Yasintha Heva Thesis PDF

BEHAVIOUR AND DESIGN OF COLD-
FORMED STEEL COMPRESSION MEMBERS

AT ELEVATED TEMPERATURES
YASINTHA BANDULA HEVA

Behaviour and Design of Cold-formed Steel
Compression Members at Elevated
Temperatures
By
Yasintha Bandula Heva
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
MARCH 2009
KEYWORDS
Light gauge cold-formed steel, local buckling, flexural-torsional buckling,

compression members, elevated temperatures, axial compression load, reduced yield
strength, reduced elasticity modulus, stress-strain model, fire safety design, fire test,
finite element analysis.
Behaviour and Design of Cold-formed Steel Compression Members at Elevated Temperatures iii
ABSTRACT
Cold-formed steel members have been widely used in residential, industrial and
commercial buildings as primary load bearing structural elements and non-load
bearing structural elements (partitions) due to their advantages such as higher
strength to weight ratio over the other structural materials such as hot-rolled steel,
timber and concrete. Cold-formed steel members are often made from thin steel
sheets and hence they are more susceptible to various buckling modes. Generally
short columns are susceptible to local or distortional buckling while long columns to
flexural or flexural-torsional buckling. Fire safety design of building structures is an
essential requirement as fire events can cause loss of property and lives. Therefore it
is essential to understand the fire performance of light gauge cold-formed steel
structures under fire conditions. The buckling behaviour of cold-formed steel
compression members under fire conditions is not well investigated yet and hence
there is a lack of knowledge on the fire performance of cold-formed steel
compression members. Current cold-formed steel design standards do not provide
adequate design guidelines for the fire design of cold-formed steel compression
members. Therefore a research project based on extensive experimental and
numerical studies was undertaken at the Queensland University of Technology to
investigate the buckling behaviour of light gauge cold-formed steel compression
members under simulated fire conditions.
As the first phase of this research, a detailed review was undertaken on the
mechanical properties of light gauge cold-formed steels at elevated temperatures and
the most reliable predictive models for mechanical properties and stress-strain
models based on detailed experimental investigations were identified. Their accuracy
was verified experimentally by carrying out a series of tensile coupon tests at
ambient and elevated temperatures.
As the second phase of this research, local buckling behaviour was investigated
based on the experimental and numerical investigations at ambient and elevated
temperatures. First a series of 91 local buckling tests was carried out at ambient and
elevated temperatures on lipped and unlipped channels made of G250-0.95, G550-
0.95, G250-1.95 and G450-1.90 cold-formed steels. Suitable finite element models
Behaviour and Design of Cold-formed Steel Compression Members at Elevated Temperatures iv

were then developed to simulate the experimental conditions. These models were
converted to ideal finite element models to undertake detailed parametric study.
Finally all the ultimate load capacity results for local buckling were compared with
the available design methods based on AS/NZS 4600, BS 5950 Part 5, Eurocode 3
Part 1.2 and the direct strength method (DSM), and suitable recommendations were
made for the fire design of cold-formed steel compression members subject to local
buckling.
As the third phase of this research, flexural-torsional buckling behaviour was

investigated experimentally and numerically. Two series of 39 flexural-torsional
buckling tests were undertaken at ambient and elevated temperatures. The first series
consisted 2800 mm long columns of G550-0.95, G250-1.95 and G450-1.90 cold-
formed steel lipped channel columns while the second series contained 1800 mm
long lipped channel columns of the same steel thickness and strength grades. All the
experimental tests were simulated using a suitable finite element model, and the
same model was used in a detailed parametric study following validation. Based on
the comparison of results from the experimental and parametric studies with the
available design methods, suitable design recommendations were made.
This thesis presents a detailed description of the experimental and numerical studies
undertaken on the mechanical properties and the local and flexural-torsional bucking
behaviour of cold-formed steel compression member at ambient and elevated
temperatures. It also describes the currently available ambient temperature design
methods and their accuracy when used for fire design with appropriately reduced
mechanical properties at elevated temperatures. Available fire design methods are
also included and their accuracy in predicting the ultimate load capacity at elevated
temperatures was investigated.
This research has shown that the current ambient temperature design methods are
capable of predicting the local and flexural-torsional buckling capacities of cold-
formed steel compression members at elevated temperatures with the use of reduced
mechanical properties. However, the elevated temperature design method in
Eurocode 3 Part 1.2 is overly conservative and hence unsuitable, particularly in the
case of flexural-torsional buckling at elevated temperatures.
Behaviour and Design of Cold-formed Steel Compression Members at Elevated Temperatures v

Publications
• Bandula Heva, D.Y.M. and Mahendran, M. (2007) Local Buckling Behaviour of

Cold-formed Steel Compression Members at Elevated Temperatures,
Proceedings of Smart System Theme Conference, Queensland University of
Technology, Brisbane, Australia, September 2007.
• Bandula Heva, D.Y.M. and Mahendran, M. (2008) Local Buckling Tests of Cold-
formed Steel Compression Members at Elevated Temperatures, Proceedings of
5th International Conference on Thin Walled Structures, Gold Coast, Australia,
June 2008, PP 745-752.
Behaviour and Design of Cold-formed Steel Compression Members at Elevated Temperatures vi

TABLE OF CONTENT
Keywords…………………………………………………………………………...iii
Abstracts……………………………………………………………………………iv
Publications………………………………………………………………………...vi
Table of Contents………………………………………………………………....vii
List of Figures……………………………………………………………………...xii
List of Tables………………………………………………………………...……xvii
Abbreviations……………………………………………………………………..xxi
Statement of Original Authorship……………………………………………...xxiii
Acknowledgments……………………………………………………………….xxiv
CHAPTER 1 Introduction ........................................................................ 1-1

1.1 Problems of Cold-formed Steel Members................................................ 1-3
1.1.1 Buckling Failure............................................................................... 1-3
1.1.2 Low Fire Resistance......................................................................... 1-3
1.2 Applications of Cold-formed Steel Structures in Building Industry........ 1-4
1.3 Fire Safety in Buildings ........................................................................... 1-5
1.4 Problem Definition................................................................................... 1-6
1.5 Objectives................................................................................................. 1-7
1.5.1 Main Objective................................................................................. 1-7
1.5.2 Specific Objectives........................................................................... 1-8
1.6 Limitations ............................................................................................... 1-9
1.7 Methodology ............................................................................................ 1-9
1.8 Content of the Thesis ............................................................................. 1-10
CHAPTER 2 Literature Review............................................................... 2-1

2.1 Buckling of Steel Compression Members................................................ 2-1
2.1.1 Local Buckling ................................................................................. 2-1
2.1.2 Distortional Buckling ....................................................................... 2-2
2.1.3 Global Buckling ............................................................................... 2-3
2.2 Mechanical Properties of Cold-formed Steels at Ambient and Elevated
Temperatures............................................................................................ 2-4
2.2.1 Previous Research ............................................................................ 2-5
2.2.2 Test Methods.................................................................................. 2-14
2.2.2.1 Specimens .................................................................................. 2-14
2.2.2.2 Test Rigs and Test Methods used by Previous Researchers ...... 2-15
2.2.3 Code Provisions ............................................................................. 2-17
2.3 Local Buckling Behaviour ..................................................................... 2-18
2.3.1 Local Buckling Behaviour at Ambient Temperature ..................... 2-18
2.3.1.1 Previous Research ...................................................................... 2-18
2.3.1.2 Test Methods.............................................................................. 2-21
2.3.1.3 Design Rules .............................................................................. 2-24
2.3.2 Local Buckling Behaviour at Elevated Temperatures ................... 2-27
2.3.2.2 Test Methods.............................................................................. 2-31
2.3.2.3 Design Rules .............................................................................. 2-33
Behaviour and Design of Cold-formed Steel Compression Members at Elevated Temperatures vii
2.4 Global Buckling Behaviour.................................................................... 2-34
2.4.1 Global Buckling Behaviour of Cold-formed Steel Members at
Ambient Temperature .................................................................................... 2-34
2.4.1.2 Test Methods.............................................................................. 2-36
2.4.1.3 Design Rules .............................................................................. 2-38
2.4.2 Global Buckling Behaviour at Elevated Temperatures.................. 2-42
2.4.2.2 Test Methods.............................................................................. 2-49
2.4.2.3 Design Rules .............................................................................. 2-51
2.5 Geometric Imperfections and Residual Stresses .................................... 2-53
2.5.1 Geometric Imperfections................................................................ 2-53
2.5.1.1 Local Imperfections ................................................................... 2-54
2.5.1.2 Global Imperfections.................................................................. 2-56
2.5.2 Residual Stresses............................................................................ 2-57
2.6 Analytical Methods ................................................................................ 2-59
2.7 Fire Safety Design.................................................................................. 2-60
2.8 Literature Review Findings.................................................................... 2-60
CHAPTER 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 .................................................................... 3-4
3.2.4 Mechanical Properties at Ambient and Elevated Temperatures .... 3-13
3.2.4.1 Determination of Yield Strength and Elastic Modulus .............. 3-13
3.2.4.2 Results ........................................................................................ 3-18
3.2.4.3 Comparison of Test Results with Available Predictive Method 3-19
3.2.4.3.1 Mechanical Property Reduction Factors .............................. 3-20
3.2.4.3.2 Stress-Strain Curves ............................................................. 3-21
3.3 Conclusions ............................................................................................ 3-26
CHAPTER 4 Experimental Study of Cold-formed Steel

Compression Members Subject to Local Buckling ......... 4-1
4.1 Experimental Investigation ...................................................................... 4-1
4.1.1 Local Buckling Tests at Ambient Temperature ............................... 4-2
4.1.1.1 Preliminary Investigation and Selection of Test Specimens........ 4-2
4.1.1.2 Test Set-up and Procedure ........................................................... 4-7
4.1.1.3 Observations and Results ........................................................... 4-10
4.1.2 Local Buckling Tests at Elevated Temperatures............................ 4-12
4.1.2.1 Preliminary Investigation and Selection of Test Specimens...... 4-13
4.1.2.2 Test Set-up and Procedure ......................................................... 4-16
4.1.2.3 Results of Local Buckling Tests at Elevated Temperatures....... 4-20
4.2 Comparison of Test Results with Predictions from the Current Design
Standards ................................................................................................ 4-23
4.2.1 Design Standards............................................................................ 4-23
4.2.2 Discussion of Results and Comparison.......................................... 4-31
4.3 Capacity Reduction Factor..................................................................... 4-34
Behaviour and Design of Cold-formed Steel Compression Members at Elevated Temperatures viii
4.4 Conclusions ............................................................................................ 4-36
CHAPTER 5 Development and Validation of Finite Element

Models of Cold-formed Steel Compression Members
Subject to Local Buckling.................................................. 5-1
5.1 Development of Experimental Finite Element Models............................ 5-2
5.1.1 Selection of Element Types ............................................................. 5-2
5.1.2 Selection of Element Size ................................................................ 5-3
5.1.3 Loading and Boundary Conditions .................................................. 5-4
5.1.4 Geometric Imperfections.................................................................. 5-7
5.1.5 Residual Stresses.............................................................................. 5-8
5.1.6 Mechanical Property Model........................................................... 5-11
5.1.7 Analysis Methods........................................................................... 5-17
5.2 Validation of Finite Element Models..................................................... 5-18
5.2.1 Validation of FEA Models Using Load-Shortening Curves .......... 5-18
5.2.2 Validation of FEA Models Using Ultimate Loads......................... 5-20
5.2.3 Validation of FEA Models Using Failure Modes .......................... 5-26
5.3 Comparison of FEA Results with Code Predictions .............................. 5-26
5.4 Development of Ideal Model for Local Buckling Simulations.............. 5-30
5.4.1 Boundary Conditions ..................................................................... 5-30
5.4.2 Model Length ................................................................................. 5-32
5.5 Conclusions ............................................................................................ 5-40
CHAPTER 6 Parametric Study and Numerical Analyses of Local

Buckling .............................................................................. 6-1
6.1 Parametric Study at Ambient Temperature Conditions ........................... 6-2
6.1.1 Lipped Channel Sections (Type A).................................................. 6-2
6.1.1.1 Comparison of FEA Results with Design Methods ..................... 6-2
6.1.2 Unlipped Channel Sections (Type B) .............................................. 6-7
6.1.2.1 Comparison of FEA Results with Design Methods ..................... 6-8
6.2 Parametric Study at Elevated Temperature Conditions ......................... 6-11
6.2.1 Lipped Channel Sections (Type A)................................................ 6-12
6.2.1.1 Comparison of FEA Results with Design Methods ................... 6-12
6.2.2 Unlipped Channel Sections (Type B) ............................................ 6-24
6.2.2.1 Comparison of FEA Results with Design Methods ................... 6-24
6.3 Conclusions ............................................................................................ 6-35
CHAPTER 7 Experimental Study of Cold-formed Steel

Compression Members Subject to Flexural-torsional
Buckling Behaviour............................................................ 7-1
7.1 Test Specimens........................................................................................... 7-2
7.2 Ambient and Elevated Temperature Test Set-up ....................................... 7-6
7.3 Test Procedure.......................................................................................... 7-14
7.4 Measurements of Geometric Imperfections ............................................. 7-16
7.5 Observations and Results ......................................................................... 7-19
7.6 Comparison of Test Results with Current Design Standards................... 7-31
7.7 Conclusion................................................................................................ 7-41
Behaviour and Design of Cold-formed Steel Compression Members at Elevated Temperatures ix

CHAPTER 8 Development and Validation of Finite Element
Models of Cold-formed Steel Columns Subject to
Flexural-Torsional Buckling ............................................. 8-1
8.1 Development of Experimental Finite Element Models............................ 8-2
8.1.1 Selection of Element Types ............................................................. 8-2
8.1.2 Selection of Element Size ................................................................ 8-3
8.1.3 Loading and Boundary Conditions .................................................. 8-4
8.1.4 Geometric Imperfections.................................................................. 8-7
8.1.5 Residual Stresses.............................................................................. 8-7
8.1.6 Mechanical Property Model............................................................. 8-8
8.1.7 Analysis Methods............................................................................. 8-8
8.2 Validation of Finite Element Models....................................................... 8-8
8.2.1 Validation of FEA Model using Load-Deflection Curves ............... 8-9
8.2.2 Validation of FEA Model Using Deflected Shape......................... 8-19
8.2.3 Validation of FEA model Using Ultimate Loads........................... 8-23
8.2.4 Development of Ideal Finite Element Models ............................... 8-26
8.3 Conclusion.............................................................................................. 8-26
CHAPTER 9 Parametric Study of Flexural-Torsional Buckling

Behaviour of Cold-formed Steel Compression
Members ............................................................................. 9-1
9.1 Details of Finite Element Model used in the Parametric Study ............... 9-1
9.2 Flexural-torsional Buckling Behaviour at Ambient Temperature ........... 9-2
9.2.1 Selection of Member Sizes for the Parametric Study ...................... 9-2
9.2.2 Comparison of Ultimate Loads with Predictions from the Current
Design Rules at Ambient Temperature ............................................................ 9-3
9.2.2.1 AS/NZS 4600 and the Direct Strength Method ........................... 9-4
9.2.2.2 Eurocode 3 Part 1.3 ...................................................................... 9-9
9.2.2.3 Eurocode 3 Part 1.2 .................................................................... 9-11
9.2.2.4 BS 5950 Part 5 ........................................................................... 9-12
9.3 Flexural-torsional Buckling Behaviour at Elevated Temperatures........ 9-14
9.3.1 Comparison of Ultimate Loads with Predictions from the Current
Design Rules .................................................................................................. 9-14
9.3.1.1 AS/NZS 4600 and the Direct Strength Method ......................... 9-37
9.3.1.2 BS5950 Part 5 ............................................................................ 9-41
9.3.1.3 Eurocode 3 Part 1.3 .................................................................... 9-42
9.3.1.4 Eurocode 3 Part 1.2 .................................................................... 9-44
9.4 Conclusion.............................................................................................. 9-45
CHAPTER 10 Conclusion and Recommendations................................. 10-1

10.1 Mechanical Properties of Cold-formed Steels at Elevated
Temperatures.......................................................................................... 10-3
10.2 Local Buckling Behaviour at Ambient and Elevated Temperatures...... 10-4
10.3 Flexural-torsional Buckling Behaviour at Ambient and Elevated
Temperatures.......................................................................................... 10-5
10.4 Future Research...................................................................................... 10-7
Behaviour and Design of Cold-formed Steel Compression Members at Elevated Temperatures x

Appendix A Stress-strain Curves of tensile Coupon Tests……………………..A-1
Appendix B Load-deflection Curves of Local Buckling Tests…………………B-1
Appendix C Sample Calculations for Local Buckling Capacities……………...C-1
Appendix D Load-deflection Curves of Flexural-torsional Buckling Tests……D-1
Appendix E Sample Calculations for Flexural-torsional Buckling capacities….E-1
Appendix F Sample Calculations for Capacity Reduction factor………………F-1
Appendix G Sample Calculations for plotting Flexural-torsional Buckling

capacities with Design Curves ……………………….…………..G -1
REFERENCES…………………………………………………………………..R-1
Behaviour and Design of Cold-formed Steel Compression Members at Elevated Temperatures xi

LIST OF FIGURES
Figure 1.1 Roll-forming Process ....................................................................... 1-2

