaching the elastic buckling moment. This type of
buckling of beams having intermediate slenderness is called the inelastic lateral
torsional buckling. If the beam is fully or adequately restrained laterally so that the
slenderness is low, it achieves the full plastic moment capacity.
Figure 2.12: Effect of Slenderness on the Buckling Resistance of Beams
(Real et al., 2004a)
The elastic critical buckling moment for lateral torsional buckling of an I-beam
subjected to pure bending is given next (Hill, 1954, Trahair, 1993, Trahair and
Bradford, 1988, Yu, 2000).
⎛ π 2 EI y ⎞⎛
⎟⎜ GJ + π EI w
2
⎞
M cr = ⎜ ⎟⎟ (2.2)
⎜ L2 ⎟⎜ L2
⎝ ⎠⎝ ⎠
In Equation 2.2, EI y , GJ and EI w are the minor axis flexural rigidity, torsional
rigidity and warping rigidity, respectively.
The unpublished data of 74 tests on lateral buckling of cold-formed I-section beams

of various shapes, spans and loading conditions have demonstrated that Equation 2.2
also applies to cold-formed steel sections with reasonable accuracy (Winter, 1970,
Yu, 2000). Hill (1954) has shown that Equation 2.2 applies to channel beams
without change, as a very satisfactory approximation. Trahair (1993) also stated that
the same equation for the elastic buckling resistance of I-beams can be used for
mono-symmetric sections bent about the axis of symmetry.
Most design codes give single beam design curve for lateral-torsional buckling of
different types of section as a function of the beam slenderness. Trahair (1998)
argued that multiple design curves for the beams against lateral buckling are needed
because there is a very wide range of beam cross-section types and variations in
factors such as residual stresses, initial imperfections and twist, web flexibility and
the material stress-strain relationship. They developed a consistent set of multiple
design curves for the lateral-torsional buckling of steel beams. Trahair (1998)
explained the relationship between the beam slenderness and the beam capacity. The
influence of elastic lateral buckling on the strength of simply supported beams in
uniform bending is shown in Figure 2.13 (Trahair, 1998). At low slenderness,
λ = M px M o the strengths of compact beams can rise above the full plastic
moment capacity, Mpx due to strain hardening effects. At high slenderness region, the
beam strengths are close to the elastic buckling strengths. Trahair (1998) stated that
some design codes take beam capacity to be less than elastic buckling capacity at
high slenderness region following the theoretical predictions of the influence of
initial curvature and twist, while other design codes take beam capacity to be equal
to elastic buckling moments. Trahair (1998) approximated the beam capacity of
beams with high slenderness to be equal to elastic buckling capacity based on a
compromise between the weakening effects of the geometrical imperfections and the
strengthening effects of the pre-buckling deformations.
Figure 2.13: Lateral Buckling Strength (Trahair, 1998)
The pre-buckling deformations in the plane of the beam increase the buckling
resistance of the beam. Trahair (1996) stated that the resistance of the beam in
uniform bending increases with the ration of Iy/Ix, and theoretically becomes infinite
when Iy=Ix. Equation 2.2 ignores the effect of the pre-buckling in-plane deflections
which resulted in increase of its buckling resistance. When the effects of pre-
buckling deflections are included, the lateral torsional buckling moment is given by
(Trahair, 1993, Pi and Trahair, 1992a,b)
M cr
M = (2.3)
{(1 − EI y ( [ ]
EI x ) 1 − GJ + π 2 EI w L2 2 EI x )}
The solution from Equation 2.2 is only valid if the beam remains elastic and a
modified equation for inelastic critical moment is given by Trahair and Bradford
(1988). If the beams are subjected to equal and opposite end moments and the
distribution of yield across the section does not vary along the beam and if there are
no residual stresses, the inelastic critical buckling moment is given by,
⎛ π 2 (EI y )e (GJ )e ⎞⎛ π 2 (EI w )e ⎞

M cr = ⎜ ⎟⎜1 + ⎟ (2.4)
⎜
⎝ L2 ⎟⎜
⎠⎝ (GJ )e L2 ⎟
⎠
in which (EI x )e , (GJ )e and (EI w )e are the reduced inelastic rigidities which are
effective at buckling.
Put et al. (1999b) performed lateral buckling tests on concentrically loaded simply
supported unbraced cold-formed steel C-section beams with different cross-section
dimensions. The experimental results showed that the strengths of beams failed in
the negative direction by lateral-distortional buckling are lower than the strengths of
beams failed in the positive direction. The results of lateral buckling tests showed
good agreement with the theoretical predictions of lateral buckling capacity of cold-
formed steel beams that has been developed by Pi et al. (1998). The results were also
compared with both hot-rolled and cold-formed steel design codes AS4100 (SA,
1998) and AS/NZS 4600 (SA, 2005). The comparison showed that the predictions by
AS4100 (SA, 1998) were generally closer to the test results while the prediction by
AS/NZS 4600 (SA, 2005) were too high for beams with thinner wall thickness, but
too low for beams with thicker wall thickness. Therefore it was suggested that the
formulations of AS4100 (SA, 1998) is more suitable for the lateral buckling capacity
design of thicker cold-formed steel beams. However, this should be further
investigated as it was based on the results of a few experimental tests.
Ungermann and Kalameya (2006) conducted fourteen experimental tests on thin-

walled channel sections in major axis bending. They have varied parameters such as
loading, slenderness ratio, b/t of the local buckling plates (for flanges and web),
global lateral-torsional buckling slenderness ratio of the beams and the type of
production process. Five types of channel sections have been tested including four
welded sections (Grade S355) and one cold-formed section (Grade S235). Beams
were tested with simply supported boundary conditions and warping about the minor
axis is prevented. The test results showed that the load-carrying behaviour of the
specimens depends strongly on the direction of the global lateral imperfections and
displacements of the member. According the test results, lateral-imperfections and
displacements in the direction of the web (negative y-direction) lead to a sudden
collapse of the beam with a significant decrease in the load while global
imperfections and the increasing displacements due to lateral-torsional buckling in
the direction of the open part of the section lead to a ductile failure with no abrupt
decline of the load. Higher failure moments were observed if the beams failed in the
positive direction comparison to those beams with negative imperfections. Brune and
Ungermann (2008) carried out numerical investigations of channel beams subject to
coupled instabilities. Numerical models were validated using the test results of
Ungermann and Kalameya (2006). They found out that the DSM and EC 3 Part 1.3
overestimate the ultimate loads for thin-walled slender channels in bending.
Linder and Aschinger (1994) checked the applicability of design methods given in
AISI specification, Eurocode 3 Part 1-1 and Part 1-3 for combination of lateral
torsional buckling and local plate buckling. They compared the test results on cold-
formed unlipped and simple lipped channel sections with the design procedures.
They found out that both design methods agree with the test results only if the
rotation of the principle axes between the gross and effective cross-sections is
insignificant. However, they concluded that the unlipped C-sections should not be
designed using these methods.
Pi and Trahair (1995) theoretically investigated the lateral buckling strengths of
cold-formed steel rectangular hollow sections. Finite element model was developed
which used realistic stress-strain curves, residual stresses and initial crookedness and
twist. They found out that the cross-section shape has significant effect on the
lateral-buckling strengths. The comparison of results obtained for RHS beam and I-
section beam showed that the capacities of I section beam is lower than for RHS
beam (Figure 2.14). The use of strain-hardening stress-strain curve instead of a tri-
linear curve for the corners of RHS was found to be less effect on ultimate moment
capacity at high slenderness region while there is moderate effect at low slenderness.
The effects of initial twist were found to be very important while the effects of initial
crookedness were less important than for hot-rolled I-sections. A different design
equation to that used for hot-rolled I-sections was developed for RHS beams.
Figure 2.14: Comparison of Lateral-torsional buckling strengths of RHS and

I-section Beams (Pi and Trahair, 1995)
Chu et al. (2004) proposed an alternative analytical model for predicting the lateral
torsional buckling of thin-walled channel section beams partially laterally restrained
when subjected to an uplift load. The model was constructed for the practical case
where the cold-formed steel member is subjected to transverse loads and is laterally
restrained partially by sheeting and interval anti-sag bars. The focus of this research
was to investigate the influence of the restraints provided by the interval anti-sag
bars and the variation of moment along the longitudinal axis on the lateral torsional
buckling behaviour of the channel section beams. From the numerical results
obtained from this research, the following conclusions were drawn,
(1). Moment gradient has a significant influence on the critical buckling
load of channel section beams. The critical buckling moment in the
pure bending case is less than half of the critical moment (maximum
moment when the lateral torsional buckling occurs) in the uniformly
distributed load case.
(2). The loading position has a remarkable influence on the critical
buckling load of channel section beams. The closer the loading point
to the shear centre, the lower the critical buckling load.
(3). The warping stress does not influence the lateral torsional buckling of
the channel section beams.
2.5 Current Design Rules for Cold-Formed Steel Beams at

Ambient Temperature
2.5.1 General
Design of cold-formed steel members is more complicated compared to hot-rolled

steel members due to many reasons. Some of them are problems associated with the
thinness of the plate elements, mono-symmetric nature, complex shapes, and
changes in material properties due to cold-forming process. Many specifications and
design standards developed in different countries are available for the design of cold-
formed steel members. They are American specification (AISI, 2007), European
Standard (Eurocode 3-Part 1-3), The Australian and New Zealand standard (AS/NZS
4600), British Standard, (BS 5950 Part-5) and Canadian Standard (CAN/CSA S136-
94, 1994). AS/NZS 4600 (SA, 2005) is based on the AISI specification (AISI, 2007).
Some of the design standards are discussed next.
2.5.2 AS/NZS 4600 (SA, 2005)
AS/NZS 4600 (SA, 2005) gives design rules for cold-formed steel members at
ambient temperature only. According to AS/NZS 4600 (SA, 2005), the design
bending moment, M * , should satisfy the following,
(a) M * ≤ φb M s (2.5)
(b) M * ≤ φb M b (2.6)
where,
φb is the capacity reduction factor for bending (Table 1.6 of AS/NZS 4600).
The nominal section moment capacity ( M s ) is given by
M s = Ze f y (2.7)
The effective section modulus ( Z e ) is calculated based on the extreme compression
or tension fibre attaining yield strength ( f y ). The effects of flexural local buckling
are accounted for by using the reduced effective width ( be ) of slender plate elements
at specified locations to calculate the effective section modulus. This approach is
based on the well known effective width method as shown in Figure 2.15.
fmax
be/2 be/2
Figure 2.15: Stress Distribution in Stiffened Compression Elements
For uniformly compressed stiffened elements, the effective width ( be ) is equal to the
flat width of the element ( b ) if λ ≤ 0.673 and otherwise equal to ( ρb ). The

effective width factor ( ρ ) is given as follows:
⎛ 0.22 ⎞
⎜1 − ⎟
⎝ λ ⎠
ρ= ≤ 1.0 (2.8)
λ
The slenderness ratio ( λ ) is determined as follows:
⎛ 1.052 ⎞⎛ b ⎞ f *
λ = ⎜⎜ ⎟⎟⎜ ⎟ (2.9)
⎝ k ⎠⎝ t ⎠ E
In Equation 2.9, k , E , ν and t are the plate buckling coefficient, Young’s

modulus, Poisson’s ratio and the thickness of the uniformly compressed stiffened
elements, respectively. The plate buckling coefficient ( k ) depends on the edge
boundary conditions and the type of stress, and the method of calculation is given in
applicable clauses in Section 2.2 of AS/NZS 4600 (SA, 2005). f * is the design
stress in the compression element calculated on the basis of the effective design
width.
The nominal member moment capacity ( M b ) of the laterally unbraced segments of

singly, doubly, and point-symmetric sections subjected to lateral buckling is given
by,
⎛M ⎞
M b = Zc ⎜ c ⎟ (2.10)
⎜Z ⎟
⎝ f ⎠
Z c and Z f are the effective section modulus calculated at a stress level (M c Z f ) in
the extreme compression fibre, and the full unreduced section modulus for the
extreme compression fibre, respectively.
The critical moment (Mc) can be calculated as follows,

For λb ≤ 0.60 Mc = M y
⎡ ⎛ 10λb2 ⎞⎤
For 0.60 < λb < 1.336 M c = 1.11M y ⎢1 − ⎜⎜ ⎟⎟⎥ (2.11)
⎢⎣ ⎝ 36 ⎠⎥⎦
⎛ 1 ⎞
For λb ≥ 1.336 M c = M y ⎜⎜ 2 ⎟
⎟
⎝ λb ⎠
where, λb = M y M 0 - Non-dimensional slenderness ratio used to
determine M c for members subject to lateral buckling
M y = Z f fy - Moment causing initial yield at the extreme
compression fibre of the full section.
Elastic buckling moment ( M o ) is given in Clause 3.3.3.2.1 of AS/NZS 4600 (SA,

2005) by;
M o = C b Arol f oy f oz (2.12)
π 2E GJ ⎛ π 2 EI w ⎞
where f oy = and f oz = ⎜1 + ⎟
(l ey ry )
2
Arol2 ⎜
⎝ GJl 2 ⎟
ez ⎠
rx and ry are radii of gyration of the cross section about x and y axes. rol is the polar
radius gyration of the section about the shear centre.
Equation 2.12 is similar to the following equation
π 2 EI y ⎡ π 2 EI w ⎤
M o = Cb 2 ⎢GJ + 2 ⎥ (2.13)
l ey ⎢⎣ l ez ⎥⎦
in which, EI y , EI w and GJ are the minor axis rigidity, warping rigidity and
torsional rigidity, respectively. C b in Equation 2.12 and 2.13 is a coefficient

depending on moment distribution in the laterally unbraced segment (see Clause
3.3.3 of AS/NZS 4600).
2.5.3 Eurocode 3 Part 1-3 (ECS, 2006)
Eurocode 3 Part 1-3 (ECS, 2006) only gives design guidelines for cold-formed steel
members at ambient temperature. The effects of local buckling are taken into
account in determining the resistance and stiffness of cold-formed members. Similar
to AS/NZS 4600 (SA, 2005), this is achieved by using effective cross-sectional
properties calculated on the basis of effective widths. The effective widths of
compression elements can be obtained in accordance with EN 1993 Part 1.5.
The section moment capacity of cold-formed steel beam is defined in Clause 6.1.4.1
of Eurocode 3 Part 1.3 as follows,
If Weff < Wel , M c ,Rd = Weff f yb / γ M 0 (2.14)
If Weff = Wel ,
⎛ ⎞⎞
M c , Rd = f yb ⎜⎜Wel + 4(W pl − Wel )⎜1 − λ e max λ e 0 ⎟ ⎟⎟ / γ M 0 ≤ W pl f yb / γ M 0 (2.15)
⎛ _ _
⎝ ⎝ ⎠⎠
Where Weff, Wel and Wpl are the effective section modulus, gross elastic section
modulus and the plastic section modulus, respectively.
_
λ max is taken as the slenderness of the element which correspond to the largest value
_ _ _
of λ e λ e 0 . The plate element slenderness, λ p is defined in EN 1993-1-5.
_ _ _
For double supported plane elements λ e = λ p and λ e 0 = 0.5 + 0.25 − 0.055(3 + ψ )
_ _ _
For outstand elements λ e = λ p and λ e 0 = 0.673
_ _ _
For stiffened elements λ e = λ p and λ e 0 = 0.65
Eurocode 3 Part 1.3 recommends the use of the design equations based on buckling
curve ‘b’ in Eurocode 3 Part 1.1 (ECS, 2005) for the design of cold-formed steel
beams subject to lateral-torsional buckling. These design equations are given as
follows:
M b , Rd = χ LT W y f y γ M 1 (2.16)
in which χ LT is defined as follows:
χ LT = 1
⎧
⎨φ LT
⎩
[
+ φ 2 − λ2
LT LT
]0. 5 ⎫
⎬
⎭
(2.17)
⎡ ⎛_ ⎞ _ 2⎤
φ LT = 0.5 ⎢1 + α LT ⎜ λ LT − 0.2 ⎟ + λ LT ⎥ (2.18)
⎣ ⎝ ⎠ ⎦
[ ]0.5
_
λ LT = f y Wy M o (2.19)
Wy is the appropriate section modulus of the compression flange depending on the
class of cross section (plastic, elastic or effective section modulus). fy in (2.15) and
(2.18) is the yield strength of steel and α LT for buckling curve ‘b’ in (2.17) is equal
to 0.34. Eurocode 3 Part 1.1 gives four buckling curves for lateral-torsional buckling
as given in Table 2.2.
Table 2.2: Recommended Values for Imperfection Factors for Lateral-torsional

Buckling Curves
Buckling Curve a b c d
Imperfection factor ( α LT ) 0.21 0.34 0.49 0.76
2.5.4 BS 5950-Part 5 (BSI, 1998)
BS 5950-Part 5 (BSI, 1998) gives design rules for cold-formed steel members at
ambient temperature. The effect of local buckling is taken into account in the
determination of moment capacity of beams and this is achieved by using the
effective cross-sectional properties which are calculated on the basis of the effective
widths of individual elements (refer Section 4 of BS 5950-Part 5). The moment
capacity of laterally braced cold-formed steel beams is calculated based on the
limiting compressive stress in the webs ( p o ) and is given by,
M c = p o Z eff (2.20)
The limiting compressive stress ( p o ) in the webs is also used in the evaluation of the
effective widths of compression elements, and hence the effective section modulus
( Z eff ). Details of determining the effective widths are given in Section 4 of BS5950-
Part 5 (BSI, 1998). The limiting compressive stress ( p o ) in a stiffened element

which results from bending in its plane should not exceed the lesser of the following
values.
⎧⎪ Dw ⎛ Ys ⎞ ⎫⎪
12
p o = ⎨1.13 − 0.0019 ⎜ ⎟ ⎬ py or (2.21)

⎪⎩ t ⎝ 280 ⎠ ⎪⎭
po = p y
In Equation 2.21, Dw is the section depth or twice the depth of the compression
zone, Dc , whichever is the greater in mm. Ys , t and p y are the material yield
strength, web thickness in mm and the design strength in MPa.
The lateral torsional buckling resistance moment ( M b ) is given by
M EMY
Mb = ≤ Mc (2.22)
φ B + φ B2 − M E M Y
M Y + (1 + η )M E
where, φB = (2.23)
2
In Equation 2.21, M c and M y are the moment capacity and the yield moment of the
section, respectively. Yield moment ( M y ) of the section is the product of the design
strength ( p y ) and the elastic modulus of the gross section with respect to the
compression flange ( Z c ). The elastic lateral buckling resistance moment ( M E ) in

Equation 2.23 for channel section beams bent in the plane of the web is given in
Clause 5.6.2.2 by,
12
π AED2 ⎧ 1 ⎛⎜ LE t ⎞⎟ ⎫⎪
2
⎪
ME = C b ⎨1 + (2.24)
2(LE ry )
2 ⎜ ⎟ ⎬
⎪⎩ 20 ⎝ ry D ⎠ ⎪⎭
In Equation 2.24 A , D , t , Le and ry are the cross-sectional area, overall depth of
the cross-section, material thickness, effective length and the radius of gyration of
the section about the y axis, respectively.
η in Equation 2.23 is the Perry coefficient, such that:
when Le ry < 40C b η =0
⎛ Le ⎞
when Le ry > 40C b η = 0.002⎜⎜ − 40C b ⎟
⎟
(2.25)
⎝ ry ⎠
C b is a coefficient which may be conservatively assumed to be unity or can be

calculated using,
C b = 1.75 − 1.05β + 0.3β 2 (2.26)
β is the ratio of the smaller end moment to the larger end moment M in the
unbraced length of a beam.
The design methods given in AS/NZS 4600 (SA, 2005), Eurocode 3-Part 1-3 (ECS,
2006) and BS 5950-Part5 (BSI, 1998) may be used to design the cold-formed steel
beams at elevated temperatures by replacing the mechanical properties at ambient
temperature with the reduced mechanical properties at elevated temperatures. But the
accuracy of these design methods based on reduced mechanical properties is not
known.
2.5.5 Direct Strength Method
The nominal member moment capacity M b is taken as the minimum of the nominal
member moment capacities for lateral-torsional buckling (M be ) , local buckling
(M bl ) and distortional buckling (M bd ) .
The nominal member moment capacity, M be for lateral-torsional buckling is given

by
For M o < 0.56M y M be = M 0
10 ⎡ ⎛ 10M y ⎞⎤
For 2.78M y ≥ M o ≥ 0.56M y M be = M y ⎢1 − ⎜⎜ ⎟⎟⎥ (2.27)
9 ⎢⎣ ⎝ 36 M o ⎠⎥⎦
For M o > 2.78M y M be = M y
where,
M o - elastic lateral-torsional buckling moment is same as Equation 2.13.
M y = Z f f y - Moment causing initial yield at the extreme compression fibre
of the full section.
The nominal member moment capacity, M bl for local buckling is given by
For λl ≤ 0.776 M bl = M be
⎡ ⎛M ⎞
0.4
⎤⎛ M ⎞
0.4
For λl > 0.776 M bl = ⎢1 − 0.15⎜⎜ ol ⎟⎟ ⎥⎜⎜ ol ⎟⎟ M be (2.28)

⎢⎣ ⎝ M be ⎠ ⎥⎦⎝ M be ⎠
where,
M be
λl = - Non dimensional slenderness used to determine M bl
M ol
M ol = Z f f ol - Elastic local buckling moment
The nominal member moment capacity, M bd for distortional buckling is given by
For λl ≤ 0.673 M b = M be
⎡ ⎛M ⎞
0.5
⎤⎛ M ⎞
0.5
For λl > 0.673 M bd = ⎢1 − 0.22⎜ od ⎟ ⎥⎜ od ⎟ My (2.29)

⎢ ⎜M ⎟ ⎥⎜⎝ M y ⎟
⎣ ⎝ y ⎠ ⎦ ⎠
where,
My
λd = - Non dimensional slenderness used to determine M bd
M od
M od = Z f f od - Elastic distortional buckling moment
2.6 Fire Safety
2.6.1 General
Buildings can catch fire accidentally or purposely which may cause heavy loss of
lives and property. Therefore ensuring the fire safety of structures is very important.
Fire deaths and property loss can be eliminated by preventing fires. However, it is
not possible to totally eliminate the fires and only the probability of occurrence can
be reduced. The main elements in fire safety engineering are providing facilities to
detention and notification of fires, providing safe escape for occupants and fire
fighters, providing barriers to control the spread of fire and smoke, and prevention of
structural collapse.
2.6.2 Fire Behaviour
The typical development of fire in a room is shown in Figure 2.16 and it can be
divided into four phases: incipient period, growth period (also known as pre-
flashover), burning period (also known as post flashover) and the decay period.
Figure 2.16: Time-Temperature Curve for Typical Fire Development Situations

(Buchanan, 2001)
In the incipient period, heating of potential fuel is taking place leading to ignition
which is the transition to the growth phase. In the growth phase temperature rise is
very small and at the end of the growth phase, temperature increases rapidly leading
to flashover which is the transition to the burning period. It is during the burning
period that the structural elements are exposed to the worst effects of fire and where
collapse or loss of integrity is likely (Purkiss, 1996). In the decay period, temperature
decreases as the rate of fuel combustion decreases.
However, the fire can be controlled by taking certain measures that prevent reaching
the burning period. Fire control can be divided into two groups: active measures and
passive measures. In the case of active measures, fire control is achieved by some
action taken by a person or an automatic device. The best active measure for fire
control is the automatic sprinkler system which sprays water over the area under fire.
In the case of passive control, effects of fire are controlled by systems that are built
into the structure or fabric of the structure, not requiring operation by people or
automatic controls (Buchanan, 2001).
2.6.3 Design for Fire Safety
Fire safety of the structure or each part of the structure is verified if the fire
resistance is greater than the fire severity. A general definition for the fire resistance
of construction elements is defined as the time after which an element, when
submitted to the action of a fire, ceases to fulfil the functions for which it has been
designed (Rodrigues et al., 2000). Fire severity is a measure of the destructive impact
of a fire, or a measure of the forces or temperatures which could cause collapse or
other failure as a result of the fire (Buchanan, 2001). There are three different
methods of comparing fire severity with fire resistance: time domain, temperature
domain, and strength domain. In the case of time domain, time to failure of the
element ( t fail ) is compared with the fire duration ( t s ) such that it satisfies, t fail ≥ t s .
To verify the fire safety in the temperature domain, the maximum temperature
( Tmax ,oC) in a part of the structure is compared with the temperature which would
cause failure ( T fail , oC) such that T fail ≥ Tmax . Fire safety in the strength domain is
verified by comparing the applied load ( U f ) at the time of the fire with the
minimum load capacity ( R f ) of structural members throughout the fire such that it
satisfies R f ≥ U f .
2.6.4 Effect of Fire on Steel Structures
When steel structures are exposed to fire, temperature increase in the structural
members causes severe reduction of strength and stiffness which can lead to
structural collapse. When a fire occurs in a building, the steel members could be
heated to temperatures in excess of 700oC and may collapse if the applied stress on
them equalled, or exceeded, the reduced strength of steel at elevated temperatures
(Kirby and Preston, 1988). The temperature increase in steel depends on the fire
severity, the area of steel exposed to fire and the amount of applied fire protection.
Several factors influence the behaviour of steel structures exposed to fire: the
elevated temperatures in the steel members, applied loads, the mechanical properties
at elevated temperatures, connections and end restraints. In the next section, details
of how the mechanical properties vary with the temperature increase are presented.
2.7 Mechanical Properties of Cold-Formed Steels at Elevated

Temperatures
The material properties that affect the fire performance of cold-formed steel
structures such as yield strength, elasticity modulus and coefficient of thermal
expansion play an important role in the fire safety design of cold-formed steel
structures. These properties are temperature dependent and sensitive to parameters
such as heating rate, strain rate and temperature gradient of the testing method.
BS 5950-Part 8 (BSI, 1990): Code of Practice for fire resistance design provides
strength reduction factors for cold-formed steel at elevated temperatures ranging
from 2000C to 6000C for yield stresses corresponding to 0.5, 1.5 and 2.0 percent
strain levels (see Table 2.3), even though the common practice is to use 0.2% proof
stress as the corresponding yield stress.
Table 2.3: Strength Reduction Factors for Cold-Formed Steels
BS5950-Part 8 (BSI, 1990)
Strain Temperature (oC)

200 250 300 350 400 450 500 550 600
%
0.5 0.945 0.890 0.834 0.758 0.680 0.575 0.471 0.370 0.269
1.5 1.000 0.985 0.49 0.883 0.815 0.685 0.556 0.453 0.349
2.0 1.000 1.000 1.000 0.935 0.867 0.730 0.590 0.490 0.390
Eurocode 3 Part 1-2 (ECS, 2005): “General rules for structural fire design” provides
reduction factors for design strength and modulus of elasticity at elevated
temperatures for both cold-formed and hot-rolled thin-walled carbon steel. However,
there is no difference between the reduction factors of cold-formed and hot-rolled
steels (see Table 2.4). Even though the code gives the same reduction factors for
both cold-formed and hot-rolled steels, it is not acceptable, because it has been found
that the behaviour of cold-formed steel members at elevated temperatures is different
from that of hot-rolled steel members. Further, the code recommends the calculation
of strength and deformation properties of both cold-formed and hot-rolled steels at
elevated temperatures based on the stress-strain relationship given in Figure 2.17 for
heating rates between 2 and 50 K/min. The code also suggests using 0.2% proof
stress as the design strength of steel structures under fire conditions. Further, it is
noted that the yield strength is constant until 400oC according to Eurocode 3 Part 1-2
(ECS, 2005). But several research projects conducted on cold-formed steels reveal
that the yield strength starts to decrease earlier at lower temperatures.
Figure 2.17: Stress-Strain Relationship for Steel at Elevated Temperatures

(Eurocode 3 Part 1-2 (ECS, 2005))
Table 2.4: Reduction Factors for the Mechanical Properties of Steels at
Elevated Temperatures (Eurocode 3 Part 1-2 (ECS, 2005))
Steel Reduction factors for cold-formed and hot-rolled steels relative to the
temperature value of f y or E a at 20oC
( θ a , oC) Effective yield strength Modulus of elasticity
k y ,θ = f y ,θ f y k E ,θ = E a ,θ E a
20 1.000 1.000
100 1.000 1.000
200 1.000 0.900
300 1.000 0.800
400 1.000 0.700
500 0.780 0.600
600 0.470 0.310
700 0.230 0.130
800 0.110 0.090
900 0.060 0.0675
1000 0.040 0.0450
1100 0.020 0.0225
1200 0.000 0.0000
Many researchers carried out experimental tests on mechanical properties of cold-

formed steel at elevated temperatures. There are two types of test methods to
investigate the fire resistance of a steel member under fire conditions: the steady
state test and the transient state test. In the steady state test method, the loading is
commenced and continued to increase until failure after the temperature of the test
specimen stabilized at the required value. The steady state tests can be carried out
either as strain-controlled or load-controlled. In the strain-controlled tests, the strain
rate is kept constant whereas in the load-controlled tests, the loading rate is kept
constant. On the other hand, in the transient state test method, the test specimen is
kept under constant load and under a constant temperature rise until failure. The
steady state tests are easier to carry out than the transient state tests, however, the
transient state tests seem to give more realistic test results, especially for low carbon
structural steel (Outinen and Makelinen, 2002a).
Gerlich (1996) derived the following expressions for yield strength ( f y ,T ) and
modulus of elasticity ( ET ) for the temperature range below 650oC using data from
several steady state tests on cold-formed steel framing members presented by
Klippstein (1980). In Equations 2.30 and 2.31, T , f y and E are the steel
temperature in oC, yield strength and the modulus of elasticity at room temperature,
respectively.
f y ,T f y = 1.0 − 5.3 × 10 −4 T + 4.0 × 10 −6 T 2 − 1.9 × 10 −8 T 3 + 1.7 × 10 −11 T 4 (2.30)
ET E = 1.0 − 3.0 × 10 −4 T + 3.7 × 10 −7 T 2 − 6.1 × 10 −9 T 3 + 5.4 × 10 −12 T 4 (2.31)
An extensive experimental research program was carried out at the Helsinki

University of Technology to investigate the mechanical properties of various
structural steels at elevated temperatures. Outinen and Makelainen (2000 and 2001)
recorded the results of the mechanical properties of the studied steel types. The steel
grades included cold-rolled hot dip zinc coated structural steel (S350GD+Z), hot-
rolled structural steel (S355), high-strength steel (S460M) and S355J2H steel taken
from SHS sections. The mechanical properties of cold-rolled steel (S350GD+Z) are
also recorded in Outinen and Makelainen (2002a,b). For the cold-formed steels, the
mechanical properties do not match with that given in Eurocode 3 Part 1-2 (ECS,
1995) while there was good agreement between Eurocode 3 Part 1-2 (ECS, 1995)
and the results of hot-rolled (S355) and high-strength (S460M) steels. Therefore they
proposed reduction factors for yield strength and modulus of elasticity of cold-
formed steels (S350GD+Z) at elevated temperatures (see Table 2.5).
Table 2.5: Reduction Factors for Mechanical Properties of S350GD+Z Steel
Proposed by Outinen et al. (2002a,b)
Steel Temperature, Reduction factors for modulus Reduction factors for yield
(oC) of elasticity, k E ,θ = E a ,θ E a strength, k y ,θ = f y ,θ f y
20 1.000 1.000
100 1.000 0.970
200 0.900 0.932
300 0.800 0.895
400 0.700 0.857
500 0.600 0.619
600 0.310 0.381
700 0.130 0.143
800 0.090 0.105
900 0.06 0.067
950 0.056 0.048
1000 0.045 0.029
Chen and Young (2004) investigated the mechanical properties of cold-formed steels
at elevated temperatures for two steel grades of G550 and G500 with nominal yield
strength of 550 and 500 MPa, respectively. The plate thicknesses were 1.0 mm and
1.9 mm. They proposed a unified equation to determine the yield strength of G550
and G500 cold-formed steel for temperatures ranging from 22oC to 1000oC. In
addition to that, a unified equation was also proposed to determine the modulus of
elasticity of G500 cold-formed steels for temperatures ranging from 22oC to 650oC
based on transient state test results. Proposed equations are given in Table 2.6.
Table 2.6: Reduction Factors for Mechanical Properties of Flat Regions of
Cold-Formed Steel at Elevated Temperatures (Chen and Young, 2004)
Temperature
Reduction factors Grade 22≤T≤300 300≤T≤450 450≤T≤1000
(oC)
a 1.0 0.9 0.021.0
G550 b 22 300 1000
For yield stress,
1.0 mm c 2.78x103 4.8x10 6
9x108
f y ,T
=a−
(T − b )n n 1 3 3
f y ,normal c Temperature 22≤T≤300 300≤T≤650 650≤T≤1000
G500 a 1.0 0.95 0.105
b 22 300 650
1.9 mm
c 5.56x103 1.45x10 5
5x103
n 1 2 1
Temperature 22≤T≤450 450≤T≤650
For elastic modulus,
G500 a 1.0 -0.11
ET
=a−
(T − b)
n
1.9 mm
b 22 860
E normal c c 1.25x103 -2.2x105
n 1 2
Chen and Young (2006) conducted an experimental investigation to study the

changes in the mechanical properties of corner regions of cold-formed steel sections
at elevated temperatures. Two kinds of corner coupon specimens, namely the inner
and outer corner coupon specimens having the steel grade of G500 (nominal 0.2%
proof stress of 500 MPa) and nominal thickness of 1.9 mm were tested at different
temperatures ranging from 20 to 1000oC. They showed that the reduction factors of
yield strength and elastic modulus of the corner coupon specimens for G500 steel are
exactly the same as those of the flat coupon specimens. Their proposed equations for
yield stress, elastic modulus, ultimate strength and ultimate strain of cold-formed
steel inner and outer coupon specimens are given in Table 2.7.
Table 2.7: Reduction Factors for Mechanical Properties of Corner Regions of
Cold-Formed Steel at Elevated Temperatures (Chen and Young, 2006)
Reduction factors Temperature (oC) a b c n
For yield stress,

22 ≤ T < 300 1.0 22 5.56x103 1
300 ≤ T < 650 0.95 300 1.45x105 2
f y ,T f y ,normal = a − (T − b ) c
n
650 ≤ T < 1000 0.105 650 5.00x103 1
For elastic modulus, 22 ≤ T < 450 1.0 22 1.25x103 1
ET E normal = a − (T − b ) c
n 450 ≤ T < 650 -0.11 860 -2.20x105 2
For ultimate strength, 22 ≤ T < 450 1.0 22 5.6x108 3
f u ,T f u ,normal = a − (T − b ) c
n 450 ≤ T < 1000 0.043 1000 -1.12x1011 4
For ultimate strain,

22 ≤ T < 1000 1.0 22 1.0x106 2
ε u ,T ε u ,,normal = a − (T − b ) c
n
Ranawaka and Mahendran (2009a) developed a set of empirical equations to

determine the yield strength and the elasticity modulus of thin (≤1 mm) cold-formed
steels for temperatures ranging from 20oC to 800oC. These empirical equations were
developed for the reduction factors based on the 0.2% proof stress method. Separate
sets of equations for the reduction factors ( f y ,T f y , 20 ) have been developed for both
low and high strength steels based on various temperatures, where f y ,T and f y , 20 are
the 0.2% proof stress at elevated and ambient temperatures, respectively. In the case
of reduction factors for elastic modulus ( ET E 20 ) , the same set of equations has been
proposed for both high strength and low strength steels because the difference was
negligible, where ET and E 20 are the elastic modulii at elevated and ambient
temperatures, respectively. Ranawaka and Mahendran’s (2009a) predictive equations
for the reduction factors of mechanical properties of thin cold-formed steels are
given next.
Reduction factors for yield strength at temperature T are given by,
For G550 steels

f y ,T
= −0.00016T + 1.0003 20 o C ≤ T ≤ 200 o C
f y , 20
f y ,T
= 0.97 −
(T − 200 )
1.81
200 o C < T < 600 o C (2.32)
f y , 20 58500
f y ,T
= −0.00037T + 0.3363 600 o C ≤ T ≤ 800 o C
f y , 20
For G250 steels

f y ,T
= −0.0007T + 1.014 20 o C ≤ T ≤ 200 o C
f y , 20
f y ,T
= 3 .7 −
(T − 74)0.15 200 o C < T ≤ 800 o C (2.33)
f y , 20 0.736
Reduction factors for elastic modulus at temperature T are given by,

ET
=1 20 o C ≤ T ≤ 100 o C
E 20
ET
= −0.0013T + 1.1297 100 o C < T ≤ 800 o C (2.34)
E 20
In addition to the reduction of mechanical properties, cold-formed steel expands with

increasing temperature. The coefficient of thermal expansion at ambient temperature
is 12x10-6/0C, but it changes with increasing temperature. Lie (1992) provides an
expression for temperature dependent thermal expansion coefficient ( α T ), which
was used by Gerlich (1996) in his research, as follows.
For T < 1000o C , αT = (0.0004T + 12) × 10−6 (2.35)
Temperature rise in a steel member is dependent on thermal properties of material

such as thermal conductivity, specific heat, and density. Lie (1992) presented an
approximation for the thermal conductivity as a function of temperature and also
suggested a constant value of 600 J/kg0C for specific heat for temperatures below
6000C.
Ramberg and Osgood (1943) proposed a simple formula for describing the stress-
strain curve at ambient temperature in terms of three parameters: modulus of
elasticity and two secant yield strengths. Olawale and Plank (1988) and Outinen et
al. (1999) proposed stress-strain models for hot-rolled steels at elevated temperature
based on Ramberg and Osgood stress-strain model. Lee et al. (2003), Ranawaka and
Mahendran (2009) and Chen and Young (2007) proposed models for stress-strain
curves for cold-formed steel at elevated temperature based on Ramberg-Osgood
stress-strain model. The stress-strain model for elevated temperature based on
Ramberg-Osgood (1943) is usually in the form of Equation 2.36 as proposed by
various researchers.
ηT
f ⎛ f y ,T ⎞⎛ f T ⎞
ε T = T + β ⎜⎜ ⎟⎟⎜ ⎟ (2.36)
ET ⎜ ⎟
⎝ T ⎠⎝ f y ,T
E ⎠
ε T in Equation 2.36 is the strain corresponding to f T , stress at elevated temperature,

T while ET and f y ,T are modulus of elasticity and yield strength at temperature T ,
respectively. Various authors have modified the parameter β and η T depending on

the steel type. The parameter β was taken as a constant of 3/7 by Olawale and Plank
(1988), 6/7 by Outinen (1999) and 0.86 by Ranawaka and Mahendran (2009) while
Lee et al. (2003) considered it as a temperature dependent parameter. The parameter
η T was considered as a temperature dependent by Olawale and Plank (1988),
Outinen (1999) and Ranawaka (2006) whereas Lee et al. (2003) considered it as a
constant.
Ranawaka and Mahendran (2009a) proposed separate equations for the parameter
η T for low grade and high grade as given in Equation 2.37 and 2.38, respectively.
For G250 steel,
350 o C ≤ T ≤ 800 o C
η T = 0.000138T 2 − 0.085468T + 19.212 (2.37)
For G550 steel,

20 o C ≤ T ≤ 800 o C
η T = −3.05 x10 −7 T 3 + 0.0005T 2 − 0.2615T + 62.653 (2.38)
The model proposed by Chen and Young (2007) is different from Equation 2.36.
Their proposed model for describing stress-strain curve of cold-formed steel is based
on the stress-strain curve model for stainless steel at normal room temperature
proposed by Mirambelll and Real (2000) and Rasmussen ( 2003) which is also
originally based on the Ramberg-Osgood (1943) equation. Proposed equation by
Chen and Young (2007) is given in Equation 2.39.
For f T ≤ f y ,T ,
nT
f ⎛ f ⎞
ε T = T + 0.002⎜⎜ T ⎟
⎟ (2.39a)
ET ⎝ f y ,T ⎠
For f T > f y ,T ,
mT
⎛ f T − f y ,T ⎞ ⎛ f − f y ,T ⎞
ε T = ⎜⎜ ⎟ + ε u ,T ⎜ T
⎟ ⎜f −f
⎟
⎟
+ ε y ,T (2.39b)
⎝ E y ,T ⎠ ⎝ u ,T y ,T ⎠
ET
and E y ,T =
1+ 0.002nT ET f y ,T
nT = 20 − 0.6 T
mT = 1 + T 350
ε
Where ε T is the strain at temperature, T C , u,T is the strain corresponding to yield
o
f
strength at temperature, T C , f T is the stress at temperature, T C , u ,T is the
o o
o f
ultimate strength at temperature, T C , y.,T is the yield strength at temperature,
T o C , ET is the elastic modulus at temperature, T o C E y ,T is the elastic modulus at

o
yield strength at temperatures, T C .
2.8 Flexural Behaviour of Steel Beams at Elevated Temperatures
When subject to fire, an unprotected steel beam looses its stiffness and strength as a
result of deterioration of mechanical properties. This will increase the complexity of
the behaviour of cold-formed steel beams under fire. However, the buckling
behaviour of cold-formed steel beams under fire conditions has not been researched
adequately. Even though past research has focused on the behaviour of hot-rolled
steel beams at elevated temperatures, no research has been undertaken to study the
behaviour of cold-formed steel beams at elevated temperatures.
When calculating the load carrying capacity of cold-formed steel beams at elevated
temperatures, it is necessary to take into account the influence of local buckling.
Ranby (1998) showed that the calculation of the plate buckling resistance based on
Eurocode 3 Part 1-3 (ECS, 2006) with reduced yield strength and elastic modulus is
sufficiently accurate at elevated temperatures, provided that the yield strength is
taken as the 0.2 percent proof stress.
Kitipornachai and Trahair (1975) performed an experimental investigation of the

inelastic flexural torsional buckling of hot-rolled steel I-beams. Tests were carried
out on full-scale simply-supported beams with central concentrated loads applied
with a gravity load simulator. Their test results were used by many researchers for
validating their numerical models.
Bailey et al. (1996) performed numerical analyses of uniformly heated unrestrained

hot-rolled steel beams subject to lateral torsional buckling. They selected six
different serial sizes of UB and UC sections with spans giving a wide range of ratios
of lateral torsional buckling resistance moment to the moment capacity of the section
about its major axis, (M b M cx ) which represent both elastic and inelastic lateral
torsional buckling failure modes at ambient temperature. They used the results of
Kitipornchai and Trahair (1975) to validate the model for inelastic lateral torsional
buckling. The analysis was carried out under both stabilized loading in which the
loading was at the shear centre and destabilizing loading in which the loading was at
the top flange. In the case of loading at the shear centre, different types of loading
conditions were used to give a wide range of bending moment distribution while in
the case of loading at the top flange only uniformly distributed loading condition was
used. The results indicated that both the design codes BS5950 Part 8 (BSI, 1990) and
Eurocode 3 Part 1-2 (ECS, 1995) over-estimate the limiting temperatures for
unrestrained simple beams in fire resistance calculations. However, no analytical
proposal has been made for the lateral torsional buckling resistance moment in fire
situation.
Yin and Wang (2003) presented the results of a numerical study of the lateral
torsional buckling behaviour and strengths of hot-rolled steel beams with non-
uniform temperature distributions. They investigated the effects of factors such as
beam slenderness, type of non-uniform temperature distribution, load type, position
and magnitude. They challenged the design methods in BS 5950 Part 8 (BSI, 1990)
and EC 3 Part 1-2 (ECS, 2002) which only consider the effect of non-uniform
temperature distributions in a steel beam for the beam’s plastic bending moment
capacity and they suggested that the beam’s slenderness is also affected. They
concluded that the design methods given in BS5950 Part 8 (BSI, 1990) and
Eurocode 3 Part 1-2 (ECS, 2002) predict much lower beam failure temperatures and
hence neither method is suitable to deal with the effects of non-uniform temperature
distributions. They also observed that finite element analysis predicts lower failure
temperatures for beams under uniform temperature distribution than BS5950-Part 8
(BSI, 1990). This was also pointed out by Bailey et al. (1996). The reason for this
was that the design method does not consider the non-linear stress-strain
relationships of steel at elevated temperatures.
Real et al. (2003) carried out a numerical study and proposed an alternative
expression for lateral torsional buckling resistance of unrestrained hot-rolled steel
beams with uniform temperature distribution across the cross-section. The proposed
method was validated using the results of 120 experimental and numerical tests
performed on IPE100 beams subjected to temperatures varying from room
temperature to 600oC. They clearly showed that the previous expression in Eurocode
3 Part 1-2 (ECS, 1995) for the case of a simply supported beam with fork supports
under uniform bending moment is unsafe for a certain range of slenderness. The
proposed change in the design method for lateral torsional buckling has been adopted
in the current version of Eurocode, Eurocode 3 Part 1-2 (ECS, 2002) (see Section
2.9.2).
However, Real et al. (2004b) showed that the beam design curve from Eurocode 3
Part 1-2 (ECS, 2002) method which was proposed by Real et al. (2003) is over-
conservative for loadings other than uniform bending and therefore proposed a new
method to check the lateral torsional buckling of beams under fire conditions by
adopting the newly proposed methodology for ambient temperature design to
elevated temperature design. A simply supported beam with fork supports was used
in the study and three values, (-1, 0, 1) for the ratio of bending moment, ψ (see
Figure 2.18) were investigated. A large number of numerical tests was conducted
using the finite element code for the geometrical and material non-linear analysis,
SAFIR. The selected section for the numerical tests was an IPE 220 steel section of
grade S235. Uniform temperatures in the cross-section of 400, 500, 600 and 700oC
were used, so that the comparison between numerical results and Eurocode
predictions were possible.
Figure 2.18: Simply Supported Beam (Real et al., 2004b)
The design method for lateral torsional buckling at uniform elevated temperatures
given in Eurocode 3 Part 1-2 (ECS, 2002) was found to give over-conservative
results for the case of non-uniform bending. The main factor responsible for the
over-conservative nature of lateral torsional buckling resistance calculated from
Eurocode 3 Part 1-2 (ECS, 2002) was identified as the loading type. Therefore a
modified reduction factor for lateral torsional buckling, χ LT , fi ,mod was proposed and
is given by,
χ LT , fi
χ LT , fi ,mod = but χ LT , fi ,mod ≤ 1 (2.40)
f
where, f = 1 − 0.5(1 − k c )
k c is a moment correction factor and its values for some common loading cases are
given in Table 2.8.
Table 2.8: Moment Correction Factors k c for the New Proposal

(Real et al., 2004b)
For other bending diagrams kc=1
Real et al. (2006) also found that the present method for lateral torsional buckling
resistance of unrestrained hot-rolled steel beams under fire conditions incorporated
in Eurocode 3 Part 1-2 (ECS, 2002) gives unsafe values for some types of profiles
and presented a new proposal which takes into account the influence of the steel
grade, the type of cross-section and the residual stresses. An improved proposal was
presented for the lateral torsional buckling of unrestrained steel beams subjected to
fire based on the newly proposed method in Eurocode 3 Part 1-1 (ECS, 2002). They
proposed the imperfection factor, α , which is given as equal to 0.65 235 f y in
Eurocode 3 Part 1-2 (ECS, 2002) to be rewritten as a function of severity factor in
order to account for the steel grade, the type of cross-section and the effect of
residual stresses as shown next,
α = β 235 f y (2.41)
The severity factor β is given in Table 2.9.
Table 2.9: Values of the Severity Factor, β (Real et al., 2006)
β
Cross-section Limits
S235, S237, S355, S420 S460
h b≤2 0.65 0.70
Rolled I section
h b>2 0.75 0.80
h b≤2 0.70 0.75

Welded I section
h b>2 0.80 0.85
Other cross-sections - 0.80 0.85
Lopes et al. (2006) found out that the formula for lateral torsional buckling resistance
given for carbon steel in Eurocode 3 Part 1-2 (ECS, 2002) does not give safe results
for stainless steels at high temperatures for some cases, even though the code
recommends designing the stainless steel members using the same formulae. They
proposed different severity factors ( β ) to be used in Equation 2.41 for the lateral
torsional buckling of stainless steel welded I-section, instead of those values given in
Table 2.9 for carbon steel.
Dharma and Tan (2007) proposed an alternative approach to predict the lateral
torsional buckling load of unrestrained hot-rolled steel beams under fire. The
proposed method is different from Eurocode 3 Part 1-2 (ECS, 2002) and provides an
accurate, safe and smooth transition from ambient temperature designs to elevated
temperature designs. The authors also provided a simple interaction formula widely
known as Rankine approach for quick calculations and more conservative
predictions to the lateral torsional buckling moment under fire conditions. The
authors used the results of Kitipornchai and Trahair (1975) to validate their finite
element model. The results of finite element analysis were used to validate both
methods: alternative approach and Rankine approach. They considered the
temperature to be uniform across the cross-section. Their proposed alternative
approach is given in Section 2.9.3.
Dharma and Tan (2007) conducted an extensive experimental program of steel I-

beams subjected to hogging bending moment by varying four parameters:
temperature, flange slenderness, web slenderness and effective length. The
objectives were to study the temperature effects and to identify key parameters
which affect the rotational capacity. Out of nine tested beams, three were welded
sections while the rest were UB sections. The test results revealed that the rotational
capacity reduces at elevated temperatures. Furthermore, the results indicated that
shorter effective length provides greater rotational capacity and change the failure
mode from lateral torsional buckling to local flange and web buckling.
All the above discussed research projects were conducted on hot-rolled steel beams
and stainless steel beams. It is thus clear that the behaviour of individual cold-formed
steel beams have not been studied yet. However, the behaviour of floor assemblies
made of cold-formed steel beams has been studied at elevated temperatures by
several researchers (Alfawakhiri et al., 2001, Kaitila, 2002, Sultan, 2004).
Alfawakhiri et al. (2001) investigated the behaviour of lightweight steel framed

unrestrained floor assemblies. The parameters investigated in the test series were
joist spacing, number of gypsum board layers in the ceiling membrane, floor cavity
installation and the presence of concrete topping in the sub floor. All the steel joists
were C-shaped cross-section having 203 mm depth, 41.3 mm width, 12.5 mm lips
and 1.21 mm thickness. Galvanised floor joists were spanned in the shorter
dimension so that the effective length was equal to 3.8 m. It was observed that the
dominant structural failure mode was the local buckling mode of joist top flange near
mid span. The heating of joists in the non-insulated floor assemblies appeared to be
closer to uniform heating. Development of higher temperature gradient in the
insulated floor assemblies causes large deflections due to associated thermal bowing
of joists compared to the non-insulated floor assemblies. Furthermore, it has been
observed that the steel joists can maintain their load bearing capacity even when the
bottom flange temperature exceeds 800oC. Therefore temperature rise in the top
flange appears to be more critical.
Kaitila (2002) presented numerical modelling of simply supported cold-formed steel
floor beams under fire conditions using the results of tests conducted at VTT
Building and Transport. He has not provided the test results for cold-formed steel
floor beams and only the method of testing was given. In the tests, cold-formed steel
floor beams were protected with two layers of gypsum plaster boards and 70 mm
thick glass wool insulation in the floor cavity. However, in the finite element
modelling only the beams were modelled simulating the transient state test method
by introducing temperature versus time field and the effect of having plaster boards
was taken into account by applying a temperature gradient over the cross-section. He
only provides the details of modelling and has not extensively studied the behaviour
of beams using the developed model.
Sultan (2004) presented the results of 28 full scale fire resistance test assemblies
conducted in accordance with ULC-S101 standard on wood joist, wood I-joist, steel
C-joist and wood truss floor systems as part of a collaborative research program on
the fire and acoustical performance of floor assemblies. Parameters investigated
were the effects of attachment of the gypsum board layers, joist spacing, resilient
channel spacing, sub floor type and structural load on the fire resistance performance
of floor assemblies. The results of this study were used as the basis for the fire rated
floor assemblies published in Part 9 Appendix A of the National Building Code of
Canada (NBCC) which included 200 fire rated floor assemblies. Twenty eight floor
assemblies of 4.8 m long by 3.9 m wide were constructed and tested in accordance
with CAN/CSA-A82.31-M91 (1991) in this study. They showed that the installation
of glass fibre insulation in the floor cavity with two layers of gypsum boards reduced
the fire resistance by 8% compared to an assembly with no insulation in the floor
cavity.
2.9 Current Design Rules for Cold-Formed and Hot-Rolled Steel

Beams at Elevated Temperatures
The design standards, BS5950 Part 8 (BSI, 1990) and Eurocode 3 Part 1-2 (ECS,
2005), provide design guidelines to determine the capacity of structural elements
under fire situations. However, BS 5950 Part 8 (BSI, 1990) only provides design
guidelines for hot-rolled steel members while Eurocode 3 Part 1-2 (ECS, 2002)
provides the same design guidelines for both hot-rolled and cold-formed steel
members under fire conditions.
2.9.1 BS 5950 Part 8 (BSI, 1990)
BS 5950 Part 8 (BSI, 1990) considers the structural effects of a fire in a building or
part of a building as one of the design limit states. It gives two methods in
determining the fire behaviour of hot-rolled steel members: the limiting temperature
method and the moment capacity method. In the limiting temperature method, the
load ratio which is defined as the ratio of the load carried during the fire to the load
capacity at 20oC, is calculated first. Secondly, the limiting temperature, which is a
function of load ratio, is determined based on Table 2.10. The load ratio, R, for hot-
rolled beams can be taken as the greater of:
Mf mM f
R= or R= (2.42)
Mc Mb
M f and M b are the applied moment at the fire limit state and the lateral torsional
buckling resistance moment, respectively. M c is M cx or M cy as appropriate to the
axis of bending, where they are the section moment capacity about the major and
minor axes in the absence of axial load. m is the equivalent uniform moment factor.
The maximum permissible temperature or the limiting temperature of a member

depends on its type, its loading characteristics, and the load ratio. In this method, the
effect of slenderness has been considered in the calculation of load ratio by dividing
the applied bending moment in fire by the lateral torsional buckling resistance
moment at ambient temperature. However, this is accurate only where the beam is
under uniform heating and when the strength and stiffness of steel change at the
same rate at elevated temperatures (Yin et al. 2003).
Table 2.10: Limiting Temperatures for the Design of Protected and
Unprotected Beams (BS 5950-Part 8 (BSI, 1990))
Limiting temperature at a load

Description of Member ratio of:
0.7 0.6 0.5 0.4 0.3 0.2
Members in bending not supporting
concrete slabs:
Unprotected members, or protected
members complying with item a) or b)
520 555 585 620 660 715
of Clause 2.3 of BS5950-Part 8
Other protected members 460 510 545 590 635 690
The code suggests that when the limiting temperature is less than the design
temperature the member can be considered to have adequate fire resistance without
fire protective measures and otherwise fire protection should be provided to the
beams. The design temperatures, for unprotected hot-rolled beams having I or H
cross-section, corresponding to the required fire resistance are given in Table 2.11.
In the moment capacity method, the moment capacity of beam, M cf , is calculated
using the elevated temperature profile for the required period of fire resistance and
the appropriate values of strength reduction factor given in Table 2.3. It appears that
M cf is the section moment capacity although the design code does not clearly state
this. If the applied moment ( M f ) in the fire limit state is less than the moment
capacity ( M cf ), the steel beam is said to have adequate fire resistance without fire
protection. If the applied moment at the fire limit state exceeds M cf , the protection
should be provided and the thickness of fire protection required can be derived either
from Clause 4.3 or from the calculation given in Appendix D of BS5950 Part 8.
Table 2.11: Design Temperatures for Beams (BS 5950-Part 8 (BSI, 1990))
Flange Design temperature for fire resistance period of:

thickness, mm 30 min 60 min 90 min 120 min
≤ 6.8 810 940 1000 1045
8.6 790 939 1000 1045
9.7 776 938 1000 1045
10.9 767 938 1000 1045
11.8 755 936 1000 1045
12.7 750 936 1000 1045
13.2 746 936 1000 1045
14.8 741 936 1000 1045
17.0 739 935 1000 1045
17.7 736 933 1000 1045
18.8 730 931 1000 1045
19.7 722 929 1000 1045
20.2 719 929 1000 1045
22.1 716 928 1000 1045
23.6 694 920 1000 1045
25.4 688 919 1000 1045
26.8 676 914 1000 1045
27.9 665 908 1000 1045
32.0 625 885 1000 1045
36.6 586 849 1000 1045
2.9.2 Eurocode 3 Part 1-2 (ECS, 2005)
Eurocode 3 Part 1-2 (ECS, 2005) gives design rules for hot-rolled steel members
subjected to bending which can also be used for cold-formed steel members. The
design method given in Eurocode 3 Part 1-2 (ECS, 2005) is based on the moment
capacity method. This method incorporates the yield strength and stiffness reduction
factors in fire to account for the degradation of these properties at elevated
temperatures.
According to Eurocode 3 Part 1-2 (ECS, 2005), the design section moment capacity
of a steel beam with a uniform temperature θ a is as follows,
⎡ γ M ,1 ⎤
for Class 1 and 2 sections, M fi ,θ , Rd = k y ,θ ⎢ ⎥ M pl , Rd (2.43)
⎢⎣ γ M , fi ⎥⎦
⎡ γ M ,1 ⎤
for Class 3 sections, M fi ,θ , Rd = k y ,θ ⎢ ⎥ M el , Rd (2.44)
⎣⎢ γ M , fi ⎦⎥
M pl , Rd and M el , Rd are the plastic moment resistance and the elastic moment
resistance of the gross section for normal temperature design according to Eurocode
3 Part 1-1 (ECS, 2005), respectively. k y ,θ is the reduction factor for the yield
strength of steel at temperature θ a and γ M , fi is the partial factor for the fire situation
(usually γ M , fi =1).
The design lateral torsional buckling resistance moment of a laterally unrestrained

beam is given by,
for Class 1 and 2 sections, M b , fi ,t , Rd = χ LT , fiW pl , y k y ,θ ,com f y γ M , fi (2.45)
for Class 3 sections, M b , fi ,t , Rd = χ LT , fiWel , y k y ,θ ,com f y γ M , fi (2.46)
W pl , y and Wel , y are the plastic and the elastic section modulii of the gross section,
respectively.
χ LT , fi is given by
1
χ LT , fi = (2.47)
⎛Φ
⎜ LT ,θ ,com + [Φ LT ,θ ,com ] − [λ
2
LT ,θ ,com ]
2 ⎞
⎟
⎝ ⎠
with Φ LT ,θ ,com =
1
2
( [
1 + αλ LT ,θ ,com + λ LT ,θ ,com ]) 2
(2.48)
The imperfection factor α which depends on the yield strength of steel is given by,
α = 0.65 235 f (2.49)

y
This imperfection factor is based on the study of Real et al. (2000) and has been
obtained after calibrating against the numerical results obtained using the finite
element analysis software SAFIR.
The non-dimensional slenderness λ LT ,θ ,com is given by,
⎡k y ,θ ,com ⎤
λ LT ,θ ,com = λ LT ⎢ k ⎥ (2.50)
⎣ E ,θ , com ⎦
λ LT is the beam slenderness at ambient temperature. k y ,θ ,com and k E ,θ ,com are the
reduction factors for the yield strength and the modulus of elasticity of steel at the
maximum temperature in the compression flange θ a,com reached at time t ,
respectively.
The beam slenderness given in Equation 2.50 has been modified from the ambient
temperature beam slenderness to take into account the effect of different rates of
change in the strength and stiffness of steel at elevated temperatures. Real et al.’s
recent research has recommended further modifications to the design rules in
Eurocode 3 Part 1-2, which are described in Section 2.8.
The resistance of members with Class 4 cross-sections can be verified using the
same equations for Class 3 section in which area is replaced by the effective area and
the section modulus replaced by the effective section modulus. These design
equations are applicable only if the steel temperature at all cross-sections does not
exceed 350oC. According to Eurocode 3 Part 1.2 (ECS, 2005), if the design
temperature exceeds 350oC, fire protection should be provided regardless of the load
ratio values. This leads to uneconomical designs especially for low load ratios
(Renaud and Zhao, 2006). Most cold-formed steel sections are slender because of
thinness of individual plate elements and belong to class 4 section classification.
As an alternative to the above discussed moment resistance method Eurocode 3 Part

1-2 (ECS, 2005) gives the critical temperature method. This method uses the
utilization factor and gives critical temperature of a beam as a function of the
utilization factor. For uniform heating, the degree of utilization is equivalent to the
load ratio in BS5950 Part 8 (BSI, 1990).
The critical temperature θ a,cr of carbon steel at time t for a uniform temperature
distribution in a steel member is given by

⎡ 1 ⎤
θ a ,cr = 39.19 ln ⎢ − 1⎥ + 482 (2.51)
⎣ 0.9674 μ o
3.833
⎦
The degree of utilization μ o at time t = 0 for members with Class 1, Class 2, and
Class 3 cross-sections can be obtained by
E fi ,d
μo = (2.52)
R fi ,d ,0
where R fi ,d ,0 is the design resistance of the steel member for the fire design situation
for time t = 0 and E fi ,d is the design effect of actions for the fire design situation.
2.9.3 Alternative Approach Proposed by Dharma and Tan (2007)
Dharma and Tan (2006) proposed the following equations to find the ultimate
buckling moment capacity ( M b , fi ) at elevated temperatures.
M b , fi = pb , fi S x (2.53)
where,
S x - the plastic section modulus about the major axis.
k y p y p ET
pb , fi = (2.54)
T2
φ T
LT + φ LT − ky py p T
E
in which, φ LT
T
=
(
k y p y + η LT
T
+ 1 p ET ) and (2.55)
2
π 2kE E
p ET = (2.56)
λ2LT
π 2 E M cx
λ LT = (2.57)
py ME
In Equations 2.54-56, k y and k E are the yield strength and elastic modulus reduction
factors at elevated temperatures, respectively. M E is the elastic buckling moment

capacity considering major axis curvature given by Kirby and Nethercot (1979) as,
M cr
ME = (2.58)
1 − EI y EI x
The initial curvature and twist term η LT

T
in Equation 2.55 at an elevated temperature
is given by
⎛ 1.5 ⎞
π 2 Ek p ⎜ M cx ⎛ kp ⎞ kE ⎟
η T
= 0.007 − 0.4⎜ ⎟ ≥0 (2.59)
p y k y ⎜⎜ M E ⎜k ⎟ k y ⎟⎟
LT
⎝ ⎝ y ⎠ ⎠
2.10 Analysis Methods
To study the behaviour of steel members, various analysis methods are available
today. Finite element and finite strip analysis programs are widely adopted in
research works all over the world. Out of them, ABAQUS and CU-FSM analysis
were used in this research and therefore a brief description of each is given next.
2.10.1 CU-FSM Program
CU-FSM is a finite strip analysis program developed by Cornell University

researchers. It provides a convenient and efficient way to determine the elastic
buckling stresses and corresponding buckling modes (local, distortional and lateral
torsional) of thin-walled structures. However, this analysis can only be performed for
simply supported thin-walled members. CU-FSM program provides the buckling
plot of load factor versus half wave length of the buckled shape as shown in Figure
2.19. Figure 2.19 also shows the range of half-wave lengths for different buckling
modes. The minima points of the curve can be considered to represent the critical
buckling loads and buckling modes for a member. The load factor, in the case of
bending, is defined as the ratio of elastic buckling moment to the yield moment.
CUFSM program can also be used to calculate the section properties of thin-walled
cross-sections.
Lateral torsional buckling

Local buckling
Distortional buckling
Figure 2.19: Buckling Load Factor Versus Half-Wave Length
2.10.2 Finite Element Analysis (ABAQUS)
At present, finite element analysis programs are extensively used and have greater
importance in the field of research. This trend is increasing as the use of finite
element analysis programs is relatively inexpensive and time efficient compared with
large number of full scale tests. The finite element analysis program (ABAQUS) is a
very important tool that is widely used in engineering applications. It can be used to
solve problems ranging from relatively simple linear analyses to the complex non-
linear analyses. Finite element analysis using ABAQUS usually consists of three
major stages: pre-processing, analysis, and post processing. In the pre-processing
stage, the model of the physical problem is created which includes creating suitable
finite element mesh, assigning appropriate material properties and applying relevant
boundary conditions in the form of restraints or loads. In the analysis phase, the
numerical problem defined in the model is solved. Once the analysis is completed,
results are evaluated in the post-processing stage. In the following section, details of
the finite element models developed to investigate the buckling behaviour of steel
beams are presented.
2.10.2.1 Modelling Details
Bailey et al. (1996) conducted finite element analysis of restrained and unrestrained
simply supported beams subjected to loading at the shear centre of the cross-section
under uniform elevated temperature conditions. The finite element software they
used was based on a non-linear spread-of-yield program originally written by El-
Zanaty and Murray (1983) to study the two dimensional behaviour of steel frames at
ambient temperature. This software was later modified firstly by Saab and Nethercot
(1991) to include elevated temperature effects and more recently extensively
reformulated by Najjar (1994) and Najjar and Burgress (1996) to analyse three
dimensional frame behaviour including warping behaviour in fire conditions. This
program used two noded one dimensional line elements with eight degrees of
freedom in local co-ordinates. In the finite element analysis facilities have been
provided to incorporate geometric non-linearity together with the variation in cross-
sectional properties. This software was further modified by Bailey et al. (1996) to
include flexural shear stresses to enhance the accurate prediction of lateral torsional
buckling effects for loads placed at any level on a cross-section. The ability of
software to predict lateral torsional buckling has been checked at ambient
temperature with theoretical bifurcation solutions in the elastic range and with
experimental and computational results in the inelastic range. They used the results
of full scale tests carried out by Kitipornachai and Trahair (1975) to validate the
model for inelastic lateral torsional buckling.
Yu and Schafer (2002) developed a finite element model to study the local buckling
of cold-formed steel beams while Yu and Schafer (2004) developed a finite element
model simulating the distortional buckling behaviour of simply supported cold-
formed steel beams. The overall view and the details of the finite element model
developed are presented in Figures 2.20 and 2.21, respectively. They used the finite
element software ABAQUS 6.2 (HKS, 2001) in the analysis. Four node linear shell
elements with reduced integration (S4R) were used for the purlin, panel and tubes
and 8-node linear solid elements (C3D8) for the loading beam. The tube and purlin
were connected through four nodes to simulate the four bolts from purlin to the
surface of the tube. The same method was used for the connection between the panel
and the purlin. The loading beam was simply connected to the tubes by using a pin
connection. The tubes and the loading beam were modelled as rigid bodies by setting
an arbitrarily high elastic modulus. The finite element analysis was performed in a
displacement control mode, simulating the same loading method of the actual
testing.
Figure 2.20: Finite Element Modelling of Local Buckling Tests

(Yu and Schafer, 2004)
Figure 2.21: Details of Finite Element Modelling (Yu and Schafer, 2004)
Yin and Wang (2003) performed a parametric study using ABAQUS to investigate
the effects of a number of design factors on the lateral torsional buckling bending
moment resistance of a steel beam with non-uniform temperature distributions. They
used S4R type shell elements to allow for different loading positions and non-
uniform temperature distributions. The length and width of the elements were
selected so that the aspect ratio is in the range of 3 to 5. Two types of boundary
conditions were considered as shown in Figure 2.22. Both allow the beam to expand
freely along the beam’s longitudinal direction but prevent twisting. The boundary
condition without warping restraint has only positional restraints applied in the
middle of the web at the two beam ends, which allows both flanges to rotate freely in
plan.
Figure 2.22: Boundary Conditions (Yin and Wang, 2003)
Mahaarachchi and Mahendran (2005a) developed a finite element model to

investigate the lateral distortional buckling behaviour of Light Steel Beam (LSB)
sections. The model accounts for all significant behavioural effects including
material inelasticity, local buckling and lateral distortional buckling deformations,
member instability, web distortion, residual stresses, and geometric imperfections.
They modelled half of the beam because of the symmetry (see Figure 2.23). The
S4R5 elements which are thin, shear flexible, isoparametric quadrilateral shell with
four nodes and five degrees of freedom per node, utilizing reduced integration and
bilinear interpolation schemes, were used in the model.
According to Zhao et al. (1994) and Mahaarachchi and Mahendran (2005b),

idealized simply supported boundary conditions are needed for beams to satisfy the
following requirements.
Simply supported in-plane – This is both ends fixed against in-plane vertical
deflection but unrestrained against in-plane rotation, and one end fixed against
longitudinal horizontal displacement.
Simply supported out-of-plane – This is both ends fixed against out-of-plane
horizontal deflection, and twist rotation, but unrestraint against minor axis rotation
and warping displacements.
Therefore Mahaarachchi and Mahendran (2005a) modelled the support conditions

accordingly as shown in Figure 2.24. Rigid beam type Multiple Point Constraint
(MPC) elements were used to spread the concentrated moment evenly to the web and
flanges at the shear centre of the cross-section. Two explicit type MPCs connecting
the web and flange MPC elements linking only the x translational and y and z
rotational degrees of freedom were used in order to allow the flanges to rotate
independently about the minor axis without twisting (i.e. warping restraints
eliminated). Explicit type MPC elements were used to link the rigid beams to the
corresponding nodes on the edge of the section. For the nodes on the web, only x and
z translational degrees of freedom were linked whereas for the flanges, only x
translation was linked. This system of MPC was connected to the shear centre using
rigid beam MPC with simply supported boundary conditions.
Figure 2.23: Plan View of the Model (Mahaarachchi and Mahendran, 2005a)
Figure 2.24: End Support Boundary Conditions
(Mahaarachchi and Mahendran, 2005a)
Dharma and Tan (2007) developed a finite element model to investigate the lateral-
torsional buckling behaviour of hot-rolled steel I beam using commercial finite
element program, MSC.MARC Mentat. Their model was first validated against the
theoretical bifurcation elastic solution and published experimental results at ambient
temperature to verify ability of the model to simulate lateral torsional buckling
failure. The validated numerical models are then used to simulate various section
sizes and beam lengths with mid-span point load and uniform bending at elevated
temperatures. Rectangular thick shell elements which have three global
displacements and three global rotations as degrees of freedom were used to model
the beam. The beam is simply supported at both ends so that lateral deflection and
twist rotation at the supports are prevented, while the flange ends are free to warp.
This support system was modelled by applying boundary conditions as given in
Table 2.12 and Figure 2.25 shows the different types of end nodes and the axis
system.
Left Cross Section Right Cross Section
Figure 2.25: Axis System and Different Types of End Nodes
Table 2.12: Applied Boundary Conditions
The magnitude of initial imperfection was used as L/1000 and residual stresses
haven’t included in the analysis. The stress strain relationship given in Eurocode 3
Part 1.2 (ECS, 1995) has been used as the material properties. The load was applied
in two stages. Temperature was first increased to the desired value and then the load
was increased until the member fails. Temperatures considered in their analysis were
400oC, 500oC, 600oC, 700oC and 800oC. Failure point have been assumed to be
occurred when the maximum in-plane deflection exceeds L/20, or when the
deflection exceeds L/30, the beam reaches a maximum rate of deflection L2/(9000
depth) mm/min (L and depth are in mm)
2.10.2.2 Initial Conditions
Buckling failures of cold-formed steel beams are very sensitive to the initial
geometric imperfections and residual stresses of the member. Therefore the initial
state of a cold-formed steel member such as geometric imperfections and residual
stresses must be considered in the advanced finite element analysis to increase the
accuracy of the results.
2.10.2.3 Residual Stresses
The residual stresses of cold-formed steel members affect the member stiffness and
also cause premature initial yielding which causes the reduction of ultimate strength.
Residual stresses can be divided into two types: flexural and membrane residual
stresses. In cold-formed steel members, membrane residual stresses are generally
small compared to flexural residual stresses (Schafer and Pekoz, 1998, and Young
and Rasmussen, 1999). Therefore membrane residual stresses can be neglected. The
flexural residual stresses are compressive on the inside surface and tensile on the
outside surface of the section. On the other hand the membrane residual stresses are
tensile or compressive and are the same across the thickness.
Pi et al. (1998) adopted longitudinal normal residual stresses recommended by Weng

and Pekoz (1988) for cold-formed steel channel sections. The recommended
magnitudes and distributions of the residual stresses are shown in Figure 2.26. These
residual stress distributions satisfy the bending moment equilibrium condition but
not the axial force equilibrium condition because the outside corner area with tensile
residual stresses is larger than the inside corner area with compressive residual
stresses. Therefore the maximum residual stresses on the inside and outside surfaces
of the corners, equal to − 0.65σ y and + 0.65σ y , were modified to − 0.65σ y (2t i t )
and + 0.65σ y (2t o t ) as shown in Figure 2.26 (c), where t i = (r 2

)
+ rt + t 2 2 and
to = t − ti .
(a) (b) (c)
(a) Around Channel Section

(b) Through Section Wall at Flats and Lips
(c) Through Section Wall at Corners
Figure 2.26: Residual Stress Distributions (Pi et al., 1998)
Schafer and Pekoz (1998) gives distributions of flexural residual stresses as a

percentage of yield stress for roll-formed and press-braked cold-formed steel lipped
channel section based on their measurements (see Figure 2.27). However, they have
not provided the residual stresses for the lips. It can be observed that there is a
considerable difference of residual stresses in the web and flange, although they
were affected by the same cold-working.
Figure 2.27: Residual Stresses in Cold-Formed Steel C-Section

(Schafer and Pekoz, 1998)
Considering the above mentioned shortcomings, Ranawaka (2006) proposed a new

set of residual stresses based on Schafer and Pekoz’s (1998) model for press-braked
cold-formed steel channel section as shown in Figure 2.28.
0.17fy
outside +σr
0.08fy
0.17fy Residual stress

through thickness
inside -σr
Figure 2.28: Assumed Flexural Residual Stress Distribution (Ranawaka and

Mahendran, 2009b)
2.10.2.4 Geometric Imperfections
In reality, every beam has some deviations from a perfect geometry caused by the
fabrication process, storage, transport, handling and installations. The geometric
imperfections of a member include bowing, warping, twisting and local deviations.
Local deviations include the dents and regular undulations in the plate elements of
the section. Geometric imperfections reduce the strength of the member
significantly. Further, the post buckling of cold-formed steel members is difficult to
predict if geometric imperfections and material nonlinearities are present. However,
it is very difficult to quantify the geometric imperfections of cold-formed steel
members. The guidelines given in design codes or product standards propose to use
conservative upper limits for the geometric imperfections used in the designs. The
distribution of initial crookedness and twist may vary for different cold-formed steel
beams (Kwon, 1992, Schafer and Pekoz, 1998).
Earlier researchers such as Thomasson (1978), Lau (1988), Kwon (1992) and
Bernard (1993) measured the geometric imperfections of cold-formed steel
members, and Schafer and Pekoz (1998b) sorted them under two categories: Type 1,
maximum local imperfection in a stiffened element and Type 2, maximum deviation
from straightness for a lip stiffened or unstiffened flange as shown in Figure 2.29.
For plate thicknesses less than 3 mm, Schafer and Pekoz (1998) presented a rule of
thumb to calculate the magnitudes of imperfection which can be applied for
width/thickness ratio ( w t ) less than 200 for Type 1 imperfection and w t < 100 for
Type 2 imperfection. For Type 1 imperfection, a simple linear regression based on
the plate width yields an approximate expression, d1 ≈ 0.006 w and an alternative
rule based on an exponential curve fit to the thickness is d1 ≈ 6te −2t ( d and t are in
mm). For Type 2 imperfections, the maximum deviation from straight is
approximately equal to the plate thickness ( d 2 ≈ t ).
Type 1 Type 2
Figure 2.29: Definition of Geometric Imperfection (Schafer and Pekoz, 1998)
A simple method of modelling the initial crookedness and twist is to assume the
same distribution as that of the lowest buckling mode. Pi et al. (1998) used the
lowest buckling mode as the distribution of the initial crookedness and twist. The
initial crookedness and twist may exist in the negative or positive directions. Pi et al.
(1998, 1997) found out in their numerical studies beams with negative imperfections
have lower beam capacities compared to those with positive imperfections.
Kurniawan and Mahendran (2009a,b) also made similar observations regarding the
direction of initial geometric imperfection and the failure direction in their numerical
studies. The experimental studies of Put et al. (1999b), Poutre et al.’s (2000) and
Ungermann and Kalameya (2006) also confirmed this finding. However, the initial
crookedness and twist are more likely to be in the negative than in the positive
direction according to the measurements of Kwon (1992). The magnitudes of the
initial crookedness and twist vary randomly and should be treated probabilistically
(Schafer and Pekoz, 1998).
Bailey et al. (1996) assumed in their study that all the imperfections represented by
an equivalent initial out-of-plane bow, which gave an imperfection parameter value
at mid span equal toη = 0.003λ , where λ is the slenderness ratio. However,
according to Bailey et al. (1996), recorded displacements during the tests indicated
that the above assumption has underestimated the actual initial imperfections.
Real et al. (2001) and Real et al. (1999) used characteristic values for initial out of
straightness (L/1000) in their numerical simulations to study the lateral torsional
buckling behaviour of hot-rolled steel beams. However, Real et al., (2003) measured
two types of geometric imperfections, one relating to the cross-section dimension
measured by digital callipers and the other relating to the longitudinal lateral distance
from an imaginary straight line measured by a laser beam method. The measured
cross-section geometry and the measured longitudinal imperfections of the beam,
approximated by the following expression, were used in the numerical modelling.
⎛π × z ⎞
u ( z ) = a sin ⎜ ⎟ (2.60)
⎝ L ⎠
where a is the measured maximum amplitude of the lateral imperfection
Real et al. (2004) used L 1000 as the maximum amplitude of the lateral
imperfection in Equation 2.60. Dharma and Tan (2006) also assumed a maximum
initial out-of-plane bow of L/1000, in the form of a half sine curve. An initial
rotation about the longitudinal axis with a maximum value of 1/1000 radians at mid
span was also considered by Real et al. (2004).
2.10.2.5 Experimental Procedures
Real et al. (2003) carried out a series of tests on IPE100 beams at elevated
temperatures. The tested beams were simply supported with fork supports as shown
in Figure 2.30. In Figure 2.30, qb is the self weight of the beam and q is the
additional distributed load due to the weight of the ceramic mat and electro-ceramic
resistances used. The beams were electrically heated by a power unit of 70 kVA by
means of ceramic mat elements. A ceramic mat was used around the beam and the
heating elements to ensure the maximum thermal efficiency. The experimental set up
is shown in Figure 2.31. In order to ensure a uniform temperature distribution along
the beam, automatic control of separated heating elements were used. The
temperature was measured at different places using thermocouples welded to the
beam. Tests were conducted at different temperatures of 200, 300, 400, 500, and
600oC. The loads were applied after the required temperature was stabilized by
means of two hydraulic jacks with 600 kN of capacity each. Three types of mid span
displacement were measured during the loading as shown in Figure 2.32. They are
vertical displacement (DV), the lateral bottom displacement (DLB), and the lateral
top displacement (DLT).
Figure 2.30: Simply Supported Beam with Forks Supports (Real et al., 2003)
Test Arrangement (b) Fork Support and Hydraulic Jack
Figure 2.31: Experimental Set up (Real et al., 2003)
Figure 2.32: Measured Mid Span Displacement (Real et al., 2003)
Put et al. (1999b) conducted a series of 10 lateral buckling tests on simply supported
unbraced cold-formed steel beams of two different cross-sections. Figure 2.33 shows
the test set up and the details of the support. The support system was developed to
prevent in and out-of-plane deflections and twist rotation but not to restrain in and
out-of-plane rotations and warping displacements. One of the supports was movable
and the other is fixed. The running tracks and side guides prevented in and out-of
plane displacement and twist rotations. The tapered (bottom) and ball (top) bearing
races allowed out-of plane rotations while the side ball bearing races allowed in-
plane rotations. The two supports were aligned to ensure that the vertical deflections
remained in the same plane, which reduced the end-to-end initial twist of the web.
The attachment of the top and bottom bearing races was intended to allow
differential flange rotations associated with warping displacements. The loading
system applied vertical load in the loading drum to a point at or below the shear
centre. Put et al. (1999a) used the same test arrangement to test unbraced simply
supported cold-formed steel channel beams loaded eccentrically at mid span.
(a) Support
(b) Test Set-up