Figure 1.2 Press Braking Process ...................................................................... 1-2
Figure 1.3 Cold-formed Steel Sections ............................................................. 1-2
Figure 1.4 Use of Cold-formed Steel Members in Residential Construction.... 1-4
Figure 1.5 Use of Cold-formed Steel Members in Trusses ............................... 1-5
Figure 1.6 Fire in a Steel Building .................................................................... 1-6
Figure 2.1 Local Buckling of a Column ............................................................ 2-2
Figure 2.2 Distortional Buckling of a Short Column (Ranawaka, 2006) .......... 2-2
Figure 2.3 Flexural Buckling of a Column (Gardner and Baddoo, 2006)......... 2-3
Figure 2.4 Flexural and Flexural-torsional Buckling (Young, 2005)................ 2-4
Figure 2.5 Converting Transient State Test Results to Stress-Strain Curves
(Outinen and Makelainen, 2004) ..................................................... 2-9
Figure 2.6 Comparison of Mechanical Properties of Heated Steel
Element After Cooling (Outinen and Makelainen, 2004).............. 2-10
Figure 2.7 Tensile Test Coupon from Dolamune Kankanamge (2009) .......... 2-15
Figure 2.8 Test Setup Used by Outinen and Makelainen (2004) .................... 2-15
Figure 2.9 Test Setup Used by Lee (2004)...................................................... 2-16
Figure 2.10 Test setup used by Ranawaka (2006)............................................. 2-17
Figure 2.11 Effective Width Concept from AS/NZS 4600 (SA, 2005) ............ 2-19
Figure 2.12 Locally Buckled Columns from Popovic et al. (2001) .................. 2-20
Figure 2.13 Tested Specimens and Finite Element Models from
Dhanalakshmi and Shanmugam (2001) ......................................... 2-21
Figure 2.14 Test Rig Used by Popovic et al. (2000) for Compression Tests .... 2-22
Figure 2.15 INSTRON Servo-controlled Universal Testing Machine
Used by Dhanalakshmi and Shanmugam (2001) ........................... 2-22
Figure 2.16 Tinius Olsen Testing Machine used by Lee (2004) and
Ranawaka (2006) ........................................................................... 2-23
Figure 2.17 Fixed-End Achieved by Ranawaka (2006) Using Pyrocrete
165 Coil Grout ............................................................................ 2-24
Figure 2.18 Fixed-end Support used by Dhanalakshmi and
Shanmugam (2001) ........................................................................ 2-24
Figure 2.19 Re-distribution of Stress and Effective Width Concept................. 2-25
Figure 2.20 Tests of Outinen and Myllymaki (1995)........................................ 2-29
Figure 2.21 Local Buckling Failure at Elevated Temperatures
(Yang et al., 2006a)........................................................................ 2-31
Figure 2.22 Electric Furnace used by Lee (2004) and Ranawaka (2006) ......... 2-32
Figure 2.23 Sketch of the Test Rig Used by Feng et al. (2003a)....................... 2-33
Figure 2.24 Test Rig used by Yang et al. (2002)............................................... 2-36
Figure 2.25 Test Rig used by Liu and Young (2003)........................................ 2-37
Figure 2.26 End Conditions of Specimens used by Ali and
O’Connor (2001)............................................................................ 2-44
Figure 2.27 Axial Force Generated in Columns at Elevated
Temperatures (Ali and O’Connor, 2001) ....................................... 2-45
Figure 2.28 Variation of Reduction Factor for Axial Strength of
Steel Columns with Temperature (Yang et al., 2006).................... 2-46
Figure 2.29 Effect of Slenderness Ratio on Axial Load Carrying
Capacity at Elevated Temperatures (Yang et al., 2006)................. 2-47
Behaviour and Design of Cold-formed Steel Compression Members at Elevated Temperatures xii
Figure 2.30 Analytical Results of Burgress et al. (1992) .................................. 2-48
Figure 2.31 Tests of Ali and O’Connor (2001) ................................................. 2-50
Figure 2.32 Horizontal Furnace used by Wang and Davies (2003) .................. 2-50
Figure 2.33 Fire Test Arrangement used by Aasen (1985) ............................... 2-51
Figure 2.34 Imperfection Types Defined by Schafer and Pekoz (1998) ........... 2-55
Figure 2.35 Imperfection Model used by Sivakumaran and Rahman (1998) ... 2-56
Figure 2.36 Sequence of Measuring Residual Stress from Batista and
Rodrigues (1992)............................................................................ 2-58
Figure 2.37 Measurement of Residual Stress by EDM method
(Weng et al, 1990).......................................................................... 2-58
Figure 2.38 Residual Stress Models Proposed by Schafer and Pekoz (1998)... 2-59
Figure 3.1 Tension Test Specimen and its Dimensions..................................... 3-3
Figure 3.2 Test Rig ............................................................................................ 3-5
Figure 3.3 Details of Test Arrangement ............................................................ 3-7
Figure 3.4 EDCAR (Experimental Data Collection And Recorder) ................. 3-8
Figure 3.5 Laser Speckle Extensometer ............................................................ 3-9
Figure 3.6 Strain Measurements...................................................................... 3-10
Figure 3.7 Typical Speckle Output for Strain Measurements ......................... 3-11
Figure 3.8 Stress-strain Curve for 0.42 mm Thick G550 Steel at 20oC .......... 3-12
Figure 3.9 Stress-strain Curve of High Strength Steels................................... 3-14
Figure 3.10 Stress-strain Curves at Elevated Temperatures.............................. 3-16
Figure 3.11 Stress-strain Curves at Different Temperatures ............................. 3-17
Figure 3.12 Comparison of Yield Strength Reduction Factors
with the Available Predictive Methods .......................................... 3-20
Figure 3.13 Comparison of Elastic Modulus Reduction Factors
with the Available Predictive Methods .......................................... 3-21
Figure 3.14 Comparison of Test and Predicted Stress-Strain
Curves at 20oC ............................................................................... 3-22
Curves at 100oC ............................................................................. 3-23
Curves at 200oC ............................................................................. 3-23
Curves at 300oC ............................................................................. 3-24
Curves at 400oC ............................................................................. 3-24
Curves at 500oC ............................................................................. 3-25
Curves at 600oC ............................................................................. 3-25
Curves at 700oC ............................................................................. 3-26
Figure 4.1 Selected Cross Section Types for Local Buckling Tests ................. 4-3
Figure 4.2 Buckling Plot of a Lipped Channel Section..................................... 4-4
Figure 4.3 Buckling Plot of an Unlipped Channel Section ............................... 4-4
Figure 4.4 Test Set-up using Tinius Olsen Testing Machine ........................... 4-8
Figure 4.5 Displacement Measurements using LVDTs..................................... 4-9
Figure 4.6 Load-Deflection Curves for G550-0.95-20-B................................ 4-10
Figure 4.7 Elastic Buckling Waves ................................................................. 4-11
Figure 4.8 Sequence of Local Buckling Failure .............................................. 4-11
Behaviour and Design of Cold-formed Steel Compression Members at Elevated Temperatures xiii
Figure 4.9 Variation of Elastic Buckling Loads with Increasing
Temperatures for G550-0.95-B...................................................... 4-13
Figure 4.10 Electrical Furnace and its Attachments.......................................... 4-18
Figure 4.11 Aligning of Loading Shafts............................................................ 4-18
Figure 4.12 Observed Local Buckling Failure at Elevated Temperatures ........ 4-20
Figure 4.13 Typical Load-Deflection Curves for G550-0.95-300-B ................ 4-21
Figure 4.14 Comparison of Test and Predicted Ultimate Loads at Ambient
Temperature ................................................................................... 4-27
Figure 4.15 Comparison of Test and Predicted Ultimate Loads at 200oC ........ 4-28
Figure 4.21 Variation of (E/fy)0.5 with Temperature for Different
Grades and Thickness of Cold-formed Steel ................................. 4-32
Figure 5.1 Comparison of FEA Simulations Using Different Shell
Element Types with Test Results of G550–0.95–20–A................... 5-3
Figure 5.2 Effect of Element Size on Elastic Buckling and Ultimate
Loads of G550-0.95-20-A Specimen ............................................... 5-4
Figure 5.3 Experimental Models Used in Finite Element Analyses ................. 5-5
Figure 5.4 Loading and Boundary Conditions for Experimental Model........... 5-7
Figure 5.5 Residual Stress Models Proposed by Schafer and Pekoz (1998)..... 5-9
Figure 5.6 Proposed Residual Stress Models for this Research ........................ 5-9
Figure 5.7 Residual Stress Model for the 0.95mm G550 Cold-formed
Steel Section at Ambient Temperature .......................................... 5-11
Figure 5.8 Stress-Strain Models ...................................................................... 5-12
Figure 5.9 Load-Shortening Curves of G550-0.95-A Specimen for
Different Mechanical Property Models.......................................... 5-14
Figure 5.10 Comparison of Load-Shortening Curves from FEA and
Experimental Tests......................................................................... 5-20
Figure 5.11 Comparison of Failure Modes as a Function of
Imperfection Direction in a Tested Specimen................................ 5-22
Figure 5.12 Comparison of Failure Modes from Finite Element
Analyses and Experimental Tests of Type B Specimens............... 5-24
Figure 5.13 Quarter Wave Length Model Developed with Rigid Surface........ 5-31
Figure 5.14 Quarter Wave Length Model Developed with MPCs .................... 5-32
Figure 5.15 Buckling Plot for a Type A Section from CUFSM Analyses ........ 5-33
Figure 5.16 Development of Quarter Wave Model Considering Symmetry.... 5-33
Figure 5.17 Variation of Elastic Buckling Load with the Model Length.......... 5-34
Figure 5.18 Different Model Types Used for Model Length Study.................. 5-38
Figure 5.19 Elastic Buckling Load of Models with Varying Model Lengths ... 5-39
Figure 5.20 Ultimate Load of Models with Varying Model Lengths............... 5-39
Figure 6.1 Comparison of FEA Results with DSM for Type A
Specimens at Ambient Temperature ................................................ 6-6
Figure 6.2 Effect of Web Width to Thickness Ratio – High Strength
Steels at Ambient Temperature........................................................ 6-6
Figure 6.3 Effect of Web Width to Thickness Ratio – Low Strength
Steels at Ambient Temperature........................................................ 6-7
Behaviour and Design of Cold-formed Steel Compression Members at Elevated Temperatures xiv
Figure 6.4 Comparison of FEA results with DSM for Type B
Specimens at Ambient Temperature .............................................. 6-10
Figure 6.5 Effect of Flange Width to Thickness Ratio – High
Strength Steels at Ambient Temperature ....................................... 6-10
Figure 6.6 Effect of Flange Width to Thickness Ratio – Low
Strength Steels at Ambient Temperature ....................................... 6-11
Figure 6.7 Variation of E/fy ratio with temperature......................................... 6-14
Figure 6.8 Variations of Mean Values of Type A Section
for AS/NZS 4600 ........................................................................... 6-22
for the Direct Strength Method ...................................................... 6-22
for BS 5950 Part 5.......................................................................... 6-23
for Eurocode 3 Part 1.2 .................................................................. 6-23
Figure 6.12 Variations of Mean Values of Type B Section
for AS/NZS 4600 ........................................................................... 6-33
for BS 5950 Part 5.......................................................................... 6-33
for Direct Strength Method ............................................................ 6-34
for Eurocode 3 Part 1.2 Method..................................................... 6-34
Figure 7.1 Buckling Plot of Lipped Channels ................................................... 7-3
Figure 7.2 Three Segment Furnace and Temperature Control system .............. 7-7
Figure 7.3 Two Segment Furnace ..................................................................... 7-7
Figure 7.4 Heating Coil Arrangement of the Furnace ....................................... 7-8
Figure 7.5 Location of Thermocouples ............................................................. 7-8
Figure 7.6 Control Panel of the Furnace............................................................ 7-9
Figure 7.7 Two Segment Furnace with a Test Column..................................... 7-9
Figure 7.8 Three Segment Furnace with a Test Column................................. 7-10
Figure 7.9 Loading Arrangement at the Upper End of the Furnace ................ 7-11
Figure 7.10 Loading Arrangement at the Bottom End of the Furnace .............. 7-12
Figure 7.11 Hydraulic Pump and Pressure Transducer ..................................... 7-13
Figure 7.12 Jacks and Jack Bases Used for the Tests........................................ 7-13
Figure 7.13 End Plates used for Making Fixed End Support ............................ 7-14
Figure 7.14 Measurement of Out-of-Plane Displacement................................. 7-16
Figure 7.15 Imperfection Measurement Device ................................................ 7-17
Figure 7.16 Variation of Imperfection along Specimen G450-1.90-2800-20 ... 7-17
Figure 7.17 Measurement of Temperature Profile along the Specimen............ 7-20
Figure 7.18 Flexural-torsional Buckling of G250-1.95 mm Specimens ........... 7-22
Figure 7.21 Interaction of Flexural-torsional Buckling with Local Buckling... 7-25
Figure 7.22 Load-Deflection Curves for Series 1 Specimen at Ambient
Temperature (G450-1.90-2800-20)................................................ 7-26
Figure 7.23 Load-Deflection Curves for Series 1 Specimen at Elevated
Temperature-700oC (G450-1.90-2800-700) .................................. 7-27
Figure 7.24 Load-Deflection Curves for Series 2 Specimen at Ambient
Temperature (G450-1.90-1800-20)................................................ 7-28
Behaviour and Design of Cold-formed Steel Compression Members at Elevated Temperatures xv

Figure 7.25 Load-Deflection Curves for Series 2 Specimen at Elevated
Temperature-700oC (G450-1.90-1800-700) .................................. 7-29
Figure 7.26 Comparison of Test Results with AS/NZS 4600 Design Curve .... 7-33
Figure 7.27 Design Curves of Eurocode 3 Part 1.3 and Part 1.2....................... 7-34
Figure 8.1 Ultimate Load of G450-1.90-2800-20 with Different Mesh Sizes .. 8-4
Figure 8.2 Idealisation of Half Column Model and Boundary Conditions ....... 8-5
Figure 8.3 Loading and Boundary Conditions Applied to the Half
Length Column Model ..................................................................... 8-6
Figure 8.4 Comparison of Load-Deflection Curves from Tests and Finite
Element Analyses for G450-1.90-2800-20 Specimens.................. 8-10
Element Analyses for G450-1.90-2800-700 Specimens................ 8-11
Element Analyses for G4250-1.905-1800-700 Specimens............ 8-17
Figure 8.13 Comparison of Failure Modes for Validation Purpose .................. 8-12
Figure 9.1 Comparison of FEA Results with AS/NZS 4600 Prediction ........... 9-4
Figure 9.2 Comparison of FEA Ultimate Load of G250-1.95-60-40-15
Series with AS/NZS 4600 Design Curves
at Ambient Temperature .................................................................. 9-8
Figure 9.3 Comparison of FEA Results of G450-1.90-60-40-15 Series
with AS/NZS 4600 Design Curve at Ambient Temperature .......... 9-9
Figure 9.4 Comparison of FEA Ultimate Loads with
Eurocode 3 Part 1.3 Design Curves .............................................. 9-11
Figure 9.5 Comparison of FEA Ultimate Loads with AS/NZS 4600
Design Curve at Elevated Temperature ......................................... 9-39
Figure 9.6 Comparison of Stress-strain Curves at Elevated Temperatures..... 9-41
Figure 9.7 Comparison of FEA Ultimate loads with
Eurocode 3 Part 1.3 Buckling Curves ............................................ 9-44
Behaviour and Design of Cold-formed Steel Compression Members at Elevated Temperatures xvi
LIST OF TABLES
Table 2.1 Mechanical Properties of Cold-formed Steel at Elevated
Temperatures (Outinen et al., 2000) ................................................ 2-7
Table 2.2 Mechanical Properties Obtained from Transient State Tests
(Outinen and Makelainen, 2004) ..................................................... 2-9
Table 2.3 Reduction Factors of Mechanical Properties of Cold-formed
Steels at Elevated Temperatures (Chen and Young, 2006) ........... 2-11
Table 2.4 Yield strength Reduction Factors for Cold-formed Steel from
BS 5950 Part 8 (BSI, 1990) ........................................................... 2-18
Table 2.5 Experimental Results of Ali and O’Connor, 2001 ......................... 2-45
Table 3.1 Dimensions of Tensile Test Specimens ........................................... 3-4
Table 3.2 Yield Strength and Elastic Modulus at Ambient Temperature ...... 3-13
Table 3.3 Yield Stress of G550-0.42 mm Cold-formed Steel........................ 3-18
Table 3.4 Elastic Modulus of G550-0.42 mm Cold-formed Steel ................. 3-18
Table 3.5 Mechanical Property Reduction Factors for G550-0.42 mm
Cold-formed Steels ........................................................................ 3-19
Table 4.1 Effective Widths of Stiffened and Unstiffened Elements ................ 4-3
Table 4.2 Nominal Dimensions of Test Specimens ......................................... 4-5
Table 4.3 Base Metal Thickness of Selected Cold-formed Steels ................... 4-6
Table 4.4 Measured External Dimensions of Type A Specimens Tested at
Ambient Temperature ...................................................................... 4-6
Table 4.5 Measured External Dimensions of Type B Specimens Tested at
Ambient Temperature ...................................................................... 4-7
Table 4.6 Results of Local Buckling Tests at Ambient Temperature ............ 4-12
Table 4.7 Measured Dimensions of Type A Specimens ................................ 4-14
Table 4.8 Measured Dimensions of Type B Specimens ................................ 4-15
Table 4.9 Comparison of Ultimate Loads using the Two
Loading Set-ups ............................................................................. 4-19
Table 4.10 Ultimate Loads of Type B Specimens at Elevated
Temperatures.................................................................................. 4-22
Table 4.11 Ultimate Loads of Type A Specimens at Elevated
Temperatures.................................................................................. 4-23
Table 4.12 Comparison of Ultimate Loads from Tests and
Design Standards for Type A Sections .......................................... 4-25
Table 4.13 Comparison of Ultimate Loads from Tests and
Design Standards for Type B Sections .......................................... 4-26
Table 5.1 Mechanical Properties of the Selected Cold-Formed Steels .......... 5-16
Table 5.2 Finite Element Analysis Results for Type A Specimens ............... 5-23
Table 5.3 Effects of Positive and Negative Imperfections on the
Ultimate Loads of Type A Sections ............................................... 5-24
Table 5.4 Finite Element Analysis Results for Type B Specimens ............... 5-25
Table 5.5 Comparison Code Predictions with FEA Results for
Type A-Experimental Modelling ................................................... 5-28
Table 5.6 Comparison Code Predictions with FEA Results for
Type B-Experimental Modelling ................................................... 5-29
Table 5.7 Comparison of Elastic Buckling Loads from CUFSM and
FEA for Quarter Wave Length Model ........................................... 5-35
Behaviour and Design of Cold-formed Steel Compression Members at Elevated Temperatures xvii
Table 5.8 Comparison of Elastic Buckling and Ultimate Loads for
Type A Specimens with Different Model Lengths ........................ 5-36
Table 5.9 Comparison of Elastic Buckling and Ultimate Loads for
Type B Specimens with Different Model Lengths ........................ 5-36
Table 6.1 Comparison of FEA Results with Code Predictions for
Type A Sections at Ambient Temperature....................................... 6-5
Type B Sections at Ambient Temperature ....................................... 6-9
Type A Sections at 100oC .............................................................. 6-15
Type B Sections at 100oC .............................................................. 6-26
Table 7.1 Nominal Dimensions of Test Specimens ......................................... 7-2
Table 7.2 Base Metal Thickness of Selected Cold-formed Steels ................... 7-3
Table 7.3 Measured External Dimensions of 1800 Test Series Specimens ..... 7-4
Table 7.4 Measured External Dimensions of 2800 Test Series Specimens ..... 7-5
Table 7.5 Measured Maximum Imperfections of 1800 Test Series
Specimens ...................................................................................... 7-18
Table 7.6 Measured Maximum Imperfections of 2800 Test
Series Specimens............................................................................ 7-19
Table 7.7 Ultimate Compression Loads Obtained for the Shorter
Column Series ................................................................................ 7-30
Table 7.8 Ultimate Compression Loads Obtained for the Longer
Column Series ................................................................................ 7-31
Table 7.9 Comparison of Test Results with Code Prediction
for the G250-1.95-2800 Test Series ............................................... 7-35
Behaviour and Design of Cold-formed Steel Compression Members at Elevated Temperatures xviii
Table 8.1 Comparison of Test and FEA Ultimate Loads of
G450-1.90-2800-20 for Different Element Types ........................... 8-3
Table: 8.2 Comparison of Ultimate Loads for G450-1.90-2800 Series .......... 8-23
Table 8.3 Comparison of Ultimate Loads for G250-1.95-2800 Series .......... 8-24
Table 9.1 Comparison of FEA Ultimate loads with Code
Predictions for G250-1.95 Series at Ambient Temperature............. 9-5
Table 9.2 Comparison of FEA Ultimate with Code Predictions
for G450-1.90 Series at Ambient Temperature................................ 9-6
Table 9.3 Comparison of FEA Ultimate with Code Predictions
for G550-0.95 Series at Ambient Temperature................................ 9-7
Table 9.4 Comparison of Ultimate Loads from the Current and
Modified Design Methods of BS5950 Part 5................................. 9-13
Table 9.5 Comparison of FEA Ultimate Loads with Code
Predictions for G250-1.95 Series at 100oC .................................... 9-15
Table 9.9 Comparison of FEA Ultimate Loads Results with
Code Predictions for G450-1.90 Series at 200oC........................... 9-19
Behaviour and Design of Cold-formed Steel Compression Members at Elevated Temperatures xix
Table 9.26 Summary of the Comparative Analyses of FEA
Ultimate Loads with Code Predictions .......................................... 9-36
Table 9.27 Proposed Buckling Curves from Eurocode 3 Part 1.3
for the Fire Design of Low Strength Steel columns....................... 9-42
Table 9.28 Proposed Buckling Curves from Eurocode 3 Part 1.3
for the Fire Design of Low Strength Steel columns....................... 9-42
Behaviour and Design of Cold-formed Steel Compression Members at Elevated Temperatures xx

ABBREVIATIONS
A,Ag cross sectional area

Ae,Aeff effective cross section area
b plate width or (flange width)
be effective plate width
BMT Base Metal Thickness
Cp Correction factor depending on the number of tests
Cw warping constant
E (E20) elasticity modulus at ambient temperature
ET elasticity modulus at temperature T
fcrl critical elastic buckling stress
Fm mean of the fabrication factor
fn ultimate stress
foc elastic flexural, torsional or flexural-torsional buckling stress
fox elastic flexural buckling stress of the member about x -axis
foxz elastic flexural-torsional buckling stress of the member
foy elastic flexural buckling stress of the member about y -axis
fu ultimate tensile stress
fy (fy20) yield strength at ambient temperature
fyT yield strength at temperature T
G shear modulus
J torsional constant
Ix second moment of area about x-x axis
Iy second moment of area about y-y axis
k Plate buckling coefficient
kET reduction factor of elasticity modulus
kyT reduction factor of yield strength
L length
LE,le effective length
m degree of freedom (n-1)
Mm Mean of the material factor
n number of tests
N Section capacity
N* design axial compression force
Nb,Rd design buckling resistance for axial compression
Nc member axial capacity
Behaviour and Design of Cold-formed Steel Compression Members at Elevated Temperatures xxi
Pc member capacity
Pcs short strut capacity
PE elastic buckling load
PEX elastic flexural buckling load
PT torsional buckling load
Pm mean value of the tested to predicted load ratio
Pu applied axial load
Py crushing load
r radius of gyration
s stiffener length
T temperature
t thickness
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
α imperfection factor
β coefficient
β0 Target reliability index
Δ magnitude of initial imperfection
δ out-of-plane displacement
ε strain
εlnpl logarithmic plastic strain
εnom nominal strain
εT strain at temperature T
εy strain at yield stress
εyT strain at yield stress at temperature T
ηT parameter for Ramberg-Osgood model at temperature T
ηT Perry coefficient
λ slenderness ratio
ρ ratio of effective width to actual width of the plate
σ stress
σnom nominal stress
σtrue true stress
Φ the capacity reduction factor
Behaviour and Design of Cold-formed Steel Compression Members at Elevated Temperatures xxii
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:
Behaviour and Design of Cold-formed Steel Compression Members at Elevated Temperatures xxiii
ACKNOWLEDGEMENTS
First and foremost I offer my sincerest gratitude to my supervisor, Prof. Mahan

Mahendran, who has supported me throughout this research with his patience and
knowledge whilst providing constructive suggestions, encouragements and the
valuable assistance in many ways over the past three and a half years. This study
would not have been success to this level without such assistance. Dr. Mat
Humphreys to be mentioned with thanks for kindly agreeing to serve as an associate
supervisor. Special thanks to Dr. Ranawaka who provides valuable guldens to make
success in this project.
I would like to thank the Sri Lankan Government for offering me Presidential Fund
Scholarship which provided tuition fee for my first semester of this research project
and also Queensland University of Technology (QUT) and the Department of
Education, Science and Training, Australia for providing financial support of my
research project through the International Postgraduate Research Scholarship (IPRS).
Thanks also to the School of Urban Development at QUT for providing the
necessary facilities and technical support. Special thanks to Prof. John Bell and
faculty of Built Environment and Engineering for providing financial support for the
furnace built for this research.
Many thanks the Structural Laboratory staff members for assistance with fabricating,
preparing and testing the cold-formed steel specimens.
It is my pleasure to thank my fellow post-graduate students, friends for their support

and contribution to this research.
I would like to express my sincere appreciation to my wife and son for providing
enormous supports, encouragements and their understanding.
Finally I would like to express thanks to my parents, my mother-in-law and my

brothers and sisters for providing endless support and encouragement.
Behaviour and Design of Cold-formed Steel Compression Members at Elevated Temperatures xxiv
1 Introduction
Cold-formed steel members have been widely used in residential, industrial and
commercial buildings as primary load bearing structural elements, non-load bearing
structural elements (partitions), and curtain wall applications. Cold-formed steel
members are made of thin steel sheets. They are formed into required shapes either
by roll-forming (see Figure 1.1) or by press-breaking (see Figure 1.2) through a
series of dies. Press braking is often used in the production of a small quantity of
simple shapes. Cold roll-forming is the most popular method of mass production of
the cold-formed structural members as well as the claddings. In this process, the steel
sheet is fed into a series of rolls, each of which form the shape progressively until it
reaches the required shape. Cold-formed steel members are available in a variety of
complex sections due to ease of forming at ambient temperature (see Figure 1.3).
Cold-forming increases the strength of steel by about 20% to 50% depending on the
type of forming and thickness of steel. Furthermore, cold-formed and thin-walled
steel sections have the following advantages.
♦ Availability
♦ Ease of prefabrication and mass production
♦ Uniform quality
♦ Light weight
♦ Design flexibility
♦ Economy in transportation and handling
♦ Fast and simple erection or installation
♦ Pleasing appearance
♦ Easy to later modifications
♦ Easy to connect members.
Due to these advantages, the use of cold-formed steel members has increased
dramatically in recent times in the building industry as well as in the other structural
applications.
_________________________________________________________________
Behaviour and Design of Cold-formed Steel Compression Members at Elevated Temperatures
1-1
This Figure is not available online.
Please consult hardcopy thesis available
from QUT library
Figure 1.1: Roll-forming Process

(http://www.hbfxywj.com/ENGLISH/info/147-1.htm)

from QUT library
Figure 1.2: Press Braking Process (http://www.weldall.co.uk/Press_Brake.jpg)

from QUT library
Figure 1.3: Cold-formed Steel Sections

(http://imgs.tootoo.com/ed/64/ed643ec92d855ca11721cb35a0be5f46_m.jpg)
_________________________________________________________________
1-2
1.1 Problems of Cold-formed Steel Members
1.1.1 Buckling Failure
Since the thickness of cold-formed steel members is small, they are subjected to
local buckling, distortional buckling, flexural buckling, flexural-torsional buckling or
their combinations. Generally short columns fail due to local buckling and/or
distortional buckling while long columns fail due to flexural buckling, flexural
torsional buckling or their combinations. Lack of design rules is also a major issue
with the design of cold-formed steel members as they are limited to ambient
temperature conditions.
1.1.2 Low Fire Resistance
One of the major issues of steel structures is that they can be subjected to fire which
will cause a loss of lives and properties, not only because of fire but also due to the
structural failure of buildings due to deterioration of the mechanical properties of the
steel at elevated temperatures. Therefore fire safety design of building structures has
received greater attention in recent times and there is a need to assess the
performance of light gauge cold-formed steel structures under fire conditions.
Although some past research have focussed on fire, they are predominantly focussed
on heavier, hot-rolled steel members, and not on light gauge cold-formed steel
structures.
The structural behaviour of cold-formed steel compression members submitted to

fire has been studied by other researchers recently. Most of their research has
focused on local and distortional buckling of cold-formed steel compression
members at ambient and elevated temperatures. Only limited literature could be
found for flexural and flexural-torsional buckling of cold-formed steel compression
members at elevated temperatures.
_________________________________________________________________
1-3
1.2 Applications of Cold-formed Steel Structures in Building
Industry
(a) Deck/Cladding Applications

Cold-formed steel is used for deck, roof cladding and wall cladding applications. In
the case of roof and wall cladding usually cold-formed steel with a thickness of
0.42mm or 0.48mm is used. The design criteria for roof cladding include wind load,
snow load and imposed load due to foot traffic. Cold-formed steel is also used for
composite decks. These deck systems should have sufficient strength to withstand
the wet concrete load, construction load, reinforcement load etc. Usually a
trapezoidal section is used for composite slabs.
(b) Light Weight Floor Systems
Cold-formed steel members are used as floor beams with plywood boards or oriented
strand-boards. The use of such floor systems leads to light weight structures.
(c) Residential Uses
The use of cold-formed steel in residential construction instead of timber has

increased dramatically in recent times (see Figure 1.4). This trend is due to the
higher architectural and structural performance of light gauge cold-formed steels
compared to timber and other constructional materials.

from QUT library
Figure 1.4: Use of Cold-formed Steel Members in Residential Construction

(http://www.eng.mcmaster.ca/civil/faculty/sivakumaran/CFS-building.gif)
_________________________________________________________________
1-4
(d) Use as Truss Members
There is an increasing trend of usage of cold-formed steel in truss systems in

industrial applications (see Figure 1.5). This is basically due to the better quality of
cold-formed steel members in terms of strength, availability of various shapes,
pleasing aesthetics, ease of connection etc. compared to the hot-rolled steel or timber
members.

from QUT library
Figure 1.5: Use of Cold-formed Steel Members in Trusses

(http://www.marinoware.com/Images/project_desktop.jpg)
(e) Steel Stud Wall Panel Systems
Cold-formed steel studs are used in wall panels made of lining boards such as
gypsum plasterboard. These lining boards provide lateral restraints to the steel studs
and at the same time they provide significant fire and weather resistance.
(f) Other Applications
Cold-formed steel members are used as door and window frames, storage racks,
frames for display boards, etc.
1.3 Fire Safety in Buildings
Fire is a major issue in buildings. This issue becomes serious when a building is not
designed for unwanted fires. Structural systems should have sufficient fire resistance
for evacuation and further, it should have been properly designed to prevent the
spread of fire. Otherwise, spreading of fire causes severe damage to structures (see
_________________________________________________________________
1-5
Figure 1.6). Both the structural system and the fire safety system such as the
sprinkler, fire extinguisher, fire exit door etc. are equally important for a building to
minimize fire damage. Integration of all fire safety components gives rise to a safe
system. Design of structural systems to resist the fire is important because failure in
the structural system before the specified time gives rise to loss of lives and property
due to the collapse of buildings. Therefore the primary objective of fire safety in
terms of the structural point of view is to reduce the level of fire damage to the
structural system of a building by proper designs and detailing with increased
knowledge in fire safety.

from QUT library
Figure 1.6: Fire in a Steel Building (http://cache.boston.com/bonzai-

fba/Globe_Photo/2007/12/15/1197726448_0520.jpg)
1.4 Problem Definition
Cold-formed steel members provide high strength to weight ratio and are also easy to
construct when compared with thicker hot-rolled steel members. Further, their
pleasing appearance gives rise to achieving better architectural concepts. However,
cold-formed steel members have reduced fire resistance because of rapid temperature
rise due to their higher thermal conductivity and higher section factor. Therefore
cold-formed steel structures should be properly designed. But available information
in the literature on the behaviour of cold-formed steel structures at elevated
_________________________________________________________________
1-6
temperatures is scarce and hence an extensive set of fire test data is not available to
form valid design methods at elevated temperatures.
BS 5950 Part 8 (BSI, 1990) and Eurocode 3 Part 1.2 (ECS, 2005) provide design
rules for steel structures subjected to fire. These codes mainly focus on the design of
hot-rolled steel members at elevated temperatures. However, Eurocode 3 Part 1.2
recommends to use the same guidelines for the design of cold-formed steel members
at elevated temperatures with some limitations. It recommends ambient temperature
effective area for all temperatures and limits the maximum temperature to 350oC.
Past research showed that the limited design guidelines of Eurocode 3 Part 1.2 for
fire design of cold-formed steel members are not sufficient and they lead to
uneconomical designs.
Designers and researchers have used ambient temperature design guidelines for the
elevated temperature design of cold-formed steel members with the use of reduced
mechanical properties at elevated temperatures. Sine the behaviour of hot-rolled and
cold-formed steel members is not identical, the use of ambient temperature design
method needs to be validated using extensive experimental and numerical results. To
date extensive numerical and experimental research data is not available to validate
the use of ambient temperature design method for elevated temperature design of
cold-formed steel compression members.
To date, no detailed research that involves both experimental and numerical

investigations could be found on local, flexural and flexural-torsional buckling of
light gauge cold-formed steel compression members at elevated temperatures. Now,
this problem has become critical because the usage of light gauge cold-formed steel
members is rising while the threat of fire disasters in buildings is increasing.
1.5 Objectives
1.5.1 Main Objective
The main objective of this research is to assess the accuracy of current elevated
temperature design methods and the use of ambient temperature design methods with
_________________________________________________________________
1-7
reduced mechanical properties at elevated temperatures with respect to local and
flexural-torsional buckling of cold-formed steel compression members.
1.5.2 Specific Objectives
1. To carry out an extensive literature review on mechanical properties and

buckling behaviour of cold-formed steel compression members at ambient
and elevated temperatures.
2. To investigate the mechanical properties of cold-formed steels at elevated

temperatures. This involves the review of current predictive methods and to
assess the accuracy of them using tensile coupon tests of cold-formed steel
at elevated temperatures.
3. To develop suitable experimental designs including test sections, end

supports, and test setup for local and flexural-torsional buckling tests of light
gauge cold-formed steel members under ambient and elevated temperatures.
4. To develop a new electric furnace capable of testing the cold-formed steel

compression members having different heights up to 3 m and temperatures
up to 1000oC. The furnace height should be adjustable to suit the specimen
height. Further, it should be able to regulate the temperature throughout the
furnace to ensure uniform temperature in the columns.
5. To develop experimental finite element models capable of simulating local

and flexural-torsional buckling behaviour of light gauge cold-formed steel
compression members at ambient and elevated temperatures using the finite
element program ABAQUS and to validate them using experimental results.
6. To develop ideal finite element models to carry out a parametric study for
local and flexural-torsional buckling behaviour of light gauge cold-formed
steel compression members at ambient and elevated temperatures.
_________________________________________________________________
1-8
7. To investigate the local and flexural torsional buckling behaviour of light
gauge cold-formed steel compression members at ambient and elevated
temperatures using simulated fire tests (electric furnace) and numerical
modelling (ABAQUS).
8. To assess the accuracy of current ambient temperature design method for

elevated temperature design with the use of reduced mechanical properties
using the results experimental and numerical analyses for cold-formed steel
compression members.
9. To assess the accuracy of elevated temperature design methods for cold-

formed steel compression members using the results of experimental and
numerical analyses.
1.6 Limitations
This research was limited to channel sections as they are the most common sections
used for cold-formed structural steel columns. In the case of local buckling both
lipped and unlipped channels were considered whereas only lipped channel section
was considered for flexural-torsional buckling.
Effect of corner in these sections was not considered due to the following reasons:
• Corner radius are is small for the sections considered in this research
• Both yield stress and residual stress at the corners reduce at similar rate at
elevated temperatures.
1.7 Methodology
A detailed literature review was undertaken to understand the buckling behaviour

and variation of mechanical properties of cold-formed steel compression members at
ambient and elevated temperatures. A set of tensile coupon tests was carried out to at
ambient and elevated temperatures to verify the most recent and accurate mechanical
_________________________________________________________________
1-9
property predictive methods by Ranawaka and Mahendran (2009) and Dolamune
Kankanamge (2009). A series of local buckling tests on short lipped and unlipped
channel columns was carried out at ambient and elevated temperatures to validate the
finite element models for local buckling behaviour. Validated finite element models
were then converted to ideal finite element models to simulate ideal local buckling
behaviour to carry out a detailed parametric study. Similar methodology was
followed for flexural-torsional buckling behaviour of cold-formed steel lipped
channel columns with variable column height. A detailed parametric study was then
undertaken and their results were compared with the available design methods to
assess their accuracy.
This research is based on the strength domain used in fire safety design. In the
strength domain approach, the strength of fire exposed structural members is
assessed using the maximum temperature in the member during the design fire event.
The maximum temperature can be obtained using conventional thermal modeling
whereas the strength of structural members can be determined based on this
maximum temperature and assuming it to be uniform. This research has provided
useful outcomes that can be used in the latter, and thus will help improve fire safety
design of cold-formed steel compression members.
1.8 Content of the Thesis
The outline of this thesis is as follows:
Chapter 1 presents the general introduction of cold-formed steel structures and their
applications. Also it provides the current methods for ambient and
elevated temperature design. Finally it explains the objectives of this
research and the methodology of achieving them.
Chapter 2 presents the literature review findings based on this investigation. It
describes the mechanical properties and buckling behaviour of cold-
formed steel columns at ambient and elevated temperatures. Further, it
describes the current design methods of the cold-formed steel
compression members at ambient and elevated temperatures.
_________________________________________________________________
1-10
Chapter 3 presents the experimental investigation of mechanical properties at
ambient and elevated temperatures. It assesses the accuracy of different
mechanical property predictive methods.
Chapter 4 presents the details of the experimental investigation on the local
buckling behaviour of cold-formed steel columns at ambient and
elevated temperatures and the results.
Chapter 5 presents the validation of the experimental finite element models
developed to simulate local buckling behaviour and the development of
ideal models to carry out a detailed parametric study.
Chapter 6 presents the results of parametric study on local buckling behaviour and
the acceptability of current design methods of cold-formed steel
compression members subject to local buckling at ambient and elevated
temperatures.
Chapter 7 presents the experimental investigation on the flexural-torsional buckling
behaviour of cold-formed steel columns at ambient and elevated
temperatures and the results.
Chapter 8 presents the validation of the experimental finite element models
developed to simulate flexural-torsional buckling behaviour and the
development of ideal models to carry out a detailed parametric study.
Chapter 8 presents the results of parametric study on flexural-torsional buckling
behaviour and the acceptability of current design methods of cold-formed
steel compression members subject to flexural-torsional buckling at
ambient and elevated temperatures.
Chapter 10 presents the presents the significant findings from this research and
recommendations for further research.
_________________________________________________________________
1-11
2 Literature Review
This research focussed on the buckling behaviour of cold-formed steel compression

members under fire conditions. The key elements of this research are the different
types of buckling modes, available design methods and the essential mechanical
properties at elevated temperatures. Therefore literature review was carried out on
each key element. This chapter describes the different types of buckling modes, past
research on fire design, available design rules and the mechanical properties at
2.1 Buckling of Steel Compression Members
Buckling is one of the failure criteria of compression members that can occur well
below its yield capacity. It can be subdivided into two major categories: local and
global buckling. Local buckling is the deflection of individual plate elements relative
to the corners or stiffened elements. Global buckling is the deflection of whole
section relative to the ends of the member. These two major buckling modes can be
further sub-divided into several modes depending on the deflected shapes. Those
buckling modes are local buckling, distortional buckling, flexural buckling, torsional
buckling and flexural-torsional buckling. In addition to these buckling modes,
combinations of the above buckling modes, in particular local buckling with other
modes, can be frequently observed in the light-gauge cold-formed steel compression
members basically due to the low thickness.
2.1.1 Local Buckling
Local buckling (see Figure 2.1) is the flexural displacement of plate elements, with
the line junctions between them remaining straight (Hancock, 1998). This is
common with cold-formed steel members as they have a small thickness and hence,
large width to thickness ratios. Thinner plate elements are more susceptible to local
buckling.
__________________________________________________________________
Behaviour and Design of Cold-formed Steel Compression Members at Elevated Temperatures 2-1
from QUT library
Figure 2.1: Local Buckling of a Column

(http://911research.wtc7.net/mirrors/guardian2/fire/SCI-4-1.jpg)
2.1.2 Distortional Buckling
Distortional buckling (see Figure 2.2) is the rotation of flange at the flange-web
junction of the member with edge stiffeners and displacement of the intermediate
stiffeners normal to the plane of the element. This is a very common buckling mode
in short channel columns with thinner section.
Rotation of flange at flange-

web junction
Figure 2.2: Distortional Buckling of a Short Column (Ranawaka, 2006)
__________________________________________________________________
2.1.3 Global Buckling
There are three global Buckling has three types of failure modes. They are flexural,
torsional and flexural-torsional buckling modes. Global buckling mostly occurs in
long columns. Characteristics of these buckling modes are given next.
Flexural Buckling
Flexural buckling (see Figure 2.3) is the deflection caused by bending or flexure and
occurs about the axis with the largest slenderness ratio, (Le/r). This failure mode is
common in long columns.
Figure 2.3: Flexural Buckling of a Column (Gardner and Baddoo, 2006)
Torsional Buckling
This type of buckling only occurs in compression members that are doubly-
symmetric and have very slender cross sectional elements. It is caused by twisting of
the cross section about the longitudinal axis. Torsional buckling occurs mostly in
built-up sections.
__________________________________________________________________
Flexural-torsional Buckling
This type of buckling only occurs in compression members that have a cross section
with only one axis of symmetry. Flexural-torsional buckling is the simultaneous
bending and twisting of a member and mostly occurs in channels, structural tees,
double-angle shapes and equal-leg single angles.
Figure 2.4: Flexural and Flexural-torsional Buckling (Young, 2005)
2.2 Mechanical Properties of Cold-formed Steels at Ambient and

Elevated Temperatures
Mechanical properties of steels are characterised by yield strength, elastic modulus,

Poisson’s ratio, coefficient of thermal conductivity, specific heat, density, etc. Yield
strength and elastic modulus are the most important properties as they contribute to
the strength of the steel members. They play an important role in the finite element
analyses. Finite element analysis with accurate mechanical properties simulates real
life testing saving time and resources. Mechanical properties of hot-rolled steel are
well known and design standards accurately give the mechanical properties at
ambient and elevated temperatures. Use of cold-formed steel for structural members
has increased recently and therefore researchers focussed their attention on the
__________________________________________________________________
mechanical properties of cold-formed steel. Cold-formed steels have higher yield
strength compared to the hot-rolled steels. However, yield stress and elastic modulus
deteriorate with the increase of temperature. Researchers showed that the reduction
of mechanical properties of cold-formed steels is different from those of hot-rolled
steels.
2.2.1 Previous Research
Mechanical properties of steel are taken from the tensile and compression tests. As
steel is a homogeneous material, tensile property and compression properties are
similar for both cases. Therefore most researchers carried out only tensile tests to
find the mechanical properties. Methods of tensile tests are simple compared to the
compression tests because buckling of the specimen decreases the accuracy of the
results. From these tests, stress-strain curve is obtained. Stress-strain curve is then
used to find the yield stress and elastic modulus.
Makelainen and Miller (1983) carried out tensile tests of cold-formed galvanised
sheet steel Z33 at elevated temperatures using transient state and steady state test
methods. Steady state tests were carried out at a loading rate of 3.5 MPa/s and 0.35
MPa/s. They have observed that higher loading rates gave higher yield stress and
ultimate stresses. Reduction factors for elastic modulus obtained from steady state
tests were fitted to an equation as given in Equations 2.1.
ET T − 550
= −0.46 tanh S + 0.56 (2.1)
E20 o C 250
In the case of transient state tests, both reduction factors for yield stress and elastic
modulus were fitted to the equations as given in Equations 2.2 to 2.3.
σ y ,T
= 1.008 − 0.1314 exp(0.0047(Ts − 148.3o C )) 20 ≤ Ts ≤ 500 o C (2.2a)
σ y , 20 C
o
σ y ,T 104o C (1 − Ts 1135o C )
= Ts > 500 o C (2.2b)
σ y , 20 C
o Ts − 356o C
ET
E20 o C
( (
= 1.01 − 0.139 exp 0.007 Ts − 346o C )) (2.3)
__________________________________________________________________
Gerlich et al. (1996) also developed a polynomial, analysing the experimental works
of Klippstein (1980) to find the reduction factors of yield stress and elastic modulus
(see Equations 2.4 and 2.5).
Fy ,T 5.3T 4.0T 2 1.9T 3 1.7T 4

= 1− + − + (2.4)
Fy , 0 10 4 10 6 10 8 1011
Where,
FyT is the yield stress (MPa) at temperature T (oC)
Fy 0 is the yield stress (MPa) at room temperature (20oC)
T is the temperature of steel (oC)
ET 3.0T 3.7T 2 6.1T 3 5.4T 4

=1− 4 + − + (2.5)
E0 10 107 109 1011
Where,
ET is the modulus of elasticity (MPa) at temperature T (oC)
E0 is the modulus of elasticity (MPa) at room temperature (20oC)
T is the temperature of steel (oC)
Outinen et al. (2000) also carried out tensile tests of cold-formed steel. They
considered the following materials.
• S350GD+Z cold-rolled hot dip zinc coating structural steel (2 mm)
• S355 hot-rolled steel (4 mm)
• S460M high-strength steel specimen was made from 20 mm thick steel plate
• S355J2H specimen were taken from SHS of 3 mm thickness
Both transient and steady state tests were carried out that led to reduction factors for
yield stress and elastic modulus. The reduction factors for elastic modulus of
S350GD+Z steel are close to that of Eurocode 3 Part 1.2 (ECS, 2005) reduction
factors for carbon steel. Their reduction factors for yield strength did not follow that
of Eurocode 3 Part 1.2 (ECS, 2005). Since they did not agree with the code
predictions, they proposed yield stress reduction factors (see Table 2.1). Recent
version of Eurocode 3 Part 1.2 provides yield stress reduction factors for cold-
__________________________________________________________________
formed steels. However, they also do not agree with the proposed reduction factors
of Outinen et al. (2000).
Table 2.1: Mechanical Properties of Cold-formed Steel at Elevated