Figure 2.33: Test Arrangement Used by Put et al. (1999a)
Yu et al. (2002) conducted four point bending tests to study the local buckling
behaviour of cold-formed C and Z section beams. The test setup is illustrated in
Figure 2.34 (a). A pair of 18ft. long C or Z specimens having 16 ft span loaded at
one third points. The arrangement of rollers at the support is shown in Figure 2.34
(b), which provides the simply supported end condition. Machined, quarter round
aluminium blocks were used as guides for rollers at the loading points as shown in
Figure 2.34 (c). Thin Teflon sheets were also added at the load points and support
points to limit unwanted friction and to ensure the required boundary conditions.
Tests were performed in displacement control at a rate of 0.0015in./sec. The
deflections at one third points were measured using two LVDTs while the force was
monitored using MTS 407 controller and a load cell. The longitudinal strain on the
lip and the top of the web at mid span were also monitored by using strain gauges for
a limited number of tests. Yu et al. (2004) used the same test set up to study the
distortional buckling behaviour of cold-formed steel C and Z section beams.
(a) Elevation View of Overall Test Arrangement for Four Point Bending Test
(b) Support Configuration (c) Loading Configuration
Figure 2.34: Test Arrangement (Yu et al, 2002)
Mahaarachchi and Mahendran (2005a) conducted a series of lateral distortional

buckling tests on a new hollow flange channel section called LSB. The support
system was specially developed to satisfy the simply supported boundary conditions
described in Section 2.10.2.1. Supports at both ends were the same except that at one
end the support was free to move in the longitudinal direction of the beam as shown
in Figure 2.35 while movement of other end was fixed by preventing the side bearing
from rolling along the running track by using horizontal stops. The loading system
was designed so that the load was applied through the shear center as shown in
Figure 2.36.
Figure 2.35: Support System Used by Mahaarachchi and Mahendran (2005a)
Figure 2.36: Loading System Used by Mahaarachchi and Mahendran (2005a)
2.11 Literature Review Findings
In this chapter, an extensive literature review was conducted gathering information

in the following areas: effects of cold-forming, different failure modes of cold-
formed steel beams, current cold-formed steel design standards at ambient
temperature, previous research on cold-formed steel beams at ambient temperature,
fire safety of steel structures, change in mechanical properties at elevated
temperatures, previous research on cold-formed and hot-rolled steel beams at
elevated temperatures, current design standards available at elevated temperatures,
different analysis methods, initial conditions of beams including geometric
imperfections and residual stresses, and the experimental investigations on steel
beams. A summary of the literature review findings is presented.
The use of cold-formed steel in various constructions is increasing because its light
weight makes it easier and more economical compared to hot-rolled steel.
Several researchers have found that the cold-forming process particularly by cold-
rolled forming can alter the mechanical properties of cold-formed steel sections due
to strain hardening and strain aging. The design standards allow the use of the
increased yield strength resulting from cold-working.
Cold-formed steel flexural members may subject to complex stability problems

because of thinness of the plate elements. Generally, cold-formed steel beams may
undergo four types of buckling modes: local buckling, distortional buckling, lateral
distortional buckling and lateral torsional buckling. The lateral torsional buckling is
the most complex buckling mode at ambient temperature, and at elevated
temperatures this buckling mode becomes more complex because of the reduction of
mechanical properties.
Eurocode 3 Part 1-3 (ECS, 2006), BS5950-Part 5 (BSI, 1998) and AS/NZS 4600
(SA, 2005) have addressed the behaviour of cold-formed steel beams at ambient
temperature. However, none of these standards have addressed the behaviour of
cold-formed steel members at elevated temperatures.
Put et al. (1999b) suggested that the formulations of standard for hot-rolled steel, AS
4100, are more suitable for the lateral buckling capacity of cold-formed steel flexural
members. However, this should be further investigated as it was based on the results
of few experimental tests.
The structural behaviour of cold-formed steel beams depends on its mechanical

properties such as its yield strength and modulus of elasticity. These mechanical
properties of cold-formed steels deteriorate as temperature increases. The studies on
the mechanical properties of cold-formed steel at elevated temperatures are very
limited and they only cover some steel grades and thicknesses. BS5950 Part 8 (BSI,
1990) and Eurocode 3 Part 1-3 (ECS, 2006) give the mechanical properties of cold-
formed steels at elevated temperatures.
In the current design codes, the fire behaviour of cold-formed steel members is not
covered well. According to Eurocode 3 Part 1-2 (ECS, 2005), designers have been
allowed to use the same design guidelines for cold-formed steel members as that
given for hot-rolled steel members under fire conditions. But the accuracy of these
guidelines should be checked for cold-formed steel members because the variations
of mechanical properties with increasing temperature are different from that of hot-
rolled steel. BS5950 Part 8 (BSI, 1990) provides design guidelines for hot-rolled
members only, and not for cold-formed steel members.
The design equations given in Eurocode 3 Part 1-2 (ECS, 2005) are applicable for
members with Class 4 cross-sections only if the steel temperature at all the cross-
sections does not exceed 3500C. If the design temperature exceeds 3500C, fire
protection is to be provided regardless of the load ratio values leading to
uneconomical designs, especially for low load ratios.
It is necessary to take into account the influence of local buckling when calculating
the load carrying capacity of cold-formed steel beams at elevated temperatures.
Ranby (1998) showed that the calculation of the plate buckling resistance according
to Eurocode 3 Part 1-3 (ECS, 2006) with reduced yield strength and elastic modulus
is accurate enough at elevated temperatures, provided that the yield strength is taken
as the 0.2% proof stress.
Studies of Bailey et al. (1996) and Yin and Wang (2003) on lateral torsional
buckling behaviour of hot-rolled steel beams revealed that BS5950 Part 8 (ECS,
1990) and Eurocode 3 Part 1-2 (ECS, 2002) overestimate the limiting temperatures
for unrestrained steel beams.
Real et al. (2003) proposed an alternative design method for Eurocode 3 Part 1-2
(ECS, 1995) and that method is incorporated in the current version of Eurocode 3
Part 1-2 (ECS, 2002). Real et al. (2004) extended the research into study the effect of
different moment distribution and found out that the proposed method gives
overconservative results for loadings other than uniform bending. Real et al. (2006)
again found that the present method in Eurocode 3 Part 1-2 (ECS, 2002) gives unsafe
values for some types of profiles. They made a proposal to rectify that. The
applicability of these modified proposals for cold-formed steel flexural members at
elevated temperatures has to be verified.
Although extensive research has been carried out to study the behaviour of cold-
formed steel columns at ambient and elevated temperatures, only limited number of
studies has been carried out on the flexural behaviour of cold-formed steel beams
especially on lateral torsional buckling even at ambient temperature. At elevated
temperatures, no research was found on the lateral torsional buckling of cold-formed
steel beams. The above studies show that some researchers have focused on the
behaviour of hot-rolled steel beams at elevated temperatures. But no research has
been carried out to study the lateral torsional buckling and the section moment
capacities of cold-formed steel beams subjected to uniform elevated temperatures
(uniform across the section).
Some researchers have focused their study on the behaviour of steel joists in floor
assemblies. In floor assemblies cold-formed steel beams are protected with gypsum
boards and other insulation materials placed inside the cavities and the fire resistance
of these floor assemblies were found to be governed by the ability of the gypsum
boards to remain in place. In these studies, the effect of the loss of lateral restraints
provided by plasterboards and the lateral steel members at high temperatures has not
been taken into account in the numerical simulations of cold-formed steel beams. In
these studies, the floor assemblies were subjected to fire from one side and hence
temperature distribution was not uniform across the steel section.
Earlier researchers have used finite element analysis methods to investigate the
flexural behaviour of steel beams. They have shown that the behaviour of cold-
formed steel flexural members can be accurately studied using finite element
analyses provided that the end support and loading conditions, residual stresses,
geometric imperfections and mechanical properties are included in the finite element
model accurately. Suitable element size and type must be used and the model has to
be validated with sufficient experimental results.
The residual stresses of cold-formed steel affect the member stiffness and also cause
premature initial yielding which causes the reduction of ultimate strength. In cold-
formed steel flexural residual stresses are more important to consider than the
membrane residual stresses because membrane residual stresses are generally small
compared to flexural residual stresses. Ranawaka (2006) proposed residual stresses
for press braked cold-formed steel lipped C-section based on Schafer and Pekoz’s
(1998) model.
Geometric imperfections in cold-formed steel flexural members cause considerable

reduction of the ultimate strength. Therefore it is needed to consider the geometric
imperfections in design capacity calculations and also in finite element analysis.
However, it is very difficult to quantify the geometric imperfections of cold-formed
steel members. Kwon (1992) and Shafer and Pekoz (1998) stated that the distribution
of initial crookedness and twist may vary for different cold-formed steel beams.
Therefore it is better to use measured imperfections in design capacity calculations.
It is obvious that no research has been conducted, to date, to study the behaviour of
cold-formed steel beams in relation to lateral torsional buckling and section moment
capacity at uniform elevated temperatures. Hence the lateral torsional buckling and
the section moment capacity of cold-formed steel beams should be studied in order
to develop new design guidelines and to check whether the available design
guidelines are accurate for cold-formed steel structures.
3 Mechanical Properties of Cold-formed
Steels at Elevated Temperatures
3.1 Introduction
Mechanical properties play an important role in the fire safety design of cold-formed
steel structures since temperature rise in steel members reduces the mechanical
properties such as yield strength and elastic modulus. This reduction of mechanical
properties greatly affects the structural behaviour of cold-formed steel members and
reduces the load carrying capacity at high temperatures. Therefore the mechanical
properties at elevated temperatures are required for the fire safety design of steel
members. Hence it is important to have a good knowledge and understanding of the
reduction of mechanical properties with increasing temperatures. BS5950-Part 8
(BSI, 1990) and Eurocode 3 Part 1.2 (ECS, 2005) provide reduction factors for
mechanical properties of cold-formed steels at elevated temperatures. However,
BS5950-Part 8 only provides reduction factors for yield strengths corresponding to
0.5%, 1.5% and 2.0% strain levels although the common practice is to use 0.2%
proof stress as the yield stress. Further, it does not provide reduction factors for
elastic modulus at elevated temperatures. On the other hand, Eurocode 3 Part 1.2
(ECS, 2005) provides reduction factors for both yield strength and elastic modulus
for cold-formed steels, which are similar to hot-rolled steels despite the fact that the
reduction of mechanical properties of both types of steel is considered different.
Sidey and Teague (1988) stated that the strength reduction of cold-formed steel at
elevated temperatures may be 20% higher than that of hot-rolled steel. Past research
on mechanical properties at elevated temperatures was mainly focussed on hot-rolled
steel. However, in recent times some studies have been carried out on the reduction
of mechanical properties of cold-formed steels at elevated temperatures (Makelainen
and Miller, 1983, Outinen et al., 2000, Outinen and Makelainen, 2001, Lee et al.,
2003, Ranawaka and Mahendran, 2009a, Chen and Young, 2007). More studies are
required to determine the mechanical properties of cold-formed steels covering
different thicknesses and steel grades in order to gain a thorough knowledge and
thereby use them in fire safety design.
The mechanical properties of cold-formed steels with thicknesses 1.55 and 1.95 mm
belonging to G250 and 1.5 and 1.9 mm belonging to G450 at various elevated
temperatures are needed in the finite element analysis and finite strip analysis of
cold-formed steel beams in this research. Therefore tensile coupon tests of 1.55 mm
and 1.95 mm thick G250 and 1.5 mm and 1.9 mm thick G450 cold-formed steels at
various elevated temperatures were undertaken. Further, tensile coupon tests were
carried out on 0.95 mm thick G550 cold-formed steel at elevated temperatures in
order to check the accuracy of the test results based on which equations have been
proposed for the yield strength and elastic modulus reduction factors by Ranawaka
and Mahendran (2009a). This chapter presents the details of this experimental study,
results, comparison of the results with the currently available information on the
reduction factors of mechanical properties obtained from previous research as well
as from design codes. Finally, equations developed for yield strength and elastic
modulus based on this research and all the companion QUT research on mechanical
properties of cold-formed steel are presented.
3.2 Experimental Investigation
3.2.1 Test Method
The mechanical properties of steels are commonly determined using tensile coupon
tests. Since the mechanical properties of steels obtained from tensile coupon tests are
used in the numerical modelling of steel beams and columns, some researchers have
also used compression coupon tests to determine their mechanical properties.
However, Ranawaka (2006) showed that the differences between tension and
compression coupon test results for yield strength and elastic modulus of cold-
formed steels are very small. Compression coupon tests are also very difficult to
perform, especially at elevated temperatures. Therefore tensile coupon tests were
used in this research for both ambient and elevated temperatures. From tensile
coupon tests, the stress-strain curve is obtained and thereby the required material
properties such as yield strength, elastic modulus and the ultimate strength can be
determined. There are three types of test methods to investigate the mechanical
properties of steel under fire conditions: the steady state test, the transient state test
and ISO test (ISO, 1999). In the steady state test, the specimen is heated up to the
required temperature and then the tensile load is applied at a constant rate until
failure while maintaining the constant temperature. The steady state tests can be
carried out either as strain-controlled or load-controlled. In the strain-controlled tests,
the strain rate is kept constant whereas in the load-controlled tests, the loading rate is
kept constant. On the other hand, in the transient state and ISO tests, the test
specimen is kept under constant load while the temperature is increased at a
specified rate. It has been stated by several researchers that the transient state test
method is more realistic since it simulates the behaviour of a real fire including the
creep effect (Outinen and Makelinen, 2002a, Chen and Young, 2007, Lee et al.,
2003). The creep effect which is time dependent is influenced by both the applied
load and temperature. This effect can be considered negligible since both steady state
and transient state tests are usually completed within an hour and thus undergo
limited amount of creep effect.
The steady state tests are easier to carry out than the transient state tests. The steady
state tests directly give the stress-strain curve whereas the transient state tests give
the temperature-strain curve. The temperature-strain curve obtained from transient
state tests is then converted to the stress-strain curve. This process is difficult and
can easily lead to errors in the results. Therefore many researchers have used the
steady state test method because of its simplicity and accuracy. Hence in this
research the steady state test method was adopted to determine the mechanical
properties of 1.55 and 1.95 mm thick Grade 250 and 1.5 and 1.9 mm thick Grade
450 cold-formed steels.
3.2.2 Test Specimen
The shapes and size of tensile test specimens were in accordance with AS 1391 (SA,
1991) as illustrated in Figure 3.1. One hole was provided at both ends of each
specimen in order to fix the specimen to the loading shafts located at the top and
bottom ends of the furnace using M10 bolts. The hole size was determined so that
there would not be any premature failure of the specimen at the holes. The test
specimens were cut in the longitudinal direction of the cold-formed steel sheet since
the mechanical properties of steel may vary depending on the direction, ie.,
longitudinal or transverse. All the specimens of 1.95 mm thick G250 and 1.9 mm
thick G450 steel were cut from the same batch of steel that was used in beam testing.
35 mm 10.5 mm Diameter
35 mm
25 mm Radius
13 mm 50 mm
60 mm
70 mm 25mm 25mm 70 mm
(a) Dimensions of tensile coupons
(b) Tensile test coupons
Figure 3.1: Tensile Test Coupons and Dimensions
Tensile coupon tests were carried out to determine the mechanical properties (tensile
strength, yield strength, elastic modulus etc.) of cold-formed steels with 1.55 and
1.95 mm thick G250 and 1.5 and 1.9 mm thick G450 at different temperatures.
Specimens were tested at different uniform temperatures from ambient temperature
to 700oC. At least two tests were carried out for each temperature which gave a total
of 88 tensile coupon tests.
The coating of each specimen was removed by immersing it in diluted hydrochloric

acid (acid to water = 1:3) and then the specimen thickness was measured at three
points within the gauge length using a micrometer. The average of these three
readings was taken as the base metal thickness. The specimen width in the middle
was also measured at three locations using a vernier calliper and the average of the
three readings was taken as the width. The measured dimensions were used in the
calculation of mechanical properties. Table 3.1 shows the dimensions of the tensile
coupons used in the tests.
Table 3.1: Dimensions of Tensile Test Coupons
Steel Minimum yield Nominal Test temperature Measured base metal Measured
Grade strength (MPa) thickness (mm) (oC) thickness (mm)* width (mm)*
20 1.483 12.895
100 1.498 12.877
200 1.498 12.872
300 1.498 12.947
1.55
400 1.500 12.960
500 1.490 12.880
600 1.486 12.938
G250 250 700 1.498 12.893
20 1.943 12.846
100 1.947 12.833
200 1.947 12.859
300 1.939 12.863
1.95
400 1.945 12.848
500 1.942 12.878
600 1.947 12.848
700 1.932 12.886
20 1.480 12.870
100 1.497 12.899
200 1.483 12.877
300 1.488 12.948
1.5 400 1.501 12.863
500 1.496 12.875
550 1.498 12.873
600 1.484 12.973
G450 450 700 1.493 12.932
20 1.891 12.746
100 1.871 12.749
200 1.887 12.740
300 1.889 12.760
400 1.888 12.752
1.9
450 1.888 12.743
500 1.889 12.754
550 1.893 12.897
600 1.880 12.754
700 1.882 12.773
0.95 20 0.945 12.863
100 0.918 12.900
450 0.935 12.967
G550 550
500 0.945 12.957
550 0.935 12.880
600 0.928 12.930
*Mean value based on three measurements.
3.2.3 Test Rig and Procedure
All the tensile coupon tests at both ambient and elevated temperatures were
conducted in the Structures Laboratory in the Faculty of Built Environment and
Engineering, Queensland University of Technology. Tensile coupon tests at elevated
temperatures were carried out by using an electrical furnace shown in Figure 3.2.
Furnace is heated by four glow bars fitted inside the furnace. The two internal
thermocouples fitted inside the furnace give the air temperature of the furnace. Since
the two internal thermocouples are located a distance away from the specimen, an
additional thermometer was placed in contact with the specimen as a means of
independent verification of the specimen temperature.
Tensile Load
Hydraulic actuator
Load cell
Top rod
Glow bars
Thermocouples
Specimen
Glass
Bottom rod
Thermometer
Figure 3.2: Test Rig
Eight temperatures were selected in this study: 20oC, 100oC, 200oC, 300oC, 400oC,
500oC, 600oC and 700oC. For Grade 450 steels, 450oC and 550oC were also selected.
A heating rate of 10-20oC/min was used. It was observed that specimen temperature
measured by the thermometer differed from the air temperature measured by the
thermocouples by about 10 to 20oC. This difference was dependant on the set
temperature and was due to the radiative heating of the glow bars. Therefore the
temperature shown by the thermometer was considered as the temperature reading.
The specimen temperature was assumed to be uniform because of short specimens
and the rapid transfer of heat in cold-formed steels.
The steady state test method was used in all the tensile coupon tests. Initially, the
temperature inside the furnace was increased to a pre-selected value with the
specimen inside the furnace. After reaching the pre-selected temperature, it was
allowed to remain for 10 min before starting the loading in order to attain the
uniform temperature throughout the specimen. The length increase due to the
thermal expansion of the specimen may result in developing compressive load in the
specimen. Therefore during the heating phase, the specimen was maintained under a
small tensile load while allowing free upward movement. Once stabilizing the
specimen at the required temperature, a tensile load was applied at a constant strain
rate until the failure of the specimen. The strain rate should be within the range of
0.001 to 0.005/min according to high temperature testing standard for metallic
materials, SFS-EN10002-5 (ECS, 1992). In this study, the displacement rate used
was 0.15 mm/min, which was equivalent to a strain rate of 0.0025/min (0.000042/s)
and thus satisfies the requirement of SFS-EN100002-5 (ECS, 1992). Adopted strain
rate also satisfies the requirement of Australian standard for tensile testing of metals
at elevated temperatures, AS 2291 (SA, 1979) which specify the allowable range of
strain rate as 0.00002/s to 0.00005/s.
Load cell
(c) Top end
(a) Furnace
Base of the loading Camera
arrangement
(d) Side view of laser speckle

extensometer with cameras
(b) Bottom end
(e) Specimen inside the furnace
Figure 3.3: Details of Test Arrangement
Details of the test set-up are shown in Figure 3.3. The specimen was connected to
two vertical end rods which were accurately aligned with each other. The bottom end
was fixed (Figure 3.3b) while the top end was free to move upwards (Figure 3.3c).
The tensile load was applied by using a hydraulic actuator connected to the top end
rod. Multi-purpose TestWare System was used to control the hydraulic actuator and
also act as the data logger system. The applied load was measured by using a load
cell of 1 tonne for high temperatures and a load cell of 5 tonnes for low
temperatures. Since the vertical alignment of the specimen is the key factor in the
accuracy of results, special care was taken to align two end rods accurately by
following the procedure explained next. First, hydraulic actuator was fixed to the
crosshead beam and then a plumb bob was hanged from the centre of the loading
shaft of the hydraulic actuator to find the corresponding vertical point at the base.
Thereafter load cell and top loading shaft were fixed to the hydraulic actuator while
bottom shaft was fixed to the base so that they are aligned with each other.
Verticality of bottom and top loading shaft was checked by sprit level. Alignment of
two rods was further checked by observing the movement of the magnified laser
images of a specimen connected to two rods while loading.
Since it was not possible to use strain gauges at elevated temperatures, a laser
speckles extensometer (Figure 3.4) developed by Austrian Company Messphysik
GmbH was used to measure the strains of the tensile test specimens. It has the
capability of measuring very small strains accurately and comprised of PC based
Video Processor, laser diodes (class 3A), video cameras, lens and frame grabber.
Monochrome video cameras, which are capable of easily adjusting to obtain the
correct image, are utilized. The frame grabber digitizes the analogue video signal and
these digitized images are displayed on the computer screen. The video processor is
capable of continuously measuring the displacements of two speckle patterns
recorded by video cameras in a master slave configuration. The two displacements
are converted to a strain signal and sent to an external control system. Interference
filters were fitted to cameras to eliminate infra-red radiation emitted by the hot
specimen above 500oC temperatures.
Pivoted cameras and Lenses of the
lens mounting device cameras
Two lasers
Base plate
Data transferring
cable
Figure 3.4: Laser Speckle Extensometer
The laser speckle extensometer was located behind the furnace so that the cameras
can be directed to the specimen gauge length through the fire resistant window of the
furnace (Figure 3.3 (d)). Two laser beams are directed to two points on the specimen
as shown in Figure 3.5 and two cameras were oriented targeting the specimen as
shown in Figure 3.3 (e). The upper camera is called the slave camera while the lower
camera is called the master camera.
Specimen
Slave Camera
50 mm
Master Camera
Laser beams
Figure 3.5: Strain Measurements
The measuring principle of extensometer is based on tracking laser speckle patterns
through a digital correlation technique. When a coherent laser beam is directed on to
an optically rough surface, the lights are diffused in different directions. If the
diffused light rays travel through the original beam, the light is spatially eliminated,
resulting in a granular looking speckle pattern, which is shown in Figure 3.6. Each
camera produces unique speckle patterns relevant to each zone to which it is
directed. Speckle patterns corresponding to two zones on the specimen, which are
separated from each other by a predetermined distance in the elongation direction are
initially stored as a reference speckle patterns. When the tensile load is applied to the
specimen, targeted zones of two cameras changes and hence the speckle pattern
produced by the laser light changes. The video processor is able to locate the new
position of a stored reference pattern and calculate the distance it has moved between
images. Figure 3.6 (a) shows the speckle patterns of master and slave cameras before
starting the test while Figure 3.6 (b) shows them during the test. Before starting the
tests, the laser-speckle extensometer was calibrated using a special calibration
method, which enables accurate strain measurements. The strain is calculated using
Equation 3.1. Since the distance between initial reference patterns given by the two
cameras is adjusted to 50 mm and is stored in the program before starting the test,
the processor is able to calculate the strain at any given time.
Σd slave − Σd master
ε= …………………………………………………… (3.1)
lo
where, Σd slave - Sum of displacements on Slave camera
Σd master - Sum of displacements on Master camera
lo - Distance between initial reference patterns
Selected speckle
patterns
Output of slave
camera Output
of slave
camera
(a) Speckle output before the test
Distance travelled from Distance travelled

master camera from slave master
(b) Speckle output during the test
Figure 3.6: Speckle Output for Strain Measurements
There were two computers used for data acquisition purposes. One computer was
used to obtain the strain data and transfer them to the second computer. The second
computer was connected to the load cell and therefore recorded the tensile load with
the relevant strain value transferred by the other computer. Hence it is directly
possible to plot the stress-strain curve. The overall test set-up for tensile coupon
testing at elevated temperatures is shown in Figure 3.7.
Loading
Thermometer (measures the Shafts
specimen temperature)
Temperature
Controller
Furnace
Load Laser Speckle
Measurement Extensometer
Strain
Measurement
Figure 3.7: Overall Test Set-up
Firstly, the tensile coupon tests of two steel grades with selected thicknesses were
carried out at ambient temperature. In this case the strain was measured by using
both 5 mm strain gauges and the laser speckle extensometer so that the accuracy of
the strain readings of laser speckles extensometer could be checked by comparing
with strain gauge measurements. Figure 3.8 shows the stress-strain curves obtained
from both strain gauges and laser speckle extensometer for 1.9 mm thick Grade 450
steel at ambient temperature. It shows that the stress-strain curves obtained from the
strain-gauge and the laser speckle extensometer are in good agreement. This
confirmed that the laser speckle extensometer could accurately measure the strains of
tensile specimens.
Table 3.2 shows the comparison of results obtained from both strain-gauges and the
laser speckle extensometer at ambient temperature. It is noted that the results of yield
strength and elastic modulus are in good agreement. Hence all the elevated
temperature tensile coupon tests were carried out by using laser speckle
extensometer.
600
500
400
Stress (MPa)
Extensometer
300
Strain Gauge
200
100
0
0 1 2 3 4 5 6 7 8 9 10 11 12
Strain (%)
600
500
400
Stress (MPa)
Extensometer
300
Strain Gauge
200
100
0
0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
Strain (%)
Figure 3.8: Stress-strain Curve for 1.9 mm thick G450 Steel at 20oC
Table 3.2: Yield Strength and Elastic Modulus at Ambient Temperature
Yield strength (MPa) Elastic modulus (MPa)

Thickness
Grade Strain Laser speckle Strain Laser speckle
(mm)
gauge extensometer gauge extensometer
1.55 250 293.50 292.00 204385 202700
1.50 450 537.11 536.25 207490 209240
1.95 250 270.51 269.73 188220 189090
1.90 450 514.5 515.25 206328 201395
3.2.4 Mechanical Properties at Ambient and Elevated Temperatures
3.2.4.1 Determination of yield strength and elastic modulus
Generally, there are two types of stress-strain curves and they are: a sharp yielding
type and a gradual yielding type (see Figure 3.9). Steel exhibits a linear stress-strain
relationship at the beginning and it is called the elastic region. The slope of the
elastic region of stress-strain curve is taken as the elastic modulus of both types of
curves. The way of extracting yield strength from both types of curves is different.
The sharp yielding type stress-strain curve gives a well defined yield point and the
strength corresponding to the flat region of stress-strain curve is taken as the yield
strength (Figure 3.9a). In the case of gradual yielding type stress-strain curve, the
yield strength is taken as the 0.2% proof stress or the strength corresponding to 0.5%
of total strain at ambient temperature. These two methods give similar yield strength
values as illustrated in Figure 3.9 (b).
At elevated temperatures, the yield strength is taken as either 0.2% proof stress or
stress corresponding to different strain levels of 0.5%, 1.5% and 2.0%. However,
most researchers commonly used the 0.2% proof stress as the yield stress. In the
0.2% proof stress method, the yield strength is taken as the stress value
corresponding to the intersection point of the stress-strain curve and the 0.2% strain
proportional line (Figure 3.9b). In the case of other strain values, the yield strength is
the stress corresponding to the intersection point of stress-strain curve and the
vertical line corresponding to the strain level specified as shown in Figure 3.9 (b). In
this research, the 0.2% proof stress values were used as the yield strength since it is
commonly used as the yield strength by many researchers.
σu
Stress, σ
σy
Δσ
E = tan α =
Δε
Strain, ε
α
(a) Sharp yielding

(b)
σu σ2.0
σ0.5 σ1.5
σ0.2
σp
Stress, σ
Δσ
E = tan α =
Δε
α
0.002 0.015 Strain, ε
0.005 0.020
(c) Gradual yielding
Figure 3.9: Stress-strain Curves
3.2.4.2 Yield Strength
Figure 3.10 (a) to (d) show the typical stress-strain curves obtained for 1.95 mm
thick low strength and 1.9 mm thick high strength cold-formed steels at ambient and
500oC temperatures. Other stress-strain curves are given in Appendix A. These
stress-strain curves were used to calculate the yield strengths using a procedure that
depends on the type of stress-strain curve as mentioned earlier.
400
350
300
250 Extensometer
Stress (MPa)
Strain Gauge
200
150
100
50
0
0 5 10 15 20 25 30 35
Strain (%)
(a) 1.95 mm G250 steel at 20oC
160
140
120
100
Stress (MPa)
Extensometer
80
60
40
20
0
0 4 8 12 16 20 24 28 32
Strain (%)
(b) 1.95 mm G250 steel at 500oC
Figure 3.10: Stress-strain Curves of Low and High Strength Steels at Ambient
and Elevated Temperatures
600
500
400
Stress (MPa)
Extensometer
300
Strain Gauge
200
100
0
0 1 2 3 4 5 6 7 8 9 10 11 12
Strain (%)
(c) 1.9 mm G450 steel at 20oC
280
240
200
Stress (MPa)
160
Extensometer
120
80
40
0
0 1 2 3 4 5 6 7 8 9 10
Strain (%)
(d) 1.9 mm G450 steel at 500oC
Figure 3.10: Stress-strain Curves of Low and High Strength Steels at Ambient
Figure 3.11 (a) and (b) show the comparison of stress-strain curves for 1.95 mm
thick G250 and 1.9 mm thick G450 cold-formed steels at elevated temperatures. The
stress-strain curves of low strength steel shows a linear elastic region followed by a
flat region at ambient temperature and the yield strength was taken as the stress
corresponding to the flat region. Temperatures at 100oC and 200oC show similar
kind of stress-strain curves but do not provide a smooth flat region as for ambient
temperature. In these cases, the yield strength was taken as the average value of
stresses in the flat region (Figure 3.11a). At temperatures beyond 200oC the stress-
strain curves of G250 steel were of the gradual yielding type and did not include a
flat yield region.
500
450
400
350
Stress (MPa)
300
250
200
150
100
50
0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 5.5 6.0 6.5 7.0 7.5 8.0
Strain (%)
20 Deg. 100 Deg. 200 Deg. 300 Deg.
400 Deg. 500 Deg. 600 Deg. 700 Deg.
(a) 1.95 mm Grade 250 cold-formed steel
600
500
400
Stress (MPa)
300
200
100
0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
Strain (%)
20 Deg. 100 Deg. 200 Deg. 300 Deg. 400 Deg.
(b) 1.9 mm Grade 450 cold-formed steel
Figure 3.11: Stress-strain Curves at Different Temperatures
Unlike the low strength steel, the high strength steel gave gradual yielding type
stress-strain curves both at ambient and elevated temperatures as shown in Figure
3.11(b). Due to the absence of a clear yield point, the 0.2% proof stress was taken as
the yield strength for gradual yielding type stress-strain curves. In addition, the
stresses at 0.5%, 1.5% and 2.0% strain levels were measured from the intersection of
stress-strain curve and non-proportional vertical line at the specified strain values
(see Figure 3.9b).
The reduction factors of yield strength at elevated temperatures were calculated as

the ratio of yield strength at elevated temperatures, σy,T, to that at ambient
temperature, σy,20. Table 3.3 gives the average yield strength at different
temperatures while Table 3.4 gives the corresponding yield strength reduction
factors for different temperatures. Measured yield strengths and corresponding yield
strength reduction factors of 1.55 and 1.95 mm thick G250, 1.5 and 1.9 mm thick
G450 steels and 0.95 mm thick G550 steels are presented in Appendix A. Figure
3.12 shows the deterioration of yield strength with increasing temperatures for cold-
formed steels of both low and high strength steels with two different thicknesses.
1.2
1.0
1.95 - G250
0.8
1.55 - G250
1.90 - G450
0.6 1.50 - G450
fy,T/fy,20
0.4
0.2
0.0
0 100 200 300 400 500 600 700 800
o
Temperature, C
Figure 3.12: Yield Strength Reduction Factors versus Temperature for

Different Steel Grades and Thicknesses
Table 3.3: Yield Strength Values (σy,T) for Cold-formed Steels at Various Strain Levels
1.55 mm Grade 250 Cold-formed Steel 1.5 mm Grade 450 Cold-formed Steel
Temperature, oC
0.2 % 0.5 % 1.5 % 2.0 % 0.2 % 0.5 % 1.5 % 2.0 %
20 293.5 294.7 293.7 293.8 537.1 536.5 538.2 543.2
100 279.0 278.5 277.1 278.2 526.6 526.7 533.2 539.2
200 274.5 265.5 275.5 282.5 534.5 523.2 555.1 568.9
300 187.9 199.7 253.3 275.4 508.7 505.0 547.0 553.0
400 144.9 155.8 187.7 198.3 373.6 375.5 402.7 401.3
500 94.9 101.0 113.0 118.2 193.1 204.9 234.9 237.3
550 - - - - 125.0 134.5 158.0 162.6
600 61.3 69.0 74.2 76.0 59.0 64.3 78.8 81.7
700 37.4 38.1 40.5 41.0 33.0 34.9 39.4 40.0
Temperature, oC
0.2 % 0.5 % 1.5 % 2.0 % 0.2 % 0.5 % 1.5 % 2.0 %
20 270.5 270.2 270.6 271.3 514.5 511.2 521.0 526.0
100 267.3 269.0 269.8 270.6 521.1 527.5 537.5 548.2
200 257.0 258.0 261.5 291.3 510.1 512.3 531.6 540.6
300 196.4 208.3 259.1 280.8 483.8 482.5 527.7 531.3
400 147.7 156.4 186.0 199.0 362.3 363.6 392.7 393
450 - - - - 271.0 284.8 313.0 313.5
500 95.8 99.2 111.0 114.6 197.0 211.7 233.1 242.0
550 - - - - 123.5 127.3 157.1 160.8
600 54.0 56.5 61.3 63.0 56.4 60.7 71.0 73.1
700 34.2 34.9 36.8 37.2 34.2 35.9 39.8 40.6
Note: Average yield strength values based on two or more tensile coupon tests
Table 3.4: Yield Strength Reduction Factors (σy,T/σy,20) for Cold-formed Steels at Various Strain Levels
Temperature, oC
0.2 % 0.5 % 1.5 % 2.0 % 0.2 % 0.5 % 1.5 % 2.0 %
20 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000
100 0.951 0.945 0.943 0.947 0.980 0.982 0.991 0.993
200 0.935 0.901 0.938 0.962 0.995 0.975 1.031 1.047
300 0.640 0.678 0.862 0.938 0.947 0.941 1.016 1.018
400 0.492 0.529 0.639 0.675 0.695 0.700 0.748 0.739
500 0.323 0.343 0.385 0.402 0.360 0.382 0.436 0.437
550 - - - - 0.233 0.251 0.294 0.299
600 0.226 0.234 0.253 0.259 0.110 0.120 0.146 0.150
700 0.127 0.129 0.138 0.140 0.061 0.065 0.073 0.074
Temperature, oC
0.2 % 0.5 % 1.5 % 2.0 % 0.2 % 0.5 % 1.5 % 2.0 %
20 1.000 1.000 1.000 1.000 1.000 1.000 1.000 1.000
100 0.988 0.996 0.997 0.997 1.013 1.032 1.032 1.042
200 0.950 0.955 0.966 1.074 0.991 1.002 1.020 1.028
300 0.726 0.771 0.958 1.035 0.940 0.944 1.013 1.010
400 0.546 0.579 0.687 0.733 0.704 0.711 0.754 0.747
450 - - - - 0.527 0.557 0.601 0.596
500 0.354 0.367 0.410 0.422 0.383 0.414 0.448 0.460
550 - - - - 0.240 0.249 0.302 0.306
600 0.200 0.209 0.226 0.232 0.110 0.119 0.136 0.139
700 0.066 0.070 0.076 0.077
0.127 0.129 0.136 0.137
Figure 3.12 demonstrates that the reduction of yield strength of low and high
strength steels is clearly different from each other. For high strength steels, the
reduction factors were high for temperatures up to 500oC (ie. less reduction in yield
strength), but were low for temperatures above 500oC compared with low strength
steels. It appears that up to 200oC, the yield strength of high strength steel (G450) is
similar to the ambient temperature yield strength. Unlike the high strength steels, the
low strength steel loses its yield strength at a lower rate up to 200oC and thereafter a
rapid reduction of yield strength is observable. At 300oC the yield strength of high
strength steels reduced slightly and thereafter it reduced rapidly up to 600oC.
Comparatively, G450 steel loses its yield strength more rapidly than low strength
steel in the temperature range of 300oC to 600oC. Figure 3.12 reveals that the
specimen thickness does not have much influence on the yield strength reduction
factor of high strength steel. The effect of steel thickness on the yield strength
reduction factor for G250 also appears to be negligible with a maximum difference
of 8% at 300oC between the results of 1.55 mm and 1.95 mm thick steels.
Figure 3.13 (a) and (b) show the variation of reduction factors with respect to the
different strain levels for low and high strength steels, respectively. It is interesting
to note that the yield strength reduction factors based on 0.5% total strain is closer to
that based on 0.2% proof stress, for both low and high strength steels. Similar
observation has been made by Ranawaka and Mahendran (2009). Further, Table 3.4
shows that both methods based on 0.2% proof stress and 0.5% total strain level give
closer yield strength values in all the cases. Both G250 and G450 steels show higher
yield strength reduction factors for yield strengths based on higher strain levels
including 1.5% and 2.0%. It is found that the yield strength corresponding to 1.5%
and 2.0% strain levels is very close to the ultimate strength in some cases especially
for G450 steel. Therefore it is not safe to use the stresses based on 1.5% or 2.0%
total strain as the yield strength for design purposes. Considering all these facts it is
recommended that either 0.2% proof stress or stress at 0.5% total strain is used as the
yield strength for elevated temperatures.
1.2
1.0
1.95 - G250 - 0.2%

0.8 1.95 - G250 - 0.5%
1.95 - G250 - 1.5%
1.95 - G250 - 2.0%
0.6
fy,T/fy,20
0.4
0.2
0.0
0 100 200 300 400 500 600 700 800
o
Temperature, C
(a) 1.95 mm G250 cold-formed steel
1.2
1.0
1.9 - G450 - 0.2%
1.9 - G450 - 0.5%
0.8
1.9 - G450 - 1.5%
1.9 - G450 - 2.0%
0.6
fy,T/fy,20
0.4
0.2
0.0
0 100 200 300 400 500 600 700 800
Temperature, oC
(b) 1.9 mm G450 cold-formed steel
Figure 3.13: Yield Strength Reduction Factor versus Temperature for Different
Strain Levels
3.2.4.3 Elastic Modulus
In designing structural steel members elastic modulus plays an important role.

Similar to yield strength, elastic modulus also deteriorates with increasing
temperatures. Elastic modulus was calculated from the initial slope of the stress-
strain curve and is a measure of material stiffness. Reduction factors were then
calculated as the ratio of the elastic modulus at elevated temperature (ET) to that of
ambient temperature (E20). The average measured values of elastic modulus and
corresponding reduction factors at different temperatures for both low and high
strength steels are presented in Figure 3.14. The measured elastic modulus values
and the corresponding reduction factors obtained at different temperatures for 1.55
and 1.95 mm thick G250 steels, 1.9 and 1.95 mm thick G450 steels and 0.95 mm
thick G550 steels are presented in Appendix A.
Table 3.5: Elastic Modulus and Reduction Factors of Cold-formed Steel at

Elevated Temperatures
ET ET/E20
Temperature,
o
G250 G450 G250 G450
C
1.55mm 1.95mm 1.5mm 1.9mm 1.55mm 1.95mm 1.5mm 1.9mm
20
204385 188220 207490 206328 1.000 1.000 1.000 1.000
100
191505 179640 189375 196540 0.937 0.954 0.913 0.953
200
160555 171745 172000 173337 0.786 0.912 0.829 0.840
300
142470 154330 157380 148395 0.697 0.820 0.758 0.719
400
128220 121230 126030 118533 0.627 0.644 0.607 0.574
450
- - - 101330 - - - 0.491
500
81096 90631 96800 77100 0.397 0.482 0.467 0.374
550
- - 78904 72222 - - 0.380 0.350
600
62066 57777 63126 52536 0.304 0.307 0.304 0.255
700
24851 31363 47559 24286 0.122 0.167 0.229 0.118
Figure 3.14 shows that the elastic modulus of cold-formed steel decreases in a
similar trend irrespective of steel grades and thicknesses. In most cases the
discrepancy in the reduction factors of elastic modulus at a particular temperature is
less than 10%. Therefore it is considered that the effect of steel thickness and steel
grade on the reduction factors of elastic modulus at elevated temperatures is
insignificant.
1.2
1.0
1.95 - G250
1.90 - G450
0.8
1.55 - G250
1.50 - G450
0.6
ET/E20
0.4
0.2
0.0
0 100 200 300 400 500 600 700 800
Temperature, oC
Figure 3.14: Elastic Modulus versus Temperature for Different Steel Grades
and Thicknesses
3.2.4.4 Ultimate Strength
The ultimate strength is also an important factor in the fire design of steel structures.
It is defined as the maximum tensile strength that a material is capable of sustaining.
The ultimate strength reduction factor was calculated based on the ratio of ultimate
strength at a particular temperature (σu,T) to that at temperature (σu,20). Table 3.6
shows the average measured ultimate strength values and the corresponding
reduction factors of ultimate strength for different temperatures.
Figure 3.15 shows how the reduction factors of ultimate strength vary with
temperature. Reduction factors for ultimate strength of G250 and G450 steel are
closer to each other except at 200oC and 300oC. The effect of thickness on the
reduction factors of ultimate strength is insignificant for both low and high strength
steels as shown in Figure 3.15.
Table 3.6: Ultimate Strengths (σu,T) and Reduction Factors (σu,T/σu,20) of Cold-
formed Steel at Elevated Temperatures
Temp. (σu,T) (σu,T/(σu,20)

o G250 G450 G250 G450
C
20 361.0 356.1 561.8 542.5 1.000 1.000 1.000 1.000
100 379.5 369.0 561.5 559.0 1.051 1.036 1.000 1.030
200 442.0 435.2 597.0 562.0 1.224 1.222 1.063 1.036
300 388.2 385.0 560.0 536.7 1.075 1.081 0.997 0.989
400 250.0 240.0 402.6 392.7 0.693 0.674 0.717 0.724
450 - - - 313.5 - - - 0.578
500 144.0 137.5 239.5 236.3 0.399 0.386 0.426 0.435
550 - - 202.5 165.0 - - 0.360 0.304
600 85.5 71.4 89.3 88.9 0.237 0.200 0.159 0.164
700 41.9 37.7 41.7 41.8 0.116 0.106 0.074 0.077
1.4
1.2
1.0 1.95 - G250

1.90 - G450
1.55 - G250
0.8
1.50 - G450
0.6
fu,T/fu,20
0.4
0.2
0.0
0 100 200 300 400 500 600 700 800
o
Temperature, C
Figure 3.15: Ultimate Strength Reduction Factors versus Temperature for

Different Steel Grades and Thickness
It is interesting to note that there is almost 20% increase in the ultimate strength of
low strength steel at 200oC compared to that at ambient temperature. This can be
explained in relation to the chemical composition of cold-formed steel. Typical
chemical composition of cold-formed steel used in this research is given in Table 3.7
(BlueScope Steel, 2005). Increase in the ultimate strength of low strength steel can
be attributed to the transformations taking place in the steel base at low temperatures.
At lower temperatures, the effect of transformations taking place is predominant
compared to the reduction of ultimate strength due to temperature increase. With
increasing temperature, these transformations are retarded and therefore ultimate
strength reduces. Ultimate strengths of high strength steel are also slightly higher
compared to that at ambient temperature. Because of the same chemical composition
of both types of steel, high strength steel also exhibit the transformation of steel base
at low temperatures which leads to increase in the ultimate strength. However,
parallel with the increase of ultimate strength due to these transformations, the
strength increase due to cold-working drops with increasing temperature. Therefore
the increase in ultimate strength at low temperatures is not significant as for low
strength steel.
Table 3.7: Typical chemical composition of tested steel
Carbon Prosporous Manganese Sulphur Silicon Titanium Nitrogen

Grade
(C) (P) (Mn) (S) (SI) (TI) (N)
250 0.035 – 0.00 – 0.02 0.2 – 0.3 0.00 - 0.02 0.00 – 0.02 0.02 – 0.07 0.000 – 0.008
0.070
450 0.035 – 0.01 – 0.02 0.2 – 0.3 0.00 - 0.02 0.00 – 0.02 0.02 – 0.07 0.000 – 0.008
0.070
3.2.4.5 Ductility
Ductility of steel is defined based on the level of deformation that steel can undergo
plastically before fracture. This is also an important parameter in the design of steel
structures. The ductility of steel changes with increasing temperature compared to
the ductility at ambient temperature. In tensile coupon tests, the tensile strain was
measured until fracture in order to study the effect of temperature on ductility.
Stress-strain curves are plotted in the same graph for different temperatures as shown
in Figure 3.16 (a) and (b) for low and high strength cold-formed steels, respectively.
The effect of temperature and the steel grade on the ductility of steel were studied by
comparing the strain values at fracture.
500
450
400
350
Stress (MPa)
300
250
200
150
100
50
0
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40 42 44
Strain (%)
20 Deg. 100 Deg. 200 Deg. 300 Deg.
400 Deg. 500 Deg. 600 Deg. 700 Deg.
(a) 1.95 mm thick G250 steel
600
500
400
Stress (MPa)
300
200
100
0
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34
Strain (%)
(b) 1.90 mm thick G450 steel
Figure 3.16: Stress-strain Curves at Various Temperatures
Low strength steel shows higher ductility compared to high strength steel at ambient
temperature. This can be attributed to the comparatively high strain hardening caused
by the cold working in the case of high strength steel. It is noteworthy that low
strength steel has reduced ductility at temperatures in the range of 100oC to 500oC
compared to the ductility at ambient temperature while its ductility increased beyond
600oC. Ductility of high strength steel is similar at ambient temperature and 300oC
while at 100oC and 200oC it has been slightly reduced compared to the ductility at
ambient temperature. For both low and high strength steels, the lowest ductility is at
100oC.
It is interesting to note that ductility of low and high strength steels has reduced by
about 50% at 100oC. However, it improved when the temperature was increased
beyond 100oC for G250 steels and beyond 100oC to 400oC for G450 steels. Decrease
in ductility at 100oC can be attributed to chemical transformations taking place in the
steel base. With increasing temperature, these chemical transformations retard and
ductility increase due to temperature becomes predominant which leads to
improvement in the ductility at temperatures above 100oC compared to that of
100oC. At 450oC ductility of G450 drops significantly compared to 400oC and
thereafter it starts to improve again. Both G250 and G450 steels possess higher
ductility beyond 600oC compared to ductility at ambient temperature.
Typical failure modes of tensile test specimen of 1.95 mm thick G250 and 1.9 mm
thick G450 cold-formed steels at different temperatures are shown in Figure 3.17 (a)
and (b), respectively. Failure modes for 1.55 mm thick G250 and 1.50 mm thick
G450 cold-formed steels are shown in Appendix A. Up to 300oC, G450 steels show
less ductile failures (brittle with no necking) and thereafter their failures became
more ductile. Brittle failure can be seen in G250 steel only at 100oC and it shows
some ductile behaviour even at ambient temperature. For temperatures beyond
200oC, low strength steel becomes more ductile. At 700oC, both low and high grade
steels have high failure strains/elongation compared to that at ambient temperature.
700oC 600oC 500oC 400oC 300oC 200oC 100oC 20oC
(a) Grade 250 1.95 mm thick cold-formed steel
o
700oC 600oC 550oC 500oC 450 C 400oC 300oC 200oC 100oC 20oC
(c) Grade 450 1.90 mm thick cold-formed steel
Figure 3.17: Failure Modes of Tensile Specimens at Elevated Temperatures
3.2.5 Comparison of Measured Yield Strength and Elastic Modulus with
Those Available in Design Codes for Cold-formed Steel Structures
The reduction factors of yield strength and elastic modulus obtained from the tensile
coupon tests of cold-formed steels at different temperatures were compared with
those specified in the design standards, BS5950 Part 8 (BSI, 1990) and Eurocode 3
Part 1.2 (ECS, 2005). BS5950 Part 8 (BSI, 1990) provides reduction factors of yield
strength at total strain levels of 0.5%, 1.5% and 2.0% whereas Eurocode provides
them only based on 0.2% proof stress. Reduction factors given in BS5950 Part 8
(BSI, 1990) corresponding to 0.5% total strain level was used in the comparison
since the 0.2% proof stress and the stress corresponding to 0.5% of total strain are
very close to each other. Figure 3.18 shows the comparison of yield strength
reduction factors for both low and high strength steels from this research with
corresponding values given in BS5950 Part 8 (BSI, 1990) and Eurocode 3 Part 1.2
(ECS, 2005).
Figure 3.18 shows that there is a significant difference between the reduction factors
obtained from this research and the current cold-formed steel design standards for
temperatures beyond 100oC. The results are in good agreement at 100oC with the
design code predictions while the reduction factors at 200oC are on the safe side.
However, the discrepancy of reduction factors given in both design codes and the
measured values varies from 0 to 20%. It appears that most of the reduction factors
beyond 200oC given in both Eurocode 3 Part 1.2 (ECS, 2005) and BS5950 Part 8
(BSI, 1990) are unconservative while at 200oC and 300oC both design codes under-
predict the reduction factors for G450 steels. Yield strength reduction factors
obtained from this research are very much dependant on the grade of steel. However,
none of the design standards considered this in providing the reduction factors and
the reduction factors are the same for all steel grades. Therefore it is recommended
to use the yield strength reduction factors obtained from this research in the finite
element analysis and design.
1.90 - G450
1.2
1.50 - G450
1.95 - G250
1.0
1.55 - G250
0.95 - G550
0.8 Eurocode 3 Part 1.2 (SA, 2005)
fy,T/fy,20 BS5950 Part 8 (BSI, 1990)
0.6
0.4
0.2
0.0
0 100 200 300 400 500 600 700 800
o
Temperature, C
Figure 3.18: Comparison of Yield Strength Reduction Factors at Elevated

Temperatures with Current Design Standards
Since BS5950 Part 8 (BSI, 1990) does not provide reduction factors for elastic
modulus, only those given in Eurocode 3 Part 1.2 (ECS, 2005) was used in the
comparison. Reduction factors obtained from this study for elastic modulus
reduction factors are presented in Figure 3.19 together with those given in Eurocode
3 Part 1.2 (ECS, 2005). It is notable that Eurocode 3 Part 1.2 (ECS, 2005)
overestimates the elastic modulus reduction factors for both low and high strength
steels for all the temperatures except 700oC. Maximum difference of reduction
factors between Eurocode and the measured values is 23% and occurs at 500oC.
Figure 3.19 clearly shows that the use of reduction factors given in Eurocode 3 Part
1.2 (ECS, 2005) is not accurate and therefore it is recommended to use the reduction
factors obtained from this research for 1.55 mm and 1.95 mm thick G250 and 1.5
mm and 1.9 mm thick G450 cold-formed steels.
1.90 - G450
1.2
1.50 - G450
1.95 - G250
1.0
1.55 - G250
0.95 - G550
0.8 Eurocode 3 Part 1.2 (SA, 2005)
0.6
ET/E20
0.4
0.2
0.0
0 100 200 300 400 500 600 700 800
o
Temperature, C
Figure 3.19: Comparison of Elastic Modulus Reduction Factors at Elevated

Temperatures with Current Design Standards
3.2.6 Comparison of Yield Strength and Elastic Modulus Results with

Available Research Results
Outinen (1999) conducted tensile coupon tests on 2 mm thick cold-rolled hot dip
zinc coated structural steel S350GD+Z (Z35) and provided reduction factors for both
yield strength and elastic modulus. Mecozzi and Zhao (2005) conducted steady state
and transient state tests of 0.6 mm S280 and 1.2 mm and 2.5 mm S350 cold-formed
steels. They proposed reduction factors for effective yield strength, proportional
limit, 0.2% yield strength and elastic modulus. A series of tensile coupon tests was
also undertaken by researchers from the Queensland University of Technology
(Ranawaka and Mahendran, 2009a, Kolarkar, 2006, Bandula Heva, 2009). Ranwaka
and Mahendran (2009a) conducted tensile coupon tests covering steel thicknesses of
0.6 mm, 0.8 mm and 0.95 mm of two steel grades of Grade 550 and Grade 250 and
developed equations based on the steady state test results. Kolarkar (2006) conducted
tensile coupon tests on 1.15 mm G500 cold-formed steels and Bandula Heva (2009)
conducted tensile coupon tests on 0.42 mm G550 cold-formed steels. Chen and
Young (2007) provided a unified equation for yield strength reduction factors based
on the steady state tests carried out on 1.9 mm Grade 450 and 1.0 mm G550 cold-
formed steels. They also proposed a unified equation for the reduction factors for
elastic modulus based on the transient state test results carried out on 1.9 mm thick
Grade 450 steel. Reduction factors for yield strength are compared with those given
by other researchers in Figure 3.20
1.90 - G450
1.50 - G450
1.2 1.95 - G250
1.55 - G250
Ranawaka & Mahendran (2009a) - G550
1.0 Ranawaka & Mahendran (2009a) - G250
Chen & Young (2007)- G550
Chen & Young (2007) - G450
Outinen (1999) - 2.0 S350
0.8 Mecozzi & Zhao 2005 - S280
Mecozzi & Zhao 2005 - S350
0.6
fy,T/fy,20
0.4
0.2
0.0
0 100 200 300 400 500 600 700 800 900
o
Temperature, C
Figure 3.20: Comparison of Yield Strength Reduction Factors with Those

Obtained by Other Researchers
Ranawaka and Mahendran (2009a) noted a big difference in the reduction of yield
strength at elevated temperatures depending on the steel grade. This is also
confirmed by the results of Chen and Young (2007), Mecozzi and Zhao (2005) and
Outinen (1999). It was found that their results for high strength steels were higher
than that of low strength steels. The maximum difference exists with Chen and
Young’s (2007) results for G450 and it is about 40% at 400oC in comparison to the
results for Grade 250 steel.
Proposed yield strength reduction factors of Mecozzi and Zhao (2005) based on
0.2% proof stress for S280 and S350 steels do not agree with the test results of this
study for G250 or G450 cold-formed steel. Their proposed reduction factors for
S280 are overconservative in the temperature range of 200oC to 500oC in comparison
to the results of this study for G250 steel. Proposed reduction factors by Mecozzi and
Zhao (2005) for S380 steel are overconservative for results of high strength steel in
the temperature range of 200oC to 400oC while they are unconservative for high
strength steel in the temperature range of 500oC to 700oC. Chen and Young’s (2007)
predictions for 1.9 mm G450 cold-formed steel are in closer agreement up to 300oC.
Thereafter their predictions are unconservative up to 600oC compared with the
results from this study, even though, they are for the same steel grade and thickness.
There is a big difference of about 20% between the results from this research and
Chen and Young’s (2007) predictions for Grade 550 steel at 450oC. Outinen’s (1999)
results are too unconservative beyond 300oC for high strength steels while there is a
closer agreement up to 300oC. Therefore neither predictive equations of Chen and
Young (2007) for Grade 450 and Grade 550 steels nor the results of Outinen (1999)
for S355 steel and Mecozzi and Zhao (2005) for S280 and S350 cold-formed steels
are accurate for 1.55 mm and 1.95 mm thick Grade 250 and 1.5 mm and 1.9 mm
thick G450 cold-formed steels at elevated temperatures.
Ranawaka and Mahendran (2009a) proposed separate equations for both low and
high strength steels. Figure 3.20 shows that their proposed reduction factors for both
low and high strength steels agree well with the results from this research. However,
further comparison was made to find the accuracy of the predictive equations by
Ranawaka and Mahendran (2009a) for low and high strength steels separately
together with the results of Kolarkar (2006) and Bandula Heva (2009). Comparison
of results for low and high strength steels are shown in Figure 3.21 (a) and (b),
respectively. Figure 3.21 (a) clearly shows that the results from this research for
Grade 250 cold-formed steel are in good agreement with the reduction factors
obtained using the proposed equations for G250 steel by Ranawaka and Mahendran
(2009a). However, reduction factors from Ranawaka and Mahendran (2009a) are
slightly overconservative in the temperature range of 200oC to 400oC. The reduction
factors obtained from this study for Grade 450 steel are also in good agreement with
the results of Kolarkar (2006), Bandula Heva (2009) and the predictive equations by
Ranawaka and Mahendran (2009a) for Grade 550 cold-formed steel. However, at the
temperature ranging from 400oC to 600oC, reduction factors from Ranawaka and
Mahendran (2009a) for G550 steel are slightly unconservative for G450 steel. It is
further noticed that the proposed equation by Ranawaka and Mahendran (2009a) for
G550 steel are slightly unconservative in the same temperature range based on the
results of Kolarkar (2006) for G500 steels, Bandula Heva’s (2009) results and the
results of this study for G550 steels.
1.2
1.0 1.95 - G250
1.55 - G250
0.8 Ranawaka & Mahendran (2009a) - Eqn
0.6
fy,T/fy,20
0.4
0.2
0.0
0 100 200 300 400 500 600 700 800 900
Temperature, oC
(a) Low strength steel
1.2
1.90 - G450
1.50 - G450
1.0 0.95 - G550
Bandula Heva (2009) - 0.42 - G550
0.8 Kolarkar (2006) - 1.15 - G500
Ranawaka & Mahendran (2009a) - Eqn
fy,T/fy,20
0.6
0.4
0.2
0.0
0 100 200 300 400 500 600 700 800 900
o
Temperature, C
(b) High strength steel
Figure 3.21: Comparison of Yield Strength Reduction Factors
Reduction factors for elastic modulus are compared with those given by other
researchers in Figure 3.22. Results of Kolarkar (2006) and Bandula Heva (2009) are
closer to the results from this study except at 400oC in the case of Kolarkar’s (2006)
results. Further, the results of Outinen (1999) overestimate the elastic modulus in
most cases except at 700oC. It is also notable that the predictive equations of
Ranawaka and Mahendran (2009a) overestimate the results at 100oC and beyond
450oC. It appears that Chen and Young’s results accurately predict the reduction
factors from 100oC to 300oC while their results overestimate the reduction factors
between 400oC to 500oC. Mecozzi and Zhao’s (2005) results for S280 steel
overconservative for the temperature range of 300oC to 700oC while their results for
S350 steel are not accurate mostly at 100oC and 400oC. Therefore none of the
proposed equations can be used to calculate the reduction factors for 1.55 mm and
1.95 mm thick G250 and 1.5 mm and 1.9 mm thick G450 cold-formed steels at
elevated temperatures, especially at high temperatures.
1.90 - G450
1.2 1.50 - G450
1.95 - G250
1.55 - G250
0.95 - G550
1.0 Chen & Young (2007) - G450
Kolarkar (2006) - 1.15 - G500
Outinen (1999) - 2.0 - S350
Mecozzi & Zhao (2005) - S280
0.8 Mecozzi & Zhao (2005) - S350
Bandula Heva (2009) - 0.42 - G550
Ranawaka & Mahendran (2009a) - G550 and G250
0.6
ET/E20
0.4
0.2
0.0
0 100 200 300 400 500 600 700 800 900
Temperature, oC
Figure 3.22: Comparison of Elastic Modulus Reduction Factors with Those

Obtained by Other Researchers
3.2.7 Comparison of Reduction Factors for Yield Strength and Elastic
Modulus for Hot-rolled and Cold-formed Steels
All types of steel lose their strength when subjected to temperature increase.
However, there are differences in how their mechanical properties vary with
increasing temperature. In order to study the difference of reducing mechanical
properties of cold-formed steel in comparison to hot-rolled steel, the results from this
study were compared with available yield strength and elastic modulus reduction
factors for hot-rolled steels in the relevant design standards.
The design standards considered are Eurocode 3 Part 1.2 (ECS, 2005), AS 4100 (SA,
1998), and BS5950 Part 8 (BSI, 1990). BS 5950 Part 8 provides strength reduction
factors corresponding to 0.5%, 1.5% and 2.0% total strain while Eurocode 3 Part 1.2
provides reduction factors for hot-rolled steel based on 2.0% total strains and
proportional limit. Since the yield strength based on 0.5% total strain is closer to
0.2% proof stress, the former values were used in the comparison in the case of
BS5950 Part 8. Since the reduction factors based on 2% total strains and the
proportional limit are totally different from 0.2% proof stress, Eurocode 3 Part 1.2
values were not used in the comparison.
Figure 3.23 shows the comparison of reduction factors for yield strength obtained
from this research with that given in design codes. Figure 3.23 clearly shows that the
reduction factors for yield strength given in the design codes for hot-rolled steel
overestimates the reduction factors for cold-formed steel in the temperatures beyond
200oC, with a reasonable agreement up to 200oC. The difference between the yield
strength reduction factors from design codes and the results of this research varies
from 0 to 35%. This unconservative nature of code predictions also agrees well with
SCI publication (1993), which states that the loss of strength of cold-formed steel at
elevated temperatures exceeds that of hot-rolled steel by between 10 and 20% based
on data obtained from tests performed by British Steel (Welsh Laboratories). The
results of this research and Ranawaka and Mahendran (2009a) showed that the
reduction of yield strength with increasing temperature is dependent on the type of
steel grade. In contradiction, SCI publication (1993) stated that the performance of
hot-rolled steel in fire does not vary significantly with steel grade. In summary it is
considered that it is unsafe to use the reduction factors given for hot-rolled steel in
the fire safety design of cold-formed steel structures.
1.2
1.90 - G450
1.50 - G450
1.0 1.95 - G250
1.55 - G250
0.95 - G550
0.8 BS5950 Part 8 (BSI, 1990)
fy,T/fy,20 AS 4100 (SA,1998)
0.6
0.4
0.2
0.0
0 100 200 300 400 500 600 700 800
o
Temperature, C
Figure 3.23: Comparison of Yield Strength Reduction Factors with Those Given
in Hot-rolled Steel Design Standards
1.90 - G450
1.2
1.50 - G450
1.95 - G250
1.0 1.55 - G250
0.95 - G550
AS 4100 (SA, 1998)
0.8
0.6
ET/E20
0.4
0.2
0.0
0 100 200 300 400 500 600 700 800
o
Temperature, C
Figure 3.24: Comparison of Elastic Modulus Reduction Factors with Those

Given in Hot-rolled Steel Design Standards
BS5950 Part 8 does not provide reduction factors for elastic modulus while
Eurocode 3 Part 1.2 provides similar reduction factors for both cold-formed steel and
hot-rolled steel. Comparison of Eurocode 3 Part 1.2 reduction factors were carried
out in Section 3.2.9. Therefore only the AS4100 provisions for the elastic modulus
reduction factors were compared with the results of this study. The elastic modulus
reduction factors obtained from this research were compared with those predicted
from equations given in AS4100 in Figure 3.24. It can be seen that the predicted
values of AS4100 overestimate the reduction factors and the difference varies from 0
to 20%. This confirms that the AS4100 predictions are unsafe to be used in the fire
design of cold-formed steel structures.
3.2.8 Stress-Strain Model
Stress-strain curves of G450 cold-formed steel at ambient and elevated temperatures

are in the form of gradual yielding type while it is the same for G250 steel at
elevated temperature beyond 200oC. Ramberg and Osgood (1943) proposed a simple
formula to describe the stress-strain curve at ambient temperature in terms of three
parameters: elastic modulus and two secant yield strengths. Olawale and Plank
(1988) and Outinen et al. (1999) proposed stress-strain models for hot-rolled steels at
elevated temperatures based on Ramberg and Osgood stress-strain model. Lee et al.
(2003), Ranawaka and Mahendran (2009a) and Chen and Young (2007) proposed
models for stress-strain curves of cold-formed steels at elevated temperatures.
Models proposed by Lee et al. (2003) and Ranawaka and Mahendran (2009a) are
based on Ramberg-Osgood (1943) stress-strain model whereas the proposed model
by Chen and Young (2007) is based on the model developed by Mirambelll and Real
(2000) and Rasmussen (2003) for stainless steel at normal room temperature, which
is also originally based on the Ramberg and Osgood’s (1943) equation. Chapter 2
provides details of the models.
Olawale and Plank’s (1988) and Outinen’s (1999) models were calibrated with hot-
rolled steel test results while Lee et al.’s (2003), Ranawaka and Mahendran’s
(2009a) and Chen and Young’s (2007) models were calibrated with cold-formed
steel test results. The models proposed by Olawale and Plank (1988) and Outinen
(1999) based on hot-rolled steel test results are not suitable to represent the stress-
strain curves for cold-formed steel because of their different behaviour. Ranawaka
and Mahendran (2009a) stated that the stress-strain model proposed by Lee et al.
(2003) is not suitable to predict the stress-strain model of light-gauge cold-formed
steel at elevated temperatures. Chen and Young (2007) showed that the proposed
models by Olawale and Plank (1988), Outinen (1999) and Lee et al. (2003) do not
accurately predict the stress-strain curves of cold-formed steel for temperatures in
the range of 22oC to 660oC. Therefore only the models proposed by Ranawaka and
Mahendran (2009a) and Chen and Young (2007) were checked for accuracy against
the test results obtained from this research to find out the most suitable stress-strain
curve model at elevated temperatures for cold-formed steel.
In Figure 3.25 (a) and (b), the predicted stress-strain curves by Ranawaka and
Mahendran (2009a) at selected temperatures were compared against the stress-strain
curves obtained in this study for low and high strength cold-formed steels,
respectively. The comparison was made between the proposed stress-strain curve by
Chen and Young (2007) for high strength steel with the results of this study for low
and high strength steels in Figure 3.26 (a) and (b), respectively. The comparison of
predicted stress-strain curves with experimental results for 1.55 mm thick G250 and
1.5 mm thick G450 steels is shown in Appendix A. It can be clearly seen that the
stress-strain curves predicted by Ranawaka and Mahendran (2009a) for G450 steel
are in good agreement with the test results. Chen and Young’s (2007) predicted
stress-strain curves are in good agreement with the test results of G450 steel up to
about 0.5% strain. It can be seen that there is a significant difference thereafter at
temperatures 300oC, 400oC and 500oC. Therefore it is concluded that the stress-strain
curve model proposed by Ranawaka and Mahendran (2009a) for high strength steel
is more accurate compared to that of Chen and Young’s (2007) model.
It appears that there is a good agreement in the elastic region of the stress-strain
curves obtained from Ranawaka and Mahendran (2009a) for G250 steel. However it
is notable that the stress-strain model does not accurately predict the stresses in the
plastic strain range especially at 400oC and 500oC. The same observation is made
with Chen and Young’s predictions. However, the stress-strain curves obtained from
Ranawaka and Mahendran’s (2009a) equations are more closer to the stress-strain
curves obtained from the tensile coupon tests compared to those obtained from Chen
and Young’s (2007) equations.
300
250
200
Stress (MPa)
150
100
50
0
0.0 0.4 0.8 1.2 1.6 2.0 2.4 2.8 3.2 3.6 4.0
Strain (%)
Exp - 400 Deg. Exp - 500 Deg. Exp - 600 Deg. Exp - 700 Deg.
R - 400 Deg. R - 500 Deg. R - 600 Deg. R - 700 Deg.
600
500
400
Stress (MPa)
300
200
100
0
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0
Strain (%)
Exp - 600 Deg. Exp - 700 Deg. R - 20 Deg. R - 300 Deg.
Note: R – Ranawaka and Mahendran (2009a)

Figure 3.25: Comparison of Predicted Stress-strain Curves by Ranawaka and

Mahendran (2009a) with Test Results
300
250
200
Stress (MPa)
150
100
50
0
0.0 0.4 0.8 1.2 1.6 2.0 2.4 2.8 3.2 3.6 4.0
Strain (%)
C - 400 Deg. C - 500 Deg. C - 600 Deg. C - 700 Deg.
600
500
400
Stress (MPa)
300
200
100
0
0.0 0.5 1.0 1.5 2.0
Exp - 20 Deg. Exp - 300 Deg.
StrainExp
(%)- 400 Deg. Exp - 500 Deg.
Exp - 600 Deg. Exp - 700 Deg. C - 20 Deg. C - 300 Deg.
Note: C – Chen & Young (2007)

Figure 3.26: Comparison of Predicted Stress-strain Curves by Chen and Young

(2007) with Test Results
3.3 Predictive Equations for Mechanical Properties
3.3.1 Yield Strength
Outinen (1999), Lee et al. (2003), Mecozzi and Zhao (2005), Ranawaka and
Mahendran (2009a) and Chen and Young (2007) have proposed suitable equations
for yield strength reduction factors of cold-formed steels. Their proposals determine
the yield strength reduction factors as a function of temperature. Based on the yield
strength results obtained from tensile coupon tests at various temperatures in this
research, it was found that none of the proposals can be used to predict the yield
strength reduction factors of 1.55 mm and 1.95 mm thick G250 and 1.5 mm and 1.9
mm thick G450 cold-formed steels at elevated temperatures except the proposals of
Ranawaka and Mahendran (2009a). Although Ranawaka and Mahendran’s (2009a)
equations closely predict the yield strength reduction factors for cold-formed steels,
certain modifications were identified to be necessary in some cases. These
modifications to Ranawaka and Mahendran’s (2009a) equations were based on the
reduction factors obtained from the 0.2% proof stress method.
As observed by Ranawaka and Mahendran (2009a), there were differences between

the low and high strength steels in the reduction of yield strength with temperature
based on the results of this study. Ranawaka and Mahendran (2009a) have developed
separate equations for low and high strength steels. In Section 3.2.6, it was
demonstrated that the proposed equations by Ranawaka and Mahendran (2009a)
accurately predict the yield strength reduction factors of low strength steels.
However, it was identified that their proposed equations have to be modified in the
temperature range of 200oC to 400oC and therefore a new set of equations was
developed as given in Equation 3.2a and Equation 3.2b.
For low strength steels,

f y ,T
20 ≤ T ≤ 200 o C = −0.0005T + 1.01 (3.2a)
f y , 20
f y ,T
200 < T ≤ 800 o C = 25(1.16 − T 0.022 ) (3.2b)
f y , 20
Equations 3.2 (a) and (b) present the proposed equations for reduction factors
( f y ,T f y , 20 ) of low strength steels, where f y ,T and f y , 20 are the 0.2% proof stresses
at elevated and ambient temperatures, respectively, and T is the temperature. The

predictions from Equations 3.2 (a) and (b) are compared with the results from this
study in Figure 3.27 (a) and with the results from Ranawaka and Mahendran (2009a)
including this research in Figure 3.27 (b). As shown in Figure 3.27 (a) there is a
good agreement between the test results of this study and the values obtained from
the proposed equations while there is a good agreement between the results of
Ranawaka and Mahendran (2009a) and the proposed equations according to Figure
3.27 (b). Therefore it is recommended to use the modified equations 3.2 (a) and (b)
in the future to determine the yield strength reduction factors of low strength steels at
any given temperature.
1.2
1.95 - G250
1.0
1.55 - G250
Eqn 3.2
0.8
fy,T/fy,20
0.6
0.4
0.2
0.0
0 100 200 300 400 500 600 700 800 900
o
Temperature, C
(a) Low strength steels - This research
Figure 3.27: Comparison of Predicted Yield Strength Reduction Factors from

Equation 3.2 with Test Results
1.2 1.95 - G250
1.55 - G250
1.0 Ranawaka & Mahendran (2009a) - 0.60 - G250
Ranawaka & Mahendran (2009a) - 0.80 - G250
0.8 Eqn 3.2
0.6
fy,T/fy,20
0.4
0.2
0.0
0 100 200 300 400 500 600 700 800 900
o
Temperature, C
(b) Low strength steels – QUT research

In Section 3.2.6, it was shown that none of the proposals of previous researchers can
be used to calculate the reduction factors for 1.5 mm and 1.9 mm thick G450 cold-
formed steels. Although Ranawaka and Mahendran’s (2009a) equations predicted the
yield strength reduction factors reasonably well, it was identified that they have to be
improved in some regions to increase the accuracy. Therefore it was decided to
develop a new set of equations to determine the yield strength reduction factors of
high strength steels by considering all the results obtained by QUT researchers
including the results of this study. The reduction factors of high strength steels show
three main regions: two nonlinear regions (20oC – 300oC and 300oC – 600oC) and
one linear region (600oC – 800oC). Three different equations were therefore
developed for these three main regions as the first option. Equations 3.3(a) to (c)
present the proposed equations for reduction factors ( f y ,T f y , 20 ) of high strength
steels.
For high strength steels (Option 1),
f y ,T ⎧ (T − 20)4.56 ⎫
20 ≤ T < 300 o C = ⎨1 − ⎬ (3.3a)
f y , 20 ⎩ 1x1010 T ⎭
300 ≤ T < 600 C o

f y ,T ⎧
= ⎨0.95 −
(T − 300) ⎫
1.45
(3.3b)
⎬
f y , 20 ⎩ 7.76T ⎭
f y ,T
600 ≤ T ≤ 800 o C = −0.0004T + 0.35 (3.3c)
f y , 20
In Figures 3.28 (a) and 3.28 (b), the predictions from Equations 3.3 (a) to (c) are
compared with the results from this study and with the results from all the QUT
research including this research, respectively. As shown in Figures 3.27(a) there is a
good agreement between the test results of this study and the values obtained from
the proposed equations. Figure 3.28 (b) shows that a good agreement also exists
between the results of other QUT research and the proposed equations. Therefore it
is recommended to use the modified equations 3.3 (a), (b) and (c) to determine the
yield strength reduction factors of high strength steels (G450, G500 and G550) at
any given temperature.
1.2
1.0 1.90 - G450

1.50 - G450
0.95 - G550
0.8
Eqn 3.3
fy,T/fy,20
0.6
0.4
0.2
0.0
0 100 200 300 400 500 600 700 800 900
o
Temperature, C
(a) High strength steels - This research

1.90 - G450
1.50 - G450
1.2 0.95 - G550
Bandula Heva (2009) - 0.42 - G550
Kolarkar (2006) - 1.15 - G500
0.8 Eqn 3.3
0.6
fy,T/fy,20
0.4
0.2
0.0
0 100 200 300 400 500 600 700 800 900
o
Temperature, C
(b) High strength steels – QUT research

In the second option linear equations for 20oC to 300oC and 600oC to 800oC
temperature ranges and one non-linear curve for 300oC to 600oC were proposed
(Equations 3.4a, b and c). According to Figures 3.29(a) and (b), the results from all
the research conducted at QUT are in good agreement, and therefore Equation 3.4
can be used in predicting the yield strength reduction factors of high strength steels.

f y ,T
20 ≤ T < 300 o C = −0.000179T + 1.00358 (3.4a)
f y , 20
300 ≤ T < 600 o C

f y ,T ⎧
= ⎨0.95 −
(T − 300)1.45 ⎫ (3.4b)
⎬
f y , 20 ⎩ 7.76T ⎭
f y ,T
600 ≤ T ≤ 800 o C = −0.0004T + 0.35 (3.4c)
f y , 20
1.2
1.0 1.90 - G450

1.50 - G450
0.95 - G550
0.8
Eqn 3.4
0.6
fy,T/fy,20
0.4
0.2
0.0
0 100 200 300 400 500 600 700 800 900
o
Temperature, C
1.2 1.90 - G450

1.50 - G450
0.95 - G550
0.8 Bandula Heva (2009) - 0.42 - G550
Kolarkar (2006) - 1.15 - G500
Eqn 3.4
0.6
fy,T/fy,20
0.4
0.2
0.0
0 100 200 300 400 500 600 700 800 900
o
Temperature, C

As an alternative to Equations 3.3 and 3.4, three simple linear equations were
developed for the three main regions: 20oC – 300oC, 300oC – 600oC and 600oC –
800oC as given in Equations 3.5(a) to (c). Figures 3.30(a) and (b) show that there
exists a good agreement between the predicted values and the test results obtained
from this study and all the other QUT research. Therefore Equations 3.5(a) to (c) can
also be used to predict the yield strength reduction factors of high strength steels
accurately.

f y ,T
20 ≤ T < 300 o C = −0.000179T + 1.00358 (3.5a)
f y , 20
f y ,T
300 ≤ T < 600 o C = −0.0028T + 1.79 (3.5b)
f y , 20
f y ,T
600 ≤ T ≤ 800 o C = −0.0004T + 0.35 (3.5c)
f y , 20
1.2
1.0 1.90 - G450

1.50 - G450
0.8 0.95 - G550
fy,T/fy,20 Eqn 3.5
0.6
0.4
0.2
0.0
0 100 200 300 400 500 600 700 800 900
o
Temperature, C

1.90 - G450
1.2
1.50 - G450
0.95 - G550
1.0
Bandula Heva (2009) - 0.42 - G550
0.8
Kolarkar (2006) - 1.15 - G500
Eqn 3.5
0.6
fy,T/fy,20
0.4
0.2
0.0
0 100 200 300 400 500 600 700 800 900
o
Temperature, C

3.3.2 Elastic modulus
Deterioration of elastic modulus with increasing temperature directly influences the

performance of the structural member as it reduces the stiffness. Several researchers
(Outinen, 1999, Lee et al., 2003, Ranawaka and Mahendran, 2009a and Chen and
Young, 2007) have developed equations for elastic modulus reduction factors as a
function of T . It was found that none of these equations accurately predicted the
elastic modulus reduction factors. Therefore new empirical equations were
developed for elastic modulus with respect to the temperature. There are two main
regions in which reduction factors vary linearly: 20oC-200oC and 200oC–800oC.
Section 2.3.6 showed that the influence of steel grade and thickness on the modulus
of elasticity reduction factors is negligible and that there was not any identifiable
trend of reduction of elastic modulus with respect to the steel thickness or grade.
Hence neither steel thickness nor steel grade was included in developing the
predictive equations. Two linear equations were developed for the two identified
temperature regions to predict the elastic modulus reduction factors at elevated
temperatures (Equation 3.6).
For low and high strength steels,
ET
20 ≤ T ≤ 200 o C = −0.000835T + 1.0167 (3.6a)
E20
ET
200 < T ≤ 800 o C = −0.00135T + 1.1201 (3.6b)
E 20
The calculated elasticity modulus values from the above equations and the results of
this study are compared in Figure 3.31(a) while Figure 3.31(b) compares them with
the results obtained from all the QUT research. The predicted values agree well with
the test results of this study as well as the results of all the other QUT research.
1.2
1.0
1.90 - G450
1.50 - G450
0.8 1.95 - G250
1.55 - G250
0.95 - G550
0.6 Eqn 3.6
ET/E20
0.4
0.2
0.0
0 100 200 300 400 500 600 700 800 900
o
Temperature, C
(a) This research
Figure 3.31: Comparison of Predicted Elastic Modulus Reduction Factors with

Test Results
1.90 - G450
1.50 - G450
1.2
1.95 - G250
1.55 - G250
0.95 - G550
1.0 Bandula Heva (2009)- 0.42 - G550
Kolarkar (2006) - 1.15 - G500
0.8
Eqn 3.6
ET/E20
0.4
0.2
0.0
0 100 200 300 400 500 600 700 800 900
o
Temperature, C
(b) QUT research
Figure 3.31: Comparison of Predicted Elastic Modulus Reduction Factors

with Test Results
3.3.3 Stress-strain Curves
The stress-strain model at elevated temperature is usually based on the Ramberg-

Osgood stress-strain model and is given as shown in Equation 3.7 where ε T is the
strain corresponding to a given stress f T at temperature ( T ), ET and f y ,T are elastic
modulus and yield strength, respectively, and η T and β are two parameters.
Ranawaka and Mahendran (2009a) proposed β to be taken as 0.86 and two
equations for η T depending on the grade of steel as given in Equation 3.8 (a) and
(b).
ηT
f ⎛ f y ,T ⎞⎛ f T ⎞
ε T = T + β ⎜⎜ ⎟⎟⎜ ⎟ (3.7)
ET ⎜ ⎟
⎝ T ⎠⎝ f y ,T
E ⎠
where,
For high strength steels (G550), 20 ≤ T ≤ 800 o C
η T = −3.05 x10 −7 T 3 + 0.0005T 2 − 0.2615T + 62.653 (3.8a)
β = 0.86
For low strength steels (G250), 350 ≤ T ≤ 800 o C

η T = 0.000138T 2 − 0.085468T + 19.212 (3.8b)
β = 0.86
In Section 3.2.8, it was shown that the stress-strain curve model proposed by
Ranawaka and Mahendran (2009a) accurately predicted the stress-strain curves for
cold-formed high strength steel based on the results of this study. Hence it is
recommended their proposed stress-strain model for G550 steels to be used for 1.5
mm and 1.9 mm thick G450 cold-formed steels. However, it was shown that their
stress-strain model for low strength steels does not accurately predict the stresses in
the plastic strain range especially at 400oC and 500oC. Therefore the parameter β
was modified to 1.5 in Equation 3.7 and used with Equation 3.8 (b). The temperature
range that the Equation 3.8b valid is also changed based on the results of this
research from 350 ≤ T ≤ 800 o C to 300 ≤ T ≤ 800 o C . Figures 3.32 (a) and (b) show
the comparison of stress-strain curves obtained from the modified equations and the
experimental study for low strength steels. These figures clearly show that these
stress-strain curves are in good agreement. Therefore Equations 3.7, 3.8 (a) and 3.8
(b) together with β equal to 0.86 for high strength steels and 1.5 for low strength
steels are recommended for the determination of the stress-strain curves.
300
250
200
Stress (MPa)
150
100
50
0
0.0 0.4 0.8 1.2 1.6 2.0 2.4 2.8 3.2 3.6 4.0
Strain (%)
M - 400 Deg. M - 500 Deg. M - 600 Deg. M - 700 Deg.
(a) 1.55 mm thick G250 steels
300
250
200
Stress (MPa)
150
100
50
0
0.0 0.4 0.8 1.2 1.6 2.0 2.4 2.8 3.2 3.6 4.0
Strain (%)
M - 400 Deg. M - 500 Deg. M - 600 Deg. M - 700 Deg.
Note: M – with Modification

(b) 1.95 mm thick G250 steels
Figure 3.32: Stress-strain Curves from Equations 3.7 and 3.8 with
Modified β Value
3.4 Conclusion
This chapter has presented a detailed experimental study of the mechanical

properties of light gauge cold-formed steels at elevated temperatures. The
experimental study included tensile coupon tests carried out on two steel thicknesses
and two steel grades of light gauge cold-formed steels at various elevated
temperatures. The yield strength, ultimate strength, elastic modulus and the stress-
strain curve were determined from each tensile coupon test. The results showed that
the steel grade has an influence on the yield strength of steel, while there is no
observable influence of steel thickness on the results. There was no clear relationship
between the elastic modulus and the steel grade or thickness. Neither current design
standards nor proposals by other researchers provide accurate reduction factors for
both yield strength and the modulus of elasticity of light gauge cold-formed steels at
elevated temperatures. By considering the results of this study it was identified that
certain modifications are necessary for yield strength and elastic modulus reduction
factors of low and high strength steels. Accordingly a new set of predictive equations
was developed for yield strength and elastic modulus reduction factors of low and
high strength steels at elevated temperatures based on all the results obtained for the
mechanical properties of cold-formed steel at QUT and the results of this study.
Ranawaka and Mahendran’s (2009a) equations predict the stress-strain curves for
light-gauge high strength cold-formed steels reasonably accurately. However, an
improvement was made to these equations in the case of low strength steels. The use
of accurate mechanical properties as suggested by this research will lead to safe and
economical design of light gauge cold-formed steel structures under fire conditions.
4 Experimental Investigation of Cold-formed
Steel Flexural Members
The objectives of this research were to investigate the behaviour of cold-formed steel
beams at elevated temperatures and to check the applicability of the present design
methods given in fire and ambient temperature design codes for lateral-torsional
buckling. Finite element analysis was used extensively in this research considering
the practical difficulties, excessive time and resources required with experimental
tests at elevated temperatures. However, finite element models must be validated
before using them in detailed parametric studies. For this purpose, eight full scale
cold-formed steel beam tests were carried out at ambient temperature. This chapter
presents the details, results and the discussion of this experimental study.
4.1 Section Geometry and Specimen Sizes
A commonly used mono-symmetric lipped channel section was selected for the
beam tests (Figure 4.1). Elastic buckling analyses were first carried out on lipped
channel sections using a well known finite strip analysis program Thin-Wall to select
suitable cross-section sizes and beam spans. Figure 4.2 shows the results of finite
strip analyses using Thin-Wall in a buckling plot format.
Figure 4.1: Cross-section of a Lipped Channel Section