Temperatures (Outinen et al., 2000)
Steel temperature, Ea ,θ Ea Ea ,θ Ea f y ,θ f y f y ,θ f y
(oC) EC 3 :Part 1.2 Proposal EC 3 :Part 1.2 Proposal
20 1.0000 1.0000 1.0000 1.000
100 1.0000 1.0000 1.0000 0.970
200 0.9000 0.9000 1.0000 0.932
300 0.8000 0.8000 1.0000 0.895
400 0.7000 0.7000 1.0000 0.857
500 0.6000 0.6000 0.7800 0.619
600 0.3100 0.3100 0.4700 0.381
700 0.1300 0.1300 0.2300 0.143
800 0.0900 0.0900 0.1100 0.105
900 0.0675 0.0675 0.0600 0.067
1000 0.0450 0.0450 0.0400 0.029
Lee et al. (2003) also developed a set of equations to determine the reduction factors
for yield strength and elastic modulus of light gauge cold-formed steel at elevated
temperatures. Both low and high strength grades with the thicknesses of 0.4, 0.6, 1.0
and 1.2 mm were used. Temperatures from 100oC to 800oC at intervals of 100oC
were used with the steady state method. Equations 2.6 to 2.7 are for the mechanical
property reduction factors at all temperatures as derived by Lee et al. (2003).
Reduced yield strength at temperature T , f y ,T
f y ,T
= 1.0 20 o C ≤ T < 100 o C (2.6a)
f y , 20
f y ,T
= 0.964 + 0.00045T − 3.08 ⋅ 10 −6 T 2 + 1.969 ⋅ 10 −9 T 3
f y , 20
100 o C ≤ T ≤ 350 o C (2.6b)
__________________________________________________________________
f y ,T 0.0144 × T
= 1.514 − 1/ 5
400 o C ≤ T ≤ 750 o C (2.6c)
f y , 20 f y , 20 + 4.72
f y ,T
= 0.1 T = 800 o C (2.6d)
f y , 20
Reduced elasticity modulus at temperature T, ET

ET
= 1 .0 20 o C ≤ T ≤ 100 o C (2.7a)
E 20
ET
= 1 − 0.0014 (T − 100) 100 o C < T ≤ 500 o C (2.7b)
E 20
T
1−
ET 1200
= − 0.203 500 o C < T ≤ 800 o C (2.7c)
E 20 0.00122 T + 0.3
Outinen and Makelainen (2004) described the mechanical properties of cold-formed

steel at elevated temperatures and after cooling down following steady state and
transient state test methods. Steady state test results are similar to Outinen et al.
(2000). In the transient state test method, temperature and strain were measured and
then they were converted into stress-strain curves (see Figure 2.5). The results of
transient state test methods are given in Table 2.9. Further, Outinen and Makelainen
(2004) tested the mechanical properties of steel after the cooling of steel members
subjected to a fire. Thereafter, the results were compared with original mechanical
properties obtained at the ambient temperature. They observed that some mechanical
properties return to the nominal values (see Figure 2.6).
__________________________________________________________________
Figure 2.5: Converting Transient State Test Results to Stress-Strain Curves
(Outinen and Makelainen, 2004)
Table 2.2: Mechanical Properties Obtained from Transient State Tests

(Outinen and Makelainen, 2004)
Nominal yield
Measured yield
Steel grade strength, fy Material standard
2
strength, fy (N/mm2)
(N/mm )
S350GD+Z 350 402 SFS-EN 10 147
S355 355 406 SFS-EN 10 025
S460M 460* 445* SFS-EN 10 113
S355J2H 355 539-566 SFS-EN 10 219-1
__________________________________________________________________
Figure 2.6: Comparison of Mechanical Properties of Heated Steel Element
After Cooling (Outinen and Makelainen, 2004)
Chen and Young (2006b) investigated the effect on the mechanical properties of
corner region in cold-formed steel sections at elevated temperatures. They concluded
that the reduction factors of yield strength, elastic modulus, and ultimate strength of
corner regions are similar to that of flat regions. Further they concluded that
elongation is also similar, provided steel temperature is greater than 180oC. But
below 180oC, total elongation of corner part is smaller than that of flat region. Their
proposed equations for reduction factors for yield stress, elastic modulus, ultimate
strength and strain corresponding to ultimate strength are given in Table 2.3.
__________________________________________________________________
Table 2.3: Reduction Factors of Mechanical Properties of Cold-formed Steels 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
=a−
(T − b ) n
f y ,normal c
650 ≤ T < 1000 0.105 650 5.00x103 1
For elastic modulus,

22 ≤ T < 450 1.0 22 1.25x103 1
ET
=a−
(T − b)
n
450 ≤ T < 650 -0.11 860 -2.20x105 2
Tnormal c
For ultimate strength,

22 ≤ T < 450 1.0 22 5.6x108 3
f u ,T
=a−
(T − b )n 450 ≤ T < 1000 0.043 1000 -1.12x1011 4
f u ,normal c
For ultimate strain,

ε u ,T (T − b)
n
22 ≤ T < 1000 1.0 22 1.0x106 2
=a−
ε u , normal c
Ranawaka and Mahendran (2008) also developed a set of empirical equations to

determine the reduction factors of yield stress and elasticity modulus for
temperatures ranging from 200C to 8000C. These empirical equations were developed
using reduction factors obtained from 0.2% proof stress method. Two separate sets
of equations for yield strength reduction factors ( f y ,T f y , 20 ) have been developed for
high strength steels and for low strength steels based on various temperature ranges,
where f y ,T and f y , 20 are the 0.2% proof stress at elevated and ambient temperatures,
respectively. Further one set of equations has been proposed for reduction factors of
elastic modulus (ET E 20 ) for both high strength and low strength cold-formed steels
because the difference was insignificant. ET and E 20 are the elastic modulii at
__________________________________________________________________
elevated and ambient temperatures, respectively. Predictive equations for reduction
factors of mechanical properties are given next.
Reduction factors for yield strength at temperature T is given by,
For G550 steels

f y ,T
= −0.00016T + 1.0003 20 o C ≤ T ≤ 200 o C (2.8a)
f y , 20
f y ,T
= 0.97 −
(T − 200 )
1.81
200 o C < T < 600 o C (2.8b)
f y , 20 58500
f y ,T
= −0.00037T + 0.3363 600 o C ≤ T ≤ 800 o C (2.8c)
f y , 20
For G250 steels

f y ,T
= −0.0007T + 1.014 20 o C ≤ T ≤ 200 o C (2.9a)
f y , 20
f y ,T
= 3 .7 −
(T − 74)0.15 200 o C < T ≤ 800 o C (2.9b)
f y , 20 0.736
Reduction factors for elastic modulus at temperature T is given by,

ET
=1 20 o C ≤ T ≤ 100 o C (2.10a)
E 20
ET
= −0.0013T + 1.1297 100 o C < T ≤ 800 o C (2.10b)
E 20
Dolamune Kankanamge (2009) also carried out tensile tests for cold-formed steel at
ambient and elevated temperatures and developed a set of formulae for the
mechanical property reduction factors based on the tensile coupon tests carried out
on G450-1.90 mm, G450-1.50 mm, G250-1.50 mm and G250-1.95 mm cold-formed
steels.
__________________________________________________________________
Yield strength reduction factors for G450-1.9 mm and G450-1.50 mm thick
cold-formed steels,
f y ,T
= 1−
(T − 20)
4.56
20 ≤ T < 300 o C (2.11a)
10
f y , 20 1x10 T
f y ,T
= 0.95 −
(T − 300)1.45 300 ≤ T < 600 o C (2.11b)
f y , 20 7.76T
f y ,T
= −0.0004T + 0.35 600 ≤ T ≤ 800 o C (2.11c)
f y , 20
Yield strength reduction factors for G250-1.95 mm and G250-1.50 mm thick

cold-formed steels,
f y ,T
= −0.0005T + 1.01 20 ≤ T ≤ 200 o C (2.12a)
f y , 20
f y ,T
f y , 20
(
= 25 1.16 − T 0.022 ) 200 < T ≤ 800 o C (2.12b)
Elasticity modulus reduction factors for G450-1.9 mm, G450-1.50 mm, G250-
1.95 mm and G250 -1.50 mm thick cold-formed steels,
ET
= −0.000833T + 1.0167 20 ≤ T ≤ 200 o C (2.13a)
E 20
ET
= −0.00135T + 1.1201 200 < T ≤ 800 o C (2.13b)
E 20
The stress-strain relationship of steels is given by Ramberg Osgood stress-strain

model (see Equation 2.14). Based on the tensile tests carried out on both low and
high strength steels up to 0.95 mm thickness at ambient and elevated temperatures,
Ranawaka and Mahendran (2008) proposed Equations 2.15 (a) and (b) for the
parameterηT . They proposed 0.86 for β for both high and low strength steels.
Based on the tensile coupon tests carried out on G450-1.90 mm, G450-1.50 mm,
G250-1.5 mm and G250-1.95 mm cold-formed steels, Dolamune Kankanamge
(2009) proposed 1.5 for β for low strength steels and 0.86 for high strength steels.
__________________________________________________________________
ηT
f f y ,T fT
εT = T + β (2.14)
ET ET f y ,T
For G550 steel

η T = −3.05 X 10 −7 T 3 + 0.0005T 2 − 0.2615T + 62.653 20 o C ≤ T ≤ 800 o C (2.15a)
For G250 steel
η T = 0.000138T 2 − 0.085468T + 19.212 350 o C ≤ T ≤ 800 o C (2.15b)
2.2.2 Test Methods
There are three types of test methods available to investigate the mechanical
properties of steel at elevated temperatures. They are 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 load is applied. Load can be either
tension or compression. The differences between the mechanical properties from
tensile coupon tests and compression coupon tests are small (Ranawaka, 2006).
Therefore most researchers used tensile load as it is easy to apply and there is no
buckling effect in the specimens. Tensile load can either be applied 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. This
method directly produces the stress-strain curve. Therefore most researchers prefer
steady state tests.
In the transient state and ISO tests, the specimen is kept under constant load and
temperature is increased according to a fire curve. 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 Makelainen, 2004,
Chen and Young, 2007, Lee et al., 2003).
2.2.2.1 Specimens
Australian standard AS 1391 (SA, 1991) gives recommendations for the size and
shape of the specimen. Typical tensile coupon consists of gripping parts on both
sides and gauge length at the middle. The width of the gauge length is smaller than
__________________________________________________________________
the gripping part to ensure the failure will occur in the gauge length. A curvature is
defined to smooth transform from gauge width to the gripping width. As shown in
Figure 2.7 two holes are provided to fix the specimens to the loading shafts. A strain
gauge is attached at the middle of gauge length for the ambient temperature tests. In
the case of rolled-formed sections, mechanical properties vary across the section.
Therefore specimens are taken from different places.
Figure 2.7: Tensile Test Coupon from Dolamune Kankanamge (2009)
2.2.2.2 Test Rigs and Test Methods used by Previous Researchers
Outinen and Makelainen (2004) used a smaller oven to heat up the specimens. Strain
was measured using a special high temperature strain gauge. The tips of the ceramic
strain measuring rod of the strain measuring device were placed against the
measuring points in the test specimens (see Figure 2.8).
Figure 2.8: Test Setup Used by Outinen and Makelainen (2004)
__________________________________________________________________
Lee (2003) used an electric furnace to carryout tensile coupon tests at elevated
temperatures (see Figure 2.9). Lee (2004) used mechanical devices to measure the
strain at elevated temperatures. He used an extensometer made of metal rod to
measure the strain. The method of measuring the strain by contact method induces
errors due to the thermal expansion of the extensometer. In addition to this error, the
opening provided to insert the arms of extensometer causes heat lost and hence it
was difficult to maintain uniform temperature inside the furnace.
By eliminating the difficulties of measuring the strain using contact method at

elevated temperatures, Ranawaka and Mahendran (2009) used a laser speckle
extensometer to measure the strain (see Figure 2.10). This method used laser image
to measure the strain. This method is contact free and therefore accuracy of the strain
measurements was higher. Ranawaka and Mahendran (2009) also used an electric
furnace to heat up the specimen.
Figure 2.9: Test Setup Used by Lee (2004)
__________________________________________________________________
(a) Furnace and loading setup (b) Laser Speckle Extensometer
Figure 2.10: Test setup used by Ranawaka (2006)
2.2.3 Code Provisions
Mechanical and thermo-physical properties of hot-rolled steel vary with the

temperature. Mechanical properties i.e. yield strength, elastic modulus, coefficient of
thermal expansion etc. do not vary linearly and their variation at elevated
temperature is complex. Usually mechanical properties of steel at elevated
temperatures are described in terms of reduction factors of the ambient temperature
properties. Poisson’s ratio and density of steel are assumed to be constant at any
temperature (Ranby, 1999 and Outinen and Myllymaki, 1995).
Eurocode 3 Part 1.2 (ECS, 2005) provides yield strength reduction factors for Class
4 sections. But there is no difference in yield strength reduction factors between hot-
rolled and cold-formed steels. BS 5950: Part 8 (BSI, 1990) also provides yield
strength reduction factors for cold-formed steels (see Table 2.4). However, yield
__________________________________________________________________
strength reduction factors are given at 0.5%, 1.5% and 2.0% strain levels even
though the 0.2% proof stress is commonly used. No reduction factors are given for
elastic modulus.
Table 2.4: Yield strength Reduction Factors for Cold-formed Steel from
BS 5950 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 0.949 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
2.3 Local Buckling Behaviour
The major advantage of cold-formed steel members over the other structural
materials (concrete, timber etc.) is the higher strength. Cold-formed steel members
are often made with thin sheets. Thinness of steel sections causes them to fail by
local buckling unless the plate elements are compact. This issue is more critical with
cold-formed steel because its strength is higher than the conventional hot-rolled
steel. Therefore cold-formed steel sections are often made with thin steel sheets.
Hence, cold-formed steel sections are more prone to local buckling.
2.3.1 Local Buckling Behaviour at Ambient Temperature
2.3.1.1 Previous Research
The stress pattern of a slender plate element subjected to local buckling is not
uniform. It follows a parabolic stress pattern. Therefore local buckling capacity
needs rigorous calculations. To overcome this issue, Von Karman (1932) introduces
a concept called effective width method. This method assumes uniform stress
distribution over the effective width instead of nonlinear stress distribution over the
entire element. Figure 2.11 illustrates the concept of effective width.
__________________________________________________________________
Actual Width Effective Width
(a) Effective Width of Stiffened Element
Actual Width Effective Width

(b) Effective Width of Unstiffened Element
Figure 2.11: Effective Width Concept from AS/NZS 4600 (SA, 2005)
Von Karman (1932) proposed a simple equation as given in Equation 2.16 to find the
effective width of plate element based on the design stress and elastic critical local
buckling stress.
be f cr
= (2.16)
b fy
Where,
f y is the design stress
kπ 2 E
f cr is the elastic local buckling stress given by
12(1 − υ 2 )(b / t ) 2
Winter (1947) modified Von Karman’s formula based on experimental
investigations. Winter proposed Equations 2.17 (a) and (b) to find the effective width
of stiffened elements based on the plate slenderness ( ).
be
=1 0.673 (2.17a)
b
be (1 − 0.22 / λ )
= ≤ 1 .0 > 0.673 (2.17b)
b λ
Where,
__________________________________________________________________
f max
λ=
f cr and f max is the maximum design stress in the compression element.
Winter (1947) proposed Equation 2.18 to find the effective width of unstiffened
element based on test results. Winter’s equation for unstiffened elements is as
follows.
be 1 1
= 1.19 1.0 − 0.298 (2.18)
b λ λ
Most of current design standards use Equations 2.17 (a) and (b) to find the effective
width of stiffened and unstiffened elements.
Popovic et al. (1999) reported compression tests on stub columns made from cold-
formed steel angle columns. Tests were carried out with short and long columns of
2.4, 3.8 and 4.7 mm in thickness. All angle sections were 50 x 50 mm equal angles.
Their results for stub columns showed that the predicted capacities from AS 4100
and AS 4600 (also similar to AISI) are conservative. Conservatism increases with
the slenderness of the section. Figure 2.12 shows the angle sections used by Popovic
et al. (2001).
Figure 2.12: Locally Buckled Columns from Popovic et al. (2001)
Dhanalakshmi and Shanmugam (2001) conducted local buckling tests on cold-

formed steel angle stub columns. Two types of angles, namely perforated and non-
perforated were used. Finite element analyses and a parametric study were carried
out for these two types of sections. Figure 2.13 shows the tested specimens and finite
element models for simulation of local buckling. Tests on non-perforated columns
__________________________________________________________________
showed that load eccentricity (away from the corner) reduces the capacity. Further,
when slenderness increases, eccentricity has less influence on the ultimate load
capacity.
Figure 2.13: Tested Specimens and Finite Element Models from Dhanalakshmi
and Shanmugam (2001)
Some researchers carried out parametric study to gain a proper understanding of

local buckling behaviour. The major advantage of parametric study is the possibility
of understanding the effects of different parameters. In the case of pure local
buckling, axial strength should be the same as the section capacity. Therefore
parametric study can be carried out with minimum length of model that represents
the local buckling wave. By considering the symmetry, Lee (2004) used the quarter
wave model for the parametric study. Dubina and Ungureanu (2002) used three half
waves for their simulation. Some researchers consider the width of plate element as
the variable and keep the model length same as in the experimental model. The
accuracy of model length for the local buckling simulation is not fully clear.
2.3.1.2 Test Methods
Test Rigs
Due to high strength to weight ratio, the effect of self weight is insignificant.
Therefore steel columns can also be tested horizontally without any significant
errors. The major advantage of test rig, which can test the columns horizontally, is
the ability to test short and long columns. Vertical test rigs can also test long
columns. However, head room and stiffness of the test rig are the disadvantages of
__________________________________________________________________
the tall test rigs. Popovic et al. (2000) used a horizontal test rig for testing short and
long columns (see Figure 2.14). This test rig was fitted with special arrangements to
measure the rotations and the out of plane displacements.
Figure 2.14: Test Rig Used by Popovic et al. (2000) for Compression Tests
Dhanalakshmi and Shanmugam (2001) used an INSTRON servo-controlled

universal testing machine (see Figure 2.15). Displacement transducers were attached
to measure the displacements. Stub columns were rigidly fixed to the loading arm of
the machine through plates welded to the ends of the column.
Figure 2.15: INSTRON Servo-controlled Universal Testing Machine Used by

Dhanalakshmi and Shanmugam (2001)
__________________________________________________________________
Lee (2004) and Ranawaka (2006) used the Tinius Olsen Machine to test short
columns (see Figure 2.16). This test rig can be used for short columns vertically.
Linear variable displacement transducers were fitted to the specimen and the cross-
head of the machine to measure out of plane and axial displacements.
Figure 2.16: Tinius Olsen Testing Machine used by Lee (2004) and
Ranawaka (2006)
End Supports
Some researchers used fixed-end supports while some other researchers used pinned-
end supports for column tests. Ranawaka (2006) used fixed ends for distortional
buckling. Fixed ends were achieved by fixing the specimens to a groove made on a
thicker steel plate (see Figure 2.17 ) and fixing using Pyrocrete 165 coil grout.
Dhanalakshmi and Shanmugam (2001) welded the column to steel plates to acheive
fixed-end support (see Figure 2.18). In the case of local buckling, end conditions do
not affect the ultimate load them as local buckling capacity is the section capacity.
Therefore Lee (2004) and Feng et al. (2003a) used thicker steel plates on either ends
of the column to load. These plates were simply kept on the test on the columns.
__________________________________________________________________
(a) Grove Cast on End Pate (b) Specimen Fixed to End Plate
Figure 2.17: Fixed-End Achieved by Ranawaka (2006) Using Pyrocrete 165

Coil Grout
Figure 2.18: Fixed-end Support used by Dhanalakshmi and Shanmugam (2001)
2.3.1.3 Design Rules
AS/NZS 4600 (SA, 2005), Eurocode 3: Part 1.3 (ECS, 1996), North American
Specification (NAS, 2003), BS 5950 Part 5 (BSI, 1998) and Direct Strength Method
(DSM) provide design guidelines for the local buckling of cold-formed steel
compression members. AS/NZS 4600 (SA, 2005), Eurocode 3: Part 1.3 (ECS, 1996)
North American Specification (NAS, 2003) use effective width method and the
design equations and limitations are the same. Even though BS 5950 Part 5 (BSI,
1998) also recommends the effective width method, the design equations and
limitations are slightly different from the other standards. Most of the design codes
recommend the effective width method for local buckling. Schafer and Pekoz (1998)
proposed direct strength method by considering the gross section instead of the
effective section. This method allows the interaction of local buckling with flexural
or flexural-torsional buckling. Details of the above design standards are given next.
__________________________________________________________________
AS/NZS 4600
AS/NZS 4600 (SA, 2005) recommends the effective width/area method for the
calculation of local buckling capacity. The effective width concept based on the
stress redistribution over a reduced plate width was first derived by von Karman
(1932). In this method, maximum uniform stress is applied to the effective width
instead of nonlinear stress distribution over the entire section (see Figure 2.19). The
effective width of plate elements is calculated using Equations 2.19 (a), 2.19(b) and
2.20 based on the plate slenderness ratio ( λ ).
Figure 2.19: Re-distribution of Stress and Effective Width Concept
For λ ≤ 0.673 : be = b (2.19a)
For λ > 0.673 : be = ρ b (2.19b)
Where,
0.22
1−
λ
Effective width factor (ρ) = ≤ 1.0 (2.20)
λ
Plate slenderness ratio is expressed in terms of design compressive stress ( f * ) and
plate elastic buckling stress ( f cr ).
f*
λ= (2.21)
f cr
Plate elastic buckling stress ( f cr ) is given by following equation,
__________________________________________________________________
2
kπ 2 E t
f cr = (2.22)
(
12 1 − υ 2 ) b
E = Young’s modulus of elasticity (200 x 103 MPa)