Local buckling
Lateral-torsional buckling
Distortional buckling
Figure 4.2: Buckling Plot of Selected Channel Section from Thin-Wall
The selection of beam spans was made based on Figure 4.2 so that lateral-torsional
buckling would occur in the tests. The beam span was varied from 1500 mm to 2900
mm in order to produce a range of beam slenderness ratios in the tests. Section
dimensions of 110 mm web depth, 40 mm flange width and 15 mm lip were selected
so that the beams with selected spans would fail by lateral-torsional buckling without
any local buckling effects. Table 4.1 shows the lateral-torsional buckling test
program for lipped channel beams. All the test specimens were labelled as follows to
indicate their grade, thickness and span.
G250 - 1.9 - 2000
Span (mm)
Thickness (mm)
Steel grade
Table 4.1: Test Program
Specimen Nominal Cross-section Span
Designation Grade
No. Dimensions (mm) (mm)
1 G250 - 1.95 – 1500 d - 110 1500
2 G250 - 1.95 – 2000 b - 40 2000
250
3 G250 - 1.95 – 2500 l - 15 2500
4 G250 - 1.95 – 2900 t - 1.95 2900
5 G450 - 1.90 – 1500 d - 110 1500
6 G450 - 1.90 – 2000 b - 40 2000
450
7 G450 - 1.90 – 2500 l - 15 2500
8 G450 - 1.90 – 2900 t - 1.90 2900
Eight beam tests were undertaken covering two steel grades and four beam spans.
All the cross-section dimensions were measured using a vernier calliper. The coated
thickness of each beam was measured using a micrometer screw gauge. The base
metal thickness of each steel grade and thickness was taken as the average of
measured base metal thicknesses (t) given in Table 3.1 of Chapter 3, and is given in
Table 4.2. The centre line dimensions of the cross-section were calculated based on
the measured external dimensions and are also given in Table 4.2. They were used
to obtain the elastic buckling and ultimate strengths of beams based on finite element
analyses.
Table 4.2: Measured Section Dimensions
Test Test Beam dimensions (mm)

Test Beam Designation
No. t b d l
1 G250 - 1.95 – 1500 38.37 108.56 13.57
2 G250 - 1.95 – 2000 38.23 108.52 13.51
1.943
3 G250 - 1.95 – 2500 38.15 108.42 14.49
4 G250 - 1.95 – 2900 38.00 108.55 13.79
5 G450 - 1.90 – 1500 37.82 107.85 13.48
6 G450 - 1.90 – 2000 38.40 108.87 13.46
1.886
7 G450 - 1.90 – 2500 39.78 107.97 13.77
8 G450 - 1.90– 2900 37.98 107.97 13.81
4.2 Measurement of Initial Geometric Imperfections
Two types of global initial geometric imperfections may exist in mono-symmetric

cold-formed steel C-section beams, namely negative and positive imperfections. The
negative imperfection is identified by the lateral movement of the cross-section
outward with anticlockwise twist without any cross-sectional deformation as shown
in Figure 4.3 (a) while the cross-section deformed inward with a clockwise twist in
the case of positive imperfection as shown in Figure 4.3 (b).
The direction and magnitude of initial geometric imperfection present in a beam is

very important especially for mono-symmetric sections as its effect on the ultimate
failure capacity is considerably high (Pi et al., 1997). Hence it is important to
measure the initial imperfections generated as a result of press braking process used
in making the lipped channel beams for the test program.
Outward deformation
Inward deformation
(a) Negative Imperfection (b) Positive Imperfection
Figure 4.3: Negative and Positive Geometric Imperfections
The member out-of-plane imperfections were measured for each test beam using the
equipment shown in Figure 4.4. The imperfection measuring equipment comprises
of a level table with guided rails with an accuracy of 0.01 mm, a laser sensor,
travelator to move the sensor and a data logger. The laser sensor attached to the
travelator could be moved in-plane and normal to the plane of the beam. The test
beams were kept on the table as shown in Figure 4.4 and levelled using the
adjustable screws of the legs and clamped. The laser sensor was moved along the
beam in a straight line and readings were recorded at every 100 mm intervals. The
readings were taken along three lines in the longitudinal direction of web and two
lines along the flanges of the beam in order to measure the initial crookedness and
twist.
Travelator
Test Beam
Laser sensor
Figure 4.4: Imperfection Measuring Device
Figures 4.5 (a) and (b) show how the beam web and flanges are initially deformed
along a 2500 mm span test beam, respectively. All the beams were deformed in the
negative direction as shown in Figure 4.5. The measured initial imperfections for
each test of specimen are given in Table 4.3 together with the fabrication tolerance
limit L/1000; where L is the beam span as specified in AS4100 (SA, 1998).
Measured imperfections are well below the fabrication tolerance limit and are always
closer to L/2000.
0.2
Length (mm)
0.0
0 500 1000 1500 2000 2500 3000
-0.2
-0.4
Im p e rfe ctio n (m m )
-0.6 Web - bottom

Web - Central
Web - top
-0.8 Web - Average
-1.0
-1.2
Initially Deformed
Shape and Direction
-1.4
(a) Along Web

Negative Imperfection
Figure 4.5: Measured Initial Geometric Imperfections
0.5
0.4
0.3
0.2
Im pe rfec tion (m m )
0.1
0.0
0 500 1000 1500 2000 2500 3000
-0.1
Length (mm)
-0.2
-0.3 Flange 1 - Line 1 Flange 1 - Line 2

Flange 1 - Average Flange 2 - Line 1
-0.4 Flange 2 - Line 2 Flange 2 - Average
(b) Along Flange
Figure 4.5: Measured Initial Geometric Imperfections
Table 4.3: Measured Initial Geometric Imperfections of Test Beams and

Fabrication Tolerance Limit (AS4100, 1998)
Test Measured* Fabrication Tolerance

Beam Designation
No. Imperfection (mm) Limit (mm)
1 G250 - 1.95 – 1500 -0.64 -1.5

2 G250 - 1.95 – 2000 -0.81 -2.0
3 G250 - 1.95 – 2500 -0.86 -2.5
4 G250 - 1.95 – 2900 -1.50 -2.9
5 G450 - 1.90 – 1500 -0.28 -1.5
6 G450 - 1.90 – 2000 -0.59 -2.0
7 G450 - 1.90 – 2500 -1.18 -2.5
8 G450 - 1.90 – 2900 -1.66 -2.9
* - All the measured imperfections were in the negative direction.
4.3 Lateral-torsional Buckling Tests
Mahaarachchi and Mahendran (2005) designed and built a test rig to conduct lateral
distortional buckling tests of LiteSteel Beams (LSB). The same test rig was used in
this study to investigate the lateral-torsional buckling of simply supported mono-
symmetric lipped channel beams. The test rig includes specially designed support
and loading systems that facilitate testing of simply supported mono-symmetric
beams subjected to quarter point loadings applied through the shear centre. The
schematic and overall views of the test set up are shown in Figures 4.6 and 4.7.
P P
L/4 L/2 L/4
Y
X Test beam
Bending moment distribution M=PL/4
Figure 4.6: Schematic View of Test Arrangement and Bending Moment

Distribution in the Test Beam
The test rig included an internal frame which contains four columns and an external
frame consisting of two main beams and four columns. The internal frame is
positioned in the external frame and rigidly attached to the main beams of the
external frame on top while the bottom is connected rigidly to the ground. The main
beams were positioned horizontally at 2 m height with a 1.8 m gap between each
other and connected to each pair of columns which are firmly fixed to the ground.
The support system is connected to the internal frame at either end. The loading
systems were suspended from a specially made wheel system that rested on SHS
beams positioned on top of the main beams directly over the loading points of the
test beams.
Wheel system
External frame
Loading system
Test beam
Support System
Internal frame
(a) Overall view of test rig
Test beam
(b) Beam supported at two ends
Figure 4.7: Test Rig
4.3.1 Support System
The support system developed by Maharachchi and Mahendran (2005) was used in
this experimental study. This support system allows the test beam to be simply
supported in-plane and out-of-plane at its ends. In other words, the beam ends were
fixed against in-plane vertical displacement, out-of-plane horizontal displacement
and twist rotations while they were unrestrained against in-plane and out-of-plane
rotations and warping displacements. Figure 4.8 shows the overall view of the
support system used in lateral-torsional buckling tests of cold-formed steel lipped
channel beams.
Box frame
Internal frame
Side guide and

running track
Test beam
Ball bearing
Figure 4.8: Support System
The two support systems at either ends are the same except one of them is prevented
from moving in the longitudinal direction while the other is allowed to move freely.
The box frame has two ball bearings on its two sides and another two at the top and
bottom of the box frame (Figure 4.8). The side ball bearings allows beam to rotate
freely about the major axis of the test beam (in-plane rotations) while the top and
bottom bearings allow rotations about the minor axis (out-of-plane rotations) and
warping displacements. The in-plane vertical movements and out-of-plane horizontal
movements at both support ends were prevented by the running tracks and side
guides. The two supports were aligned with each other to ensure that the vertical
deflections remained in the same plane.
4.3.2 Loading System
The test beam was loaded at two quarter points of the beam using the loading system
developed by Mahaarachchi and Mahendran (2005). This loading system ensured
that the test beam was loaded through the shear centre of the mono-symmetric
sections. It eliminated the effects of load height and torsional loading effects. The
overall loading system is shown in Figure 4.9 (a).
The loading system included two hydraulic rams connected to a wheel system, load
cell and a series of other components as shown in Figure 4.9. The hydraulic rams
were operated under displacement control using a single hydraulic pump to apply
equal loads at quarter points of test beam. Maintaining the same load at the two
loading points enables a symmetric and uniform bending moment distribution
between the two loading points as shown in Figure 4.6. The loads were applied
vertically upward. The hydraulic jacks were supported on a rail system which had
the capability of moving in either parallel or perpendicular directions to the beam
span (Figure 4.9c). Therefore this system was able to apply the loads to the test beam
without any restraint to their displacements or rotations at the loading points in any
direction. Therefore all the six degrees of freedoms at the loading position of the test
beam can be considered as unrestrained. A universal joint at the bottom of load cell
and at the connecting arm ensured that the load was applied at the shear centre
without applying a torque to the beam (Figure 4.9b).
Hydraulic jack
(d) Wheel system
(c) Hydraulic jack connected

to wheel system
Load cell connected to

hydraulic jack
Universal joint
Loading arm
Loading at
shear centre
(a) Overall view (b) Loading arm
Figure 4.9: Loading System and its Components
4.3.3 Measuring system
In each test the applied loads and the deformations were measured until failure. The
EDCAR (Experimental Data Collection and Recorder) system was used to record all
the measurements automatically (Figure 4.10). Test beam was loaded at quarter
points using two hydraulic jacks, each with 50 kN capacity. To measure the applied
loads, two 60 kN load cells were attached to the loading arms as shown in Figure
4.9. The horizontal deflections at the top and bottom of the flanges at mid-span were
measured. The vertical deflections of the compression flange at mid-span and at both
loading points of the test beam were also measured. Deflection measurement points
are shown in Figure 4.11. Wire potentiometer type displacement transducers (WDT)
were connected to measure the displacements at these points. Calibration factors of
the WDTs (i.e. unit of voltage per mm) were determined and input to the EDCAR
UNIT before starting the test.
Figure 4.10: EDCAR (Experimental Data Collection And Recorder)
Load cell connected on top

to hydraulic jack
Horizontal deflection
Universal joint measuring points
Loading arm
Vertical deflection
measuring points
Figure 4.11: Measuring System
4.3.4 Test Procedure
The cross-section dimensions and thickness of each specimen were measured using a
vernier calliper and a micrometer, respectively. The test beam length was 60 mm
longer than the beam span to allow for the connections at each support. Three 10 mm
diameter holes were drilled on the web at the support assembly and the loading
points. Deflection measuring points were marked at mid-span and loading points of
the beam. Nuts were glued to the beam at the required points to attach the wire
displacement transducers (WDTs).
The test rig (support and loading systems) was set for the required beam span. The
ends of the test beam were connected to the support system by using M10 bolts and a
12 mm thick clamping plate. The use of clamping plate was expected to prevent web
crippling and twisting of the section at the supports. The flanges were free to warp
independently. Figure 4.12 shows the end connection of the test beam to the support
system.
Test beam
Clamping plate
Figure 4.12: End Connection of Test Beam
The loading arms were bolted to the web at the quarter points of test beam (Figure
4.13). Additional plates were used between the clamping plate and the loading arm
in order to apply the loads through the shear centre of the cross-section. The support
frames were aligned with each other to avoid any initial twisting. The loading jacks
and the arms were aligned with each other in order to prevent any eccentricities and
to ensure the application of equal vertical loads at quarter points through the shear
centre of the test beam. Wire displacement transducers (WDTs) were attached to the
test beam at the required positions. The WDTs for horizontal displacements were
held by tripod while those for vertical displacements were held by weights on the
ground. The load cells and displacement transducers were connected to the data
logger. Each data channel was individually checked to ensure correct operation.
Load applied
through the
shear centre
Loading arm
Test beam
Figure 4.13: Connection at Loading Point
A small load was applied first to allow the loading and support system components
to settle evenly on wheels and bearings. Then the measuring system was initialised
to make all the readings equal to zero. A trial load of 10% of the expected ultimate
load was applied and released in order to remove any slackness in the system and to
check the functionality. However, it may not work for long beams since a 10% load
is very small and therefore it is not enough to remove the slackness of the system.
Therefore the results may not be accurate.
After preparing the system the load was applied gradually using manual pumping to
the hydraulic system. Load and displacement readings were recorded automatically
at every 2 seconds by the data logger, EDCAR. The load-displacement curves were
plotted and displayed on the computer screen automatically during the test and until
the test beam failed. The applied load first increased gradually and then reached its
maximum when the test beam buckled out-of-plane, and finally started to unload. In
all the tests the loading was continuously applied until the test beam failed by lateral-
torsional buckling.
4.4 Experimental Results
Test beams were simply supported at both ends and loaded at quarter points and
therefore they were subject to a uniform bending moment distribution between the
two quarter points with a linear moment distribution over the end segments (Figure
4.6). The developed maximum uniform moment was calculated using PL/4, where L
is the beam span and P is the applied load at each quarter point. The mean value of
the two load cell readings at quarter points was used to calculate the failure moment.
Table 4.4 presents the mean values of the two quarter point loads (P) and the
corresponding moments at failure. It also includes the failure directions of the test
beam. Figure 4.14 shows the test beam at failure.
Table 4.4: Ultimate Lateral-torsional Buckling Moment Capacities
Beam Designation P (kN) Ultimate Moment (kNm) Failure Direction
G250 - 1.95 – 1500 9.83 3.69

G250 - 1.95 – 2000 7.10 3.55
G250 - 1.95 – 2500 5.22 3.26 All the beams
G250 - 1.95 – 2900 4.09 2.97 failed in the
G450 - 1.90 – 1500 13.68 5.13 Positive Direction
G450 - 1.90 – 2000 8.88 4.44
G450 - 1.90 – 2500 6.41 4.01
G450 - 1.90 – 2900 4.70 3.41
Mono-symmetric beams could fail in two different directions, namely positive and
negative directions as shown in Figure 4.15. The failure direction is mainly
dependent on the magnitude and direction of initial geometric imperfections. Beams
with negative geometric imperfection always fail in the negative direction while
beams with positive geometric imperfection can fail in either direction depending on
the magnitude of imperfection as shown by the preliminary finite element analyses
and further discussed in Chapter 5.
Tension flange
Compression flange
moved inward
(Failure in the
Positive Direction)
Figure 4.14: Cold-formed Steel Lipped Channel Beam at Failure
Compression flange
(a) Failure in the Positive Direction (b) Failure in the Negative Direction
Figure 4.15: Possible Failure Directions of Lipped Channel Beam
In this study all the test beams failed in the positive direction by lateral-torsional
buckling although their measured geometric imperfections were in the negative
direction. Since beams with negative imperfections always deform in the negative
direction, the two loading arms used in the test rig might have introduced some
restraints preventing the failure of test beam in the negative direction. Therefore the
test results obtained in this research may not represent accurate beam capacities.
Similar observations were made by Poutre et al. (2000) in their experimental study
on hot-rolled steel unlipped channel beams. In their tests of four identical beams with
positive imperfections, only one of them failed in the positive direction while the
remaining three failed in the negative direction. This confirms that test beams subject
to lateral-torsional buckling can fail in the opposite direction to their initial
geometric imperfection direction due to various unavoidable factors associated with
testing, and hence lead to lower or higher moment capacities than numerical or
theoretical capacities.
During the tests the vertical and horizontal deflections of the compression flange at
mid-span were measured together with the corresponding applied load. Figure 4.16
(a) to (d) illustrate the applied moment versus deflection curves obtained for G450
and G250 steel beams of different spans.
4
M (kN m )
G450 - 1500
2
G450 - 2000
G450 - 2500
1 G450 - 2900
0
0 5 10 15 20 25 30 35 40
Vertical Deflection (mm)
(a) Applied Moment (M) vs. Vertical Deflection

(G450 Steel Beams)
Figure 4.16: Moment-Deflection Curves
4.0
3.5
3.0
2.5
M (kN m )
2.0
G250 - 1500
1.5 G250 - 2000
G250 - 2500
1.0 G250 - 2900
0.5
0.0
0 5 10 15 20 25 30
-0.5
(b) Applied Moment (M) vs. Vertical Deflection

(G250 Steel Beams)
6.0
5.0
4.0
M (k N m )
3.0
2.0
G450 - 1500
G450 - 2000
1.0 G450 - 2500
G450 - 2900
0.0
-5 0 5 10 15 20 25 30
Horizontal Deflection (mm)
(c) Applied Moment (M) vs. Horizontal Deflection

(G450 Steel Beams)
4.0
3.5
3.0
2.5 G250 - 1500

G250 - 2000
M (kN m )
2.0
G250 - 2500
G250 - 2900
1.5
1.0
0.5
0.0
-5 0 5 10 15 20 25 30
-0.5
Horizontal Deflection (mm)
(d) Applied Moment (M) vs. Horizontal Deflection

(G250 Steel Beams)
4.5 Discussion of Test Results
The ultimate failure moments obtained from eight lateral-torsional buckling tests in
this research were compared with the predicted ultimate moment capacities based on
cold-formed steel design standards, AS/NZS 4600 (SA, 2005), North American
Specification (NAS) (AISI, 2007) EC3 Part 1.3 (ECS, 2006), BS5950 Part 5 (BCS,
1998) and fire design code EC 3 Part 1.2 (ECS, 2005). In addition to them, the hot-
rolled steel design standard, AS4100 (SA, 1998), was also considered. The design
rules in the NAS (AISI, 2007) and AS/NZS 4600 are identical and hence only
AS/NZS 4600 is mentioned in the discussions. Appendix C shows the sample
calculations of moment capacities.
The predicted moment capacities from the design codes are for beams subjected to a
uniform bending moment distribution. However, the test beam was subjected to a
non-uniform bending moment distribution over the span due to the quarter point
loading method adopted here. This type of loading develops a uniform moment
distribution within the two loading points while a linear moment distribution exists
between the supports and loading points. The failure moments from experiments
may be affected by the non-uniform moment distribution. In order to compare the
test results with code predictions, it was necessary to account for the non-uniform
moment distribution. Therefore the failure moments from experiments were divided
by a moment modification factor (αm) of 1.09 as recommended by AS 4100 (SA,
1998). Table 4.5 presents the experimental failure moments and the moment
capacities predicted by the current design codes.
It can be seen that the experimental failure moments are much higher compared with
the predicted moment capacities using the design standards. As the length increases,
the difference is also increasing and in some cases experimental results were more
than twice the predicted capacity. The experimental failure moments for long beams
were found to exceed the elastic buckling moment capacities. Higher moment
capacities were considered to be mainly due to the failure direction of the beam.
Being a mono-symmetric section, cold-formed steel lipped channel beams may fail
either in the positive or negative direction (Figure 4.3). The strength of such beams is
affected by the failure direction, which is determined by the direction and magnitude
of initial geometric imperfections. In all the tested beams initial geometric
imperfections were negative. However, all of them failed in the positive direction.
The reason for this may be that the beam’s failure in the negative direction was
restricted by the two loading arms. Normally mono-symmetric lipped channel beams
failing in the positive direction have higher ultimate moment capacities compared to
those which failed in the negative direction and this is further discussed in Section
5.1.8.
Higher ultimate moment capacities were achieved in the tests relative to elastic
buckling moment capacities because all the test beams in this study failed in the
positive direction. Similar observations were made in the experimental studies of
Poutre et al. (2000) and Ungermann and Kalameya (2006) as reported by Snijder et
al. (2002) and Brune and Ungermann (2008), respectively. Poutre et al. (2000) tested
2.8 m long hot-rolled steel unlipped channel beams (UPE160 section) with loads at
quarter points. Snijder et al. (2002) conducted experimental and numerical studies of
lipped channel beams. In Poutre et al.’s (2000) lateral-torsional buckling tests of four
identical beams with positive initial imperfection, only one of them failed in the
positive direction and led to 10% higher failure load than others which failed in the
negative direction. This again confirms that failure in the positive direction is the
reason for the occurrence of higher failure loads in lateral-torsional buckling tests.
Ungermann and Kalameya (2006) conducted 14 experimental tests of thin-walled

channel shaped welded and cold-formed steel beams. Their cold-formed beams were
made of 3 mm Grade S235 steel and had web and flange sizes of 162 mm and 83
mm. Some of their test beams also experienced higher failure loads when the beam
failed in the positive direction. The failure load of one test beam that failed in the
positive direction was 58% higher than the numerical results based on negative
imperfections according to Brune and Ungermann (2008). Brune and Ungermann
(2008) state that the direction of global lateral imperfection and displacement of the
test beam have a strong influence on its load carrying capacity. They also state that if
the global imperfections and the increasing displacements due to lateral-torsional
buckling are in the positive direction, a ductile failure appears with a slowly
increasing load. Pi et al.’s (1997) numerical study and Put et al.’s (1999b)
experimental study also reported higher failure loads for lateral-distortional buckling
of lipped channel beams with positive imperfections.
In summary, past experimental and numerical studies confirm that the ultimate
failure loads of mono-symmetric sections subject to lateral buckling are very
sensitive to the failure direction and that failure loads were significantly increased
with failures in the positive direction. They also showed that test beams could fail in
a certain direction independent of their initial geometric imperfections. Higher
moment capacities were achieved in this experimental study due to the possible
lateral restraints at the loading points which forced the test beam with negative
imperfections to deform and buckle in the positive direction.
Table 4.5: Comparison of Experimental Moment Capacity Results with Code Predictions
Member moment capacities, Mb Mu/Mb
Failure Direction
Exp Mu / 1.09
Designation
AS/NZS 4600
AS/NZS 4600
Test No
EC 3 PART
EC 3 PART
EC 3 PART
EC 3 PART
Exp Mu
Length
Beam
PART 5
PART 5
AS4100
AS4100
BS5950
BS5950
DSM
DSM
1.3
1.2
1.3
1.2
1 G250 - 1.95 - 1500 1500 + 3.684 3.38 2.612 3.157 2.845 2.667 1.873 3.16 1.29 1.07 1.19 1.27 1.80 1.07
2 G250 - 1.95 - 2000 2000 + 3.548 3.255 1.883 2.439 2.043 1.939 1.404 2.442 1.73 1.33 1.59 1.68 2.32 1.33
3 G250 - 1.95 - 2500 2500 + 3.267 2.997 1.439 1.771 1.516 1.472 1.115 1.771 2.08 1.69 1.98 2.04 2.69 1.69
4 G250 - 1.95 - 2900 2900 + 2.971 2.726 1.141 1.351 1.192 1.164 0.913 1.351 2.39 2.02 2.29 2.34 2.99 2.02
5 G450 - 1.90 - 1500 1500 + 5.128 4.705 3.3 4.27 3.853 3.394 2.741 4.275 1.43 1.10 1.22 1.39 1.72 1.10
6 G450 - 1.90 - 2000 2000 + 4.436 4.07 2.254 2.693 2.525 2.3 1.953 2.693 1.81 1.51 1.61 1.77 2.08 1.51
7 G450 - 1.90 - 2500 2500 + 4.002 3.671 1.706 1.974 1.872 1.744 1.523 1.974 2.15 1.86 1.96 2.10 2.41 1.86
8 G450 - 1.90 - 2900 2900 + 3.405 3.124 1.248 1.418 1.351 1.281 1.144 1.418 2.50 2.20 2.31 2.44 2.73 2.20
Note: ‘+’ indicates failure in positive direction.
It was noted that different design codes predict varying capacities. Both AS/NZS
4600 (SA, 2005) and DSM predictions are very close to each other and their
predictions are greater than others’ predictions. Eurocode 3 Part 1.2 (ECS, 2005)
predictions are the lowest predictions. The predictions of BS5950 Part 5 (BCS,
1998) and AS4100 (SA, 1998) are of the same order as Eurocode 3 Part 1.3 (ECS,
2006).
This experimental study has shown that the design code predictions are very
conservative in the case of cold-formed steel lipped channel beams that failed by
lateral-torsional buckling in the positive direction. However, the design code rules
are likely to have been developed based on the negative imperfections since they
give the lowest failure moments. Therefore negative geometric imperfections should
be adopted in the parametric study on the lateral-torsional buckling of lipped channel
4.6 Conclusions
Eight lateral-torsional buckling tests of cold-formed steel lipped channel beams were
conducted at ambient temperature in this study. Failure moments from these tests
appeared to have been increased by the undesirable loading restraints and initial
geometric imperfections. Hence they did not agree well when compared with the
corresponding moment capacities predicted by various steel design codes. Further
research is needed based on finite element analyses of lipped channel beams.
5 Finite Element Modelling of Cold-formed
Steel Flexural Members
Problem solving based on computer simulations such as finite element analysis has
become increasingly popular in current research. Computer simulations are efficient
and save time and resources compared to physical experiments. This research needs
a wide range of data to develop and verify accurate design rules. However, there are
practical difficulties in conducting a large number of physical experiments.
Therefore finite element analysis provides a good alternative to physical
experiments, especially in the parametric study.
In this research few experimental tests were conducted to verify the accuracy of the
finite element models. Eight beam tests were conducted for a typical lipped channel
beam with four different spans and two steel grades. All the test beams failed by
lateral-torsional buckling. The ultimate moment capacities and moment versus
deflection curves were obtained from these tests. Chapter 4 presented the details of
these lateral-torsional buckling tests conducted on cold-formed steel lipped channel
flexural members at ambient temperature.
Abaqus Version 6.8 (HKS, 2007) provided by the High Performance Computing and
Research Support Section of the Queensland University of Technology was used for
finite element analysis. Experimental results were used in the validation of finite
element models, which are to be used in a detailed parametric study aimed at
investigating the effects of a range of parameters on the flexural behaviour of cold-
formed steel beams at ambient and elevated temperatures.
This chapter presents the details of developing suitable finite element models to
simulate the structural behaviour of cold-formed steel lipped channel beams subject
to local and lateral-torsional buckling effects at ambient and elevated temperatures,
and the validation process of the developed models.
5.1 Development of Finite Element Models
Two types of finite element models were required in this research: ideal finite
element models to undertake a parametric study and experimental finite element
model for validation purposes.
Two types of ideal finite element models were developed: one for studying member
moment capacities relating to lateral-torsional buckling and the other for section
moment capacities relating to local buckling. The main ideal model is capable of
analysing beams that fail by lateral-torsional buckling. Use of this model is limited
to beam spans having capacities less than yield moment capacity. The section
moment capacity is governed by local buckling or yielding and can be obtained by
analysing short beams. It may exceed the yield moment capacity depending on the
compactness of the section. Therefore a separate finite element model was developed
to study the section moment capacity of lipped channel beams.
The main ideal finite element model investigating the lateral-torsional buckling
behaviour incorporates idealised simply supported boundary conditions with a
uniform bending moment within the span. The other ideal model simulates a beam
subject to four-point loading. Developed models were used in the parametric study
for developing suitable design rules. The schematic diagrams of these two models
are shown in Figures 5.1 (a) and (b).
Experimental finite element model simulates the actual conditions used in the
experimental study such as geometry, loads, constraints and mechanical properties as
closely as possible. In the experiments, the beams were simply supported at both
ends with quarter point loading (Figure 5.1(c)). All these beams failed in lateral-
torsional buckling. Therefore these experimental results were used in validating the
main ideal model developed to investigate the lateral-torsional buckling behaviour.
Although the loading type of experimental model is different from the ideal model
for lateral-torsional buckling, it can be used to validate the ideal model indirectly in
terms of element type, mesh density, material model and boundary conditions.
The ideal model for lateral-torsional buckling was further validated by using Pi et
al.’ (1997) non-linear finite element analysis results for lateral-torsional buckling of
cold-formed steel beams subject to a uniform moment at ambient temperature and by
also using Poutre et al.’s (2000) experimental results on unlipped hot-rolled channel
beams undergoing lateral-torsional buckling failures as alternatives to the validation
based on experimental results reported in Chapter 4.
L
M M
(a) Ideal model for member moment capacity (lateral-torsional buckling)
Laterally and 0.05 m Laterally and

torsionally torsionally
restrained restrained
L
0.55 m
(b) Ideal model for section moment capacity (local buckling and yielding)
L/4 L/2 L/4
(c) Experimental model
Figure 5.1: Physical Arrangements of Finite Element Models
The finite element modelling was carried out using MSC/PATRAN pre-processing
facility and the analysis was carried out by submitting the output file from
MSC/PATRAN pre-processing facility to ABAQUS (HKS, 2007). The results were
viewed by the post-processing facility provided in MSC/PATRAN.
5.1.1 Element Type
The geometry of the cross-section was modelled using the centreline dimensions of
the lipped channel beam as illustrated in Figure 5.2. A wide range of elements is
available in ABAQUS element library, which can be used to solve different types of
problems. The most suitable element type has to be selected based on the problem to
be solved. Shell elements have been successfully used in studies related to buckling
of thin-walled steel beams (Avery and Mahendran, 2000, Kurniawan and
Mahendran, 2009a,b). Use of shell elements in the finite element models provides
sufficient degrees of freedom to explicitly model the buckling deformations and
spread of plasticity. Four noded shell elements, S4, S4R, S4R5 and nine noded shell
elements, S9R5, are often used in the buckling analysis. Both S4 and S4R are
general purpose shell elements while S4R5 and S9R5 are thin-only shell elements.
First letter in the element name indicated the type of element (‘S’ for shell element)
and the first number indicates the number of nodes in the element. R in the element
name stands for reduced integration with hourglass control and the next number
indicates the degrees of freedom at each node.
Figure 5.2: Lipped Channel Beam Cross-section Used in Finite Element

Modelling
The most suitable element type for this research was selected based on elastic
buckling analyses carried out using the main ideal model and comparing their results
with those from Thin-Wall finite strip analysis program (Figure 5.3). The cross-
section dimensions of the G450 cold-formed steel beam selected for these analyses are
110 mm in web depth, 40 mm in flange width and 12 mm lip. Figure 5.3 shows that
the most accurate elastic lateral-torsional buckling moments are attained when using
S4 elements while the use of S4R5 elements yields the lowest accuracy. However, the
accuracy gain using S4 element type over S4R5 element type is small (< 1%
difference). On the other hand the disk space and processing time involved with S4
element type is considerably high compared to S4R5 element type. Therefore it was
decided to use S4R5 element type in the finite element models. S4R5 element can be
described as a four noded thin, shear flexible, quadrilateral, reduced integration shell
element with five degrees of freedom per each node: three translations and two in-
plane rotations (i.e. no rotations about the shell normal). Shell thickness was included
as a section property to describe the shell cross-section. Nine section points were
selected through the shell thickness to describe the flexural residual stresses.
1.000
0.995
0.990
Mo,FEA/Mo,TW
0.985 S4R5
S4R
S4
0.980
S9R5
0.975
0.970
0.965
1.0 1.5 2.0 2.5 3.0 3.5
Span, m
Figure 5.3: Comparison of Elastic Lateral-torsional Buckling Moments

Obtained from FEA and Thin-Wall
5.1.2 Finite Element Mesh
The accuracy of finite element analysis results also depends on the mesh size. Use of
as many elements as possible increases the accuracy. However, as the number of
elements used increases the required processor time and disk usage also increase
making it uneconomical. Therefore it is desirable to select the most economical
element size considering the accuracy, processor time and the disk usage.
Preliminary investigations were carried out by conducting elastic buckling analyses
to find the most suitable element size by using different element sizes while keeping
the element aspect ratio constant at 1:2 (width:length). The elastic lateral-torsional
buckling moments obtained from finite element analyses with different element sizes
were compared with those obtained from Thin-Wall computer software. Based on
the results, the width and length of the elements were selected as 5mm and 10mm,
respectively. Figure 5.4 shows the typical finite element mesh used in the analyses of
cold-formed steel lipped channel beams.
Table 5.1: Comparison of Elastic Lateral-torsional Buckling Moments

for Different Element Sizes
L - 1.5 m G450 20oC L - 1.5 m G450 500oC

Element Size
TW FEA TW/FEA TW FEA TW/FEA
2.5 x 5.0 6.31 6.28 1.005 2.81 2.81 1.000
5.0 x 10.0 6.31 6.26 1.008 2.81 2.80 1.004
7.5 x 15.0 6.31 6.23 1.013 2.81 2.79 1.008
Beam d – 115 mm l – 15 mm
Grade 450
Dimensions b – 45 mm t – 1.9 mm
Note: TW – Thin-Wall Results FEA – Finite Element Analysis Results
Figure 5.4: Typical Finite Element Mesh for Lipped Channel Beams
5.1.3 Symmetry
Simply supported beams subjected to either four-point loading (experimental model

for member moment capacity and ideal model for section moment capacity) or
uniform bending moment (ideal model for member moment capacity - lateral-
torsional buckling) can be considered as symmetrical about the plane perpendicular
to the longitudinal axis (Z-axis) passing through the mid-span of the beam because
symmetry can be seen with respect to the geometry and loading patterns before and
after the deformations. In the case of support conditions, only one support provides
restraints against lateral translations while keeping other degrees of freedom same.
However, it can be considered as symmetrical about the mid-plane. Therefore it was
possible to model a half span beam for both ideal and experimental models which
reduces the computing requirements significantly. Figure 5.5 illustrates the applied
boundary conditions at mid-span and supports with the load application for both
ideal and experimental models.
L/2
ux = 0 uz = 0
M X Z θx = 0
uy = 0
θz = 0 θy = 0
(a) Ideal model for member moment capacity (lateral-torsional

buckling)
Y
L/2
ux = 0 uz = 0
uy = 0 X Z
θx = 0
θy = 0 θy = 0
θz = 0
Laterally Restrained
ux = 0
θy = 0
θz = 0
(b) Ideal model for section moment capacity (local buckling and
yielding)
Figure 5.5: Structural Arrangement and Applied Boundary Conditions
Y
L/4 L/4
ux = 0 uz = 0
uy = 0 X Z
θx = 0
θz = 0
θy = 0
(c) Experimental model for member moment capacity (lateral-

torsional buckling)
Note: un - Displacement in the direction of axis ‘n’
θn – Rotation about the axis ‘n’
n – Axis X, Y or Z
Figure 5.5: Structural Arrangement and Applied Boundary Conditions
5.1.4 Ideal Model for Member Moment Capacity (Lateral-torsional Buckling)
Member moment capacity of beams subjected to lateral-torsional buckling depends

on the moment distribution along the beam which is determined by the applied
loading type: uniform bending moment, concentrated loads or distributed loads, etc.
Real beams rarely undergo uniform bending. However, simply supported beam with
a uniform bending moment is the worst case that yields the lowest beam capacity in
lateral-torsional buckling. Hence this case is commonly used in generating beam
design curves (member capacity) for lateral-torsional buckling.
The ideal model for lateral-torsional buckling was developed with idealised simply
supported boundary conditions and a uniform bending moment over the entire span
(Figure 5.6). The idealized boundary conditions should satisfy the following
requirements according to Trahair (1993).
• Simply supported in-plane; both ends fixed against in-plane vertical

deflection, but unrestrained against in-plane rotation, and one end fixed
against longitudinal horizontal displacement.
• Simply supported out-of-plane; both ends fixed against out-of-plane
horizontal deflection and twist rotation, but unrestrained against minor axis
rotation and warping displacement.
Uniform moment
simulation
M
Support End (SPC 126

boundary condition applied
to the end)
(b) Support end
(a) Half span beam model
Symmetric plane
(SPC 345 boundary
condition applied to all
the nodes)
(c) Symmetric plane
Figure 5.6: Ideal Model for Member Moment Capacity (lateral-torsional

buckling)
Figure 5.6 shows the developed finite element model of a half-span beam for lateral-
torsional buckling. Axis system of the ideal model is such that Z axis lies in the
longitudinal direction of the beam while X and Y axes lie in the major and minor
axes of channel beam, respectively. The idealized simply supported boundary
condition was modelled by restraining degrees of freedom “1,2,6” of all the nodes at
the support end of the beam (Figure 5.6 (b)). This allows major and minor axis
rotations and warping displacements while restraining in-plane and out-of-plane
translations and twisting. The degrees of freedoms “3,4,5” was restrained in all the
nodes at the other end of the model representing symmetric plane (Figure 5.6 (c)).
The degrees of freedom notations “1,2,3” correspond to translations in X, Y, Z
directions while “4,5,6” represent the rotations about X, Y and Z axes, respectively.
The uniform bending moment distribution was achieved by applying a series of

linear forces to all the nodes at one end creating triangular distribution of forces
across the section as shown in Figure 5.6 (b). One half of the section is under tension
while other half is under compression creating a uniform bending moment
distribution throughout the beam. The effects of undesirable stress concentrations
were eliminated by including a 20 mm long strip of elastic elements at the support
end of the ideal model.
The method used to create a uniform bending moment at the end in this model
restricts the development of plastic moment capacity in the beam. Beams with very
short span may fail with moment capacities higher than the yield moment capacity
and thus this finite element model cannot be used to investigate the section moment
capacity. This finite element model was used to analyse beams with intermediate and
long spans, which failed by lateral-torsional buckling. Following section describes
the ideal model developed to study the section moment capacity of short span beams.
5.1.5 Ideal Model for Section Moment Capacity (Local Buckling)
The moment capacity of short beams is governed by yielding or local buckling

failures and is independent of the applied loading type or the support conditions of
the beam. In other words short span beams yield the same capacity irrespective of
the loading type or the support conditions used. Therefore the ideal model of short
span beams was developed as simply supported with a four-point loading so that the
model could be used to study the section moment capacity (Figure 5.5b). The central
part of the beam between the two loading points is under uniform bending and most
vulnerable to failure. Half of the beam was modelled due to symmetry in loading and
geometry. A rigid body of 100 mm in length and web height was modelled using
R3D4 elements as shown in Figure 5.7 in order to transfer the load applied at the
shear centre. A rigid body is used to model very stiff components and is a collection
of nodes and elements whose motion is governed by the motion of a single node
which is called as rigid body reference node. Rigid body reference node in this case
is the node at the shear centre. The end beam segment from support end to the end of
the rigid body was restricted to lateral and torsional deformations and modelled as
purely elastic since the objective of using the end segments was to create a uniform
moment between the loading points. This was achieved by restraining degrees of
freedoms “1,5,6” to all the nodes except the support end nodes, which prevent lateral
and torsional deformations while degrees of freedom “1,2,5,6” of all the end nodes
were restrained. For clarity application of “1,5,6” boundary condition was not shown
and only the region is indicated (Figure 5.7). Symmetric plane was applied with the
“3,4,5” boundary conditions similar to the previous model.
Figure 5.7 illustrates the developed half span model to study the section moment
capacity of short beams. The beam span was considered as the length between the
ends of two rigid bodies within the two loading points as the failure was restricted to
this region. The length of the end segments was kept as 0.5 m. The length between
the ends of two rigid bodies was determined as one or three times the half wave
buckling lengths obtained from Thin-Wall elastic buckling analysis corresponding to
the local buckling mode. S4R5 element type was used in this model as for the main
ideal model. A finer mesh with 5 mm x 5 mm elements was used because of the
short beam span.
Support End (SPC Symmetric plane
1256 boundary (SPC 345 boundary
condition applied to condition applied to
all the nodes at the all the nodes)
end support) Rigid body
Laterally and torsionally restrained

(SPC156 boundary condition
applied to all the nodes) Span/2
Figure 5.7: Ideal Model for Section Moment Capacity (local buckling)
5.1.6 Experimental Finite Element Model
Experimental finite element model was developed by simulating the physical

condition in the experimental test set up as accurately as possible. Figure 5.8
illustrates the experimental finite element model. In the experiments, simply
supported lipped channel cold-formed steel beams were loaded at the quarter points
through the shear centre until failure. The support system of the beams prevents in-
plane and out-of-plane deformations and twisting while allowing in-plane rotations,
out of plane rotations, z axis translations and warping displacements. Test beams
included a rigid plate at each support to prevent distortion and twisting of the
section. These plates were modelled as rigid body using R3D4 elements in
ABAQUS (HKS, 2007). The motion of the rigid body was controlled by applying
simply supported boundary conditions (SPC 126) to the shear centre, which acted as
the rigid body reference node.
Rigid body
SPC 126 boundary

condition applied to
the node at shear
centre
Figure 5.8: Experimental Finite Element Model
In testing, the load applied at the shear centre of the beam was transferred through
three bolts located as shown in Figure 4.13 in Chapter 4. Two 12 mm plates of 25 x
130 mm were used to strengthen the loading area. These plates were modelled as
rigid body using R3D4 elements in ABAQUS (HKS, 2007). The load applied at the
shear centre acts as the rigid body reference node and transfers the load to the beam.
5.1.7 Material Model and Properties
It is the objective of this research to conduct finite element analyses of cold-formed

steel lipped channel beams at different temperatures varying from ambient
temperature to 700oC. For this purpose mechanical properties such as yield strength,
elastic modulus, Poisson’s ratio and the stress-strain graph are required as input.
ABAQUS classical metal plasticity model was adopted in this research to include the
material non-linearity effects. The material model can be either elastic-perfect plastic
or isotropic stress hardening type. Elastic-perfect plastic type means that the yield
stress remains unchanged as the plastic strain increases whereas isotropic hardening
model allows the stress to increase with the plastic strain. Figure 5.9 shows the
elastic-perfect plastic and isotropic hardening material models. Normally higher
grade cold-formed steels show isotropic hardening behaviour at ambient temperature
as well as at elevated temperatures as confirmed by tensile coupon tests (see Chapter
3). Low grade cold-formed steels show the same above 300oC and below that a well
defined yield point is observable with a constant yield stress up to a certain plastic
strain. Therefore isotropic stress hardening model simulates the material behaviour
more accurately relative to the elastic-perfect plastic model. This is needed mostly at
elevated temperatures since the stress-strain curve becomes highly nonlinear.
Although the elastic-perfect plastic material model is sufficient at ambient
temperature to model the lateral-torsional buckling, isotropic hardening model was
selected over elastic-perfect plastic model since it simulates the actual material
behaviour accurately, especially at elevated temperatures. However, the elastic-
perfect plastic model was assumed for G250 steel below 300oC since its stress-strain
curve has a well defined yield point and a yield plateau.
σ
fy
Strain Hardening Model
Elastic-perfect Plastic Model
ε
Figure 5.9: Material Models
The elastic-perfect plastic material model is simple and only requires the yield
strength and the elastic modulus as input. The nominal yield strength was converted
to true yield strength using Equation 5.1 and used in the elastic-plastic material
model. Material plastic behaviour was defined in the isotropic stress hardening
model by inputting true stress and corresponding true strain values at different points
in the stress-strain curve. ABAQUS approximates the smooth stress-strain behaviour
of the material by joining the given points in the true stress versus true strain curve.
The nominal stress-strain data obtained from the output of tensile coupon tests were
converted to true stress and true strain values using the following equations.
σ true = σ nom (1 + ε nom ) (5.1)
σ true
ε lnpl = ln(1 + ε nom ) − (5.2)
E
where σ nom , σ true and E are the nominal stress, true stress and the elastic modulus,
respectively. ε nom and ε lnpl are the nominal strain and logarithmic plastic strain,
respectively.
In order to obtain the required mechanical properties, tensile coupon tests of 1.95
mm and 1.55 mm thick G250 steels and 1.9 mm and 1.5 mm thick G450 steels were
carried out. Based on the results obtained from this research (Chapter 3) and a
previous research study carried out by Ranawaka and Mahendran (2009), equations
for the reduction factors of yield strength and elastic modulus and the stress-strain
curve model were developed. In Chapter 3 all the measured values of yield strength
and elasticity modulus, developed equations for the reduction factors and the model
developed for stress-strain curves are given. The yield strength and the elastic
modulus at different temperatures that were needed for the ideal model were
obtained from the developed equations of the reduction factors based on the ambient
temperature mechanical properties as given in Table 5.2. The stress-strain curve was
also obtained from the developed model. It must be noted that it is conventional
practice to use the yield strength, elastic modulus and stress-strain curve of steels
obtained from tensile coupon tests in the numerical modelling of steel beams and
columns instead of compression coupon tests. The Poisson’s ratio was taken as 0.3
and was assumed to remain unchanged with increasing temperature as stated in
Ranby (1999).
In the experimental models (ambient temperature only), the isotropic hardening

model was implemented for high strength cold-formed steels while the elastic-
perfect plastic material model was used for low strength steels. The mechanical
properties that were obtained from the tensile coupon tests were used. This is
important because in the experimental models the measured mechanical properties
must be used since the results are used to validate the numerical models by
comparing with the experimental results.
Table 5.2 presents the mechanical properties at ambient temperature while Equations
5.3 to 5.7 present the reduction factors for yield strength and modulus of elasticity
and the stress-strain curve equations used for all the temperatures.
Table 5.2: Ambient Temperature Mechanical Properties used in the

Parametric Study
Grade Thickness (mm) fy (MPa) E (GPa)

1.55 294 204
G250
1.95 271 188
1.50 537 207
G450
1.90 515 206
Reduction Factors for Yield Strength
For G250 steels,
f y ,T
20 ≤ T ≤ 200 o C = −0.0005T + 1.01 (5.3a)
f y , 20
f y ,T
200 < T ≤ 800 o C
f y , 20
(
= 25 1.16 − T 0.022 ) (5.3b)
For G450 steels (Option 1),

f y ,T ⎧ (T − 20)4.56 ⎫
20 ≤ T < 300 C o
= ⎨1 − ⎬ (5.4a)
f y , 20 ⎩ 1x1010 T ⎭
300 ≤ T < 600 o C

f y ,T ⎧
= ⎨0.95 −
(T − 300) ⎫
1.45
(5.4b)
⎬
f y , 20 ⎩ 7.76T ⎭
f y ,T
600 ≤ T ≤ 800 o C = −0.0004T + 0.35 (5.4c)
f y , 20
Reduction Factors for Elastic Modulus

For G250 and G450 steels,
ET
20 ≤ T ≤ 200 o C = −0.000835T + 1.0167 (5.5a)
E20
ET
200 < T ≤ 800 o C = −0.00135T + 1.1201 (5.5b)
E 20
Equation for Stress-Strain Curve

ηT
f ⎛ f y ,T ⎞⎛ f T ⎞
ε T = T + β ⎜⎜ ⎟⎟⎜ ⎟ (5.6)
ET ⎜ ⎟
⎝ T ⎠⎝ f y ,T
E ⎠
For G450 steels, 20 ≤ T ≤ 800 o C
β = 0.86
η T = −3.05 x10 −7 T 3 + 0.0005T 2 − 0.2615T + 62.653 (5.7a)
For G250 steels, 300 ≤ T ≤ 800 o C

β = 1.5
η T = 0.000138T 2 − 0.085468T + 19.212 (5.7b)
5.1.8 Geometric Imperfections
Presence of imperfections in a beam as a result of transportation, handling and

installation processes directly affects its load carrying capacity. Initial geometric
imperfections of a cold-formed steel member can be categorised into two:
imperfections in the cross-section (local imperfections) and member deviations
without changing the cross-section (global imperfections). The magnitude and shape
of the imperfection are considered to be important because they influence the
ultimate moment capacity. The magnitude and shape of imperfections present in a
beam may vary from member to member and only limited data is available. The
magnitude of imperfections for the ideal short span finite element model subjected to
local buckling was estimated by using Equation 5.8 as proposed by Schafer and
Pekoz (1998). They have proposed two types of imperfections: Type 1 imperfection
which governs the local buckling failure mode and Type 2 imperfection which
governs the distortional buckling failure mode (Figure 5.10 (a)).
For Type 1 imperfection, Δ 1 = 0.006 w (5.8)
For Type 2 imperfection, Δ2 = t (5.9)
where w and t are the plate width and the thickness, respectively.
In the case of ideal finite element model for lateral-torsional buckling, the fabrication
tolerance of L/1000 as recommended by AS4100 (SA, 1998) for flexural members
was used as the overall member imperfection. Different shapes of possible
imperfections in a lipped channel beam are shown in Figure 5.10.
The initial imperfection was introduced in the non-linear analysis by

*IMPERFECTION option in ABAQUS using the lowest eigen mode obtained from
an elastic buckling analysis. The lowest eigen mode in an elastic buckling analysis
gives the most probable failure mode of a beam. This allows the input of the shape of
the imperfection. Accordingly the magnitude of imperfection to be used is decided.
The largest magnitude of the imperfection is assigned to the node that has the
maximum deformation obtained from the elastic buckling analysis as shown in
Figure 5.11. All the other nodes displace in relation to the maximum deformed node,
thus keeping the shape of the deformed beam similar to the shape of the lowest eigen
mode.
Δ3
Δ2
Δ1
Local buckling Flange distortional buckling Lateral-torsional buckling

(a) Imperfection in the cross-section (b) Imperfection in the member
Figure 5.10: Different Imperfection Types in a Lipped Channel Beam
Maximum imperfection
Maximum imperfection (+L/1000) at mid-span
(-L/1000) at mid-span Compression Flange
(a) Negative Imperfection (b) Positive Imperfection
Figure 5.11: Positive and Negative Geometric Imperfections
The imperfection for lateral-torsional buckling includes both lateral displacement
and twist rotation to the beam. In the case of mono-symmetric sections, geometric
imperfection relating to lateral-torsional buckling can be either negative or positive.
Negative imperfection in the case of lateral-torsional buckling is identified by lateral
deflection in a direction to the left with anticlockwise twist rotation while the
positive imperfection is identified by lateral deflection in a direction to the right with
clockwise twist rotation as shown in Figure 5.11. In Figure 5.11 the top flange is the
compression flange.
Since lipped channel beam is a mono-symmetric section, its ultimate moment

capacity is affected by the direction of initial imperfection as discussed in the next
section and Chapter 4. According to the comparison of ultimate failure moments
with negative and positive imperfections, the beams with negative imperfections had
lower moment capacities. The same observation was also made by Pi et al. (1997)
and Kurniawan and Mahendran (2009a) in their numerical studies. Ungermann and
Kalameya (2006) and Poutre et al. (2000) also observed higher failure moments in
their experimental study when the beam failed in the positive direction in
comparison to those failed in the negative direction. According to the geometric
imperfection measurements by Kwon (1992), negative imperfections are more likely
to be present than positive initial imperfections. Therefore a negative imperfection
with a magnitude of L/1000 was adopted in the nonlinear analysis of ideal models
since it yields lower bound results of lateral-torsional buckling strengths.
In the lateral-torsional buckling tests, the magnitudes of member imperfections were

measured for each test specimen, and are presented in Chapter 4. The measured
overall imperfection was in the negative direction with its magnitude always less
than L/2000. It is considerably less than the fabrication tolerance of L/1000
recommended by AS4100 (SA, 1998). In the experiments all the beams failed by
lateral-torsional buckling in the positive direction although the initial imperfections
were on the negative side. The reason was considered to be due to the possible lateral
restraints provided by the loading shaft against the failure towards the negative side.
Considering the smaller negative imperfection present in the test beams and the
ultimate failure direction of the beam (positive), the measured negative imperfection
was ignored. A positive imperfection of L/1000 was used in the case of experimental
model in order to simulate the observed failure in the positive direction.
5.1.8.1 Effect of Initial Imperfections
The ultimate moment capacity of mono-symmetric beams is affected by the

magnitude and direction of their initial geometric imperfections. Therefore their
effects were studied by varying the magnitude of negative and positive imperfections
from L/500 to L/10000 for a lipped channel section with three beam spans. Selected
dimensions of this channel section are given in Table 5.3. The ultimate moment
capacity results are given in Table 5.3 and Figure 5.12. It can be seen that as the
magnitude of negative imperfection increases the ultimate moment capacity
decreases. When the applied negative imperfection is twice the fabrication tolerance
of L/1000 the ultimate moment capacity decreases by 11%, 10% and 9% for 1.5 m,
2.0 m and 2.5 m beam spans, respectively. When the applied negative imperfection
is half of the fabrication tolerance (L/1000) the ultimate moment capacity increases
by 7%, 6% and 5.5% for 1.5 m, 2.0 m and 2.5 m beam spans, respectively. Therefore
it can be concluded that the effect of imperfection magnitude on the ultimate
moment capacity is significant.
It could be noted that all the beams with negative imperfections have failed in the
negative direction while the beams with positive imperfections fail on either side
depending on the magnitude of the imperfections (Table 5.3). When the positive
imperfections are small (L/5000 and L/10000) the beam fails on the negative side
and reaches only lower moment capacities compared to those failed on the positive
side. As the initial positive imperfection increases the beam starts to fail in the
positive direction with higher moment capacity. Figure 5.12 shows how the ultimate
moment capacity changes as the initial imperfection changes from a negative to
positive value.
Table 5.3: Effect of Imperfection Magnitude
(a) Negative Imperfection
L - 1.5 m L - 2.0 m L - 2.5 m

Imperfection
Magnitude
Mu Failure Mu Failure Mu Failure

(kNm) Direction (kNm) Direction (kNm) Direction
L/10000 5.95 3.62 2.52

L/5000 5.83 3.69 2.48
Negative Negative Negative
L/2000 5.59 3.56 2.38
L/1000 5.22 3.36 2.26
L/500 4.65 3.04 2.06
NI - Negative Beam D – 115 mm L – 15 mm Grade 450
Imperfection Dimensions B – 45 mm T – 1.9 mm
(b) Positive Imperfection
L - 1.5 m L - 2.0 m L - 2.5 m

Imperfection
Magnitude
Mu Failure Mu Failure Failure

Mu (kNm)
(kNm) Direction (kNm) Direction Direction
L/10000 6.22 Negative 3.86 2.62 Negative

Negative
L/5000 6.35 3.93 2.68
L/2000 7.65 4.29 4.45
Positive Positive
L/1000 7.19 5.39 Positive 4.42
L/500 6.84 5.29 4.37
PI - Positive Beam D – 115 mm L – 15 mm Grade 450
Imperfection Dimensions B – 45 mm T – 1.9 mm
M u (kNm) .
L ‐ 1.5 m 9
Failure in Negative Failure in Positive
L ‐ 2.0 m Direction
8 Direction
L ‐ 2.5 m
6 Failure in Positive
Failure in Negative Direction
Direction
5
Failure in Negative Failure in Positive
2 Direction Direction
0
‐L/500 ‐L/1000 ‐L/2000 ‐L/5000 ‐L/10000 0 L/10000 L/5000 L/2000 L/1000 L/500
‐5 ‐4 ‐3 ‐2 ‐1 0 1 2 3 4 5
Imperfection
Figure 5.12: Effect of Imperfection Magnitude
Effect of the direction of initial geometric imperfection was studied by carrying out
non-linear finite element analysis on a G450 cold-formed steel lipped channel beam
whose dimensions are web depth (110 mm), flange (40 mm) and lip (8 mm). The
magnitude of both negative and positive initial geometric imperfections used was
L/1000. In Figure 5.13, the ultimate moment capacities of beams with negative and
positive global imperfections are compared. It can be seen that the ultimate moment
capacities with negative imperfections are significantly lower compared to those
with positive imperfections. In other words the beams which failed in the negative
direction have lower moment capacities compared to those which failed in the
positive direction. Pi et al. (1997, 1998) and Kurniawan and Mahendran (2009a,b)
also made similar observations regarding the direction of initial geometric
imperfection and the failure direction in their numerical studies while the
experimental studies of Pi et al. (1999b), Poutre et al. (2000) and Ungermann and
Kalameya (2006) also confirmed this finding. From Figure 5.13 it can be seen that
for slender beams elastic buckling moments are slightly higher than the moment
capacities obtained with negative imperfections while the moment capacities with
positive imperfections have increased well beyond the elastic buckling moments.
This phenomenon has also been observed by Pi et al. (1997) in their numerical study
on lateral distortional buckling of cold-formed lipped channel beams.
8
Negative Imperfection
7 Positive Imperfection
Elastic Buckling Moment
6
Mu (kNm) 4
0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
Length (mm)
Figure 5.13: Effects of Negative and Positive Geometric Imperfections on

the Ultimate Moment Capacity of Lipped Channel Beams
The reasons for the higher moment capacities with positive imperfections compared
to negative imperfections are discussed next. When a cold-formed steel lipped
channel beam, subjected to a uniform moment (M) about x axis (major axis)
displaces laterally, the beam is subjected to component bending actions about both x
(Mx) and y axes (My) and torsion actions (Figure 5.14). Beams with negative
imperfections failed in the negative direction and have undergone a lip failure. In the
case of positive imperfections they failed in the positive direction and have
undergone web-flange failure.
My My
y
M
x Mx
Mx
Before failure Negative twist rotations Positive twist rotations
Figure 5.14: Failure Directions
When the beam failure is in the negative direction, compression flange lip of the
beam is subjected to compressive stresses due to the component moment about x
axis, component moment about y axis and warping effect. This resulted in
developing first yield in the compression flange lip. It is obvious that the lip is not
strong in compression and that the development of additional compressive stresses
makes the beam fail at a lower moment capacity. Figure 5.15 (a) shows the von-
Mises yield stress distribution along the beam, and it could be seen that the
compression flange-lip junction is subjected to higher stresses.
Compression flange
(a) Beam failed in the negative direction (Negative imperfection)
Figure 5.15: Effect of Imperfection Direction
Compression flange
(b) Beam failed in the positive direction (Positive imperfection)
Figure 5.15: Effect of Imperfection Direction
When the beam failure is in the positive direction, the web-compression flange
junction is subjected to compressive stresses due to the component moments about
x-axis and y-axis and warping effect. Therefore the first yield occurs at this junction.
In this case the lip is under compression due to Mx and under tension due to warping
effect and My. Therefore the effect of compressive stresses developed in the lip due
to Mx is reduced due to the tensile stresses developed due to My and the warping
effect. Failure at the flange-web junction does not make the beam weaker compared
to lip failure. Therefore beams with positive imperfections fail at higher moments
and have higher post-buckling capacities. Figure 5.15 (b) shows the von-Mises stress
distribution along the beam at failure in the case of positive imperfection
From Figure 5.13, it can be seen that as the beam span increases the difference
between moment capacities of beams with negative and positive imperfections
increases. When the beam span increases both lateral displacement and twist
rotations also increase. This causes the component of moment about y axis (My) to be
high. In this case the effect of My increases as the beam span increases. This explains
why the difference of moment capacities with negative and positive imperfections
increases as the beam span increases.
5.1.9 Residual Stresses
The presence of residual stresses, originated as a result of cold working process, may
cause structural members to fail prematurely. Therefore it is important to use
appropriate residual stresses in finite element analysis in order to obtain accurate
results. In cold-formed steel members, membrane residual stresses are small
compared to flexural residual stresses and therefore can be neglected. Flexural
residual stresses distribute linearly across the thickness with compression on the
inside surface and tension on the outside surface of the section. Ranawaka (2006)
proposed a residual stress distribution for lipped channel sections based on Schafer
and Pekoz (1998), and this was adopted in this research (see Figure 5.16).
0.17fy
outside +σr
0.08fy
Residual stress
through thickness
0.17fy
inside -σr
0.08fy
0.17fy
Figure 5.16: Flexural Residual Stress Distribution
Residual stresses are affected by the temperature of the steel member, i.e. they
decrease with increasing temperature. Therefore the effect of elevated temperature
on the residual stresses must be considered. Lee (2004) presented a linear equation to
determine the reduction factors (RT R20 ) of residual stresses at elevated temperature
as given in Equation 5.10.
RT
= 1.0181 − 0.00128T (5.10)
R20
Where RT and R20 are the residual stresses at elevated and ambient
temperatures, respectively, and T is the temperature.
Flexural residual stresses shown in Figure 5.16 were included in the finite element
model using *INITIAL CONDITION option in ABAQUS with TYPE=STRESS,
USER. The initial stresses were created in the model using SIGINI Fortran
subroutine which defines the local components of the initial stress as a function of
the global co-ordinates. To include the variation of flexural residual stresses through
the thickness, they were applied as a function of the integration point numbers
through the thickness. An example of the coding of the subroutine defining residual
stresses is given in Appendix B. The flexural residual stresses are introduced to the
finite element model as a *STATIC step with no loading and the standard model
boundary conditions to allow equilibration of the initial stress field before starting
response history. Figure 5.17 shows the distribution of residual stresses along the
half span of the beam used in finite element analysis.
Figure 5.17: Residual Stress Contours for Lipped Channel Beams
Although the flexural residual stresses in a member are in self equilibrium, there
were some deformations following the application of residual stresses. These
deformations were negligibly small and therefore their effect on the ultimate moment
capacities was neglected.
5.1.9.1 Effect of Residual Stresses
The presence of residual stresses in a beam can affect its ultimate moment capacity.
Hence the effect of residual stresses was studied at ambient temperature by
performing non-linear analysis with and without residual stresses on a G450 cold-
formed steel lipped channel beam whose dimensions are web depth (110 mm),
flange width (40 mm) and lip (8 mm). Figure 5.18 compares the ultimate moment
capacities of beams with and without residual stresses. In these analyses a negative
geometric imperfection was included with a magnitude of L/1000.
Figure 5.18 clearly shows that the ultimate moment capacities obtained with and
without residual stresses are almost the same. According to Lee (2004), residual
stresses diminish with increasing temperature. Therefore the effect of residual
stresses on the ultimate moment capacity can also be considered as negligible at
elevated temperatures. According to Figure 5.18, short beams with residual stresses
fail with slightly lower ultimate moment capacities while longer beams fail with
slightly higher moment capacities compared to beams without residual stresses.
9.0
8.0
7.0
With Residual Stresses
6.0 Without Residual Stresses
5.0
Mu (kNm)
4.0
3.0
2.0
1.0
0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
Length (m)
Figure 5.18: Effect of Residual Stresses on the Ultimate Moment

Capacity of Lipped Channel Beams
5.1.10 Pre and Post Processing
Two types of analysis were employed by using the developed finite element model:
elastic buckling and nonlinear static analyses. Elastic buckling analysis gives the
elastic buckling loads and corresponding buckling modes. The lowest eigen vector
obtained from elastic buckling analysis was used to input the geometric
imperfections in the nonlinear analysis. Non-linear analysis gave the ultimate failure
moments and simulated the behaviour of lipped channel beams until failure.
The physical model of a simply supported cold-formed steel lipped channel beam
subjected to a uniform bending moment distribution was created, and input files for
bifurcation buckling and non-linear static analysis were produced using MSC/Patran
pre-processor facility. These input files were submitted to ABAQUS and the results
were obtained. MSC/Patran was also used as a post processor to read the results files
obtained from finite element analysis carried out with ABAQUS.
ABAQUS uses the Newton-Raphson method to solve non-linear problems which is

an incremental-iterative procedure. The RIKS method in ABAQUS was included in
the non-linear analysis. This method is generally used to predict unstable or
geometrically non-linear collapse of a structure. This method uses the load
magnitude as an additional unknown and solves simultaneously for loads and
displacements. ABAQUS uses “arc length” along the static equilibrium path in load-
displacement space. This approach provides solutions regardless of whether the
response is stable or unstable.
In the non-linear analysis the maximum number of load increments was used as 100
with an initial increment size of 0.001 and the automatic incrementation control was
enabled.
1. Create geometric surfaces for web, flanges and lips.

2. Mesh all the web, flange and lip surfaces.
3. Define material properties and assign material properties for elements
4. Create nodal load field and apply the load at ends.
5. Apply boundary conditions for end and mid-planes.
6. Define bifurcation buckling parameters and generate input file for bifurcation
buckling analysis.
7. Run bifurcation buckling analysis using ABAQUS (Version 6.8) to obtain the
first Eigen mode.
8. Define non-linear static analysis parameters.
9. Enter initial geometric imperfection to the input file using the first Eigen modes
from the bifurcation buckling analysis.
10. Generate input file for the non-linear static anlaysis.
11. Run non-linear static analysis using input file along with residual stress input
subroutine.
12. Import nonlinear static analysis results into Patran and view the results using
Patran post-processing facilities.
13. The data required for the load-deflection plot were imported from Patran to an
Excel worksheet.
5.2 Validation of Ideal Finite Element Models
An extensive parametric study using the developed ideal finite element models is
proposed in order to develop a suitable design curve for the lateral-torsional buckling
of cold-formed steel beams subjected to fire conditions. The accuracy of the results
depends on how accurately the ideal model simulates the flexural behaviour of cold-
formed steel beams at elevated temperatures. Therefore the validation of the ideal
beam models is a vital part of this research.
The validation was intended to verify the accuracy of the developed finite element
models in terms of finite element type, mesh density, load and boundary conditions
and the material model. Validation process of ideal model involves two series of
comparisons. First one is to validate using the elastic buckling moments obtained
from the bifurcation buckling analysis. Second involves the validation of the ideal
model using the non-linear analysis results.
5.2.1 Elastic Buckling Analysis
5.2.1.1 Ideal Model for Member Moment Capacity (Lateral-torsional Buckling)
A series of elastic buckling analysis using the ideal model for lateral-torsional
buckling was conducted for different lengths of simply supported cold-formed steel
beams at ambient temperature and at 500oC in order to determine the critical elastic
buckling moments and the corresponding buckling modes. Beam lengths were varied
from intermediate to long spans. Four cold-formed steel sections were selected for
this investigation: G450-110-40-12-1.9, G450-90-35-10-1.5, G250-150-60-12-1.95
and G250-120-50-12-1.55 which are classified as compact sections at ambient
temperature based on AS/NZS 4600 (SA, 2005). The elastic buckling moments
obtained from bifurcation buckling analyses are presented in Table 5.4 (a) and (b)
together with corresponding buckling modes. Two types of buckling modes, namely
lateral distortional buckling and lateral-torsional buckling, were observed.
The elastic buckling moment results obtained from finite element analyses (FEA)
were compared with the solutions obtained from the well known finite strip analysis
program, Thin-Wall (Hancock and Papangelis, 1994) and the equation for elastic
lateral-torsional buckling moment (Mo). The equation for elastic lateral-torsional
buckling is given as Equation 2.2 in Chapter 2. The Mo equation only provides
elastic buckling moments for lateral-torsional buckling mode and therefore finite
element analysis results for lateral-distortional buckling mode cannot be compared
with that obtained from Mo equation. However, lateral-distortional buckling
moments obtained from FEA would agree well with those calculated using the Mod
equation (Pi et al., 1998). The results are presented in Table 5.4 together with the
ratios of finite element analysis results to that of Thin-wall (TW) and Mo equation. It
can be seen that the difference of finite element analysis results with that of Thin-
Wall and Mo Equation are small and are always less than 1.5% and 2.5%,
respectively. This confirms the accuracy of the selected finite element type, mesh
density, boundary conditions and the method used to apply the uniform moment
loading.
Table 5.4: Comparison of Elastic Buckling Moments of Lipped Channel
Beams (a) 20oC
Beam Span (kNm) Buckling
TW/FEA Mo-eqn/FEA
Designation (m) Mode
FEA TW Mo-Eqn
1.0 9.763 9.762 10.199 1.000 1.045 LDB
1.5 4.661 4.668 4.734 1.002 1.016 LTB
G450-110-40-12-1.9
2.0 2.790 2.797 2.814 1.003 1.009 LTB

2.5 1.906 1.912 1.918 1.003 1.006 LTB
3.0 1.414 1.422 1.425 1.006 1.008 LTB
3.5 1.114 1.121 1.122 1.006 1.007 LTB
4.0 0.915 0.921 0.921 1.006 1.007 LTB
1.0 4.134 4.148 4.301 1.003 1.040 LTB
1.5 1.987 1.991 2.013 1.002 1.013 LTB
G450-90-35-10-1.5
2.0 1.199 1.202 1.207 1.003 1.007 LTB

2.5 0.826 0.828 0.830 1.003 1.005 LTB
3.0 0.620 0.621 0.622 1.002 1.003 LTB
3.5 0.492 0.493 0.494 1.002 1.004 LTB
4.0 0.406 0.408 0.408 1.005 1.005 LTB
2.0 8.717 8.76 9.009 1.005 1.033 LTB
G250-150-60-12-1.95
2.5 5.802 5.82 5.901 1.003 1.017 LTB

3.0 4.167 4.177 4.21 1.002 1.010 LTB
3.5 3.165 3.172 3.187 1.002 1.007 LTB
4.0 2.507 2.513 2.521 1.002 1.006 LTB
1.5 6.380 6.442 6.768 1.010 1.061 LDB
G250-120-50-12-1.55
2.0 3.808 3.823 3.899 1.004 1.024 LTB

2.5 2.538 2.544 2.569 1.002 1.012 LTB
3.0 1.830 1.834 1.844 1.002 1.008 LTB
3.5 1.398 1.401 1.406 1.002 1.006 LTB
4.0 1.115 1.117 1.119 1.002 1.004 LTB
Table 5.4: Comparison of Elastic Buckling Moments of Lipped Channel
Beams (b) 500oC
Beam Span (kNm) Buckling
TW/FEA Mo-Eqn/FEA
Designation (m) Mode
FEA TW Mo-Eqn
1.0 4.346 4.343 4.540 0.999 1.045 LDB
1.5 2.075 2.077 2.107 1.001 1.015 LTB

G450-110-40-12-1.9
2.0 1.242 1.244 1.252 1.002 1.008 LTB
2.5 0.848 0.851 0.854 1.004 1.007 LTB
3.0 0.631 0.633 0.634 1.003 1.005 LTB
3.5 0.497 0.499 0.500 1.004 1.006 LTB
4.0 0.408 0.410 0.410 1.005 1.005 LTB
1.0 1.840 1.845 1.914 1.003 1.040 LDB
1.5 0.884 0.886 0.896 1.002 1.014 LTB

G450-90-35-10-1.5
2.0 0.534 0.535 0.537 1.002 1.006 LTB
2.5 0.368 0.369 0.370 1.003 1.005 LTB
3.0 0.276 0.277 0.277 1.004 1.004 LTB
3.5 0.219 0.220 0.220 1.005 1.005 LTB
4.0 0.181 0.181 0.182 1.000 1.006 LTB
2.0 3.880 3.899 4.010 1.005 1.034 LDB

G250-150-60-12-1.95
2.5 2.583 2.590 2.627 1.003 1.017 LTB
3.0 1.855 1.859 1.874 1.002 1.010 LTB
3.5 1.409 1.412 1.419 1.002 1.007 LTB
4.0 1.116 1.119 1.122 1.003 1.005 LTB
1.5 2.881 2.867 3.012 0.995 1.045 LDB

G250-120-50-12-1.55
2.0 1.720 1.702 1.735 0.990 1.009 LTB
2.5 1.146 1.132 1.143 0.988 0.997 LTB
3.0 0.828 0.816 0.821 0.986 0.992 LTB
3.5 0.631 0.624 0.626 0.989 0.992 LTB
4.0 0.503 0.497 0.498 0.988 0.990 LTB
5.2.1.2 Ideal Model for Section Moment Capacity (Local Buckling)
Elastic buckling analyses were conducted using the ideal model for section moment
capacity at ambient temperature and at 500oC in order to determine the critical
elastic local buckling moments. Thin-wall computer program gives the elastic
buckling moments for simply supported beams. In this study the section of beam
subjected to local buckling failure is continuous and therefore cannot be considered
as simply supported, although the beam ends are simply supported. The beam span
was selected as three half wave lengths as determined based on Thin-Wall computer
program. Three half wave lengths were selected to avoid the end effect. In some
cases the range of half wave lengths that the local buckling failure could happen was
less than three half wave lengths. In such cases, only one half wave length was used
as the beam span. One half wave span beam does not give accurate elastic buckling
moments because of the end effects. Therefore suitable sections were selected so that
the range of half wave length that dominates local buckling failure is at least three
half wave lengths. However, there is no end effect in the ultimate moment results
from nonlinear analyses.
Four cold-formed steel sections were selected for this investigation: G250-150-60-
12-1.95 and G250-120-50-12-1.55 which are classified as compact and G450-120-
80-15-1.9, G450-180-50-15-1.5 which are classified as slender at ambient
temperature based on AS/NZS4600 (SA, 2005). The elastic local buckling moments
obtained from bifurcation analyses are presented in Table 5.5 together with those
obtained from Thin-Wall. It can be seen that the differences between the finite
element analysis and Thin-Wall results are insignificant in most cases and are less
than 3%. This confirms the accuracy of the finite element model for section moment
capacity in terms of element type and size, material model and the boundary
conditions.
Table 5.5: Comparison of Elastic Local Buckling Moments of Lipped
Channel Beams
Half Elastic Local

Temp. wave Buckling TW
Beam Designation Compactness Moment (kNm)
(oC) Length FEA
(m) FEA TW
20 80 C 20.983 21.070 1.004
G450-110-40-12-1.9
500 80 C 9.344 9.388 1.005
20 65 C 11.738 11.384 0.970
G450-90-35-10-1.5
500 65 C 5.224 5.075 0.971
20 76 S 14.287 12.850 0.899
G250-150-60-12-1.95
500 76 S 6.359 5.721 0.900
20 100 S 10.205 10.205 1.000
G250-120-50-12-1.55
500 100 S 4.429 4.542 1.026
Note: C – Compact Section, S – Slender Section, TW – Thin-Wall
5.2.2 Non-linear Analysis
5.2.2.1 Using Experimental Results of this Research
Ideal finite element model for member moment capacity simulates a cold-formed
steel beam subjected to a uniform bending moment distribution over the span. In
order to validate the ideal model, lateral buckling tests of beams have to be
undertaken under a uniform bending moment. However, it is impossible to test a
simply supported beam subjected to a uniform bending moment condition. Therefore
it was decided to conduct beam tests with simply supported boundary conditions and
quarter point loading conditions. This type of loading creates a uniform bending
moment distribution between the two loading points while a linear bending moment
distribution is created between the loading point and the support. The experimental
finite element model was developed by simulating the experimental conditions as
closely as possible. The differences between the experimental and ideal finite
element models are due to the loading and support conditions and the direction of
initial imperfection. The element types, element size, the material model and the
modeling of residual stresses were all the same. The objective was to validate the
experimental model using experimental results and this comparison can indirectly
establish the validity of finite element mesh density, type of shell element and the
material model in the ideal finite element model for lateral-torsional buckling.
The accuracy of the experimental finite element model was first checked by
comparing the ultimate moment capacities with the corresponding results from the
experimental study reported in Chapter 4 (Table 5.6). It can be seen that the
experimental results are considerably higher than the finite element analysis results.
The percentage difference of ultimate moment capacities from experimental study
and the finite element analysis varies from 10.5% for 1.5 m long G450 beam to 47%
for 2.9 m long G250 beam. It can be seen that finite element model predicts the
ultimate moment capacities for 1.5 m long beams of both grades reasonably well
with 10.5% and 13% difference. The overall mean ratio of the ultimate moment
capacities from experiments and finite element analyses is 1.27 with a COV of 0.09.
The individual mean and COVs are 1.25 and 0.08, and 1.29 and 0.10 for G450 and
G250 steel beams, respectively.
Table 5.6: Comparison of Experimental and Finite Element Analysis

Results
EXP FEA
Grade Length (mm) EXP/FEA
(kNm) (kNm)
1500 3.686 3.258 1.1315
2000 3.550 2.940 1.2075
G250
2500 3.266 2.406 1.3571
2900 2.969 2.016 1.4730
1500 5.132 4.643 1.1052
2000 4.443 3.615 1.2289
G450
2500 4.006 3.063 1.3082
2900 3.408 2.472 1.3783
Mean 1.274
COV 0.094
Possible reasons for the differences between the results are described next. The
lateral buckling test set-up used in this research was developed for testing of cold-
formed steel LSBs with higher moment capacities. The capacity of load cells used in
the failure load measurement was 60 kN. This is very high compared to the failure
loads of the channel beams, which varied from 4 kN to 14 kN. Since most of our
measured loads were comparatively small compared to the load cell capacity, an
error might have been involved. Most importantly, the loading arms connected to the
test beams could have introduced lateral restraints and thus increased the ultimate
failure loads. This effect is considerable when the ultimate failure loads are low, ie.
when the beam span is increasing.
All the tested beams had a negative geometric imperfection but failed in the positive
direction. Beams having negative imperfections always displace laterally in the
negative direction as shown in Section 5.1.8.1. Therefore the two loading arms might
have introduced some restraints preventing the failure of the test beam in the
negative direction and causing the failure to occur in the positive direction. As a
result of this, finite element analyses for beam tests were carried out with positive
imperfections in order to simulate the observed failure in experiments in the positive
direction. However, this model does not simulate the actual beam with negative
imperfection that failed in the positive direction. This may be another reason for the
results from finite element analysis and experiments to be different.
In the experiments, the horizontal deflection was measured at the compression

flange-web corner at mid-span while the vertical deflection was obtained at the
tension flange-web corner. The accuracy of the experimental finite element models
was then checked by comparing the moment-deflection curves obtained from
experimental finite element model and experiments. Figure 5.19 (a) to (d) show the
comparison of moment versus vertical and horizontal deflection curves obtained
from experiments and finite element models for 1.5 m and 2.0 m long G250 and
G450 steel beams.
6.0
6.0
5.0
5.0
4.0
4.0
Exp 3.0 Exp

3.0
FEA Mu (kNm)
Mu (kNm) FEA
2.0 2.0
1.0 1.0
0.0 0.0
0 10 20 30 40 0 5 10 15 20 25
Vertical Deflection (mm) Horizontal Deflection (mm)
(a) G450-1500
5.0
5.0
4.0
4.0
3.0
3.0
Mu (kNm) Mu (kNm)
Exp 2.0 Exp
2.0
FEA FEA
1.0 1.0
0.0 0.0
0 10 20 30 40 50 60 0 10 20 30 40 50 60
(b) G450-2000
4.0 4.0
3.5 3.5
3.0 3.0
2.5 2.5
2.0 2.0
Mu (kNm) Exp
1.5 Mu (kNm) 1.5
FEA
1.0 1.0
0.5 0.5
0.0 0.0
0 5 10 15 20 25 30 0 5 10 15 20 25 30
(c) G250-1500
4.0
4.0
3.5
3.5
3.0
3.0
2.5
2.5
Mu(kNm) 2.0 Exp
Mu(kNm) 2.0 Exp FEA
1.5
1.5 FEA
1.0
1.0
0.5
0.5
0.0
0.0
0 10 20 30 40
0 10 20 30 40
(d) G250-2000
Figure 5.19: Comparison of Moment-Deflection Curves
Due to the above-mentioned problems in the comparison with experimental results, it
was decided to use nonlinear finite element analysis results of a numerical study
conducted by Pi et al. (1997) for the lateral-torsional buckling of simply supported
cold-formed steel lipped channel beams and the results of an experimental study
conducted by Poutre et al. (2000) on unlipped hot-rolled channel beams undergoing
lateral-torsional buckling failures, in order to validate the ideal finite element model
for lateral-torsional buckling.
5.2.2.2 Using Pi et al’s (1997) Finite Element Analysis Results for Lateral-
torsional Buckling
Pi et al. (1997) investigated the lateral-buckling of simply supported cold-formed

steel lipped channel beams subjected to a uniform bending moment distribution.
Their focus was mainly on lateral-distortional buckling mode under ambient
temperature conditions. They developed a non-linear inelastic finite element model
to investigate the lateral-distortional buckling strengths of cold-formed steel lipped
channel beams. In their finite element model for lateral distortional buckling, each
flange maintains its shape during the deformation while web can distort into a cubic
curve. They developed a seperate model for lateral-torsional buckling in which each
flange maintains its shape and the web cannot distort as in the lateral-distortional
buckling model. They compared the results obtained for lateral-torsional buckling
strengths (flexural-tortional buckling) and lateral distortional buckling strengths
obtained from finite element analysis in Figure 5.20. The results are presented in
Figure 5.20 as the variations of the dimensionless ultimate moment ( M M px )
against the modified slenderness, λd = [ ]