ν = Poisson’s ratio
t = thickness of the uniformly compressed stiffened elements
k = plate buckling coefficient
The buckling coefficient k is conservatively recommended as 4 for stiffened

elements and 0.43 for unstiffened elements in the design codes AS/NZS 4600 (SA,
2005), Eurocode 3: Part 1-3 (ECS, 1996) and North American Specification (NAS,
2003).
BS5950: Part 5
BS 5950 Part 5 (BSI, 1998) also recommends the use of effective width concept for
calculating the local buckling capacities of cold-formed steel members. However,
design equations are different from the design equations of AS/NZS 4600 (SA,
2005), Eurocode 3: Part 1-3 (ECS, 1996), North American Specification (NAS,
2003). Instead of plate slenderness ratio of AS/NZS 4600, British standard uses a
ratio between design compressive stress and elastic local buckling stress to find the
effective width of a plate element. Limiting value for the above ratio is different
from the AS/NZS 4600 design standard. Equations 2.23 (a) and (b) are used to find
the effective width of the cold-formed steel section.
beff
For f c / p cr ≤ 0.123 , =1 (2.23a)
b
For f c / p cr > 0.123 ,

beff
b
[ {
= 1 + 14 ( f c / p cr )
1/ 2
}]
− 0.35
4 −0.2
(2.23b)
Where,
b is the full flat width
fc is the compressive stress on the effective element and is equal
to the yield stress
__________________________________________________________________
p cr is the local buckling stress of the element given by:
2
t
p cr = 0.904 EK
b
Buckling coefficient K is 4 for the stiffened element which is same as the AS/NZS
4600 recommendation. However, K value for unstiffened element is slightly
conservative in the British standard compared to the AS/NZS 4600 and is equals to
0.425.
Direct Strength Method (DSM)

Schafer (2001) proposed Direct Strength Method considering the gross section
instead of the effective section. This method can be used to predict the capacity of
cold-formed steel column failure in local buckling interaction with other buckling
modes. The direct strength method gives the following set of equations to calculate
the local buckling capacity.
0.4 0.4
P Pcrl P
Pn = 1 − 0.15 crl P for > 0.776 , (2.24a)
P P Pcrl
Else Pn = P (2.24b)
Where,
Pn = Local buckling capacity of the section.
Pcrl = Elastic critical buckling load of the section using rational buckling
analyses. Most researchers use Thin-wall, CUFSM or ABAQUS softwares to

find the elastic buckling load.
P = Squash load except when interaction with other buckling modes is
considered.
2.3.2 Local Buckling Behaviour at Elevated Temperatures
To date only limited experimental research was carried out on the local buckling
behaviour of cold-formed steel member at elevated temperatures. Most of them were
finite element simulations. However, some research were carried out on hot-rolled
__________________________________________________________________
steel members subjected to local and global buckling failures. Therefore this
literature review was not limited to the research on cold-formed steel member.
Limited research was carried on the local buckling of cold-formed steel columns
subjected to fire (Outinen and Myllymaki 1995, Ranby 1999, Feng et al. 2003a,
2003b and Lee 2004). Some researchers carried out their research with non-uniform
temperatures (Kaitila 2002, Feng et al, 2003c). This situation arises in wall panels
where columns were protected with lining boards.
Outinen and Myllymaki (1995) tested cold-formed steel short columns of hollow
sections. Columns were loaded concentrically and eccentrically. They were
rectangular hollow sections (RHS 200x200x5 and RHS 150x100x3) with a length of
900 mm. The top of the columns was free to rotate about the main axis and free to
expand axially. The bottom of the columns was connected to a plate and rested on
the pack of steel and hence gave rotational restraint (see Figure 2.20). Transient state
test method was adopted in their testing. Temperature was increased from 20oC to
300oC in three minutes and subsequently by 10oC per minute. After comparing the
test results with numerical calculations, Outinen and Myllymaki (1995) concluded
that yield strength corresponding to the total strain of 2% leads to higher prediction
of strength while 0.2% proof stress leads to safe results at elevated temperatures.
They commented that the limiting temperature of 350oC for Class 4 section is too
conservative. Tensile coupon tests were also carried out at ambient temperature and
then the Eurocode 3 Part 1.2 mechanical property reduction factors were used to
predict the mechanical properties at elevated temperatures in their numerical
analysis. Effective widths of plate elements at elevated temperatures were
determined according to the design guidelines given in Eurocode 3 Part 1.3 while
using reduced mechanical properties at elevated temperatures. Many researchers
(Lee, 2004, Ranawaka and Mahendran, 2009, Dolamune kankanamge, 2009) showed
that the reduction factors in the Eurocode 3 Part 1.2 are not accurate enough,
resulting incorrect modelling results.
__________________________________________________________________
(a) End Condition of Specimen (b) Failed Specimen
Figure 2.20: Tests of Outinen and Myllymaki (1995)
Ranby (1999) carried out numerical simulations of local and global buckling. He
concluded that the ambient temperature design guidelines for cold-formed steel
structures from Eurocode 3 Part 1.3 can be used provided yield stress is taken as the
0.2% proof stress. Further, he concluded that the limitation for the maximum
temperature of 350oC is conservative. However, these conclusions are mostly based
on numerical simulations. Therefore it needs an extensive test data to verify this
conclusion.
A series of fire tests and numerical simulations on cold-formed steel short columns
were carried out by Feng et al. (2003a). Tests were carried out on lipped and
unlipped columns with thicknesses of 1.2 mm and 2.0 mm. Some of lipped channels
were provided with a service hole at the web. Most of the columns failed in local
buckling while some of the lipped channel columns failed in distortional buckling.
They observed that most of the columns failed in local buckling for the test
temperatures less than or equal to 400oC. At temperatures greater than or equal to
550oC, failure mode was predominantly distortional buckling. As expected, service
holes reduced the axial strength. However, at 550oC, service holes enhanced the
axial load carrying capacity for both thicknesses and the axial load carrying capacity
was greater than that of columns without a service hole for the thickness of 2.0 mm.
In addition to enhancing the capacity, the buckling mode also changed. Columns
with service holes failed in distortional buckling while columns without a service
hole failed in local buckling. However, this scenario is not properly clarified yet.
Feng et al. (2003b) numerically simulated the test carried out by Feng et al. (2003a)
and concluded that the ambient temperature design guidelines can be extended for
__________________________________________________________________
elevated temperature design provided 0.2% proof stress is used. However, they used
reduced mechanical properties from Eurocode 3 Part 1.2 and Outinen et al. (2000).
Recent researchers proposed mechanical property reduction factors which do not
fully agree with the above two predictions. Therefore further research is needed to
validate the use of ambient temperature design guidelines for fire design.
Lee (2004) also carried out local buckling tests of cold-formed steel members at
ambient and elevated temperatures as well as compression tests of the lipped and un-
lipped channels of low and high strength grades cold-formed steel short columns.
Specimens were selected to fail only by local buckling. Lee (2004) concluded that
the simplified method of using the reduced mechanical properties can be used for
predicting the local buckling capacities at elevated temperatures in the case of low
strength grades. But this can not be used for high strength grades. However,
experimental setup had some problems in the temperature measurement and hence
the accuracy of the results is not known.
Yang et al. (2006a) conducted compression tests of hot-rolled steel columns at

elevated temperatures using the steady state test method, with both local and global
buckling tests. Local buckling tests were carried out with 300x300x10x15 and
175x175x7.5x11 universal columns. The first section showed local buckling at
ambient and elevated temperatures. But the buckling mode of the second section
changes the buckling modes at elevated temperatures from global buckling to local
buckling (see Figure 2.21). They observed a rapid reduction of strength after 500oC.
Up to 500oC, hot-rolled steel columns retained 70% of their strength at ambient
temperature and thereafter, decreased rapidly. However, the changing of buckling
modes at elevated temperatures is not well understood.
__________________________________________________________________
(a) 300x300x10x15 Sections (b) 175x175x7.5x11 Sections
Figure 2.21: Local Buckling Failure at Elevated Temperatures

(Yang et al., 2006a)
Unlike ambient temperature tests, elevated temperature tests are difficult to carry out
because loading should be done while heating. Most researchers loaded the columns
inside a furnace. Some researchers attached heating element to the column and the
column was covered by insulation blanket. In the case of column test inside a
furnace, loading shafts should be made from a metal that has good fire resistance and
higher stiffness at elevated temperatures. Stainless steel has good fire resistance.
Therefore most researchers used stainless steel for loading shafts. Some researchers
used concrete filled steel tubes for loading shafts.
Lee (2004) and Ranawaka (2006) used an electrical furnace for their tests at elevated
temperatures (see Figure 2.22). Four glow bars fitted to the furnace heated the
specimens while a micro computer based temperature control system controlled the
pre-set temperatures and heating rates. A maximum heating rate of up to 30 oC/min
can be achieved with this furnace. Loading shafts were made of stainless steel. These
loading shafts were connected to the Tinius Olsen Testing machine and the load was
applied to the specimen via the loading shafts.
__________________________________________________________________
Extended
loading shaft
Fixed-end
conditions
Thermocouples
LVDT
Specimen
Four glow bars
Figure 2.22: Electric Furnace used by Lee (2004) and Ranawaka (2006)
Feng et al. (2003a) used an electric furnace for local buckling tests (see Figure 2.23).
Loading shafts were made of concrete filled tubular sections. A hydraulic jack was
used to apply the loading and a load cell fixed underneath the hydraulic jack was
used to measure the axial load.
__________________________________________________________________
Figure 2.23: Sketch of the Test Rig Used by Feng et al. (2003a)
None of the design codes provide specific design guidelines for local buckling of
cold-formed steel members at elevated temperatures. Eurocode 3 Part 1.2 (ECS,
2005) and BS 5950 Part 8 (BSI, 1990) provide guidelines for the design of steel
members at elevated temperatures. These guidelines are mainly focussed on the hot-
rolled steel members at elevated temperatures. In addition to the above standards, AS
4100 (SA, 1998) also gives guidelines for the fire design of steel members. But these
guidelines are not in detail like European or British standards. However, Eurocode 3
Part 1.2 recommends the use of Eurocode 3 Part 1.3 recommendations with reduced
mechanical properties and ambient temperature effective area.
__________________________________________________________________
2.4 Global Buckling Behaviour
Cold-formed steel long columns are often made with open sections. The long
columns of open sections are more prone to flexural and flexural-torsional buckling.
When increasing the length of a cold-formed steel column, it will first experience
local buckling and then distortional buckling. If it further increases the length, in
most cases, flexural buckling occurred, followed by flexural-torsional buckling. But
the column capacity that subject to flexural buckling is significantly less than the
section capacity. Therefore designers limit the length of cold-formed steel columns.
Flexural-torsional buckling is the most common failure mode of long columns. Use
of cold-formed steel for structural members has increased in recent past. Therefore
researchers focussed their attention on the behaviour of cold-formed steel. Cold-
formed steels are used as medium scale load bearing members such as columns in
low-rise buildings. Since fire is a threat to low-rise buildings, recent researchers
focussed their attention on the behaviour of cold-formed steel columns under fire.
Design standards also do not provide clear guidelines for the fire design of cold-
formed steel compression members. Therefore a detail review was carried out on the
behaviour of long cold-formed steel columns.
2.4.1 Global Buckling Behaviour of Cold-formed Steel Members at

Ambient Temperature
Young and Rasmussen (1998a) tested cold-formed steel plain channel columns to
assess the accuracy of design standards. Fixed-end condition was used in their tests
with base metal thickness of 1.5 mm. The column height was varied from 280 mm to
3500 mm covering the full range of slenderness. They concluded that AS/NZS 4600
(SA, 1996) and American Iron and Steel Institute (AISI, 1996) specifications slightly
over-estimate the design strength. Further, they concluded that the failure modes
predicted by the above specifications agree with the columns having long effective
lengths, but not for shorter effective lengths. They observed that local buckling does
not induce overall bending to fixed–end columns. They recommended that the fixed
__________________________________________________________________
end columns can be designed assuming an effective length equal to half the column
length. Young and Rasmussen (1998b) tested lipped channels. They concluded that
AS/NZS 4600 (SA, 1996) and American Iron and Steel Institute (AISI, 1996)
specifications conservatively predict the buckling capacity.
Young and Rasmussen (1999) experimentally investigated the effect of local

buckling on the behaviour of cold-formed steel columns with fixed and pin-end
conditions at ambient temperature. Both lipped and unlipped channels were tested.
Effective length of fixed-end columns was considered as half of the column height.
Ultimate loads of pin-end columns are less than that of fixed-end columns having the
same effective length. Hence they concluded that local buckling does not induce
overall bending to the fixed-end singly symmetric columns. However, local buckling
induced overall bending to the pin-end columns.
Popovic et al. (1999) tested cold-formed steel angle columns at ambient temperature
with fixed-end and pin-end supports. Their results were compared with AS/NZS
4600 (SA, 1996) and AS 4100 (SA, 1998) and it was concluded that prediction of
stub column strength by the above codes are conservative. Conservatism increased
with slenderness of the section. AS/NZS 4600 (SA, 1996) recommends the design of
concentrically loaded angles by considering an additional bending moment about
minor axis due to an eccentricity of L/1000. But this was found to be more
conservative (Popovic et al., 1999 and Young, 2005) and there was no need to
consider additional moment about minor axis. Young (2004b) also tested cold-
formed steel plane (unstiffened) angle columns at ambient temperature. Specimen
size was 70x70 mm equal angles with 1.2 mm, 1.5 mm and 1.9 mm thicknesses.
Fixed-end condition was adopted in the testing. Specimen lengths were varied from
250 mm to 3500 mm. Young (2004b) also considered the results of Popovic et al.
(1999). Results were compared with Australian/ New Zealand standards AS/NZS
4600 (SA, 1996) and North American Standards NAS (2001). As given in Equations
2.25a and b, Young (2004b) proposed a modification for the current design rules of
AS/NZS 4600 (SA, 1996) for the design of plane angle columns. This indicates that
the available design rules need modifications based on the geometry of sections.
__________________________________________________________________
( )
f n = 0.5λ2c f y λ c ≤ 1 .4 (2.25a)
f n = 0.5 fy λ c > 1 .4 (2.25b)

λc2
Test Rigs
Cold-formed steel long columns were tested horizontally as well as vertically. Most
of the test rigs fitted with a moveable cross-head to apply axial load with a stationary
base. Test rigs are almost similar to those used for local buckling tests. The only
difference is that they can accommodate long columns. Figures 2.24 and 2.25 show
the test rigs used by Yang et al. (2002) and Liu and Young (2003), respectively.
Figure 2.24: Test Rig used by Yang et al. (2002)
__________________________________________________________________
Figure 2.25: Test Rig used by Liu and Young (2003)
End Supports
Many researchers used fixed-end conditions (Young and Rasmussen, 1998a and b,
Liu and Young, 2003, Yan and Young, 2002, Young and Ellobody, 2000) while
some other researchers used pin-end conditions (Yang et al., 2002). The major
advantage of pin-end is that global buckling failure can be achieved with shorter
__________________________________________________________________
columns. Figure 2.24 shows a typical pin-end. Methods of achieving fixed-end
support are similar to that explained under local buckling (Section 2.3.2.1).
Global buckling has three major modes. They are flexural, torsional and flexural-
torsional buckling. All these buckling modes depend on the geometry of sections,
end conditions and member lengths.
AS/NZS 4600 (SA, 2005)
AS/NZS 4600 (SA, 2005) gives the design guidelines for flexural or flexural-
torsional buckling capacity calculations based on the non-dimensional slenderness,
λc , of the column as given in the following Equations
N c = Ae f n (2.26)
( 2
f n = 0.658 λc f y ) λ c ≤ 1 .5 (2.27a)
f n = 0.877 fy λ c > 1 .5 (2.27b)

λc2
Where,
λc = non-dimensional slenderness used to determine f n
fy
λc = (2.28)
f oc
f oc = least of the elastic flexural buckling stress ( f ox ) from Equation 2.29
and torsional and flexural-torsional buckling stress ( f oxz ) from Equations
2.30 (a) and (b).

π 2E
f ox = (2.29)
(le r )2
f oxz =
1
2β
( f ox + f oz ) − (( f ox + f oz ) − 4 β f ox f oz
2
) (2.30a)
Alternatively f oxz = f oz f ox ( f oz + f ox ) (2.30b)
Where, l e = effective length of member
__________________________________________________________________
r = radius of gyration of the full, unreduced cross-section
f oz = Elastic torsional buckling stress
β = 1 − (xo rol )2
rol = Polar radius of gyration of the cross section about the shear
centre
xo = Distance between shear centre and geometric centre
BS 5950 Part 5 (BSI, 1998)
BS 5950 Part 5 (BSI, 1998) provides Equation 2.31 to calculate the flexural buckling
capacity. In the case of flexural-torsional buckling, PE in Equation 2.31 needs to be
replaced with the minimum value of elastic buckling load, PE about the axis of
symmetry and the elastic flexural torsional buckling load, PTF from Equation2.33.
BS 5950 Part 5 uses an enhanced elastic torsional buckling load (PT) than AS/NZS
4600 (SA, 2005)
PE Pcs
Pc = (2.31)
φ + φ 2 − PE Pcs
Where,
Pcs + (1 + η )PE
φ= (2.32a)
2
PCS = Aeff Py (2.32b)
Aeff is the effective cross sectional area,
p y is the design strength,
E is the modulus of elasticity,

I is the second moment of area of the cross-section about the critical axis,
Le is the effective length of the member about the critical axis,
For Le r ≤ 20, η =0 (2.32c)
For Le r > 20, η = 0.002(LE r − 20 ) (2.32d)
Where,
r is the radius of gyration of the gross cross-section corresponding to PE
__________________________________________________________________
Where,
PTF =
1
2β
[ {
(PEx + PT ) − (PEx + PT )2 − 4βPEx PT }
1/ 2
] (2.33)
PEx is the elastic flexural buckling load for a column about the x axis given
π 2 EI x
by ,
L2E
PT is the torsional buckling load of a column given by,
1 π 2 EC w
PT = GJ + 2 (2.34)
ro2 L2E
2
x
β is a constant given by β = 1 − o ,
ro
ro is the polar radius of gyration about the shear centre given by:
G is the shear modulus,

xo is the distance from the shear centre to the centroid measured along the x
axis,
J is the St Venant torsion constant which may be taken as the summation of
bt 3 3 for all elements, where b is the element flat width and t is the
thickness,
I x is the second moment of area about the x axis,
C w is the warping constant for the cross section.
Eurocode 3 Part 1.3 (ECS, 1996)
Eurocode 3 Part 1.3 (ECS, 1996) provides Equation 2.35 for both flexural and
flexural-torsional buckling resistance (N b , Rd ) . However, relative slenderness (λ )

_
changes depending on the failure type. Buckling resistance is defined as a function of

gross area (Ag ) , effective area (Aeff ) , yield stress ( f y ), buckling coefficient (χ ) and
partial safety factor (γ M 1 ) .
N b , Rd = χAeff f y γ M 1 ≅ χβ A Ag f y γ M 1 (2.35)
__________________________________________________________________
Where,
β A is a reduction factor given by: β A = (Aeff Ag ) ,
1
χ= 0.5
but χ ≤ 1.0 (2.36a)
_
φ + φ2 −λ 2
_ _
and φ = 0 .5 1 + α λ − 0 .2 + λ 2 (2.36b)
Where,
α is an imperfection factor depending on the appropriate buckling
curve
For flexural buckling, relative slenderness is defined as follows;
[ ( )]
_
λ = (l i ) π E f y βA (2.36c)
Where,
l is the buckling length for flexural buckling about the relevant axis
i is the radius of gyration about the corresponding axis of the gross cross-
section
For flexural-torsional buckling, relative slenderness is defined as follows;
λ = ( f yb σ cr ) β A
_
(2.36d)
with: σ cr = σ cr ,TF but σ cr ≤ σ cr ,T

Where,
σ cr ,TF is the elastic critical stress for flexural-torsional buckling,
σ cr ,T is the elastic critical stress for torsional buckling,

For cross-sections that are symmetrical about the major axis, the elastic critical stress
σ cr,TF for flexural-torsional buckling is defined as follows,
σ cr ,TF =
1
(σ cr , y + σ cr ,T ) − (σ cr , y + σ cr ,T ) − 4 βσ cr , y σ cr ,T
2
(2.36e)
2β
with: σ cr , y = π 2 E (l y i y )2 (2.36f)
β = 1 − ( y o i o )2 (2.36g)
Where,
σ cr ,T is the elastic torsional buckling stress,
__________________________________________________________________
io2 is the polar radius gyration,
y o is the distance from the shear centre to the centroid measured along the x
axis.
2.4.2 Global Buckling Behaviour at Elevated Temperatures
Only limited research can be found in literature on the global buckling behaviour of
cold-formed steel compression members at elevated temperatures. Therefore the
literature review focused on research carried out on hot-rolled steel members to have
a sound knowledge on testing methods and behaviour at elevated temperatures. A
large number of research projects were carried out on hot-rolled steel compression
members at elevated temperatures for buckling failure during the 1990s. Limited
research could be found during the 1980s or after 2000. Most of them focussed on
the global buckling of hot-rolled steel members.
Aasen (1985) tested long hot-rolled steel columns at elevated temperatures. A total
of 20 universal columns with four lengths (3100, 2210, 1750 and 1700 mm) were
tested with transient method. The columns were tested with three different support
conditions: pin-end, rotationally restrained and axially restrained end conditions. End
rotational restraint was given by connecting the ends of column to a restraining
beam. Heating rate was 20oC/minute for most tests and 10oC/minute for few other
tests. Most columns failed in minor axis global buckling while two columns failed in
flexural-torsional buckling. Aasen (1985) concluded the column strength of axially
restrained columns is determined by preheated load level and geometrical
imperfections. The columns with axial restraint failed suddenly while others failed
gradually. The heating rate slightly influenced the maximum column strength.
Rodrigues et al. (2000) experimentally investigated the critical temperature of

compressed steel members with restrained thermal elongation. All specimens were
__________________________________________________________________
rectangular with the thickness varying from 5 mm to 20 mm while the width of the
section was 50 mm. Different values of restraint and eccentricity were considered in
their testing. They experienced a considerable temperature gradient along the
specimen reaching a maximum temperature at the middle of the specimen and lowest
temperatures at either ends. This gradient increased with the increasing of
temperature. They concluded that restraint to thermal elongation of centrally loaded
columns having a slenderness ratio greater than 80 can reduce its critical temperature
by up to 200oC. Further, for an eccentrically loaded column with restraint to thermal
elongation, the magnitude of eccentricity did not affect their critical temperature
significantly. Wang (1997) also concluded that restraint to thermal elongation of a
steel column reduced the limiting temperature.
Ali and O’Conner (2001) investigated the effects of rotational restraint on the failure
of columns subjected to different loading levels (0%, 20%, 40%, 60% and 80%). All
the columns were subjected to an axial restraint degree of approximately 0.29
(imposed axial stiffness approximately 57 kN/mm). Two levels of rotational restraint
0.18 and 0.93 were imposed in their testing. Rotational restraint was provided by
fixing steel plates with different thicknesses to the column ends as shown in Figure
2.26. The degree of rotational restraint is defined by Equation 2.37.
ρs
ρ= (2.37)
ρs + ρc
ρ − degree of rotational restraint.
ρ s − rotational stiffness of the surrounding (to column) structures.
ρ c − rotational stiffness of the column.
They concluded that the addition of rotational restraint had a relatively minor effect
on the value of generated restraint forces but failure temperatures were greatly
increased under the same loading level (see Figures 2.13a,b and Table 2.2). Also,
they observed that changing the value of applied rotational restraint has an
insignificant effect on the value of generated restraint forces. Further, they observed
that the increase in loading level caused a significant drop in failure temperature.
Furthermore, a sudden failure pattern was observed with few cases when the applied
load was significantly higher than the generated forces.
__________________________________________________________________
Figure 2.26: End Conditions of Specimens used by Ali and O’Connor (2001)
(a) Axial Force Generated Under High Rotational Restraint
Figure 2.27: Axial Force Generated in Columns at Elevated Temperatures

(Ali and O’Connor, 2001)
__________________________________________________________________
(b) Axial Force Generated Under Low Rotational Restraint
Figure 2.27: Axial Force Generated in Columns at Elevated Temperatures (Ali

and O’Connor, 2001)
Table 2.5: Experimental Results of Ali and O’Connor, 2001
Wang and Davies (2003a) tested the rotationally restrained columns subjected to fire.
The main objectives of this research were to investigate the change of bending
moment in restrained columns with the temperature and how these changes affect the
columns’ failure temperature. Wang and Davies (2003b) analysed the results and
concluded that the failure temperature or failure load calculated using both BS 5950
Part 8 and Eurocode 3 Part 1.2 follow better agreement with the test results, when
column bending moments are ignored and column effective length factor is taken as
0.7 for one end fixed and the other end pinned column. Ali (2001) also
experimentally investigated the effective length of steel columns subjected to fire
__________________________________________________________________
using the geometry of cooled columns after a fire. However, geometry may change
during the cooling stage and hence the determination of effective length from cooled
columns may not be accurate.
Yang et al. (2006a) carried out compression tests of hot-rolled steel columns at
elevated temperatures using the steady state test method. As reported under local
buckling (Section 2.3.2.1), they observed that up to 500oC, hot-rolled steel columns
retained 70% of their strength at ambient temperature and thereafter decreased
rapidly (see Figure 2.18). Depending on the slenderness ratio of steel columns,
failure mode of steel columns changed from global buckling at ambient temperature
to local buckling at elevated temperatures (Figure 2.21). Compressive strength of
steel columns decreases with the increase of the slenderness ratio, and it decreases
dramatically when the slenderness ratio is greater than 50 at elevated temperatures
(Figure 2.29). However, such a limit has not been defined the cold-formed steel
columns.
Figure 2.28: Variation of Reduction Factor for Axial Strength of Steel Columns
with Temperature (Yang et al., 2006)
__________________________________________________________________
Figure 2.29: Effect of Slenderness Ratio on Axial Load Carrying Capacity at
Elevated Temperatures (Yang et al., 2006)
Burgress et al. (1992) numerically analysed the failure of hot-rolled steel columns in
fire. Analyses were based on finite strip method while considering the non-linear
material characteristics at elevated temperatures given in BS 5950 Part 8. They
analysed 203x203x52 universal columns with slenderness ratio ranging from 20 to
180. From their results, they observed that flexural buckling is predominant when
the slenderness ratio is greater than 100 at ambient temperature. At 200oC, the failure
pattern is similar to that of ambient temperature, but the behaviour of steel columns
follows a similar pattern when the temperature is beyond 200oC (see Figure 2.30a).
They concluded that columns with different slenderness ratios lose their capacity at
different rates as temperature increases (see Figure 2.30b).
__________________________________________________________________
(b) Failure Stress for Different
(a) Failure Stress at Elevated
Slenderness with Different Temperatures
Temperatures with the Slenderness
Figure 2.30: Analytical Results of Burgress et al. (1992)
Talamona et al. (1997) numerically modelled a large number of steel universal

columns at elevated temperatures. Two analytical softwares, namely SAFIR and
LEANS, both are specifically developed for analysing structures subjected to fire,
were used. After analysing a large number of tests, a proposal was made for
calculating the capacity of columns subjected to axial loading, for temperatures
between 400oC to 800oC. The proposed method was similar to the methods given in
Eurocode 3 Part 1.2. Franssen et al. (1995) and Franssen et al. (1996) also proposed
a similar method to Talamona et al. (1997). Franssen et al. (1996) compared the
proposed method by Franssen et al. (1995) with 59 experimental results found in
literature and 21 original test results. They proposed a modification to the severity
factor to obtain safer results.
Ranby (1999) modelled flexural torsional buckling of cold-formed steel columns at

elevated temperatures. But the modelling was limited to columns inside a wall panel.
Kaitila (2002) also numerically simulated columns restrained on one flange with
__________________________________________________________________
thermal gradient. Therefore conclusive remarks cannot be drawn for unrestrained
columns under uniform temperature.
Chen and Young (2007) numerically simulated cold-formed steel columns subjected
to local and global buckling of lipped channel columns. However, their validation
was based on the ambient temperature tests.
Test Rigs
There are two methods used for the fire tests of long columns. The most common
method is the test inside a furnace. The other method is heating the column by
attaching heating elements to it while covering it with an insulation blanket made of
ceramic fibber. Unlike local buckling, global buckling requires a long furnace.
Standard loading machines cannot accommodate tall furnaces. Therefore most
researchers developed a loading mechanism to suit the dimensions of the furnace.
Figure 2.31 shows the test arrangement used by Ali and O’Connor (2001). They used
a load cell and hydraulic ram above the furnace. However, most researchers prefer to
have a loading system below the furnace to protect load cells from over heating.
They used pin-end ands hence global buckling can be observed with shorter
columns. Wang and Davies (2003) used a horizontal furnace (see Figure 2.32).
Horizontal furnace can be used for long columns and working is also easy. However,
the reaction frame needs to be made with higher bending stiffness. Figure 2.33
shows the test arrangement used by Aasen (1985). Low voltage heating elements
were attached to the columns and the major advantage is the possibility of having
thermal gradient across the sections by controlling the power supply to the heating
elements.
__________________________________________________________________
Figure 2.31: Tests of Ali and O’Connor (2001)
Figure 2.32: Horizontal Furnace used by Wang and Davies (2003)
__________________________________________________________________
Figure 2.33: Fire Test Arrangement used by Aasen (1985)
Eurocode 3 Part 1.2 (ECS, 2005)
Eurocode 3 Part 1.2 (ECS, 2005) provides design guidelines for steel compression
members. These guidelines are specified for Class 1, 2 and 3 steel sections.
However, in the case of Class 4 sections, same guidelines are recommended with
some limitations. It is recommended to use ambient temperature effective area
according to Eurocode 3 Part 1.3. Eurocode 3 Part 1.2 provides a separate table of
yield stress reduction factors for Class 4 sections. However, there is no difference
between the reduction factors of hot-rolled and cold-formed steels. No reduction
factors are given for the elastic modulus of cold-formed steels at elevated
temperatures. Therefore the accuracy of the design method should be assessed
__________________________________________________________________
because the behaviour and properties of cold-formed steels are different from hot-
rolled steels.
The design buckling resistance N b , fi ,t , Rd at time t and at temperature θ of a
compression member can be calculated from Equation 2.37.

N b , fi ,t , Rd = χ fi Ak y ,θ f y γ M , fi (2.37)
where:
χ fi is the reduction factor for flexural buckling in a fire design situation
(Buckling coefficient);
k y ,θ is the reduction factor for the yield strength of steel at the steel
temperature θ a reached at time t ;
Ag is the gross area of the section;
γ M , fi is the partial safety factor for fire situation;
The value of χ fi should be taken as the lesser of the values of χ y, fi and
χ z, fi determined according to;
Where, χ y , fi and χ z , fi are buckling coefficients about principal axes
1
χ fi = (2.38a)
ϕθ + ϕθ2 − λθ2
ϕθ =
1
2
[
1 + αλθ + λθ2 ] (2.38b)
Imperfection factor α = β 235 f y (2.38c)
β is the severity factor used to provide an appropriate safety level. Eurocode 3 Part
1.2 (ECS, 2005) recommends the use of 0.65 as the severity factor. However,
Franssen et al. (1996) showed that a severity factor of 0.65 led to 50% of safe results
in their analysis. Further, they concluded that a value of 1.2 for the severity factor is
very safe because it led to 82% of experimental loads and for 95% of safe results,
needs to be given a value of 1.46, which led to very conservative results.
The non-dimensional slenderness λθ for the temperature θ a is given by:
λθ = λ [k y ,θ k E ,θ ]0.5
__________________________________________________________________
Where,
λ is the relative slenderness at ambient temperature and is defined as
Le
rmin
λ =
E
π
fy
Where, Le is the effective length of the column;
rmin is the minimum radius of gyration;

E is the Elastic modulus at ambient temperature;
f y is the yield stress at ambient temperature.
k E ,θ is the reduction factor for the modulus of elasticity at the steel temperature θ a
reached at time t .
BS 5950 Part 8 (BSI, 1990)
BS 5950 Part 8 (BSI, 1990) recommended two methods for calculation of the fire
resistance of hot-rolled steel members. They are namely the limiting temperature
method and the moment capacity method. In the case of steel columns, only limiting
temperature method is recommended. Limiting temperature of a column can be
found based on the load ratio and slenderness. Based on the limiting temperature and
flange thickness of the column, fire resistance, in terms of time, can be found.
2.5 Geometric Imperfections and Residual Stresses
2.5.1 Geometric Imperfections
Imperfection of steel members is very important as it reduces the capacity of the

member. Two types of imperfection, namely local imperfection and global
imperfection, can be observed with cold-formed steel members. Local imperfections
mainly affect the local or distortional buckling capacities while global imperfection
mostly affects the member capacity. Local imperfections are expressed in terms of
dimensions of the section and thickness while global imperfections are expressed in
terms of the member length.
__________________________________________________________________
Finite element analyses need imperfections to initiate the failure pattern. Finite
element softwares assume that imperfections follow a similar pattern of buckling
wave with the amplitude of appropriate imperfection. Therefore accurate location of
the maximum imperfection cannot be included in the analyses even if it measured.
However, the imperfection magnitude is more important as the ultimate load is
sensitive to the imperfection magnitude.
2.5.1.1 Local Imperfections
Local imperfection is the deformation of the plate elements of the section. These
deformations may occur while folding or handling. Most researchers measure the
imperfections of their specimens (Ranawaka, 2006, Sivakumaran and Abdel-
Rahman, 1998, Dubina and Ungureanu, 2002). However, most researchers use
standard values from the literature because measurement of local imperfection needs
accurate instrumentations.
Walker (1975) derived an equation to predict the plate imperfection as given next.
He recommended 0.3 for β to obtain the actual local imperfections.
py
∆=β t
pcr
where ∆ is the magnitude of initial imperfection.

t is the thickness of the plate.
β is a constant that can be adjusted to fit experimental results.
py is the yield or crushing load.
pcr is the elastic buckling load.
Walker’s (1975) proposal can be rearranged by substituting the critical buckling load
as given next.
(
12f y 1 −ν 2 )
∆ = 0.3 b
kπ 2 E
__________________________________________________________________
This equation is independent of the thickness of the plate element. Imperfections of
cold-formed steel sections occur during the cold-forming process and handling.
These deformations depend on the bending stiffness of the plate elements. Bending
stiffness of the plate elements is a function of width, (thickness)3, elastic modulus
and supporting type of the plate elements. Hence, imperfections should be described
in terms of all those parameters. But the elastic modulus of steel is almost constant.
Therefore imperfection should be a function of width, thickness and supporting type
of the plate elements.
Schafer and Pekoz (1998) developed models for local imperfections after sorting
available data from literature. These models consider the thickness, width and type
of section. There are some limitations for the width to thickness ratio and for the
thickness. The width to thickness ratio should be less than 200 for stiffened elements
and less than 100 for unstiffened elements while thickness should also be less than 3
mm. Figure 2.34 shows the two types of imperfections considered. Following
equations give the imperfection:
∆ = 0.006w or ∆ = 6te−2t for stiffened elements
∆ = t for unstiffened elements

Where; ∆ is the imperfection
w is the width of the stiffened element
t is the plate thickness
Figure 2.34: Imperfection Types Defined by Schafer and Pekoz (1998)
Recent research by Sivakumaran and Rahman (1998) used a double sine wave
(Figure 2.35) for the imperfections ( δ ) of stiffened elements where δ0 is the upper
__________________________________________________________________
limit of the imperfections. In the analyses, they used half of the upper limit to reflect
the average imperfection. However, their method was limited to stiffened elements.
Therefore the method proposed by Schafer and Pekoz (1998) is widely used because
it covers both stiffened and unstiffened elements.
πX πY
δ = δ0 sin sin
W W
Where;
δ0 W Fy
= 0.145
t t E
Figure 2.35: Imperfection Model used by Sivakumaran and Rahman (1998)
Kaitila (2002) and Feng et al. (2004) investigated the sensitivity of imperfections on
cold-formed steel columns at elevated temperatures. Kaitila used local imperfection
as a function of web width (h) and global imperfection as a function of the member
height (L). Kaitila (2002) concluded that an imperfection of h/200 gives accurate
estimation. Feng et al. (2004) used local imperfection as a function of thickness (t)
and web width. They used h/200, t, 0.1t, and 0.5t as the local imperfections.
2.5.1.2 Global Imperfections
Global imperfection is the out of straightness of the member. It is expressed in terms

of member length. Some researchers used measured imperfection values while some
__________________________________________________________________
others researchers used a standard value of L/1000 where L is the height of the
column. Yan and Young (2002) used a theodolite to measure the imperfections. at
the middle span and at both ends. Some researchers used a standard value of L/1000
where L is the height of the column.
2.5.2 Residual Stresses
There are two types of residual stresses, namely flexural residual stresses and
membrane residual stresses that are presented in steel members. Flexural residual
stresses develop in the steel members due to the folding of section and are dominant
in the cold-formed steel members. Cold-forming induces flexural residual stresses in
the cold-formed steel members whereas membrane residual stresses develop due to
the uneven cooling of the steel section during welding process. Usually cold-formed
steel section is not subject to welding or heat treatment. Therefore membrane
residual stresses are negligibly small in cold-formed steel members. Some
researchers have measured the residual stresses.
Destructive and non-destructive methods are used to determine the residual stresses
of steel sections. The most common method is the strip sectioning. In this method,
strain is taken as the measurement after each cut. Then reduction in strain is
converted to the reduction in stress. The released stresses are the residual stresses.
Figure 2.36 shows the sequence of measuring the residual stress.
Saw cutting method has some disadvantages such as heat generation and disturbance
to the strain gauges due to vibrations. These issues can be avoided by using the
Electrical Discharge Machining (EDM) method (Weng et al, 1990). In this method,
sectioning is carried out by removing the material at the cut by electro osmosis. The
steel section is immersed in a dielectric fluid to provide a path to flow the ions and
cutting tool and a pole is moved closer to the specimen. For each cut, reduction in
strain is measured and hence the release of residual stresses can be calculated. Figure
2.37 illustrates the EDM method of measuring the residual stresses.
__________________________________________________________________
Figure 2.36: Sequence of Measuring Residual Stress from Batista and
Rodrigues (1992)
Figure 2.37: Measurement of Residual Stress by EDM method

(Weng et al, 1990)
Residual stresses in cold-formed steel members have been well investigated by

previous researchers. They depend on the method of the production of cold-formed
members. Rolled-formed steel members have higher flexural residual stresses than
press-braked members. Schafer and Pekoz (1998) reviewed the past research results
__________________________________________________________________
and proposed two residual stress models for rolled-formed and press-braked
members (see Figure 2.38). Corners of the cold-formed profiles were subjected to
more cold work than the flat regions. Therefore residual stress in those regions is
higher than the flat regions. Recent researchers (Dubina and Ungureanu 2002, Lee
2004, Ranawaka 2006) used these models for their research because they gave
accurate predictions. Therefore the residual stress model for press-braked profile as
recommended by Schafer and Pekoz (1998) is used in this research after making the
necessary modifications to the corner stresses.
Figure 2.38: Residual Stress Models Proposed by Schafer and Pekoz (1998)
Residual stresses reduce with increasing temperature (Lee, 2004). Therefore reduced
residual stresses were used in the analyses at elevated temperatures. At about 800oC,
residual stresses fully relieve in steel containing 0.2% of carbon. Based on this fact,
Lee (2004) proposed an equation to find the reduced residual stress at elevated
temperatures.
α =1.0181 − 0.00128 T
Where
α is the residual stress reduction factor
T is the temperature in oC and 20 ≤T ≤ 800
2.6 Analytical Methods
Numerous analysis methods are currently available. Finite Element Method (FEM)
and Finite Stripe Method (FSM) are the most common methods used in the analyses.
__________________________________________________________________
FSM is commonly used for the calculation of elastic buckling load while FEM is
used for obtaining elastic buckling and ultimate failure loads. Thin-wall and CUFSM
are the most common finite strip analysis program. Finite strip analyses produce
buckling plot that can be used to find the elastic buckling load, buckling modes and
buckling half wave length. Further, it can be used to find the essential section
properties such as shear centre, torsional constant and warping constant. There are a
number of finite element softwares available (ABAQUS, ANSYS, SIMULA,
ALGOR etc.). In addition to these general finite element softwares, specific
softwares are also available for fire design (SAFIR, LEANS). Finite element
analyses need a pre-processor to create input files and post processor to read the
results. PATRAN is one of the common pre/post processors used infinite element
analyses.
2.7 Fire Safety Design
Buchanan (2001) explained three major approaches for the fire safety design. They
are design for time, temperature and strength domains. Time domain needs failure
time to be greater than the fire resistant time defined for the structure. In the case of
temperature domain, rated temperature of the structure that causes failure should be
greater than the maximum temperature reached during a fire. BS 5950 Part 8
recommends time and temperature domains. According to the strength domain, load
bearing capacity of the structure should be greater than the applied load in a fire.
Eurocode 3 Part 1.2 uses this method.
2.8 Literature Review Findings
The use of cold-formed steel members for structural purposes has increased recently
due to a number of advantages over the other types of structural materials such as
hot-rolled steel, timber and aluminium in light structural applications. Therefore
researchers have focused their attention on the behaviour of cold-formed steel
structural members. However, the knowledge on the behaviour of cold-formed steel
members is still developing. There is no extensive set of data in the literature on the
__________________________________________________________________
design of cold-formed steel members for various structural applications, in particular
for fire design.
AS/NZS 4600 (SA, 1998), BS 5950 Part 5 (BSI, 1990) and Eurocode 3 Part 1.3
(ECS, 1996) give design guidelines for cold-formed steel members. However, these
design guidelines are given only at ambient temperature. Researchers argue the
validity of some design guidelines. BS 5950 Part 8 (BSI, 1990) and Eurocode 3 Part
1.2 (ECS, 2005) give design guidelines for steel members subjected to fire. These
two codes are heavily focused on hot-rolled steel members and only minor attention
has been paid to the design of cold-formed steel members. Eurocode 3 Part 1.2 limits
the temperature to 350oC for Class 4 steel sections. Most research concluded that this
is a very conservative approach.
BS 5950 Part 8 (BSI, 1990) and Eurocode 3 Part 1.2 (ECS, 2005) provide elastic
modulus and yield stress reduction factors for hot-rolled steel while they provide
only yield strength reduction factors for cold-formed steels. Most researchers used
0.2% proof stress as the yield stress for the fire design. However, BS 5950 Part 8
provides yield stress reduction corresponding to the 0.5% strain instead of 0.2%
proof stress reduction factors for the cold-formed steels. In addition, it provides
stresses corresponding to the stress levels of 1.5% and 2% highlighting the fact that
steel is more ductile at elevated temperatures. Recent researchers experimentally
verified that the code provisions on the mechanical properties of cold-formed steels
at elevated temperatures are not accurate.
Extensive research could be found in the literature on hot-rolled steel compression