M px (α m M yzd ) for lateral-distortional
buckling and λ = [ ]
M px (α m M yz ) for lateral-torsional buckling. α m is the moment
modification factor whose values for specific bending moment distributions are
given by Trahair and Bradford (1991) and used in AS 4100 (1990). It is equal to 1.0
for the case of uniform bending moment distribution. M px is the full plastic moment
capacity of the cross-section about the major x axis given by,
M px = S pxσ y (5.11)
where S px is the plastic section modulus about the major axis ox and is given by,
π
S px = h[B − 2(r + t )]t +
1 2
4
(
d c t + [D − l − (r + t )][l − (r + t )]t + t 2 + 2rt
2
)⎡⎢d c
⎛ t ⎞⎤
+ 2 ⎜ r + ⎟⎥
2 ⎠⎦
⎣ ⎝
(5.12)
where the dimensions D, B, l , r , t , h and d c are shown in Figure 5.21.
M yz is the elastic lateral-torsional buckling moment (Hill, 1954 , Trahair, 1993)
given by Equation 2.2 in Chapter 2 and M yzd is the elastic lateral-distortional
buckling moment (Pi et al., 1997).
Figure 5.20: Comparison of Lateral-Distortional and Flexural-

Torsional Buckling Strengths (Pi et al., 1997)
13.55
100.1
13.55
49.1
(a) Rounded corner (b) Two flat corner (c) Centreline dimensions of the
channel section section cross-section without corners
Figure 5.21: Dimensions of Lipped Channel Beams Used by

Pi et al. (1997)
Their results for flexural torsional strengths of lipped channel were used to validate
the present ideal model for lateral-torsional buckling in this research. The
dimensions of the beam cross-section they used are shown in Figure 5.21. They
replaced the corners with two flats with the same thickness in their model. A trilinear
stress-strain curve was assumed for the flats while the Ramberg-Osgood stress-strain
curve was assumed for the corners as shown in Figure 5.22. They adopted the
longitudinal flexural residual stresses shown in Figure 5.23 as recommended by
Weng and Pekoz (1998). Initial crookednesses and twists are taken as − L 1000 and
− (N y L ) (1000M yz ), respectively, where L , M yz and N y are the beam span, elastic
lateral-torsional buckling moment, and the column elastic buckling load.
(a) Tri-linear (Flats) (b) Ramberg-Osgood (Corners)
Figure 5.22: Sress-Strain Curves Adopted by Pi et al. (1997)
Figure 5.23: Residual Stress Distributions used by Pi et al. (1997)
The same beam was modelled using the developed ideal model for lateral-torsional
buckling described in Section 5.4 with their proposed material properties, initial
imperfections and initial residual stresses. The corners were neglected when
modeling this section, and the centerline dimensions of the cross-section used here
are shown in Figure 5.21 (c). Since the corners were neglected in this model the
Ramberg-Osgood stress-strain curve (Figure 5.22 (a)) for flats was used for the
whole beam. In addition the residual stress distribution was assumed to be similar
over the cross-section and varies through the thickness as given in Figure 5.23(b).
Initial imperfection was taken as negative with a magnitude of L 1000 and was
included in the finite element model as described in Section 5.1.8.
Pi et al.’s (1997) results for lateral-torsional buckling are reproduced in Figure 5.24
and are compared with the non-linear analysis results using the developed ideal
model for lateral-torsional buckling. Figure 5.24 shows that the results agree well in
the non-dimensional format. Therefore the ideal finite element model developed for
lateral-torsional buckling in this research can be considered accurate in terms of
mesh density, element type and size, boundary conditions, method of applying initial
geometric imperfections and residual stresses.
1.0
0.9
0.8
0.7 Pi et al. (1997)

FEA
0.6
0.5
(Mb/MPX) 0.4
0.3
Lateral-torsional buckling
failure
0.2
0.1
0.0
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0
(MPX/MO)1/2
Figure 5.24: Comparison of Nonlinear Finite Element Analysis Results

from the Developed Ideal Model with those of Pi et al. (1997)
5.2.2.3 Using Poutre et al.’s (2000) Experimental Results for Lateral-torsional

Buckling
In order to further validate the finite element model for non-linear analyses of cold-
formed steel lipped channel beams subject to lateral-torsional buckling, Poutre et
al.’s (2000) experimental results for the lateral-torsional buckling of hot-rolled
channel beams at ambient temperature were used. Poutre et al.’s (2000) experimental
study was followed by a numerical study (Snijder et al. 2002) of the same beams.
Test program was carried out on simply supported beams with quarter point loading.
Load was applied at the shear centre and also eccentrically along the line of the web.
In the case of eccentric loading the point of loading was also varied from top to
bottom height of the web. In this validation, only the results for the case of shear
centre loading were used. The dimensions of channel used (UPE160) in their
experiments are shown in Figure 5.25. The steel grade used was S235 with a
measured yield strength of 296 MPa and the material model used with the elastic
modulus value is shown in Figure 5.26 based on Snijder et al.’s (2002) finite element
modelling. Tests of beams with 2.8 m span and shear centre loading were repeated
four times by them, and their results are given in Table 5.7. In the testing two plates
were welded to the beam at quarter points to facilitate the load application at the
shear centre. This caused the test beam to deform during the cooling phase due to
shrinking of weld, and Figure 5.27 shows the deformed shape of the beam after
welding. The channel can be considered to be straight between the welds and also
between the welds and supports according to Snijder et al. (2002). The measured
initial geometric imperfections relative to the supports caused by welding are given
in Table 5.7. All the initial imperfections in the tested beams were on the positive
side (Negative side according to their sign convention). The sign convention used by
them is opposite to what is used in this research. Therefore the failure and
imperfection directions are marked in figures and tables based on the sign
convention used in this research. Beam designated by 1B1 has failed in the positive
direction (+), ie. in the same direction as the initial geometric imperfection, while all
the other beams failed in the negative direction (-), ie., in the opposite direction to
the initial geometric imperfection (see Table 5.7). They observed higher moment
capacities when the beam failed in the positive direction compared to those failed in
the negative direction.
Figure 5.25: Dimensions of Hot-rolled Channel Beam Used by Poutre et

al. (2000) and Snijder et al. (2002)
Figure 5.26: Sress-Strain Curves Used by Snijder et al. (2002)
Table 5.7: Magnitude and Direction of Initial Geometric Imperfections
and Experimental Results
Initial Failure
Initial Imperfection Snijder et al. (2002)
Exp. Imperfection Direction
Average Direction Mo Mu,FEA Mu,EXP
e1 e2
(eav) (kN) (kN) (kN)
1B1 3.19 3.19 3.190 + 50.85 55.64 +
1B2 3.47 2.54 3.005 + 50.98 49.8 -
66.7
1b3 3.49 3.14 3.315 + 50.76 50.02 -
1B4 1.98 2.09 2.035 + 51.69 48.37 -
+ Positive Direction
- Negative Direction
Figure 5.27: Shape of Initial Imperfection in Test Beams
Snijder et al. (2002) developed finite element models to simulate the tested beams
using 8-node quadratic shell elements (Figure 5.28a). The root radius at the flange-
web junction was ignored and this might have underestimated the torsional rigidity
of the section considerably. An additional beam element with sufficient torsional
rigidity was placed in the corners to compensate for this. They also used additional
stiffening elements to simulate the support reaction (Figure 5.28a).
Rigid plate representing
Simply supported boundary stiffening element at the
condition applied to the shear support
centre
Load application at shear

centre
Rigid plate representing

welded plate
(a) FEA Model used by Snijder et al. (2002) (b) FEA Model used in this Research
Figure 5.28: Finite Element Models used by Snijder et al. (2002) and
Developed Model in this Research
All four test beams of Poutre et al. (2000) were modelled using the experimental
model described in Section 5.1.6 with the mechanical properties shown in Figure
5.26 and the initial average measured imperfection given in Table 5.7. An
approximation was made relating to the imperfection magnitude by taking the
average of the two values at quarter points and applying it at mid-span. Two rigid
bodies were used for the loading plate at quarter points on to which the load was
transferred and at the support plane. The corners were neglected when modeling the
section, and the centerline dimensions of the cross-section were used. Residual
stresses were not considered here as for Snijder et al. (2002). Non-linear analyses of
test beams with both positive and negative initial imperfections were undertaken.
According to the finite element analysis results, beams with negative imperfections
failed in the negative diretion while those with positive imperfections failed in the
positive direction. Snijder et al. (2002) used positive imperfections in their finite
element models. Table 5.8 presents the finite element analysis results with Snijder et
al.’s (2002) experimental and numerical results. When the observed positve
geometric imperfections were used, finite element analysis results agree well with
experimental results with less than 2% difference. However, when negative
geometric imperfections were used the difference in results was 12% on average.
Table 5.8: Comparison of Finite Element Analysis Results with
Experimental Results from Snijder et al. (2002)
Snijder et al. (2002) This Research

Negative
Exp. Positive Imperfection
Mo,FEA
Mu,FEA
Mu,EXP
Mo,FEA
Imperfection
Mu,FEA Mu,EXP/Mu,FEA Mu,FEA Mu,EXP/Mu,FEA
50.85 55.64 49.40 43.70
1B1 1.13 1.27
(+) (+) (+) (-)
50.98 49.80 49.60 44.10
1B2 1.00 1.13
(+) (-) (+) (-)
66.7 66.35
50.76 50.02 49.30 43.50
1B3 1.01 1.15
(+) (-) (+) (-)
51.69 48.37 50.70 45.90
1B4 0.95 1.05
(+) (-) (+) (-)
Average 1.02 1.12
Note: Failure direction is indicated within brackets.
Since the failure direction has a signifcance influence on the capacity, comparison
was also made by considering the same failure direction. When the experimental
capacity of beams that failed in the positive direction is compared with FEA results
for the positive failure direction, there is about 13% difference while it was on
average 11% when the experimental capacity of beams that failed in the negative
side are compared with FEA results for the negative direction. The difference
between FEA and experimental results can be attributed to the strength reduction due
to the neglection of corners, which results in the underestimation of the torsional
rigidity of beams. However, it can be seen from Table 5.8 that Snijder et al.’s (2002)
FEA results with positive imperfections are very closer to the FEA results of this
research with positive imperfections. Further FEA results for elastic buckling
moments are also very closer to each other. The load versus deflection graphs were
obtained at the point where deflections were measured in the experiment, ie. at the
middle of the web element, 850 mm away from the suppport. The load-deflection
graphs obtained from FEA with positive geometric imperfections are given in Figure
5.29 together with experimental graphs obtained by Poutre et al. (2000) for beam
1B4. It can be seen that the vertical and lateral deflections are accurately simulated
by the finite element model developed here.
Lateral- Lateral-
deflection to deflection to
the positive the negative
side side
(a) Results of (Snijder et al. 2002)

60
.
50
40
30
20
Load (kN)
10
0
0 2 4 6 8 10 12 14 16 18 20 22 24
Vertical Deflection, mm
Vertical Deflection
.
60
50
40
Lateral-deflection to the positive side
30
Load (kN)
20
10
0
0 4 8 12 16 20
Lateral Deflection, mm
Lateral Deflection
(b) Results of This Research

Figure 5.29: Load versus Deflection Plots
The finite element model used here has the same element type, mesh density and
boundary conditions used in the ideal finite element model. Therefore it can be
concluded that the ideal finite element model developed for member moment
capacities accuratly simulates the lateral-torsional buckling behaviour in terms of
element type and size, mesh density, boundary conditions and also the method of
applying boundary conditions.
5.3 Conclusion
This chapter has described the finite element models developed for the investigation
of section and member moment capacities of cold-formed steel lipped channel beams
subject to local buckling and yielding, and lateral-torsional buckling. Two ideal
finite element models were developed; one for lateral-torsional buckling and the
other for local buckling and yielding. A separate finite element model was developed
to simulate the experimental tests for the lateral-torsional buckling of lipped channel
beams in order to verify the accuracy of the ideal model developed for lateral-
torsional buckling. All the models were based on half-span beam due to the
symmetry in loading and geometry, and included the following features: S4R5
elements, isotropic stress hardening material model, residual stresses, geometric
imperfections, and reduced mechanical properties at elevated temperatures.
The ideal model for member moment capacity (lateral-torsional buckling)

incorporates idealized simply supported boundary conditions and a uniform moment
loading while the model for section moment capacity (local buckling and yielding)
incorporates four-point loading with simply supported boundary conditions and
laterally and torsionally restrained end segments. The ideal models for section and
member moment capacities were validated against the elastic buckling results
obtained from an established finite strip analysis program, Thin-Wall, both at
ambient and elevated temperatures. A good comparison of elastic lateral-torsional
buckling moment results obtained from finite element analysis and Thin-wall
indicated that these models could accurately predict elastic buckling moments at
both ambient and elevated temperatures.
The main ideal finite element model was validated for non-linear analysis using
experimental results and the results of Pi et al. (1997) for lateral-torsional buckling.
The comparison of nonlinear analysis results with Pi et al.’ (1997) results showed
that the ideal model is able to predict the lateral-torsional buckling capacities
accurately. However, there was significant difference between the experimental
results of this study and the non-linear analysis results obtained from the
experimental finite element model, which were identified as due to the shortcomings
in the experiments. The ideal finite element model was also validated for non-linear
analysis based on Poutre et al.’s (2000) experimental results of lateral-torsional
buckling of hot-rolled steel unlipped channel beam at ambient temperature. Although
the developed experimental finite element model was validated only for the lateral-
torsional buckling of cold-formed steel lipped channel beams at ambient
temperature, it could also be used to predict their lateral-torsional buckling behaviour
at elevated temperatures accurately.
6 Behaviour and Design of Cold-formed Steel
Beams Subject to Local Buckling and
Lateral-torsional Buckling at Ambient
Temperature
Past research on lateral-torsional buckling of steel beams has been limited to hot-
rolled steel beams despite the increased usage of cold-formed steel beams. Cold-
formed steel beams are commonly made of thinner steels and their cross-sections are
often mono-symmetric or unsymmetric. Hence their lateral-torsional buckling
behaviour is more complicated than that of doubly-symmetric hot-rolled beams. A
detailed parametric study is needed to fully understand the lateral-torsional buckling
behaviour and to evaluate and develop suitable design guidelines that can accurately
represent the structural behaviour of cold-formed steel beams. In recent times finite
element analysis has been increasingly used in such parametric studies by many
researchers. In this research on cold-formed steel beams, a detailed parametric study
was undertaken based on finite element analyses to investigate the lateral-torsional
behaviour of compact lipped channel beams as well as their local buckling
behaviour. Two types of idealised finite element models were developed for this
purpose. The first model was developed to simulate the behaviour of short beams or
fully laterally restrained beams subject to yielding or local buckling while the second
model was developed to simulate the lateral-torsional buckling behaviour of
unrestrained beams. Chapter 5 has presented the details of these two ideal finite
element models and their validation.
Beams subject to uniform bending represent the worst case in lateral-torsional

buckling and are therefore used in developing their member capacity design rules.
The validated ideal finite element model of simply supported cold-formed steel
beams subject to lateral-torsional buckling under uniform bending was used in this
parametric study to obtain the ultimate moment capacities of cold-formed steel
lipped channel beams by varying parameters such as steel grade and thickness,
section geometry and beam span.
The main aim of this research is to verify the applicability of current design rules for
the fire design of cold-formed steel beams based on available elevated temperature
design guidelines, and also ambient temperature design guidelines but using
appropriate mechanical properties at elevated temperatures. As the first step, the
accuracy of currently available ambient temperature design guidelines (AS/NZS
4600 (SA, 2005), NAS (AISI, 2001), Direct Strength Method (DSM), EC 3 Part 1.3
(ECS, 2006) and BS5950 Part 5 (BS, 1998)) was investigated based on the
parametric study of cold-formed steel beams at ambient temperature. Since the
design rules in AS/NZS 4600 (SA, 2005) and NAS (AISI, 2001) are identical, this
chapter makes reference to only one of them (AS/NZS 4600) in the discussions.
This chapter presents the results of the parametric study carried out on the lateral-
torsional buckling behaviour of cold-formed steel lipped channel beams at ambient
temperature and the development of new design rules. Also presented in this chapter
is the behaviour of short lipped channel beams subject to inelastic local buckling and
yielding and their section moment capacities. The results of the parametric study
carried out at elevated temperatures are discussed in the next chapter.
6.1 Details of Finite Element Model Used in the Parametric Study
The ideal model for short span beams was modelled with four point loading and
simply supported boundary conditions at the ends. The segment between the support
and the end of the loading plate was modelled as non-deformable part with restrained
lateral and torsional deformations, and were modelled with elastic material
properties. The central segment of the beam between the ends of two loading plates
was under uniform bending and most vulnerable to failure. Due to the presence of
symmetry in loading, support and geometry, half the beam length was considered in
modelling. A finer mesh with 5 mm x 5 mm S4R5 type elements was used in the
model. The loading plate was modelled as a rigid body of 100 mm in length and web
height with R3D4 elements. The shear centre was used as the rigid body reference
node and the load applied at this point was transferred to the beam through two rigid
plates. The end segments were modelled by restraining degrees of freedoms “1,5,6”
at all the nodes except the support end nodes to prevent lateral and torsional
deformations while degrees of freedom “1,2,5,6” of all the end nodes were
restrained. Symmetric conditions were simulated at mid-span using “3,4,5” boundary
conditions.
A simply supported cold-formed steel lipped channel beam under uniform bending
was used in order to investigate the lateral-torsional buckling behaviour at ambient
temperature. Due to the presence of symmetric conditions in loading, support and
geometry of the beams, only half the span was modelled in order to reduce
processing time and memory requirements. Four noded S4R5 type shell elements of
5 mm x 10 mm were used in the elastic buckling and non-linear analyses of lipped
channel beams. A series of tensile and compressive forces was applied to the nodes
at one end creating a triangular distribution of forces across the section and thus a
uniform bending moment distribution throughout the span length of the beam. The
beam was modelled with idealized simply supported boundary conditions at the
support end which allows major and minor axis rotations and warping displacement
while preventing in-plane and out-of-plane translations and twisting. It was achieved
by providing “126” boundary condition to all the nodes at the support end of the
beam. The symmetric plane was simulated by applying boundary condition “345” to
all the nodes at the other end of the model.
The measured yield strength and modulus of elasticity values for both finite element
models were taken as ambient temperature mechanical properties and the stress-
strain graphs were based on the developed equations in Chapter 3. A strain hardening
material model was used for G450 steels with gradual yielding type stress-strain
curve while an elastic-perfect plastic material model was used for G250 steels that
have a stress-strain relationship with a well defined yield point.
The initial geometric imperfections and residual stresses were included in the non-
linear analysis. The corresponding buckling mode, local buckling for short span
beams and lateral-torsional buckling mode for long span beams obtained from the
elastic buckling analyses, was used to input the initial geometric imperfection in the
non-linear analysis. The initial geometric imperfection for local buckling was taken
as equal to 0.006w as proposed by Schafer and Pekoz (1998). Care was taken to
select the local buckling eigen mode since the lowest eigen mode did not always
relate to local buckling.
In the case of lateral-torsional buckling a maximum initial imperfection of L/1000 in

the form of a half sine wave was used. It was found that the finite element analysis
with negative imperfections yields the lowest failure moments compared to positive
imperfections. Similar observations have been made by Pi et al. (1997 and 1998) and
Kurniawan and Mahendran (2009) based on their research on other mono-symmetric
beams. Therefore negative imperfections were used in all the non-linear analyses.
Further Pi et al. (1998) stated that the negative initial crookedness and twist are most
likely to occur than positive initial imperfection and twist based on the
measurements of Kwon (1992), and therefore the use of negative imperfection and
twist in the parametric study is justified.
There are two types of residual stresses: membrane and flexural. Only flexural
residual stresses were included since membrane residual stresses are generally small
compared to flexural residual stresses in cold-formed steel members. In this research,
the flexural residual stress was assumed to be equal to 0.17fy along the flanges and
webs and 0.08fy along the lips, based on Schafer and Pekoz (1998) and Ranawaka
(2006). It was assumed to vary linearly across the thickness with compression on the
inside surface and tension on the outside surface of the section.
The finite element models were first created by using MD/PATRAN pre-processing
facilities and the analysis was then carried out by using ABAQUS. The results were
viewed by using MD/PATRAN post-processing facility. Further details of finite
element models used here in the parametric study are given in Chapter 5.
6.2 Selection of Member Sizes for the Parametric Study
Twelve cross sections of cold-formed steel beams made of 1.5 mm and 1.9 mm thick
G450 steels and 1.55 mm and 1.95 mm thick G250 steels were selected with a range
of web depth to flange width (d/b) ratios that represent the lipped channel beams
currently used in the industry. The d/b ratio of currently available lipped sections in
the industry is in the range of 2 to 3.3 while the d/b ratio of the selected sections
ranges from 1.67 to 3.75. The cross section dimensions were selected so that all the
plate elements are compact based on AS/NZS 4600 (SA, 2005) in order to eliminate
the occurrence of local buckling at ambient temperature. Table 6.1 shows the details
of the selected cross sections of lipped channel beams in this research. The
nomenclature of the lipped channel beams used is shown next.
G250-1.95-150-60-15-L2500
Span (mm)
Lip (mm)
Flange (mm)
Web (mm)
Thickness (mm)
Grade
Table 6.1: Selected Lipped Channel Beam Sections and their Dimensions
Thickness Web Flange Lip

No. Beam Designation Grade b/t d/t d/b
(mm) (mm) (mm) (mm)
B1 G-250-1.55-90-45-15 1.55 90 45 15 29.03 58.06 2.00
B2 G-250-1.55-100-40-15 250 1.55 100 40 15 25.81 64.52 2.50
B3 G-250-1.55-120-45-15 1.55 120 45 15 29.03 77.42 2.67
B4 G-250-1.95-100-40-15 1.95 100 40 15 20.51 51.28 2.50
B5 G-250-1.95-100-60-15 1.95 100 60 15 30.77 51.28 1.67
B6 G-250-1.95-125-50-15 250 1.95 125 50 15 25.64 64.10 2.50
B7 G-250-1.95-150-40-15 1.95 150 40 15 20.51 76.92 3.75
B8 G-250-1.95-150-60-17 1.95 150 60 17 30.77 76.92 2.50
B9 G-450-1.5-90-30-14 450 1.5 90 30 14 20.00 60.00 3.00
B10 G-450-1.9-80-40-15 1.9 80 40 15 21.05 42.11 2.00
B11 G-450-1.9-100-35-15 450 1.9 100 35 15 18.42 52.63 2.86
B12 G-450-1.9-115-40-15 1.9 115 40 15 21.05 60.53 2.88
The short span beam model was used to find the section moment capacities of beams
having effective spans 100 mm, 500 mm and 800 mm or 1000 mm depending on the
cross section. In total 32 finite element analyses were carried out using short span
beam models. Effective span was taken as the length of the short beam finite element
model between the two ends of the loading plates as the failure of the two end
segments was prevented.
Finite element analyses using the ideal model for lateral-torsional buckling were
carried out for the selected beams in Table 6.1 with their spans in the range of 1000
mm to 5000 mm. The beam spans were selected for each cross section so that they
fail by lateral-torsional buckling. The results of beams that failed by local,
distortional or lateral-distortional buckling were not considered here. Convergence
problems were experienced when analysing very slender beams (in some cases,
spans larger than 5 m). A total of about 100 different cases of lipped channel beams
was analysed.
6.3 Section Moment Capacity
Elastic buckling and non-linear analyses of short span beams with at least two
different beam spans were carried out for each selected beam. Elastic buckling and
ultimate failure moments (Mo and Mu) were obtained and presented in Table 6.2. The
lowest eigen mode of very short beams was always local buckling whereas for spans
of 500 mm and above the lowest eigen mode was either local or distortional
buckling. It is believed that the lowest eigen mode obtained in an elastic buckling
analysis is the most critical one. However, in the non-linear analysis, the eigen mode
corresponding to local buckling was always used to input the initial geometric
imperfections in the non-linear analysis even if it was not the lowest eigen mode.
Figures 6.1 (a) to (d) show the elastic buckling and ultimate failure modes for two
different spans. It can be seen from the elastic buckling plot that multiple local
buckling waves are present in the beam with 1000 mm span and their wavelengths
increased towards the symmetric plane (mid-span). The length of buckling wave that
developed closer to the mid-span was taken as the buckling wavelength. When
multiple buckling waves were observed in the elastic buckling analysis, the half-
wave buckling length of the wave near the centre of beam was equal to that obtained
from Thin-Wall elastic buckling analyses for local buckling. Table 6.2 gives the
relevant half-wave lengths observed in elastic local buckling analysis results.
Table 6.2: Section Moment Capacity Results from FEA

Half wave
My Span Mo Mu Mp/My
Section Length Mu/My
(kNm) (mm) (kNm) (kNm) (=Sx/Zx)
(mm)
100 50 12.035 3.130 1.082
G250-1.55-90-45-15 2.8939 1.13
500 45 10.199 2.900 1.002
800 45 10.165 2.960 1.023
100 50 12.803 3.465 1.124
G250-1.55-100-40-15 3.0814
500 50 11.769 3.275 1.063 1.15
800 50 11.901 3.210 1.042
100 90 12.629 4.520 1.080
G250-1.55-120-45-15 4.1869
500 65 11.575 4.405 1.052 1.15
800 65 11.755 4.345 1.038
100 60 23.173 4.065 1.138
G250-1.95-100-40-15 3.5735
500 55 21.347 3.920 1.097 1.15
800 50 21.573 3.775 1.056
100 90 18.943 4.990 1.078
G250-1.95-100-60-15 4.6305
500 60 14.894 4.670 1.009 1.12
1000 60 14.728 4.630 1.000
G250-1.95-125-50-15 5.4513 100 100 23.306 6.100 1.119
1.15
1000 70 21.233 5.600 1.027
G250-1.95-150-40-15 6.1196 100 100 21.842 7.000 1.144
1.18
500 80 18.914 6.700 1.095
100 100 24.338 8.450 1.097
G250-1.95-150-60-17 7.7051
500 80 20.891 8.100 1.051 1.14
1000 80 21.432 8.000 1.038
G450-1.5-90-30-14 3.9944 100 50 11.310 4.405 1.103
1.17
500 50 10.791 4.110 1.029
100 45 21.363 5.550 1.118
G450-1.9-80-40-15 4.9644
500 45 18.813 5.250 1.058 1.14
800 45 18.737 4.975 1.002
G450-1.9-100-35-15 6.1275 100 60 23.064 7.000 1.142
1.16
500 55 22.236 6.500 1.061
100 90 23.043 8.900 1.120
G450-1.9-115-40-15 7.9442
500 65 21.303 8.150 1.026 1.16
800 65 21.644 7.850 0.988*
Average
1.068 1.15
* did not consider in calculating the average.
½ x Half wave
length
Symmetric plane
(a) G250-1.95-125-50-15-L100 – Elastic Local Buckling
Fully yielded flange

and lip
Partly yielded web
(b) G250-1.95-125-50-15-L100 – Ultimate Failure Mode
Figure 6.1: Elastic Buckling and Ultimate Failure Modes
½ x Half wave
length
Symmetric plane
(c) G250-1.95-125-50-15-L1000 – Elastic Local Buckling
Fully yielded
flange and lip
Partly yielded web
(d) G250-1.95-125-50-15-L1000 –Ultimate Failure Mode
Figure 6.1: Elastic Buckling and Ultimate Failure Modes
The ultimate failures of all the beams occurred by cross section yielding and the
failure moment capacities were always above the yield moment capacity except in
one case since all the selected sections are compact according to AS/NZS 4600. The
average inelastic bending capacity of cold-formed steel lipped channel beams is
1.068 times the yield moment capacity (1.068My). Very short beams of 100 mm
effective span are capable of achieving 1.112 times the yield moment capacity on
average. Experimental studies carried out by Yu and Schafer (2003, 2006) also
confirmed the existence of inelastic reserve capacity in cold-formed steel thin-walled
beams. With increasing beam spans, lateral-deformation is always possible and
hence may be reducing the capability of achieving higher inelastic bending
capacities of compact cross-sections. In one case, the ultimate moment capacity is
slightly below the yield moment capacity, and this may be due to additional lateral
deformations.
Theoretically beams with compact cross-sections are capable of achieving full plastic
moment capacity (Mp=Sxfy) at failure. The ratios of theoretical plastic moment
capacity and the yield moment capacity (Mp/My) are presented in Table 6.2. It can be
seen that even very short beams are not able to reach the full plastic moment
capacity. The average ratio of theoretical full plastic moment capacity to the yield
moment capacity is 1.15 whereas practically achievable inelastic bending capacity
from FEA relative to the yield moment capacity is only 1.112 for short beams.
Figure 6.1 shows the von Mises yield stress distribution at failure, and it can be seen
that the flange and lip elements of the section are fully yielded while the web
element is only partly yielded. The stress distribution along the web at mid-span in
Figure 6.2 also shows that the web is partly yielded, ie. fully yielded near the two
flanges but not yielded near the neutral axis. Shifferaw and Schafer (2008) reported
that the maximum inelastic bending capacity obtained from experiments is 1.18
times the yield moment capacity in Yu and Schafer (2003) while it was 1.06 times
the yield moment capacity in Yu and Schafer (2006). The average maximum reserve
capacity of their test results is 10% as given in Shifferaw and Schafer (2008). It was
also reported that the maximum shape factor (Sx/Zx) of the C and Z beams used in
the experiments by Yu and Schafer (2003) and (2006) are 1.20 and 1.23,
respectively. According to Shifferaw and Schafer (2008) and this research, full
plastic moment capacity cannot be achieved by thin-walled cold-formed steel
members as shown by both experiments and finite element analyses. However, it is
possible that finite element analyses have underestimated the inelastic bending
capacities to some extent due to some of its limitations as reported by Greiner
(2001).
In order to investigate whether symmetric compact sections can achieve their full
plastic moment capacity, a doubly symmetric hot-rolled universal beam
(360UB56.7) was analysed. The finite element model was similar except for the
method of load application. Instead of applying the load at the shear centre and
transferring it to the beam through a rigid body in the case of mono-symmetric
beams, the load in this case was applied to all the nodes in the web in the plane of the
loading. The theoretical plastic moment capacity of 360UB56.7 without corners was
298.64 kNm while the section moment capacity predicted by FEA was 298.08 kNm.
This shows that the doubly symmetric and compact hot-rolled steel beams can reach
their full plastic moment capacities. A hot-rolled and compact parallel flange channel
beam (150PFC) was then analysed to determine whether it can achieve its full plastic
moment capacity. The theoretical plastic moment capacity without considering its
corners was 40.27 kNm while the section moment capacity obtained from FEA was
38.6 kNm, ie. the ratio of FEA section moment capacity to full plastic moment
capacity is 0.959. Therefore it appears that FEA is not able to predict the full plastic
moment capacity of mono-symmetric steel beams (hot-rolled and cold-formed).
Experimental study of these sections is needed to verify these observations.
Effect of strain hardening on the moment capacity was investigated for one section
by changing the strain hardening model used in FEA to elastic-perfect plastic model.
It was found that the use of strain hardening model increased the moment capacity
by only 1.4% when compared with the elastic-perfect plastic model. In this research
the strain hardening model that represents the true behaviour was used when the
stress-strain curves did not exhibit a well defined yield point.
.
125
100
Section of 75
Distance (mm)
web not
yielded
50
25
0
-300 -250 -200 -150 -100 -50 0 50 100 150 200 250 300
Stress (MPa)
Figure 6.2: Von Mises Stress Distribution over the Cross-section
6.3.1 Comparison of Ultimate Capacities from Finite Element Analysis and

Current Design Rules
6.3.1.1 AS/NZS 4600 (SA, 2005)
Based on Initiation of Yielding
The section moment capacity of cold-formed steel beams is given by,

M s = Ze f y (6.1)
where Ze and fy are the effective section modulus and yield stress, respectively.
The effective section modulus used in the equation accounts for the reduction in
strength due to local buckling and is calculated based on the effective widths of
individual plate elements. All the sections considered in this study are classified as
compact based on AS/NZS4600 and therefore the effective section modulus is equal
to the full section modulus (Zf). Section moment capacity of cold-formed steel beams
is based on initiation of yielding and does not take in to account the inelastic reserve
capacity. Therefore the section moment capacity is limited to the yield moment
capacity (Zffy). However, the FEA moment capacities of all the selected sections with
very short spans are higher than the yield moment capacity and there is inelastic
reserve capacity of 11.2% on average for very short beams. Based on FEA moment
capacity results for all the selected beam spans, AS/NZS 4600 design rules
underestimate the section moment capacity by 6.8% as seen in Table 6.2.
Based on Inelastic Reserve Capacity
Finite element analyses of short cold-formed steel lipped channel beams in this
research confirmed the existence of inelastic reserve capacity. Shifferaw and Schafer
(2008) stated that beams that are symmetric about the axis of bending or have first
yield in compression are capable of achieving the inelastic reserve capacity through
their ability to sustain higher compressive strains. However, the maximum inelastic
reserve capacity that a beam can achieve is restricted based on the longitudinal
membrane strain that it can develop in the compression flange.
AS/NZS 4600 also allows the use of inelastic reserve capacity based on a limit on
the maximum compressive strain (Clause 3.3.2.3). The maximum compressive strain
is defined as, ε max = C y ε y , where C y (≤ 3) is the compression strain factor and
depends on the b/t ratio of compression elements, and ε y is the yield strain which is
equal to f y E . The section moment capacity is then calculated based on the allowed
maximum compressive strain ( ε max = C y ε y ). However, AS/NZS 4600 has another
requirement that the ratio of the depth of the compressed portion of the web to its
1.11
thickness (dw/t) does not exceed the slenderness ratio λ1 defined as . Most of
fy / E
the cross-sections chosen in this study (Table 6.1) do not meet this requirement for
web slenderness and hence the inelastic bending capacity was not calculated based
on AS/NZS 4600 design rules. Clause 3.3.2.3 appears to be quite restrictive as λ1
value for the chosen sections is less than 30. Other codes such as Eurocode 3 Part 1.3
(ECS, 2006) do not have such limits.
The calculated yield strain (εy) together with the maximum longitudinal membrane
compressive strain (εFEA) observed in the finite element analyses of 100 mm short
beams are given in Table 6.3. The value of εFEA is less than 3 x εy for only two beams
and all the other beams have developed higher strains at failure. The εFEA values
increase as the slenderness of compression plate elements decrease. This implies that
it is possible for compact lipped channel beams to achieve higher inelastic bending
capacities. The maximum longitudinal compressive strain is less than 5 times the
yield strain in most cases except in three beams (B1, B5, B6 and B10). Shifferaw
and Schafer (2008) has observed the average membrane strain to reach as high as 14
times the yield strain for local buckling. In this research the maximum strain
developed was 8.8 times the yield strain.
Table 6.3: Maximum Strains Observed in Finite Element Analyses
Beam Yield Longitudinal

Section My Mu/My ε /ε
No. Strain, εy Membrane Strain, εFEA FEA y
1 G250-1.55-904515 2.8939 1.082 1.44 12.7 8.8

2 G250-1.55-1004015 3.0814 1.124 1.44 3.153 2.2
3 G250-1.55-1204515 4.1869 1.080 1.44 4.595 3.2
4 G250-1.95-1004015 3.5735 1.138 1.44 5.619 3.9
5 G250-1.95-1006015 4.6305 1.078 1.44 12.691 8.8
6 G250-1.95-1255015 5.4513 1.119 1.44 8.782 6.1
7 G250-1.95-1504015 6.1196 1.144 1.44 5.737 4.0
8 G250-1.95-1506017 7.7051 1.097 1.44 6.835 4.7
9 G450-1.5-903014 3.9944 1.103 2.59 6.156 2.4
10 G450-1.9-804015 4.9644 1.118 2.50 14.237 5.7
11 G450-1.9-1003515 6.1275 1.142 2.50 11.75 4.7
12 G450-1.9-1154015 7.9442 1.120 2.50 9.245 3.7
Maximum Von Mises Maximum compressive strain
compressive strain developed developed at the middle of
at the middle of the web flange
(a) von Mises Strain (b) Longitudinal Strain
Figure 6.3: Strain Distribution
Figure 6.4 shows the strain variation along the compression flange and the maximum
strain in the middle of the flange is 11.6 x 10-3. Figure 6.5 shows how the
longitudinal membrane strain varied with the applied moment at the point of
maximum strain. Considering the strain development in the short beams, the
maximum strain in all cases has exceeded the limiting strain of 3ε y in AS/NZS 4600.
If the limits for the longitudinal membrane strain and the plate slenderness ratios
(dw/t) are increased in AS/NZS 4600, higher inelastic reserve capacities can be
realised for cold-formed steel lipped channel beams. Shifferaw and Schafer (2008)
observed that for a typical normalized strength prediction, typical average membrane
strain limits fall under 5 for local buckling failures. However, there are concerns
about possible fracture in the section if excessive compressive strains are allowed
and this may be the reason for the restrictions based on slenderness in AS/NZS 4600
design rules.
0.014
At compression
.
0.012 flange and web

junction
0.010
Compressive strain
Tensile strain
0.008
Strain
At compression
0.006 flange and lip
junction
0.004
0.002
0.000
0 5 10 15 20 25 30 35 40 45 50
Distance along the flange (mm)
Figure 6.4: Von Mises Strain Variation along the Compression Flange –
G250-1.95-125-50-15
7
6
Applied Moment (kNm)
0
0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.1
Strain
Figure 6.5: Longitudinal Membrane Strain Variation with the Applied Moment
at the Point of Maximum Strain - G250-1.95-125-50-15
6.3.1.2 Design Method Proposed by Shifferaw and Schafer (2008)
According to Shifferaw and Schafer (2008) considerable test data are available to
support the use of inelastic reserve capacity in the design of cold-formed flexural
members. However, they also noted that full plastic moment capacity was not
achievable based on test results. Same observation was made with finite element
analysis results of this research. Shifferaw and Schafer (2008) state that the presence
of reduced inelastic bending capacity in cold-formed steel beams in comparison to
hot-rolled steel beams is due to higher web to flange area, unsymmetric sections
resulting in first yield occurring in the tension flange and the inability of cold-formed
steel sections to sustain high compressive strains.
Shifferaw and Schafer (2008) developed suitable design equations based on a

comprehensive finite element analysis study and available experimental results of
cold-formed steel sections. Their first set of equations was based on the average
membrane compressive strains (Cy). They developed separate equations for Cy as a
function of slenderness for local and distortional buckling cases based on the back-
calculated strain values corresponding to the ultimate moments obtained from FEA.
Their equation in the case of local buckling is given next.
0.544
⎛ λly ⎞
If λl < λly , C y = ⎜ ⎟ (6.2)
⎝ λl ⎠
where, λly = 0.776 and λl = M y M crl . Here, Mcrl is the elastic local buckling
moment.
They then developed the following equation for the inelastic bending capacity (Mn)
that lie between the yield moment (My) and plastic moment (Mp) capacities.
2.42
Mn − M y ⎛ 1 ⎞
=1− ⎜ ⎟ for Cy >1 (6.3)
Mp − My ⎝ Cy ⎠
Shifferaw and Schafer (2008) also gave relevant equations for the inelastic bending
capacity as a function of slenderness using the direct strength method (DSM) format
by combining Equations 6.2 and 6.3.
⎧ ⎛ λ ⎞1.32 ⎫
M nl = M y + (M p − M y )⎨1 − ⎜ l ⎟ ⎬
⎪ ⎪
If λl < λly , (6.4)
⎜ ⎟
⎪⎩ ⎝ λly ⎠ ⎪⎭
where, Mnl and Mp are local buckling strength and plastic section moment capacity,
respectively.
Predicted inelastic bending capacities using this method are compared with FEA
moment capacities for 100 mm effective span in Table 6.4. This proposed design
method accurately predicts the inelastic bending capacity of cold-formed steel lipped
channel beams as reflected by the mean and COV of the ratio of predicted and FEA
moment capacities of 1.053 and 0.014, respectively. The capacity reduction factor is
0.97, which is slightly higher than the value of 0.95 recommended in AS/NS 4600.
Table 6.4: Comparison of Section Moment Capacities from FEA and

Design Methods Given in Shifferaw and Schafer (2008) and EC3 Part 1.3
Shifferaw
Shiffera EC 3
EC 3 and
w and Part 1.3
Section My Mu Part 1.3 Schafer
Schafer Mb
Mb (2008)
(2008)
Mu/Mb Mu/Mb
G250-1.55-904515 2.8939 3.130 3.042 3.109 1.029 0.993
G250-1.55-1004015 3.0814 3.465 3.281 3.512 1.056 1.014
G250-1.55-1204515 4.1869 4.520 4.381 4.538 1.032 1.004
G250-1.95-1004015 3.5735 4.065 3.883 4.406 1.047 1.084
G250-1.95-1006015 4.6305 4.990 4.816 4.817 1.036 0.965
G250-1.95-1255015 5.4513 6.100 5.795 6.206 1.053 1.017
G250-1.95-1504015 6.1196 7.000 6.511 7.35 1.075 1.050
G250-1.95-1506017 7.7051 8.450 8.025 8.087 1.053 0.957
G450-1.5-903014 3.9944 4.405 4.188 4.287 1.052 0.973
G450-1.9-804015 4.9644 5.550 5.248 5.444 1.058 0.981
G450-1.9-1003515 6.1275 7.000 6.532 7.051 1.072 1.007
G450-1.9-1154015 7.9442 8.900 8.312 8.824 1.071 0.991
Mean 1.053 1.003
COV 0.014 0.034
φ 0.970 0.918
6.3.1.3 Eurocode 3 Part 1.3 (ECS, 2006)
The section moment capacity of cold-formed steel beam is defined in Eurocode 3

Part 1.3 as follows, and these equations do not include a conservative limit for web
slenderness as in AS/NZS 4600.
If Weff < Wel , M c ,Rd = Weff f yb / γ M 0 (6.5)
If Weff = Wel ,
⎛ ⎞⎞
M c , Rd = f yb ⎜⎜Wel + 4(W pl − Wel )⎜1 − λ e max λ e 0 ⎟ ⎟⎟ / γ M 0 ≤ W pl f yb / γ M 0 (6.6)
⎛ _ _
⎝ ⎝ ⎠⎠
Where Weff, Wel and Wpl are the effective section modulus, gross elastic section
modulus and the plastic section modulus, respectively.
_
λ max is taken as the slenderness of the element which correspond to the largest value
_ _ _
of λ e λ e 0 . The plate element slenderness, λ p is defined in EN 1993-1-5.
_ _ _
For double supported plane elements λ e = λ p and λ e 0 = 0.5 + 0.25 − 0.055(3 + ψ )
_ _ _
For outstand elements λ e = λ p and λ e 0 = 0.673
_ _ _
For stiffened elements λ e = λ p and λ e 0 = 0.65
Section moment capacity predictions using Eurocode 3 Part 1.3 were compared with
the finite element analysis results obtained for very short beams of 100 mm effective
span in Table 6.4. It can be seen that the capacity predictions are accurate with a
mean of 1.003 and a COV of 0.034. This design code does not specify any
compressive strain limit and therefore it can be concluded that this design method
accurately predicts the inelastic reserve capacity of cold-formed steel lipped channel
beams.
In summary, inelastic reserve capacity exists for cold-formed steel lipped channel
beams. It can be predicted using Eurocode 3 Part 1.3 design rules or the proposed
direct strength method based equations of Shifferaw and Schafer (2008). Further
experiments of compact lipped channel beams are useful to investigate the concerns
over fracture of sections and the possible inability of finite element models in
predicting the plastic moment capacity of mono-symmetric beams, and also the
accuracy of currently available design equations.
6.4 Member Moment Capacity of Lipped Channel Beams Subject
to Lateral-torsional Buckling
A series of elastic buckling and non-linear finite element analyses was carried out
using the developed finite element model of simply supported cold-formed steel
lipped channel beams subjected to lateral-torsional buckling. Elastic lateral-torsional
buckling moments (Mo) and ultimate moment capacities (Mu) were obtained from
these analyses. Lateral-torsional buckling of mono-symmetric lipped channel beams
is likely to depend on a range of parameters such as beam span, initial geometric
imperfections, residual stresses, yield stress, modulus of elasticity, steel thickness,
web depth to flange width (d/b) ratio, etc. Table 6.5 presents the finite element
analysis (FEA) results of Mo and Mu in kNm. Figures 6.6 shows the typical lateral-
torsional buckling failures of lipped channel beams. They confirm that failures
occurred by lateral deflection and twisting in the anti-clockwise direction without
any cross-sectional distortion. The absence of cross-sectional distortion indicates that
the failure is governed by pure lateral-torsional buckling.
Undeformed
Deformed Shape
Figure 6.6: Lateral-Torsional Buckling Failures of Cold-formed Steel

Lipped Channel Beams (G250-1.95-100-60-15-5000)
Table 6.5: Finite Element Analysis Results
Designation
Designation
Designation
FEA (kNm) FEA (kNm) FEA (kNm)
Span (mm)
Span (mm)
Span (mm)
Beam
Beam
Beam
Mo Mu Mo Mu Mo Mu
1600 3.680 2.373 1800 6.964 4.083 800 5.411 3.049

1800 2.993 2.163 2000 5.885 3.887 1000 3.596 2.457
G250-1.95-100-60-15
G250-1.55-90-45-15
G450-1.5-90-30-14
2000 2.485 1.925 2500 4.053 3.211 1200 2.574 1.994
2500 1.681 1.384 3000 2.978 2.428 1400 1.945 1.564
3000 1.232 1.012 3500 2.305 1.788 1600 1.532 1.268
3500 0.957 0.768 4000 1.857 1.459 1800 1.246 1.027
4000 0.774 0.634 4500 1.544 1.281 2000 1.040 0.865
4500 0.647 0.532 5000 1.316 1.076 2500 0.719 0.611
5000 0.555 0.458 1400 10.039 4.701 3000 0.541 0.680
1200 5.503 2.689 1600 7.902 4.439 1200 5.485 3.882
1400 4.165 1800 6.384 4.110 1400 4.211 3.370
G250-1.95-125-50-15
2.489
1600 3.267 2.248 2000 5.273 3.743 1600 3.347 2.825
G250-1.55-100-40-15
G450-1.9-80-40-15
1800 2.637 1.986 2500 3.531 2.810 1800 2.738 2.346
2000 2.181 1.731 3000 2.566 2.049 2000 2.293 1.962
2500 1.470 1.207 3500 1.976 1.628 2500 1.596 1.367
3000 1.077 0.883 4000 1.588 1.303 3000 1.206 1.037
3500 0.837 0.689 4500 1.318 1.083 3500 0.963 0.834
4000 0.679 0.561 5000 1.122 0.925 4000 0.799 0.698
4500 0.568 0.471 1000 14.639 5.324 4500 0.683 0.603
5000 0.487 0.437 1200 10.430 4.986 5000 0.596 0.533
1400 6.518 3.462 1400 7.815 4.578 1000 7.450 4.617
1600 5.103 3.192 1600 6.090 4.099 1200 5.349 3.728
G250-1.95-150-40-15
G450-1.9-100-35-15
1800 4.105 2.877 1800 4.894 3.602 1400 4.049 3.274

G250-1.55-120-45-15
2000 3.380 2.540 2000 4.033 3.123 1600 3.191 2.603

2500 2.247 1.810 2500 2.698 2.218 1800 2.597 2.193
3000 1.622 1.326 3000 1.966 1.638 2000 2.167 1.829
3500 1.241 1.008 3500 1.519 1.276 2500 1.499 1.273
4000 0.991 0.805 4000 1.226 1.042 3000 1.127 0.972
4500 0.818 0.787 4500 1.021 0.889 1200 8.193 5.317
5000 0.693 0.665 5000 0.872 0.788 1400 6.176 4.621
G450-1.9-115-40-15
1200 6.577 3.130 1800 12.198 6.887 1600 4.839 3.926

1400 4.990 2.923 2000 10.081 6.533 1800 3.910 3.230
G250-1.95-150-60-17
1600 3.931 2.668 2500 6.696 5.293 2000 3.239 2.687

G250-1.95-100-40-15
1800 3.192 2.386 3000 4.796 4.030 2500 2.197 1.839

2000 2.657 2.096 3500 3.631 3.056 3000 1.622 1.363
2500 1.823 1.517 4000 2.867 2.392 3500 1.269 1.067
3000 1.361 1.145 4500 2.339 1.940 4000 1.035 0.878
3500 1.075 0.910 5000 1.958 1.617
4000 0.885 0.752
4500 0.751 0.652
Figures 6.7 (a) and (b) present the curves of applied moment versus vertical and
horizontal deflections, respectively, for lipped channel beams of three beam spans
subject to lateral-torsional buckling. It can be seen from Figure 6.7 (a) that vertical
deflection at failure increases with increasing span. The same could be seen with
lateral deflection. It can also be observed that the beam failure occurs suddenly in the
case of shorter spans.
3.5
3.0
.
2.5 2.0 m
3.0 m
2.0 4.0 m
Moment (kNm)
1.5
1.0
0.5
0.0
0 5 10 15 20 25 30 35 40 45 50
(a) Vertical Deflection

.
3.5
3.0
2.5 2.0 m
3.0 m
2.0 4.0 m
1.5
Moment (kNm)
1.0
0.5
0.0
0 20 40 60 80 100 120
Lateral Deflection (mm)
(b) Lateral Deflection
6.4.1 Discussion
The finite element analysis results of ultimate moment capacities for short and long
span lipped channel beams (ie. both section and member moment capacities) are
plotted in Figure 6.8 in a non-dimensional format. On the vertical axis the ratio of
ultimate moment capacity (Mu) to the yield moment capacity, which is given by My
= Zx x fy (Product of yield stress, fy and elastic section modulus, Zx), is used while the
horizontal axis represents the beam slenderness, which is given by λb = (My/Mo)0.5,
where Mo is the elastic lateral-torsional buckling moment. In calculating the beam
slenderness for the non-dimensional plot, elastic lateral-torsional buckling moments
obtained from finite element analyses were used. All the lipped channel beam
sections considered in this study are compact and therefore their section moment
capacities are higher than or equal to the yield moment capacity (Msx ≥ My).
1.2
Elastic Buckling
G250-1.55-90-45-15
Yielding G250-1.55-100-40-15
1.0 G250-1.55-120-45-15
G250-1.95-100-40-15
G250-1.95-100-60-15
0.8 G250-1.95-125-50-15
G250-1.95-150-40-15
G250-1.95-150-60-17
G450-1.9-80-40-15
0.6 G450-1.9-100-35-15
M u/M y
G450-1.9-115-40-15
G450-1.5-90-30-14
0.4 G450-1.5-90-30-12
0.2
Plastic range Inelastic range Elastic Range
0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
1/2
Slenderness regions are shown based (My/Mo)
on AS/NZS 4600 (SA, 2005)
Figure 6.8: Ultimate Moment Capacity Curves Based on Finite Element

Analyses
The moment capacity curves can be divided into three main regions depending on
the beam slenderness, namely: plastic (low slenderness), inelastic buckling
(intermediate slenderness) and elastic buckling (high slenderness). The beam
slenderness ranges of these three regions are different depending on the design code.
Short span beams or long span beams with lateral restraints at close intervals belong
to the plastic region and their capacity is governed by local buckling or yielding.
Short beams with compact cross sections fail by yielding while semi-compact or
slender sections fail by local buckling. Some cold-formed steel design codes have
limited the section moment capacity to the yield moment capacity even for compact
sections while other codes allow the use of inelastic reserve capacity. For beams with
slender cross sections the section moment capacity is influenced by local buckling
and is therefore reduced below the yield moment capacity. The local buckling case is
not further discussed in the following sections. But the results are to be used in
deciding the slenderness range of the plastic region as part of developing design
guidelines in the following sections.
The moment capacity of beams with intermediate slenderness is reduced below the
section moment capacity due to lateral-torsional buckling effects. It is strongly
influenced by the presence of residual stresses and initial geometric imperfections
and thus drops well below the elastic buckling capacity. This can be seen in Figure
6.8 for the intermediate slenderness region. The presence of residual stresses in a
beam causes premature yielding of the cross section, thus reducing the effective
stiffness of the member, and therefore also reduces the member capacity below its
elastic buckling capacity. Pi et al. (1997, 1998) and Brune and Ungeruman (2006)
have described in their studies on lipped and unlipped channel beams the reasons for
having lower capacities with negative imperfection. The presence of initial negative
geometric imperfection (crookedness and twist) causes the beam to deform laterally
and twist in the anti-clockwise direction. Twisting in the anti-clockwise direction
results in out-of-plane bending in addition to in-plane bending, and therefore
additional stresses develop in the section. As the loading increases both the
deformations and stresses in the beam increase. Stresses are also induced in the
section due to warping effect. Warping effect is critical in the intermediate
slenderness region and this effect decreases as the span length increases. The stresses
in the lipped channel beam are maximum at the lip-compressive flange junction
since the stresses developed due to in-plane bending, out-of-plane bending and
warping at the lip-compression flange are all compressive. For the intermediate
spans the presence of additional compressive stresses at the lip-flange junction due to
in-plane bending, out-of-plane bending and warping causes beam to fail in lateral-
torsional buckling at a lower capacity compared to its elastic buckling capacity.
Third region is the elastic buckling region (high slenderness) in which the beam
capacity is approximated by elastic lateral-torsional buckling capacity. However,
past research has shown that there is some deviation from the elastic buckling
capacity at high slenderness. Trahair (1998) stated that beams with high slenderness
may fail at lower moment capacities in the high slenderness region due to initial
crookedness and twist. This behaviour is highly dependent on the section type. Pi
and Trahair (1995) compared the lateral buckling capacity results of RHS sections
and equivalent I-sections. The comparison showed that as the beam slenderness
increases, the moment capacities of RHS attain their elastic buckling moments.
However, their results show that the moment capacities of I-section beams are low
and have not converged to their elastic buckling moment capacities with increasing
slenderness. Their results for I-section beams are considerably low in comparison to
RHS beams in the non-dimensional plot, proving the high dependence of the results
on the section type. Pi et al. (1997 and 1998) also observed similar behaviour for
beams with high slenderness relating to lateral distortional buckling in which beam
capacities were lower than elastic buckling capacities. The results of this research
have also confirmed this in Figure 6.8, which show that the member capacities of
beams with high slenderness are lower than their elastic buckling capacities.
However, as the slenderness increases the member capacity approaches the elastic
buckling capacity. The dependence on initial imperfections at high slenderness is
also proven correct by the imperfection sensitivity analysis in the next section, which
showed that the beam capacity increased with the reduction of initial imperfection
from L/1000 to L/3000. One can argue that the ultimate moment capacities in the
elastic buckling region has been reduced below elastic buckling moments because of
the web distortion caused by very thin plates. However, Pi et al. (1997) stated that
the effect of web distortion on the ultimate capacity is less than 2% and Figure 6.6
clearly shows that no web distortion was associated with the ultimate failure mode
and that pure lateral-torsional buckling failure occurred. However, web distortion
was observed in the case of lower span lengths in the inelastic region and these
results were not included in the comparison.
Another aspect of lateral-torsional buckling is the effect of pre-buckling deflections.
Elastic buckling analyses using Thin-wall, CUFSM and ABAQUS do not take into
account the effect of pre-buckling deflections. The lateral-torsional buckling
moments obtained by including the effects of pre-buckling deflections may exceed
the elastic buckling moments obtained without taking into account the effect of pre-
buckling deflections. The effects of pre-buckling deflections are significant if the
ratio of minor axis flexural stiffness to major axis flexural stiffness is considerably
high. This ratio of the selected cross sections varies from 0.037 to 0.20 (most of
them below 0.10). Trahair (1998) stated that the moment capacity of beams at high
slenderness can be approximated by the elastic buckling capacity based on a
compromise between the weakening effects of geometrical imperfections and the
strengthening effects of pre-buckling deflections. However, the ultimate moment
capacity may be lower than the elastic buckling capacity in slender beams if the
weakening effect of the geometric imperfection cannot be compromised by the effect
of pre-buckling deflections as observed in this research.
6.4.1.1 Effect of Cross Section
In Figure 6.8 the results are plotted as a set of buckling curves for different sections.
Data points are scattered in the region of intermediate slenderness while they are
converging as the slenderness increases. Figure 6.9 shows that the moment capacity
curves in the elastic and inelastic regions for different sections plot in an orderly
manner as a function of d/b ratio. As the d/b ratio decreases the moment capacity
curves are moving upwards in the region of intermediate slenderness and as the
slenderness increases these curves are converging.
1.2
G250-1.95-100-60-15 - d/b = 1.67

1.0
G250-1.95-125-50-15 - d/b = 2.50
G250-1.95-150-40-15 - d/b = 3.75
Elastic Buckling
0.8
M b/M y or M u /M y
0.6
0.4
0.2
0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
(My/Mo)0.5
Figure 6.9: Effect of d/b ratio on Ultimate Moment Capacity
Figure 6.8 shows that some points belonging to the G450-1.5-90-30-12 section are
located below others in the non-dimensional plot, and interestingly, this was found to
be due to the effect of lip size. To further study this phenomenon, one lipped channel
beam section (G250-1.95-150-60) was analysed with different lip sizes while
ensuring that the lip did not fail by local buckling. Other section dimensions
remained the same and the moment capacity results are plotted in a non-dimensional
format in Figure 6.10. Figure 6.10 shows that the non-dimensional ultimate moment
capacities increase with increasing lip size. The reason for this behaviour is that the
beam stiffness about Y-Y axis is increasing at a higher rate as the lip size increases,
thus increasing the lateral-torsional buckling resistance in comparison to the increase
in section moment capacity. Therefore with increasing lip size, the beam slenderness
also reduces as seen in Figure 6.10. A review of lipped channel beam sections
currently used in the industry (Stramit, 2009; Lysaght, 2009) showed that the
minimum lip size they used was 14 mm. Therefore the section G450-1.5-90-30-12
was replaced with G450-1.5-90-30-14 section. In addition the lip sizes were also
selected so that local buckling of lips was avoided, ie. the maximum lip size was
limited to 17 mm. Table 6.1 shows the selected lip sizes and other section
dimensions.
1.2
Elastic Buckling
1.0
G250-1.95-150-60-17
G250-1.95-150-60-16
0.8 G250-1.95-150-60-15
G250-1.95-150-60-14
M b/M y
0.6
0.4
0.2
0.0
0.0 0.5 1.0 1.5 2.0 2.5
(My/Mo)0.5
Figure 6.10: Effect of Lip Size on Ultimate Moment Capacity
6.4.1.2 Effect of Steel Thickness
In order to study the effect of thickness, G250-1.95-100-60-15 and G250-1.95-150-

40-15 beams were analysed with higher thicknesses of 3, 4 and 5 mm. By increasing
the thickness the section becomes more compact. It could be seen from Figures 6.11
(a) and (b) that as the thickness increases the member slenderness at which beams
first reach the elastic buckling moment reduces. In other words more compact and
stable sections reach elastic buckling moments at lower slenderness while beams
with smaller thicknesses reach elastic buckling moments at high slenderness values.
The reason for this kind of behaviour may be explained in relation to extra stress
development at the compression flange and lip junction due to minor axis bending
that resulted from twisting. As the thickness increases their torsional rigidity
increases considerably and this results in reducing the twisting angle of the beam
before failure. As a result of this, the compressive stresses at the flange and lip
junction due to minor axis bending decreases resulting in higher beam capacities that
allow elastic buckling moments to be reached.
1.2
.
Elastic Buckling
1.0
G250-1.95-1006015T20
G250-3.0-1006015T20
0.8
G250-4.0-1006015T20
G250-5.0-1006015T20
0.6
M u/M y
0.4
0.2
0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
0.5
(My/Mo)
(a) G250-t-100-60-15 Section
1.2
1.0 Elastic Buckling
G250-1.95-1504015T20
G250-3.0-1504015T20
0.8
G250-4.0-1504015T20
G250-5.0-1504015T20
0.6
Mu/My
G250-6.0-1504015T20
0.4
0.2
0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
0.5
(My/Mo)
(b) G250-t-150-40-15 Section
Figure 6.11: Effect of Steel Thickness
6.4.1.3 Effect of mechanical properties
The effects of steel grade, ie. yield strength, elastic modulus and stress-strain
relationship, were investigated by plotting the results for G250 and G450 steel
sections in Figure 6.12. The yield strengths of these two steel grades are different
whereas their elastic modulus values are of the same order. The ultimate moment
capacity plots in the case of these two steel grades appear to follow a similar trend as
seen in Figure 6.12. Pi and Trahair (1995) also studied the effect of yield stress on
the ultimate lateral-torsional buckling capacity of cold-formed steel hollow flange
beams. As shown in Figure 6.12, their results in the non-dimensional plot (Mb/Mpx
vs. λ=√(Mp/αmMy), where αm is the moment modification factor, did not depend on
the yield stress. They stated that it is because the effect of yield stress has been
eliminated by non-dimensionalising using Mp.
1.2
1.0
Elastic Buckling
G250
0.8 G450
0.6
M u /M y
0.4
0.2
0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
0.5
(My/Mo)
Figure 6.12: Effect of Mechanical Properties on Ultimate Moment

Capacity
6.4.1.4 Effect of initial geometric imperfections
As shown in Chapter 5, cold-formed steel flexural members are very sensitive to

initial geometric imperfections. Steel structures design standards worldwide have
used a nominal initial imperfection value of L/1000 as the maximum value for cold-
formed and hot-rolled steel flexural members. However, laboratory measurements
have shown that the measured imperfections are often considerably less than L/1000
(Zeinoddini and Schafer, 2008). Therefore an imperfection sensitivity study was
undertaken to investigate the effect of imperfection magnitude on the ultimate
moment capacity of lipped channel beams. Some lipped channel beam sections were
selected and analysed with different initial negative imperfection magnitudes of
L/1000, L/2000 and L/3000. Figure 6.13 (a) presents the variation of ultimate
moment capacity as the imperfection reduces for G250-1.95-125-50-15 (B6) section
while Figure 6.13 (b) shows the results for all the cross sections (B2, B3, B4, B6,
B8, B11, B12). The reduction in ultimate moment capacity was found to be the
greatest in the inelastic region (intermediate spans) as the magnitude of initial
imperfection was increased. As the magnitude of initial imperfection was reduced
from L/1000 to L/3000 the ultimate moment capacity was reaching the elastic
buckling curve for beams with higher slenderness. However, it appears that the
initial imperfection magnitude does not have a significant impact on the lateral-
torsional buckling capacity of beams with higher slenderness. Effect of initial
imperfection was also investigated by Pi and Trahair (1995) for cold-formed steel
rectangular hollow sections and observed significant effects of initial crookedness
and twist on the ultimate lateral-torsional buckling capacity.
1.2
1.0
Elastic Buckling
L/1000
L/2000
.
0.8
L/3000
0.6
M b/M y
0.4
0.2
0.0
0.0 0.5 1.0 1.5 2.0 2.5
0.5
(My/Mo)
(a) G250-1.95-125-50-15 Section

Capacity
. 1.2
1.0
Elastic Buckling
L/1000
0.8
L/2000
L/3000
0.6
M u/M y
0.4
0.2
0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
0.5
(My/Mo)
(b) Different Beam Sections

Capacity
Figures 6.14 (a) and (b) present the percentage increase in ultimate moment capacity
as the initial geometric imperfection reduces from L/1000 to L/2000 and L/1000 to
L/3000 for G250 and G450 steel sections, respectively. For grade 250 steel sections
the maximum percentage increase in moment capacity of about 7% occurs when the
beam slenderness is about 1.00 as the imperfection magnitude changes from L/1000
to L/3000, and it is about 5% when it changes from L/1000 to L/2000. However,
there is no clear trend for some sections in the case of G450 steels.
6.4.1.5 Effect of Residual Stresses
Effects of residual stresses were investigated in Section 5.1.9.1 of Chapter 5 by

comparing the moment capacity results with and without residual stresses. The
moment capacity results of beams with residual stresses were slightly lower than
those of beams without residual stresses and the influence of residual stresses on
lateral-torsional buckling capacities was found to be insignificant. However, in the
parametric study, residual stresses were included in the nonlinear analyses.
.
8
B2-L/2000
7 B3-L/2000
B4-L/2000
6
B6-L/2000
5 B2-L/3000
( M b,imp -M b,L/1000 ) / M y
B3-L/3000
4 B4-L/3000
B6-L/3000
3
0
0.0 0.5 1.0 1.5 2.0 2.5
0.5
(My/Mo)
(a) G250 Steel Beams

.
10
8
B11-L/2000
7 B12-L/2000
( M b,imp -M b,L/1000 ) / M y
6 B11-L/3000
5 B12-L/3000
0
0.0 0.5 1.0 1.5 2.0 2.5
0.5
(My/Mo)
(b) G450 Steel Beams
Figure 6.14: Ultimate Moment Capacity Increase as a Function of Initial

Geometric Imperfection Magnitude
6.4.2 Comparison of Ultimate Moment Capacities from FEA with Predictions
from the Current Design Rules at Ambient Temperature
The member capacity design rules of cold-formed steel beams subject to lateral-
torsional buckling are not the same in the well known national and international
cold-formed steel structures design codes such as AS/NZS 4600 (SA, 2005), NAS
(AISI, 2001), BS 5950 Part 5 (BS, 1998) and Eurocode 3 Part 1.3 (ECS, 2006),
except for AS/NZS 4600 and NAS (AISI, 2001) design rules that are identical. The
design rules based on the new Direct Strength Method (DSM) for lateral-torsional
buckling are presented differently, but are still based on the member capacity design
rules given in AS/NZS 4600 (SA, 2006) and NAS (AISI, 2001). However, local
buckling effects are included differently by the DSM. For this purpose AS/NZS 4600
(SA, 2005) uses the conventional effective width method for individual plate
elements and the resulting effective cross section whereas the DSM method uses the
elastic local buckling capacity of the section. In this study local buckling effect was
eliminated by using compact sections and therefore the predictions from both
methods (DSM and AS/NZS 4600) are the same.
The moment capacity curves based on AS/NZS 4600 (SA, 2005), DSM and EC 3
Part 1.3 (ECS, 2006) are plotted in Figures 6.15 (a) and (b) in a non-dimensional
format. However, BS5950 Part 5 (BS, 1998) design method cannot be plotted in this
non-dimensional format as the moment capacity curve changes depending on the
section dimensions. Therefore it was not included in this comparison. It should be
noted that AS/NZS 4600 and DSM design curves are non-dimensionalised using the
yield moment capacity, My whereas EC3 Part 1.3 design rules are non-
dimensionalised using the section moment capacity, Msx. There are considerable
differences between AS/NZS 4600 (SA, 2005), DSM and EC3 Part 1.3 (ECS, 2006)
in the plastic, inelastic and elastic buckling regions even if they are for cold-formed
steel members. As stated by Trahair (1998) these differences are due to differences
in bending moment distributions, capacity factors and load factors. All the moment
capacity curves considered here have a constant moment capacity region in the low
slenderness range, where Mu = Msx or My, although the slenderness range is different.
This slenderness range is 0 to 0.6 for AS/NZS 4600 (SA, 2005), NAS (AISI, 2001)
and DSM and the section moment capacity is assumed to be equal to the yield
moment capacity. This region is 0 to 0.202 for EC3 Part 1.3 and unlike the
Australian code this code allows inelastic reserve capacity of compact cross-sections
and hence the section moment capacity of compact section can be higher than the
yield moment capacity. AS/NZS 4600 (SA, 2005) and NAS (AISI, 2001) adopt a
capacity reduction factor of 0.9 while it is 1.0 for EC 3 Part 1.3 (ECS, 2006).
1.2
1.2
1.0
1.0
Eurocode 3 Part 1.3 - 'b'

0.8
AS/NZS4600 / DSM
0.6
M b/M sx
M b/M y
0.4
0.4
0.2 0.2
0.0 0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 0.0 0.5 1.0 1.5 2.0 2.5 3.0
(Msx/Mo)0.5 (My/Mo)0.5
(a) Eurocode 3 Part 1.3 (b) AS/NZS 4600 - DSM
Figure 6.15: Comparison of Moment Capacity Design Curves in Cold-

formed Steel Design Codes
In the intermediate slenderness (inelastic buckling) region, all the design codes
propose reduced member capacities when compared with elastic lateral-torsional
buckling capacities due to the presence of residual stresses and initial imperfections.
For beams with high slenderness (elastic buckling), the member capacity is equal to
the elastic lateral-torsional buckling moment based on AS/NZS 4600 (SA, 2005) and
DSM, ie. Mb = Mo. However, in other design codes the member capacity is less than
the elastic buckling moment, due to the influence of initial crookedness and twist
(Trahair, 1998). Figure 6.15 shows the differences in member capacity curves of
cold-formed steel beams and therefore it is necessary to investigate the accuracy of
these design rules.
The ultimate moment capacity results of finite element analyses (FEA) from this
research are compared in Tables 6.6 to 6.9 with the corresponding predictions using
AS/NZS 4600, DSM, EC3 Part 1.3 (ECS, 2006) and BS 5950 Part 5 (BS, 1998) for
cold-formed steel beams made of varying steel thicknesses and yield stresses (1.55
and 1.95 mm thick G250 steels and 1.50 and 1.90 mm thick G450 steels). Sample
calculations of member moment capacities based on these design codes are presented
in Appendix B.
Predictions for G250-1.55 Series at Ambient Temperature
FEA (kNm) Mpred. (kNm) Mu/Mpred.
Beam Designation
Span (mm)
EC 3 PART 1.3
EC 3 PART 1.3
BS5950 Part 5
BS5950 Part 5
AS/NZS 4600
AS/NZS 4600
(ECS, 2006)
(ECS, 2006)
(SA, 2005)
(SA, 2005)
(BS, 1998)
(BS, 1998)
Mo
Mu
1600 3.680 2.373 2.545 2.063 2.333 0.932 1.150 1.017
1800 2.993 2.163 2.380 1.854 2.120 0.909 1.167 1.020
G250-1.55-90-45-15
2000 2.485 1.925 2.201 1.658 1.897 0.875 1.161 1.015

2500 1.681 1.384 1.700 1.256 1.410 0.814 1.102 0.981
3000 1.232 1.012 1.246 0.978 1.074 0.812 1.035 0.942
3500 0.957 0.768 0.965 0.787 0.850 0.795 0.975 0.903
4000 0.774 0.634 0.780 0.653 0.696 0.812 0.970 0.911
4500 0.647 0.532 0.652 0.556 0.585 0.817 0.957 0.909
5000 0.555 0.458 0.558 0.483 0.503 0.820 0.948 0.909
1200 5.503 2.689 2.916 2.592 2.731 0.922 1.037 0.985
1400 4.165 2.489 2.742 2.318 2.529 0.908 1.074 0.984
G250-1.55-100-40-15
1600 3.267 2.248 2.547 2.045 2.284 0.882 1.099 0.984

1800 2.637 1.986 2.332 1.792 2.020 0.851 1.108 0.983
2000 2.181 1.731 2.099 1.571 1.769 0.824 1.102 0.978
2500 1.470 1.207 1.486 1.155 1.280 0.812 1.045 0.943
3000 1.077 0.883 1.087 0.887 0.966 0.812 0.995 0.914
3500 0.837 0.689 0.843 0.710 0.761 0.817 0.971 0.905
4000 0.679 0.561 0.683 0.587 0.622 0.821 0.955 0.902
4500 0.568 0.471 0.572 0.498 0.522 0.824 0.945 0.901
1400 6.518 3.462 3.855 3.244 3.620 0.898 1.067 0.956
1600 5.103 3.192 3.622 2.919 3.350 0.881 1.094 0.953
G250-1.55-120-45-15
1800 4.105 2.877 3.363 2.599 3.033 0.856 1.107 0.949

2000 3.380 2.540 3.078 2.302 2.701 0.825 1.103 0.940
2500 2.247 1.810 2.274 1.708 1.980 0.796 1.059 0.914
3000 1.622 1.326 1.636 1.307 1.486 0.810 1.015 0.892
3500 1.241 1.008 1.250 1.036 1.158 0.807 0.974 0.870
4000 0.991 0.805 0.998 0.847 0.934 0.807 0.950 0.862
4500 0.818 0.787 0.823 0.711 0.775 0.956 1.106 1.016
5000 0.693 0.665 0.697 0.610 0.658 0.955 1.091 1.012
Mean 0.850 1.047 0.947
COV 0.058 0.066 0.049
φ 0.769 0.942 0.861

Beam Designation
EC 3 PART 1.3
EC 3 PART 1.3
Span (mm)
BS5950 Part 5
AS/NZS 4600
AS/NZS 4600
(ECS, 2006)
(ECS, 2006)
BS5950
(SA, 2005)
(SA, 2005)
(BS, 1998)
Part 5
Mo
Mu
(BS,
1998)
1200 6.577 3.130 3.393 3.189 3.173 0.923 0.982 0.987

1400 4.990 2.923 3.201 2.840 2.948 0.913 1.029 0.992
1600 3.931 2.668 2.987 2.501 2.677 0.893 1.067 0.997
G250-1.95-100-40-15
1800 3.192 2.386 2.755 2.193 2.386 0.866 1.088 1.000

2000 2.657 2.096 2.507 1.928 2.107 0.836 1.087 0.995
2500 1.823 1.517 1.843 1.435 1.555 0.823 1.057 0.975
3000 1.361 1.145 1.373 1.120 1.195 0.834 1.022 0.958
3500 1.075 0.910 1.083 0.909 0.957 0.840 1.001 0.951
4000 0.885 0.752 0.891 0.763 0.794 0.843 0.985 0.947
4500 0.751 0.652 0.756 0.656 0.676 0.862 0.993 0.964
1800 6.964 4.083 4.282 3.517 4.032 0.954 1.161 1.013
2000 5.885 3.887 4.094 3.269 3.811 0.949 1.189 1.020
G250-1.95-100-60-15
2500 4.053 3.211 3.565 2.662 3.144 0.901 1.206 1.021

3000 2.978 2.428 2.966 2.155 2.505 0.819 1.127 0.970
3500 2.305 1.788 2.338 1.768 2.013 0.765 1.011 0.888
4000 1.857 1.459 1.877 1.480 1.653 0.777 0.985 0.882
4500 1.544 1.281 1.558 1.264 1.390 0.822 1.013 0.922
5000 1.316 1.076 1.326 1.099 1.192 0.812 0.979 0.903
1400 10.039 4.701 5.188 4.630 4.901 0.906 1.015 0.959
1600 7.902 4.439 4.936 4.226 4.630 0.899 1.050 0.959
1800 6.384 4.110 4.656 3.816 4.300 0.883 1.077 0.956
G250-1.95-125-50-15
2000 5.273 3.743 4.352 3.423 3.927 0.860 1.093 0.953

2500 3.531 2.810 3.492 2.600 3.004 0.805 1.081 0.936
3000 2.566 2.049 2.592 2.019 2.301 0.790 1.015 0.891
3500 1.976 1.628 1.992 1.618 1.814 0.818 1.006 0.898
4000 1.588 1.303 1.599 1.335 1.475 0.815 0.976 0.883
4500 1.318 1.083 1.326 1.129 1.232 0.817 0.959 0.879
5000 1.122 0.925 1.129 0.975 1.052 0.819 0.948 0.879
Predictions for G250-1.95 Series at Ambient Temperature – Contd.