members at elevated temperatures. However, only limited research is available on
cold-formed steel compression members. Most of them were focussed on local or
distortional buckling behaviour. Some researchers recommended using ambient
temperature design rules with reduced mechanical properties for the fire design.
Most of their research were based on finite element modelling, which were validated
with ambient temperature tests. Hence extensive detailed research data does not exist
on the buckling behaviour of cold-formed steel compression members to assess the
accuracy of current design methods or form valid design rules. Therefore it can be
concluded that a research gap currently exist on the buckling behaviour of cold-
formed steel compression members at elevated temperatures.
__________________________________________________________________
3 Mechanical Properties of Cold-formed
Steels at Elevated Temperatures
3.1 Introduction
Yield strength and elastic modulus are the most important mechanical properties in
the fire design of steel structures because the variation of these properties with
temperature is significant compared to others. These two properties deteriorate with
increasing temperature. Therefore the mechanical properties at various temperatures
are required for the fire safety design of steel members. BS5950-Part 8 and Eurocode
3 Part 1.2 provide suitable reduction factors for mechanical properties of cold-
formed steels at elevated temperatures. However, BS5950-Part 8 only provides
reduction factors for yield strength 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. Eurocode 3 Part 1.2 provides suitable reduction factors for both yield
strength and elastic modulus for cold-formed steels but they are similar to hot-rolled
steels despite the fact that their reduction of mechanical properties are considered to
be different. Sidey and Teague (1988) stated that the strength reduction of cold-
formed steels at elevated temperatures may be 20% higher than that of hot-rolled
steels. Further, past research on mechanical properties have been mainly focussed on
hot-rolled steels. However, some studies have been carried out on the reduction of
mechanical properties of cold-formed steel at elevated temperatures (Makelainen and
Miller, 1983, Outinen et al. 2000, Outinen and Makelainen, 2001, Lee et al., 2003,
Outinen and Makelainen, 2004, Chen and Young, 2007a, Ranawaka and Mahendran,
2009, Dolamune Kankanamge, 2009). To date the mechanical properties of cold-
formed steels with 0.55 mm to 1.95 mm thicknesses have been investigated by QUT
researchers, and suitable equations have been developed to determine the reduction
factors at elevated temperatures. To investigate the accuracy of these reduction factor
equations for very thin cold-formed steels, a series of tensile coupon tests was
carried out on G550-0.42 mm cold-formed steel. This thickness was considered here
since it was to be used in making the lipped and unlipped C-section columns for the
study on their flexural and flexural torsional buckling at elevated temperatures.
____________________________________________________________________
3.2 Experimental Investigation
3.2.1 Test Method
The most commonly adopted method to determine the mechanical properties of steel
is to perform tension and compression coupon tests. Past research (Ranawaka, 2006)
shown that there are minimal differences between the mechanical properties obtained
from tension and compression coupon tests. Therefore the easier tensile coupon tests
are commonly used in most research project. Tensile tests can be carried out under
steady state and transient state conditions. In the case of steady state tests, the
specimen is heated up to the required temperature and then a tensile load is applied
at a constant rate until failure while maintaining the pre-set 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 was kept constant. In the case of transient state tests, the
applied load was kept constant and the specimen is heated until failure. From tensile
coupon tests, the stress-strain curves are obtained and thereby the required
mechanical properties such as yield strength, ultimate strength and elastic modulus
can be determined.
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, 2004, 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. Since they
are carried out at a uniform temperature, thin test specimens will also have a uniform
temperature during the tests. In the case of transient tests, additional effort is needed
to be taken to ensure that the measured temperature is the same as the specimen
temperature. 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 need to be converted to a stress-strain curve. This
process is difficult and can easily lead to errors in the results. Therefore many
researchers used the steady state test method for the mechanical property tests.
Hence, in this research the steady state test method was adopted to determine the
mechanical properties of 0.42 mm thick Grade 550 cold-formed steel at elevated
temperatures.
3.2.2 Test Specimen
The shape and size of test specimens were in accordance with AS 1391 (SA, 1991)
as illustrated in Figure 3.1. A 10.5 mm diameter 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 chosen so that
there was not any premature bearing 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.
35 mm 10.5 mm Diameter 35 mm
25 mm Radius
50 mm
60 mm
13 mm
70 mm 25mm 25mm 70 mm
(a) Dimensions of Tensile Test Specimen
(b) Tensile Test Specimen
Figure 3.1: Tension Test Specimen and its Dimensions

____________________________________________________________________
Cold-formed steel sheets are generally coated with zinc based metal, which provides
corrosion protection. This coating does not provide any additional strength to the
cold-formed steel. Therefore this metal coating was removed to determine the Base
Metal Thickness (BMT). The other advantage in removing the coating is to facilitate
the taking of laser images that are used to measure the strains at elevated
temperatures. Therefore the coating of each specimen was removed by immersing it
in diluted hydrochloric acid (acid to water = 1:3). The specimen thickness was then
measured at three points within the gauge length using a micrometer and the average
of these three readings was taken as the base metal thickness. The specimen width
was also measured at three locations using a vernier calliper and their average was
taken as the width. The measured specimen dimensions were used in the calculation
of mechanical properties. Table 3.1 shows the dimensions of the tensile test
specimens.
Table 3.1: Dimensions of Tensile Test Specimens
Temperature (oC) Test Number BMT (mm) Width (mm)

20 1 0.40 12.86
2 0.40 12.78
100 1 0.41 12.78
2 0.41 12.78
200 1 0.41 12.78
300 1 0.40 12.78
2 0.41 12.78
400 1 0.41 12.76
2 0.40 12.74
500 1 0.40 12.76
2 0.40 12.78
600 1 0.40 12.67
2 0.40 12.61
1 0.41 12.62
700 2 0.40 12.62
3 0.40 12.74
3.2.3 Test Rig and Procedure
Tensile coupon tests were conducted inside a small electrical furnace in the structural
laboratory (see Figure 3.2). This furnace is fitted with four glow bars to uniformly
____________________________________________________________________
heat the test specimens. Tensile specimen was located at the geometric centre of the
four glow bars. This ensured maximum heat transfer to the specimen since glow bars
produce radiative heat.
Tensile load
Hydraulic actuator
Load cell
Top
shaft
Glow bars
Thermocouples
Specimen
Glass
window
Bottom
shaft
Figure 3.2: Test Rig
A micro-computer based temperature control system controlled the pre-set

temperature and heating rate while two internal thermocouples fitted inside the
furnace gave the air temperature of the furnace to the control system. Since the two
internal thermocouples were located a short distance away from the specimen, an
additional thermometer was used in contact with the specimen as a means of
independent verification of the specimen temperature. It was observed that the
independent specimen temperature measurement gave a slightly higher temperature
than that indicated by the furnace. This difference is considered to be due to radiative
____________________________________________________________________
heating. Therefore the independently measured temperature was taken as the pre-set
temperature of the test. An observation window located on one side of the furnace
was used sight the specimen for the use of laser speckle extensometer. Tensile
specimen was mounted between the two loading shafts made of 253MA stainless
steel, which has a higher strength at elevated temperatures.
Loading arrangement consisted of a hydraulic actuator connected to the upper

loading shaft through a load cell. A one ton load cell was used to measure the tensile
load. The hydraulic actuator was rigidly connected to the crosshead made of a deeper
beam. The bottom loading shaft was rigidly connected to the basement of the
structural lab. Special care was taken to align the loading shaft to avoid any eccentric
loading. Details of the test set-up are shown in Figure 3.3. The hydraulic actuator
was connected to the Multi-purpose TestWare System. A data logger, EDCAR, was
used to process the data and to obtain the stress-strain curves (see Figure 3.4). Data
channels from load cell, strain gauges and control device of laser speckle
extensometer were connected to EDCAR.
The steady state test method was adopted for the elevated temperature tests. Selected
temperatures in this study were 20oC, 100oC, 200oC, 300oC, 400oC, 500oC, 600oC
and 700oC. A heating rate of 20oC/min was used. The specimen temperature along
the length was assumed uniform because of long glow bars (700 mm) compared to
the specimen gauge length of 60 mm and the fact the specimen was located at mid
height. Once the furnace temperature reached its target temperature, another 10
minutes was allowed to ensure a uniform temperature throughout the specimen
before the commencement of loading. During the heating phase, special care was
taken to eliminate the development of any compression load in the specimen due to
thermal expansion. Therefore a small tensile load was maintained during the heating
phase. Once a uniform temperature condition was ensured in the specimen, a tensile
load was applied at a constant strain rate constant until the failure of the specimen.
The strain rate should be within the range of 0.001 to 0.005/min according to the
high temperature testing standard for metallic materials, SFS-EN10002-5 (ECS,
1992). In this study, a displacement rate of 0.2 mm/min was used, which is
equivalent to a strain rate of 0.0033/min and thus satisfies the requirement of SFS-
EN100002-5 (ECS, 1992).
____________________________________________________________________
Load cell
(c) Top End
Glass Window
(a) Furnace
Camera
(d) Side View of Laser Speckle
Extensometer with Cameras
(b) Bottom End
(e) Specimen Inside the Furnace
Figure 3.3: Details of Test Arrangement

____________________________________________________________________
Figure 3.4: EDCAR (Experimental Data Collection And Recorder)
Strains are usually measured using strain gauges. Accuracy of this method is
significantly higher and hence this method is widely used in engineering
applications. However, they can not be used for elevated temperature tests, because
of the burning the circuit. Although elevated temperature strain gauges are available,
they are very expensive. Therefore researchers used mechanical extensometers to
measure the strains during elevated temperature tests (Lee, 2003). However, the
accuracy of this method is dependent on the precision of extensometer and other
connecting devices. Recent researchers of QUT used an advanced method of strain
measurement using laser speckle extensometer which avoids the difficulties in the
strain measurement using mechanical extensometer at elevated temperatures.
Therefore a laser speckle extensometer (see Figure 3.5) was used in this research to
measure the strains at elevated temperatures. This method needs a precise calibration
to obtain accurate results. Strains measured by strain gauges at ambient temperature
were used to validate the strain measurements using the laser speckle extensometer.
The laser speckle extensometer was developed by an Austrian Company Messphysik
GmbH, Laser speckle extensometer has the capability of measuring very small
strains accurately and comprised of a PC based Video Processor, laser diodes (class
3A), video cameras, lens and a frame grabber. Monochrome video cameras, which
are capable of easily adjusting to obtain the correct images, 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 at temperatures above 500oC.
Pivoted cameras and Lenses of the

lens mounting device cameras
Two lasers
Base plate
Data transferring
cable
Figure 3.5: 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 special window made from
fire resistant glass (Figure 3.3(d)). Two laser beams and two cameras were oriented
targeting the specimen as shown in Figure 3.3(e). The upper camera is referred to as
the slave while the lower camera is master.
____________________________________________________________________
Specimen
Slave camera
50 mm
Master camera
Laser beams
Figure 3.6: Strain Measurements
The measuring principle of the 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 as shown in
Figure 3.7. Each camera produces unique speckle patterns relevant to the zones 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 reference speckle patterns. When the tensile load is applied to the
specimen, targeted zones of the two cameras change 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 3.7(a) shows typical speckle patterns of master and slave cameras
before the test while Figure 3.7(b) shows typical speckle patterns during the test.
Before the tests, the laser-speckle extensometer was calibrated with a special
calibration method, which enables accurate strain measurements. 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
testing, the processor is able to calculate the strain at any 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 from

master camera slave master camera
(b) Speckle Output During the Test
Figure 3.7: Typical Speckle Output for Strain Measurements
____________________________________________________________________
Mechanical properties at ambient temperature were measured using both the strain
gauge and the laser speckle extensometer. Figure 3.8 compares the ambient
temperature stress-strain curves obtained from both methods, which are in good
agreement. In addition to the graphical comparison, Table 3.2 compares the
mechanical properties of yield strength and elastic modulus obtained from both
methods and the agreement is good. Therefore it can be concluded that the laser
speckle extensometer accurately measures the strain of tensile specimen under
loading. Hence all the elevated temperature tensile coupon tests were carried out by
using the laser speckle extensometer.
800
700
600
500
Stress (MPa)
400
300 Laser Speckle Extensometer
Strain Gauge
200
100
0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5
Strain (%)
800
700
600
500
Stress (MPa)
400
300 Laser Speckle Extensometer
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 0.42 mm Thick G550 Steel at 20oC
____________________________________________________________________
Table 3.2: Yield Strength and Elastic Modulus at Ambient Temperature
Yield Strength (MPa) Elastic Modulus (MPa)

Test
Number Strain Laser speckle Strain Laser speckle
gauge extensometer gauge extensometer
1 673.82 664.80 210568 207500
2 678.80 676.84 218968 218770
3.2.4 Mechanical Properties at Ambient and Elevated

Temperatures
3.2.4.1 Determination of Yield Strength and Elastic Modulus
Unlike the low strength steels, high strength steels do not show a well defined yield
point in the stress-strain curve at ambient temperature. It shows a linear stress-strain
relationship in the initial part of the stress-strain curve and thereafter a gradual
yielding pattern (see Figure 3.8). The gradual yielding pattern is more dominant in
the elevated temperature tests. Therefore the yield stress is defined as either 0.2%
proof stress or strength corresponding to 0.5%, 1.5% and 2.0% total strain.
Figure 3.9 illustrates the different definitions of yield stress. However, the most
common method is to use 0.2% proof stress as the yield stress. Therefore 0.2% proof
stress was taken as the yield stress in this research. Elastic modulus was taken as the
slope of the linear region of the stress-strain curve.
____________________________________________________________________
σu
σ0.5 σ1.5 σ2.0
σ0.2
Stress, σ
∆σ
E = tan α =
∆ε
0.002 0.015 Strain, ε

0.005 0.020
Figure 3.9: Stress-strain Curve of High Strength Steels
Figure 3.10(a) to (d) show the typical stress-strain curves obtained for 0.42 mm thick
G550 cold-formed steel at 300 oC and 500oC temperatures. Other stress-strain curves
are given in Appendix A. Figure 3.11 (a) and (b) show the comparison of stress-
strain curves for G550-0.42 mm thick cold-formed steel at ambient and elevated
temperatures. They show that ductility has increased with increasing temperature. A
significant increase in the ductility can be seen for temperatures beyond 500oC.
However, the lowest ductility can be seen at 200oC. Other researchers noted that low
strength cold-formed steels achieved the higher ultimate tensile strength at 200oC
(Ranawaka and Mahendran, 2009 and Dolamune Kankanamge, 2009).
____________________________________________________________________
800
700
600
Stress (MPa) 500
400
300
200
100
0
0.00 2.00 4.00 6.00 8.00 10.00 12.00
Strain (%)
(a) G550-0.42 Steel at 300oC (Full Stress-strain Curve)
800
700
600
500
Stress (MPa)
400
300
200
100
0
0.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80 0.90 1.00
Strain (%)
(b) G550-0.42 Steel at 300oC (Initial Region)
Figure 3.10: Stress-strain Curves at Elevated Temperatures
____________________________________________________________________
350
300
250
Stress (Mpa)
200
150
100
50
0
0.00 2.00 4.00 6.00 8.00 10.00 12.00
Strain (%)
(c) G550-0.42 Steel at 500oC (Full Stress-strain Curve)
350
300
250
Stress (Mpa)
200
150
100
50
0
0.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80 0.90 1.00
Strain (%)
(d) G550-0.42 Steel at 500oC (Initial Region)
Figure 3.10: Stress-strain Curves at Elevated Temperatures
____________________________________________________________________
700
600
500 20 100
Stress (N/mm )
200 300
2
400
400 500
300
600 700
200
100
0
0.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0 40.0
Strain (%)
(a) Full Stress-strain Curves
700
600
500 20 100
Stress (N/mm )
200 300
2
400
400 500
300
600 700
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 (%)
(b) Initial Regions
Figure 3.11: Stress-strain Curves at Different Temperatures
____________________________________________________________________
3.2.4.2 Results
Both yield stress and elastic modulus were calculated for all the temperatures. Most
of the tests were repeated to obtain accurate results and the average yield stress and
elastic modulus values were determined. Table 3.3 gives the yield strength while
Table 3.4 gives the elastic modulus at temperatures in the range of 20oC to 700oC.
Table 3.3: Yield Stress of G550-0.42 mm Cold-formed Steel
Temperature Yield stress (MPa)

o
C Test 1 Test 2 Test 3 Average
20 673.82 678.80 - 676.31
100 663.00 667.40 - 665.20
200 649.90 - - 649.90
300 615.70 609.20 - 612.45
400 440.00 445.80 - 442.90
500 240.00 230.50 - 235.25
600 65.40 65.20 - 65.30
700 35.00 30.00 34.62 33.20
Table 3.4: Elastic Modulus of G550-0.42 mm Cold-formed Steel
Temperature Elastic Modulus, E (GPa)

o
C Test 1 Test 2 Test 3 Average
20 210.56 218.97 - 214.77
100 211.36 213.33 - 212.35
200 205.00 - - 205.00
300 164.94 160.00 - 162.47
400 127.00 130.00 - 128.50
500 84.30 88.60 - 86.45
600 52.60 51.80 - 52.20
700 43.95 41.66 45.45 43.69
Since the mechanical properties of cold-formed steel vary with thickness,
appropriate reduction factors are defined for comparison purposes. Reduction factors
help to predict the elevated temperature mechanical properties of different batch
steels in the same strength category. They are defined as the ratio of mechanical
property at a particular temperature to that at ambient temperature. The yield
strength reduction factor at temperature ToC, k(Y,T), is defined as the ratio of yield
strength at temperature ToC to that at ambient temperature. Similarly, the elastic
____________________________________________________________________
modulus reduction factor at temperature ToC, k(E,T), is defined as the ratio of elastic
modulus at temperature ToC to that at ambient temperature. Table 3.5 shows the
yield strength and elastic modulus reduction factors for G550-0.42 mm cold-formed
steel at ambient and elevated temperatures. Both yield strength and elastic modulus
were found to deteriorate with increasing temperature. However, the rate of
reduction was different. Up to 300oC, the yield strength did not reduce much.
Thereafter, it reduced at a higher rate up to 600oC beyond which it reduced at a
slower rate. However, at 600oC, it has lost almost 90% of ambient temperature
values. Unlike yield strength, elastic modulus shows a linear reduction pattern. The
calculated reduction factors were compared with the available results from other
researchers and design standards.
Table 3.5: Mechanical Property Reduction Factors for G550-0.42 mm

Cold-formed Steels
Temperature Reduction Factor
o
C Elastic Modulus, k(E,T) Yield Strength, k(Y,T)
20 1.000 1.000
100 0.989 0.984
200 0.955 0.961
300 0.756 0.906
400 0.598 0.655
500 0.403 0.348
600 0.243 0.097
700 0.203 0.049
3.2.4.3 Comparison of Test Results with Available Predictive Method
Test results were compared with the available predictive methods proposed by
different researchers. First mechanical property reduction factors were compared and
then stress-strain test curves were compared with the model.
____________________________________________________________________
3.2.4.3.1 Mechanical Property Reduction Factors
Eurocode 3 Part 1.2 and BS 5950 Part 5 provide mechanical property reduction
factors for cold-formed steels at elevated temperatures. However, Chen and Young
(2007a), Ranawaka and Mahendran (2009) and Dolamune Kankanamge (2009) also
developed new predictive equations for the mechanical property reduction factors
based on tensile test results at elevated temperatures. Since they found that the code
reduction factors were not accurate. These new predictive equations are given in
Chapter 3. In this study the results are compared only with the predictive equations
proposed by these researchers.
1.2
1.0
0.8
fy,T/fy,20
0.6
Test
Dolamune Kankanamge (2009)
0.4
Ranawaka & Mahendran (2009)
Chen & Young (2007) 1.9 mm
0.2 Chen & Young (2007) 1.0 mm
0.0
0 100 200 300 400 500 600 700 800
Temperature, oC
Figure 3.12: Comparison of Yield Strength Reduction Factors with the

Available Predictive Methods
Figure 3.12 compares the yield strength reduction factors while Figure 3.13
compares the elastic modulus reduction factors with four different predictive
equations proposed for higher strength steels. It can be observed that equations
proposed by Dolamune Kankanamge (2009) and Ranawaka and Mahendran (2009)
predict closer results. However, the yield strength reduction factors for 0.42 mm
____________________________________________________________________
G550 steel appear to be slightly below the predictions as the predictive equations
were developed based on the results for thickness steels ( 0.95 mm).
1.2
TEST
1.0
Dolamune Kankanamge (2009)
Ranawaka & Mahendran (2009)

0.8
Chen & Young (2007) - G450
0.6
ET/E20
0.4
0.2
0.0
0 100 200 300 400 500 600 700 800 900
Temperature, oC
Figure 3.13: Comparison of Elastic Modulus Reduction Factors with the

Available Predictive Methods
3.2.4.3.2 Stress-Strain Curves
The stress-strain curves at elevated temperatures are usually based on the Ramberg-
Osgood stress-strain model, which is defined by Equation 3.2 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 (2009) proposed β to be taken as 0.86 and the parameter
η T from Equation 3.3.
ηT
f f y ,T fT
εT = T + β (3.2)
ET ET f y ,T
η T = −3.05 x10 −7 T 3 + 0.0005T 2 − 0.2615T + 62.653 20 ≤ T ≤ 800 o C (3.3)
____________________________________________________________________
All the test curves were compared with the stress-strain model. Figures 3.14 to 3.21
show the comparison of test curves with the Ramberg Osgood model. They showed
a good agreement in the initial region of all the curves.
It is proposed to use the stress-strain relationships based on Ramberg Osgood model

based on above equations in our numerical analyses using ABAQUS (Chapter 5 and
8). It is assumed that the nonlinearity and strain hardening phenomena simulated by
Ramberg Osgood models together with the traditional Von Mises yield criteria do
not lead to any inaccuracies in the numerical predictions.
800
700
600
Stress (N/mm )
500
2
400
Test
300
Ramberg Osgood Model
200
100
0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
Strain (%)
Figure 3.14: Comparison of Test and Predicted Stress-Strain Curves at 20oC
____________________________________________________________________
800
700
Stress (N/mm ) 600
500
2
400
Test
300
200
100
0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
Strain (%)
800
700
600
Stress (N/mm2)
500
400
Test
300 Ramberg Osgood Model
200
100
0
0.0 0.5 1.0 1.5 2.0 2.5
Strain (%)
____________________________________________________________________
700
600
500
Stress (N/mm2)
400
Test
300
200
100
0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
Strain (%)
600
500
400
Stress (N/mm )
2
300 Test