Beam Designation
BS5950 Part 5 (BS,

AS/NZS 4600 (SA,
AS/NZS 4600 (SA,

Span (mm)
EC 3 PART 1.3
EC 3 PART 1.3
(ECS, 2006)
(ECS, 2006)
BS5950
2005)
1998)
2005)
Part 5
Mo
Mu
(BS,
1998)
1200 10.430 4.986 5.721 5.207 5.363 0.872 0.958 0.930

1400 7.815 4.578 5.350 4.584 4.948 0.856 0.999 0.925
1600 6.090 4.099 4.932 3.985 4.448 0.831 1.029 0.922
G250-1.95-150-40-15
1800 4.894 3.602 4.469 3.450 3.919 0.806 1.044 0.919

2000 4.033 3.123 3.966 2.995 3.421 0.788 1.043 0.913
2500 2.698 2.218 2.724 2.166 2.455 0.814 1.024 0.903
3000 1.966 1.638 1.981 1.645 1.838 0.827 0.995 0.891
3500 1.519 1.276 1.529 1.305 1.438 0.834 0.977 0.887
4000 1.226 1.042 1.233 1.072 1.166 0.845 0.971 0.893
4500 1.021 0.889 1.027 0.905 0.973 0.866 0.982 0.913
5000 0.872 0.788 0.877 0.781 0.831 0.898 1.008 0.948
1800 12.198 6.887 7.127 5.891 6.725 0.966 1.169 1.024
2000 10.081 6.533 6.805 5.458 6.367 0.960 1.197 1.026
G250-1.95-150-60-17
2500 6.696 5.293 5.876 4.390 5.261 0.901 1.206 1.006

3000 4.796 4.030 4.788 3.493 4.161 0.842 1.154 0.969
3500 3.631 3.056 3.667 2.813 3.298 0.833 1.086 0.927
4000 2.867 2.392 2.890 2.311 2.667 0.828 1.035 0.897
4500 2.339 1.940 2.354 1.937 2.205 0.824 1.001 0.880
5000 1.958 1.617 1.969 1.655 1.860 0.821 0.977 0.869
Mean 0.852 1.044 0.941
COV 0.057 0.068 0.049
φ 0.772 0.939 0.856
FEA
Mpred. (kNm) Mu/Mpred.
(kNm)
Beam Designation
BS5950 Part 5 (BS,

AS/NZS 4600 (SA,
AS/NZS 4600 (SA,
BS5950 Part 5 (BS,

Span (mm)
EC 3 PART 1.3
EC 3 PART 1.3
(ECS, 2006)
(ECS, 2006)
2005)
1998)
2005)
1998)
Mo
Mu
800 5.411 3.049 3.574 2.928 3.537 0.853 1.041 0.862

1000 3.596 2.457 3.112 2.364 2.936 0.789 1.039 0.837
G450-1.5-90-30-14
1200 2.574 1.994 2.568 1.882 2.287 0.776 1.059 0.872

1400 1.945 1.564 1.982 1.514 1.787 0.789 1.033 0.875
1600 1.532 1.268 1.555 1.241 1.431 0.815 1.022 0.886
1800 1.246 1.027 1.262 1.037 1.175 0.814 0.990 0.875
2000 1.040 0.865 1.052 0.883 0.986 0.823 0.980 0.878
2500 0.719 0.611 0.726 0.631 0.686 0.841 0.968 0.890
Mean 0.813 1.017 0.872
COV 0.031 0.030 0.018
φ 0.744 0.931 0.802
FEA
Mpred. (kNm) Mu/Mpred.
Beam Designation
Span (mm) (kNm)
EC 3 PART 1.3
EC 3 PART 1.3
BS5950 Part 5
AS/NZS 4600
AS/NZS 4600
BS5950 Part 5
(ECS, 2006)
(ECS, 2006)
(SA, 2005)
(SA, 2005)
(BS, 1998)
(BS, 1998)
Mo
Mu
1200 5.485 3.882 4.206 3.365 4.049 0.923 1.154 0.959
1400 4.211 3.370 3.776 2.866 3.446 0.892 1.176 0.978
1600 3.347 2.825 3.303 2.439 2.876 0.855 1.158 0.982
G450-1.9-80-40-15
1800 2.738 2.346 2.794 2.089 2.412 0.840 1.123 0.973

2000 2.293 1.962 2.334 1.809 2.050 0.841 1.084 0.957
2500 1.596 1.367 1.615 1.326 1.453 0.846 1.031 0.941
3000 1.206 1.037 1.217 1.032 1.106 0.852 1.005 0.937
3500 0.963 0.834 0.971 0.840 0.887 0.860 0.993 0.940
4000 0.799 0.698 0.805 0.707 0.738 0.867 0.987 0.945
4500 0.683 0.603 0.688 0.611 0.632 0.877 0.988 0.954
5000 0.596 0.533 0.600 0.537 0.552 0.888 0.991 0.965
1000 7.450 4.617 5.312 4.424 5.216 0.869 1.044 0.885
G450-1.9-100-35-15
1200 5.349 3.728 4.699 3.643 4.402 0.793 1.024 0.847

1400 4.049 3.274 4.003 2.991 3.576 0.818 1.095 0.916
1600 3.191 2.603 3.246 2.482 2.912 0.802 1.049 0.894
1800 2.597 2.193 2.634 2.091 2.409 0.833 1.049 0.910
2000 2.167 1.829 2.194 1.790 2.030 0.834 1.022 0.901
2500 1.499 1.273 1.513 1.288 1.420 0.841 0.988 0.896
3000 1.127 0.972 1.137 0.992 1.071 0.855 0.980 0.907
1200 8.193 5.317 6.530 5.164 6.389 0.814 1.030 0.832
1400 6.176 4.621 5.750 4.318 5.378 0.804 1.070 0.859
G450-1.9-115-40-15
1600 4.839 3.926 4.878 3.614 4.435 0.805 1.086 0.885

1800 3.910 3.230 3.974 3.052 3.676 0.813 1.058 0.879
2000 3.239 2.687 3.284 2.609 3.089 0.818 1.030 0.870
2500 2.197 1.839 2.219 1.858 2.127 0.829 0.990 0.864
3000 1.622 1.363 1.635 1.410 1.577 0.833 0.966 0.864
3500 1.269 1.067 1.278 1.124 1.235 0.835 0.950 0.864
4000 1.035 0.878 1.042 0.929 1.006 0.842 0.944 0.872
Mean 0.842 1.038 0.910
COV 0.035 0.059 0.047
φ 0.771 0.938 0.828
6.4.2.1 AS/NZS 4600 and Direct Strength Method (DSM)
As shown in Tables 6.6 to 6.9 the mean and COV values of the ratio of ultimate
moment capacities obtained from FEA and AS/NZS 4600 (SA, 2005) are 0.850 and
0.058, respectively for 1.55 mm thick G250 steel while they are 0.852 and 0.057,
respectively for 1.95 mm thick G250 steel, respectively. The mean and COV values
are 0.813 and 0.031, and 0.842 and 0.035 for 1.5 mm and 1.9 mm thick G450 steels.
This means that the predictions of AS/NZS 4600 are higher than the FEA results and
are considered unsafe. This is also confirmed by the corresponding capacity
reduction factor of about 0.7 calculated from these mean and COV values, which is
less than the currently used factor of 0.9.
Figures 6.16 (a) and (b) compare the ultimate moment capacity results from FEA
with the moment capacity curve based on the design rules in AS/NZS 4600 (SA,
2005) for G250 and G450 steels, respectively. The results are plotted in a non-
dimensional format of ultimate moment capacity (Mb/My) versus beam slenderness
(My/Mo)0.5. Figures 6.16 (a) and (b) show that AS/NZS 4600 (SA, 2005) design rules
over-predict the moment capacity of intermediate and long span beams subject to
lateral-torsional buckling. Although the section moment capacity results of shorter
beams are also plotted in Figures 6.16 (a) and (b), the mean and COV values
reported in this section are based on the moment capacity results for lateral-torsional
buckling only.
Further, according to AS/NZS 4600 (SA, 2005) the ultimate moment capacity of
long spans beams with a member slenderness, λ>1.336, is equal to their elastic
lateral-torsional buckling moments (Mu = Mo). However, Figures 6.16 (a) and (b)
show that the moment capacities have been reduced below the elastic buckling
moments. Pi and Trahair (1995) have observed the same behaviour for other types of
sections. Pi et al.’s (1997 and 1998) research also observed that the moment
capacities of lipped channel beams subject to lateral-distortional buckling were
below the elastic buckling moment curve even at high slenderness values.
1.2
AS/NZS 4600 (SA, 2005)
.
Elastic Buckling
1.0 G250-1.55-90-45-15
G250-1.55-100-40-15
G250-1.55-120-45-15
0.8
G250-1.95-100-40-15
M b/M y or M u/M y
G250-1.95-100-60-15
0.6 G250-1.95-125-50-15
G250-1.95-150-40-15
G250-1.95-150-60-17
0.4
0.2
0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
0.5
(My/Mo)

.
1.2
AS/NZS 4600 (SA, 2005)
Elastic Buckling
1.0
G450-1.5-1004015T20
G450-1.9-804015T20
0.8 G450-1.9-1003515T20
G450-1.9-1154015T20
Mb/My or M u/My
G450-1.5-903014T20
0.6
0.4
0.2
0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
0.5
(My/Mo)
Figure 6.16: Comparison of FEA Results with AS/NZS 4600 Moment

Capacity Curve for G250 and G450 Steel Beams
Based on Tables 6.6 to 6.9 and Figures 6.16 (a) and (b) it can be concluded that the
AS/NZS 4600 (SA, 2005) moment capacity predictions are unsafe in both elastic and
inelastic buckling regions for cold-formed steel lipped channel beams subject to
lateral-torsional buckling. Brune and Ungermann (2006) have also observed that
DSM design rules (NAS, 2001) which are based on AS/NZS 4600(SA, 2005)
overestimate the ultimate moment capacities of unlipped channel beams subject to
lateral-torsional buckling. Appendix C shows the sample calculations of moment
capacities based on AS/NZS 4600 (SA, 2005) design rules including non-
dimensional moment capacity ratios. Non-dimensional slenderness values required
in plotting Figures 6.16 (a) and (b) are also presented in Appendix D.
The measured initial imperfection magnitudes of lipped channel beams were about
L/2000 (see Chapter 4). However, a larger imperfection value of L/1000 was used in
finite element analyses as used or recommended by many other researchers and
international codes. It is not known what imperfection magnitude was used in the
development of AS/NZS 4600 design rules. Effect of imperfection magnitude on the
member moment capacity is considered significant in the intermediate slenderness
region. Therefore it was decided to undertake an imperfection sensitivity study to
determine whether AS/NZS 4600 (SA, 2005) design rules give accurate predictions
if a lower geometric imperfection magnitude is used. The results of this imperfection
sensitivity study are presented in Section 6.4.1.4 Figure 6.17 compares the moment
capacity results obtained for varying imperfection magnitudes (L/1000, L/2000 and
L/3000) with the moment capacity curve of AS/NZS 4600. It can be seen from
Figure 6.17 that as the imperfection magnitude is reduced from L/1000 to L/3000 the
moment capacities are increasing, but are still below the moment capacity curve of
AS/NZS 4600 (SA, 2005). At high member slenderness values of 2.5, the ultimate
moment capacities of beams with an imperfection of L/3000 approach the elastic
buckling curve. However, they are noticeably below the elastic buckling moment
curve in the member slenderness range of 1.336 to 2.1. These observations confirm
that the moment capacity design rules in AS/NZS 4600 (SA, 2005) are to be
considered unsafe even for beams with a smaller initial imperfection of L/3000, and
therefore a suitable modification is needed.
. 1.2
1.0
AS/NZS4600
Elastic Buckling
0.8
M u / M y or M b / M y
L/1000
L/2000
0.6 L/3000
0.4
0.2
0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
0.5
(My/M0)
Figure 6.17: Comparison of FEA Results for Beams with Varying Initial
Imperfections with AS/NZS 4600 Moment Capacity Curve
Observations made in relation to AS/NZS 4600 (SA, 2005) are also applicable for
the Direct Strength Method (DSM) since the latter is based on the same moment
capacity equations for lateral-torsional buckling.
Eurocode 3 Part 1.3 (ECS, 2006) recommends the use of the design equations based
on buckling curve ‘b’ in Eurocode 3 Part 1.1 (ECS, 2005) for the design of cold-
formed steel beams subject to lateral-torsional buckling. These design equations are
given in Section 2.5.3 of Chapter 2. The ultimate moment capacities from finite
element analyses are compared with the predicted moment capacities of Eurocode 3
Part 1.3 in Tables 6.6 to 6.9 for cold-formed steel beams made of 1.5 and 1.9 mm
thick G450 steels and 1.55 and 1.95 mm thick G250 steels. The mean values of the
ratio of moment capacities from FEA and EC 3 Part 1.3 are 1.017 and 1.038 for
G450 1.5 and 1.9 mm thick steels, respectively, while they are 1.047 and 1.044 for
G250 1.55 and 1.95 mm thick steels, respectively. The corresponding COVs are
0.066 and 0.068 for 1.55 and 1.95 mm thick G250 steels and 0.030 and 0.059 for 1.5
and 1.9 mm thick G450 steels. Figures 6.18 (a) and (b) compare the FEA results with
the moment capacity design curve of Eurocode 3 Part 1.3 (ECS, 2006) for G250 and
G450 steels, respectively. The moment capacity results are plotted in a non-
dimensionalised format using the section moment capacity Msx since Eurocode 3 Part
1.3 allows the use of inelastic reserve capacity. The section moment capacity was
taken as the finite element analysis results obtained for the shortest span length
(beam span of 100 mm). It appears that Eurocode 3 Part 1.3 accurately predicts the
moment capacities of beams with high slenderness values as defined by
(Msx/Mo)0.5≥1.5, while its predictions are slightly overconservative for beams having
intermediate slenderness values in the range of (Msx/Mo)0.5<1.5.
1.2
EC 3 Part 1.3 - b
Elastic Buckling
1.0 G250-1.55-90-45-15
G250-1.55-100-40-15
G250-1.55-120-45-15
0.8
G250-1.95-100-40-15
M b/M sx or M u/M sx
G250-1.95-100-60-15
0.6 G250-1.95-125-50-15
G250-1.95-150-40-15
G250-1.95-150-60-17
0.4
0.2
0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
0.5
(Msx/Mo)
Figure 6.18: Comparison of FEA Results with Eurocode 3 Part 1.3 (ECS,
2006) Moment Capacity Curve
1.2
EC 3 Part 1.3 (ECS, 2006) - 'b'
Elastic Buckling
1.0
G450-1.9-80-40-15
G450-1.9-100-35-15
0.8
G450-1.9-115-40-15
G450-1.5-90-30-14
0.6
0.4
0.2
0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
0.5
(Msx/Mo)
Figure 6.18: Comparison of FEA Results with Eurocode 3 Part 1.3

(ECS, 2006) Moment Capacity Curve
Eurocode 3 Part 1.1 suggests different buckling curves depending on the type of
section and d/b ratios and Eurocode 3 Part 1.3 recommends the use of buckling curve
‘b’ for cold-formed steel beams with equations given in Eurocode 3 Part 1.1. Based
on the comparison with FEA results it is clear that buckling curve ‘b’ is
overconservative for beam slenderness values {(My/Mo)0.5} below 1.5 while it gives
accurate predictions for high slenderness values (above 1.5). Therefore all the
buckling curves given in Eurocode 3 Part 1.1 are compared with FEA results in
Figure 6.19 to determine the most suitable buckling curve for cold-formed steel
members. It can be seen that the buckling curve ‘a’ predicts the moment capacities
accurately in comparison to FEA results throughout the elastic and inelastic
slenderness ranges. Comparison of buckling curve ‘a’ with the results is also shown
in Figure 6.20.
1.2
Eurocode 3 Part 1.3 - 'a'
Eurocode 3 Part 1.3 - 'b'
Eurocode 3 Part 1.3 - 'c'
1.0 Eurocode 3 Part 1.3 - 'd'
Elastic Buckling
G250-1.55-90-45-15
G250-1.55-100-40-15
0.8 G250-1.55-120-45-15
G250-1.95-100-40-15
G250-1.95-100-60-15
G250-1.95-125-50-15
G250-1.95-150-40-15
0.6 G250-1.95-150-60-17
G450-1.9-80-40-15
G450-1.9-100-35-15
G450-1.9-115-40-15
0.4 G450-1.5-90-30-14
0.2
0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
0.5
(Msx/Mo)
Figure 6.19: Comparison of FEA Results with Different Buckling Curves

of Eurocode 3 Part 1.1 (ECS, 2005)
1.2
Eurocode 3 Part 1.3 - 'a'
Elastic Buckling
G250-1.55-90-45-15
1.0
G250-1.55-100-40-15
G250-1.55-120-45-15
G250-1.95-100-40-15
0.8 G250-1.95-100-60-15
G250-1.95-125-50-15
G250-1.95-150-40-15
G250-1.95-150-60-17
0.6
G450-1.9-80-40-15
G450-1.9-100-35-15
G450-1.9-115-40-15
0.4 G450-1.5-90-30-14
0.2
0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
0.5
(Msx/Mo)
Figure 6.20: Comparison of Buckling Curve ‘a’ of Eurocode 3 Part 1.3

(ECS, 2006) with FEA Results
The mean, COV and capacity reduction factors are presented in Table 6.10 for the
current and proposed buckling curves in Eurocode 3 Part 1.3. The buckling curve ‘a’
of Eurocode 3 Part 1.1 predicts moment capacities of steel beams made of both grade
steels with an overall mean of 0.995 and associated COV of 0.066. It is noted the
predictions based on buckling curve ‘a’ are equally accurate for cold-formed steel
members irrespective of their thickness or steel grade. Therefore it is recommended
that Eurocode 3 Part 1.1 design equations are used with buckling curve ‘a’ to predict
the moment capacities of cold-formed steel beams subject to lateral-torsional
buckling, instead of buckling curve ‘b’ as proposed in the current version of
Eurocode 3 Part 1.3. It should be also noted that as shown in Section 6.4.1.1 the
effect of d/b ratios on the moment capacity is not too high to demand different
buckling curves depending on the d/b ratios of cold-formed steel sections as
proposed for hot-rolled I beams in EC 3 Part 1.1. Figure 6.9 on the effect of d/b
ratios shows that a single buckling curve is adequate for the commonly used lipped
channel beams with d/b ratios in the range of 2.0 to 3.3
.
Factors of the Ratio of FEA to Predicted Moment Capacities using Buckling
Curve ‘a’ of Eurocode 3 Part 1.1
Mult./Mpred.
Grade and EC 3 Part 1.3 (ECS, 2006)/ EC 3 Part 1.1 (ECS, 2005)
Thickness EC3 Part 1.1 (ECS, 2005) Buckling curve ‘a’
Mean COV Φ Mean COV Φ
G250-1.55 1.047 0.066 0.942 0.990 0.063 0.893
G250-1.95 1.044 0.068 0.939 0.997 0.060 0.901
G450-1.5 1.017 0.030 0.931 0.952 0.023 0.875
G450-1.9 1.038 0.059 0.938 0.982 0.050 0.893
G250 1.055 0.070 0.948 1.004 0.065 0.905
G450 1.038 0.054 0.942 0.982 0.047 0.895
Overall 1.048 0.066 0.944 0.995 0.061 0.899
6.4.2.3 BS5950 Part 5 (BS, 1998)
Tables 6.6 to 6.9 compare the ultimate moment capacities from finite element
analyses and the predicted moment capacities of BS5950 Part 5 (BS, 1998). The
equation given in BS5950 Part 5 (BS, 1998) for elastic lateral-torsional buckling
moment capacity of mono-symmetric cold-formed beams bending in the plane of
web is different from that given in other design codes (AS/NZS 4600 (SA, 2005) and
EC 3 Part 1.3 (ECS, 2006)), and gives higher elastic buckling moments. As a result
of this the moment capacity predictions of BS5950 Part 5 are somewhat higher than
the FEA results. The mean and COV values of the ratio of ultimate moment
capacities obtained from FEA and BS5950 Part 5 are 0.947 and 0.049 for 1.55 mm
thick G250 steels, respectively, while they are 0.941 and 0.049 for 1.95 mm thick
G250 steels. The associated capacity reduction factors for 1.55 and 1.95 mm thick
G250 steels are 0.861 and 0.856, respectively. The mean and COV values of the
same ratio are 0.872 and 0.018 for 1.5 mm thick G450 steels while they are 0.910
and 0.047 for 1.9 mm thick steels. The capacity reduction factors are 0.802 and
0.828 for 1.5 and 1.9 mm thick G450 steels, respectively. Therefore it can be
concluded that BS5950 Part 5 (BS, 1998) design equations for lateral-torsional
buckling give unsafe predictions for cold-formed steel lipped channel beams.
6.5 Development of Design Rules
Based on the comparison with the ultimate moment capacities of cold-formed steel
beams obtained from finite element analyses, it was identified that some
modifications are needed for the AS/NZS 4600 (SA, 2005) moment capacity
equations for lateral-torsional buckling. In this section a suitable modification is
proposed for the AS/NZS 4600 (SA, 2005) design rules for lateral-torsional
buckling. Since the design rules in NAS (AISI, 2001) are identical to AS/NZS 4600
design rules, any modifications to AS/NZS 4600 design rules are equally applicable
to the former.
6.5.1 AS/NZS 4600 (SA, 2005) Design Rules
The design rules in AS/NZS 4600 (SA, 2005) are based on the following
assumptions as discussed next. Fully laterally restrained or short beams made of
compact sections can attain a full plastic moment capacity, which has been assumed
to be equal to 10/9 times their first yield moment capacity (My = fyZx), where 10/9 is
the shape factor for wide flange beams based on the partial plastification of the
section in bending (Galambos, 1963). It has been assumed that long beams with a
higher slenderness ratio undergo elastic buckling when their moment capacity is
equal to or less than half of the maximum expected section moment capacity of 10/9
My (= 0.56My). Hence for beams with slenderness values (λb=(Msx/Mo)0.5) larger
than 1.336, their moment capacity is assumed to be equal to their elastic lateral-
torsional buckling moments (see Figure 6.21). When the beam slenderness is in the
range of 0.6 and 1.336, the beam capacities are considered to be governed by
inelastic buckling and a Johnson parabolic equation was used between Points A and
B representing moment capacities of 0.56My to 10/9My, as shown in Figure 6.21.
AS/NZS 4600 (SA, 2005) design method has limited the maximum moment capacity
to first yield moment (My) for cold-formed steel beams as a conservative measure.
Therefore a plateau is proposed for the low slenderness region with a moment
capacity of My until Point D on the parabolic curve shown in Figure 6.21. According
to Yu (2000) the inelastic buckling curve for single, double and point-symmetric
sections has been confirmed by research in beam-columns (Pekoz and Sumer, 1992)
and wall studs (Kian and Pekoz, 1994).
1.2 Corresponding
A to 10/9f y
C D
1.0 AS/NZS4600 (SA, 2005)
Elastic Buckling
Yielding
0.8
0.6 B
M b/Msx
Corresponding
to 0.56M y
0.4
0.2
Yielding Inelastic Buckling Elastic Buckling
0.0
0.0 0.5 0.6 1.0 1.336 1.5 2.0 2.5 3.0
(Msx/Mo)0.5
Figure 6.21: AS/NZS 4600 Moment Capacity Curve for Lateral-torsional

Buckling
The design method given in AS/NZS 4600 (SA, 2005) is given next. The nominal
member moment capacity ( M b ) of the laterally unbraced segments of singly, doubly,
and point-symmetric sections subjected to lateral buckling is given by,
⎛M ⎞
M b = Zc ⎜ c ⎟ (6.7)
⎜Z ⎟
⎝ f ⎠
Z c and Z f are the effective section modulus calculated at a stress level (M c Z f ) in
the extreme compression fibre, and the full unreduced section modulus for the
extreme compression fibre, respectively.
The critical moment (Mc) is calculated as follows,
For λb ≤ 0.60 Mc = M y (6.8a)
⎡ ⎛ 10λb2 ⎞⎤
For 0.60 < λb < 1.336 M c = 1.11M y ⎢1 − ⎜⎜ ⎟⎟⎥ (6.8b)
⎣⎢ ⎝ 36 ⎠⎦⎥
⎛ 1 ⎞
For λb ≥ 1.336 M c = M y ⎜⎜ 2 ⎟
⎟ (6.8c)
⎝ λb ⎠
where, λb = M y M 0 - Non-dimensional slenderness ratio used to determine
M c for members subject to lateral buckling
M y = Z f fy - Moment causing initial yield at the extreme
compression fibre of the full section (first yield moment).
As explained in Section 6.4.2.1, these design equations over-predict the moment

capacities of cold-formed steel lipped channel beams. Therefore a new set of
moment capacity equations was proposed as given in the next Section.
6.5.1.1 New Equations for AS/NZS 4600 (SA, 2005)
Option 1
The critical moment (Mc) in Equation 6.7 can be calculated as follows,
For λb ≤ 0.60 Mc = M y (6.9a)
For 0.60 < λb < 1.336 [

M c = 1.132M y 1 − 0.326λb2 ] (6.9b)
⎛ 0.84 ⎞
For λb ≥ 1.336 M c = M y ⎜⎜ 2 ⎟⎟ (6.9c)
⎝ λb ⎠
The moment capacity results from FEA showed that the ultimate moment capacities
of beams in the elastic buckling region (λb>1.336) could not attain their elastic
buckling capacity in all the cases. When initial geometric imperfections were
reduced, their moment capacities approached the elastic-lateral-torsional buckling
capacity, indicating the reasons for lower capacities as initial crookedness and twist
of beams. Therefore a reduction factor of 0.84 was introduced to Equation 6.8c to
allow for the deviation of the moment capacity results from elastic buckling
moments (Equation 6.9c). A factor of 0.877 is used in the column design equation of
AS/NZS 4600 (SA, 2005) to allow for the reduced capacities of columns of high
slenderness below the elastic buckling capacity. The slenderness range for the plastic
region was maintained up to 0.6 as used in AS/NZS 4600 and by considering the
section moment capacity results from finite element analyses. Equation 6.9b was
developed for the inelastic buckling region based on the same parabolic format.
Figure 6.22 shows the comparison of moment capacities predicted by the proposed
equations and AS/NZS 4600 (SA, 2005), and FEA results. It can be seen from Figure
6.22 that the new design equations has improved the accuracy in the inelastic and
elastic buckling regions in comparison to AS/NZS 4600 (SA, 2005) and DSM, but is
still unsafe in the inelastic buckling (intermediate slenderness) region.
Table 6.11 presents the mean and COV values and the capacity reduction factors of
the ratio of finite element analysis results to the moment capacities obtained from the
proposed design equations. The use of new equations has improved the mean value
from 0.852 in the case of current design rules in AS/NZS 4600 to 0.964 while the
corresponding COVs remained the same at 0.056. However, it can be seen from
Figure 6.22 that the predictions of Equations 6.9 (a) to (c) are unsafe in the
intermediate slenderness region and therefore more accurate rules were developed
(Option 2).
1.2 Option 1
Elastic Buckling
G250-1.55-904515T20
1.0 G250-1.55-1004015T20
G250-1.55-1204515T20
G250-1.95-1004015T20
G250-1.95-1006015T20
0.8 G250-1.95-1255015T20
G250-1.95-1504015T20
M b/M y or M u/M y
G250-1.95-1506017T20
0.6 G450-1.9-804015T20
G450-1.9-1003515T20
G450-1.9-1154015T20
G450-1.5-903014T20
0.4
0.2
0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
0.5
(M y/M o)
Figure 6.22: Comparison of New Design Equations with FEA Results -

Option 1
Factors of the Ratio of FEA to Predicted Moment Capacities using Option 1
Option 1
Grade and
Mu/Mpred.
Thickness
Mean COV Φ
G250-1.55 0.958 0.058 0.867
G250-1.95 0.955 0.046 0.870
G450-1.5 0.930 0.062 0.832
G450-1.9 0.971 0.057 0.879
G250 0.965 0.055 0.876
G450 0.971 0.060 0.878
Overall 0.964 0.056 0.874
Option 2
Since the beam design curve proposed under Option 1 is not sufficiently accurate in
the inelastic region an attempt was made to improve Equation 6.9b in the inelastic
region and this was proposed as Option 2.
The critical moment (Mc) in Equation 6.7 can be calculated as follows,
For λb ≤ 0.60 Mc = M y (6.10a)
For 0.60 < λb < 1.336 M = M ⎡0.2λ2 − 1.11λ + 1.59⎤ (6.10b)

c y ⎢⎣ b b ⎥⎦
⎛ 0.84 ⎞
For λb ≥ 1.336 M c = M y ⎜⎜ 2 ⎟⎟ (6.10c)
⎝ λb ⎠
Figure 6.23 shows the comparison of finite element analysis results with the
proposed beam design curve. As observed in Figure 6.23 this design method
accurately predicts the moment capacities of cold-formed steel lipped channel
beams. The mean, COV and capacity reduction factors given in Table 6.12 also
confirm the accuracy of the moment capacity equations proposed under Option 2.
1.2 Option 2
Elastic Buckling
G250-1.55-904515T20
G250-1.55-1004015T20
1.0
G250-1.55-1204515T20
G250-1.95-1004015T20
G250-1.95-1006015T20
0.8 G250-1.95-1255015T20
G250-1.95-1504015T20
M b/M y or M u/M y
G250-1.95-1506017T20
0.6 G450-1.9-804015T20
G450-1.9-1003515T20
G450-1.9-1154015T20
G450-1.5-903014T20
0.4
0.2
0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
0.5
(M y/M o)

Option 2
The mean and COV values of the ratio of ultimate moment capacities obtained from
FEA and the new equations proposed under Option 2 are 1.000 and 0.048 for 1.55
mm thick G250 steels, respectively, while they are 0.997 and 0.039 for 1.95 mm
thick G250 steels. The capacity reduction factors are 0.910 and 0.912 for 1.55 and
1.95 mm thick G250 steels, respectively. The mean and COV values of the same
ratio for 1.5 mm thick G450 steels are 0.962 and 0.021 while they are 1.003 and
0.032 for 1.9 mm thick G450 steels. The associated capacity reduction factors for 1.5
and 1.9 mm thick G450 steels are 0.884 and 0.920, respectively. Therefore it can be
concluded that the new design equations proposed under Option 2 accurately predict
the moment capacities of lipped channel beams subject to lateral-torsional buckling.
However, there is a drawback in the proposed design method in which the member
moment capacities in the elastic buckling region are given as 0.84 times the elastic
lateral-torsional buckling moment. The member moment capacities from finite
element analyses in the elastic buckling region are also below the elastic buckling
capacities. However, as the member slenderness increases they approach the
theoretical elastic lateral-torsional buckling moments. The proposed design equation
for the elastic buckling region (Equation 6.10c) reduces the member capacities by a
factor of 0.84 even for very slender beams, and this may not be acceptable.
Therefore another set of design equations was developed, which overcomes this
problem (Option 3).
Option 2
Grade and
Mult./Mpred.
Thickness
Mean COV Φ
G250-1.55 1.000 0.048 0.910
G250-1.95 0.997 0.039 0.912
G450-1.5 0.962 0.021 0.884
G450-1.9 1.003 0.032 0.920
G250 1.007 0.048 0.918
G450 1.003 0.034 0.919
Overall 1.003 0.044 0.915
Option 3
New design equations were developed under Option 3 based on the finite element
analysis results. The new equation is similar to the moment capacity equation in the
Australian hot-rolled steel structures code, AS 4100 (SA, 1998). The critical moment
(Mc) in Equation 6.7 can be calculated as follows,
For λb ≤ 0.6 Mc = M y (6.11a)
⎡ ⎤
For λb > 0.6 M = 0.9⎢ λ4 + 2.05 − λ2 ⎥ M (6.11b)
c ⎣ b b ⎦ y
a b
Figure 6.24 compares the FEA results with the proposed moment capacity curve. It
appears that the new equation accurately predicts the moment capacities of cold-
formed steel beams in all the slenderness regions, and is therefore recommended for
the design of cold-formed steel lipped channel beams subject to lateral-torsional
buckling.
1.2 Option 3
Elastic Buckling
G250-1.55-904515T20
G250-1.55-1004015T20
1.0
G250-1.55-1204515T20
G250-1.95-1004015T20
G250-1.95-1006015T20
0.8 G250-1.95-1255015T20
G250-1.95-1504015T20
M b/M y or M u/M y
G250-1.95-1506017T20
0.6 G450-1.9-804015T20
G450-1.9-1003515T20
G450-1.9-1154015T20
G450-1.5-903014T20
0.4
0.2
0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
0.5
(M y/M o)
Figure 6.24: Comparison of FEA Results with New Design Beam Curve
– Option 3
The mean, COV and capacity reduction factors of FEA to predicted moment
capacities based on Option 3 are given in Table 6.13. The mean of the ratio of
moment capacities from FEA and Equations 6.11 (a) and (b) are 1.004 and 0.981 for
1.55 mm thick G250 steels and 1.95 mm thick G250 steels, respectively. The mean
values of the same ratio are 0.944 and 0.978 for G450 1.5 and 1.9 mm thick steels,
respectively. The corresponding COV’s are 0.067 and 0.053 for 1.55 and 1.95 mm
thick G250 steels, respectively, and 0.018 and 0.046 for 1.5 and 1.9 mm thick G450
steels, respectively. The overall capacity reduction factor of the new design
equations is 0.892, which is very close to the recommended value of 0.9 in AS/NZS
4600. The mean and COV values and the capacity reduction factors indicate the
accuracy of the proposed design method.
Grade and
Mu/Mpred.
Thickness
Mean COV Φ
G250-1.55 1.004 0.067 0.902
G250-1.95 0.981 0.053 0.891
G450-1.5 0.944 0.018 0.868
G450-1.9 0.978 0.046 0.892
G250 0.990 0.060 0.896
G450 0.978 0.044 0.893
Overall 0.984 0.056 0.892
Considering Figure 6.24 and Table 6.13 it is concluded that the new design equations
are able to accurately predict the moment capacities of cold-formed steel lipped
channel beams subject to lateral-torsional buckling.
The new design equations proposed under Option 3 have many advantages in
comparison to the AS/NZS 4600 design equations in three slenderness regions.
Equation 6.11b provides a simple way of calculating lateral-torsional buckling
capacities in the inelastic and elastic buckling regions. This equation eliminates the
problem of using 0.84 x Mo for beams with very high slenderness under Option 2.
The lateral-torsional buckling capacity prediction using this method tends to coincide
with the elastic buckling moment as the beam slenderness increases. Another
advantage of this proposal is that this equation can be modified by changing ‘a’ and
‘b’ as shown in Equation 6.11b to develop multiple beam curves. Moment capacities
are dependent on the type of section and it is useful to develop multiple beam curves
for different section types such as lipped channel sections, hollow flange sections,
RHS, back to back C-sections etc. Trahair (1998) has also shown the need to have
multiple design curves to allow for varying section types, residual stresses and initial
imperfections. Further equations 6.11a and 6.11b can also be easily adopted to
calculate the lateral-torsional buckling capacities at elevated temperatures when
modifications are needed to allow for the non-linearity in stress-strain relationship.
This will be discussed in the next chapter.
The new equations are also presented in the DSM format as shown next. The nominal
member moment capacity, M be , for lateral-torsional buckling is given by
For M o > 6.25M y M be = M y (6.12a)
⎡ ⎛My ⎞
2
⎛M ⎞⎤⎥
For 6.25M y ≥ M o M be = 0 .9 ⎢ ⎜⎜ ⎟⎟ + 2.05 − ⎜⎜ y ⎟⎟ M y (6.12b)
⎢ ⎝ Mo ⎠ ⎝ Mo ⎠⎥⎦
⎣
6.6 Capacity Reduction Factor Calculations
The American cold-formed steel structures code (AISI, 2007) recommends a

statistical model to evaluate the structural performance in terms of capacity reduction
factor. This capacity reduction factor allows for the variations in material, fabrication
and loading effects. The capacity reduction factor, φ, is given by,
− β 0 Vm 2 +V f 2 + C pV p 2 +Vq 2
φ = 1.52 M m Fm Pm e (6.13)
where, Mm, Vm = Mean and coefficient of variation of the material factor (1.1, 0.1)
Fm, Vf = Mean and coefficient of variation of the fabrication factor (1, 0.05)
Vq = coefficient of variation of load effect = 0.21
β0 = Target reliability index = 2.5
⎛ 1 ⎞⎛ m ⎞
Cp = Correction factor depending on the number of tests = ⎜1 + ⎟⎜ ⎟
⎝ n ⎠⎝ m − 2 ⎠
Pm = mean value of the tested to predicted load ratio
Vp = Coefficient of variation of the tested to predicted load ratio
n = Number of tests
m = Degree of freedom = n-1
Substituting all the values in Equation 6.13,
− 2.5 0.0566 + C pV p 2
φ = 1.672 Pm e (6.14)
The values of Pm and Vp have to be determined from experimental or numerical

analysis results. In this investigation finite element analysis results provided the base
set of results. Therefore Pm and Vp are the mean and COV of the ratio of ultimate
moment capacities from finite element analyses and cold-formed steel design
equations.
The capacity reduction factors determined using Equation 6.14 for various design
codes are summarised next. Table 6.14 presents the summary of mean, COV and
capacity reduction factors calculated based on the comparison of ultimate moment
capacities from finite element analyses and design codes. The results are also
presented separately for each steel thickness and grade in Table 6.14. They clearly
show that the accuracy of design rules in AS/NZS 4600 (SA, 2005), EC 3 Part 1.3
(ECS, 2006) and BS5950 Part 5 (BS, 1998) does not rely on the type of steel grade
or thickness. It appears that predictions of BS5950 Part 5 for G250 steel beams are
unconservative with the mean and COV values of 0.936 and 0.059 with an
associated capacity reduction factor of 0.847. Eurocode 3 Part 1.3 (ECS, 2006)
design rules are the best in comparison to other codes with its overall mean and COV
values of 1.048 and 0.066 and a capacity reduction factor of 0.944. As seen in Table
6.14, AS/NZS 4600 (SA, 2005) and DSM design rules give mean and COV values of
0.852 and 0.056 and are the most unconservative. AS/NZS 4600 (SA, 2005)
recommends a capacity reduction factor of 0.9 for flexural members subjected to
lateral buckling. It can be seen from Table 6.14 that in all the cases the capacity
reduction factor is less than 0.9 indicating the unconservatism of the design rules.
Hence the moment capacity design equations and curves in AS/NZS 4600 (SA,
2005), DSM and EC 3 Part 1.3 (ECS, 2006) were modified in this chapter based on
the available finite element analysis results.
Table 6.14: Summary of Mean and COV Values and Associated

Capacity Reduction Factors of the Ratio of FEA to Predicted Moment
Capacities
Mu/Mpred.
Grade and
AS/NZS 4600 EC 3 Part 1.3 BS 5950 Part 5
Thickness
Mean COV Φ Mean COV Φ Mean COV Φ
G250-1.55 0.850 0.058 0.769 1.047 0.066 0.942 0.947 0.049 0.861
G250-1.95 0.852 0.057 0.772 1.044 0.068 0.939 0.941 0.049 0.856
G450-1.5 0.813 0.031 0.744 1.017 0.030 0.931 0.872 0.018 0.802
G450-1.9 0.842 0.035 0.771 1.038 0.059 0.938 0.910 0.047 0.828
G250 0.860 0.061 0.777 1.055 0.070 0.948 0.952 0.056 0.864
G450 0.842 0.037 0.771 1.038 0.054 0.942 0.910 0.046 0.829
Overall 0.852 0.056 0.773 1.048 0.066 0.944 0.936 0.059 0.847
It was found that Eurocode 3 Part 1.3 design equations could be used with buckling
curve ‘a’ for which the imperfection factor is 0.21, instead of curve ‘b’. New design
equations under Option 3 were proposed for AS/NZS 4600 (SA, 2005) and DSM.
Table 6.15 demonstrates the improvements provided by the new equations by
presenting the mean and COV values and capacity reduction factors obtained for
AS/NZS 4600 (SA, 2005) and EC 3 Part 1.3 (ECS, 2006) design rules following the
proposed modifications. The proposed new equations and EC 3 Part 1.3 capacity
predictions with buckling curve ‘a’ have similar mean and COVs.
Factors for the New Equations and EC3 Part 1.1 – Buckling Curve ‘a’
Mu/Mpred.
Grade and
EC 3 Part 1.3 – Buckling Curve ‘a’ Option 3 – New Proposal
Thickness
Mean COV Φ Mean COV Φ
G250-1.55 0.990 0.063 0.893 0.979 0.059 0.885
G250-1.95 0.997 0.060 0.901 0.981 0.053 0.891
G450-1.5 0.952 0.023 0.875 0.944 0.018 0.868
G450-1.9 0.982 0.050 0.893 0.978 0.046 0.892
G250 1.004 0.065 0.905 0.990 0.060 0.896
G450 0.975 0.047 0.888 0.971 0.044 0.886
Overall 0.995 0.061 0.899 0.984 0.056 0.892
6.7 Conclusion
This chapter has described an extensive parametric study on the section and member
moment capacities of simply-supported cold-formed steel lipped channel beams
subject to uniform bending about the major axis at ambient temperature. This
parametric study used a validated ideal finite element model of unrestrained lipped
channel beams subject to lateral-torsional buckling effects and another ideal model
of short lipped channel beams subject to inelastic local buckling and yielding effects.
The ultimate moment capacities of lipped channel beams made of different steel
thicknesses and grades were obtained from finite element analyses performed using
ABAQUS. All the selected sections are classified as compact according to AS/NZS
4600 (SA, 2005). The finite element analysis results for section moment capacity
showed that cold-formed steel lipped channel beams possess inelastic reserve
capacity but the full plastic moment capacity is not achievable. The maximum
membrane strain in the compression flange developed at failure was less than five
times the yield strain in most cases, but exceeded the maximum recommended strain
in AS/NZS 4600. It was found that the proposed design method by Shifferaw and
Schafer (2008) and the design method given in EC 3 Part 1.3 accurately predict the
inelastic bending capacity of short lipped channel beams. EC 3 Part 1.3 predicts the
inelastic bending capacities more accurately than Shifferaw and Schafer (2008).
However, the design method in EC 3 Part 1.3 is lengthy and complex while
Shifferaw and Schafer’s (2008) design method is simple and allows easier
calculations of inelastic bending capacity.
The ultimate moment capacities in the case of lateral-torsional buckling of cold-

formed steel lipped channel beams made of different steel thicknesses and grades
were obtained from finite element analysis. The results show that the current design
rules in AS/NZS 4600 (SA, 2005) are unsafe for lipped channel beams subject to
lateral-torsional buckling and therefore new design equations were proposed in this
chapter. The new design equations were shown to predict the moment capacities
accurately. They can also be used with the direct strength method accurately. The
capacity reduction factor of 0.9 recommended in AS/NZS 4600 (SA, 2005) can be
used with the new equations.
The predictions of BS 5950 Part 5 (BS, 1998) design equations were also found to be
unsafe. However, the design method given in EC 3 Part 1.3 (ECS, 2006) (Design
Method given in EC 3 Part 1.1 with buckling curve ‘b’) was found to be accurate for
cold-formed steel beams with high slenderness values while its predictions for beams
with intermediate slenderness are over-conservative. Therefore the bucking curve ‘a’
is proposed instead of ‘b’ in the equations given in EC 3 Part 1.1 for cold-formed
steel beams.
7 Behaviour and Design of Cold-formed Steel
Beams Subject to Local Buckling and
Lateral-Torsional Buckling at Elevated
Temperatures
The section and member moment capacities of cold-formed steel lipped channel
beams subject to local buckling and yielding, and lateral-torsional buckling were
investigated under ambient temperature conditions and presented in the previous
chapter. Behavioural and moment capacity changes to cold-formed steel lipped
channel beams under elevated temperature conditions in a fire situation are
investigated in this chapter. No research has been reported on the lateral-torsional
buckling behaviour of cold-formed steel beams under fire conditions. However,
considerable research has been carried out on the lateral-torsional buckling
behaviour of hot-rolled steel members at elevated temperatures (Real et al. (1999,
2001, 2002, 2003, 2004a, 2004b, 2006), Bailey et al. (1996), Yin and Wang (2003)
and Dharma and Tan (2007)). Lopes et al. (2006) conducted research on the lateral-
torsional buckling behaviour of stainless steel beams at elevated temperatures.
Lateral-torsional buckling behaviour becomes more complicated at elevated

temperatures in comparison to that at ambient temperature due to the factors such as
material non-linearity, different yield strength to elastic modulus ratios etc.
Therefore a good understanding and knowledge of lateral-torsional buckling
behaviour of cold-formed steel beams at elevated temperature is very important.
Changes to mechanical properties of cold-formed steels such as yield strength,
elastic modulus and stress-strain relationship at elevated temperatures have been
investigated by conducting tensile coupon tests and the details are presented in
Chapter 3. A new set of equations was developed for the reduction factors of yield
strength and elastic modulus and is presented in Chapter 3.
The section and member moment capacities of cold-formed steel lipped channel
beams at elevated temperatures were investigated by carrying out an extensive
parametric study using finite element analyses. Applicability of ambient temperature
design methods (Eurocode 3 Part 1.3 (ECS, 2006) and Shifferaw and Schafer (2008))
and fire design method (Eurocode 3 Part 1.2 (ECS, 2005)) for the section moment
capacity of cold-formed steel beams was investigated based on the results of the
parametric study. The applicability of design method given in Eurocode 3 Part 1.3
(ECS, 2006) and the new design method developed based on AS/NZS 4600 (SA,
2005) was verified for fire situations with reduced mechanical properties based on
the results of the parametric study of lateral-torsional buckling. The accuracy of fire
design method given in Eurocode 3 Part 1.2 (ECS, 2005) and the proposed method
by Dharma and Tan (2007) for hot-rolled steel beams was also verified using the
results of the parametric study.
This chapter presents the results of the parametric study carried out on the flexural
behaviour of cold-formed steel lipped channel beams at elevated temperatures,
comparison of the results with various design methods and the development of new
design guidelines for lateral-torsional buckling at elevated temperatures.
7.1 Details of Finite Element model used in the Parametric Study
A simply supported beam subject to uniform bending was used in the parametric
study of cold-formed steel lipped channel beams subject to lateral-torsional buckling
and a short beam model with four-point loading was used to study the section
moment capacity of cold-formed lipped channel beams at elevated temperatures. The
finite element models used in the parametric study are similar to those used in the
previous chapter in terms of element type, element size, load application, boundary
conditions, magnitude, type and the direction of the initial imperfection, etc. The
yield strength and elastic modulus of steel decrease with increasing temperature and
therefore appropriately reduced mechanical properties were used to account for the
elevated temperature conditions.
The yield strength and elastic modulus at elevated temperatures were obtained using
the developed equations in Chapter 3 for the reduction factors and the ambient
temperature mechanical properties. The measured values from tensile coupon tests
were taken as the yield strength and elastic modulus at ambient temperature. For the
sake of completeness, these details are provided next.
Table 7.1: Ambient and Elevated Temperature Mechanical Properties

Used in the Parametric Study
fy (MPa) E (MPa)
Temp
G250 G450 G250 G450
(oC)
20 294.0 204.0 271.0 188.0 537.0 207.0 515.0 206.0

100 282.2 190.4 260.2 175.4 536.7 193.2 514.8 192.2
200 267.5 173.3 246.6 159.7 531.8 175.9 510.0 175.0
300 193.3 145.9 178.2 134.4 510.2 148.0 489.3 147.3
400 140.4 118.3 129.4 109.1 372.7 120.0 357.5 119.5
500 99.2 90.8 91.4 83.7 209.8 92.1 201.2 91.7
600 65.3 63.3 60.2 58.3 59.1 64.2 56.7 63.9
700 36.6 35.7 33.7 32.9 37.6 36.2 36.1 36.1
Reduction Factors for Yield Strength

For G250 steels,
f y ,T
20 ≤ T ≤ 200 o C = −0.0005T + 1.01 (7.1a)
f y , 20
f y ,T
200 < T ≤ 800 o C
f y , 20
(
= 25 1.16 − T 0.022 ) (7.1b)
For G450 steels (Option 1),

f y ,T ⎧ (T − 20)4.56 ⎫
20 ≤ T < 300 o C = ⎨1 − ⎬ (7.2a)
f y , 20 ⎩ 1x1010 T ⎭
300 ≤ T < 600 o C

f y ,T ⎧
= ⎨0.95 −
(T − 300)1.45 ⎫ (7.2b)
⎬
f y , 20 ⎩ 7.76T ⎭
f y ,T
600 ≤ T ≤ 800 o C = −0.0004T + 0.35 (7.2c)
f y , 20
Reduction Factors for Elastic Modulus

For G250 and G450 steels,
ET
20 ≤ T ≤ 200 o C = −0.000835T + 1.0167 (7.3a)
E20
ET
200 < T ≤ 800 o C = −0.00135T + 1.1201 (7.3b)
E 20
The stress-strain relationship was established using Ranawaka and Mahendran’s
(2009) predictive equations and appropriate modifications were made in Chapter 3.
The strain hardening material model was used for steels with gradual yielding type
stress-strain curve except for G250 steels at 200oC which have a stress-strain
relationship with a well defined yield point. For G250 steels at 200oC elastic-perfect
plastic material model was used. For the sake of completeness, the details of stress-
strain material model used here are given next.
Equation for Stress-Strain Curve

ηT
f ⎛ f y ,T ⎞⎛ f T ⎞
ε T = T + β ⎜⎜ ⎟⎟⎜ ⎟ (7.4)
ET ⎜ ⎟
⎝ T ⎠⎝ f y ,T
E ⎠
For G450 steels, 20 ≤ T ≤ 800 o C

β = 0.86
η T = −3.05 x10 −7 T 3 + 0.0005T 2 − 0.2615T + 62.653 (7.5a)
For G250 steels, 300 ≤ T ≤ 800o C

β = 1.5
η T = 0.000138T 2 − 0.085468T + 19.212 (7.5b)
As used in the finite element models for the ambient temperature case, a negative
imperfection of L/1000 was used for lateral-torsional buckling while the magnitude
of imperfection used for local buckling was 0.006w. Beams with a negative overall
geometric imperfection yield lower moment capacities compared to positive
imperfections in the case of lateral-torsional buckling at ambient temperature. This is
also applicable to elevated temperatures. Residual stresses were also included in the
non-linear analysis although they rapidly decrease with increasing temperature. The
residual stresses at elevated temperatures were calculated using an equation
developed by Lee (2004) and the ambient temperature residual stresses proposed by
Ranawaka and Mahendran (2006) and Schafer and Pekoz (1998). These details are
presented in Chapter 5.
7.2 Selection of Member Sizes for the Parametric Study
Nine lipped channel beam cross-sections out of the twelve cross-sections analysed at
ambient temperature were selected for the parametric study at elevated temperatures.
Three sections used in the ambient temperature study were not used here as there
were only minor differences in the results between some sections of the same steel
grade. The beams are made of 1.5 mm and 1.9 mm thick G450 steels and 1.55 and
1.95 mm thick G250 cold-formed steels. This selection was made according to
AS/NZS 4600 (SA, 2005) so that plate elements are compact at all the elevated
temperatures based on the calculations using the mechanical properties at
corresponding temperatures. Selected cross-sections are listed in Table 7.2. The
nomenclature of the lipped channel beams used is shown below.
G250-1.95-150-60-17-2500-T200
Temperature (oC)
Span (mm)
Lip (mm)
Flange (mm)
Web (mm)
Thickness (mm)
Grade
The parametric study of section moment capacities at elevated temperatures was

carried out using the short span beam model with three effective beam spans (100,
500 and 800 or 1000 mm). Effective beam span is termed here as the length of the
short beam finite element model between the two ends of the loading plates as the
failure of the two end segments is restricted. In total 284 elastic and non-linear finite
element analyses were carried out using the short span beam model. Beam spans
selected for the parametric study of lateral-torsional buckling vary from 1000 to
5000 mm which represents the range of beam lengths failing by lateral-torsional
buckling. Beams undergoing local, distortional or lateral-distortional buckling were
carefully avoided by checking the failure mode at the ultimate capacity. Beams
having convergence problems were not included. In total 1060 elastic buckling and
nonlinear analyses were carried out in the case of lateral-torsional buckling.
Table 7.2: Dimensions of Selected Cross-sections for the Parametric

Study at Elevated Temperatures
Thickness Web Flange Lip

No. Beam Designation Grade b/t d/t d/b
(mm) (mm) (mm) (mm)
B1 G-250-1.55-90-45-15 1.55 90 45 15 29.03 58.06 2.00

250
B2 G-250-1.55-120-45-15 1.55 120 45 15 29.03 77.42 2.67
B3 G-250-1.95-100-60-15 1.95 100 60 15 30.77 51.28 1.67
B4 G-250-1.95-125-50-15 1.95 125 50 15 25.64 64.10 2.50

250
B5 G-250-1.95-150-40-15 1.95 150 40 15 20.51 76.92 3.75
B6 G-250-1.95-150-60-17 1.95 150 60 17 30.77 76.92 2.50
B7 G-450-1.5-90-30-14 450 1.5 90 30 14 20.00 60.00 3.00
B8 G-450-1.9-80-40-15 1.9 80 40 15 21.05 42.11 2.00

450
B9 G-450-1.9-115-40-15 1.9 115 40 15 21.05 60.53 2.88
7.3 Section Moment Capacity of Cold-formed Lipped Channel

Elastic buckling and non-linear analyses of short span beams were carried out for the
selected beam spans of each beam cross-section at different elevated temperatures.
The finite element analysis results obtained for the elastic buckling and ultimate
failure loads are given in Table 7.3 in the case of 100 mm effective beam span. The
results for other spans are presented in Appendix F.
It was observed that at elevated temperatures the failure mode was the same as that
obtained at ambient temperature and was always local buckling for beams with 100
mm effective span. The lowest eigen mode of the beams with effective spans of 500
mm or above was either local or distortional buckling. In the non-linear analysis,
eigen mode corresponding to local buckling was always used to input the initial
geometric imperfections in obtaining the relevant ultimate moment capacities. Figure
7.1 shows the elastic buckling and ultimate failure modes of a beam having 100 mm
effective span at temperatures of 300oC and 600oC. It can be seen from the elastic
buckling and ultimate failure modes that the failure was caused by inelastic local
buckling at both 300oC and 600oC.
Table 7.3: Section Moment Capacities at Elevated Temperatures –

100 mm Effective Span
FEA Msx (kNm) Mu/Msx
Schafer (2008)
Schafer (2008)
Shifferaw and
Shifferaw and
EC 3 Part 1.3
EC 3 Part 1.2
EC 3 Part 1.2
EC 3 Part 1.3
Section
Temp.
My
(oC)
Mo Mu Mu/My
(kNm)
(kNm) (kNm)
20 2.894 12.035 3.130 1.082 3.109 3.109 3.042 1.007 1.007 1.029
G250-1.55-90-45-15
200 2.633 10.226 2.830 1.075 2.787 2.829 2.759 1.015 1.000 1.026
300 1.903 8.606 2.200* 1.156 2.081 2.044 2.009 1.057 1.076 1.095
400 1.382 6.982 1.650* 1.194 1.544 1.485 1.467 1.069 1.111 1.125
500 0.976 5.357 1.125* 1.153 1.107 1.049 1.040 1.016 1.073 1.082
600 0.643 3.732 0.710 1.105 0.729 0.690 0.686 0.974 1.028 1.035
700 0.360 2.107 0.396 1.099 0.408 0.387 0.384 0.969 1.023 1.030
20 4.187 12.629 4.570 1.092 4.538 4.538 4.381 1.007 1.007 1.043
G250-1.55-120-45-15
200 3.810 10.731 4.155 1.091 4.061 4.130 3.967 1.023 1.006 1.047
300 2.753 9.031 3.250* 1.180 3.043 2.984 2.895 1.068 1.089 1.123
400 2.000 7.326 2.395* 1.198 2.260 2.168 2.117 1.060 1.105 1.131
500 1.412 5.621 1.610 1.140 1.626 1.531 1.502 0.990 1.052 1.072
600 0.930 3.916 1.065 1.145 1.070 1.008 0.992 0.995 1.057 1.074
700 0.520 2.211 0.590 1.133 0.599 0.564 0.556 0.985 1.046 1.061
20 4.631 18.943 4.990 1.078 4.817 4.817 4.816 1.036 1.036 1.036
G250-1.95-100-60-15
200 4.214 16.096 4.305 1.022 4.322 4.383 4.368 0.996 0.982 0.986
300 3.045 13.547 3.610* 1.186 3.221 3.168 3.179 1.121 1.140 1.136
400 2.212 10.989 2.655* 1.200 2.387 2.301 2.320 1.112 1.154 1.144
500 1.562 8.432 1.805* 1.156 1.710 1.625 1.644 1.056 1.111 1.098
600 1.028 5.875 1.145 1.113 1.137 1.070 1.085 1.007 1.070 1.055
700 0.576 3.317 0.635 1.103 0.637 0.599 0.608 0.997 1.060 1.044
20 5.451 23.306 6.100 1.119 6.206 6.206 5.795 0.983 0.983 1.053
G250-1.95-125-50-15
200 4.961 19.803 5.550 1.018 5.571 5.647 5.255 0.996 0.983 1.056
300 3.585 16.666 4.605* 1.285 4.103 4.081 3.827 1.122 1.128 1.203
400 2.604 13.520 3.325* 1.277 2.980 2.964 2.794 1.116 1.122 1.190
500 1.839 10.373 2.230* 1.213 2.104 2.093 1.980 1.060 1.065 1.126
600 1.211 7.227 1.425* 1.177 1.386 1.378 1.307 1.028 1.034 1.090
700 0.678 3.825 0.785* 1.158 0.776 0.772 0.738 1.012 1.017 1.064
Table 7.3: Section Moment Capacities at Elevated Temperatures –
100 mm Span (Contd.)
FEA Msx (kNm) Mu/Msx
Section
and Schafer
and Schafer
Temp.
My
Shifferaw
Shifferaw
EC 3 Part
EC 3 Part
EC 3 Part
EC 3 Part
(oC)
Mu/My
(2008)
(2008)
(kNm) Mo Mu
1.3
1.2
1.3
1.2
(kNm) (kNm)
20 6.120 21.842 7.000 1.144 7.213 7.213 6.511 0.970 0.970 1.075
G250-1.95-150-40-15
200 5.569 18.559 6.350 1.140 6.564 6.564 5.895 0.967 0.967 1.077
300 4.024 15.619 5.250* 1.305 4.743 4.743 4.306 1.107 1.107 1.219
400 2.923 12.671 3.805* 1.302 3.445 3.445 3.150 1.104 1.104 1.208
500 2.064 9.722 2.525* 1.223 2.433 2.433 2.235 1.038 1.038 1.130
600 1.359 6.773 1.605* 1.181 1.602 1.602 1.477 1.002 1.002 1.087
700 0.761 3.825 0.895 1.176 0.897 0.897 0.827 0.998 0.998 1.082
20 7.705 24.338 8.450 1.097 8.087 8.087 8.025 1.045 1.045 1.053
G250-1.95-150-60-17
200 7.012 20.680 7.650 1.091 7.233 7.359 7.271 1.058 1.040 1.052
300 5.067 17.404 5.950* 1.174 5.426 5.318 5.304 1.097 1.119 1.122
400 3.680 14.119 4.255* 1.156 4.039 3.863 3.878 1.053 1.102 1.097
500 2.599 10.833 3.030* 1.166 2.902 2.728 2.751 1.044 1.111 1.101
600 1.711 7.547 1.955* 1.142 1.932 1.796 1.816 1.012 1.089 1.077
700 0.958 4.262 1.085 1.133 1.084 1.005 1.017 1.001 1.079 1.067
20 3.994 11.310 4.405 1.103 4.287 4.287 4.188 1.028 1.028 1.052
G450-1.5-90-30-14
200 3.956 9.610 4.325 1.093 4.058 4.246 4.035 1.066 1.019 1.072
300 3.795 8.088 4.040 1.065 3.728 4.073 3.805 1.084 0.992 1.062
400 2.772 6.561 3.020 1.089 2.821 2.976 2.690 1.071 1.015 1.123
500 1.561 5.034 1.795 1.150 1.734 1.675 1.606 1.035 1.072 1.118
600 0.439 3.507 0.555* 1.263 0.513 0.472 0.500 1.082 1.177 1.110
700 0.280 1.980 0.342* 1.221 0.326 0.300 0.308 1.048 1.138 1.109
20 4.964 21.363 5.550 1.118 5.444 5.444 5.248 1.019 1.019 1.058
G450-1.9-80-40-15
200 4.917 18.152 5.450 1.108 5.213 5.392 5.102 1.045 1.011 1.068
300 4.716 15.277 5.200 1.103 4.843 5.172 4.838 1.074 1.005 1.075
400 3.446 12.393 3.860 1.120 3.631 3.779 3.547 1.063 1.022 1.088
500 1.939 9.509 2.250* 1.160 2.183 2.127 2.024 1.031 1.058 1.112
600 0.546 6.625 0.685* 1.254 0.622 0.599 0.606 1.101 1.144 1.130
700 0.348 3.741 0.428* 1.232 0.396 0.381 0.380 1.081 1.123 1.126
20 7.944 23.043 8.900 1.120 8.824 8.824 8.312 1.009 1.009 1.071
G450-1.9-115-40-15
200 7.868 19.580 8.550 1.087 8.411 8.739 8.140 1.017 0.978 1.050
300 7.547 16.478 8.100 1.073 7.779 8.383 7.722 1.041 0.966 1.049
400 5.514 13.367 6.050 1.097 5.854 6.125 5.691 1.033 0.988 1.063
500 3.104 10.257 3.540 1.141 3.550 3.447 3.276 0.997 1.027 1.081
600 0.874 7.146 1.100* 1.259 1.012 0.971 0.963 1.087 1.133 1.142
700 0.556 4.035 0.685* 1.232 0.644 0.618 0.610 1.064 1.109 1.123
‘*’ – Ultimate failure load exceeding plastic section moment capacity.
Elastic Local Buckling
Ultimate Failure Shape
(a) At 300oC (b) At 600oC
Figure 7.1: Elastic Buckling and Ultimate Failure Modes at Elevated

Temperatures – G250-1.55-120-45-15
The ultimate failure in the non-linear analyses of all the beams with 100 mm
effective span occurred by cross-section yielding and associated inelastic local
buckling. The ultimate failure capacities were above the yield moment capacity for
very short beams with an effective span of 100 mm except in a few cases since all
the selected sections are compact at elevated temperatures. The ultimate capacities of
beams with spans of 500 mm and above are less than the yield moment capacity in
most cases. The reason for this seems to be the presence of lateral deformations with
increasing span. The average of the inelastic moment capacities of cold-formed steel
lipped channel beams with 100 mm span is 1.15 My. In other words the average
inelastic reserve capacity achievable in the cold-formed steel beams according to the
finite element analysis results is 15% of the yield moment capacity at elevated
temperatures. The results of other spans were not considered in calculating the
average. At ambient temperature very short beams achieved 11.6% of the yield
moment capacity as the inelastic reserve capacity. Although the beams with compact
cross-sections are capable of achieving the full plastic moment capacity (Mp=Sxfy,T)
with fully yielded cross-section at failure, the beams have not achieved the full
plastic moment capacity in most cases. Interestingly it can be seen that some beams
exceeded their full plastic moment capacity (see Table 7.3). This can be explained
with the stress-strain relationship at elevated temperature. At elevated temperatures,
the stress-strain relationship becomes more and more non-linear and the yield
strength is approximately taken as the 0.2% proof stress. The beams were analysed
using the strain hardening material model which represents the true material
behaviour whereas in calculating the plastic moment capacities, the approximately
determined yield stress (0.2% proof stress) was used. Therefore it is possible for the
ultimate moment capacities from FEA to exceed the full plastic moment capacity.
7.3.1 Comparison of Ultimate Capacities from Finite Element Analyses and

Current Design Rules
It was found that the design method given in Eurocode 3 Part 1.3 (ECS, 2005) and
the method proposed by Shifferaw and Schafer (2008) based on the inelastic reserve
capacity are accurate for ambient temperature design. The non-linear finite element
analyses carried out on the section moment capacity revealed that inelastic reserve
capacity also exists at elevated temperatures. The accuracy of these two methods was
investigated in fire design by using the elevated temperature mechanical properties
in this section. Accuracy of AS/NZS 4600 design method based on inelastic reserve
capacities was not investigated since most of the selected beams did not satisfy the
plate slenderness conditions in AS/NZS 4600 (SA, 2005). The accuracy of fire
design code, Eurocode 3 Part 1.2 (ECS, 2005), was also assessed against the finite
element analysis results for section moment capacity.
Eurocode 3 Part 1.3 (ECS, 2006) allows the use of inelastic reserve capacity for
section moment capacity at ambient temperature and the relevant design equations
were given in Section 6.3.1.3 of Chapter 6. This design method was used to calculate
the section moment capacities at elevated temperatures with the reduced mechanical
properties. The yield strength and elastic modulus were calculated using the
developed equations reported in Chapter 3 for the reduction factors and the measured
mechanical properties at ambient temperature. Appendix B provides the sample
calculations of section moment capacities using this design method. Table 7.4
presents the summary of mean, COV and the capacity reduction factors of the ratio
of FEA to predicted section moment capacities at different elevated temperatures and
their overall values. Only the results for very short beams with an effective span of
100 mm were used in the comparison.

FEA with Predicted Capacities using Eurocode 3 Part 1.3 (ECS, 2006)
Temp. oC 20 200 300 400 500 600 700 Overall

Mean 1.012 1.020 1.086 1.076 1.030 1.032 1.017 1.039
COV 0.022 0.030 0.025 0.025 0.022 0.042 0.035 0.039
φ 0.929 0.934 0.997 0.987 0.946 0.939 0.929 0.950
It can be seen that the capacity predictions are accurate with an overall mean of
1.039 with less than 4% error and a COV of 0.039 for all the temperatures. The
capacity reduction factor calculated based on a procedure recommended by AISI
(2007) is 0.950. Therefore this design method can be accurately used to calculate the
inelastic reserve capacity of lipped channel beams at elevated temperatures.
The section moment capacity based on the fire design code, Eurocode 3 Part 1.2
(ECS, 2005) is determined using the product of the ambient temperature section
moment capacity calculated based on the ambient temperature design code for cold-
formed steel members, Eurocode 3 Part 1.3 (ECS, 2006) and the yield strength
reduction factors to account for elevated temperatures. The design equation for the
section moment capacity, Mfi,θ,Rd in Eurocode 3 Part 1.2 (ECS, 2005) is given next.
M fi ,θ ,Rd = k y ,θ M Rd (7.6)
where ky,θ is the yield strength reduction factor while MRd is the ambient
temperature section moment capacity calculated using Eurocode 3 Part 1.3.
The predicted section moment capacities using this method are given in Table 7.3
while the summary of the ratio of FEA to predicted section moment capacities is
given in Table 7.5. The sample calculations for the section moment capacity at 20oC
using this method are shown in Appendix B, and the same method was followed in
calculating the elevated temperature section moment capacities by replacing
mechanical properties at ambient temperature with those at elevated temperatures.
The overall mean of the ratio of FEA of predicted capacities is 1.053 with a COV of
0.051. The mean and COV values indicate the accuracy of this design method. If the
predictions of Eurocode 3 Part 1.3 and Eurocode 3 Part 1.2 are compared it can be
seen that Eurocode 3 Part 1.3 design method gives accurate predictions at different
elevated temperatures compared to the latter. However, Eurocode 3 Part 1.2 method
is simple to use and does not require the calculations of plate element slenderness at
each elevated temperature as needed in Eurocode 3 Part 1.3.

FEA with Predicted Capacities using Eurocode 3 Part 1.2 (ECS, 2005)
Temp. oC 20 200 300 400 500 600 700 Overall

Mean 1.012 0.996 1.067 1.078 1.065 1.079 1.064 1.053
COV 0.022 0.025 0.056 0.049 0.028 0.055 0.047 0.051
φ 0.929 0.915 0.961 0.976 0.977 0.973 0.965 0.958
7.3.1.3 Design Method Proposed by Shifferaw and Schafer (2008)
The design method proposed by Shifferaw and Shafer (2008) for the calculation of
inelastic bending capacity of cold-formed steel members based on test data and
numerical results at ambient temperature was found to give accurate results for cold-
formed steel lipped channel sections at ambient temperature. The same method was
used to calculate the inelastic bending capacity of lipped channel sections at elevated
temperatures with corresponding reduced yield strength and elastic modulus values.
Details of this design method are given in Section 6.3.1.2 of Chapter 6. Elastic local
buckling moment capacities were obtained using CUFSM analysis and are given in
Table B.1 of Appendix B. The predicted section moment capacities are given in
Table 7.3 while the summary of the ratio of FEA to predicted section moment
capacities is compared in Table 7.6. Sample calculations using the proposed design
method of Shifferaw and Schafer (2008) are presented in Appendix B for ambient
temperature. This design method predicts the inelastic bending capacity of cold-
formed steel lipped channel beams with an overall mean and COV of the ratio of the
FEA to predicted section moment capacities of 1.089 and 0.042, respectively. The
capacity reduction factor is 0.995, which is greater than the value of 0.95
recommended in AS/NZS 4600 for ambient temperature section moment capacity. It
seems that design methods in Eurocode 3 Part 1.3 (ECS, 2005) and Eurocode 3 Part
1.2 (ECS, 2005) are more accurate compared to this design method. However, the
advantage of this method is that there is no need to calculate the plate element
slenderness as for the other design methods, which is a lengthy procedure. Only the
elastic local buckling moments have to be calculated using one of the available finite
strip analysis programs such as CUFSM or Thin-Wall.
Table 7.6: Comparison of Section Moment Capacities

Obtained from FEA with Predicted Capacities using the Design Method
Proposed by Shifferaw and Schafer (2008)
Temp. oC 20 200 300 400 500 600 700 Overall

Mean 1.052 1.048 1.120 1.130 1.102 1.089 1.078 1.089
COV 0.013 0.025 0.050 0.039 0.018 0.030 0.030 0.042
φ 0.969 0.962 1.014 1.030 1.014 0.998 0.988 0.995
7.3.2 Compressive Strain
Although the design methods considered in this chapter predicts elevated

temperature section moment capacities accurately, careful consideration should be
made regarding the compressive strain development in the beams. Eurocode 3 Part
1.3 (ECS, 2006) design code does not specify any strain limit while AS/NZS 4600
(SA, 2005) specifies a maximum strain limit in the use of inelastic reserve capacity.
The maximum strain defined in AS/NZS 4600 depends on the b/t ratio and is less
than or equal to 3 x εy at ambient temperature. However, the cold-formed steel
members may sustain higher strains without fracturing and this needs to be further
investigated based on experimental study to enforce some kind of a strain limit both
at ambient and elevated temperatures.
The level of strain in the extreme compression fibre was investigated to study the
strain development in G250-1.55-120-45-15 beam with 100 mm effective span.
Figure 7.2 shows how the longitudinal membrane strain varied with the applied
moment at the point of maximum strain at different temperatures.
5.0
4.5
Applied Moment (kNm)
4.0 20
200
3.5
300
3.0 400
.
500
2.5 600
700
2.0
1.5
1.0
0.5
0.0
0.00 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.10
Strain
Figure 7.2: Longitudinal Membrane Strain Variation with Applied

Moment at the Point of Maximum Strain for Different Temperatures
G250-1.55-120-45-15 Beam
The maximum longitudinal membrane strains at the time of failure for one of the
lipped channel beams were obtained using the results of non-linear analyses at
different elevated temperatures. The calculated yield strain (εy) together with the
maximum membrane yield strain (εFEA) observed in the finite element analyses of
100 mm short beams at different temperatures are given in Table 7.7. The observed
maximum compressive membrane strain (εFEA) is always more than that observed at
ambient temperature. It is always more than 3 x εy. In the temperature range of 300
to 500oC it exceeds 15 x εy. Before coming to a conclusion of using the inelastic
reserve capacities at elevated temperatures, elevated temperature tests should be
undertaken to determine whether cold-formed steel beams can sustain the observed
maximum strains without any fracture. The section moment capacity of compact
cold-formed steel lipped channel members can always be assumed safely as equal to
the yield moment capacity at elevated temperatures as given in AS/NZS 4600 (AS,
2005) for ambient temperature design.
Table 7.7: Maximum Longitudinal Membrane Strains Observed in the

Finite Element Analyses of G250-1.55-120-45-15 Beam at Elevated
Temperatures
Yield Membrane
Temperature Mb/My Strain, εy Strain, εFEA εFEA/εy
(1x10-3) (1x10-3)
20 1.080 1.441 4.595 3.2
200 1.091 1.543 9.823 6.4
300 1.180 1.325 20.546 15.5
400 1.198 1.187 21.306 18.0
500 1.140 1.092 15.855 14.5
600 1.145 1.032 7.617 7.4
700 1.134 1.023 5.085 5.0
7.3.3 Use of Higher Strain Based Yield Strength
According to the FEA section moment capacity results in Table 7.3, some of the
beams have exceeded their full plastic moment capacities. Table 7.3 shows that
beams have exceeded their full plastic moment capacities, at 300oC, 400oC and
500oC in the case of G250 steels and at 600oC and 700oC in the case of G450 steels.
At elevated temperatures the stress-strain relationship of steels becomes more non-
linear. Since there is no well defined yield point in this case, the yield strength is
usually approximated either based on 0.2% proof stress (fy,0.2%) or based on different
levels of total strains such as 0.5%, 1.5% and 2.0%. The yield strengths based on
0.2% proof stress and 0.5% total strain give similar values for yield strengths. In
finite element modelling a strain hardening material model was used that represents
the true material behaviour at elevated temperatures. Therefore it is possible for
beam capacity to exceed the plastic moment capacity calculated based on an
approximate yield strength of 0.2% proof stress. In calculating the inelastic reserve
capacity using the available design methods, 0.2% proof stress was used as the yield
strength at elevated temperatures. Therefore their section moment capacity
predictions were also found to be considerably lower than the FEA results in the
cases when the plastic moment capacities based on 0.2% proof stress (Mp,0.2%) were
exceeded. To eliminate these problems, it may be desirable to use higher yield
strengths based on a higher total strain at elevated temperatures. As the temperature
increases ductility of steel also increases and this allows steel to reach higher strains.
This also indicates the possibility of allowing higher strain based yield strengths in
capacity calculations. However, it is very difficult to define a single strain limit for
all the elevated temperatures because the ductility and non-linearity in stress-strain
relationship of steels vary significantly with temperature and steel grades. The results
of this study also showed the inability of using higher strain levels for some
temperatures because the beam capacities are less than Mp,0.2% in most cases.
BS5950 Part 8 (BSI, 1990) recommends using yield strengths based on 1.5% total
strain for the fire design of hot-rolled steel beams. Therefore it was decided to
investigate the use of higher yield strengths based on 1.5% total strain in the strength
predictions for those which exceeded Mp,0.2%. Eurocode 3 Part 1.3 (ECS, 2006)
design method was selected since it gave the most accurate moment capacity
predictions compared to others. Table 7.8 gives the ratios of yield strengths based on
1.5% total strain to 0.2% proof stress. The use of yield strengths based on 1.5% total
strain increases the yield strength by about 30% for G250 steels at 300oC and 400oC
while it is about 18% at 500oC. For G450 steels the increase in yield strength is less
than 10% up to 400oC, but increases by about 20% at higher temperatures.
Table 7.8: Ratios of Yield Strength Based on 1.5% Total Strain and
0.2% Proof Stress
G250-1.55 G250-1.95 G450-1.5 G450-1.9

Temperature
fy,1.5/fy,0.2
fy,1.5/fy,0.2
fy,1.5/fy,0.2
fy,1.5/fy,0.2
C
fy,0.2
fy,1.5
fy,0.2
fy,1.5
fy,0.2
fy,1.5
fy,0.2
fy,1.5
o
20 293.5* 293.7 1.00 270.5* 270.6 1.00 537.1 538.2 1.00 514.5 521.0 1.01
200 274.5* 275.5 1.00 257.0* 261.5 1.02 534.5 555.1 1.04 510.1 531.6 1.04
300 187.9 253.3 1.35 196.4 259.1 1.32 508.7 547.0 1.08 483.8 527.7 1.09
400 144.9 187.7 1.30 147.7 186.0 1.26 373.6 402.7 1.08 362.3 392.7 1.08
500 94.9 113.0 1.19 95.8 111.0 1.16 193.1 234.9 1.22 197.0 233.1 1.18
600 61.3 74.2 1.21 54.0 61.3 1.14 59.0 78.8 1.34 56.4 71.0 1.26
700 37.4 40.5 1.08 34.2 36.8 1.08 33.0 39.4 1.19 34.2 39.8 1.16
*-Yield strengths obtained from well defined yield plateau.
The results of beams whose FEA section moment capacities (Mu) exceeded Mp,0.2%
were compared with the new moment capacities predicted based on the yield
strengths based on 1.5% total strain (Tables 7.9 (a) and (b)). When 0.2% proof stress
was used, the design code predictions are conservative by about 8% in most cases for
both G250 and G450 steel beams. However, when the yield strengths based on 1.5%
total strain were used, the design code predictions were unconservative by about
15% in the cases of G250 steel beams at 300oC, 400oC and 500oC. In the case of
G450 steel beams, the predictions were unconservative by about 15% for 600oC,
which was reduced to 2% (unconservative) for 700oC. In summary the use of yield
strengths based on 1.5% total strain in the section moment capacity calculations is
not recommended as it over-predicts the section moment capacities even for those
sections whose FEA capacities exceeded Mp,0.2%. Therefore it is recommended to
use 0.2% proof stress as the yield strength for all the temperatures as it predicts the
section moment capacities more accurately and also on the safe side.
Table 7.9: Comparison of Section Moment Capacity Predictions using
0.2% Proof Stress and 1.5% Total Strain Based Yield Strength
Temp. Msx -
Section Mu,FEA Msx-0.2% Mu/Msx-0.2% Mu/Msx-1.5%
(oC) 1.5%
300 2.200 2.081 2.566 1.057 0.857

G250-1.55-90-45-15 400 1.650 1.544 1.943 1.069 0.849
500 1.125 1.107 1.231 1.016 0.914
300 3.250 3.043 3.727 1.068 0.872
G250-1.55-120-45-15
400 2.395 2.260 2.829 1.060 0.847
300 3.610 3.221 4.314 1.121 0.837
G250-1.95-100-60-15 400 2.655 2.387 3.189 1.112 0.833
500 1.805 1.710 2.006 1.056 0.900
300 4.605 4.103 5.575 1.122 0.826
G250-1.95-125-50-15 400 3.325 2.980 4.115 1.116 0.808
500 2.230 2.104 2.555 1.060 0.873
300 5.250 4.743 6.557 1.107 0.801
G250-1.95-150-40-15 400 3.805 3.445 4.855 1.104 0.784
500 2.525 2.433 2.955 1.038 0.854
300 5.950 5.426 7.135 1.097 0.834
G250-1.95-150-60-17 400 4.255 4.039 5.310 1.053 0.801
500 3.030 2.902 3.379 1.044 0.897
Mean 1.076 0.846
COV 0.030 0.042
Φ 0.988 0.772
Temp.
Section Mu Msx-0.2% Msx - 1.5% Mu/Msx-0.2% Mu/Msx-1.5%
(oC)
600 0.555 0.513 0.684 1.082 0.811
G450-1.5-90-30-14
700 0.342 0.326 0.342 1.048 0.999
600 0.685 0.622 0.779 1.101 0.879
G450-1.9-80-40-15
700 0.428 0.396 0.437 1.081 0.979
600 1.100 1.012 1.269 1.087 0.867
G450-1.9-115-40-15
700 0.685 0.644 0.711 1.064 0.963
Mean 1.077 0.916
COV 0.016 0.074
Φ 0.992 0.819
7.4 Member Moment Capacity of Lipped Channel Beams Subject
to Lateral-torsional Buckling
The developed finite element model described in Chapter 5 was used to conduct an
extensive number of finite element analyses of simply supported cold-formed steel
lipped channel beams at elevated temperatures. The ultimate failure moments and the
elastic buckling moments were obtained from the finite element analyses at different
temperatures and are presented in Table 7.10. Figure 7.3 shows the lateral-torsional
buckling failures of the lipped channel beams at elevated temperatures of 300oC and
600oC.
At 300oC At 600oC
(a) Elastic Lateral-torsional Buckling
Failure in the negative direction
At 300oC At 600oC
(b) Ultimate Failure Modes
Figure 7.3: Elastic Lateral-torsional Buckling and Ultimate Failure

Modes of Cold-formed Steel Lipped Channel Beams at Elevated Temperatures
The ultimate member moment capacities are plotted in a non-dimensional format in
Figure 7.4 together with the results of the section moment capacities. Since the
selected cross-sections for the parametric study are compact type cross-sections at
each temperature, the section moment capacity is more than the yield moment
capacity. However, it is difficult to allow the use of inelastic reserve capacity of
beams at elevated temperatures without a thorough experimental study including the
level of maximum strains reached by them at elevated temperatures without any
fracture. Therefore the section moment capacity was taken as equal to the yield
moment capacity at corresponding temperature and this approach was used in
plotting the moment capacity results in a non-dimensional format. It can be seen that
the moment capacity results are scattered around the intermediate slenderness region,
but converge as the slenderness increases. The ultimate moment capacities of beams
with intermediate slenderness are reduced considerably relative to the elastic
buckling moment capacities due to the influence of residual stresses and initial
imperfections (Figure 7.4). As the beam slenderness increases the beam capacity is
approaching the elastic buckling moment capacity.
1.2 Elastic Buckling G250-20
G250-200 G250-300
.
1.0
G250-400 G250-500
G250-600 G250-700
0.8
G450-20 G450-200
M u/M y or M b/M y
0.6 G450-300 G450-400
G450-500 G450-600
0.4 G450-700
0.2
Plastic Range Inelastic Range Elastic Range

0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
Slenderness regions are marked according (My/Mo)0.5
to AS/NZS 4600 (SA, 2005)
Figure 7.4: Finite Element Analysis Results at Elevated Temperatures -

For All the Cross-sections
Capacities at Elevated Temperatures from FEA
(a) 1.55 mm Thick G250 Steels
Temp. Mo Mu My,T Temp. Length Mo Mu
Section My (kNm) Length (m) Section
(oC) (kNm) (kNm) (kNm) (oC) (m) (kNm) (kNm)
2.8939 20 1600 3.680 2.373 20 1400 6.518 3.462
20 1800 2.993 2.163 20 1600 5.103 3.192
20 2000 2.485 1.925 20 1800 4.105 2.877
20 2500 1.681 1.384 20 2000 3.380 2.540
20 3000 1.232 1.012 4.1869 20 2500 2.247 1.810
20 3500 0.957 0.768 20 3000 1.622 1.326
20 4000 0.774 0.634 20 3500 1.241 1.008
20 4500 0.647 0.532 20 4000 0.991 0.805
20 5000 0.555 0.458 20 4500 0.818 0.787
2.6334 200 1600 3.127 2.128 20 5000 0.693 0.665
200 1800 2.543 1.919 200 1400 5.539 3.125
200 2000 2.112 1.686 200 1600 4.336 2.855
200 2500 1.429 1.180 200 1800 3.488 2.540
200 3000 1.047 0.861 200 2000 2.872 2.214
200 3500 0.813 0.640 3.8101 200 2500 1.909 1.551
200 4000 0.658 0.539 200 3000 1.378 1.123
200 4500 0.550 0.451 200 3500 1.054 0.858
200 5000 0.471 0.388 200 4000 0.842 0.683
1.9030 300 1600 2.631 1.233 200 4500 0.695 0.564
300 1800 2.140 1.163 200 5000 0.589 0.479
300 2000 1.777 1.093 300 1400 4.661 1.765
300 2500 1.202 0.896 300 1600 3.649 1.675
300 3000 0.881 0.704 300 1800 2.936 1.574
300 3500 0.684 0.558 300 2000 2.417 1.383
300 4000 0.554 0.452 2.7533 300 2500 1.607 1.169
300 4500 0.463 0.470 300 3000 1.160 0.909
300 5000 0.397 0.326 300 3500 0.887 0.711
1.3823 400 1600 2.135 1.006 300 4000 0.709 0.634
G250-1.55-90-45-15
400 1800 1.736 0.959 300 4500 0.585 0.559

400 2000 1.442 0.901 300 5000 0.496 0.475
400 2500 0.975 0.733 400 1400 3.781 1.450
G250-1.55-120-45-15
400 3000 0.715 0.575 400 1600 2.960 1.338

400 3500 0.555 0.452 400 1800 2.382 1.293
400 4000 0.449 0.367 400 2000 1.960 1.203
400 4500 0.376 0.291 1.9999 400 2500 1.303 0.955
400 5000 0.322 0.263 400 3000 0.941 0.741
0.9761 500 1600 1.638 0.756 400 3500 0.720 0.578
500 1800 1.332 0.721 400 4000 0.575 0.464
500 2000 1.106 0.686 400 4500 0.475 0.386
500 2500 0.748 0.560 400 5000 0.402 0.326
500 3000 0.549 0.441 500 1400 2.901 1.090
500 3500 0.426 0.347 500 1600 2.271 1.041
500 4000 0.345 0.283 500 1800 1.827 0.962
500 4500 0.288 0.236 500 2000 1.504 0.885
500 5000 0.247 0.204 1.4122 500 2500 1.000 0.720
0.6427 600 1600 1.141 0.519 500 3000 0.722 0.565
600 1800 0.928 0.498 500 3500 0.552 0.443
600 2000 0.771 0.468 500 4000 0.441 0.356
600 2500 0.521 0.382 500 4500 0.364 0.294
600 3000 0.382 0.306 500 5000 0.308 0.250
600 3500 0.297 0.241 600 1400 2.021 0.738
600 4000 0.240 0.195 600 1600 1.582 0.714
600 4500 0.201 0.164 600 1800 1.273 0.665
600 5000 0.172 0.141 600 2000 1.048 0.611
0.3598 700 1600 0.644 0.297 0.9298 600 2500 0.697 0.493
700 1800 0.524 0.283 600 3000 0.503 0.393
700 2000 0.435 0.267 600 3500 0.385 0.308
700 2500 0.294 0.220 600 4000 0.307 0.248
700 3000 0.216 0.173 600 4500 0.254 0.206
700 3500 0.168 0.136 600 5000 0.215 0.174
700 4000 0.136 0.110 700 1400 1.141 0.425
700 4500 0.113 0.092 700 1600 0.893 0.398
700 5000 0.097 0.080 700 1800 0.719 0.383
Mo – Elastic lateral-torsional buckling moment 700 2000 0.592 0.354
0.5205 700 2500 0.393 0.283
Mu – Ultimate moment capacity 700 3000 0.284 0.223
700 3500 0.217 0.173
700 4000 0.174 0.141
My – Yield moment capacity 700 4500 0.143 0.116
700 5000 0.121 0.098
(b) 1.95 mm Thick G250 Steels
Mo Mu My Temp. Length Mo Mu
Section My (kNm) Temp. (oC) Length (m) Section
(kNm) (kNm) (kNm) (oC) (m) (kNm) (kNm)
20 1800 6.964 4.083 200 1400 8.530 4.255
20 2000 5.885 3.887 200 1600 6.714 3.979
20 2500 4.053 3.211 200 1800 5.424 3.651
20 3000 2.978 2.428 200 2000 4.481 3.296
4.6305
20 3500 2.305 1.788 200 2500 3.001 2.430
4.9607
20 4000 1.857 1.459 200 3000 2.180 1.786
20 4500 1.544 1.281 200 3500 1.679 1.379
20 5000 1.316 1.076 200 4000 1.349 1.107
200 1800 5.917 3.700 200 4500 1.136 0.921
200 2000 5.000 3.496 200 5000 0.954 0.785
200 2500 3.443 2.811 300 1400 7.179 2.416
200 3000 2.530 2.117 300 1600 5.651 2.298
4.2138
200 3500 1.958 1.628 300 1800 4.565 2.180
200 4000 1.578 1.263 300 2000 3.771 2.062
200 4500 1.312 1.085 300 2500 2.525 1.733
3.5848
200 5000 1.118 0.925 300 3000 1.835 1.392
300 1800 4.980 2.046 300 3500 1.413 1.131
300 2000 4.208 1.992 300 4000 1.136 0.923
300 2500 2.898 1.788 300 4500 0.943 0.779
300 3000 2.129 1.557 300 5000 0.803 0.661
3.0450
300 3500 1.648 1.308 400 1400 5.824 1.904
300 4000 1.328 1.076 400 1600 4.584 1.812
300 4500 1.104 0.907 400 1800 3.693 1.733
300 5000 0.941 0.772 400 2000 3.059 1.628
400 1800 4.040 1.637 400 2500 2.049 1.366
G250-1.95-100-60-15
2.6038
400 2000 3.414 1.583 400 3000 1.489 1.127
400 2500 2.351 1.441 400 3500 1.146 0.915
G250-1.95-125-50-15
400 3000 1.727 1.254 400 4000 0.921 0.746

2.2118
400 3500 1.337 1.050 400 4500 0.765 0.625
400 4000 1.077 0.872 400 5000 0.651 0.534
400 4500 0.896 0.737 500 1400 4.468 1.405
400 5000 0.764 0.626 500 1600 3.517 1.353
500 1800 3.100 1.210 500 1800 2.841 1.291
500 2000 2.619 1.156 500 2000 2.347 1.225
500 2500 1.804 1.076 500 2500 1.572 1.052
1.8386
500 3000 1.325 0.952 500 3000 1.142 0.860
1.5618
500 3500 1.026 0.807 500 3500 0.915 0.699
500 4000 0.827 0.672 500 4000 0.707 0.573
500 4500 0.687 0.564 500 4500 0.587 0.478
500 5000 0.586 0.479 500 5000 0.500 0.410
600 1800 2.160 0.847 600 1400 3.113 0.977
600 2000 1.825 0.820 600 1600 2.450 0.947
600 2500 1.257 0.757 600 1800 1.980 0.902
600 3000 0.923 0.666 600 2000 1.635 0.864
1.0284
600 3500 0.715 0.564 600 2500 1.095 0.733
1.2107
600 4000 0.576 0.470 600 3000 0.796 0.600
600 4500 0.479 0.392 600 3500 0.613 0.550
600 5000 0.408 0.334 600 4000 0.492 0.398
700 1800 1.219 0.487 600 4500 0.409 0.334
700 2000 1.030 0.475 600 5000 0.348 0.286
700 2500 0.710 0.430 700 1400 1.758 0.563
700 3000 0.521 0.381 700 1600 1.384 0.542
0.5757
700 3500 0.404 0.319 700 1800 1.118 0.517
700 4000 0.325 0.265 700 2000 0.923 0.486
700 4500 0.270 0.221 700 2500 0.618 0.410
0.6777
700 5000 0.230 0.188 700 3000 0.449 0.340
20 1400 10.039 4.701 700 3500 0.346 0.276
20 1600 7.902 4.439 700 4000 0.278 0.225
G250-1.95-125-50-15
20 1800 6.384 4.110 700 4500 0.231 0.188

20 2000 5.273 3.743 700 5000 0.197 0.160
20 2500 3.531 2.810
5.4513
20 3000 2.566 2.049
20 3500 1.976 1.628
20 4000 1.588 1.303
20 4500 1.318 1.083
20 5000 1.122 0.925
(b) 1.95 mm Thick G250 Steels (Contd.)
Mo Mu Mo Mu
Section My (kNm) Temp. (oC) Length (m) Section My (kNm) Temp. (oC) Length (m)
(kNm) (kNm) (kNm) (kNm)
20 1000 14.639 5.324 600 3000 0.610 0.483
20 1200 10.430 4.986 600 3500 0.471 0.383
20 1400 7.815 4.578 1.3591 600 4000 0.380 0.314
20 1600 6.090 4.099 600 4500 0.317 0.264
20 1800 4.894 3.602 600 5000 0.271 0.229
G250-1.95-150-40-15
20 2000 4.033 3.123 700 1000 2.563 0.580
6.1196
20 2500 2.698 2.218 700 1200 1.826 0.605
20 3000 1.966 1.638 700 1400 1.368 0.520
20 3500 1.519 1.276 700 1600 1.066 0.518
20 4000 1.226 1.042 700 1800 0.857 0.479
20 4500 1.021 0.889 700 2000 0.706 0.445
0.7608
20 5000 0.872 0.788 700 2500 0.473 0.353
200 1000 12.439 4.827 700 3000 0.344 0.273
200 1200 8.862 4.489 700 3500 0.266 0.216
200 1400 6.641 4.081 700 4000 0.215 0.175
200 1600 5.175 3.620 700 4500 0.179 0.149
200 1800 4.159 3.141 700 5000 0.153 0.129
200 2000 3.427 2.715 20 2000 10.081 6.533
5.5689
200 2500 2.293 2.129 20 2500 6.696 5.293
200 3000 1.670 1.397 20 3000 4.796 4.030
200 3500 1.291 1.088 7.7051 20 3500 3.631 3.056
200 4000 1.041 0.889 20 4000 2.867 2.392
200 4500 0.868 0.763 20 4500 2.339 1.940
200 5000 0.741 0.726 20 5000 1.958 1.617
300 1000 10.468 2.768 200 2000 8.566 5.868
300 1200 7.459 2.591 200 2500 5.749 4.628
300 1400 5.589 2.396 200 3000 4.075 3.455
300 1600 4.355 2.254 7.0116 200 3500 3.085 2.591
300 1800 3.500 2.076 200 4000 2.436 2.035
G250-1.95-150-40-15
300 2000 2.884 1.810 200 4500 1.987 1.648

4.0242
300 2500 1.930 1.460 200 5000 1.664 1.373
300 3000 1.406 1.121 300 2000 7.209 3.300
300 3500 1.086 0.891 300 2500 4.788 2.967
300 4000 0.876 0.729 300 3000 3.429 2.525
300 4500 0.730 0.614 5.0668 300 3500 2.596 2.062
300 5000 0.624 0.531 300 4000 2.050 1.676
400 1000 8.492 2.147 300 4500 1.672 1.380
400 1200 6.051 1.970 300 5000 1.400 1.152
G250-1.95-150-60-17
400 1400 4.534 1.881 400 2000 5.848 2.635

400 1600 3.533 1.766 400 2500 3.884 2.325
400 1800 2.839 1.620 400 3000 2.782 2.044
400 2000 2.339 1.489 3.6803 400 3500 2.106 1.674
2.9230
400 2500 1.565 1.176 400 4000 1.663 1.360
400 3000 1.140 0.907 400 4500 1.357 1.118
400 3500 0.881 0.720 400 5000 1.136 0.932
400 4000 0.711 0.591 500 2000 4.487 1.980
400 4500 0.592 0.497 500 2500 2.980 1.780
400 5000 0.506 0.429 500 3000 2.135 1.544
500 1000 6.516 1.592 2.5988 500 3500 1.616 1.273
500 1200 4.642 1.508 500 4000 1.276 1.041
500 1400 3.479 1.423 500 4500 1.041 0.855
500 1600 2.711 1.334 500 5000 0.871 0.715
500 1800 2.178 1.230 600 2000 3.126 1.377
500 2000 1.795 1.121 600 2500 2.076 1.200
2.0640
500 2500 1.201 0.880 600 3000 1.487 1.083
500 3000 0.875 0.696 1.7112 600 3500 1.126 0.890
500 3500 0.676 0.550 600 4000 0.889 0.724
500 4000 0.545 0.451 600 4500 0.725 0.596
500 4500 0.455 0.380 600 5000 0.607 0.498
500 5000 0.388 0.328 700 2000 1.765 0.782
600 1000 4.540 1.102 700 2500 1.172 0.711
600 1200 3.234 1.054 700 3000 0.840 0.618
600 1400 2.424 0.999 0.9579 700 3500 0.636 0.505
1.3591 600 1600 1.888 0.937 700 4000 0.502 0.410
600 1800 1.518 0.864 700 4500 0.410 0.337
600 2000 1.251 0.788 700 5000 0.343 0.281
600 2500 0.837 0.749
(c) 1.5 mm Thick G450 Steels
Mo Mu
Section My (kNm) Temp. (oC) Length (m)
(kNm) (kNm)
20 1000 3.596 2.457
20 1200 2.574 1.994
20 1400 1.945 1.564
3.9944 20 1600 1.532 1.268
20 1800 1.246 1.027
20 2000 1.040 0.865
20 2500 0.719 0.611
20 3000 0.541 0.680
200 1000 3.055 2.142
200 1200 2.187 1.707
200 1400 1.653 1.356
3.9560 200 1600 1.302 1.073
200 1800 1.059 0.874
200 2000 0.883 0.736
200 2500 0.611 0.560
300 1000 2.571 1.911
300 1200 1.840 1.485
300 1400 1.391 1.143
3.7947 300 1600 1.095 0.907
300 1800 0.891 0.736
300 2000 0.744 0.620
300 2500 0.514 0.523
G450-1.5-90-30-14
400 1000 2.086 1.457

400 1200 1.493 1.161
400 1400 1.128 0.921
2.7725 400 1600 0.889 0.731
400 1800 0.723 0.597
400 2000 0.603 0.500
400 2500 0.417 0.409
500 1000 1.600 0.967
500 1200 1.146 0.814
500 1400 0.866 0.671
1.5605 500 1600 0.682 0.551
500 1800 0.555 0.451
500 2000 0.463 0.382
500 2500 0.320 0.267
600 1000 1.115 0.331
600 1200 0.798 0.316
600 1400 0.603 0.304
0.4394 600 1600 0.475 0.279
600 1800 0.386 0.250
600 2000 0.322 0.228
600 2500 0.223 0.175
700 1000 0.630 0.217
700 1200 0.451 0.202
700 1400 0.341 0.185
0.2796 700 1600 0.268 0.167
700 1800 0.218 0.148
700 2000 0.182 0.134
700 2500 0.126 0.100
(d) 1.9 mm Thick G450 Steels
Mo Mu Mo Mu
Section My (kNm) Temp. (oC) Length (m) Section My (kNm) Temp. (oC) Length (m)
(kNm) (kNm) (kNm) (kNm)
20 1200 5.485 3.882 700 2000 0.402 0.240
20 1400 4.211 3.370 700 2500 0.279 0.204
G450-1.9-80-
20 1600 3.347 2.825 700 3000 0.211 0.170
40-15
20 1800 2.738 2.346 0.3475 700 3500 0.169 0.142
20 2000 2.293 1.962 700 4000 0.140 0.120
4.9644 20 2500 1.596 1.367 700 4500 0.120 0.105
20 3000 1.206 1.037 700 5000 0.104 0.092
20 3500 0.963 0.834 20 1200 8.193 5.317
20 4000 0.799 0.698 20 1400 6.176 4.621
20 4500 0.683 0.603 20 1600 4.839 3.926
20 5000 0.596 0.533 20 1800 3.910 3.230
200 1200 4.660 3.527 7.9442 20 2000 3.239 2.687
200 1400 3.578 2.949 20 2500 2.197 1.839
200 1600 2.844 2.354 20 3000 1.622 1.363
200 1800 2.326 1.999 20 3500 1.269 1.067
200 2000 1.949 1.664 20 4000 1.035 0.878
4.9167 200 2500 1.356 1.161 200 1200 6.961 5.012
200 3000 1.024 0.880 200 1400 5.247 4.164
200 3500 0.818 0.706 200 1600 4.112 3.383
200 4000 0.679 0.591 200 1800 3.322 2.782
200 4500 0.580 0.512 7.8678 200 2000 2.752 2.296
200 5000 0.507 0.450 200 2500 1.867 1.563
300 1200 3.922 3.155 200 3000 1.378 1.153
300 1400 3.012 2.577 200 3500 1.078 0.907
300 1600 2.393 2.065 200 4000 0.880 0.866
300 1800 1.958 1.681 300 1200 5.858 4.412
300 2000 1.640 1.400 300 1400 4.416 3.554
4.7162 300 2500 1.141 0.950 300 1600 3.460 2.868
300 3000 0.862 0.739 300 1800 2.796 2.344
7.5470
300 3500 0.688 0.595 300 2000 2.316 1.944
300 4000 0.572 0.500 300 2500 1.570 1.315
G450-1.9-80-40-15
300 4500 0.488 0.430 300 3000 1.160 0.972

300 5000 0.426 0.380 300 3500 0.907 0.763
400 1200 3.182 2.329 400 1200 4.752 3.106
G450-1.9-115-40-15
400 1400 2.443 1.958 400 1400 3.583 2.782

400 1600 1.942 1.648 400 1600 2.807 2.258
400 1800 1.588 1.359 400 1800 2.268 1.887
400 2000 1.330 1.132 5.5140 400 2000 1.879 1.572
3.4458 400 2500 0.926 0.756 400 2500 1.274 1.067
400 3000 0.699 0.599 400 3000 0.941 0.786
400 3500 0.558 0.479 400 3500 0.736 0.617
400 4000 0.464 0.403 400 4000 0.601 0.587
400 4500 0.396 0.348 500 1200 3.646 2.077
400 5000 0.346 0.307 500 1400 2.749 1.858
500 1200 2.441 1.404 500 1600 2.154 1.620
500 1400 1.874 1.313 500 1800 1.740 1.382
3.1036
500 1600 1.490 1.156 500 2000 1.442 1.163
500 1800 1.219 1.004 500 2500 0.978 0.812
1.9395 500 2000 1.021 0.863 500 3000 0.722 0.602
500 2500 0.710 0.603 500 3500 0.565 0.472
500 3000 0.537 0.458 600 1200 2.541 0.734
500 3500 0.428 0.368 600 1400 1.915 0.702
500 4000 0.356 0.308 600 1600 1.501 0.670
600 1200 1.701 0.487 0.8739 600 1800 1.212 0.628
600 1400 1.306 0.475 600 2000 1.004 0.595
600 1600 1.038 0.454 600 3000 0.503 0.392
600 1800 0.849 0.438 600 3500 0.393 0.319
600 2000 0.711 0.417 700 1200 1.435 0.444
0.5461 600 2500 0.495 0.361 700 1400 1.081 0.422
600 3000 0.374 0.301 700 1600 0.847 0.394
600 3500 0.298 0.252 700 1800 0.685 0.365
600 4000 0.248 0.213 700 2000 0.567 0.337
600 4500 0.212 0.185 0.5561 700 2500 0.385 0.275
600 5000 0.185 0.163 700 3000 0.284 0.221
700 1200 0.960 0.304 700 3500 0.222 0.180
700 1400 0.738 0.286 700 4000 0.181 0.150
0.3475
700 1600 0.586 0.271 700 4500 0.153 0.128
700 1800 0.479 0.256 700 5000 0.132 0.111
Figure 7.5 shows the ultimate moment capacity results separately at different
temperatures for the lateral-torsional buckling failure of all the nine selected
sections. As seen in Figure 7.5 the results are plotted as a set of buckling curves for
each temperature. In the case of G450 steels, the results for 20oC are located above
other results while those for 700oC provide the lower bound. The results for 20oC
provide the upper bound while those for 300oC provide the lower bound for G250
steels. Therefore it can be concluded that the ultimate member moment capacity
results are not dependent purely on the temperature and the reasons for this kind of
behaviour needs to be investigated. The reasons for the variations seen in the
intermediate slenderness region are likely to be due to the higher non-linearity in
stress-strain curves at different elevated temperatures and the varying ratios of yield
strength to elastic modulus, which need to be investigated.
20
.
200
300
1.0
400
500
600
0.8 700
M u/M y or M b/M y
0.6
0.4
0.2
0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
(My/Mo)0.5
(a) G250-1.55-90-45-15 Section

20
.
200
300
1.0
400
500
600
0.8 700
M u/M y or M b/M y
0.6
0.4
0.2
0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
(My/Mo)0.5
(b) G250-1.55-120-45-15 Section
Figure 7.5: Finite Element Analysis Results at Elevated Temperatures
20
.
200
300
1.0
400
500
600
0.8 700
M u/M y or M b/M y
0.6
0.4
0.2
0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
(My/Mo)0.5
(b) G250-1.95-100-60-15 Section

20
.
200
300
1.0
400
500
600
0.8 700
M u/M y or M b/M y
0.6
0.4
0.2
0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
0.5
(My/Mo)
(c) G250-1.95-125-50-15 Section

20
.
200
300
1.0
400
500
600
0.8 700
M u/M y or M b/M y
0.6
0.4
0.2
0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
0.5
(My/Mo)
(d) G250-1.95-150-40-15 Section
20
.
200
300
1.0
400
500
600
0.8 700
M u/M y or M b/M y
0.6
0.4
0.2
0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
0.5
(My/Mo)
(e) G250-1.95-150-60-17 Section

20
.
200
300
1.0
400
500
600
0.8 700
M u/M y or M b/M y
0.6
0.4
0.2
0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
(My/Mo)0.5
(f) G450-1.5-90-30-14 Section

20
200
1.0
300
.
400
0.8
500
600
0.6 700
0.4
M u/M y or M b/M y
0.2
0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
(My/Mo)0.5
(g) G450-1.9-80-40-15 Section
20
.
200
300
1.0
400
500
600
0.8 700
M u/M y or M b/M y
0.6
0.4
0.2
0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
0.5
(My/Mo)
(h) G450-1.9-115-40-15 Section
7.4.1 Effect of Elevated Temperatures
As the temperature increases the yield strength and elastic modulus decrease and as
expected the lateral-torsional buckling capacity also decreases. Figure 7.6 shows
how the lateral-torsional buckling capacity reduces with increasing temperature for
Grade 250 and 450 steel sections. It can be seen that at 300oC the member moment
capacity of G250 steel beam reduces considerably at lower lengths and as the length
increases this strength reduction decreases. On the other hand for Grade 450 steels
the member moment capacity of steel beams at temperatures above 300oC is
considerably high and shows different behaviour relative to G250 steels. Beyond
400oC the member moment capacity of G450 steel beams has decreased
considerably.
3.0
.
20
200
2.5
300
400
2.0 500
600
M u (kNm)
1.5 700
1.0
0.5
0.0
500 1000 1500 2000 2500 3000 3500 4000 4500 5000
Span (mm)
(a) G250-1.55-90-45-15 Section
5.0
.
20
4.5
200
4.0 300
400
3.5
500
3.0 600
M u (kNm)
2.5 700
2.0
1.5
1.0
0.5
0.0
500 1000 1500 2000 2500 3000 3500 4000 4500 5000
Span (mm)
(b) G250-1.95-125-50-15 Section

Capacity
3.0 20
,
200
300
2.5
400
500
2.0
600
700
1.5
M u (kNm)
1.0
0.5
0.0
500 1000 1500 2000 2500
Span (mm)
(c) G450-1.5-90-30-14 Section

5.0 20
,
200
4.5
300
4.0 400
3.5 500
600
3.0
700
2.5
M u (kNm)
2.0
1.5
1.0
0.5
0.0
500 1000 1500 2000 2500 3000 3500 4000 4500 5000
Span (mm)
(d) G450-1.9-80-40-15 Section

Capacity
Figure 7.7 shows the relative reduction in ultimate moment capacities at different
temperatures for G250 and 450 steel beams in comparison to the ambient
temperature ultimate moment capacity (Mu,T/Mu,20). At 200oC both G250 and 450
steel beams have about 90% of their ambient temperature moment capacity. The
ultimate moment capacity of G250 steel beams at 300oC has reduced almost by 50%
at lower lengths but is about 70% of the ambient temperature moment capacity at
higher lengths. In contrast to G250 steel beams at 300oC, G450 steel beams retain
85% of the ambient temperature moment capacity at lower lengths while their
capacity is about 70% at higher lengths. The reason for this kind of behaviour will be
discussed in the next sections. Beams of both grades of steel with longer spans retain
about 45% of its ambient temperature moment capacity at 500oC, which is
considerably high. At 700oC the moment capacities have reduced considerably and
are about 18% on average for both G250 and G450 steel beams. These results
indicate that the Eurocode 3 Part 1.2 (ECS, 2005) temperature limit of 350 oC is too
conservative for cold-formed steel beams.
1.0
0.9
.
0.8 200
0.7 300
0.6
M u,T /M u,20
400
0.5
500
0.4
600
0.3
700
0.2
0.1
0.0
1500 2000 2500 3000 3500 4000 4500 5000
Span (mm)
(a) G250-1.55-90-45-15 Section

1.0
0.9
.
200
0.8
0.7 300
0.6
400
M u,T /M u,20
0.5
500
0.4
0.3 600
0.2
700
0.1
0.0
500 1000 1500 2000 2500 3000 3500 4000 4500 5000
Span (mm)
(b) G450-1.9-80-45-15 Section

Figure 7.7: Reduction in Moment Capacities at Elevated Temperatures
Relative to Ambient Temperature Moment Capacity
7.4.2 Effect of Steel Grade
The effects of steel grade, in other words, the yield strength, elastic modulus and the
stress-strain relationship, were investigated using Figure 7.8 where the moment
capacity results for G250 and G450 steel beams at different temperatures are plotted
separately. The results for G250 steel beams seem to be more scattered in the
intermediate slenderness range in comparison to G450 steel beams. The reasons for
this kind of behaviour are discussed in the next sections. The moment capacity data
points for ambient temperature are plotted as the upper bound for both G250 and
G450 steel beams while those for 300oC in the case of G250 steel beams and 600oC
and 700oC in the case of G450 steel beams are plotted as the lower bound.
1.2
Elastic Buckling
.
1.0 G250-20
G250-200
0.8 G250-300
G250-400
M u/M y or M b/M y
0.6 G250-500
G250-600
0.4 G250-700
0.2
0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
(My/Mo)0.5
Figure 7.8: Effect of Steel Grade
G450-20
.
1.0 G450-200
G450-300
0.8 G450-400
G450-500
M u/M y or M b/M y
0.6 G450-600
G450-700
0.4
0.2
0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
(My/Mo)0.5
Figure 7.8: Effect of Steel Grade
7.4.3 Effect of Non-linear Stress-strain curves
Low grade steel has a well defined yield point at ambient temperature and therefore
elastic-perfect plastic stress-strain relationship was used in FEA. Low grade steel
maintains this type of stress-strain relationship up to 200oC and becomes non-linear
thereafter (Curve B of Figure 7.9). The stress-strain relationship of high grade steel
is gradual yielding type at ambient temperature but is very close to an elastic-perfect-
plastic model (Curve A of Figure 7.9). This stress-strain relationship changes with
increasing temperature and becomes non-linear (Curve B of Figure 7.9). It is
possible to assess the level of non-linearity of stress-strain relationship by comparing
the ratio of the limit of proportionality, fp (up to the point the curve is linear) to the
0.2% proof stress, fy. Proportional limit of G250 steels up to 200oC was assumed to
be equal to their yield strengths although there is a small difference. For G250 steels
above 200oC and G450 steel for all the temperatures fp,T values were obtained by
plotting the stress-strain curve using Equations 7.4 and 7.5 with calculated values of
yield strength and elastic modulus at elevated temperatures. Figure 7.10
demonstrates the method of determining the fp,T value at elevated temperatures.
Figure 7.11 shows how the fpT/fyT ratio changes with increasing temperature for
different grades of steels while Table 7.11 gives the values of fpT/fyT ratio at different
temperatures.
Stress, σ
Curve A
T=T1=20oC
fy1
fp1 Curve B
T=T2
fy2
fp2
fp1/fy1 > fp2/fy2
Strain, ε
Figure 7.9: Non-linearity in Stress-strain Graphs
250
200
fy,300
Stress (MPa)
150
Stress-strain curve plotted using
Equations 7.4 and 7.5 with
calculated yield strength and elastic
100
fp,300 modulus at elevated temperatures
50
0
0.000 0.002 0.004 0.006 0.008 0.010
Strain
Figure 7.10: Determining the fp,T Value at 300oC
The fp,T/fy,T ratio of G250 steels is constant at 1.0 up to 200oC since the stress-strain
graph has a well defined yield point. This ratio of G250 steels suddenly drops to 0.48
on average at 300oC, which means the material is linear up to only half of the
measured yield stress and thereafter it is non-linear. As the temperature increases
beyond 300oC, the fp,T/fy,T ratio also increases until it reaches 700oC. The lowest
fp,T/fy,T ratio occurs at 300oC for G250 steels, which is considerably low relative to
others. This is also reflected in the ultimate moment capacity results from FEA and
the corresponding data points are also plotted well below others in Figure 7.4. The
fp,T/fy,T ratio decreases for G450 steels with increasing temperature except at 600oC.
There is not much variation of fp,T/fy,T ratio for G450 steels as seen in G250 steels.
1.2
1.0
0.8
f p,T / f y,T
0.6 G250 - 1.55

G250 - 1.95
G450 - 1.5
0.4 G450 - 1.9
0.2
0.0
0 100 200 300 400 500 600 700 800
o
Temperature, C
Figure 7.11: Variation of fp,T/fy,T Ratio with Elevated Temperatures
Table 7.11: fp,T/fy,T Ratios of G250 and G450 steels at Elevated

Temperatures
Temp. fp,T/fy,T Average fp,T/fy,T
o
C
G250-1.95 G250-1.55 G450-1.9 G450-1.5 G450 G250
20 1.000 1.000 0.913 0.933 0.923 1.000
100 1.000 1.000 0.893 0.904 0.898 1.000
200 1.000 1.000 0.863 0.848 0.856 1.000
300 0.449 0.517 0.838 0.826 0.832 0.483
400 0.579 0.570 0.783 0.805 0.794 0.575
500 0.711 0.686 0.795 0.788 0.792 0.698
600 0.864 0.766 0.812 0.846 0.829 0.815
700 0.890 0.848 0.777 0.825 0.801 0.869
Figure 7.11 shows a different relationship between the fp,T/fy,T ratios of G250 and
G450 steels. This difference is also reflected in the ultimate moment capacities
obtained from the finite element analyses of G250 and G450 steel beams in Figures
7.4 and 7.8 for different temperatures. The difference could be seen in the
intermediate beam slenderness region. The moment capacity data points are scattered
in the case of G250 steel beams, which have a larger variation in the fp,T/fy,T ratio
compared to G450 steels. It could be seen from Figure 7.4 that the results of non-
dimensional moment capacities are the lowest at 300oC for G250 steels, at which it
has the lowest fp,T/fy,T ratio. As the fp,T/fy,T ratio increases the moment capacity data
points plot higher in the intermediate slenderness range. Therefore the effect of
fp,T/fy,T ratio on the ultimate moment capacity is clearly evident.
The influence of non-linear stress-strain relationship at elevated temperatures was

studied by carrying out finite element analyses using the strain hardening material
model, which represents the true material behaviour, and the elastic-perfect plastic
model with 0.2% proof stress at 300oC for both G250 and G450 steels. The ultimate
moment capacities obtained based on these two types of stress-strain relationships
are compared in Figure 7.12 (a) and (b) for G250 and G450 steel beams,
respectively. The ultimate moment capacities dropped when the actual material
behaviour is used in the intermediate slenderness range in comparison to elastic-
perfect plastic material model for both G250 and G450 steel beams. A large
reduction is seen in the case of Grade 250 steel beams, which has a very low fp,T/fy,T
ratio whereas the difference is very small in the case of G450 steel beam, which has
a considerably high fp,T/fy,T ratio. This behaviour could be further explained using
Figure 7.13 in which the strain hardening and the elastic-perfect material models are
illustrated.
Elastic Perfect Plastic Material Model
1.0 Strain Hardening Material Model

.
0.8
M u/M y or M b/M y
0.6
0.4
0.2
0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
(My/Mo)0.5
(a) G250 Steel Beam
Elastic Perfect Plastic Material Model
1.0 Strain Hardening Material Model

.
0.8
M u/M y or M b/M y
0.6
0.4
0.2
0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
(My/Mo)0.5
(b) G450 Steel Beam
Figure 7.12: Effect of Stress-strain Relationship
It can be seen that if the stress-strain curve is non-linear the tangent modulus of steel
is changing considerably with the stress level of steel member (Figure 7.13). Beyond
the elastic region the change in tangent modulus is directly affecting the beam
stiffness and therefore reduces the lateral-torsional buckling capacity considerably.
Therefore it is recommended that the stress-strain relationship at elevated
temperatures is investigated in detail before adopting the elastic-perfect plastic
material model over strain hardening material model for non-linear analyses. If the
stress-strain relationship is highly non-linear, the use of elastic-perfect plastic
material model will lead to unsafe strength predictions. This study was carried out
with strain hardening material model due to the presence of highly nonlinear stress-
strain relationship at elevated temperatures.
Stress, σ Actual stress-strain relationship
fy,T
E4 Assumed elastic-perfect plastic stress-strain relationship
E3
E2
E1 > E2 > E3 > E4
fp,T
fy,T – 0.2% proof stress at elevated temperatures
fp,T –limit of proportionality at elevated temperatures
E1
Strain, ε
Figure 7.13: Strain Hardening and Elastic-perfect Plastic Material

Models
Recently, Dharma and Tan (2007) discussed the effect of temperature related non-
linear stress-strain relationship on the lateral-torsional buckling capacity of hot-
rolled steel I-beams. They stated that the tangent modulus is very important to
consider, especially in the intermediate slenderness range as the ultimate failure may
occur beyond the limit of proportionality. They presented a new design approach by
taking into account the non-linear stress-strain relationship in terms of fp,T/fy,T ratio at
7.4.4 Effect of ky,T/kE,T Ratio
As the temperature increases the yield strength and elastic modulus of steels reduce
at different rates. The yield strength reduction rate is different from that of elastic
modulus reduction rate in relation to the ambient temperature mechanical properties.
Therefore the ratio of yield strength and elastic modulus reduction factors also varies
with increasing temperature. Figure 7.14 shows how the ky,T/kE,T ratio, where ky,T and
kE,T are the yield strength and elastic modulus reduction factors at temperature T,
respectively, varies with temperature. Table 7.12 shows the ky,T/kE,T ratios together
with the corresponding reduction factors for yield strength and elastic modulus. The
ky,T/kE,T ratio of G250 steels increases initially up to 200oC and thereafter reduces
while the same ratio of G450 steels increases up to 300oC and thereafter reduces up
to 600oC. At 700oC ky,T/kE,T ratio of G450 steels increases slightly in comparison to
that at 600oC.
1.4
1.2
G250
G450
1.0
0.8
k y,T/k E,T
0.6
0.4
0.2
0.0
0 100 200 300 400 500 600 700 800
o
Temperature, C
Figure 7.14: Variation of ky,T/kE,T Ratio at Elevated Temperatures
Table 7.12: Yield Strength and Elastic Modulus Reduction Factors and
their Ratios
Temperature oC 20 200 300 400 500 600 700
G250 ky,T = fy,T/fy,20 1.000 0.910 0.658 0.478 0.337 0.222 0.124
G450 ky,T = fy,T/fy,20 1.000 0.990 0.950 0.694 0.391 0.110 0.070
G250 & G450 kE,T = ET/E20 1.000 0.850 0.715 0.580 0.445 0.310 0.175
G250 ky,t/kE,T 1.000 1.071 0.920 0.823 0.758 0.716 0.710
G450 ky,t/kE,T 1.000 1.166 1.328 1.197 0.878 0.355 0.400
Since the non-dimensional member slenderness (λLT,fi=√(My,T/Mo,T)) is a function of

fy,T/ET, the variation of their ratios influence the beam slenderness. In order to study
the effect of variations of this ratio (ky,T/kE,T) on the non-dimensional member
slenderness (λLT,fi) obtained from FEA for the two beams of G250 and G450 steels
with different lengths are plotted against increasing temperature in Figure 7.15. It is
interesting to note that the non-dimensional member slenderness for a particular span
which directly affects the moment capacity of the beams subjected to lateral-
torsional buckling does not reduce as the temperature increases. Instead it increases
up to 300oC and then reduces up to 600oC in the case of G450 steel. Again it
increases slightly when temperature reaches 700oC. The non-dimensional member
slenderness of G250 steel beam increases up to 200oC and is thereafter gradually
reducing. This is due to the variation of the ratio of yield strength, fy,T to elastic
modulus, ET at elevated temperature, ie. the ratio of yield strength reduction factor
(kyT=fy,T/fy,20) to elastic modulus reduction factor (kET=ET/E20) (see Table 7.12 and
Figure 7.14). The non-dimensional member slenderness parameter closely follows
the variation of kyT/kET ratio. As shown in Figure 7.15, the lowest non-dimensional
slenderness value occurs at 600oC for G450 steels, and this temperature corresponds
to the lowest value of kyT/kET ratio. The highest non-dimensional slenderness for
G250 steel beam occurs at 200oC, which has the highest ky,T/kE,T ratio.
. 2.5
2.0
1200 mm
1400 mm
1600 mm
0.5
(M y,T/M o,T)
1.5 1800 mm
2000 mm
2500 mm
3000 mm
1.0 3500 mm
4000 mm
4500 mm
5000 mm
0.5
0.0
0 100 200 300 400 500 600 700 800
o
Temperature, C
(a) G250-1.55-90-45-15
.
3.5
3.0
1200 mm
2.5
1400 mm
1600 mm
2.0
1800 mm
0.5
(M y,T/M o,T)
2000 mm
1.5
2500 mm
1.0 3000 mm
3500 mm
0.5
0.0
0 100 200 300 400 500 600 700 800
o
Temperature, C
(b) G450-1.9-1154015
Figure 7.15: Variation of Non-Dimensional Member Slenderness with

Temperature
Figures 7.16 (a) and (b) display the results obtained from the non-linear finite
element analyses of G250 and G450 beams with different lengths under various
temperatures, respectively. The non-dimensionalised ultimate member moment
capacity ratio (i.e. Mb,T/My,T) is plotted against the span of beams for various elevated
temperatures. The maximum non-dimensionalised ultimate member moment
capacity ratio is achieved at 600oC which has the lowest kyT/kET ratio while the
minimum non-dimensionalised ultimate member moment capacity is achieved at
300oC which has the highest kyT/kET ratio in the case of G450 steel. At longer beam
spans, the non-dimensional ultimate member moment capacity ratio of beams is
increasing as the ky,T/kET ratio is decreasing. This order of member moment capacity
curves is seen, especially for very long beams and the trend changes for beams with
intermediate spans. In the case of G250 steels this behaviour can be clearly seen.
However, clear change in trend is seen for intermediate spans in Figure 7.16 (b). The
trend was changed in comparison to 20oC and 200oC, which had stress-strain
relationships with a well defined yield point and for others, which had a non-linear
stress-strain relationship. The reason for this is due to increasing non-linearity of
stress-strain relationship at elevated temperatures reducing the member moment
capacity and this behaviour was discussed in detail in the previous section. Finally it
can be concluded that for a particular span the member moment capacity achievable
relative to the section moment capacity is dependent on the kyT/kET ratio. However,
this behaviour can be altered if the non-linearity in the stress-strain relationship
becomes important.
7.4.5 Effect of Residual Stresses
The residual stresses decrease with increasing temperature and vanish after 700oC.
Effects of residual stresses at ambient temperature was found to be negligibly small
as demonstrated in Section 5.2.6 of Chapter 5 by comparing the ultimate member
moment capacity results with and without residual stresses. Therefore the effect of
residual stresses was not studied in detail at elevated temperatures and is considered
to be negligibly small as observed at ambient temperature.
1.2 20
200
300
1.0
400
500
0.8
600
700
0.6
M u/M y
0.4
0.2
0.0
500 1000 1500 2000 2500 3000 3500 4000 4500 5000
Span (mm)
(a) G450-1.9-80-40-15
1.0
20
0.9
200
0.8 300
400
0.7
500
0.6 600
0.5 700
M u/M y
0.4
0.3
0.2
0.1
0.0
500 1000 1500 2000 2500 3000 3500 4000 4500 5000
Span (mm)
(b) G250-1.55-90-45-15
Figure 7.16: Effect of ky,T/kE,T Ratio on the Non-Dimensional Member

Moment Capacity
7.5 Comparison of Ultimate Member Moment Capacities with
Predictions from the Current Design Rules
In the last chapter the accuracy of available ambient temperature design guidelines
(AS/NZS 4600 (SA, 2005), DSM, BS 5950 Part 5 (BS, 1998) and Eurocode 3 Part
1.3 (ECS, 2006)) was checked by comparing with the results obtained from finite
element analysis. These design rules may be used for elevated temperature design by
replacing the ambient temperature mechanical properties with those of elevated
temperatures. However, the accuracy of such an approach needs to be verified.
AS/NZS 4600, DSM and BS5950 Part 5 were found to be giving unsafe predictions
for lateral-torsional buckling at ambient temperature and therefore their accuracy
was not checked for elevated temperature conditions. Eurocode 3 Part 1.3
predictions at ambient temperature were overconservative in the intermediate
slenderness region but gave accurate results in the higher slenderness region.
Eurocode 3 Part 1.3 directly refers to Eurocode 3 Part 1.1 which gives different
imperfections factors. This research recommended buckling curve ‘a’ for the design
of cold-formed steel beams subject to lateral-torsional buckling at ambient
temperature.
Three options for a new design method were proposed based on AS/NZS 4600 for
lipped channel beams subject to lateral-torsional buckling at ambient temperature
and are presented in Chapter 6. Option 3 was selected as the best design method over
other two options and was used in this chapter. Eurocode 3 Part 1.2 provides design
guidelines for elevated temperature design of hot-rolled steel members and also
recommends using them for cold-formed steel members. The accuracy of the
predictions using the design methods given in Eurocode 3 Part 1.3/Part 1.1,
Eurocode 3 Part 1.2 and the new developed equations is assessed by comparing their
predicted member moment capacities with the finite element analysis results
obtained at different elevated temperatures. Dharma and Tan’s (2007) design method
proposed for hot-rolled beams was also assessed for cold-formed steel beams under
fire conditions. Appendix C presents the sample calculations for the member
moment capacities using different design methods at 400oC. Predicted member
moment capacities using all these design methods are given in Appendix G together
with finite element analysis results.
7.5.1 Eurocode 3 Part 1.3 / Eurocode 3 Part 1.1
The ambient temperature design code for cold-formed steel members, Eurocode 3
Part 1.3, refers to the design equations given in the design code for hot-rolled steel
members, Eurocode 3 Part 1.1, to design for the lateral-torsional buckling capacity of
cold-formed steel beams using buckling curve ‘b’. These design equations are given
in Section 2.9.2 of Chapter 2. It was found that the use of buckling curve ‘b’ gives
conservative predictions for the intermediate slenderness range while giving accurate
results in the high slenderness region at ambient temperature. A proposal was made
to use buckling curve ‘a’ instead of buckling curve ‘b’ for ambient temperature
design based on the results of this study. In Figure 7.17 the applicability of buckling
curves ‘a’ and ‘b’ for fire design was investigated. It can be seen that both buckling
curves are not accurate for some temperatures in the case of G450 steel beams. In the
case of G250 steel beams the moment capacity data points are scattered in the
intermediate slenderness range and thus the use of any of these curves will give both
unsafe and overconservative predictions for some temperatures. For example, the
data points for 300oC and 400oC are plotted below even the buckling curve ‘b’ and
therefore predictions of Eurocode 3 Part 1.3 using buckling curve ‘b’ are unsafe.
Therefore considering all the buckling curves in Eurocode 3 Part 1.1, suitable
buckling curves were proposed for different temperature ranges and G250 and G450
steel members as given in Tables 7.13 (a) and (b).
Table 7.13: Proposed Buckling Curves from Eurocode 3 Part 1.1
(a) Low Grade Steel (G250)
Temperature oC 20 ≤ T < 300 300 ≤ T ≤ 400 400 < T ≤ 700
Buckling Curve a d c
(b) High Grade Steels
Temperature oC 20 [ T [ 300 300 < T [ 500 500 < T [ 700

Buckling Curve a b c
1.2 Eurocode 3 Part 1.3 (ECS, 2006) - 'a'
Eurocode 3 Part 1.3 (ECS, 2006) - 'b'
Elastic Buckling
1.0
G250-20
G250-200
0.8 G250-300
G250-400
M u/M sx or M b/M sx
0.6 G250-500
G250-600
G250-700
0.4
0.2
0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
(Msx/Mo)0.5
1.2 Eurocode 3 Part 1.3 - Buckling Curve 'a'
Eurocode 3 Part 1.3 - Buckling Curve 'b'

.
Elastic Buckling
1.0
G450-20
G450-200
0.8
G450-300
G450-400
0.6 G450-500
G450-600
0.4 G450-700
0.2
0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
(Msx/Mo)0.5
Figure 7.17: Comparison of FEA Results with Buckling Curves ‘a’ and
‘b’ of Eurocode 3 Part 1.3 / Eurocode 3 Part 1.1
Figures 7.18 (a) and (b) compare the finite element analysis results for G250 and
G450 cold-formed steel beams, respectively, at elevated temperatures with all the
beam design curves obtained from EC 3 Part 1.1 (ECS, 2005). Based on this
comparison, it is proposed that buckling curve ‘a’ is used in the temperature range 20
≤ T ≤ 300 for G450 steel beams while using buckling curve ‘b’ is recommended in
the range of 300<T≤500oC. Above 500oC, the buckling curve ‘c’ is recommended. In
the case of G250 steel beams, the buckling curves ‘a’ and ‘c’ are recommended for
the temperature ranges of 20 ≤ T < 300 and 400 < T ≤ 700, respectively. Between
300oC and 400oC, lower member moment capacities were obtained from FEA and
therefore the buckling curve‘d’ is recommended.
Eurocode 3 Part 1.3 - Buckling Curve 'c'

.
1.0 Eurocode 3 Part 1.3 - Buckling Curve 'd'
Elastic Buckling
G250-20
0.8
G250-200
G250-300
0.6 G250-400
G250-500
G250-600
0.4 G250-700
0.2
0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
(Msx/Mo)0.5

Eurocode 3 Part 1.3 - Buckling Curve 'c'
.
1.0 Eurocode 3 Part 1.3 - Buckling Curve 'd'
Elastic Buckling
G450-20
0.8
G450-200
G450-300
0.6 G450-400
G450-500
G450-600
0.4 G450-700
0.2
0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
(Msx/Mo)0.5

Figures 7.19 (a) to (c) show the comparison of ultimate member moment capacities
of G450 steel beams from FEA with the proposed buckling curves for the three
different temperature ranges as shown in Table 7.13. Similarly Figures 7.20 (a) to (c)
show the comparison of member moment capacities for G250 steel beams. Based on
these figures, the chosen buckling curves in Table 7.13 appear to give accurate
predictions, but further improvements are needed for some temperatures.
Table 7.14 shows the mean, COV and the capacity reduction factors of the FEA to
predicted member moment capacities using the new proposal made based on
different buckling curves for different temperature ranges as given in Table 7.13. On
average the proposal has a mean of 1.046 with a COV of 0.103. The capacity
reduction factor is 0.915.
Elastic Buckling
.
1.0
G450-20
G450-200
0.8
G450-300
0.6
0.4
0.2
0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
(Msx/Mo)0.5
(a) 20 ≤ T ≤ 300 oC
1.2 Eurocode 3 Part 1.3 - Buckling Curve 'b'
Elastic Buckling
.
1.0
G450-400
G450-500
0.8
0.6
0.4
0.2
0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
0.5
(Msx/Mo)
(b) 300 < T ≤ 500 oC

1.2
.
1.0 Eurocode 3 Part 1.3 - Buckling Curve 'c'
Elastic Buckling
0.8 G450-600
G450-700
0.6
0.4
0.2
0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
0.5
(Msx/Mo)
(c) 500 < T ≤ 700oC

Figure 7.19: Comparison of Ultimate Member Moment Capacities with
Predictions based on Selected Buckling Curves in Different Temperature
Ranges – G450 Steel Beams
1.2
.
1.0
Eurocode 3 Part 1.3 - Buckling Curve 'a'
Elastic Buckling
0.8
G250-20
G250-200
0.6
0.4
0.2
0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
0.5
(Msx/Mo)
(a) 20 ≤ T < 300oC

1.2
.
Eurocode 3 Part 1.3 - Buckling Curve 'd'

1.0
Elastic Buckling
0.8 G250-300
G250-400
0.6
0.4
0.2
0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
(Msx/Mo)0.5
(b) 300 ≤ T ≤ 400 oC

1.2
.
1.0 Eurocode 3 Part 1.3 - Buckling Curve 'c'
Elastic Buckling
G250-500
0.8
G250-600
G250-700
0.6
0.4
0.2
0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
0.5
(Msx/Mo)
(c) 400 < T ≤ 700 oC

Figure 7.20: Comparison of Ultimate Member Moment Capacities with
Predictions based on Selected Buckling Curves in Different Temperature
Ranges – G250 Steel Beams
Table 7.14: Mean, COV and Capacity Reduction Factors of the
Ratio of FEA to Predicted Member Moment Capacities using the Proposed
Buckling Curves
Temp 20 200 300 400 500 600 700 Overall

Mean 0.995 0.991 0.909 1.111 1.109 1.040 1.046 1.046
COV 0.061 0.072 0.117 0.097 0.148 0.096 0.107 0.103
Φ 0.899 0.889 0.781 0.976 0.917 0.914 0.909 0.915
This proposal is simply based on the finite element analysis results and does not
include the effect of identified non-linearity effect that influences the ultimate
capacities considerably. Therefore a new design method is required which includes
the effects of non-linear stress-strain relationship. This will allow the same design
method to be used for beams made of any type of steel.
7.5.2 New Design Method Proposed in Chapter 6
The ultimate moment capacities obtained from finite element analyses at elevated
temperatures are compared in Figure 7.21 with the new design method referred to as
Option 3 in Section 6.5.1.1 of Chapter 6 for ambient temperature design. According
to Figure 7.21 the new design method is accurate for some temperatures while giving
unsafe strength predictions for other temperatures. The new design method is unsafe
for temperatures at 300oC, 400oC and 600oC for G250 steel beams and 600oC and
700oC for G450 steel beams. The scattered nature of the results in the intermediate
slenderness region does not allow the use of a single beam design curve for all the
temperatures. Therefore separate beam design curves are required, which represent
the behaviour accurately at elevated temperatures.
1.2 New Proposal for ambient temperature
Elastic Buckling
.
1.0 G250-20
G250-200
0.8 G250-300
G250-400
M u/M y or M b/M y
0.6 G250-500
G250-600
0.4 G250-700
0.2
0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
0.5
(My/Mo)
1.2 New Proposal for ambient temperature
Elastic Buckling
.
1.0 G450-20
G450-200
0.8 G450-300
G450-400
M u/M y or M b/M y
0.6 G450-500
G450-600
0.4 G450-700
0.2
0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
0.5
(My/Mo)
Figure 7.21: Comparison of FEA results with the New Design Method
Proposed in Chapter 6
7.5.3 Eurocode 3 Part 1.2
Eurocode 3 Part 1.2 (ECS, 2005) provides design guidelines for the fire design of
hot-rolled steel members. The design method given in Eurocode 3 Part 1.2 is based
on studies carried out on hot-rolled steel members by Real et al. (1999, 2001, 2003).
This code recommends the same design guidelines for cold-formed steel members.
The lateral-torsional buckling capacity of beams at elevated temperatures is affected
by residual stresses, initial crookedness and twist, material reduction factors and
nonlinear stress-strain relationship. Most of these factors are different in cold-formed
steel sections in comparison to hot-rolled steel sections and therefore considerable
differences in the behaviour of these two types of steel beams are possible. Therefore
the applicability of hot-rolled steel based design method for cold-formed steel
members needs to be verified. Verification of the design method to cold-formed steel
members at elevated temperatures was carried out using the non-linear analysis
results in this section.
1.2 Eurocode 3 Part 1.2 - G250
Elastic Buckling
.
1.0 G250-20
G250-200
0.8 G250-300
G250-400
0.6 G250-500
G250-600
0.4 G250-700
0.2
0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
0.5
(Msx/Mo)
Figure 7.22: Comparison of FEA results with Eurocode 3 Part 1.2
1.2 Eurocode 3 - Part 1.2 - G450
Elastic Buckling
.
1.0 G450-20
G450-200
0.8 G450-300
G450-400
0.6 G450-500
G450-600
0.4 G450-700
0.2
0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
0.5
(Msx/Mo)
Figure 7.22: Comparison of FEA results with Eurocode 3 Part 1.2
Figure 7.22 (a) and (b) show the comparison of FEA results obtained for different
temperatures with the beam design curve obtained from the fire design code,
Eurocode 3 Part 1.2. Most of the FEA results are above the design curves for G250
and G450 steel beams except for very few results. Eurocode 3 Part 1.2 predicts the
moment capacities reasonably well for beams with higher slenderness (i.e.
(Msx/Mo)0.5≥2.0) for both G250 and G450 steel beams. Therefore it can be concluded
that for intermediate slenderness values, (Msx/Mo)0.5<2.0, Eurocode 3 Part 1.2
predictions are over-conservative with a maximum difference of approximately 30%
at 200oC for G250 steel beams. Dharma and Tan (2007) observed that this design
method is too conservative, especially for temperatures less than 500oC for hot-rolled
steel I-section beams. Same observation can be made from this research for cold-
formed steel members based on the results of G450 steel beams. Dharma and Tan
(2007) identified the reason for this conservatism in the case of temperatures less
than 500oC as that due to some very low experimental results at 500oC and 600oC
obtained by Real et al. (2003) based on which these design rules were developed.
The reason for the over-conservative results in this research can also be attributed to
the different residual stresses and stress-strain relationship at elevated temperatures
and the use of negative geometric imperfections in the case of cold-formed steel
mono-symmetric beams compared to hot-rolled steel I-beams based on which
Eurocode 3 Part 1.2 design guidelines have been developed. In addition to being
over-conservative in the case of intermediate slenderness values, it can also be seen
that design curve in the low slenderness region is steep, which gives uneconomical
capacity predictions for all the selected beams according to FEA results of this study.
Figure 7.22 (a) and (b) present the evidence of considerably high capacities at very
low slenderness values. However, this method could be safely used as a conservative
fire design method in all the member slenderness regions.
The design method given in Eurocode 3 Part 1.2 was developed by Real (1999, 2001
and 2003). This study was based on experimental and numerical studies carried out
on single IPE100 section with spans varying from 0.5 m to 6.5 m. Figure 7.23 shows
their experimental and numerical results for temperatures 200oC, 400oC, 500oC and
600oC. Close observation of Real et al.’s (2003) experimental and numerical results
reveals that their numerical results are considerably lower in comparison to test
results at 200oC, 300oC, 400oC. It can also be seen that there are large differences in
the failure loads of beams having the same slenderness with similar beam
dimensions, especially at 500oC and 600oC (Figure 7.23). According to Dharma and
Tan (2007) this discrepancy may be due to the large degradation of mechanical
properties at higher temperatures and even a small error made in the temperature
measurement can cause large difference in the test results. Finally, it appears that
their proposed design method underestimates the lateral-torsional buckling capacities
with a maximum of 50% in comparison to their own experimental results, especially
at 200oC, 300oC and 400oC. This strongly raises the need for multiple design curves
for steel beams subject to lateral-torsional buckling as a function of varying elevated
temperatures.
(a) Experimental Results – 200oC (b) FEA Results – 200oC
(c) Experimental Results – 400oC (d) FEA Results – 400oC
(e) Experimental Results – 500oC (f) FEA Results – 500oC
(g) Experimental Results – 600oC (h) FEA Results – 600oC
Figure 7.23: Comparison of Experimental Results of Real et al. (2003)

with their Finite Element Analysis Results
7.5.4 Design Methods Proposed by Dharma and Tan (2007)
Dharma and Tan (2007) proposed two design methods for lateral-torsional buckling
at elevated temperatures based on finite element analysis results carried out on
simply supported hot-rolled steel universal I beams subject to uniform bending
moment and mid-span loading conditions. The first design method called
“Alternative Approach” provides safe and more accurate capacity predictions for
lateral-torsional buckling given in Section 2.9.3 of Chapter 2. This design method
can also be used at ambient temperature and gives the same results as in BS 5950
Part 1.1 (Dharma and Tan, 2007). The development of design method was based on
experimental results of Real et al. (2003) and numerical results carried out on UB
and UC sections of Grade S275 and S355 steels using MSC MARC Mentat. They
proposed another design method called Rankine approach, which is a simpler and
quicker method. However, this design method does not give accurate results as the
Alternative Approach gives. Both design methods proposed by Dharma and Tan
(2007) take into account the effect of major axis curvature on elastic buckling
moment using an approximate solution given by Kirby and Nethercot (1979).
Dharma and Tan (2007) discussed the effect of non-linear stress-strain relationship
on the beam capacity. At elevated temperatures the stress-strain curves become
highly non-linear with tangent modulus changing from initial tangent. They
mentioned about the unsafe capacity predictions by simply applying material
reduction factors and highlighted the importance of using tangent modulus instead of
the initial tangent at elevated temperatures. However, the use of tangent modulus is
difficult without any knowledge at which stress the failure would happen and
therefore an iterative process is needed. They used the term called non-linearity
factor to account for the highly non-linear stress-strain relationship at elevated
temperatures. The nonlinearity factor (ς ) is given in Equation 7.2 as proposed by
Dharma and Tan (2007).
f pT f pT f y kp
ς= T
= T
= (7.7)
f y f y fy ky
where f pT , f yT and f y are the elevated temperature limit of proportionality, the
elevated temperature yield strength and the ambient temperature yield strength. k p
and k y are the limit of proportionality reduction factor and the yield strength
reduction factor at elevated temperatures.
The finite element analysis results from this study are compared in Figure 7.24 for
one cross-section for several temperatures (20oC, 200oC, 300oC, 500oC, 600oC and
700oC) using Dharma and Tan’s alternative approach. It can be seen that the capacity
predictions using alternative approach are unsafe at some temperatures especially at
600oC for G450 steel beams and 300oC for G250 steel beams. Dharma and Tan
developed their rules based on doubly symmetric hot-rolled steel beams whose
stress-strain relationships at elevated temperatures are very different to those in the
case of cold-formed steel beams. This is the possible reason for the disagreement in
the results.

20
200
300
1.0 400
500
600
700
0.8 Dharma and Tan (2007) - 20oC
Dharma and Tan (2007) - 200oC

0.6
0.4
0.2
0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
0.5
(Msx/Mo)
(a) G450-1.9-80-40-15 Section
Figure 7.24: Comparison of FEA Results at Elevated Temperatures with

the Alternative Approach Proposed by Dharma and Tan (2007)
20
200
300
1.0 400
500
600
0.8 700

0.2
0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
0.5
(Msx/Mo)
(b) G250-1.55-120-45-15 Section

the Alternative Approach Proposed by Dharma and Tan (2007)
Figure 7.25 shows the comparison of FEA with the Rankine design approach
proposed by Dharma and Tan (2007). It appears that the maximum error involved
with Rankine method is about 20% at 200oC and provides conservative capacity
predictions in the intermediate slenderness range for both G250 and G450 steel
beams. According to Dharma and Tan (2007) Rankine approach provides
comparatively good results to that of Eurocode 3 Part 1.2. They observed a
maximum conservatism error of approximately 18% in the results with the capacity
predictions using Rankine approach. This approach does not take into consideration
the effect of stress-strain relationship.
1.2
Dharma and Tan (2007) - Rankine Approach
.
Elastic Buckling
1.0
G250-20
G250-200
G250-300
0.8
G250-400
G250-500
G250-600
0.6
G250-700
0.4
0.2
0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
0.5
(Msx/Mo)
1.2 Dharma and Tan (2007) - Rankine Approach
Elastic Buckling
.
1.0 G450-20
G450-200
0.8 G450-300
G450-400
0.6 G450-500
G450-600
0.4 G450-700
0.2
0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
0.5
(Msx/Mo)

the Rankine Approach Proposed by Dharma and Tan (2007)
7.5.5 New Proposal
In the previous sections the need for a new design method for the fire design of cold-
formed steel beams subject to lateral-torsional buckling was identified. The current
design method given in the fire design code, Eurocode 3 Part 1.2 (ECS, 2005), was
found to be over-conservative and it was found that according to Real et al.’s (2003)
experimental results, separate buckling curves are needed for different elevated
temperatures. The FEA results of ultimate member moment capacities were found to
be dependent on the level of non-linearity in the material stress-strain graph. Dharma
and Tan (2007) used a factor of fp,T/fy,T to account for the non-linearity in stress-strain
graphs in hot-rolled steel beams. According to this study it was found that the non-
linearity in stress-strain graphs affects the moment capacity results in the
intermediate slenderness region where failure takes place with higher stress
development closer to yield stress. It was found that the effect of non-linearity can be
accounted for by using the fp,T/fy,T factor and therefore new design equations were
proposed by including this factor in the proposed design equations in Section 6.5.1.1
as Option 3 for ambient temperature design.
The nominal member moment capacity ( M b ,T ) of the laterally unbraced segments of
cold-formed steel lipped channel beams subject to lateral-torsional buckling at

elevated temperatures is given by,
⎛M ⎞
M b ,T = Z c ⎜ c ,T ⎟ (7.8)
⎜ Z ⎟
⎝ f ⎠
Mc,T in Equation 7.3 is given by,
For λb ,T ≤ 0.6 M c ,T = M y ,T (7.9a)
⎛ f p ,T ⎞
0.4 ⎡ −0.55 ⎤
For λb,T > 0.6 M = 0.9⎜ ⎟ ⎢ λ4 + 2.05⎛⎜ f p ,T ⎞
⎟ − λ2 ⎥ M
c, T ⎜ f ⎟ ⎢ b, T ⎜ f ⎟ b, T ⎥ y , T
⎝ y ,T ⎠ ⎣⎢ ⎝ y ,T ⎠ ⎦⎥
a b (7.9b)
where,
⎛ f p ,T ⎞
⎜ ⎟ - the ratio of proportional limit and yield strength at temperature T.
⎜ f ⎟
⎝ y ,T ⎠
0.5
⎛M ⎞
λb ,T = ⎜⎜ y ,T ⎟
⎟ (7.10)
⎝ M o ,T ⎠
M y ,T in Equations 7.4 (b) and 7.5 is the yield moment capacity at temperature T, which
is given by,
M y ,T = Z f f y ,T (7.11)
Elastic lateral-torsional buckling moment at temperature T is given by,
⎛ π 2 ET I y ⎞⎛ π 2 ET I w ⎞
M o ,T = ⎜ ⎟⎜ G J + ⎟ (7.12)
⎜ L2 ⎟⎜ T L2 ⎟⎠
⎝ ⎠⎝
ET and GT are the elastic modulus and the shear modulus at temperature T.
Elastic modulus and yield strength at temperature T are given by,
ET = k E ,T E20
f y ,T = k y ,T f y , 20
k E ,T and k y ,T are the yield strength and elastic modulus reduction factors.
The ultimate moment capacities at elevated temperatures from FEA were compared
with the predicted moment capacities with appropriately reduced mechanical
properties for each temperature using the proposed new design equations. The results
are plotted in the non-dimensionalysed format for different temperatures in Figure
7.26 and 7.27 for Grade 250 and Grade 450 steel beams, respectively. Suitability of
the proposed design method was investigated for each temperature and it could be
seen that the new proposal with reduced mechanical properties is able to predict the
ultimate member moment capacities accurately.
1.2
1.0
Elastic Buckling
0.8 G250
New Proposal
M u/M y or M b/M y
0.6
0.4
0.2
0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
0.5
(My/Mo)
(a) 200 oC
1.2
1.0
Elastic Buckling
G250
0.8
New Proposal
M u/M y or M b/M y
0.6
0.4
0.2
0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
0.5
(My/Mo)
(b) 300 oC
1.2
1.0
Elastic Buckling
0.8
G250
M u/M y or M b/M y
New Proposal
0.6
0.4
0.2
0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
(My/Mo)0.5
(c) 400 oC
Figure 7.26: Comparison of the New Proposal with FEA Results
– G250 Steel Beams
1.2
1.0
Elastic Buckling
0.8 G250
New Proposal
M u/M y or M b/M y
0.6
0.4
0.2
0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
0.5
(My/Mo)
(d) 500 oC
1.2
1.0
Elastic Buckling
0.8 G250
New Proposal
M u/M y or M b/M y
0.6
0.4
0.2
0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
0.5
(My/Mo)
(e) 600 oC
1.2
1.0
Elastic Buckling
0.8 G250
New Proposal
M u/M y or M b/M y
0.6
0.4
0.2
0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
(My/Mo)0.5
(f) 700 oC
– G250 Steel Beams (Contd.)
1.2
1.0
Elastic Buckling
0.8 G450
New Proposal
M u/M y or M b/M y
0.6
0.4
0.2
0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
(My/Mo)0.5
(a) 200 oC
1.2
1.0
Elastic Buckling
0.8 G450
M u/M y or M b/M y
New Proposal
0.6
0.4
0.2
0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
(My/Mo)0.5
(b) 300 oC
1.2
1.0
Elastic Buckling
0.8
G450
M u/M y or M b/M y
New Proposal
0.6
0.4
0.2
0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
(My/Mo)0.5
(c) 400 oC
1.2
1.0
G450
M u/M y or M b/M y
0.6 New Proposal
0.4
0.2
0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
(My/Mo)0.5
(c) 500 oC
1.2
1.0
G450
M u/M y or M b/M y
0.6 New Proposal
0.4
0.2
0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
0.5
(My/Mo)
(d) 600 oC
1.2
1.0
Elastic Buckling
0.8
G450
M u/M y or M b/M y
New Proposal
0.6
0.4
0.2
0.0
0.0 0.5 1.0 1.5 2.0 2.5 3.0
(My/Mo)0.5
(f) 700 oC
Table 7.15 presents the mean, COV and the capacity reduction factors of finite
element analysis results to predicted capacities using the developed design equations
for different temperatures. It can be seen that good mean values are obtained for all
the temperatures and the overall mean is 0.983. The corresponding COV is 0.068
while the capacity reduction factor is 0.886. Therefore it is recommended to use the
proposed design method for cold-formed steel lipped channel beams subject to
lateral-torsional buckling at elevated temperatures.
Table 7.15: Mean, COV and Capacity Reduction Factors of FEA to

Predicted Moment Capacities
Temp 20 200 300 400 500 600 700 Overall
Mean 0.984 0.983 1.013 0.995 0.981 0.964 0.957 0.983

COV 0.056 0.062 0.071 0.122 0.059 0.067 0.061 0.068
Φ 0.892 0.888 0.910 0.850 0.888 0.868 0.865 0.886
The new proposed design method for elevated temperatures has many advantages in
comparison to other design methods, especially the fire design method given in
Eurocode 3 Part 1.2. The new proposed equation is a modification of ambient
temperature design method and is the same if used for ambient temperature design
because at ambient temperature the material non-linearity factor is almost equal to
1.0. Another advantage of this proposal is that this equation can be modified by
changing ‘a’ and ‘b’ as shown in Equation 6.4 to have different buckling curves and
provides a unique way of developing design methods for other types of cross-
sections. This design method provides a simpler way of calculating lateral-torsional
buckling capacities at elevated temperatures in comparison to the proposed design
method by Dharma and Tan (2007) for hot-rolled steel flexural members.
7.6 Conclusion
This chapter has described detailed parametric studies carried out on the behaviour
of cold-formed steel lipped channel beams subject to local buckling and lateral-
torsional buckling at elevated temperatures. The parametric studies were carried out
for nine lipped channel sections using validated finite element models reported in
Chapter 5.
The finite element analysis results for the section moment capacity at elevated
temperatures indicated that cold-formed steel lipped channel beams possess inelastic
reserve capacity. The capacity predictions using Eurocode 3 Part 1.3 are more
accurate if used independently at different temperatures compared to the predictions
based on the ambient temperature capacity and the yield strength reduction factor
which accounts increased temperature as proposed in Eurocode 3 Part 1.2. Eurocode
3 Part 1.2 design method is easier to use compared to Eurocode 3 Part 1.3. EC 3 Part
1.3 predictions of the inelastic reserve capacities at elevated temperatures are more
accurate compared to Shifferaw and Schafer (2008). However, the design method in
EC 3 Part 1.3 is lengthy and complex while Shifferaw and Schafer (2008) design
method provides a simple and faster method of calculating the inelastic reserve
capacity. It was found that the use of yield strengths based on 1.5% total strains
gives unsafe capacity predictions at elevated temperatures and therefore it is
recommended to use 0.2% proof stress as the yield strength for all the elevated
temperatures.
Lateral-torsional buckling capacity is strongly influenced by the level of non-

linearity in the stress-strain curves at elevated temperatures. It was found that the
level of non-linearity can be demonstrated by the ratio of the limit of proportionality
to the yield strength of steel. The lowest moment capacities of G250 steel beams
were obtained at 300oC, which correspond to the lowest fp,T/fy,T ratio of G250 steel.
The results of the parametric study were compared with the predicted moment
capacities using the ambient temperature design code, Eurocode 3 Part 1.3 (ECS,
2006), the new proposed design method in Chapter 6, fire design code, Eurocode 3
Part 1.2 (ECS, 2005) and the proposed design method by Dharma and Tan (2007).
This comparison showed that Eurocode 3 Part 1.3 design method (design equations
of Eurocode 3 Part 1.1 with bending curve ‘b’) is unsafe and overconservative for
some temperatures especially in the intermediate slenderness region. Proposed
buckling curve ‘a’ for ambient temperature design based on this research is also
unsafe for elevated temperature design. Therefore separate buckling curves were
proposed for different temperature ranges.
The new design method developed for cold-formed steel beams at ambient
temperature was also found to be unsuitable at elevated temperatures because a
single beam curve does not satisfy the range of results in the intermediate
slenderness region.
Eurocode 3 Part 1.2 (ECS, 2005) predictions were found to be overconservative for
all the temperatures except for beams with very high slenderness. Eurocode 3 Part
1.2 has enforced a limitation on the maximum temperatures to 350oC for cold-
formed steel members. This temperature limitation was found to be overly
conservative since higher moment capacities are available until 500oC in the case of
G450 steel beams.
The design method called Alternative Approach proposed by Dharma and Tan
(2007) provided reasonably accurate capacity predictions for some temperatures.
Rankine approach is also overconservative with a maximum error of 18%.
Finally, a new design method was developed for the design of cold-formed steel
lipped channel beams subject to lateral-torsional buckling at elevated temperatures.
This design proposal enables safe structural designs of cold-formed steel beams in
fire situations.
8 Conclusions and Recommendations
This thesis has described a detailed investigation into the behaviour of cold-formed
steel lipped channel beams subject to local and lateral-torsional buckling effects at
both ambient and elevated temperatures.
In the first phase of this research, more than 80 tensile coupon tests were carried out
with 1.55 and 1.95 mm thick G250 steels and 1.5 and 1.9 mm thick G450 steels at
temperatures ranging from 20oC to 700oC. Elevated temperature tests were carried
out in an electrical furnace using the static method. Strains measurements at elevated
temperatures were made using a non-contact strain measuring device called Laser
Speckle Extensometer. Accurate mechanical properties of cold-formed steels (yield
stress and elastic modulus) and stress-strain relationships were obtained at both
ambient and elevated temperatures.
In the second phase, eight lateral-torsional buckling tests were carried out at ambient
temperature. Simply supported beams were tested with quarter point loading until
failure. In the third phase of the research, ideal and experimental finite element
models were developed using Patran and ABAQUS. Two ideal models were
developed: one for studying the section moment capacities of compact cold-formed
lipped channel beams subject to inelastic local buckling and yielding and the other
model for the member moment capacities of beams subject to lateral-torsional
buckling. The ideal models for section and member moment capacities were first
validated using the elastic buckling analysis results from a well established finite
strip analysis program Thin-wall. The ideal model for the member moment
capacities was then validated in relation to the nonlinear analysis capability using the
results from the experimental study of this research and the numerical and
experimental studies of past research.
In the fourth phase of this research, detailed parametric studies were undertaken for
the section and member moment capacities of cold-formed steel lipped channel
beams subject to local and lateral-torsional buckling at ambient temperature (12
compact cross-sections) and at elevated temperatures (9 compact cross-sections).
Chapter 3 of this thesis presented the details of the experimental investigation on the
mechanical properties of cold-formed steels at ambient and elevated temperatures
using tensile coupon tests and the new developed equations for the reduction factors
of yield strength and elastic modulus. Chapter 4 presented the experimental
investigation carried out on the lateral-torsional buckling of simply supported cold-
formed steel lipped channel beams at ambient temperature. Chapter 5 gave the
details of finite element models developed for section and member moment
capacities of cold-formed steel lipped channel beams including the validation of
developed finite element models. Chapter 6 presented the results of the parametric
study carried out at ambient temperature on the section and member moment
capacities including the new developed design equations for the member moment
capacity of cold-formed steel beams subject to lateral-torsional buckling. Chapter 7
presented the detailed parametric study carried out on the section and member
moment capacities of cold-formed steel lipped channel beams at elevated
temperatures including the new design equations for the fire design of cold-formed
steel lipped channel beams subject to lateral-torsional buckling. This final chapter
presents the most significant findings of this research and the recommendations
made for further research.
The most significant findings obtained from this research are as follows:
Improved knowledge and understanding of the mechanical properties of cold-

formed steels at elevated temperatures and the development of predictive
equations for yield strength and elastic modulus reduction factors at elevated
temperatures with a proposed change for the current stress-strain model
developed by Ranawaka and Mahendran (2009a). These equations provide a
method of calculating the mechanical properties at any elevated temperature
provided that the ambient temperature mechanical properties are known.
Significantly improved knowledge and understanding of the section and

member moment capacities of cold-formed steel lipped channel beams
subject to inelastic local buckling and yielding, and lateral-torsional buckling
at ambient temperature.
Assessment of the accuracy and suitability of the currently available ambient

temperature design methods for section and member moment capacities of
cold-formed steel beams associated with inelastic local buckling and
yielding, and lateral-torsional buckling effects, respectively.
Development of new design rules to predict the member moment capacity of

cold-formed steel lipped channel beams subject to lateral-torsional buckling
at ambient temperature.
Significantly improved knowledge and understanding of the section and

member moment capacities of cold-formed steel lipped channel beams
subject to local buckling and yielding, and lateral-torsional buckling effects
under simulated fire conditions.
Assessment of the accuracy and suitability of the currently available fire

design rules and ambient temperature design rules based on reduced
mechanical properties for the section and member moment capacities of cold-
formed steel lipped channel beams at elevated temperatures.
Development of a new design method for cold-formed steel lipped channel

beams subject to lateral-torsional buckling at elevated temperatures.
Following important conclusions and recommendations have been drawn based on

the specific topics investigated in this research.
8.1 Mechanical Properties of Cold-formed Steels at Elevated

Temperatures
The mechanical properties of cold-formed steels with two thicknesses under

two steel grades were obtained by conducting tensile coupon tests at
temperatures ranging from 20oC to 700oC. The yield strength, ultimate
strength, elastic modulus and the stress-strain curve were all determined from
tensile coupon tests at each selected temperature. It was found that the
mechanical properties of cold-formed steels decreased considerably with
increasing temperature.
The stress-strain relationship of Grade 250 cold-formed steels has a linear

elastic region followed by a plateau at ambient temperature and up to 200oC
while beyond 200oC it becomes a gradual yielding type. G450 steels show a
stress-strain relationship with strain hardening at ambient and elevated
temperatures.
The yield strength reduction factors based on 0.5% total strain is closer to
that based on 0.2% proof stress for both low and high strength steels. Similar
observation was also made by Ranawaka and Mahendran (2009a). Therefore
it is recommended that either 0.2% proof stress or stress at 0.5% total strain
is used as the yield strength at elevated temperatures.
The steel grade has an influence on the yield strength of steel at elevated
temperatures while there is no observable influence of steel thickness on the
yield strength results.
The elastic modulus of cold-formed steels decreases in a similar trend

irrespective of steel grade and thickness with increasing temperature.
The ultimate strength reduction factors of G250 and G450 steels are closer to
each other except at 200oC and 300oC. The effect of thickness on the
reduction factors of ultimate strength is insignificant for both low and high
strength steels. It is interesting to note that there is almost 20% increase in the
ultimate strength of low strength steel at 200oC compared to that at ambient
temperature.
Low strength steel shows higher ductility compared to high strength steel at
ambient temperature due to the comparatively high strain hardening caused
by cold working in the case of high strength steel. Up to 300oC, G450 steels
show less ductile failures (brittle with no necking) and thereafter their
failures became more ductile. Brittle failure can be seen in G250 steel only at
100oC. However, it shows some ductile behaviour at ambient temperature,
and increased ductile behaviour at temperatures beyond 200oC.
Based on the yield strength results obtained from tensile coupon tests at
elevated temperatures in this research, it was found that none of the proposals
made by Outinen (1999), Lee et al. (2003), Mecozzi and Zhao (2005) and
Chen and Young (2007) can be used to predict the yield strength reduction
factors of 1.55 mm and 1.95 mm thick G250 and 1.5 mm and 1.9 mm thick
G450 cold-formed steels at elevated temperatures. Ranawaka and
Mahendran’s (2009a) equations closely predict the yield strength reduction
factors. However, certain modifications were identified as necessary in some
cases. Therefore based on the results of this research and other research
carried out at QUT including Ranawaka and Mahendran (2009a), a new set
of equations for yield strength reduction factors was developed for cold-
formed steels at elevated temperatures.
Several researchers (Outinen, 1999, Lee et al., 2003, Ranawaka and

Mahendran, 2009a and Chen and Young, 2007) developed equations for the
elastic modulus reduction factors as a function of temperature. It was found
that none of these equations accurately predict the elastic modulus reduction
factors. Therefore new linear equations were developed for elastic modulus
reduction factors for two identified temperature regions (20oC-200oC and
200oC–800oC).
Ranawaka and Mahendran’s (2009a) equations predict the stress-strain

curves of light-gauge high strength cold-formed steels reasonably accurately.
However, an improvement was made to these equations in the case of low
strength steels.
8.2 Section and Member Moment Capacities of Cold-formed Steel
Lipped Channel Beams at Ambient Temperature
Ideal finite element models were developed to study the section and member
moment capacities of cold-formed steel lipped channel beams subject to
inelastic local buckling and yielding, and lateral-torsional buckling. They
were validated using the elastic buckling analysis results from a well
established finite strip analysis program, Thin-Wall. The ideal model for
lateral-torsional buckling was also validated for its non-linear analysis
capability using the results from the experimental study of this research and
the numerical and experimental studies of past research.
The ultimate section moment capacities of 12 compact and short cold-formed

steel beams were successfully investigated at ambient temperature by using
the validated finite element model. The ultimate failures of all the beams
occurred by cross section yielding and it was found that inelastic reserve
capacity exists in compact cold-formed steel beams. However, based on finite
element analysis results, it appears that full plastic moment capacity cannot
be achieved by cold-formed steel beams.
Effect of using the strain hardening material model over elastic perfect plastic
material model was found to be insignificant with only 1.4% difference.
Since all the selected compact cold-formed steel beams in this research
exhibited inelastic reserve capacities, the AS/NZS 4600 design rules based on
initiation of yielding underestimated the section moment capacities
considerably. This design code allows the use of inelastic reserve capacity
with restrictive limits based on the slenderness of compression elements.
Most of the compact cross-sections chosen in this study did not meet these
plate slenderness requirements. Considering the strain development in short
beams, the maximum strain has always exceeded the limiting strain of 3ε y in
AS/NZS 4600. If the limits for the longitudinal membrane strain and the plate
slenderness ratios (dw/t) are increased in AS/NZS 4600, higher inelastic
reserve capacities can be used for cold-formed steel lipped channel beams.
Shifferaw and Schafer’s (2008) design rules accurately predicted the inelastic
bending capacities of cold-formed steel lipped channel beams at ambient
temperature.
The design method given in Eurocode 3 Part 1.3 (ECS, 2006) accurately
predicted the section moment (inelastic bending) capacities of lipped channel
beams at ambient temperature.
A series of elastic buckling and non-linear finite element analyses was

carried out using the developed finite element model of simply supported
cold-formed steel lipped channel beams subjected to lateral-torsional
buckling at ambient temperature. The ultimate member moment capacity of
beams with intermediate slenderness is strongly influenced by the presence of
residual stresses and initial geometric imperfections and thus drops well
below the elastic buckling capacity. The member capacities of beams with
high slenderness are also lower than their elastic buckling capacities.
However, as the member slenderness increases the member capacity
approaches the elastic buckling capacity.
It was found that cold-formed steel beams are very sensitive to initial
geometric imperfections. Finite element analysis conducted with positive and
negative initial geometric imperfections revealed that beams with negative
imperfections failed at considerably lower failure loads compared to that with
positive imperfections. Imperfection sensitivity analysis revealed that as the
initial imperfection reduces the beam capacities increase. As the magnitude
of initial imperfection was reduced from L/1000 to L/3000 the ultimate
moment capacity approached the elastic buckling curve for beams with
higher slenderness.
The moment capacity curves in the elastic and inelastic regions for different
sections plot in an orderly manner as a function of d/b (web height to flange
width) ratio so that with increasing d/b ratio the moment capacity curve move
downwards in the region of intermediate slenderness and as the slenderness
increases these curves converge.
The ultimate moment capacities of highly compact sections (thicker sections)

reach their elastic buckling moments at lower member slenderness values
while other sections reach their elastic buckling moments at higher
slenderness values (λ>2.5).
Steel grade does not have a significant influence on the ultimate moment
capacities plotted in the non-dimensional format.
The moment capacity results of beams with residual stresses were only
slightly lower than those of beams without residual stresses and therefore the
influence of residual stresses on lateral-torsional buckling capacities was
considered to be insignificant.
AS/NZS 4600 (SA, 2005) member moment capacity predictions for lateral-
torsional buckling were found to be unsafe in both elastic and inelastic
buckling regions for cold-formed steel lipped channel beams. It was also
found that as the initial geometric imperfection magnitude was reduced from
L/1000 to L/3000 the moment capacities increased, but the member moment
capacity design rules in AS/NZS 4600 (SA, 2005) are still unsafe even for
beams with a smaller initial geometric imperfection of L/3000.
The design method given in EC 3 Part 1.3 (ECS, 2006) (Design Method
given in EC 3 Part 1.1 with buckling curve ‘b’) was found to be accurate for
cold-formed steel beams with high slenderness values while its predictions
for beams with intermediate slenderness are over-conservative. Therefore the
bucking curve ‘a’ is proposed instead of ‘b’ in the equations given in EC 3
Part 1.1 for cold-formed steel beams.
The predictions of BS 5950 Part 5 (BS, 1998) design equations were also
found to be unsafe for the member moment capacities of cold-formed steel
beams.
Based on the comparison with the ultimate moment capacities of cold-formed

steel lipped channel beams obtained from finite element analyses, it was
identified that some modifications are needed for the AS/NZS 4600 (SA,
2005) moment capacity equations for lateral-torsional buckling. New design
equations were developed to accurately predict the ultimate member moment
capacities of cold-formed steel lipped channel beams. The comparison of the
ultimate moment capacity predictions from the new design equations showed
that the new design equation can accurately predict the lateral torsional
buckling capacities. They can also be used with the direct strength method.
The capacity reduction factor of 0.9 recommended in AS/NZS 4600 (SA,
2005) can be used with the new equations.
The observations made in this section in relation to AS/NZS 4600 (SA, 2005)
are equally applicable to NAS (AISI, 2001) as NAS and AS/NZS 4600 have
identical design rules for section and member moment capacities.
8.3 Section and Member Moment Capacities of Cold-formed Steel

Similar finite element models used in the study on the behaviour of cold-
formed steel beams at ambient temperature were used to study the section
and member moment capacities of cold-formed steel lipped channel beams at
elevated temperatures. The ultimate section and member moment capacities
of nine compact cold-formed steel sections with a range of spans were
successfully obtained at various elevated temperatures by using the
developed ideal models. The section moment capacities of very short beams
with compact cross sections are more than their yield moment capacities at
elevated temperatures, indicating the existence of inelastic reserve capacity at
Some finite element analysis results based on true stress-strain model (strain
hardening) showed that the ultimate section moment capacity exceeded the
plastic moment capacity (Mp) calculated using the yield stress determined
based on 0.2% proof stress. This indicates that the use of 0.2% proof stress as
the yield strength may not be very accurate in the case of steels with non-
linear stress-strain characteristics. With increasing temperatures, non-
linearity of stress-strain curves and ductility increase and hence the use of
0.2% proof stress as the yield strength may not be accurate. However, the use
of higher yield strengths based on 1.5% total strain over-predicted the
available section moment capacities even for those sections for which the
ultimate moment capacities from finite element analyses exceeded Mp, ie.
unconservative. Therefore it is recommended to use 0.2% proof stress as the
yield strength in the moment capacity calculations for cold-formed steel
Eurocode 3 Part 1.3 (ECS, 2006) design rules with appropriately reduced
mechanical properties accurately predicted the section moment capacities of
cold-formed steel lipped channel beams at elevated temperatures.
The proposed design rules of Shifferaw and Schafer (2008) for the section
moment capacity of cold-formed steel beams at ambient temperature were
found to give conservative capacity predictions when used in elevated
temperature design with reduced mechanical properties.
Eurocode 3 Part 1.2 section moment capacity design rules are obtained based
on ambient temperature section moment capacity design rules in Eurocode 3
Part 1.3 modifies with the yield strength reduction factor. These simplified
design rules of Eurocode 3 Part 1.2 also accurately predicted the section
moment capacities at elevated temperatures.
Although the design methods considered in this research predicted the section
moment capacities of cold-formed steel beams accurately at elevated
temperatures, careful consideration should be given regarding the
compressive strain development in the beams in order to keep it within
acceptable limits. Before making a decision on using the inelastic reserve
capacity at elevated temperatures, a detailed experimental study should be
undertaken to determine whether there is a need to limit the maximum
compressive strain in cold-formed steel beams to eliminate possible fractures.
Elastic buckling and ultimate moment capacities of compact cold-formed

steel lipped channel beams subject to lateral-torsional buckling were obtained
using finite element analyses carried out with elevated temperature
mechanical properties. As the temperature increases the yield strength and
elastic modulus decrease and hence the lateral-torsional buckling capacity
also decreased. However, the results are not dependent only on the
temperature. Noticeable differences were seen between the results of cold-
formed beams made of two steel grades.
The member moment capacity of cold-formed steel beam subject to lateral-

torsional buckling is strongly influenced by the level of non-linearity in the
stress-strain curves at elevated temperatures. It was found that the level of
non-linearity can be modelled by the ratio of the limit of proportionality to
the yield strength of steel (fp,T/fy,T). The ultimate moment capacities obtained
using the strain hardening material model were lower than those obtained
using the elastic perfect material model, indicating the influence of non-linear
stress-strain relationship at elevated temperature. Larger difference is seen in
the case of Grade 250 steel beam, which has very low fp,T/fy,T ratio whereas
the difference is very small in the case of G450 steel that has a considerably
high fp,T/fy,T ratio.
The non-dimensional member slenderness for a particular span which

directly affects the moment capacity of beams subjected to lateral-torsional
buckling does not reduce as the temperature increases. The non-dimensional
member slenderness appears to follow the variation of kyT/kET ratio.
For a particular span length the member moment capacity achievable relative
to the section moment capacity is dependent on the kyT/kET ratio. However,
this behaviour can be altered if it involves the non-linear stress-strain
relationship.
The results showed that Eurocode 3 Part 1.3 member moment capacity design
rules (design equations of Eurocode 3 Part 1.1 with buckling curve ‘b’) are
unsafe and overconservative for some temperatures, especially in the
intermediate slenderness region. Proposed buckling curve ‘a’ for ambient
temperature design based on this research is also unsafe for elevated
temperature design. Therefore separate buckling curves were proposed for
different temperature ranges.
The new member moment capacity design rule proposed in the case ambient
temperature was also found to be unsuitable because a single beam curve
does not satisfy the range of results in the intermediate slenderness region.
Eurocode 3 Part 1.2 member capacity predictions were found to be

overconservative for all the temperatures except for beams with very high
slenderness. Eurocode 3 Part 1.2 has enforced a limitation on the maximum
temperatures to 350oC for cold-formed steel members. This temperature
limitation was found to be overly conservative since higher moment
capacities were present until 500oC in the case of G450 steel beams.
The design method called Alternative approach as proposed by Dharma and

Tan (2007) provided comparatively accurate member moment capacity
predictions for some temperatures, but overconservative in other cases.
Rankine approach was also found to be overconservative with a maximum
error of 18%.
The need for a new design rule for the fire design of cold-formed steel beams
subject to lateral-torsional buckling was identified using the comparisons
made with available design methods. Therefore a new design rule was
developed for the design of cold-formed steel lipped channel beams at
elevated temperatures. This design rule takes the same format as the
developed equations for ambient temperature design. Finite element analysis
results of ultimate member moment capacities were found to be dependent on
the level of non-linearity in the stress-strain curve and hence the factor,
fp,T/fy,T, was used in the new design rule to allow for the effect of non-
linearity. This design method provides different beam design curves for
varying elevated temperatures as a function of the fp,T/fy,T ratio.
The comparisons of ultimate lateral-torsional buckling capacities obtained

from the new design rules with the finite element analysis results showed that
the new design rules developed in this research accurately predict the
ultimate failure loads of cold-formed steel lipped channel beams subject to
lateral-torsional buckling at elevated temperatures. Therefore it is
recommended to use the new design rules developed in this research to
predict the ultimate member moment capacities of cold-formed steel lipped
channel beams subject to lateral-torsional buckling at elevated temperatures.
8.4 Future Research
Further research is recommended in the area of inelastic reserve capacity and

associated strains at failures using a detailed experimental study both at ambient and
elevated temperatures in order to verify the accuracy of currently available design
rules and to develop suitable strain limits if required.
This research was based on a detailed numerical study of cold-formed steel beams
subject to lateral-torsional buckling at elevated temperatures. It is recommended that
an experimental study of the analysed cold-formed steel beams is undertaken to
confirm the numerical findings.
The behaviour of cold-formed steel beams subject to distortional buckling has not
been investigated at elevated temperatures. This research only investigated the local
buckling behaviour of compact lipped channel sections. Local buckling of slender
cold-formed steel beams is another area to be investigated. Interaction of local,
distortional and lateral-torsional buckling can be expected in beams at elevated
temperatures and therefore investigations on the interaction buckling behaviour are
also recommended.
This research was limited to lipped channel cold-formed steel sections. Further
research is needed to investigate the behaviour of other types of cold-formed steel
sections that are commonly used as flexural members.
This research was limited to cold-formed steel beams subject to a uniform

temperature distribution throughout the beam. But in practice the beams are
subjected to non-uniform temperature distributions in a fire situation. Therefore
further research is recommended in this area. Research work has already commenced
at the Queensland University of Technology to investigate the fire performance of
floor assemblies, which will focus on the behaviour of channel beams subjected to
non-uniform temperature distributions.
APPENDIX A: Tensile Coupon Test Results at Ambient
A.1 Tensile Coupon Test Results
400
350
300
250 Extensometer
Stress (MPa)
Strain Gauge
200
150
100
50
0
0 5 10 15 20 25 30 35
Strain (%)
1.95 mm G250 at 20oC
400
350
300
250
Stress (MPa)
200 Extensometer
150
100
50
0
0 2 4 6 8 10 12 14 16 18 20 22
Strain (%)
1.95 mm G250 at 100oC
Structural Behaviour and Design of Cold-Formed Steel Beams at Elevated Temperatures A-1
500
450
400
350
Stress (MPa)
300
Extensometer
250
200
150
100
50
0
0 2 4 6 8 10 12 14 16 18 20 22
Strain (%)
1.95 mm G250 at 200oC
450
400
350
300
Stress (MPa)
Extensometer
250
200
150
100
50
0
0 2 4 6 8 10 12 14 16 18 20 22
Strain (%)
1.95 mm G250 at 300oC
300
250
200
Stress (MPa)
Extensometer
150
100
50
0
0 2 4 6 8 10 12 14 16 18 20 22 24 26
Strain (%)
1.95 mm G250 at 400oC
160
140
120
100
Stress (MPa)
Extensometer
80
60
40
20
0
0 4 8 12 16 20 24 28 32
Strain (%)
1.95 mm G250 at 500oC
80
70
60
50
Stress (MPa)
40 Extensometer
30
20
10
0
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40
Strain (%)
1.95 mm G250 at 600oC
40
36
32
28
Stress (MPa)
24
20 Extensometer
16
12
0
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40 42
Strain (%)
1.95 mm G250 at 700oC
400
350
300
250
Stress (MPa)
Extensometer
200 Strain Gauge
150
100
50
0
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34
Strain (%)
1.55 mm G250 at 20oC
450
400
350
300
Stress (MPa)
250
Extensometer
200
150
100
50
0
0 2 4 6 8 10 12 14
Strain (%)
1.55 mm G250 at 100oC
500
450
400
350
Stress (MPa)
300
Extensometer
250
200
150
100
50
0
0 2 4 6 8 10 12 14 16 18 20 22
Strain (%)
1.55 mm G250 at 200oC
450
400
350
300
Stress (MPa)
250 Extensometer
200
150
100
50
0
0 2 4 6 8 10 12 14 16 18 20 22
Strain (%)
1.55 mm G250 at 300oC
300
250
200
Stress (MPa)
Extensometer
150
100
50
0
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28
Strain (%)
1.55 mm G250 at 400oC
160
140
120
100 Extensometer
Stress (MPa)
80
60
40
20
0
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34
Strain (%)
1.55 mm G250 at 500oC
100
90
80
70
60
Stress (MPa)
Extensometer
50
40
30
20
10
0
0 2 4 6 8 10 12 14 16 18 20 22 24
Strain (%)
1.55 mm G250 at 600oC
45
40
35
30
Stress (MPa)
25 Extensometer
20
15
10
0
0 2 4 6 8 10 12 14 16 18 20
Strain (%)
1.55 mm G250 at 700oC
600
500
400
Stress (MPa)
Extensometer
300
Strain Gauge
200
100
0
0 1 2 3 4 5 6 7 8 9 10 11 12
Strain (%)
1.9 mm G450 at 20oC
600
500
400
Stress (MPa)
Extensometer
300
200
100
0
0 1 2 3 4 5 6 7
Strain (%)
1.9 mm G450 at 100oC
600
550
500
450
400
Stress (MPa)
350 Extensometer
300
250
200
150
100
50
0
0 1 2 3 4 5 6 7 8 9
Strain (%)
1.9 mm G450 at 200oC
600
550
500
450
400
Stress (MPa)
350
Extensometer
300
250
200
150
100
50
0
0 1 2 3 4 5 6 7 8 9 10 11 12
Strain (%)
1.9 mm G450 at 300oC
450
400
350
300
Stress (MPa)
250
Extensometer
200
150
100
50
0
0 2 4 6 8 10 12 14 16 18 20 22
Strain (%)
1.9 mm G450 at 400oC
350
300
250
Stress (MPa)
200 Extensometer
150
100
50
0
0 1 2 3 4 5 6 7 8
Strain (%)
1.9 mm G450 at 450oC
280
240
200
Stress (MPa)
160
Extensometer
120
80
40
0
0 1 2 3 4 5 6 7 8 9 10
Strain (%)
1.9 mm G450 at 500oC
200
160
Stress (MPa)
120
Extensometer
80
40
0
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14
Strain (%)
1.9 mm G450 at 550oC
90
80
70
60
Stress (MPa)
50 Extensometer
40
30
20
10
0
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28
Strain (%)
1.9 mm G450 at 600oC
50
40
30
Stress (MPa)
Extensometer
20
10
0
0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34
Strain (%)
1.9 mm G450 at 700oC
600
550
500
450
400
Stress (MPa)
350 Extensometer
Strain Gauge
300
250
200
150
100
50
0
0 1 2 3 4 5 6 7 8 9 10 11
Strain (%)
1.5 mm G450 at 20oC
600
550
500
450
400
Stress (MPa)
350
Extensometer
300
250
200
150
100
50
0
0 1 2 3 4 5 6 7 8
Strain (%)
1.5 mm G450 at 100oC
650
600
550
500
450
400
Stress (MPa)
Extensometer
350
300
250
200
150
100
50
0
0 1 2 3 4 5 6 7
Strain (%)
1.5 mm G450 at 200oC
600
500
400 Extensometer
Stress (MPa)
300
200
100
0
0 1 2 3 4 5 6 7 8 9 10 11 12
Strain (%)
1.5 mm G450 at 300oC
450
400
350
300
Stress (MPa)
250 Extensometer
200
150
100
50
0
0 1 2 3 4 5 6 7 8 9 10 11
Strain (%)
1.5 mm G450 at 400oC
300
250
200
Stress (MPa)
Extensometer
150
100
50
0
0 1 2 3 4 5 6 7 8 9 10 11 12
Strain (%)
1.5 mm G450 at 500oC
200
180
160
140
Stress (MPa)
120
Extensometer
100
80
60
40
20
0
0 1 2 3 4 5 6 7 8 9 10 11 12 13
Strain (%)
1.5 mm G450 at 550oC
100
90
80
70
Stress (MPa)
60 Extensometer
50
40
30
20
10
0
0 2 4 6 8 10 12 14 16 18 20
Strain (%)
1.5 mm G450 at 600oC
50
40
30
Stress (MPa)
Extensometer
20
10
0
0 2 4 6 8 10 12 14 16
Strain (%)
1.5 mm G450 at 700oC
A.2 Typical Failure Modes of Tensile Specimens at Various

Temperatures
700oC 600oC 550oC 500oC 400oC 300oC 200oC 100oC 20oC
1.5mm G450 steel
700oC 600oC 500oC 400oC 300oC 200oC 100oC 20oC
1.5mm G250 steel
A.3 Measured Mechanical Properties and Reduction Factors
Table A.1: Yield Strengths in MPa at Elevated Temperatures for 1.5mm and 1.9mm thick Grade 250 and G450 cold-formed steel
1.55 mm G250 1.95 mm G250
Temperature σy,T σy,T
(oC) Specimen Number σy,T/σy,20 Specimen Number σy,T/σy,20
Average Average
1 2 3 1 2 3 4
20 297.0 290.0 293.5 1.000 270.9 269.0 273.1 269.0 270.5 1.000
100 279.0 279.0 279.0 0.951 265.0 278.0 259.0 267.3 0.988
200 279.0 270.0 274.5 0.935 260.0 256.0 255.0 257.0 0.950
300 184.8 191.0 187.9 0.640 198.0 193.0 205.0 189.4 196.4 0.726
400 144.5 144.5 0.492 148.0 142.0 153.0 147.7 0.546
500 94.9 94.9 0.323 98.0 96.6 92.7 95.8 0.354
600 66.3 66.3 0.226 55.5 52.0 54.5 54.0 0.200
700 34.0 37.4 35.7 0.122 32.2 36.3 34.2 0.127
1.5 mm G450 1.9 mm G450
Temperature σy,T σy,T
o
C Specimen Number σy,T/σy,20 Specimen Number σy,T/σy,20
Average Average
1 2 3 1 2 3 4
20 538.8 538.0 534.6 537.1 1.000 514.0 515.0 514.5 1.000
100 523.0 528.0 528.9 526.6 0.980 525.6 502.5 538.0 518.2 521.1 1.013
200 534.5 534.5 0.995 499.5 512.0 518.7 510.1 0.991
300 505.0 512.4 508.7 0.947 498.5 476.0 477.0 483.8 0.940
400 372.8 374.3 373.6 0.695 356.0 358.5 372.5 362.3 0.704
450 - 271.0 271.0 0.527
500 187.2 199.0 193.1 0.360 198.0 196.0 197.0 0.383
550 125.0 125.0 0.233 123.5 123.5 0.240
600 58.0 60.0 59.0 0.110 56.2 56.5 56.4 0.110
700 32.2 33.8 33.0 0.061 33.3 33.5 35.8 34.2 0.066
Table A.2: Modulus of Elasticity in MPa at Elevated Temperatures for 1.5mm and 1.9mm thick Grade 250 and G450 cold-formed steel
1.5 mm G250 1.9 mm G250
Temperature ET ΕΤ
o
C Specimen Number ET/E20 Specimen Number ET/E20
Average Average
1 2 3 1 2 3 4
20 200550 208220 204385 1.000 184550 189290 193330 185710 188220 1.000
100 189040 193970 191505 0.838 174400 184800 179720 179640 0.954
200 161110 160000 160555 0.744 173910 169580 171745 0.912
300 142470 142470 0.670 158220 155040 149730 154330 0.820
400 128220 128220 0.423 123690 120000 120000 121230 0.644
500 81096 81096 0.324 90556 90705 90631 0.482
600 62066 62066 0.130 55556 59998 57777 0.307
700 23014 26688 24851 0.000 30000 32725 31363 0.167
1.5 mm G450 1.9 mm G450
Temperature ET ΕΤ
o
C Specimen Number ET/E20 Specimen Number ET/E20
Average Average
1 2 3 1 2 3 4
20 205480 210960 206030 207490 1.000 211090 204430 200680 209110 206328 1.000
100 186830 191920 189375 0.913 194230 194740 197810 199380 196540 0.953
200 172000 172000 0.829 171250 173130 175630 173337 0.840
300 157380 157380 0.758 146790 150000 148395 0.719
400 126030 126030 0.607 115080 118440 122080 118533 0.574
450 - 101330 101330 0.491
500 100000 93600 96800 0.467 74300 79900 77100 0.374
550 78356 79452 78904 0.380 72222 72222 0.350
600 62000 64252 63126 0.304 51500 53571 52536 0.255
700 48000 47117 47559 0.229 23571 25000 24286 0.118
Table A.3: Ultimate Strengths in MPa at Elevated Temperatures for 1.5mm and 1.9mm thick Grade 250 and G450 cold-formed steel
1.5 mm G250 1.9 mm G250
σu,T σu,T
Temperature oC
Specimen σu,T/σu,20 Specimen Number σu,T/σu,20
Average Average
1 2 1 2 3 4
20 358.0 364.0 361.0 1.000 357.7 351.2 361.7 354.0 356.1 1.000
100 372.0 387.0 379.5 1.051 369.0 377.0 361.0 369.0 1.036
200 444.0 440.0 442.0 1.224 439.5 430.9 435.2 1.222
300 388.0 388.4 388.2 1.075 385.0 385.0 385.0 1.081
400 250.0 250.0 0.693 246.0 234.0 240.0 0.674
500 144.0 144.0 0.399 141.0 134.0 137.6 137.5 0.386
600 85.5 85.5 0.237 71.8 71.0 71.4 0.200
700 37.0 41.9 39.4 0.109 37.5 36.2 39.5 37.7 0.106
1.5 mm G450 1.9 mm G450
σu,T σu,T
Temperature oC
Specimen σu,T/σu,20 Specimen Number σu,T/σu,20
Average Average
1 2 1 2 3 4
20 563.8 559.7 561.8 1.000 548.0 545.0 540.0 537.0 542.5 1.000
100 565.0 558.0 561.5 1.000 564.0 541.0 577.0 554.0 559.0 1.030
200 590.0 604.0 597.0 1.063 555.9 565.0 565.0 562.0 1.036
300 560.0 559.9 560.0 0.997 553.0 528.0 529.0 536.7 0.989
400 403.0 402.2 402.6 0.717 388.0 385.0 405.0 392.7 0.724
450 - 313.5 313.5 0.578
500 238.0 241.0 239.5 0.426 235.0 237.5 236.3 0.435
550 237.0 168.0 202.5 0.360 165.0 165.0 0.304
600 89.0 89.5 89.3 0.159 95.5 82.4 88.9 0.164
700 40.0 43.3 41.7 0.074 40.9 42.0 42.6 41.8 0.077
Table A.4: Yield Strength and Elastic Modulus in MPa at Elevated Temperatures for 0.95 mm thick Grade 550 cold-formed steel
Yield Strength Elastic Modulus

Temperature oC
σy,T σy,T/σ20,T ET ET/E20
20 616.0 1.000 210960 1.000

100 617.7 1.003 184000 0.872
450 318.0 0.516 107290 0.509
500 223.0 0.362 92800 0.440
550 118.0 0.192 76000 0.360
600 57.0 0.093 55200 0.262
A.4 Comparison of Stress-Strain Models
300
250
200
Stress (MPa)
150
100
50
0
0.0 0.4 0.8 1.2 1.6 2.0 2.4 2.8 3.2 3.6 4.0
Strain (%)
Note: R – Ranawaka (2006)
Comparison of predicted Stress-strain curves by Ranawaka (2006) with test

results for 1.5 mm G250 Steel
300
250
200
Stress (MPa)
150
100
50
0
0.0 0.4 0.8 1.2 1.6 2.0 2.4 2.8 3.2 3.6 4.0
Strain (%)
Note: C – Chen and Young (2007)
Comparison of predicted Stress-strain curves by Chen and Young (2007) with

test results for 1.5 mm G250 Steel
650
600
550
500
450
400
Stress (MPa)
350
300
250
200
150
100
50
0
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0
Strain (%)
Note: R – Ranawaka (2006)
Comparison of predicted Stress-strain curves by Ranawaka (2006) with test

results for 1.5 mm G450 Steel
650
600
550
500
450
400
Stress (MPa)
350
300
250
200
150
100
50
0
0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0
Strain (%)
Note: C – Chen and Young (2007)
Comparison of predicted Stress-strain curves by Chen and Young (2007) with

test results for 1.5 mm G450 Steel
Appendix B: Sample Calculations for Section Moment
Capacity
B.1 Section Moment Capacity of G450-1.9-115-40-15 Cold-formed

Steel Beam at Ambient Temperature
Centreline Dimensions
d – 115 mm
b – 40 mm
b l – 15 mm
t - 1.9 mm
l Mechanical properties at 20oC,

f y , 20 = 515 MPa
E 20 = 206000 MPa
d Mechanical properties at 400oC,
k y , 400 = 0.694
k E , 400 = 0.580
l f y , 400 = k y , 400 × f y , 20 = 357.5 MPa
E 400 = k E , 400 × E 20 = 119500 MPa
E400 119500
G400 = = = 45961.5 MPa
2(1 + ν ) 2 × (1 + 0.3)
Section properties of the full section

I y = 106145 mm 4
I x = 886970 mm 4
I w = 312980629 mm6 Using CUFSM computer program

J = 514.42 mm4
Ix
Z= = 15425.6 mm3
d 2
S x = 17871.6 mm3
A = 427.5 mm 2
Structural Behaviour and Design of Cold-Formed Steel Beams at Elevated Temperatures B-1
B.1.1 Eurocode 3 Part 1.3 (ECS, 2006) Design Method – (Clause 6.1.4)
Calculation of effective section modulus

(Cl. 4.4 & Annex A.1(2) of Eurocode 3 Part 1.5)
Flange - Assuming flange is doubly supported at both edges.
ψ =1 kσ = 4 Uniformly compressed stiffened element
(Table 4.1 of Eurocode 3 Part 1.5)
_
λp
⎛ bp
= ⎜⎜ ⎟⎟ ⎢
( )
⎞ ⎡12 1 − ν f yb ⎤
2 0.5
( )
⎛ 40 ⎞ ⎡12 1 − 0.3 × 515 ⎤
= ⎜ ⎟⎢ 2
2 0.5
= 0.554 < 0.673

⎥ ⎥
⎠ ⎣⎢ π Ekσ ⎦⎥ ⎝ 1.9 ⎠ ⎣ π × 206000 × 4 ⎦
2
⎝ t
∴ρ =1 beff = ρb = 1 × 40 = 40 mm
Lip – Edge Stiffener
b pc 15 − 1.9 2
= = 0.369 > 0.35
bp 40 − 1.9
kσ = 0.5 + 0.833 (b pc b p − 0.35) = 0.5 + 0.833 (0.369 − 0.35) = 0.559

2 2
_
λp
⎛ bp
= ⎜⎜ ⎟⎟ ⎢
( )
⎞ ⎡12 1 − ν f yb ⎤
2 0.5
( )
⎛ 15 ⎞ ⎡ 12 1 − 0.3 × 515 ⎤
= ⎜ ⎟⎢ 2
2 0.5
= 0.555 < 0.673

⎥ ⎥
⎠ ⎣⎢ π Ekσ ⎦⎥ ⎝ 1.9 ⎠ ⎣ π × 206000 × 0.559 ⎦
2
⎝ t
∴ρ =1 Lip is fully effective.
ceff = ρl = 1 × 15 = 15 mm
Web – Doubly supported compression element.

σ2
ψ = = −1 kσ = 23.9 (Table 4.1)
σ1
_
λp
⎛ bp
= ⎜⎜ ⎟⎟ ⎢
( )
⎞ ⎡12 1 − ν f yb ⎤
2 0.5
( )
⎛ 115 ⎞ ⎡ 12 1 − 0.3 × 515 ⎤
=⎜
2 0.5
= 0.651 < 0.673

⎥ ⎟⎢ 2 ⎥
⎠ ⎢⎣ π Ekσ ⎥⎦ ⎝ 1.9 ⎠ ⎣ π × 206000 × 23.9 ⎦
2
⎝ t
∴ρ =1 Web is fully effective.
Z eff = Z f = 15425.6mm 3
Flange
_ _
λ e = λ p = 0.554
_
λ eo = 0.5 + (0.25 − 0.055 × (3 + ψ ))
_
λ eo = 0.5 + (0.25 − 0.055 × (3 + 1)) = 0.673
Lip
_ _
λ e = λ p = 0.559
_
λ eo = 0.673
Web
_ _
λ e = λ p = 0.651
_
λ eo = 0.5 + (0.25 − 0.055 × (3 + ψ ))
_
λ eo = 0.5 + (0.25 − 0.055 × (3 + (− 1))) = 0.874
Flange Lip Web

_ _ _
λe 0.554 λe 0.555 λe 0.651
_
= = 0.823 _
= = 0.825 _
= = 0.745
0.673 λ eo 0.673 0.874
λ eo λ eo
_ _
Maximum λ e,max =0.555 corresponding λ eo =0.673
⎧ ⎞⎫
M sx = f y ⎨Wel , y + (W pl , y − Wel , y ) × 4 × ⎜1 − λ x ,max λ eo ⎟⎬
⎛ _ _
⎩ ⎝ ⎠⎭
M sx = 515 × {15425.6 + (17871.6 − 15425.6) × 4 × (1 − 0.555 0.673)}× 10−6
M sx = 8.828 kNm
Section Moment Capacity = 8.828 kNm
B.1.2 Design Method Proposed by Shifferaw and Schafer (2008)
⎧⎪ ⎛ ⎛ ⎞ ⎞⎟⎫⎪
1.32
If λl < λly M nl = ⎨M y + (M p − M y )× 1 − ⎜ λ l λ ly ⎟
⎜ ⎬
⎪⎩ ⎜ ⎝ ⎠ ⎟⎠⎪⎭
⎝
My
Where λly = 0.776 λl = where Mcrl is the elastic local buckling moment.
M crl
M y = f y Z x = 515 × 15425.6 × 10− 6 = 7.944 kNm
M crl = 7.9442 * 2.80 = 22.243 kNm (From CUFSM analysis)
Figure B.1 shows a typical output of a CUFSM analysis that shows local buckling
minima used to calculate Mcrl. All the elastic local buckling moments obtained from
CUFSM analyses at ambient temperature are given in Table B.1.
Local Buckling Minima
Figure B.1: Output of a CUFSM Analyses
Table B.1: Elastic Critical Buckling Moments at Ambient and Elevated
Temperatures using CUFSM Analyses
Section Temperature oC My (kNm) Load Factor Mcrl (kNm)
20 2.894 3.460 10.013

200 2.633 3.230 8.505
300 1.903 3.760 7.155
G250-1.55-90-45-15 400 1.382 4.200 5.804
500 0.976 4.560 4.451
600 0.643 4.830 3.106
700 0.360 4.870 1.753
20 4.187 2.900 12.142
200 3.810 2.690 10.249
300 2.753 3.130 8.617
G250-1.55-120-45-15 400 2.000 3.500 7.000
500 1.412 3.800 5.366
600 0.930 4.020 3.739
700 0.521 4.060 2.115
20 4.631 3.170 14.679
200 4.214 2.960 12.473
300 3.045 3.440 10.475
G250-1.95-100-60-15 400 2.212 3.840 8.494
500 1.562 4.180 6.529
600 1.028 4.420 4.544
700 0.576 4.460 2.569
20 5.451 3.960 21.587
200 4.961 3.700 18.356
300 3.585 4.310 15.451
G250-1.95-125-50-15 400 2.604 4.810 12.525
500 1.839 5.230 9.618
600 1.211 5.540 6.709
700 0.678 6.940 4.705
20 6.120 3.250 19.889
200 5.569 3.030 16.874
300 4.024 3.530 14.205
G250-1.95-150-40-15 400 2.923 3.940 11.517
500 2.064 4.280 8.834
600 1.359 4.530 6.156
700 0.761 4.570 3.478
Table B.1: Elastic Critical Buckling Moments at Ambient and Elevated
Temperatures using CUFSM Analyses (Contd.)
Section Temperature oC My (kNm) Load Factor Mcrl (kNm)
20 7.705 2.770 21.343

200 7.101 2.590 18.392
300 5.131 3.010 15.444
G250-1.95-150-60-17 400 3.727 3.370 12.560
500 2.632 3.660 9.633
600 1.733 3.870 6.707
700 0.970 3.900 3.783
20 3.994 2.790 11.144
200 3.956 2.390 9.455
300 3.795 2.100 7.970
G450-1.5-90-30-14 400 2.772 2.330 6.459
500 1.561 2.300 3.590
600 0.439 7.860 3.451
700 0.280 6.970 1.952
20 4.964 3.720 18.468
200 4.917 3.190 15.685
300 4.716 2.800 13.205
G450-1.9-80-40-15 400 3.446 3.110 10.717
500 1.939 4.240 8.221
600 0.546 10.490 5.728
700 0.348 9.310 3.240
20 7.944 2.800 22.244
200 7.868 2.410 18.962
300 7.547 2.110 15.924
G450-1.9-115-40-15 400 5.514 2.340 12.903
500 3.104 3.190 9.902
600 0.874 7.910 6.913
700 0.556 7.010 3.898
My 7.944
λl = = = 0.598
M crl 22.243
M p = f y S x = 515 × 17871.9 × 10− 6 = 9.204 kNm
⎧⎪ ⎛ ⎛ 1.32
⎞⎫⎪ ⎧⎪ ⎛ 0.598 ⎞1.32 ⎫⎪
M nl = ⎨M y + (M p − M y )× ⎜1 − ⎜ λ l λ ly ⎟ ⎟⎬ = 7.944 + [9.204 − 7.944] × ⎨1 − ⎜
⎞
⎜ ⎝ ⎟⎪ ⎟ ⎬
⎪⎩ ⎝ ⎠ ⎠⎭ ⎪
⎩ ⎝ 0.776 ⎠ ⎪⎭
⎧⎪ ⎛ 0.598 ⎞1.32 ⎫⎪
M nl = 7.944 + [9.204 − 7.944]× ⎨1 − ⎜ ⎟ ⎬ = 8.311 kNm
⎪⎩ ⎝ 0.776 ⎠ ⎪⎭
Section Moment Capacity = 8.311 kNm
B.1.3 Comparison of Section Moment Capacities from Various

Sources at Ambient Temperature
Sources Ms (kNm)
EC 3 Part 1.3 (ECS, 2006) 8.828
Design Method Proposed by Shifferaw and Schafer (2008) 8.311
B.2 Section Moment Capacity of G450-1.9-115-40-15 Cold-formed
Steel Beam at 400oC
B.2.1 Eurocode 3 Part 1.3 (ECS, 2006) Design Method and Design Method
Proposed by Shifferaw and Schafer (2008)
Calculations to obtain the section moment capacities at elevated temperatures are

similar to the previous section. Section moment capacities can be obtained by simply
replacing the ambient temperature mechanical properties with those at elevated
temperatures.
Mechanical properties at 20oC,

f y , 20 = 515 MPa
Centreline Dimensions E 20 = 206000 MPa

d – 115 mm Mechanical properties at 400oC,
b – 40 mm k y , 400 = 0.694
l – 15 mm
k E , 400 = 0.580
t - 1.9 mm
f y , 400 = k y , 400 × f y , 20 = 357.5 MPa
E 400 = k E , 400 × E 20 = 119500 MPa
E400 119500
G400 = = = 45961.5
2(1 + ν ) 2 × (1 + 0.3)
Section Moment Capacity according to Eurocode 3 Part 1.3 = 5.854 kNm
Section Moment Capacity according to Shifferaw and Schafer (2008) = 5.691 kNm
(CUFSM results required to calculate the section moment capacities at elevated
temperatures are given in Table B.1)
B.2.2 Eurocode 3 Part 1.2 (ECS, 2005)
M fi ,θ ,Rd = k y ,θ M Rd
Ambient temperature section moment capacity according to Eurocode 3 Part 1.3,

M Rd = 8.828 kNm (Calculations are similar to Section B.1.1)
k y , 400 = 0.694
M fi , 400,Rd = 0.694 × 8.828 = 6.127 kNm
Section Moment Capacity at 400oC = 6.127 kNm
B.1.3 Comparison of Section Moment Capacities from Various

Sources at 400oC
Sources Ms (kNm)
EC 3 Part 1.3 (ECS, 2006) 5.854

Design Method Proposed by Shifferaw and Schafer (2008) 5.691
EC 3 Part 1.2 (ECS, 2005) 6.127
Appendix C: Sample Calculations for Lateral-torsional
Buckling Capacity
Lateral-Torsional Buckling Moment Capacity of G450-1.9-1154015 cold-formed

steel beam with 1.4 m span at 400oC elevated temperature
Centreline Dimensions
d – 115 mm
b – 40 mm
b l – 15 mm
t - 1.9 mm
l Mechanical properties at 20oC,
f y , 20 = 515 MPa
E 20 = 206000 MPa
d Mechanical properties at 400oC,
k y , 400 = 0.694
k E , 400 = 0.580
l
f y , 400 = k y , 400 × f y , 20 = 357.5 MPa
E400 = k E , 400 × E20 = 119500 MPa
E400 119500
G400 = = = 45961.5 MPa
2(1 +ν ) 2 × (1 + 0.3)
Section properties of the full section
I y = 106145 mm 4
I x = 886970 mm 4
I w = 312980629 mm6 Using CUFSM computer program

J = 514.42 mm 4
Ix
Z= = 15425.6 mm3
d 2
A = 427.5 mm 2
Structural Behaviour and Design of Cold-formed Steel Beams at Elevated Temperatures C-1
C.1 AS/NZS 4600 (SA, 2005) Design Method - (Clause 3.3.3)
Cb = 1 for uniform moment distribution along the beam
π 2 EI y ⎡ π 2 EI w ⎤
M o = Cb Arol f oy f oz = Cb 2 ⎢GJ + 2 ⎥
ley ⎣ lez ⎦
π 2 × 119500 × 106145 ⎡ π 2 × 119500 × 312980629 ⎤ −6

M o , 400 = 1× 2 ⎢45961.5 × 514.42 + 2 ⎥ × 10
1400 ⎣ 1400 ⎦
= 3.68 kNm
M y , 400 = Z f f y = 15425.6 × 357.5 × 10 −6 = 5.515 kNm
My 5.515
λb = = = 1.224
Mo 3.680
⎡ ⎛ 10λb2 ⎞⎤
M c = 1.11M y ⎢1 − ⎜⎜ ⎟⎟⎥ for 0.60 < λb < 1.336
⎢⎣ ⎝ 36 ⎠⎥⎦
⎡ ⎛ 10 × 1.224 2 ⎞⎤
M c , 400 = 1.11× 5.515 × ⎢1 − ⎜⎜ ⎟⎟⎥ = 3.574 kNm
⎣ ⎝ 36 ⎠⎦
⎛ M ⎞ ⎛ 3.574 × 106 ⎞
f c , 400 = ⎜ c ⎟ = ⎜⎜ ⎟⎟ = 231.7 MPa
⎜Z ⎟ 15425.6
⎝ f ⎠ ⎝ ⎠
Effective section modulus, Z c @ f c
Flange – Uniformly compressed element with an edge stiffener (Clause 2.4.2)
⎡E⎤ 119500
S = 1.28 ⎢ * ⎥ = 1.28 × = 29.07
⎣f ⎦ 231.7
0.328S = 0.328 × 29.07 = 9.53

b 40
= = 21.05 > 0.328S
t 1.9
d 3t 153 × 1.9
Is = Sin θ =
2
= 534.4 [θ = 90 o ∴ Sinθ = 1]
12 12
3
⎡b t ⎤ ⎡ bt ⎤
I a = 399t ⎢ 4
− 0.328⎥ ≤ t 4 ⎢115 + 5⎥
⎣S ⎦ ⎣ S ⎦
3 3
⎡b t ⎤ ⎡ 40 1.9 ⎤
399t ⎢ 4
− 0.328⎥ = 399 × 1.9 4 ⎢ − 0.328⎥ = 323.4
⎣S ⎦ ⎣ 29.07 ⎦
⎡ bt ⎤ ⎡ 40 1.9 ⎤
t 4 ⎢115 + 5⎥ = 1.9 4 ⎢115 + 5⎥ = 1150.5
⎣ S ⎦ ⎣ 29.07 ⎦
∴ I a = 323.4
Is > Ia ∴ I s = I a = 323.4
⎡ b t⎤ 1
n = ⎢0.582 − ≥
⎣ 4S ⎥⎦ 3
⎡ 40 1.9 ⎤ 1
n = ⎢0.582 − ⎥ = 0.4 >
⎣ 4 × 29.07 ⎦ 3
d l 15
= = 0.375 > 0.25
b 40
n
⎡ 5d ⎤ ⎡ I ⎤
∴ k = ⎢4.82 − l ⎥ ⎢ s ⎥ + 0.43 ≤ 4
⎣ b ⎦⎣ I a ⎦
0.4
⎡ 5 × 15 ⎤ ⎡ 323.4 ⎤
k = ⎢4.82 − + 0.43 = 3.375 < 4
⎣ 40 ⎥⎦ ⎢⎣ 323.4 ⎥⎦
2
⎡ kπ 2 E ⎤⎛ t ⎞
f cr =⎢ ⎥⎜ ⎟
(
⎣12 1 − υ
2
)
⎦⎝ b ⎠
2
⎡ 3.375 × π 2 × 119500 ⎤⎛ 1.9 ⎞
f cr ( flange ) = ⎢ ⎥⎜ ⎟ = 822.44 MPa
⎣ (
12 1 − 0.3 2 ) ⎦⎝ 40 ⎠
⎡ f* ⎤ 231.7
λ=⎢ ⎥= = 0.531 < 0.673 ∴ρ =1 be ( flange ) = b = 40 mm
⎢⎣ f cr ⎥⎦ 822.44
Flange is fully effective.
Lip – Edge stiffener with stress gradient (Clause 2.3.2.2)
f 2* 85 0.578 0.578
ψ= = = 0.739 k= = = 0.536
f1* 115 ψ + 0.34 0.739 + 0.34
2
⎡ 0.536 × π 2 ×119500 ⎤⎛ 1.9 ⎞
f cr ( web ) = ⎢ ⎥⎜ 15 ⎟ = 928.2 MPa
⎣ 12 1 (
− 0.32
)⎦⎝ ⎠
⎡ f*⎤ 231.7
λ=⎢ ⎥= = 0.45 < 0.673 ∴ρ =1 be ( lip ) = d se = 15 mm
⎢⎣ f cr ⎥⎦ 928.2
Lip is fully effective.
Web – Stiffened element with stress gradient (Clause 2.2.3.2)
f 2* 231.7
ψ = * = k = 4 + 2(1 −ψ ) + 2(1 −ψ ) = 24
3
= −1
f1 − 231.7
2
⎡ 24 × π 2 ×119500 ⎤⎛ 1.9 ⎞
f cr ( web ) =⎢ ⎥⎜ 115 ⎟ = 707.6 MPa
(
⎣ 12 1 − 0.3
2
) ⎦⎝ ⎠
⎡ f*⎤ 231.7
λ=⎢ ⎥= = 0.572 < 0.673 ∴ρ =1 be ( web ) = d = 115 mm
⎣⎢ f cr ⎥⎦ 707.6
Web is fully effective.
∴ Z c = Z f = 15425.6 mm3
⎛M ⎞
Lateral-torsional buckling capacity, M b , 400 = Z c ⎜ c ⎟ = 3.574 kNm
⎜Z ⎟
⎝ f ⎠
C.2 Direct Strength Method (DSM)
Lateral-torsional buckling capacity, M be
M o = 3.68 kNm
(From AS/NZS4600 calculations)
M y = Z f f y = 5.515 kNm
∴ 2.78M y ≥ M o ≥ 0.56M y
⎡ ⎛ 10M y ⎞⎤ 10 10 × 5.515 ⎞
⎟⎟⎥ = × 5.515⎛⎜1 −
10
M be = M y ⎢1 − ⎜⎜ ⎟ = 3.577 kNm
9 ⎢⎣ ⎝ 36M o ⎠⎥⎦ 9 ⎝ 36 × 3.680 ⎠
The nominal local buckling capacity, M bl
M ol = 5.515 × 2.34 = 12.905 kNm (From CUFSM analysis)
M be 3.577
λl = = = 0.526
M ol 12.905
∴ λl (0.526) < 0.776 M bl = M be = 3.577 kNm
The distortional buckling capacity, M bd
M od = 5.515 × 1.52 = 8.383 kNm (From CUFSM analysis)
My 5.515
λd = = = 0.811
M od 8.383
⎡ ⎛M ⎞
0.5
⎤⎛ M 0.5
⎞
∴ λ d (0.881) > 0.673 M bd = ⎢1 − 0.22⎜ od ⎟ ⎥⎜ od ⎟ My
⎢ ⎜M ⎟ ⎥⎜⎝ M y ⎟
⎣ ⎝ y ⎠ ⎦ ⎠
⎡ ⎛ 8.383 ⎞ ⎤⎛ 8.383 ⎞
0.5 0.5
M bd = ⎢1 − 0.22⎜ ⎟ ⎥⎜ ⎟ × 5.515 = 4.955 kNm

⎣⎢ ⎝ 5.515 ⎠ ⎦⎥⎝ 5.515 ⎠
M b , 400 = Lesser of M be , M bl and M bd =3.577 kNm
C.3 BS5950 Part 5 (BCS, 1998) Design Method - (Clause 5.6)
M EMY
Mb = (Cl. 5.6.2.1)
φ B + φ B2 − M E M Y
β = 1 for uniform bending moment
C b = 1.75 − 1.05β + 0.3β 2 = 1.75 − 1.05 + 0.3 = 1
Iy 106145
ry = = = 15.76 mm
A 427.5
Le 1400
= = 88.85 > 40C b
ry 15.76
⎛L ⎞
∴ η = 0.002⎜ e − 40C b ⎟ = 0.002(88.85 − 40 ) = 0.0977
⎜r ⎟
⎝ y ⎠
12
π AED 2 ⎧ 1 ⎛⎜ LE t ⎞⎟ ⎫⎪
2
⎪
ME = C b ⎨1 +
2(LE ry )
2 ⎜ ⎟ ⎬
⎪⎩ 20 ⎝ ry D ⎠ ⎪⎭
12
π 2 × 427.5 × 119500 × (115 + 1.9) ⎧⎪ ⎞ ⎫⎪
2
1 ⎛ 1400 1.9
M E , 400 = × 1× ⎨1 + ⎜⎜ × ⎟⎟ ⎬ × 10 −6
2 × (1400 15.76) ⎪⎩ 20 ⎝ 15.76 (115 + 1.9) ⎠ ⎪⎭
2
M E , 400 = 3.924 kNm
M y , 400 = f y Z c = 357.5 × 15425.6 × 10 −6 = 5.515 kNm
M Y + (1 + η )M E 5.515 + (1 + 0.0977 ) × 3.924

φB = = = 4.911
2 2
M EMY 3.924 × 5.515

M b , 400 = = = 3.337 kNm
φB + φ − M E M Y
2
B 4.911 + 4.911 − 3.924 × 5.515
2
Lateral-torsional buckling capacity, M b, 400 = 3.337 kNm
C.4 Eurocode 3 Part 1.3 (ECS, 2006) Design Method - (Clause 6.3)
Calculation of effective section modulus

(Cl. 4.4 & Annex A.1(2) of Eurocode 3 Part 1.5)
Flange - Assuming flange is doubly supported at both edges.
ψ =1 kσ = 4 Uniformly compressed stiffened element (Table 4.1)
_
λp
⎛ bp
= ⎜⎜ ⎟⎟ ⎢
( )
⎞ ⎡12 1 − ν f yb ⎤
2 0.5
⎛ 40 ⎞ ⎡12(1 − 0.3 ) × 357.5 ⎤
= ⎜ ⎟⎢ 2
2 0.5
= 0.606 < 0.673

⎥ ⎥
⎠ ⎣⎢ π Ekσ ⎦⎥ ⎝ 1.9 ⎠ ⎣ π × 119500 × 4 ⎦
2
⎝ t
∴ρ =1 beff = ρb = 1 × 40 = 40 mm
b pc 15 − 1.9 2
= = 0.369 > 0.35
bp 40 − 1.9
kσ = 0.5 + 0.833 (b pc b p − 0.35) = 0.5 + 0.833 (0.369 − 0.35) = 0.559

2 2
_
λp
⎛ bp
= ⎜⎜ ⎟⎟ ⎢
( )
⎞ ⎡12 1 − ν f yb ⎤
2 0.5
⎛ 15 ⎞ ⎡ 12(1 − 0.3 )× 357.5 ⎤
= ⎜ ⎟⎢ 2
2 0.5
= 0.607 < 0.673

⎥ ⎥
⎠ ⎣⎢ π Ekσ ⎦⎥ ⎝ 1.9 ⎠ ⎣ π × 119500 × 0.559 ⎦
2
⎝ t
ceff = ρl = 1 × 15 = 15 mm
Web – Doubly supported compression element.

σ2
ψ = = −1 kσ = 23.9 (Table 4.1)
σ1
_
λp
⎛ bp
= ⎜⎜ ⎟⎟ ⎢
( )
⎞ ⎡12 1 − ν f yb ⎤
2 0.5
( )
⎛ 115 ⎞ ⎡12 1 − 0.3 × 357.5 ⎤
=⎜
2 0.5
= 0.712 > 0.673

⎥ ⎟⎢ 2 ⎥
⎠ ⎣⎢ π Ekσ ⎦⎥ ⎝ 1.9 ⎠ ⎣ π × 119500 × 23.9 ⎦
2
⎝ t
−
λ p − 0.055(3 + ψ ) ⎡ 0.712 − 0.055 × (3 − 1)⎤
ρ= −
=⎢ ⎥ = 1.188 ≥ 1
⎣ 0.712 2 ⎦
λ 2
p
∴ρ =1 Web is fully effective.

Z eff = Z f = 15425.6mm 3
Calculation of Section Moment Capacity, Mcx (Same as Msx)
Flange
_ _
λ e = λ p = 0.606
_
λ eo = 0.5 + (0.25 − 0.055 × (3 + ψ ))
_
λ eo = 0.5 + (0.25 − 0.055 × (3 + 1)) = 0.673
Lip
_ _
λ e = λ p = 0.607
_
λ eo = 0.673
Web
_ _
λ e = λ p = 0.712
_
λ eo = 0.5 + (0.25 − 0.055 × (3 + ψ ))
_
λ eo = 0.5 + (0.25 − 0.055 × (3 + (− 1))) = 0.874
Flange Lip Web

_ _ _
λe 0.606 λe 0.607 λe 0.712
_
= = 0.9 _
= = 0.9 _
= = 0.815
0.673 0.673 0.874
λ eo λ eo λ eo
Maximum λe,max=0.607 corresponding λe,o=0.673
⎧ ⎫
M = f ⎨W + ⎛W −W ⎞ × 4 × ⎛1 − λ
⎜ λ ⎞⎟⎬
cx y ⎩ el, y ⎜⎝ pl , y el , y ⎟⎠ ⎝ x, max eo ⎠⎭
M cx, 400 = 357.5{15425.6 + (17871.9 − 15425.6)× 4 × (1 − 0.607 0.673)}×10−6
M cx, 400 = 5.858 kNm
Calculation of Lateral-torsional Buckling Capacity, Mb,400
M o , 400 = 3.680 kNm (Similar to AS/NZS 4600 calculation)
_
M sx 5.858
λ LT = = = 1.262
Mo 3.68
Buckling Curve Chosen = ‘b’ α LT = 0.34
⎡ ⎛_ ⎞ _
⎤
φ LT = 0.5⎢1 + α LT ⎜ λ LT − 0.2 ⎟ + λ LT 2 ⎥ = 0.5⎡1 + 0.34(1.262 − 0.2) + 1.262 2 ⎤ = 1.477
⎣ ⎝ ⎠ ⎦ ⎢⎣ ⎥⎦
1 1
χ LT = = = 0.446 ≤ 1.0
⎡ _
⎤
φ LT + ⎢φ LT 2 − λ LT 2 ⎥
0.5
[
1.477 + 1.477 2 − 1.262 2 ]
0.5
⎣ ⎦
M b , Rd = χ LT M sx γ M 1 = 0.446 × 5.858 = 2.61 kNm
Lateral-torsional buckling capacity, M b , 400 = 2.61 kNm
C.5 Eurocode 3 Part 1.2 (ECS, 2005) Design Method
(Clause 4.2.3.3 & 4.2.3.4)
Calculation of effective section modulus in accordance with Eurocode 3 Part 1.3

(ECS, 2006) based on mechanical properties at ambient temperature [Annex E.2,
EC3 Part 1.2 (ECS, 2005).
Flange - Assuming flange is supported at both edges.
ψ =1 kσ = 4 Uniformly compressed stiffened element
(Table 4.1Eurocode 3 Part 1.5)

235 235
ε = 0.85 = 0.85 = 0.574
fy 515
−
_
λp = ⎜⎜ ⎟⎟ ⎢
( )
⎛ b p ⎞ ⎡12 1 − ν f yb ⎤
2 0.5
≅
b
t =
40 1.9
= 0.646 < 0.673
⎥
⎝ t ⎠ ⎣⎢ π Ekσ ⎦⎥ 28.4ε kσ
2
28.4 × 0.574 × 4
∴ρ =1
beff = ρb = 1 × 40 = 40 mm
Flange is fully effective.

b pc 15 − 1.9 2
= = 0.369 > 0.35
bp 40 − 1.9
kσ = 0.5 + 0.833 (b pc b p − 0.35) = 0.5 + 0.833 (0.369 − 0.35) = 0.559

2 2
−
_
λp
⎛ bp
= ⎜⎜ ⎟⎟ ⎢
( )
⎞ ⎡12 1 − ν f yb ⎤
2 0.5
≅
b
t =
15 1.9
= 0.648 < 0.673
⎥
⎠ ⎣⎢ π Ekσ ⎦⎥ 28.4ε kσ
2
⎝ t 28.4 × 0.574 × 0.559

ceff = ρl = 1 × 15 = 15 mm
Web –Internal compression element.
ψ = −1 kσ = 23.9 (Table 4.1of Eurocode 3 Part 1.5)
−
_
λ p = ⎜⎜ ⎟⎟ ⎢
(
⎛ b p ⎞ ⎡12 1 − ν
2
)f yb
⎤
0.5
≅
b
t =
115 1.9
= 0.759 > 0.673
⎥
⎝ t ⎠ ⎢⎣ π Ekσ
2
⎥⎦ 28.4ε kσ 28.4 × 0.574 × 23.9
−
λ p − 0.055(3 + ψ ) ⎡ 0.759 − 0.055 × (3 − 1)⎤
ρ= −
=⎢ ⎥ = 1.127 ≤ 1
⎣ 0.759 2 ⎦
λ 2
p
∴ρ =1
beff = ρb = 1 × 40 = 40 mm
Web is fully effective.
Weff = Wel , y = 15425.6 mm 3
Calculation of Section Moment Capacity, M sx at 20oC
Flange
_ _
λ e = λ p = 0.646
_
λ eo = 0.5 + (0.25 − 0.055 × (3 + ψ ))
_
λ eo = 0.5 + (0.25 − 0.055 × (3 + 1)) = 0.673
Lip
_ _
λ e = λ p = 0.648
_
λ eo = 0.673
Web
_ _
λ e = λ p = 0.759
_
λ eo = 0.5 + (0.25 − 0.055 × (3 + ψ ))
_
λ eo = 0.5 + (0.25 − 0.055 × (3 + (− 1))) = 0.874
Flange Lip Web
_ _ _
λe 0.646 λe 0.648 λe 0.759
_
= = 0.96 _
= = 0.963 _
= = 0.868
0.673 0.673 0.874
λ eo λ eo λ eo
Maximum λe,max=0.648 corresponding λe,o=0.673
M sx = f y {Wel , y + (W pl , y − Wel , y ) × 4 × (1 − λ x ,max λeo )}
M sx = 515 × {15425.6 + (17871.9 − 15425.6) × 4 × (1 − 0.648 0.673)}× 10 −6
M sx = 8.131 kNm
Calculation of Lateral-torsional Buckling Capacity, M b , 400 at 400oC
E 20 206000
G20 = = = 79230.8 MPa
2(1 + ν ) 2 × (1 + 0.3)
M o at 20oC
π 2 EI y ⎡ π 2 EI w ⎤
M o , 20 = 2 ⎢GJ + 2 ⎥
ley ⎣ lez ⎦
π 2 × 206000 × 106145 ⎡ π 2 × 206000 × 312980629 ⎤

M o , 20 = ⎢79230.8 × 514.42 + ⎥
1400 2 ⎣ 1400 2 ⎦
= 6.343 kNm
_
M sx 8.131
λ LT = = = 1.132
M o , 20 6.343
[ ]
0.5
⎡k ⎤ 0.5
λ LT ,θ ,com = λ LT ⎢ y ,θ ,com ⎥ = 1.132 × 0.694 = 1.238
⎣ k E ,θ ,com ⎦ 0.580
α = 0.65 235 f = 0.65 × 235 515 = 0.439

y
Φ LT ,θ ,com =
1
2
( [
1 + αλ LT ,θ ,com + λ LT ,θ ,com ] ) = 12 (1 + 0.439 × 1.238 + 1.238 ) = 1.538
2 2
1
χ LT , fi =
⎛Φ
⎜ LT ,θ ,com + [Φ LT ,θ ,com ] − [λ
2
LT ,θ ,com ]
2 ⎞
⎟
⎝ ⎠
1
=
(1.538 + 1.538 2 − 1.238 2 ) = 0.408
M b , fi ,t ,Rd = χ LT , fi k y ,θ ,com M sx , 20 γ M 1 = 0.408 × 0.694 × 8.131 = 2.303 kNm
Lateral-torsional buckling capacity at 400oC, M b , 400 = 2.303 kNm
C.6 Dharma and Tan (2007) - Alternative Approach
M cx, 20 = 8.131kNm (Calculation similar to Eurocode 3 Part 1.2 (ECS, 2005))
M o, 20 = 6.343kNm (Calculation similar to Eurocode 3 Part 1.2 (ECS, 2005))

Mo 6.343
M E , 20 = = = 6.76 kNm
(1 − EI y EI x ) (1 − 106145 886970)
M o, 400 = 3.68 kNm (Calculation similar to AS/NZS 4600 (SA, 2005))
Mo 3.68
M E , 400 = = = 3.922 kNm
(1 − EI y EI x ) (1 − 106145 886970)
f pT 280
ζ = T
= = 0.783
p y 357.5
π 2 E ⎧⎪⎡ M cx ⎤ kE ⎫⎪
η LT
T
= 0.007 ζ ⎨⎢ ⎥ − 0.4ζ
1.5
⎬
py ⎪⎩⎢⎣ M E ⎥⎦ ky ⎪⎭
π 2 × 206000 ⎧⎪⎡ 8.131 ⎤ 0.58 ⎫⎪

η T
LT = 0.007 0.783 ⎨⎢ ⎥ − 0.4 × 0.783
1.5
⎬ = 0.328
515 ⎪⎩⎣ 6.76 ⎦ 0.694 ⎪⎭
π 2 E M cx π 2 × 206000 8.131
λ LT = = = 68.91
py ME 515 6.76
π 2 k E E π 2 k E E π 2 × 0.58 × 206000
pET = = 2 = = 248.33 MPa
λ2LT λLT 68.912
k y p y + (η LT
T
+ 1) pET 0.694 × 515 + (0.328 + 1) × 248.33
φ T
LT = = = 343.6 MPa
2 2
k y p y pET 0.694 × 515 × 248.33
pb, fi = = = 172.41 MPa
φLT
T
+ φLT
T
( 2
− k y p y pET ) 343.6 + 343.6 2 − 0.694 × 515 × 248.33
M b, fi = pb , fi Z x = 172.41×15425.6 ×10−6 = 2.66 kNm
Lateral-torsional buckling capacity, M b , 400 = 2.66 kNm
C. 7 Dharma and Tan (2007) – Rankine Approach
M cx, 20 = 8.131 kNm (Calculation similar to Eurocode 3 Part 1.2 (ECS, 2005))
M o, 20 = 6.343 kNm (Calculation similar to Eurocode 3 Part 1.2 (ECS, 2005))

Mo 6.343
M E , 20 = = = 6.76 kNm
(1 − EI y EI x ) (1 − 106145 886970)
M o, 400 = 3.68 kNm (Calculation similar to AS/NZS 4600 (SA, 2005))
Mo 3.68
M E , 400 = = = 3.922 kNm
(1 − EI y EI x ) (1 − 106145 886970)
π 2 E M cx π 2 × 206000 8.131
λ LT = = = 68.91
py ME 515 6.76
λ2LT ⎛ 68.912 ⎞
ψ LT
T
= = ⎜ ⎟ = 1.203
π E p y ⎝ π × 206000 515 ⎟⎠
2 ⎜ 2
k y py 0.694 × 515
pb, fi = = = 146.51 MPa
⎛ ky ⎞ ⎛ 0.694 ⎞
⎜1 + ψ LT ⎟ ⎜1 + ×1.203 ⎟
⎜ py ⎟ ⎝ 0.580 ⎠
⎝ ⎠
M b, fi = pb , fi Z x = 146.51×15425.6 ×10−6 = 2.26 kNm
Lateral-torsional buckling capacity, M 400 = 2.26 kNm
C.8 Comparison of Lateral-Torsional Buckling Capacities from
Various Sources at 400oC
Sources Mb,400 (kNm)
AS/NZS 4600 (SA, 2005) 3.57
DSM 3.57
BS 5950 Part 5 (BSI, 1998) 3.34
EC 3 Part 1.3 (ECS, 2006) 2.61
EC 3 Part 1.2 (ECS, 2005) 2.30
Dharma and Tan (2007) - Alternate Approach 2.66
Dharma and Tan (2007) - Rankine Approach 2.26
Appendix D: Sample Calculations for the Plotting of
Lateral-torsional Buckling Capacity Curves
in Non-dimensional Format
D.1 Plotting Beam Design Curves

M y , 400 M b , 400
X Coordinate = Y Coordinate =
M o , 400 M y , 400
Calculation of coordinates for 1400 mm long G450-1.9-115-40-15 section at 400oC.

M y , 400 = 5.514 kNm
M o , 400 = 3.68 kNm (Using Mo Equations)
M b, 400 = 3.57 kNm

M y , 400 5.514
X Coordinate = = = 1.224
M o , 400 3.68
M b , 400 3.57
Y Coordinate = = = 0.648
M y , 400 5.51
⎛ M y , 400 M b , 400 ⎞
Po int ≡ ⎜ , ⎟ ≡ (1.224,0.648)
⎜ M o, 400 M y , 400 ⎟
⎝ ⎠
Mo,400 Mb,400 X Y Coordinates

Sources
(kNm) (kNm) Coordinate Coordinate of Point
AS/NZS 4600 (SA, 2005) 3.68 3.57 1.224 0.648 (1.224 , 0.648)
DSM 3.68 3.57 1.224 0.648 (1.224 , 0.648)
EC 3 Part 1.3 (ECS, 2006) 3.68 2.61 1.224 0.464 (1.224 , 0.474)
EC 3 Part 1.2 (ECS, 2005) 3.68 2.30 1.224 0.413 (1.224 , 0.417)
Dharma and Tan (2007) -
3.68 2.66 1.224 0.490 (1.224 , 0.483)
Alternate Approach
Dharma and Tan (2007) -
3.68 2.26 1.224 0.415 (1.224 , 0.410)
Rankine Approach
Refer Appendix C for the procedure to calculate Mb,400.
Behaviour and Design of Cold-Formed Steel Beams at Elevated Temperatures D-1

D.2 Plotting Finite Element Analysis Results Points
Calculation of coordinates for 1400 mm long G450-1.9-115-40-15 section at 400oC

obtained from Finite Element Analysis.
My,400 = 5.514 kNm
FEA Results
Mu,400 = 2.782 kNm
Mo,400 = 3.583 kNm
M y , 400 5.514
X Coordinate = = = 1.241
M o , 400 3.583
M b , 400 3.583
Y Coordinate = = = 0.650
M y , 400 5.514
⎛ M y , 400 M b, 400 ⎞
Po int ≡ ⎜ , ⎟ ≡ (1.241,0.650)
⎜ M o, 400 M y , 400 ⎟
⎝ ⎠
Behaviour and Design of Cold-Formed Steel Beams at Elevated Temperatures D-2

Appendix E: Sample Calculations for Capacity Reduction
Factor (ϕ)
Calculation of capacity reduction factor for predictions of AS/NZS 4600 (SA, 2005)
for 1.55 mm thick G250 steels in comparison with Finite Element Analysis Results
and with reference to Table 6.6.
− β 0 Vm 2 +V f 2 + C pV p 2 +Vq 2
φ = 1.52 M m Fm Pm e = 1.672 Pm e
− 2.5 0.0566+ C pV p 2
Mm, Vm = Mean and coefficient of variation of the material factor = 1.1, 0.1
Fm, Vf = Mean and coefficient of variation of the fabrication factor = 1, 0.05
Vq = coefficient of variation of load effect = 0.21
β0 = Target reliability index = 2.5 for cold-formed steel members
Number of specimens n
m = Degree of freedom = n-1
Cp = Correction factor depending on the number of tests
⎛ 1 ⎞⎛ m ⎞
= ⎜1 + ⎟⎜ ⎟
⎝ n ⎠⎝ m − 2 ⎠
Pm = mean value of FEA to predicted load ratio
Vp = Coefficient of variation of FEA to predicted load ratio
n = 29
Pm = 0.85
V p = 0.058
⎛ 1 ⎞⎛ m ⎞ ⎛ 1 ⎞⎛ 29 − 1 ⎞
C p = ⎜1 + ⎟⎜ ⎟ = ⎜1 + ⎟⎜⎜ ⎟⎟ = 1.114
⎝ n ⎠⎝ m − 2 ⎠ ⎝ 29 ⎠⎝ (29 − 1) − 2 ⎠
− 2.5 0.0566 + C pV p 2
φ = 1.672 Pm e
0.0566 +1.114×0.0582
φ = 1.672 × 0.85 × e −2.5 = 0.769
Structural Behaviour and Design of Cold-formed Steel Beams at Elevated Temperatures E-1
Appendix F: Finite Element Analysis Results of Section
Moment Capacities and Predictions using
Available Design Methods
FEA Msx Mu/Msx

Temp. (oC)
Span (mm)
Schafer (2008)
Schafer (2008)
Shifferaw and
Shifferaw and
Section
EC 3 Part 1.3
EC 3 Part 1.2
EC 3 Part 1.3
EC 3 Part 1.2
My
Mu/My
Mo Mu
100 12.035 3.130 1.0816 1.01 1.01 1.03

20 500 2.894 10.199 2.900 1.0021 3.109 3.109 3.042 0.93 0.93 0.95
800 10.165 2.960 1.0228 0.95 0.95 0.97
100 10.226 2.83 1.075 1.02 1.00 1.03
200 500 2.633 8.666 2.625 0.997 2.787 2.829 2.759 0.94 0.93 0.95
800 8.637 2.67 1.014 0.96 0.94 0.97
100 8.606 2.2 1.156 1.06 1.08 1.10
300 500 1.903 7.293 1.915 1.006 2.081 2.044 2.009 0.92 0.94 0.95
G250-1.55-904515
800 7.269 1.915 1.006 0.92 0.94 0.95

100 6.982 1.65 1.194 1.07 1.11 1.12
400 500 1.382 5.917 1.44 1.042 1.544 1.485 1.467 0.93 0.97 0.98
800 5.897 1.455 1.053 0.94 0.98 0.99
100 5.357 1.125 1.153 1.02 1.07 1.08
500 500 0.976 4.540 1.02 1.045 1.107 1.049 1.040 0.92 0.97 0.98
800 4.524 1.04 1.066 0.94 0.99 1.00
100 3.732 0.71 1.105 0.97 1.03 1.03
600 500 0.643 3.163 0.675 1.050 0.729 0.690 0.686 0.93 0.98 0.98
800 3.152 0.685 1.066 0.94 0.99 1.00
100 2.107 0.3955 1.099 0.97 1.02 1.03
700 500 0.360 1.786 0.3755 1.044 0.408 0.387 0.384 0.92 0.97 0.98
800 1.780 0.383 1.065 0.94 0.99 1.00
100 12.629 4.570 1.092 1.01 1.01 1.04
G250-1.55-1204515
20 500 4.187 11.575 4.405 1.052 4.538 4.538 4.381 0.97 0.97 1.01
800 11.755 4.345 1.038 0.96 0.96 0.99
100 10.731 4.155 1.091 1.02 1.01 1.05
200 500 3.810 9.835 3.985 1.046 4.061 4.130 3.967 0.98 0.96 1.00
800 9.988 3.93 1.031 0.97 0.95 0.99
Structural Behaviour and Design of Cold-formed Steel Beams at Elevated Temperatures F-1
FEA Msx Mu/Msx
Temp. (oC)
Span (mm)
Schafer (2008)
Schafer (2008)
Shifferaw and
Shifferaw and
Section
EC 3 Part 1.3
EC 3 Part 1.2
EC 3 Part 1.3
EC 3 Part 1.2
My
Mu/My
Mo Mu
100 9.031 3.25 1.180 1.07 1.09 1.12

300 500 2.753 8.277 2.85 1.035 3.043 2.984 2.895 0.94 0.96 0.98
800 8.406 2.795 1.015 0.92 0.94 0.97
100 7.326 2.395 1.198 1.06 1.10 1.13
400 500 2.000 6.714 2.135 1.068 2.260 2.168 2.117 0.94 0.98 1.01
G250-1.55-1204515
800 6.819 2.11 1.055 0.93 0.97 1.00

100 5.621 1.61 1.140 0.99 1.05 1.07
500 500 1.412 5.152 1.51 1.069 1.626 1.531 1.502 0.93 0.99 1.01
800 5.232 1.515 1.073 0.93 0.99 1.01
100 3.916 1.065 1.145 1.00 1.06 1.07
600 500 0.930 3.589 0.995 1.070 1.070 1.008 0.992 0.93 0.99 1.00
800 3.645 1 1.075 0.93 0.99 1.01
100 2.211 0.59 1.133 0.98 1.05 1.06
700 500 0.520 2.027 0.555 1.066 0.599 0.564 0.556 0.93 0.98 1.00
800 2.058 0.56 1.076 0.93 0.99 1.01
100 18.943 4.990 1.078 1.04 1.04 1.04
20 500 4.631 14.894 4.670 1.009 4.817 4.817 4.816 0.97 0.97 0.97
800 12.199 4.585 0.990 0.95 0.95 0.95
100 16.096 4.305 1.022 1.00 0.98 0.99
200 4.214 4.322 4.383 4.368
500 12.655 4.225 1.003 0.98 0.96 0.97
100 13.547 3.61 1.186 1.12 1.14 1.14
300 3.045 3.221 3.168 3.179
G250-1.95-1006015
500 10.651 2.975 0.977 0.92 0.94 0.94

100 10.989 2.655 1.200 1.11 1.15 1.14
400 2.212 2.387 2.301 2.32
500 8.640 2.22 1.004 0.93 0.96 0.96
100 8.432 1.805 1.156 1.06 1.11 1.10
500 1.562 1.71 1.625 1.644
500 6.629 1.585 1.015 0.93 0.98 0.96
100 5.875 1.145 1.113 1.01 1.07 1.06
600 1.028 1.137 1.070 1.085
500 4.618 1.055 1.026 0.93 0.99 0.97
100 3.317 0.635 1.103 1.00 1.06 1.04
500 2.608 0.595 1.034 0.93 0.99 0.98
700 0.576 0.637 0.599 0.608
500 3.730 0.64 1.151 0.99 1.04 1.05
800 3.790 0.615 1.106 0.95 1.00 1.01
FEA Msx Mu/Msx
Temp. (oC)
Span (mm)
Schafer (2008)
Schafer (2008)
Shifferaw and
Shifferaw and
Section
EC 3 Part 1.3
EC 3 Part 1.2
EC 3 Part 1.3
EC 3 Part 1.2
My
Mu/My
Mo Mu
100 23.306 6.100 1.119 0.98 0.98 1.05

20 5.451 6.206 6.206 5.795
500 16.032 5.700 1.046 0.92 0.92 0.98
G250-1.95-1255015
200 100 4.961 19.803 5.55 1.018 5.571 5.647 5.255 1.00 0.98 1.06
300 100 3.585 16.666 4.605 1.285 4.103 4.081 3.827 1.12 1.13 1.20
400 100 2.604 13.520 3.325 1.277 2.98 2.964 2.794 1.12 1.12 1.19
500 100 1.839 10.373 2.23 1.213 2.104 2.093 1.98 1.06 1.07 1.13
600 100 1.211 7.227 1.425 1.177 1.386 1.378 1.307 1.03 1.03 1.09
700 100 0.678 3.825 0.785 1.158 0.776 0.772 0.738 1.01 1.02 1.06
100 21.842 7.000 1.144 0.97 0.97 1.08
20 6.120 7.213 7.213 6.511
500 18.914 6.700 1.095 0.93 0.93 1.03
100 18.559 6.35 1.140 0.97 0.97 1.08
200 5.569 6.564 6.564 5.895
500 16.071 6.1 1.095 0.93 0.93 1.03
100 15.619 5.25 1.305 1.11 1.11 1.22
300 4.024 4.743 4.743 4.306
G250-1.95-1504015
500 13.525 4.555 1.132 0.96 0.96 1.06

100 12.671 3.805 1.302 1.10 1.10 1.21
400 2.923 3.445 3.445 3.15
500 10.972 3.37 1.153 0.98 0.98 1.07
100 9.722 2.525 1.223 1.04 1.04 1.13
500 2.064 2.433 2.433 2.235
500 8.419 2.36 1.143 0.97 0.97 1.06
100 6.773 1.605 1.181 1.00 1.00 1.09
600 1.359 1.602 1.602 1.477
500 5.865 1.54 1.133 0.96 0.96 1.04
100 3.825 0.895 1.176 1.00 1.00 1.08
700 0.761 0.897 0.897 0.827
500 3.312 0.86 1.130 0.96 0.96 1.04
Temp. (oC) FEA Msx Mu/Msx
Span (mm)
Schafer (2008)
Schafer (2008)
Shifferaw and
Shifferaw and
Section
EC 3 Part 1.3
EC 3 Part 1.2
EC 3 Part 1.3
EC 3 Part 1.2
My
Mu/My
Mo Mu
100 24.338 8.450 1.097 1.04 1.04 1.05

20 500 7.705 20.891 8.100 1.051 8.087 8.087 8.025 1.00 1.00 1.01
1000 20.820 7.950 1.032 0.98 0.98 0.99
100 20.680 7.65 1.091 1.06 1.04 1.05
200 7.012 7.233 7.359 7.271
500 17.751 7.3 1.041 1.01 0.99 1.00
G250-1.95-1506017
100 17.404 5.95 1.174 1.10 1.12 1.12

300 5.067 5.426 5.318 5.304
500 14.940 5.15 1.016 0.95 0.97 0.97
100 14.119 4.255 1.156 1.05 1.10 1.10
400 3.68 4.039 3.863 3.878
500 12.119 3.865 1.050 0.96 1.00 1.00
100 10.833 3.03 1.166 1.04 1.11 1.10
500 2.599 2.902 2.728 2.751
500 9.299 2.75 1.058 0.95 1.01 1.00
100 7.547 1.955 1.142 1.01 1.09 1.08
600 1.711 1.932 1.796 1.816
500 6.478 0 0.000 0.00 0.00 0.00
700 100 0.958 4.262 1.085 1.133 1.084 1.00541 1.017 1.00 1.08 1.07
100 11.310 4.405 1.103 1.03 1.03 1.05
20 3.994 4.287 4.287 4.188
500 10.791 4.110 1.029 0.96 0.96 0.98
100 9.610 4.325 1.093 1.07 1.02 1.07
200 3.956 4.058 4.246 4.035
500 9.169 3.915 0.990 0.96 0.92 0.97
100 8.088 4.04 1.065 1.08 0.99 1.06
300 3.795 3.728 4.073 3.805
G450-1.5-903014
500 7.717 3.63 0.957 0.97 0.89 0.95

100 6.561 3.02 1.089 1.07 1.01 1.12
400 2.772 2.821 2.976 2.69
500 6.260 2.71 0.977 0.96 0.91 1.01
100 5.034 1.795 1.150 1.04 1.07 1.12
500 1.561 1.734 1.675 1.606
500 4.803 1.625 1.041 0.94 0.97 1.01
100 3.507 0.555 1.263 1.08 1.18 1.11
600 0.439 0.513 0.472 0.5
500 3.346 0.52 1.183 1.01 1.10 1.04
100 1.980 0.3415 1.221 1.05 1.14 1.11
700 0.28 0.326 0.300 0.308
500 1.890 0.327 1.169 1.00 1.09 1.06
FEA Msx Mu/Msx
Temp. (oC)
Span (mm)
Schafer (2008)
Schafer (2008)
Shifferaw and
Shifferaw and
Section
EC 3 Part 1.3
EC 3 Part 1.2
EC 3 Part 1.3
EC 3 Part 1.2
My
Mu/My
Mo Mu
100 11.310 4.405 1.103 1.03 1.03 1.05

20 3.994 4.287 4.287 4.188
500 10.791 4.110 1.029 0.96 0.96 0.98
100 9.610 4.325 1.093 1.07 1.02 1.07
200 3.956 4.058 4.246 4.035
500 9.169 3.915 0.990 0.96 0.92 0.97
100 8.088 4.04 1.065 1.08 0.99 1.06
300 3.795 3.728 4.073 3.805
G450-1.5-903014
500 7.717 3.63 0.957 0.97 0.89 0.95

100 6.561 3.02 1.089 1.07 1.01 1.12
400 2.772 2.821 2.976 2.69
500 6.260 2.71 0.977 0.96 0.91 1.01
100 5.034 1.795 1.150 1.04 1.07 1.12
500 1.561 1.734 1.675 1.606
500 4.803 1.625 1.041 0.94 0.97 1.01
100 3.507 0.555 1.263 1.08 1.18 1.11
600 0.439 0.513 0.472 0.5
500 3.346 0.52 1.183 1.01 1.10 1.04
100 1.980 0.3415 1.221 1.05 1.14 1.11
700 0.28 0.326 0.300 0.308
500 1.890 0.327 1.169 1.00 1.09 1.06
50 21.363 5.550 1.118 1.02 1.02 1.06
20 4.964 5.444 5.444 5.248
500 17.200 5.050 1.017 0.93 0.93 0.96
100 18.152 5.45 1.108 1.05 1.01 1.07
200 4.917 5.213 5.392 5.102
500 15.985 5.1 1.037 0.98 0.95 1.00
100 15.277 5.2 1.103 1.07 1.01 1.07
300 4.716 4.843 5.172 4.838
G450-1.9-804015
500 13.453 4.755 1.008 0.98 0.92 0.98

100 12.393 3.86 1.120 1.06 1.02 1.09
400 3.446 3.631 3.779 3.547
500 10.913 3.52 1.022 0.97 0.93 0.99
100 9.509 2.25 1.160 1.03 1.06 1.11
500 1.939 2.183 2.127 2.024
500 8.374 2.08 1.072 0.95 0.98 1.03
100 6.625 0.685 1.254 1.10 1.14 1.13
600 0.546 0.622 0.599 0.606
500 5.834 0.64 1.172 1.03 1.07 1.06
100 3.741 0.428 1.232 1.08 1.12 1.13
700 0.348 0.396 0.381 0.38
500 3.294 0.4025 1.158 1.02 1.06 1.06
FEA Msx Mu/Msx
Temp. (oC)
Span (mm)
Schafer (2008)
Schafer (2008)
Shifferaw and
Shifferaw and
Section
EC 3 Part 1.3
EC 3 Part 1.2
EC 3 Part 1.3
EC 3 Part 1.2
My
Mu/My
Mo Mu
100 23.043 8.900 1.120 1.01 1.01 1.07

20 500 7.944 21.303 8.150 1.026 8.824 8.824 8.312 0.92 0.92 0.98
800 18.656 7.850 0.988 0.89 0.89 0.94
100 19.580 8.55 1.087 1.02 0.98 1.05
200 500 7.868 18.101 7.85 0.998 8.411 8.739 8.14 0.93 0.90 0.96
800 18.391 7.5 0.953 0.89 0.86 0.92
100 16.478 8.1 1.073 1.04 0.97 1.05
300 500 7.547 15.234 7.25 0.961 7.779 8.383 7.722 0.93 0.86 0.94
800 15.477 6.95 0.921 0.89 0.83 0.90
G450-1.9-1154015
100 13.367 6.05 1.097 1.03 0.99 1.06

400 500 5.514 12.358 5.4 0.979 5.854 6.125 5.691 0.92 0.88 0.95
800 12.556 5.15 0.934 0.88 0.84 0.90
100 10.257 3.54 1.141 1.00 1.03 1.08
500 500 3.104 9.482 3.23 1.041 3.55 3.447 3.276 0.91 0.94 0.99
800 9.634 3.1 0.999 0.87 0.90 0.95
100 7.146 1.1 1.259 1.09 1.13 1.14
600 500 0.874 6.606 1.025 1.173 1.012 0.971 0.963 1.01 1.06 1.06
800 6.712 0.985 1.127 0.97 1.01 1.02
100 4.035 0.685 1.232 1.06 1.11 1.12
700 500 0.556 3.730 0.64 1.151 0.644 0.618 0.61 0.99 1.04 1.05
800 3.790 0.615 1.106 0.95 1.00 1.01
Appendix G: Finite Element Analysis Results of Member
Moment Capacities and Predictions using
Available Design Methods
FEA (kNm) Mb (kNm)
Dharma and Tan (2007)

Alternative Approach -
New Design Method

Rankine Approach -
Temp. (oC)
based on Option 3
Span (mm)
My (kNm)
Section
AS/NZS 4600
New Proposal
EC 3 Part 1.3
EC 3 Part 1.2
EC 3 Part 1.3
(ECS, 2006)
(ECS, 2005)
(SA, 2005)
Mo Mu
20 1600 3.680 2.373 2.545 2.063 1.568 2.017 1.750 2.259 2.289
20 1800 2.993 2.163 2.380 1.854 1.421 1.813 1.594 2.021 2.067
20 2000 2.485 1.925 2.201 1.658 1.287 1.627 1.453 1.804 1.846
20 2500 1.681 1.384 1.700 1.256 1.013 1.250 1.166 1.365 1.379
2.8939 20 3000 1.232 1.012 1.246 0.978 0.814 0.987 0.955 1.057 1.059
20 3500 0.957 0.768 0.965 0.787 0.672 0.804 0.799 0.845 0.844
20 4000 0.774 0.634 0.780 0.653 0.567 0.673 0.682 0.695 0.695
20 4500 0.647 0.532 0.652 0.556 0.489 0.576 0.593 0.586 0.588
20 5000 0.555 0.458 0.558 0.483 0.429 0.503 0.523 0.505 0.508
200 1600 3.127 2.128 2.273 1.806 1.386 1.794 1.549 1.990 2.007
200 1800 2.543 1.919 2.112 1.614 1.251 1.606 1.406 1.770 1.800
200 2000 2.112 1.686 1.937 1.437 1.128 1.434 1.277 1.572 1.598
G250-1.55-904515
200 2500 1.429 1.180 1.451 1.081 0.881 1.093 1.018 1.178 1.184
2.6334 200 3000 1.047 0.861 1.059 0.838 0.704 0.859 0.829 0.907 0.906
200 3500 0.813 0.640 0.820 0.674 0.579 0.697 0.691 0.722 0.721
200 4000 0.658 0.539 0.663 0.558 0.488 0.582 0.589 0.593 0.593
200 4500 0.550 0.451 0.554 0.475 0.420 0.498 0.510 0.500 0.501
200 5000 0.471 0.388 0.474 0.412 0.368 0.434 0.450 0.430 0.433
300 1600 2.631 1.233 1.709 1.420 1.067 1.301 1.188 1.293 1.123
300 1800 2.140 1.163 1.609 1.284 0.972 1.197 1.087 1.177 1.006
300 2000 1.777 1.093 1.501 1.155 0.885 1.096 0.996 1.067 0.903
300 2500 1.202 0.896 1.198 0.883 0.703 0.873 0.806 0.834 0.701
1.9030 300 3000 0.881 0.704 0.891 0.691 0.569 0.703 0.664 0.659 0.560
300 3500 0.684 0.558 0.690 0.558 0.471 0.579 0.558 0.533 0.461
300 4000 0.554 0.452 0.558 0.464 0.399 0.487 0.478 0.442 0.389
300 4500 0.463 0.378 0.466 0.395 0.345 0.419 0.416 0.375 0.336
300 5000 0.397 0.326 0.399 0.343 0.303 0.367 0.368 0.324 0.295
Structural Behaviour and Design of Cold-formed Steel Beams at Elevated Temperatures G-1
FEA (kNm) Mb (kNm)

New Design Method

Rankine Approach -
Temp. (oC)
based on Option 3
Span (mm)
My (kNm)
Section
AS/NZS 4600
New Proposal
EC 3 Part 1.3
EC 3 Part 1.2
EC 3 Part 1.3
(ECS, 2006)
(ECS, 2005)
(SA, 2005)
Mo Mu
400 1400 3.781 1.450 1.908 1.700 1.247 1.453 1.360 1.552 1.374
400 1600 2.960 1.338 1.817 1.554 1.143 1.338 1.243 1.422 1.230
400 1800 2.382 1.293 1.715 1.405 1.042 1.224 1.135 1.295 1.101
400 2000 1.960 1.203 1.603 1.261 0.949 1.115 1.036 1.173 0.986
400 2500 1.303 0.955 1.285 0.956 0.751 0.877 0.830 0.911 0.759
1.9999
400 3000 0.941 0.741 0.949 0.739 0.603 0.697 0.676 0.713 0.600
400 3500 0.720 0.578 0.725 0.589 0.495 0.567 0.561 0.570 0.488
400 4000 0.575 0.464 0.579 0.484 0.415 0.472 0.474 0.467 0.407
400 4500 0.475 0.386 0.477 0.407 0.355 0.401 0.408 0.391 0.347
400 5000 0.402 0.326 0.404 0.350 0.309 0.348 0.357 0.334 0.302
500 1400 2.901 1.090 1.365 1.245 0.903 1.047 0.986 1.185 1.146
500 1600 2.271 1.041 1.305 1.145 0.832 0.964 0.905 1.090 1.044
500 1800 1.827 0.962 1.239 1.042 0.763 0.883 0.830 0.996 0.944
G250-1.55-1204515
500 2000 1.504 0.885 1.167 0.940 0.698 0.806 0.761 0.906 0.850
500 2500 1.000 0.720 0.960 0.720 0.558 0.638 0.615 0.708 0.655
1.4122
500 3000 0.722 0.565 0.728 0.560 0.451 0.510 0.504 0.557 0.515
500 3500 0.552 0.443 0.556 0.448 0.372 0.417 0.420 0.447 0.415
500 4000 0.441 0.356 0.444 0.368 0.313 0.349 0.357 0.367 0.344
500 4500 0.364 0.294 0.366 0.310 0.268 0.297 0.308 0.307 0.292
500 5000 0.308 0.250 0.310 0.267 0.234 0.258 0.270 0.263 0.252
600 1400 2.021 0.738 0.906 0.839 0.604 0.703 0.660 0.819 0.774
600 1600 1.582 0.714 0.869 0.775 0.558 0.648 0.608 0.755 0.707
600 1800 1.273 0.665 0.828 0.708 0.514 0.593 0.559 0.692 0.642
600 2000 1.048 0.611 0.783 0.642 0.472 0.542 0.514 0.631 0.580
600 2500 0.697 0.493 0.654 0.495 0.380 0.430 0.418 0.496 0.450
0.9298
600 3000 0.503 0.393 0.507 0.387 0.309 0.345 0.344 0.392 0.355
600 3500 0.385 0.308 0.388 0.310 0.255 0.283 0.288 0.315 0.287
600 4000 0.307 0.248 0.309 0.255 0.215 0.237 0.245 0.259 0.238
600 4500 0.254 0.206 0.255 0.215 0.185 0.202 0.212 0.217 0.202
600 5000 0.215 0.174 0.216 0.185 0.161 0.176 0.186 0.186 0.175
FEA (kNm) Mb (kNm)

New Design Method

Rankine Approach -
Temp. (oC)
based on Option 3
Span (mm)
My (kNm)
Section
AS/NZS 4600
New Proposal
EC 3 Part 1.3
EC 3 Part 1.2
EC 3 Part 1.3
(ECS, 2006)
(ECS, 2005)
(SA, 2005)
Mo Mu
700 1400 1.141 0.425 0.508 0.471 0.339 0.382 0.370 0.474 0.435
700 1600 0.893 0.398 0.487 0.436 0.313 0.351 0.341 0.436 0.398
700 1800 0.719 0.383 0.464 0.398 0.289 0.322 0.314 0.399 0.361
G250-1.55-1204515
700 2000 0.592 0.354 0.439 0.361 0.265 0.294 0.289 0.364 0.327
700 2500 0.393 0.283 0.368 0.279 0.214 0.235 0.235 0.286 0.253
0.5205
700 3000 0.284 0.223 0.287 0.218 0.174 0.189 0.193 0.225 0.200
700 3500 0.217 0.173 0.219 0.175 0.144 0.156 0.162 0.181 0.162
700 4000 0.174 0.141 0.175 0.144 0.121 0.131 0.138 0.149 0.134
700 4500 0.143 0.116 0.144 0.121 0.104 0.112 0.119 0.125 0.114
700 5000 0.121 0.098 0.122 0.104 0.091 0.097 0.105 0.107 0.098
20 1800 6.964 4.083 4.282 3.517 2.690 3.589 3.088 3.966 3.858
20 2000 5.885 3.887 4.094 3.269 2.495 3.327 2.878 3.650 3.620
20 2500 4.053 3.211 3.565 2.662 2.060 2.723 2.416 2.936 2.967
20 3000 2.978 2.428 2.966 2.155 1.710 2.230 2.044 2.365 2.380
4.6305
20 3500 2.305 1.788 2.338 1.768 1.439 1.853 1.750 1.931 1.932
20 4000 1.857 1.459 1.877 1.480 1.231 1.568 1.518 1.606 1.602
20 4500 1.544 1.281 1.558 1.264 1.070 1.350 1.334 1.362 1.358
20 5000 1.316 1.076 1.326 1.099 0.943 1.182 1.187 1.176 1.173
200 1800 5.917 3.700 3.841 3.097 2.388 4.214 3.793 4.214 3.408
200 2000 5.000 3.496 3.658 2.866 2.206 2.965 2.551 3.217 3.181
G250-1.95-1006015
200 2500 3.443 2.811 3.143 2.312 1.807 2.405 2.126 2.560 2.574
200 3000 2.530 2.117 2.558 1.859 1.491 1.956 1.789 2.045 2.049
4.2138
200 3500 1.958 1.628 1.986 1.519 1.249 1.617 1.525 1.661 1.656
200 4000 1.578 1.263 1.595 1.269 1.065 1.363 1.318 1.377 1.370
200 4500 1.312 1.085 1.324 1.082 0.923 1.171 1.156 1.165 1.160
200 5000 1.118 0.925 1.127 0.940 0.812 1.024 1.026 1.004 1.001
300 1800 4.980 2.046 2.861 2.403 1.821 2.810 2.780 2.769 1.938
300 2000 4.208 1.992 2.748 2.245 1.696 2.145 1.952 2.006 1.782
300 2500 2.898 1.788 2.428 1.851 1.415 1.838 1.652 1.674 1.447
300 3000 2.129 1.557 2.065 1.510 1.184 1.556 1.407 1.389 1.190
3.0450
300 3500 1.648 1.308 1.672 1.246 1.002 1.320 1.211 1.159 0.996
300 4000 1.328 1.076 1.342 1.046 0.860 1.131 1.055 0.979 0.849
300 4500 1.104 0.907 1.114 0.895 0.750 0.982 0.931 0.838 0.737
300 5000 0.941 0.772 0.948 0.779 0.662 0.864 0.830 0.729 0.649
FEA (kNm) Mb (kNm)

New Design Method

Rankine Approach -
Temp. (oC)
based on Option 3
Span (mm)
My (kNm)
Section
AS/NZS 4600
New Proposal
EC 3 Part 1.3
EC 3 Part 1.2
EC 3 Part 1.3
(ECS, 2006)
(ECS, 2005)
(SA, 2005)
Mo Mu
400 1800 4.040 1.637 2.117 1.828 1.370 2.084 2.038 2.208 1.487
400 2000 3.414 1.583 2.044 1.719 1.284 1.578 1.473 1.628 1.374
400 2500 2.351 1.441 1.836 1.440 1.084 1.359 1.260 1.370 1.127
400 3000 1.727 1.254 1.600 1.190 0.917 1.161 1.083 1.145 0.934
2.2118
400 3500 1.337 1.050 1.343 0.989 0.782 0.993 0.939 0.961 0.786
400 4000 1.077 0.872 1.089 0.834 0.676 0.857 0.823 0.815 0.673
400 4500 0.896 0.737 0.904 0.716 0.591 0.749 0.729 0.700 0.586
400 5000 0.764 0.626 0.769 0.625 0.524 0.662 0.652 0.610 0.517
500 1800 3.100 1.210 1.514 1.334 0.992 1.525 1.448 1.562 1.604
500 2000 2.619 1.156 1.466 1.260 0.933 1.139 1.069 1.253 1.154
500 2500 1.804 1.076 1.331 1.068 0.795 0.982 0.922 1.061 0.967
500 3000 1.325 0.952 1.178 0.891 0.678 0.841 0.797 0.891 0.806
1.5618
500 3500 1.026 0.807 1.011 0.745 0.582 0.722 0.695 0.751 0.678
500 4000 0.827 0.672 0.836 0.631 0.505 0.626 0.612 0.639 0.578
G250-1.95-1006015
500 4500 0.687 0.564 0.693 0.543 0.444 0.548 0.544 0.550 0.501
500 5000 0.586 0.479 0.590 0.474 0.394 0.486 0.488 0.480 0.440
600 1800 2.160 0.847 1.005 0.898 0.663 1.028 0.957 1.028 0.831
600 2000 1.825 0.820 0.975 0.850 0.626 0.778 0.716 0.888 0.780
600 2500 1.257 0.757 0.891 0.727 0.537 0.667 0.621 0.754 0.658
600 3000 0.923 0.666 0.796 0.610 0.460 0.569 0.539 0.635 0.552
1.0284
600 3500 0.715 0.564 0.692 0.512 0.397 0.488 0.472 0.536 0.466
600 4000 0.576 0.470 0.581 0.435 0.345 0.423 0.417 0.456 0.398
600 4500 0.479 0.392 0.483 0.375 0.304 0.371 0.371 0.393 0.345
600 5000 0.408 0.334 0.411 0.328 0.271 0.329 0.334 0.344 0.304
700 1800 1.219 0.487 0.563 0.504 0.372 0.568 0.536 0.576 0.467
700 2000 1.030 0.475 0.547 0.478 0.351 0.416 0.402 0.503 0.438
700 2500 0.710 0.430 0.500 0.409 0.302 0.359 0.349 0.427 0.370
700 3000 0.521 0.381 0.447 0.343 0.259 0.308 0.303 0.360 0.311
0.5757
700 3500 0.404 0.319 0.389 0.289 0.223 0.266 0.265 0.304 0.262
700 4000 0.325 0.265 0.328 0.245 0.194 0.231 0.234 0.259 0.224
700 4500 0.270 0.221 0.273 0.212 0.171 0.203 0.209 0.223 0.195
700 5000 0.230 0.188 0.232 0.185 0.152 0.181 0.188 0.195 0.171
FEA (kNm) Mb (kNm)

New Design Method

Rankine Approach -
Temp. (oC)
based on Option 3
Span (mm)
My (kNm)
Section
AS/NZS 4600
New Proposal
EC 3 Part 1.3
EC 3 Part 1.2
EC 3 Part 1.3
(ECS, 2006)
(ECS, 2005)
(SA, 2005)
Mo Mu
20 1400 10.039 4.701 5.188 4.630 3.339 4.303 3.699 4.944 5.060
20 1600 7.902 4.439 4.936 4.226 3.054 3.908 3.381 4.487 4.679
20 1800 6.384 4.110 4.656 3.816 2.785 3.528 3.087 4.048 4.252
20 2000 5.273 3.743 4.352 3.423 2.536 3.174 2.820 3.639 3.815
20 2500 3.531 2.810 3.492 2.600 2.014 2.445 2.266 2.784 2.862
5.4513
20 3000 2.566 2.049 2.592 2.019 1.625 1.926 1.853 2.164 2.191
20 3500 1.976 1.628 1.992 1.618 1.341 1.561 1.546 1.726 1.737
20 4000 1.588 1.303 1.599 1.335 1.131 1.301 1.315 1.415 1.422
20 4500 1.318 1.083 1.326 1.129 0.973 1.108 1.137 1.189 1.196
20 5000 1.122 0.925 1.129 0.975 0.851 0.963 0.999 1.019 1.028
200 1400 8.530 4.255 4.665 4.084 2.972 3.850 3.291 4.393 4.477
200 1600 6.714 3.979 4.420 3.708 2.706 3.482 2.996 3.964 4.113
200 1800 5.424 3.651 4.147 3.330 2.457 3.128 2.726 3.556 3.712
G250-1.95-1255015
200 2000 4.481 3.296 3.850 2.973 2.228 2.802 2.481 3.180 3.310
200 2500 3.001 2.430 3.012 2.240 1.755 2.140 1.980 2.408 2.460
4.9607
200 3000 2.180 1.786 2.202 1.732 1.409 1.676 1.611 1.860 1.876
200 3500 1.679 1.379 1.692 1.385 1.158 1.354 1.339 1.478 1.484
200 4000 1.349 1.107 1.359 1.141 0.974 1.125 1.135 1.208 1.214
200 4500 1.136 0.921 1.127 0.965 0.837 0.957 0.980 1.014 1.020
200 5000 0.954 0.785 0.959 0.833 0.731 0.831 0.859 0.868 0.876
300 1400 7.179 2.416 3.457 3.156 2.253 2.705 2.497 2.645 2.562
300 1600 5.651 2.298 3.305 2.900 2.072 2.521 2.293 2.446 2.303
300 1800 4.565 2.180 3.136 2.636 1.899 2.333 2.104 2.248 2.069
300 2000 3.771 2.062 2.952 2.379 1.738 2.146 1.929 2.058 1.860
300 2500 2.525 1.733 2.431 1.826 1.394 1.721 1.564 1.635 1.444
3.5848
300 3000 1.835 1.392 1.853 1.425 1.133 1.386 1.287 1.304 1.152
300 3500 1.413 1.131 1.424 1.146 0.939 1.136 1.079 1.058 0.944
300 4000 1.136 0.923 1.143 0.947 0.794 0.952 0.921 0.876 0.793
300 4500 0.943 0.779 0.949 0.802 0.685 0.814 0.799 0.741 0.681
300 5000 0.803 0.661 0.807 0.693 0.600 0.709 0.703 0.638 0.595
FEA (kNm) Mb (kNm)

New Design Method

Rankine Approach -
Temp. (oC)
based on Option 3
Span (mm)
My (kNm)
Section
AS/NZS 4600
New Proposal
EC 3 Part 1.3
EC 3 Part 1.2
EC 3 Part 1.3
(ECS, 2006)
(ECS, 2005)
(SA, 2005)
Mo Mu
400 1400 5.824 1.904 2.551 2.396 1.688 1.987 1.873 2.134 1.962
400 1600 4.584 1.812 2.452 2.219 1.563 1.854 1.731 1.981 1.775
400 1800 3.693 1.733 2.342 2.035 1.443 1.720 1.597 1.829 1.603
400 2000 3.059 1.628 2.222 1.851 1.328 1.588 1.473 1.681 1.449
400 2500 2.049 1.366 1.884 1.442 1.080 1.288 1.207 1.348 1.136
2.6038
400 3000 1.489 1.127 1.500 1.135 0.886 1.049 1.002 1.083 0.911
400 3500 1.146 0.915 1.155 0.916 0.739 0.867 0.845 0.883 0.750
400 4000 0.921 0.746 0.928 0.760 0.628 0.732 0.725 0.733 0.632
400 4500 0.765 0.625 0.769 0.645 0.543 0.629 0.632 0.621 0.543
400 5000 0.651 0.534 0.655 0.558 0.477 0.549 0.558 0.536 0.476
500 1400 4.468 1.405 1.820 1.745 1.219 1.436 1.353 1.636 1.617
500 1600 3.517 1.353 1.756 1.626 1.134 1.338 1.256 1.523 1.490
500 1800 2.841 1.291 1.684 1.500 1.051 1.241 1.164 1.410 1.363
G250-1.95-1255015
500 2000 2.347 1.225 1.607 1.373 0.972 1.147 1.077 1.300 1.242
500 2500 1.572 1.052 1.387 1.082 0.798 0.934 0.890 1.049 0.981
1.8386
500 3000 1.142 0.860 1.137 0.857 0.660 0.764 0.744 0.847 0.784
500 3500 0.915 0.699 0.887 0.695 0.554 0.635 0.631 0.693 0.641
500 4000 0.707 0.573 0.712 0.577 0.472 0.538 0.543 0.577 0.536
500 4500 0.587 0.478 0.590 0.491 0.410 0.464 0.475 0.489 0.459
500 5000 0.500 0.410 0.502 0.425 0.361 0.406 0.420 0.423 0.399
600 1400 3.113 0.977 1.207 1.172 0.814 0.985 0.904 1.158 1.089
600 1600 2.450 0.947 1.166 1.096 0.759 0.914 0.842 1.080 1.006
600 1800 1.980 0.902 1.122 1.015 0.706 0.845 0.782 1.001 0.924
600 2000 1.635 0.864 1.074 0.933 0.655 0.779 0.726 0.923 0.845
600 2500 1.095 0.733 0.937 0.741 0.542 0.632 0.603 0.747 0.672
1.2107
600 3000 0.796 0.600 0.782 0.591 0.450 0.517 0.507 0.605 0.539
600 3500 0.613 0.550 0.618 0.480 0.379 0.430 0.431 0.495 0.442
600 4000 0.492 0.398 0.496 0.400 0.324 0.364 0.372 0.413 0.371
600 4500 0.409 0.334 0.411 0.340 0.282 0.314 0.326 0.350 0.317
600 5000 0.348 0.286 0.350 0.295 0.248 0.275 0.289 0.303 0.276
FEA (kNm) Mb (kNm)

New Design Method

Rankine Approach -
Temp. (oC)
based on Option 3
Span (mm)
My (kNm)
Section
AS/NZS 4600
New Proposal
EC 3 Part 1.3
EC 3 Part 1.2
EC 3 Part 1.3
(ECS, 2006)
(ECS, 2005)
(SA, 2005)
Mo Mu
700 1400 1.758 0.563 0.676 0.658 0.457 0.526 0.507 0.656 0.611
700 1600 1.384 0.542 0.654 0.615 0.426 0.489 0.472 0.611 0.565
700 1800 1.118 0.517 0.629 0.571 0.397 0.453 0.439 0.567 0.519
G250-1.95-1255015
700 2000 0.923 0.486 0.602 0.525 0.368 0.419 0.408 0.523 0.475
700 2500 0.618 0.410 0.526 0.418 0.305 0.343 0.339 0.423 0.378
0.6777
700 3000 0.449 0.340 0.440 0.333 0.253 0.282 0.285 0.343 0.304
700 3500 0.346 0.276 0.349 0.271 0.213 0.236 0.243 0.281 0.249
700 4000 0.278 0.225 0.280 0.225 0.183 0.201 0.210 0.234 0.209
700 4500 0.231 0.188 0.232 0.192 0.159 0.174 0.184 0.199 0.179
700 5000 0.197 0.160 0.198 0.166 0.140 0.153 0.163 0.172 0.156
20 1000 14.639 5.324 6.042 5.796 4.032 5.151 4.422 5.993 6.260
20 1200 10.430 4.986 5.721 5.207 3.627 4.587 3.954 5.356 5.737
20 1400 7.815 4.578 5.350 4.584 3.239 4.034 3.522 4.729 5.105
20 1600 6.090 4.099 4.932 3.985 2.883 3.523 3.136 4.143 4.439
20 1800 4.894 3.602 4.469 3.450 2.563 3.074 2.797 3.619 3.821
20 2000 4.033 3.123 3.966 2.995 2.283 2.691 2.503 3.162 3.291
6.1196
20 2500 2.698 2.218 2.724 2.166 1.741 1.983 1.932 2.299 2.339
20 3000 1.966 1.638 1.981 1.645 1.372 1.529 1.538 1.739 1.756
20 3500 1.519 1.276 1.529 1.305 1.116 1.226 1.259 1.368 1.381
20 4000 1.226 1.042 1.233 1.072 0.934 1.015 1.057 1.115 1.128
G250-1.95-1504015
20 4500 1.021 0.889 1.027 0.905 0.799 0.863 0.907 0.934 0.948
20 5000 0.872 0.788 0.877 0.781 0.697 0.748 0.792 0.801 0.815
200 1000 12.439 4.827 5.449 5.125 3.603 4.622 3.946 5.351 5.553
200 1200 8.862 4.489 5.137 4.574 3.222 4.095 3.510 4.749 5.053
200 1400 6.641 4.081 4.776 3.999 2.862 3.580 3.112 4.163 4.458
200 1600 5.175 3.620 4.368 3.456 2.534 3.109 2.758 3.623 3.846
200 1800 4.159 3.141 3.917 2.979 2.243 2.699 2.451 3.146 3.293
200 2000 3.427 2.715 3.426 2.577 1.990 2.354 2.186 2.736 2.826
5.5689
200 2500 2.293 2.129 2.314 1.856 1.508 1.723 1.677 1.974 1.999
200 3000 1.670 1.397 1.683 1.407 1.183 1.323 1.329 1.486 1.498
200 3500 1.291 1.088 1.299 1.115 0.960 1.058 1.085 1.167 1.178
200 4000 1.041 0.889 1.048 0.915 0.802 0.875 0.909 0.949 0.961
200 4500 0.868 0.763 0.873 0.772 0.685 0.743 0.778 0.795 0.807
200 5000 0.741 0.726 0.745 0.666 0.597 0.644 0.679 0.681 0.694
FEA (kNm) Mb (kNm)

New Design Method

Rankine Approach -
Temp. (oC)
based on Option 3
Span (mm)
My (kNm)
Section
AS/NZS 4600
New Proposal
EC 3 Part 1.3
EC 3 Part 1.2
EC 3 Part 1.3
(ECS, 2006)
(ECS, 2005)
(SA, 2005)
Mo Mu
300 1000 10.468 2.768 4.013 3.939 2.708 3.179 2.974 3.156 3.264
300 1200 7.459 2.591 3.819 3.566 2.452 2.924 2.676 2.885 2.855
300 1400 5.589 2.396 3.595 3.166 2.205 2.656 2.398 2.608 2.487
300 1600 4.355 2.254 3.341 2.773 1.974 2.387 2.146 2.337 2.168
300 1800 3.500 2.076 3.062 2.415 1.765 2.131 1.923 2.084 1.897
300 2000 2.884 1.810 2.757 2.105 1.580 1.898 1.728 1.854 1.669
4.0242
300 2500 1.930 1.460 1.948 1.532 1.215 1.435 1.345 1.392 1.248
300 3000 1.406 1.121 1.417 1.167 0.962 1.118 1.076 1.072 0.974
300 3500 1.086 0.891 1.094 0.928 0.786 0.901 0.885 0.852 0.789
300 4000 0.876 0.729 0.882 0.763 0.659 0.747 0.745 0.698 0.658
G250-1.95-1504015
300 4500 0.730 0.614 0.734 0.644 0.565 0.635 0.640 0.587 0.562
300 5000 0.624 0.531 0.627 0.556 0.493 0.551 0.560 0.505 0.490
400 1000 8.492 2.147 2.923 2.981 2.019 2.337 2.221 2.538 2.493
400 1200 6.051 1.970 2.823 2.724 1.842 2.149 2.015 2.331 2.197
400 1400 4.534 1.881 2.677 2.445 1.670 1.957 1.819 2.118 1.929
400 1600 3.533 1.766 2.513 2.163 1.508 1.767 1.639 1.908 1.692
400 1800 2.839 1.620 2.331 1.899 1.358 1.587 1.478 1.710 1.488
400 2000 2.339 1.489 2.133 1.665 1.223 1.422 1.335 1.528 1.314
2.9230
400 2500 1.565 1.176 1.580 1.223 0.952 1.090 1.050 1.157 0.990
400 3000 1.140 0.907 1.149 0.936 0.760 0.857 0.846 0.896 0.776
400 3500 0.881 0.720 0.887 0.746 0.623 0.696 0.700 0.714 0.630
400 4000 0.711 0.591 0.715 0.614 0.524 0.580 0.591 0.586 0.527
400 4500 0.592 0.497 0.596 0.519 0.450 0.495 0.510 0.494 0.451
400 5000 0.506 0.429 0.509 0.449 0.394 0.431 0.447 0.424 0.393
FEA (kNm) Mb (kNm)

New Design Method

Rankine Approach -
Temp. (oC)
based on Option 3
Span (mm)
My (kNm)
Section
AS/NZS 4600
New Proposal
EC 3 Part 1.3
EC 3 Part 1.2
EC 3 Part 1.3
(ECS, 2006)
(ECS, 2005)
(SA, 2005)
Mo Mu
500 1000 6.516 1.592 2.064 2.165 1.452 1.691 1.599 1.941 2.026
500 1200 4.642 1.508 2.017 1.992 1.332 1.552 1.459 1.789 1.833
500 1400 3.479 1.423 1.922 1.802 1.215 1.413 1.324 1.631 1.639
500 1600 2.711 1.334 1.815 1.606 1.103 1.277 1.199 1.475 1.453
500 1800 2.178 1.230 1.697 1.420 0.999 1.149 1.086 1.326 1.284
500 2000 1.795 1.121 1.569 1.251 0.904 1.032 0.985 1.189 1.135
2.0640
500 2500 1.201 0.880 1.208 0.926 0.710 0.796 0.781 0.906 0.850
500 3000 0.875 0.696 0.882 0.711 0.570 0.630 0.634 0.704 0.660
500 3500 0.676 0.550 0.681 0.568 0.469 0.513 0.526 0.563 0.531
500 4000 0.545 0.451 0.549 0.468 0.396 0.429 0.446 0.463 0.441
500 4500 0.455 0.380 0.457 0.396 0.341 0.367 0.385 0.390 0.375
500 5000 0.388 0.328 0.390 0.343 0.299 0.320 0.338 0.335 0.325
600 1000 4.540 1.102 1.359 1.453 0.967 1.163 1.066 1.359 1.362
600 1200 3.234 1.054 1.338 1.342 0.890 1.062 0.976 1.267 1.238
600 1400 2.424 0.999 1.279 1.220 0.815 0.962 0.889 1.157 1.111
600 1600 1.888 0.937 1.213 1.093 0.743 0.867 0.808 1.048 0.990
G250-1.95-1504015
600 1800 1.518 0.864 1.139 0.971 0.675 0.778 0.734 0.944 0.878
600 2000 1.251 0.788 1.059 0.859 0.613 0.698 0.668 0.847 0.778
1.3591
600 2500 0.837 0.618 0.834 0.639 0.485 0.538 0.533 0.647 0.585
600 3000 0.610 0.483 0.614 0.492 0.391 0.426 0.434 0.504 0.456
600 3500 0.471 0.383 0.474 0.393 0.323 0.348 0.361 0.403 0.367
600 4000 0.380 0.314 0.382 0.325 0.273 0.291 0.307 0.331 0.305
600 4500 0.317 0.264 0.318 0.275 0.235 0.249 0.266 0.279 0.260
600 5000 0.271 0.229 0.272 0.238 0.206 0.218 0.234 0.240 0.226
700 1000 2.563 0.580 0.761 0.815 0.542 0.621 0.598 0.761 0.765
700 1200 1.826 0.605 0.750 0.754 0.500 0.568 0.548 0.717 0.695
700 1400 1.368 0.520 0.717 0.686 0.458 0.516 0.499 0.655 0.625
700 1600 1.066 0.518 0.680 0.615 0.417 0.466 0.454 0.593 0.557
700 1800 0.857 0.479 0.639 0.546 0.380 0.420 0.413 0.534 0.494
700 2000 0.706 0.445 0.595 0.484 0.345 0.379 0.376 0.480 0.438
0.7608
700 2500 0.473 0.353 0.470 0.360 0.273 0.294 0.300 0.367 0.330
700 3000 0.344 0.273 0.347 0.278 0.220 0.235 0.244 0.286 0.257
700 3500 0.266 0.216 0.268 0.222 0.182 0.192 0.203 0.228 0.207
700 4000 0.215 0.175 0.216 0.183 0.154 0.162 0.173 0.188 0.172
700 4500 0.179 0.149 0.180 0.155 0.133 0.139 0.150 0.158 0.147
700 5000 0.153 0.129 0.154 0.134 0.116 0.121 0.132 0.136 0.127
FEA (kNm) Mb (kNm)

New Design Method

Rankine Approach -
Temp. (oC)
based on Option 3
Span (mm)
My (kNm)
Section
AS/NZS 4600
New Proposal
EC 3 Part 1.3
EC 3 Part 1.2
EC 3 Part 1.3
(ECS, 2006)
(ECS, 2005)
(SA, 2005)
Mo Mu
20 1800 12.198 6.887 7.127 5.891 4.480 5.749 4.957 6.605 6.465
20 2000 10.081 6.533 6.805 5.458 4.145 5.278 4.587 6.062 6.048
20 2500 6.696 5.293 5.876 4.390 3.389 4.206 3.775 4.822 4.890
20 3000 4.796 4.030 4.788 3.493 2.776 3.356 3.127 3.821 3.851
7.7051
20 3500 3.631 3.056 3.667 2.813 2.299 2.722 2.619 3.061 3.064
20 4000 2.867 2.392 2.890 2.311 1.933 2.252 2.224 2.497 2.492
20 4500 2.339 1.940 2.354 1.937 1.651 1.901 1.915 2.077 2.073
20 5000 1.958 1.617 1.969 1.655 1.431 1.633 1.669 1.760 1.759
200 1800 10.365 6.223 6.394 5.177 3.976 5.130 4.400 5.849 5.696
200 2000 8.566 5.868 6.080 4.776 3.665 4.692 4.058 5.341 5.302
200 2500 5.749 4.628 5.175 3.804 2.971 3.704 3.315 4.202 4.234
200 3000 4.075 3.455 4.114 3.008 2.417 2.935 2.730 3.302 3.309
7.0116
200 3500 3.085 2.591 3.116 2.414 1.993 2.369 2.277 2.631 2.623
200 4000 2.436 2.035 2.455 1.979 1.670 1.954 1.927 2.138 2.129
G250-1.95-1506017
200 4500 1.987 1.648 2.000 1.657 1.423 1.645 1.654 1.774 1.769
200 5000 1.664 1.373 1.673 1.414 1.232 1.411 1.439 1.501 1.500
300 1800 8.723 3.388 4.762 4.035 3.032 3.669 3.356 3.573 3.251
300 2000 7.209 3.300 4.567 3.757 2.818 3.443 3.118 3.332 2.979
300 2500 4.788 2.967 4.006 3.058 2.329 2.877 2.589 2.754 2.391
300 3000 3.429 2.525 3.347 2.453 1.923 2.370 2.159 2.252 1.937
5.0668
300 3500 2.596 2.062 2.622 1.985 1.602 1.958 1.819 1.845 1.595
300 4000 2.050 1.676 2.066 1.636 1.353 1.638 1.551 1.528 1.337
300 4500 1.672 1.380 1.684 1.374 1.159 1.390 1.340 1.284 1.140
300 5000 1.400 1.152 1.408 1.175 1.007 1.199 1.172 1.095 0.988
400 1800 7.076 2.680 3.525 2.878 2.281 2.697 2.527 2.889 2.503
400 2000 5.848 2.635 3.398 2.715 2.133 2.535 2.360 2.704 2.304
400 2500 3.884 2.325 3.033 2.284 1.786 2.135 1.982 2.255 1.868
400 3000 2.782 2.044 2.604 1.878 1.492 1.777 1.669 1.858 1.525
3.6803
400 3500 2.106 1.674 2.122 1.544 1.253 1.482 1.416 1.532 1.262
400 4000 1.663 1.360 1.676 1.285 1.064 1.249 1.215 1.274 1.062
400 4500 1.357 1.118 1.366 1.086 0.916 1.067 1.054 1.074 0.908
400 5000 1.136 0.932 1.142 0.932 0.798 0.924 0.925 0.918 0.788
FEA (kNm) Mb (kNm)

New Design Method

Rankine Approach -
Temp. (oC)
based on Option 3
Span (mm)
My (kNm)
Section
AS/NZS 4600
New Proposal
EC 3 Part 1.3
EC 3 Part 1.2
EC 3 Part 1.3
(ECS, 2006)
(ECS, 2005)
(SA, 2005)
Mo Mu
500 1800 5.429 2.002 2.520 2.824 1.651 1.948 1.830 2.219 2.081
500 2000 4.487 1.980 2.438 2.591 1.550 1.829 1.715 2.082 1.942
500 2500 2.980 1.780 2.201 2.039 1.311 1.543 1.453 1.747 1.607
500 3000 2.135 1.544 1.922 1.601 1.104 1.289 1.232 1.447 1.318
2.5988
500 3500 1.616 1.273 1.609 1.279 0.933 1.080 1.052 1.198 1.088
500 4000 1.276 1.041 1.286 1.046 0.797 0.914 0.906 1.000 0.911
500 4500 1.041 0.855 1.048 0.874 0.688 0.784 0.789 0.844 0.774
500 5000 0.871 0.715 0.877 0.745 0.601 0.681 0.695 0.723 0.668
600 1800 3.783 1.433 1.673 1.518 1.105 1.332 1.225 1.572 1.404
600 2000 3.126 1.377 1.621 1.434 1.040 1.248 1.151 1.476 1.314
G250-1.95-1506017
600 2500 2.076 1.200 1.474 1.210 0.885 1.047 0.980 1.242 1.095
600 3000 1.487 1.083 1.301 0.998 0.750 0.873 0.835 1.031 0.903
1.7112
600 3500 1.126 0.890 1.106 0.822 0.637 0.730 0.716 0.855 0.748
600 4000 0.889 0.724 0.896 0.684 0.545 0.618 0.619 0.715 0.628
600 4500 0.725 0.596 0.730 0.579 0.472 0.530 0.540 0.604 0.535
600 5000 0.607 0.498 0.611 0.497 0.414 0.461 0.477 0.518 0.462
700 1800 2.136 0.824 0.937 0.853 0.620 0.712 0.687 0.890 0.789
700 2000 1.765 0.782 0.909 0.806 0.584 0.668 0.646 0.836 0.739
700 2500 1.172 0.711 0.827 0.681 0.497 0.564 0.551 0.703 0.616
700 3000 0.840 0.618 0.731 0.562 0.422 0.473 0.470 0.584 0.509
0.9579
700 3500 0.636 0.505 0.623 0.463 0.358 0.399 0.403 0.485 0.422
700 4000 0.502 0.410 0.506 0.386 0.307 0.339 0.348 0.405 0.354
700 4500 0.410 0.337 0.412 0.326 0.266 0.292 0.304 0.343 0.302
700 5000 0.343 0.281 0.345 0.280 0.233 0.255 0.269 0.293 0.261
20 1000 3.596 2.457 3.112 2.364 2.022 2.386 1.989 2.534 2.364
20 1200 2.574 1.994 2.568 1.882 1.652 1.912 1.649 2.019 1.882
G450-1.5-903014
20 1400 1.945 1.564 1.982 1.514 1.358 1.546 1.380 1.621 1.514
20 1600 1.532 1.268 1.555 1.241 1.132 1.271 1.170 1.322 1.241
3.9944
20 1800 1.246 1.027 1.262 1.037 0.958 1.066 1.005 1.097 1.037
20 2000 1.040 0.865 1.052 0.883 0.823 0.909 0.874 0.928 0.883
20 2500 0.719 0.611 0.726 0.631 0.597 0.652 0.647 0.652 0.631
20 3000 0.541 0.680 0.545 0.484 0.462 0.502 0.507 0.493 0.484
FEA (kNm) Mb (kNm)

New Design Method

Rankine Approach -
Temp. (oC)
based on Option 3
Span (mm)
My (kNm)
Section
AS/NZS 4600
New Proposal
EC 3 Part 1.3
EC 3 Part 1.2
EC 3 Part 1.3
(ECS, 2006)
(ECS, 2005)
(SA, 2005)
Mo Mu
200 1000 3.055 2.142 2.866 2.094 1.840 2.186 1.819 2.234 2.094
200 1200 2.187 1.707 2.238 1.645 1.478 1.724 1.488 1.750 1.645
200 1400 1.653 1.356 1.684 1.313 1.200 1.378 1.233 1.389 1.313
3.9560 200 1600 1.302 1.073 1.322 1.071 0.992 1.126 1.038 1.125 1.071
200 1800 1.059 0.874 1.073 0.893 0.835 0.939 0.886 0.929 0.893
200 2000 0.883 0.736 0.894 0.759 0.715 0.798 0.767 0.783 0.759
200 2500 0.611 0.560 0.617 0.541 0.516 0.569 0.563 0.548 0.541
300 1000 2.571 1.911 2.544 1.816 1.631 1.955 1.622 1.951 1.816
300 1200 1.840 1.485 1.886 1.414 1.292 1.518 1.313 1.507 1.414
300 1400 1.391 1.143 1.417 1.122 1.040 1.202 1.079 1.185 1.122
3.7947 300 1600 1.095 0.907 1.112 0.912 0.855 0.976 0.902 0.954 0.912
300 1800 0.891 0.736 0.903 0.759 0.717 0.811 0.767 0.785 0.759
G450-1.5-903014
300 2000 0.744 0.620 0.752 0.644 0.612 0.687 0.661 0.660 0.644
300 2500 0.514 0.523 0.519 0.458 0.440 0.488 0.482 0.461 0.458
400 1000 2.086 1.457 1.980 1.439 1.270 1.510 1.257 1.521 1.439
400 1200 1.493 1.161 1.530 1.128 1.017 1.189 1.026 1.190 1.128
400 1400 1.128 0.921 1.150 0.900 0.825 0.949 0.849 0.943 0.900
2.7725 400 1600 0.889 0.731 0.902 0.733 0.681 0.775 0.713 0.763 0.733
400 1800 0.723 0.597 0.732 0.611 0.573 0.646 0.608 0.630 0.611
400 2000 0.603 0.500 0.610 0.519 0.490 0.549 0.526 0.531 0.519
400 2500 0.417 0.409 0.421 0.370 0.353 0.391 0.386 0.371 0.370
500 1000 1.600 0.967 1.279 1.010 0.843 0.973 0.828 1.023 1.010
500 1200 1.146 0.814 1.092 0.815 0.700 0.800 0.694 0.833 0.815
500 1400 0.866 0.671 0.881 0.660 0.582 0.657 0.586 0.679 0.660
1.5605 500 1600 0.682 0.551 0.692 0.544 0.489 0.546 0.501 0.560 0.544
500 1800 0.555 0.451 0.562 0.456 0.416 0.461 0.432 0.468 0.456
500 2000 0.463 0.382 0.468 0.389 0.359 0.395 0.378 0.397 0.389
500 2500 0.320 0.267 0.323 0.279 0.262 0.286 0.282 0.281 0.279
FEA (kNm) Mb (kNm)

New Design Method

Rankine Approach -
Temp. (oC)
based on Option 3
Span (mm)
My (kNm)
Section
AS/NZS 4600
New Proposal
EC 3 Part 1.3
EC 3 Part 1.2
EC 3 Part 1.3
(ECS, 2006)
(ECS, 2005)
(SA, 2005)
Mo Mu
600 1000 1.115 0.331 0.436 0.450 0.321 0.347 0.324 0.414 0.416
600 1200 0.798 0.316 0.415 0.404 0.294 0.311 0.292 0.374 0.368
600 1400 0.603 0.304 0.391 0.356 0.267 0.277 0.263 0.335 0.322
0.4394 600 1600 0.475 0.279 0.364 0.311 0.241 0.246 0.237 0.298 0.282
600 1800 0.386 0.250 0.336 0.272 0.217 0.219 0.214 0.265 0.247
G450-1.5-903014
600 2000 0.322 0.228 0.305 0.238 0.196 0.196 0.194 0.236 0.217
600 2500 0.223 0.175 0.225 0.178 0.153 0.152 0.155 0.180 0.164
700 1000 0.630 0.217 0.273 0.274 0.199 0.207 0.199 0.253 0.252
700 1200 0.451 0.202 0.258 0.242 0.180 0.185 0.178 0.227 0.220
700 1400 0.341 0.185 0.241 0.211 0.162 0.164 0.159 0.201 0.191
0.2796 700 1600 0.268 0.167 0.222 0.182 0.145 0.145 0.142 0.177 0.165
700 1800 0.218 0.148 0.201 0.158 0.129 0.129 0.128 0.156 0.144
700 2000 0.182 0.134 0.179 0.138 0.115 0.115 0.115 0.138 0.126
700 2500 0.126 0.100 0.127 0.102 0.089 0.089 0.091 0.103 0.094
20 1000 7.428 4.295 4.585 3.896 3.187 3.909 3.254 4.138 3.896
20 1200 5.485 3.882 4.206 3.365 2.791 3.441 2.851 3.555 3.365
20 1400 4.211 3.370 3.776 2.866 2.425 2.984 2.500 3.032 2.866
20 1600 3.347 2.825 3.303 2.439 2.106 2.576 2.202 2.587 2.439
G450-1.9-804015
20 1800 2.738 2.346 2.794 2.089 1.837 2.231 1.952 2.219 2.089
20 2000 2.293 1.962 2.334 1.809 1.614 1.947 1.742 1.920 1.809
4.9644
20 2500 1.596 1.367 1.615 1.326 1.213 1.445 1.352 1.395 1.326
20 3000 1.206 1.037 1.217 1.032 0.959 1.133 1.092 1.074 1.032
20 3500 0.963 0.834 0.971 0.840 0.789 0.927 0.911 0.866 0.840
20 4000 0.799 0.698 0.805 0.707 0.669 0.783 0.780 0.723 0.707
20 4500 0.683 0.603 0.688 0.611 0.580 0.678 0.682 0.620 0.611
20 5000 0.596 0.533 0.600 0.537 0.513 0.598 0.606 0.542 0.537
FEA (kNm) Mb (kNm)

New Design Method

Rankine Approach -
Temp. (oC)
based on Option 3
Span (mm)
My (kNm)
Section
AS/NZS 4600
New Proposal
EC 3 Part 1.3
EC 3 Part 1.2
EC 3 Part 1.3
(ECS, 2006)
(ECS, 2005)
(SA, 2005)
Mo Mu
200 1000 6.311 3.944 4.389 3.566 2.989 3.710 3.048 3.793 3.566
200 1200 4.660 3.527 3.951 3.025 2.572 3.214 2.637 3.201 3.025
200 1400 3.578 2.949 3.455 2.540 2.201 2.744 2.288 2.688 2.540
200 1600 2.844 2.354 2.910 2.140 1.888 2.338 1.997 2.265 2.140
200 1800 2.326 1.999 2.377 1.821 1.632 2.006 1.757 1.925 1.821
200 2000 1.949 1.664 1.983 1.570 1.424 1.738 1.558 1.654 1.570
4.9167
200 2500 1.356 1.161 1.373 1.143 1.059 1.276 1.195 1.190 1.143
200 3000 1.024 0.880 1.034 0.887 0.832 0.994 0.958 0.912 0.887
200 3500 0.818 0.706 0.825 0.721 0.682 0.810 0.795 0.733 0.721
200 4000 0.679 0.591 0.684 0.606 0.577 0.682 0.678 0.611 0.606
200 4500 0.580 0.512 0.584 0.523 0.500 0.589 0.591 0.523 0.523
200 5000 0.507 0.450 0.510 0.460 0.441 0.519 0.524 0.458 0.460
300 1000 5.312 3.630 4.067 3.177 2.721 3.429 2.777 3.403 3.177
300 1200 3.922 3.155 3.588 2.655 2.305 2.921 2.376 2.826 2.655
300 1400 3.012 2.577 3.046 2.204 1.947 2.457 2.042 2.342 2.204
300 1600 2.393 2.065 2.462 1.843 1.654 2.070 1.769 1.953 1.843
G450-1.9-804015
300 1800 1.958 1.681 2.000 1.562 1.420 1.761 1.546 1.648 1.562
300 2000 1.640 1.400 1.669 1.342 1.233 1.518 1.364 1.408 1.342
4.7162
300 2500 1.141 0.950 1.155 0.973 0.910 1.104 1.036 1.006 0.973
300 3000 0.862 0.739 0.870 0.753 0.712 0.856 0.826 0.768 0.753
300 3500 0.688 0.595 0.694 0.611 0.582 0.696 0.683 0.616 0.611
300 4000 0.572 0.500 0.576 0.513 0.492 0.585 0.581 0.513 0.513
300 4500 0.488 0.430 0.492 0.442 0.425 0.504 0.505 0.439 0.442
300 5000 0.426 0.380 0.429 0.389 0.375 0.444 0.447 0.384 0.389
400 1000 4.309 2.602 3.056 2.463 2.074 2.562 2.115 2.562 2.463
400 1200 3.182 2.329 2.741 2.084 1.779 2.220 1.826 2.160 2.084
400 1400 2.443 1.958 2.384 1.745 1.518 1.895 1.581 1.813 1.745
400 1600 1.942 1.648 1.992 1.468 1.300 1.614 1.378 1.527 1.468
400 1800 1.588 1.359 1.623 1.248 1.122 1.384 1.210 1.297 1.248
400 2000 1.330 1.132 1.354 1.075 0.978 1.199 1.072 1.114 1.075
3.4458
400 2500 0.926 0.756 0.937 0.782 0.726 0.879 0.821 0.801 0.782
400 3000 0.699 0.599 0.706 0.607 0.570 0.684 0.657 0.614 0.607
400 3500 0.558 0.479 0.563 0.493 0.467 0.557 0.545 0.493 0.493
400 4000 0.464 0.403 0.467 0.414 0.395 0.469 0.465 0.411 0.414
400 4500 0.396 0.348 0.399 0.357 0.342 0.405 0.405 0.352 0.357
400 5000 0.346 0.307 0.348 0.314 0.302 0.357 0.359 0.308 0.314
FEA (kNm) Mb (kNm)

New Design Method

Rankine Approach -
Temp. (oC)
based on Option 3
Span (mm)
My (kNm)
Section
AS/NZS 4600
New Proposal
EC 3 Part 1.3
EC 3 Part 1.2
EC 3 Part 1.3
(ECS, 2006)
(ECS, 2005)
(SA, 2005)
Mo Mu
500 1000 3.306 1.565 1.836 1.618 1.297 1.538 1.327 1.630 1.618
500 1200 2.441 1.404 1.705 1.418 1.152 1.378 1.175 1.427 1.418
500 1400 1.874 1.313 1.558 1.224 1.014 1.218 1.040 1.238 1.224
500 1600 1.490 1.156 1.396 1.051 0.890 1.069 0.923 1.071 1.051
500 1800 1.219 1.004 1.221 0.907 0.783 0.938 0.824 0.930 0.907
1.9395
500 2000 1.021 0.863 1.039 0.789 0.693 0.827 0.739 0.811 0.789
500 2500 0.710 0.603 0.719 0.582 0.526 0.623 0.580 0.598 0.582
500 3000 0.537 0.458 0.542 0.454 0.418 0.492 0.472 0.463 0.454
500 3500 0.428 0.368 0.432 0.371 0.345 0.404 0.395 0.375 0.371
500 4000 0.356 0.308 0.358 0.312 0.293 0.342 0.340 0.314 0.312
600 1000 2.303 0.504 0.546 0.611 0.438 0.491 0.460 0.546 0.581
600 1200 1.701 0.487 0.546 0.578 0.416 0.460 0.432 0.535 0.540
600 1400 1.306 0.475 0.538 0.541 0.392 0.429 0.405 0.501 0.500
600 1600 1.038 0.454 0.520 0.503 0.369 0.399 0.378 0.467 0.459
G450-1.9-804015
600 1800 0.849 0.438 0.500 0.464 0.346 0.371 0.353 0.434 0.421
0.5461
600 2000 0.711 0.417 0.479 0.425 0.323 0.344 0.330 0.402 0.385
600 2500 0.495 0.361 0.423 0.342 0.272 0.287 0.280 0.332 0.310
600 3000 0.374 0.301 0.363 0.279 0.231 0.243 0.242 0.277 0.255
600 3500 0.298 0.252 0.301 0.233 0.199 0.209 0.212 0.235 0.215
600 4000 0.248 0.213 0.250 0.200 0.174 0.183 0.188 0.202 0.185
700 1000 1.301 0.316 0.348 0.378 0.274 0.295 0.287 0.348 0.357
700 1200 0.960 0.304 0.348 0.355 0.259 0.276 0.268 0.328 0.330
700 1400 0.738 0.286 0.337 0.330 0.243 0.257 0.249 0.305 0.303
700 1600 0.586 0.271 0.324 0.303 0.227 0.238 0.231 0.282 0.276
700 1800 0.479 0.256 0.310 0.277 0.211 0.221 0.215 0.260 0.251
700 2000 0.402 0.240 0.295 0.252 0.196 0.205 0.200 0.239 0.228
0.3475
700 2500 0.279 0.204 0.254 0.199 0.162 0.170 0.168 0.195 0.181
700 3000 0.211 0.170 0.211 0.161 0.136 0.143 0.144 0.160 0.148
700 3500 0.169 0.142 0.170 0.134 0.116 0.123 0.125 0.135 0.124
700 4000 0.140 0.120 0.141 0.115 0.101 0.107 0.111 0.115 0.106
700 4500 0.120 0.105 0.120 0.100 0.089 0.095 0.099 0.101 0.093
700 5000 0.104 0.092 0.105 0.088 0.080 0.085 0.090 0.089 0.083
FEA (kNm) Mb (kNm)

New Design Method

Rankine Approach -
Temp. (oC)
based on Option 3
Span (mm)
My (kNm)
Section
AS/NZS 4600
New Proposal
EC 3 Part 1.3
EC 3 Part 1.2
EC 3 Part 1.3
(ECS, 2006)
(ECS, 2005)
(SA, 2005)
Mo Mu
20 1200 8.193 5.317 6.530 5.164 4.283 5.085 4.234 5.423 5.164
20 1400 6.176 4.621 5.750 4.318 3.669 4.298 3.652 4.554 4.318
20 1600 4.839 3.926 4.878 3.614 3.141 3.628 3.166 3.825 3.614
20 1800 3.910 3.230 3.974 3.052 2.702 3.084 2.763 3.232 3.052
7.9442 20 2000 3.239 2.687 3.284 2.609 2.345 2.648 2.429 2.757 2.609
20 2500 2.197 1.839 2.219 1.858 1.712 1.901 1.822 1.942 1.858
20 3000 1.622 1.363 1.635 1.410 1.320 1.450 1.429 1.456 1.410
20 3500 1.269 1.067 1.278 1.124 1.062 1.160 1.163 1.147 1.124
20 4000 1.035 0.878 1.042 0.929 0.884 0.961 0.975 0.940 0.929
200 1200 6.961 5.012 6.092 4.609 3.927 4.707 3.892 4.859 4.609
200 1400 5.247 4.164 5.192 3.803 3.312 3.915 3.320 4.017 3.803
200 1600 4.112 3.383 4.196 3.155 2.801 3.265 2.852 3.333 3.155
200 1800 3.322 2.782 3.377 2.649 2.390 2.751 2.470 2.791 2.649
7.8678 200 2000 2.752 2.296 2.790 2.256 2.060 2.348 2.158 2.365 2.256
200 2500 1.867 1.563 1.885 1.598 1.490 1.670 1.601 1.651 1.598
G450-1.9-1154015
200 3000 1.378 1.153 1.389 1.210 1.143 1.267 1.246 1.233 1.210
200 3500 1.078 0.907 1.086 0.963 0.917 1.010 1.009 0.970 0.963
200 4000 0.880 0.866 0.886 0.795 0.761 0.835 0.843 0.794 0.795
300 1200 5.858 4.412 5.490 4.022 3.504 4.240 3.487 4.271 4.022
300 1400 4.416 3.554 4.505 3.285 2.917 3.476 2.947 3.483 3.285
300 1600 3.460 2.868 3.532 2.708 2.445 2.869 2.511 2.862 2.708
300 1800 2.796 2.344 2.842 2.265 2.072 2.401 2.162 2.380 2.265
7.5470
300 2000 2.316 1.944 2.348 1.923 1.778 2.040 1.879 2.008 1.923
300 2500 1.570 1.315 1.587 1.357 1.277 1.440 1.381 1.392 1.357
300 3000 1.160 0.972 1.169 1.025 0.975 1.088 1.069 1.037 1.025
300 3500 0.907 0.763 0.914 0.815 0.781 0.865 0.863 0.814 0.815
400 1200 4.752 3.106 4.220 3.171 2.714 3.251 2.692 3.279 3.171
400 1400 3.583 2.782 3.573 2.611 2.283 2.703 2.292 2.708 2.611
400 1600 2.807 2.258 2.865 2.163 1.927 2.253 1.965 2.246 2.163
400 1800 2.268 1.887 2.306 1.815 1.642 1.897 1.700 1.880 1.815
5.5140 400 2000 1.879 1.572 1.905 1.544 1.414 1.619 1.484 1.593 1.544
400 2500 1.274 1.067 1.287 1.093 1.021 1.150 1.099 1.111 1.093
400 3000 0.941 0.786 0.948 0.827 0.782 0.872 0.854 0.830 0.827
400 3500 0.736 0.617 0.741 0.658 0.627 0.694 0.691 0.653 0.658
400 4000 0.601 0.587 0.605 0.543 0.521 0.574 0.577 0.534 0.543
FEA (kNm) Mb (kNm)

New Design Method

Rankine Approach -
Temp. (oC)
based on Option 3
Span (mm)
My (kNm)
Section
AS/NZS 4600
New Proposal
EC 3 Part 1.3
EC 3 Part 1.2
EC 3 Part 1.3
(ECS, 2006)
(ECS, 2005)
(SA, 2005)
Mo Mu
500 1200 3.646 2.077 2.660 2.191 1.775 2.059 1.754 2.188 2.191
500 1400 2.749 1.858 2.393 1.855 1.541 1.776 1.528 1.870 1.855
500 1600 2.154 1.620 2.094 1.566 1.334 1.523 1.335 1.594 1.566
500 1800 1.740 1.382 1.766 1.330 1.157 1.310 1.173 1.362 1.330
3.1036
500 2000 1.442 1.163 1.462 1.142 1.010 1.134 1.037 1.171 1.142
500 2500 0.978 0.812 0.988 0.817 0.745 0.824 0.786 0.835 0.817
500 3000 0.722 0.602 0.728 0.622 0.577 0.633 0.620 0.630 0.622
500 3500 0.565 0.472 0.569 0.497 0.466 0.508 0.507 0.498 0.497
600 1200 2.541 0.734 0.874 0.930 0.654 0.707 0.667 0.840 0.865
600 1400 1.915 0.702 0.850 0.861 0.613 0.652 0.617 0.781 0.790
600 1600 1.501 0.670 0.816 0.787 0.571 0.598 0.569 0.721 0.716
G450-1.9-1154015
0.8739 600 1800 1.212 0.628 0.779 0.713 0.530 0.547 0.525 0.662 0.646
600 2000 1.004 0.595 0.739 0.643 0.490 0.500 0.484 0.606 0.581
600 3000 0.503 0.392 0.506 0.391 0.330 0.328 0.334 0.390 0.451
600 3500 0.393 0.319 0.396 0.319 0.277 0.275 0.285 0.321 0.360
700 1200 1.435 0.444 0.553 0.569 0.406 0.424 0.412 0.513 0.296
700 1400 1.081 0.422 0.531 0.522 0.378 0.390 0.378 0.473 0.526
700 1600 0.847 0.394 0.507 0.472 0.350 0.357 0.347 0.433 0.476
700 1800 0.685 0.365 0.480 0.423 0.322 0.326 0.318 0.395 0.428
700 2000 0.567 0.337 0.451 0.378 0.295 0.297 0.292 0.358 0.383
0.5561 700 2500 0.385 0.275 0.372 0.287 0.237 0.237 0.237 0.282 0.342
700 3000 0.284 0.221 0.286 0.225 0.193 0.193 0.197 0.224 0.208
700 3500 0.222 0.180 0.224 0.183 0.161 0.161 0.167 0.183 0.170
700 4000 0.181 0.150 0.182 0.153 0.137 0.137 0.144 0.153 0.143
700 4500 0.153 0.128 0.154 0.131 0.119 0.119 0.126 0.131 0.123
700 5000 0.132 0.111 0.132 0.114 0.105 0.105 0.113 0.114 0.108
REFERENCES
Abdel-Rahman N and Sivakumaran K. Material properties models for analysis of

cold-formed steel members. Journal of Structural Engineering 2006; 123(9):
1135-1143.
Acharya VV and Schuster RM. Bending tests of hat section with multiple
longitudinal stiffeners. Proceedings of the Fourteenth International Specialty
Conference on Cold-formed Steel Structures 1998; St. Louis, Missouri: 77-
92.
American Iron and Steel Institute (AISI). Specification for the design of cold-formed
steel structural members with commentary. 1996 (Suppl. 1).Washington;
1999.
Alfawakhiri F and Sultan MA. Load bearing capacity of cold-formed steel joists
subjected to severe heating. INTERFLAME, 9th International Conference
Proceedings 2001; 1: 431-442.
Allen HG and Bulson PS. Background to buckling. McGraw-Hill Book Co. Ltd.,
London; 1980.
Bailey CG, Burgess IW and Plank RJ. The lateral torsional buckling of unrestrained
steel beams in fire. Journal of Constructional Steel Research 1996; 36(2):
101-119.
Bandula Heva DYM. Behaviour and design of cold-formed steel compression

members at elevated temperatures. PhD Thesis, Queensland University of
Technology, Brisbane, Australia; 2009.
Bernard E. Flexural behaviour of cold-formed profiled steel decking. PhD

dissertation, University of Sydney, Sydney, Australia; 1993.
Bleich F. Buckling Strength of Metal Structures. McGraw-Hill, New York; 1952.

British Standard Institution (BSI). BS5950. Structural use of steel work in building -
Part 5: Code of practice for design of cold-formed thin gauge sections,
London, UK; 1998.
Structural Behaviour and Design of Cold-formed Steel Beams at Elevated Temperatures R-1
British Standard Institution (BSI). BS5950. Structural use of steel work in building -
Part 8: Code of practice for fire resistant design, London, UK; 1990.
Brune B and Ungermann D. Ultimate loads of thin-walled C-sections in coupled

instabilities. Eurosteel Conference, Graz University of Technology, Institute
for Steel Structures and Shell Structures, Lessingstrasse 25, 8010 Graz,Austria;
2008.
Buchanan A. Structural design for fire safety. John Wiley & Sons Ltd; 2001.
Bulson P. The stability of flat plates. Chatto and Windus, London; 1970.
Chajes A, Britvec SJ and Winter G. Effects of cold-straining on structural steel sheet.

Journal of the Structural Division, ASCE Proceedings 1963; 89.
Celebi N, Winter G and Pekoz T. Diaphragm braced channel and Z-section beams.
Rep. No. 344, Dep. of Struct. Engrg., Cornell University, Ithaca, N.Y; 1972.
Chen J and Young B. Mechanical properties of cold-formed steel at elevated

temperatures. Proc., 17th International Specialty Conference on Cold-Formed
Steel Structures, University of Missouri-Rolla, Rolla, Mo.; 2004: 437-466.
Chen J and Young B. Corner properties of cold-formed steel sections at elevated

temperatures. Thin-Walled Structures 2006; 44(2): 216-223.
Chen J and Young B. Experimental investigation of cold-formed steel material at
elevated temperatures. Thin-walled Structures 2007; 45: 96-110.
Chu, X.-t., Kettle, R. and Li, L.-y. Lateral torsion buckling analysis of partial-
laterally restrained thin-walled channel-section beams. Journal of
Constructional Steel Research 2004; 60(8): 1159-1175.
Cohen JM. Local buckling behaviour of plate elements. Department of Structural

Engineering Report, Cornell University, Ithaca, New York; 1987.
Canadian Standards Association (CSA). Cold-formed steel structural members.

S136-94, Ontario, Canada; 1994.
Davies J. Recent research advances in cold-formed steel structures. Journal of

Constructional Steel Research 2000; 55(1-3): 267-288.
Desmond TP. Local buckling behaviour of plate elements. Department of Structural

Engineering Report, Cornell University, Ithaca, New York; 1977.
Dharma RB and Tane, K-H. Proposed design methods for lateral Torsional buckling
of unrestrained steel beams in fire. Journal of Constructional Steel Research
2007; 63(8): 1066-1076.
Dharma RB and Tan K-H (in press). Rotational capacity of steel I-beams under fire
conditions Part I: experimental study. Engineering Structures, In Press,
Corrected Proof.
Ellifritt D, Sputo T, Haynes J and Iron A. Flexural strength and deflections of

discretely braced cold-formed steel channel and zee sections. Rep. Prepared
for American Iron and Steel Institute, University of Florida; 1991.
Ellifrit D, Yu WW and LaBoube RA. Flexural capacity of discretely braced C's and
Z's. Proceedings of the Eleventh International Specialty Conference on Cold-
Formed Steel Structures, St.Louis, MO: Department of Civil Engineering,
University of Missouri; 1992.
Ellifritt D, Glover B and Hren J. Distortional buckling of channels and zees not
attached to sheathing. Report for the American Iron and Steel Institute,
Washington DC; 1997.
El-Zanaty MH and Murray DW. Non-linear finite element analysis of steel frames.
ASCE Journal of Structural Division 1983; 109(ST2): 353-368.
European Committee for Standardization (ECS): Eurocode 3: Design of steel

structures – Part 1.1, General Rules and Rules for Buildings, Brussels, 2005.
European Committee for Standardization (ECS): Eurocode 3: Design of Steel
Structures - Part 1.2: General Rules - Structural Fire Design, Brussels, 2005.
European Committee for Standardization (ECS): Eurocode 3: Design of steel

structures – Part 1.3, General Rules-Supplementary Rules for Cold-Formed
Thin Gauge Members and Sheathing, Brussels, 2006.
Galambos TV. Inelastic buckling of beams. Journal of the Structural Division,

ASCE, 89(ST5), October 1963.
Galambos T. Guide to stability design criteria for metal structures. John Wiley and
Sons; 1998.
Gerlich JT, Collier PCR and Buchanan AH. Design of light steel framed walls for
fire resistance. Fire and Materials 1996; 20: 79-96.
Gotluru B, Schafer B and Pekoz T. Torsion in thin-walled cold-formed steel beams.
Thin-Walled Structures 2000; 37(2): 127-145.
Hancock G, Kwon Y and Bernard E. Strength design curves for thin-walled sections
undergoing distortional buckling. Journal of Constructional Steel Research
1994; 31(2-3): 169-186.
Hancock GJ. Design of cold-formed steel structures (To Australian/New Zealand

Standard AS/NZS 4600: 1996). 3rd Edition, Australian Institute of Steel
Construction; 1998.
Hancock GJ and Rogers CA. Design of cold-formed steel structures of high strength
steel. Journal of Constructional Steel Research 1998; 46(1-3): 167-168.
Hancock GJ. Cold-formed steel structures. Journal of Constructional Steel Research

2003; 59(4): 473-487.
Hill H. Lateral buckling of channels and z-beams. Trans. ASCE 1954; 119: 829-841.
Hoglund T. Design of trapezoidal sheeting provided with stiffeners in the flanges

and webs. Swedish Council for Building Research, Stockholm, Sweden,
D28:1980.
http://www.indiamart.com/valson-fabricators/
http://www.vatsalametsec.com/index.htm
http://www.whatreallyhappened.com/spain_fire_2005.html
Kaitila O. Finite element modelling of cold-formed steel members at high

temperatures. Helsinki University of Technology Laboratory of Steel
Structures Publications 24, Espoo 2002.
Karren KW. Corner properties of cold-formed steel shapes. Journal of the Structural
Division, ASCE 1967; 93(2): 401-432.
Karren KW and Winter G. Effects of cold-forming on light-gauge steel members.

Journal of Structural Division, ASCE Proceedings, 93, (ST1); 1967.
Kavanagh K and Ellifritt D. Design strengths of cold-formed channels in bending

and torsion. Journal of Structural Engineering 1994; 120(5): 1599-1607.
Key P and Hancock G. A theoretical investigation of the column behavior of cold-
formed square hollow sections. Thin-Walled Structures 1993, 16(1): 31-64.
Kian T and Pekoz TB. Evaluation of industry-type bracing details for wall stud
assemblies. Final Report, Cornell University, submitted to American Iron and
Steel Institute, January; 1994.
Kirby BR and Preston RR. High temperature properties of hot-rolled structural steels
for use in fire engineering design studies. Fire Safety Journal 1988; 13(1):
27-37.
Klippstein K. Strength of cold-formed studs exposed to fire. American Iron and Steel
Institute, Washington, USA; 1980.
Kitipornchai S and Trahair NS. Inelastic buckling of simply supported steel I-beams.
Journal of Structural Division ASCE 1975; 101(ST5): 991-1004.
Kolarkar P. Structural and fire behaviour of a new steel stud wall system using
composite panels, PhD Thesis, Queensland University of Technology,
Brisbane, Australia; 2010.
Kurniawan CW and Mahendran M. Lateral buckling strength of simply supported
liteSteel beams subject to moment gradient effects. Journal of Structural
Engineering 2009; 135(9): 1058-1067.
Kwon YB. Post buckling behaviour of thin-walled channel sections, PhD Thesis,
University of Sydney, Sydney, Australia; 1992.
Kwon YB and Hancock GJ. Tests of cold-formed channels with local and
distortional buckling. Journal of Structural Engineering 1992; 118(7): 1786-
1803.
LaBoube RA and Yu W. Structural behaviour of beam webs subjected to bending

stress. Civil Engineering Study Structural Series, 78-1, Department of Civil
Engineering, University of Missouri-Rolla, Rolla, Missouri; 1978.
Lau, S. Distortional buckling of thin-walled columns. PhD dissertation, University of

Sydney, Sydney, Australia; 1988.
Lie T. Structural fire protection. American Society of Civil Engineers, New York;
1992.
Lee J, Mahendran M and Makelainen P. Prediction of mechanical properties of light
gauge steels at elevated temperatures, Journal of Constructional Steel Research
2003; 1517-1532.
Lindner, J. and Aschinger, R. Load-carrying capacity of cold-formed beams

subjected to overall lateral-torsional buckling and local plate buckling, Journal
of Constructional Steel Research 1994; 31(2-3): 267-287.
Liu TCH, Fahad MK and Davies JM. Experimental investigation of behaviour of

axially restrained steel beams in fire. Journal of Constructional Steel
Research 2002; 58(9): 1211-1230.
Lopes N, Real P, da Silva L, Franssen J and Mirambell E. Proposal to the Eurocode

3 for the lateral-torsional buckling of stainless steel I-beams in case of fire.
Fourth International Workshop-Structures in Fire, Aveiro, Portugal 2006:
127-137.
Macdonald M, Taylor GT and Rhodes J. The effect of cold forming on the yield
strength of thin gauge steel-hardness test approach. Thin-Walled Structures
1197; 29(1): 243-256.
Mahaarachchi D and Mahendran M. Finite element analysis of litesteel beam

sections. Research Report, Queensland University of Technology; 2005a.
Mahaarachchi D and Mahendran M. Lateral buckling tests of litesteel beam sections.

Research Report, Queensland University of Technology; 2005b.
Mecozzi, E. and Zhao, B. Development of stress-strain relationships of cold-formed
lightweight steel at elevated temperatures, Proc. Eurosteel Conference (2005)
5.1-41-5.1-49.
Mirambell, E. and Real, E. On the calculation of deflections in structural stainless
steel beams: an experimental and numerical investigation. J. Construct Steel
Res 2000; 54(1):109–33.
Moreyra M. The behaviour of cold-formed lipped channels under bending. MSc

Thesis, Cornell University, Ithaca, New York; 1993.
Moss PJ, Buchanan AH, Seputro J, Wastney C and Welsh R. Effect of support
conditions on the fire behaviour of steel and composite beams. Fire and
Materials 2004; 28: 159-175.
Najjar SR. Three-dimensional analysis of steel frames and subframes in fire. PhD
Thesis, University of Sheffield; 1994.
Najjar S and Burgess I. A nonlinear analysis for three-dimensional steel frames in

fire conditions. Engineering Structures 1996; 18(1): pp.77-89.
NAS. North American Specification for the Design of Cold-Formed Steel Structural
Members. American Iron and Steel Institute, November 9 draft Ed; 2001.
Olawale AO and Plank RJ. The collapse analysis of steel columns in fire using a
finite strip method, International Journal of Numerical Methods in Engineering
1988; 26: 2755-2764.
Outinen J. Mechanical properties of structural steels at elevated temperatures,
Licentiate Thesis, Helsinki University of Technology, Finland 1999.
Outinen J and Mäkeläinen P. A study for the development of the design of steel
structures in fire conditions. Proceedings of the First International Workshop
on Structures in Fire, Copenhagen; 2000: 269–81.
Outinen J and Makelainen P. Effect of high temperature on mechanical properties of

cold-formed structural steel. Ninth International Symposium on Tubular
Structures; 2001: 439-444.
Outinen J and Makelainen P. Behaviour of cold-formed thin-walled structural steel at

high temperatures and after cooling down. 3rd European Conference on Steel
Structures, Coimbra, Portugal; 2002a.
Outinen J and Makelainen P. Mechanical properties of structural steel at elevated

temperatures and after cooling down. Second International Workshop-
Structures in Fire, Christchurch; 2002b: 273-290.
Papazian RP, Schuster RM, Sommerstein M. Multiple stiffened deck profiles.

Proceedings of the Twelfth International Specialty Conference on Cold-
formed Steel Structures, St. Louis, Missouri; 1994: 217-228.
Peköz T. Development of a unified approach to the design of cold-formed steel

members. Advisory Group on the Specification For the Design of Cold-
Formed Steel Structural Members, American Iron and Steel Institute; 1987.
Pekoz TB and Sumer O. Design provisions for cold-formed steel columns and beam-
columns, Final Report, Cornell University, submitted to American Iron and
Steel Institute, September; 1992.
Pi Y and Trahair N. Prebuckling deflections and lateral buckling I: Theory. Journal

of Structural Engineering 1992a, 118(11): 2949-2966.
Pi Y and Trahair N. Prebuckling deflections and lateral buckling II: Applications.

Journal of Structural Engineering 1992b; 118(11): 2967-2985.
Pi YL and Trahair NS, Lateral buckling strengths of cold-formed rectangular hollow

sections, Thin-Walled Structures 1995: 22: 71-95.
Pi YL, Put BM and Trahair NS. Lateral buckling strengths of cold-formed channel
section beams, Research Report No. R748, The University of Sydney,
Australia; 1997.
Pi YL, Put B and Trahair NS. Lateral buckling strengths of cold-formed channel
section beams, Journal of Structural Engineering 1998; 124(10): 1182-1191.
Poutre DB la. Lateral-torsional buckling of channel sections-experimental research.

Eindhoven University of Technology, Report No. TUE/BCO/99.06, TNO
Report 1999-CON-R5020/1, TNO Bouw, Rijswijk; 1999.
Purkiss J. Fire safety engineering design of structures. Butterworth-Heinemann Ltd;

1996.
Put BM, Pi YL and Trahair NS. Bending and torsion of cold-formed channel beams.
Journal of Structural Engineering 1999a; 125(5): 540-546.
Put BM, Pi YL and Trahair NS. Lateral buckling tests on cold-formed channel
beams. Journal of Structural Engineering 1999b; 125(5): 532-539.
Ramberg W and Osgood WR. Description of stress-strain curves by three

parameters. NACA Technical Note 902, 1943.
Ranawaka T. Distortional buckling behaviour of cold-formed steel compression

members at elevated temperatures, PhD Thesis, Queensland University of
Technology, Brisbane, Australia, (2006).
Ranawaka T and Mahendran M. Experimental Study of the Mechanical Properties of
Light Gauge Cold-formed Steels at Elevated Temperatures. Fire Safety
Journal 2009a; 44: 219-229.
Ranawaka T and Mahendran M. Numerical modelling of light gauge cold-formed

steel compression members subjected to distortional buckling at elevated
temperatures. Thin-walled Structures 2009b; Accepted.
Ramberg W and Osgood WR. Description of stress-strain curves by three
parameters. NACA Technical Note 902; 1943.
Rasmussen KJR. Full-range stress–strain curves for stainless steel alloys. J Construct
Steel Research 2003; 59 (1): 47–61.
Ranby A. Structural fire design of thin walled steel sections. Journal of

Constructional Steel Research 1998; 46(1): 303-304.
Rasmussen KJR. Full-range stress–strain curves for stainless steel alloys. J Construct
Steel Research 2003; 59 (1): 47–61.
Real PMM and Franssen JM. Lateral buckling of steel I beams under fire
conditions—comparison between the Eurocode 3 and the SAFIR code.
Relatório interno; 1999.
Real PMM and Franssen JM. Lateral torsional buckling of steel i-beams in case of
fire—numerical modeling. Proceedings of the First International Workshop
on Structures in Fire 2000; 71-93.
Real PMM and Franssen JM. Numerical modelling of lateral buckling of steel I
beams under fire conditions—comparison with Eurocode 3. Journal of Fire
Protection Engineering, USA, 2001; 11(2): 112-28.
Real PMM, Piloto PAG and Franssen JM. A new proposal of a simple model for the
lateral torsional buckling of unrestrained steel I-beams in case of fire:
experimental and numerical validation. Journal of Constructional Steel
Research 2003; 59(2): 179-199.
Real PMM, Cazeli R, Simoes da Silva L, Santiago A and Piloto P. The effect of
residual stresses in the lateral-torsional buckling of steel I-beams at elevated
temperature. Journal of Constructional Steel Research 2004a; 60(3-5): 783-
793.
Real PMM, Lopes N, Simoes da Silva L and Franssen JM. Lateral-torsional buckling
of unrestrained steel beams under fire conditions: improvement of EC3
proposal. Computers & Structures 2004b; 82(20-21): 1737-1744.
Real PMM, Lopes N, Simoes da Silva L and Franssen JM. Numerical validation of
an improved proposal to the Eurocode 3 for the lateral-torsional buckling of
steel beams subjected to fire. Fourth International Workshop Structures in
Fire, Aveiro, Portugal; 2006: 151-160.
Renaud C and Zhao B. Investigation of simple calculation method in EN-1993-1-2

for buckling of hot rolled Class 4 steel members exposed to fire. Fourth
International Workshop - Structures in Fire, Aveiro, Portugal; 2006.
Rodrigues JPC, Cabrita Neves I and Valente JC. Experimental research on the
critical temperature of compressed steel elements with restrained thermal
elongation. Fire Safety Journal 2000; 35(2): 77-98.
Rogers CA. Interaction buckling of flange, edge stiffener and web of C-sections in
bending. M.S. Thesis, University of Waterloo, Ontario, Canada; 1995.
Rondal J. Cold formed steel members and structures - General Report. Journal of
Constructional Steel Research 2000; 55(1-3): 155-158.
Saab HA and Nethercot DA. Modelling of steel frame behaviour under fire
conditions. Engineering Structures 1991; 13: 371-382.
Schafer B and Pekoz T. Computational modelling of cold-formed steel:

characterizing geometrical imperfections and residual stresses, Journal of
Constructional Steel Research 1998; 47: 193 – 210.
Schafer B and Pekoz T. Direct strength prediction of cold-formed steel members

using numerical elastic buckling solutions. IN: Proceedings, Second
International Conference on Thin-Walled Structures, Thin Walled Structures,
Research and Development, Singapore; 1998b.: 137-144.
Schafer BW and Pekoz T. Laterally braced cold-formed steel flexural members with
edge stiffened flanges. Journal of Structural Engineering 1999; 125(2): 118-
127.
Schuster R. Testing of perforated c-stud sections in bending. Report for the Canadian
Sheet Steel Building Institute, University of Waterloo, Waterloo Ontario;
1992.
Shan M, LaBoube RA and Yu W. Behaviour of web elements with openings

subjected to bending, shear and the combination of bending and shear. Civil
Engineering Study Structural Series, 94-2, Department of Civil Engineering,
University of Missouri-Rolla, Rolla, Missouri; 1994.
Shifferaw Y and Schafer B. Inelastic bending capacity in cold-formed steel

members. Research Report, American Iron and Steel Institute, Committee on
Specifications; 2008.
Sidey MP and Teague DP. Elevated temperature data for structural grades of
Galvanised steel, British Steel (Welsh Laboratories) Report, UK; 1988.
Snijder HH, Bijlaard FSK and Steenbergen HMGM. FEM simulations of lateral-
torsional buckling experiments on channel sections loaded in bending.
Proceedings of the 3rd European Conference on Steel Structures, Coimbra,
Portugal; 2002: 201-210.
Standards Australia (SA) AS 4100, Steel Structures, Sydney, Australia, 1998.
Standards Australia/New Zealand, AS/NZS 4600, Cold-formed steel structures,
Sydney, Australia, 2005.
Standards Australia (SA) AS 1391, Methods for Tensile Testing of Metal, Sydney,
Australia, 1991.
Standards Australia (SA) AS 2291, Methods for the tensile testing of metals at
elevated temperatures, Sydney, Australia, 1979.
Standard SFS-EN 10002-1, Metallic Materials. Tensile Testing, Part 1-5,
Brussels/Helsinki, 1992.
Sultan MA. Parameters affect the fire resistance performance of floor systems,
NRCC-39799. National Research Council Canada; 2004.
Thomasson PO. Thin-walled c-shaped panels in axial compression. Swedish Council

for Building Research; D1:1978.
Timoshenko S and Gere J. Theory of elastic stability. McGraw-Hill Book Co. Inc.,
New York; 1961.
Trahair NS and Bradford M. The behaviour and design of steel structures. Chapman
and Hall New York; 1988.
Trahair NS. Flexural-torsional buckling of structures. Taylor & Francis Group; 1993.
Trahair NS. Laterally unsupported beams. Engineering Structures 1996; 18 No. 10:
759-768.
Trahair N. Multiple design curves for beam lateral buckling, Proceedings of the 5 th
International Colloquiumon Stability and Ductility of Steel Structures,
Nagoya, Japan; 1998: 33-44.
Ungermann D and Kalameya J. Realitatsnhe bestimmung der tragfahigkeit beul-und

biegedrillknickgefahrdeter, dunnwandiger Sahlquerschnitte. AiF-Research
Report No. 14125, German Association of Seel Constructions DASt
(Germany); 2006.
Von Karman T, Sechler EE and Donnell LH. (1932). The strength of thin plates in
compression. Transaction, ASME 1932; 54(5).
Weng C and Pekoz T. Residual stresses in cold-formed steel members. Journal of

Structural Engineering 1988; 116(6): 1611-1625.
Willis C and Wallace B. Behavior of cold-formed steel purlins under gravity loading.
Journal of Structural Engineering 1990; 116(8): 2061-2069.
Winter G, Lansing W and McCalley Jr. RB. Performance of laterally loaded channel
beams. Proc. Symp. on Engrg. Struct., Colston Research Society, U.K.; 1949:
49-60.
Winter G. Strength of thin steel compression flanges. Transactions ASCE 1947; 112:
527-554.
Winter G. Commentary on the 1968 Edition of the specification for the design of
cold-formed steel structural members. American Iron and Steel Institute;
1970.
Wong MB. Modelling of axial restraints for limiting temperature calculation of steel
members in fire. Journal of Constructional Steel Research 2005; 61(5): 675-
687.
Young B and Rasmussen K. Behaviour of cold-formed singly symmetric columns.
Thin-Walled Structures (UK) 1999; 33(2): 83-102.
Yin YZ and Wang YC. Numerical simulations of the effects of non-uniform

temperature distributions on lateral torsional buckling resistance of steel i-
beams. Journal of Constructional Steel Research 2003; 59(8): 1009-1033.
Yu WW. Cold-Formed Steel Design, 3rd Edition, John Wiley & Sons, Inc., USA;
2000.
Yu C and Schafer BW. Local buckling tests on cold-formed steel beams. 16th
International Specialty Conference, Cold-formed Steel Structures, Orlando,
FL; 2002.
Yu C and Schafer B. Distortional buckling of C and Z members in bending, Progress

Report Submitted to American Iron and Steel Institute (AISI); 2004.
Yu C and Schafer B. Distortional buckling tests on cold-formed steel beams. Journal

of Structural Engineering 2006; 132(4): 515-528.
Yu C and Schafer BW. Simulation of cold-formed steel beams in local and

distortional buckling with application to the direct strength method. Journal
of Constructional Steel Research 2007; 63: 581-590.
Zhao X, Hancock G and Trahair N. Lateral-buckling tests of cold-formed RHS

beams. Journal of Structural Engineering 1995; 121(11): 1565-1573.
Zeinoddini VM and Schafer BW. Out-of-straightness measurements for SSMA

studs, Steel Stud Manufacturers Association (SSMA); 2008.

Thesis

Transféré par

Informations du document

Copyright

Formats disponibles

Partager ce document

Partager ou intégrer le document

Options de partage

Avez-vous trouvé ce document utile ?

Ce contenu est-il inapproprié ?

Droits d'auteur :

Formats disponibles

Thesis

Transféré par

Droits d'auteur :

Formats disponibles

STRUCTURAL BEHAVIOUR AND DESIGN

OF COLD-FORMED STEEL BEAMS AT

NIROSHA DOLAMUNE KANKANAMGE

Nirosha Dolamune Kankanamge

School of Urban Developments

A THESIS SUBMITTED TO THE SCHOOL OF URBAN DEVELOPMENTS

Light gauge cold-formed steel, lateral-torsional buckling, section moment capacity,

Structural Behaviour and Design of Cold-formed Steel Beams at Elevated Temperatures v

Structural Behaviour and Design of Cold-formed Steel Beams at Elevated Temperatures vi

Keywords ................................................................................................................ iii

1 Introduction ............................................................................................ 1-1

2 Literature Review .................................................................................. 2-1

3 Mechanical Properties of Cold-formed Steels at Elevated

4 Experimental Investigation of Cold-formed Steel

5 Finite Element Modeling of Cold-formed Steel

6 Behaviour and Design of Cold-formed Steel Beams Subject to

Structural Behaviour and Design of Cold-formed Steel Beams at Elevated Temperatures ix

7 Behaviour and Design of Cold-formed Steel Beams Subject to

8 Conclusions and Recommendations ..................................................... 8-1

Structural Behaviour and Design of Cold-formed Steel Beams at Elevated Temperatures x

Appendix A Tensile Coupon Tests Results at Ambient and

REFERENCES ................................................................................................... R-1

Structural Behaviour and Design of Cold-formed Steel Beams at Elevated Temperatures xi

Figure 1.1: Building Composed of Cold-Formed Steel Sections ..................... 1-1

Structural Behaviour and Design of Cold-formed Steel Beams at Elevated Temperatures xv

Structural Behaviour and Design of Cold-formed Steel Beams at Elevated Temperatures xx

My sincere thanks goes to Queensland University of Technology (QUT) for giving

I dedicate this thesis to my beloved father, D.K. Ariyadasa,

Figure 1.1: Building Composed of Cold-Formed Steel Sections

Three methods are generally used in the manufacturing of cold-formed steel

This Figure is not available online.

(a) Cold-roll forming sequence for a C-section (b) Cold roll-forming

Figure 1.2: Manufacturing Process of Cold-Rolled Steel Sections

Figure 1.3: Press Braking (Karren, 1967)

The interest of engineers and architects in using cold-formed steel members in

• Consistent quality – They have uniform properties throughout the member

This Figure is not available online.

Figure 1.5: Fire in a Steel Structure

A good understanding of the structural behaviour of cold-formed steel members

1.2 Research Problem

Both buckling behaviour and mechanical properties of cold-formed steel members

Kaitila (2002) presented numerical modelling of simply supported cold-formed steel

Alfawakhiri et al. (2001) presented a comprehensive thermal-structural model for the

1.3 Research Objectives and Scope

Specific Objectives and Tasks

1. To investigate the mechanical properties of cold-formed steels at

1. To develop advanced finite element models of cold-formed steel lipped

2. To investigate the lateral torsional buckling failure of long unrestrained

3. To undertake a detailed numerical parametric study using the validated

Development of Design Rules

2. To develop accurate design rules for the member moment capacity of

This research concentrates on improving the understanding of the flexural behaviour

• The member moment capacity of beams subject to lateral torsional buckling

1.5 Research Methodology

1.6 Contents of the Thesis

The flexural behaviour of cold-formed steel beams at ambient and elevated

This chapter presents the experimental investigation of mechanical properties of

2.2 Characteristics and Mechanical Properties of Cold-Formed

Generally, the depth of cold-formed steel sections ranges from 50 mm to 300 mm

Figure 2.1: Effect of Cold Work on the Mechanical Properties of Cold-Formed

Figure 2.2: Effects of Strain Hardening and Strain Aging on Stress-Strain

(a) Tri-Linear (Flats) (b) Ramberg-Osgood Type Curves (Corners)

Figure 2.3: Stress-Strain Curves (Pi and Trahair, 1995)

2.3 Flexural Behaviour of Cold-Formed Steel Beams