200
100
0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
Strain (%)
____________________________________________________________________
350
300
250
Stress (N/mm2)
200
150
Test
100
50
0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
Strain (%)
100
90
80
70
Stress (N/mm2)
60
50
40 Test
30
20
10
0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
Strain (%)
____________________________________________________________________
60
50
40
Stress (N/mm2)
30
20 Test
10
0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0
Strain (%)
3.3 Conclusions
A series of tensile coupon tests was carried out on G550-0.42 mm steels at elevated
temperatures in the range of 20 to 700oC. Stress-strain curves were obtained for each
test to obtain the mechanical properties and to understand the behaviour of cold-
formed steels at ambient and elevated temperatures. Strains were accurately
measured at elevated temperatures using a new non-contact device, the laser speckle
extensometer. Both yield strength and elastic modulus were determined form these
tests at elevated temperatures and are presented in this chapter. The results show the
deterioration of these mechanical properties with increasing temperature. The
ductility of cold-formed steels was noted to increase with increasing temperature
with significant increases beyond 500oC. However, the lowest ductility can be
observed at 200oC. A significant increase in the ductility can be seen for
temperatures beyond 500oC. Mechanical property reduction factors were calculated
to investigate the accuracy of predictive equations for yield strength and elastic
modulus reduction factors developed by other researchers in recent times. A
comparative study concluded that the methods proposed by Ranawaka and
Mahendran (2009) and Dolamune Kankanamge (2009) are accurate enough to
____________________________________________________________________
predict the mechanical properties of thin high strength cold-formed steels at elevated
temperatures. Further, the modified Ramberg Osgood stress-strain model as
proposed by Ranawaka and Mahendran (2009) and Dolamune Kankanamge (2009)
can be used to obtain accurate stress-strain curves at elevated temperatures.
____________________________________________________________________
4 Experimental Study of Cold-formed Steel
Compression Members Subject to Local
Buckling
Cold-formed steel short columns are susceptible to local and distortional buckling
effects. Many cold-formed steel codes, AS/NZS 4600 (SA, 2005), Eurocode 3: Part
1.3 (ECS, 1996), BS 5950 Part 5 (BSI, 1998) and the North American Specification
(NAS, 2003), and the Direct Strength Method (DSM) give suitable design guidelines
for cold-formed steel short columns subjected to local buckling at ambient
temperature. BS 5950 Part 8 (1990) and Eurocode 3: Part 1.2 (2005) give design
guidelines for steel members at elevated temperatures. These two design standards
mainly focus on hot-rolled steel members and recommend the same guidelines for
the design of cold-formed steel members with some limitations, which leads to
uneconomical designs, despite the fact that the behaviour of cold-formed steel
members is different from that of hot-rolled steel members. Distortional buckling
behaviour of cold-formed steel members was investigated at ambient temperature by
Schafer (2001, 2002, 2006) while further investigations at ambient and elevated
temperatures were conducted by Ranawaka (2006), who proposed new design
guidelines for elevated temperatures. However, local buckling behaviour of cold-
formed steel members at elevated temperatures is not fully investigated and hence
there is no well documented test data to form design guidelines at elevated
temperatures. Therefore a series of local buckling tests was carried out on cold-
formed steel short columns at ambient and elevated temperatures with two
thicknesses (0.95 mm and 1.9 mm) and three grades (G250, G450 and G550). This
chapter describes the selection of test specimens, preliminary investigation, method
of testing, numerical analysis and the results.
4.1 Experimental Investigation
Two series of experimental investigations were carried out to understand the

behaviour of cold-formed steel compression members subject to local buckling at
ambient temperature and different elevated temperatures up to 700oC. Suitable test
____________________________________________________________________
sections and thicknesses were selected based on the available literature and standard
sections and thicknesses commonly used for structural and architectural applications.
Elastic local buckling analyses of the chosen sections were undertaken using
CUFSM and ABAQUS (HKS, 2004) to ensure the occurrence of local buckling
during experiments. These analyses were then extended to elevated temperatures
using the reduced mechanical properties of 0.95 mm, 1.9 mm and 1.95 mm thick
cold-formed steels from Ranawaka and Mahendran (2009) and Dolamune
Kankanamge (2009). The first test series was carried out at the ambient temperature
and the specimens were loaded using the Tinius Olsen testing machine in the QUT
structural laboratory. The second test series was carried out at elevated temperatures
inside an electrical furnace and the specimens were loaded using the special loading
shafts fitted to the Tinius Olsen testing machine.
4.1.1 Local Buckling Tests at Ambient Temperature
The first series of experimental investigation and numerical analysis were carried out
at room temperature (approximately 20oC). Preliminary investigations were carried
out to find suitable test section types, thickness and steel grades of light gauge cold-
formed steel commonly used for structural members. These sections were analysed
using CUFSM, ABAQUS (HKS, 2004) and design guidelines of AS/NZS 4600 (SA,
2005) to find the most suitable geometry of the specimens. The length of specimen
was selected as a multiple of buckling half wave lengths from CUFSM analyses.
4.1.1.1 Preliminary Investigation and Selection of Test Specimens
Thicknesses of light gauge cold-formed steel vary from of 0.42 mm to 3 mm.

Therefore thicknesses of 0.95 mm and 1.9 mm were selected to represent the light
gauge cold-formed steel domain. Cold-formed steel members are available in two
strength grads, namely low strength and high strength. Therefore both grades were
selected with the above thicknesses in this research. Most commonly available
sections, unlipped and lipped channels, were selected for the testing (see Figure 4.1).
However, the chosen section dimensions were not standard sizes to ensure the
occurrence of desired buckling failures, i.e. buckling in the web element of lipped
channels while buckling in the flange elements of unlipped channels. These buckling
____________________________________________________________________
modes were ensured by calculating the effective widths of stiffened and unstiffened
elements (see Table 4.1) based on AS/NZS 4600 (SA, 2005). A slender web was
selected for the lipped channels while slender flanges were selected for the unlipped
channels to ensure that local buckling occurred in the web elements of lipped
channels and the flange elements of unlipped channels.
Table 4.1: Effective Widths of Stiffened and Unstiffened Elements
Nominal Effective width of element (mm)

Grade of steel
thickness (mm) Stiffened Unstiffened
250 39.0 12.9
0.95
550 27.7 9.1
1.95 250 65.5 21.5
1.90 450 48.0 15.7
Type B section
Flange subjected to local buckling
Type A section
Web subjected to local buckling
Figure 4.1: Selected Cross Section Types for Local Buckling Tests
In this thesis, lipped channels are referred to as Type A sections and unlipped
channels as Type B sections. A series of trial sections of Types A and B, with
different cross-sectional dimensions, were selected to find the most suitable
dimensions for the sections. All the selected sections were analysed using CUFSM to
ensure local buckling, and some buckling plots obtained in the trial analyses are
shown in Figures 4.2 and 4.3. Distortional buckling can be observed in the lipped
channels. Therefore the web width was further increased to avoid any interaction of
distortional and local buckling models. Half wave length for each buckling mode
can be easily identified from the buckling plots. Specimen lengths were selected as a
multiple of half wave lengths. In addition to the above length, an additional 20 mm
____________________________________________________________________
length was included to either ends to minimize the effect of end supports. Table 4.2
shows the buckling half wave lengths ( ) and the chosen specimen lengths for
different sections.
Distortional Flexural-torsional
Buckling Buckling
Local Buckling
Figure 4.2: Buckling Plot of a Lipped Channel Section
Flexural-
torsional
Buckling
Local Buckling
Figure 4.3: Buckling Plot of an Unlipped Channel Section
____________________________________________________________________
Table 4.2: Nominal Dimensions of Test Specimens
Nominal Grade Type Web Flange Length of

Lip
Thickness of of depth width specimens
(mm) (mm)
(mm) steel section (mm) (mm) (mm)
A 60 30 9 50 190
250
B 25 25 55 150
0.95
A 60 25 9 50 190
550
B 25 25 55 150
A 118 38 15 90 320
1.95 250
B 48 49 110 260
A 98 38 15 80 280
1.90 450
B 48 49 110 260
All the external dimensions and the thicknesses of the specimens were measured
using a vernier calliper and a micrometer screw gauge (see Tables 4.4 and 4.5).
Measured dimensions were close to the nominal dimensions proposed to construct
the columns except for Type A specimens with 0.95 mm thickness. These deviations
happened due to the folding difficulties of the proposed dimensions shown in Table
4.2. As noted in Tables 4.4 and 4.5 a total of 24 ambient temperature tests was
conducted. All the specimens were labelled according to the grade, thickness, test
temperature, type of specimen and specimen number as shown next.
G550 – 0.95 – 20 – A 1
Specimen number
Specimen type
Test temperature
Nominal thickness of specimens
Grade of steel
It is common practice to use base metal thickness in the numerical calculations. The
base metal thickness can be obtained by removing the coating of cold-formed steel.
Coating of the cold-formed steel can be removed by immersing the specimens in
diluted hydrochloric acid. Since it is difficult to remove the coating of all the
specimens, the coatings of two specimens of each grade and thickness were removed
____________________________________________________________________
to measure their base metal thicknesses. Thickness measurements were taken at
three different locations for each specimen, and the average thickness was
calculated. Table 4.3 gives the average base metal thickness of each steel grade and
thickness used for the local buckling tests.
Table 4.3: Base Metal Thickness of Selected Cold-formed Steels
Steel grade and Average base metal

nominal thickness thickness (mm)
G550 – 0.95 0.95
G250 – 0.95 0.95
G450 – 1.90 1.88
G250 – 1.95 1.95
Table 4.4: Measured External Dimensions of Type A Specimens Tested at

Ambient Temperature
Specimen Web Flange( Lip Length

(mm) mm) (mm) (mm)
G550 – 0.95 – 20 – A1 62.85 27.65 9.60 190
G550 – 0.95 – 20 – A2 62.80 27.75 9.65 190
G550 – 0.95 – 20 – A3 63.20 27.55 9.60 190
G250 – 0.95 – 20 – A1 63.12 32.50 9.65 190
G250 – 0.95 – 20 – A2 62.85 32.60 9.35 190
G250 – 0.95 – 20 – A3 63.10 32.50 9.30 190
G450 – 1.90 – 20 – A1 100.05 40.55 15.15 180
G450 – 1.90 – 20 – A2 99.80 40.10 15.10 280
G450 – 1.90 – 20 – A3 100.05 40.00 15.00 280
G250 – 1.95 – 20 – A1 120.70 40.45 17.70 320
G250 – 1.95 – 20 – A2 120.60 40.20 15.15 320
G250 – 1.95 – 20 – A3 120.00 40.10 15.10 320
____________________________________________________________________
Table 4.5: Measured External Dimensions of Type B Specimens Tested at
Ambient Temperature
Specimen Web Flange Length

(mm) (mm) (mm)
G550 – 0.95 – 20 – B1 25.35 25.00 150
G550 – 0.95 – 20 – B2 24.90 25.05 150
G550 – 0.95 – 20 – B3 24.80 24.95 150
G250 – 0.95 – 20 – B1 25.05 25.00 150
G250 – 0.95 – 20 – B2 25.00 24.95 150
G250 – 0.95 – 20 – B3 24.95 24.90 150
G450 – 1.90 – 20 – B1 50.15 50.05 260
G450 – 1.90 – 20 – B2 50.20 49.90 260
G450 – 1.90 – 20 – B3 50.35 50.00 260
G250 – 1.95 – 20 – B1 50.55 49.95 260
G250 – 1.95 – 20 – B2 50.25 50.05 260
G250 – 1.95 – 20 – B3 50.10 50.15 260
4.1.1.2 Test Set-up and Procedure
Ambient temperature tests were carried out using the Tinius Olsen testing machine in
the QUT Structural Laboratory (see Figure 4.4 ). This machine can be used to apply
the compression load with user defined strain rates. Inbuilt load cells of the machine
measure the axial load while an inbuilt Linear Variable Displacement Transducer
(LVDT) attached to the cross-head measures the axial displacement. Display unit of
the machine showed the axial load, axial displacement and also the maximum load
during the test. Since local buckling of a short column is independent of end
conditions, a simple end condition was adopted. The specimen was located on the
centre of a thicker steel plate, which was located on the base of the testing machine.
A similar size steel plate was fixed to the cross-head of the machine and the
specimen was placed between the two steel plates.
During the tests, two types of displacements, axial shortening and out-of-plane
displacements, were recorded. Out-of-plane deformations of the test specimens were
measured at the middle of the specimens using 10 mm LVDTs attached to the
____________________________________________________________________
specimens. Axial shortening of the specimens was measured using the inbuilt LVDT
of the machine. In addition to this, an independent 5 mm LVDT was employed (see
Figure 4.5) to measure the axial shortening to ensure accurate results, and both
displacements were plotted against the axial load. Both channel outputs agreed well
with each other. However, the external LVDT shows a steady variation (see Figure
4.6a) and hence only the external LVDT was used to measure the axial shortening.
All the LVDTs and the load channel were connected to a data logger system. This
data logger system was programmed to plot both axial compression load versus axial
shortening and axial compression load versus out-of-plane (buckling) displacement
curves (see Figures 4.6 a and b). All the load deflection curves are given in
Appendix B.
Cross-head
Specimen
Figure 4.4: Test Set-up using Tinius Olsen Testing Machine
____________________________________________________________________
LVDT to measure LVDTs to measure
axial displacement out of plane
displacement
Figure 4.5: Displacement Measurements using LVDTs
30
25
Axial Compression Load (kN)
20
15
10 Cross-Head of Tinius
LVDT Axial Shortening

5
0
0.00 0.25 0.50 0.75 1.00 1.25 1.50
Axial Shortening (mm)
(a) Axial Compression Load versus Axial Shortening
Figure 4.6: Load-Deflection Curves for G550-0.95-20-B
____________________________________________________________________
30
25

20
15
10
0
-1.60 -1.20 -0.80 -0.40 0.00
Out-of-Plane Deflection (mm)
(b) Axial Compression Load versus Out-of-Plane Deflection
Figure 4.6: Load-Deflection Curves for G550-0.95-20-B
4.1.1.3 Observations and Results
Elastic Buckling
Elastic buckling waves were first observed in the tests and their amplitude gradually
increased (see Figure 4.7). In most cases the bifurcation type of buckling was not
observed. Hence determining the elastic buckling load based on visual observations
was almost impossible. Since plate elements are not perfect, bifurcation buckling
load could not be identified from the load-deflection curves. Venkatarmaiah and
Roorda (1982) proposed a method to find the elastic buckling load using axial load
versus out of plane displacement curve for imperfect plates. In this method, axial
compression load versus (out of plane displacement)2 curve is used to determine the
elastic buckling load. Axial compression load versus out-of-plane deflection curves
for Type B specimens showed an approximate bifurcation behaviour while Type A
specimens did not show significant out of plane deflection in the elastic range or
inelastic range before failure. Therefore the method proposed by Venkatarmaiah and
Roorda (1982) could not be used accurately to find the elastic buckling loads.
____________________________________________________________________
Figure 4.7: Elastic Buckling Waves
Ultimate Load
After reaching the elastic buckling load, axial load was further increased owing to
the presence of considerable post-buckling capacity. During the post-buckling stage,
elastic buckling waves were reduced to a single wave as shown in Figure 4.8. The
maximum load during the test is displayed by the machine. In addition the peak
value of the axial load versus axial shortening curves also gives the ultimate load.
Loading
Figure 4.8: Sequence of Local Buckling Failure
____________________________________________________________________
Results
From the tests, only the ultimate load was taken since elastic buckling load was
difficult to be determined accurately based on visual observations (see Table 4.6).
However, both the axial compression versus axial shortening and out-of-plane
displacement curves were recorded for all the tests in order to use them in the
validation of the finite element models.
Table 4.6: Results of Local Buckling Tests at Ambient Temperature
Ultimate load (kN) Average load

Specimen
1 2 3 (kN)
G550 – 0.95 – 20 – A 54.14 53.78 53.92 53.95
G550 – 0.95 – 20 – B 27.12 26.19 26.20 26.50
G250 – 0.95 – 20 – A 37.84 37.19 36.22 37.09
G250 – 0.95 – 20 – B 17.29 16.88 16.34 16.83
G450 – 1.90 – 20 – A 150.67 153.23 149.99 151.30
G450 – 1.90 – 20 – B 95.82 97.96 96.89
G250 – 1.95 – 20 – A 105.55 100.88 99.71 102.04
G250 – 1.95 – 20 – B 67.22 64.44 67.45 66.37
4.1.2 Local Buckling Tests at Elevated Temperatures
The second series of tests was carried out to understand the behaviour of cold-
formed steel short columns subjected to local buckling at elevated temperatures.
Since cold-formed seel members lose their strength significantly after 700oC, these
tests were limited to 700oC. As reported in Chapter 3, a significant change of yield
strength occurred after 300oC. Therefore tests were carried out at every 100oC
intervals for temperatures greater than 300oC up to 700oC. In addition to the above
temperatures, tests were carried out for some sections at 100oC and 200oC to
understand the local buckling behaviour at lower elevated temperatures. The same
cold-formed steel section types, thicknesses and grades tested at ambient temperature
were used for elevated temperature tests. Most fire design codes follow the method
of assessing the capacity at elevated temperatures by using the reduced mechanical
____________________________________________________________________
properties at elevated temperatures. Therefore, the steady state test method was used
in this research. In this method, the test specimen is heated up to the required
temperature and then loaded until failure. Failure is defined as the point when the
maximum axial is reached. Test specimens were heated using the electric furnace in
the QUT structural laboratory. A special loading set-up was made to fit inside the
furnace, and the specimens were then loaded using Tinius Olsen Testing Machine.
4.1.2.1 Preliminary Investigation and Selection of Test Specimens
All the selected specimens were analysed to find the elastic buckling loads and
buckling modes using CUFSM and ABAQUS (HKS, 2004) at pre-selected
temperatures using the reduced mechanical properties reported in Chapter 3.
Buckling plots of all the selected sections of each grade were plotted at ambient
temperature and pre-selected elevated temperatures to understand the variation of
elastic buckling loads with increasing temperatures (see Figure 4.9). The yield
strength of cold-formed steel deteriorates rapidly after 500oC. Hence it is small
beyond 600oC compared to the elastic modulus. This results in a lower squash load
and a higher elastic buckling load in this range. Therefore it is difficult to find the
specimens that show elastic buckling beyond 600oC.
Figure 4.9: Variation of Elastic Buckling Loads with Increasing

Temperatures for G550-0.95-B
____________________________________________________________________
Test specimens were labelled as for ambient temperature tests. External dimensions
and thicknesses of the specimens were measured using a vernier calliper and a
micrometer screw gauge (see Tables 4.7 and 4.8). Measured dimensions were close
to the nominal dimensions proposed to fold the test specimens except for Type A
specimens with 0.95 mm thickness. The flange size of Type A specimen was slightly
higher than the proposed dimension due to the difficulty in folding. As noted in
Tables 4.7 and 4.8 a total of 64 elevated temperature tests were conducted.
Table 4.7: Measured Dimensions of Type A Specimens
Web Flange Lip Length

Specimen
(mm) (mm) (mm) (mm)
G550 – 0.95 – 100 – A1 62.85 27.65 9.60 190.0
G550 – 0.95 – 200 – A1 63.40 27.75 9.65 190.0
G550 – 0.95 – 300 – A1 63.30 27.75 9.65 190.0
G550 – 0.95 – 300 – A2 63.25 27.65 9.55 190.0
G550 – 0.95 – 400 – A1 63.30 27.55 9.65 190.0
G550 – 0.95 – 400 – A2 63.35 27.67 9.70 190.0
G550 – 0.95 – 500 – A1 63.50 27.75 9.65 190.0
G550 – 0.95 – 600 – A1 62.85 27.80 9.65 190.0
G550 – 0.95 – 600 – A2 63.30 27.65 9.70 190.0
G550 – 0.95 – 700 – A1 63.55 27.55 9.65 190.0
G250 – 0.95 – 200 – A1 64.00 32.50 9.50 188.5
G250 – 0.95 – 200 – A2 63.50 33.00 9.50 190.0
G250 – 0.95 – 300 – A1 63.40 33.50 9.80 189.0
G250 – 0.95 – 300 – A2 63.60 33.50 9.75 189.0
G250 – 0.95 – 300 – A3 63.55 33.40 9.75 190.0
G250 – 0.95 – 400 – A1 65.15 32.75 9.80 190.0
G250 – 0.95 – 400 – A2 63.55 33.25 9.70 189.5
G250 – 0.95 – 500 – A1 63.80 33.50 9.80 189.0
G250 – 0.95 – 500 – A2 63.80 33.60 9.85 189.0
G250 – 0.95 – 600 – A1 63.55 33.50 9.70 190.0
G250 – 0.95 – 700 – A1 63.80 33.50 9.80 189.0
G450 – 1.9 – 300 – A1 99.70 40.05 15.10 280.0
G450 – 1.9 – 500 – A1 99.82 40.10 15.05 280.0
G450 – 1.9 – 500 – A2 100.05 40.05 15.10 280.0
G450 – 1.9 – 600 – A1 99.60 40.20 15.00 280.0
G450 – 1.9 – 700 – A1 100.10 40.05 15.10 280.0
G250 – 1.95 – 300 – A1 120.50 40.15 15.15 320.0
G250 – 1.95 – 400 – A1 120.50 40.15 15.15 320.0
G250 – 1.95 – 500 – A1 120.70 40.20 15.00 320.0
G250 – 1.95 – 600 – A1 120.05 40.10 15.05 320.0
G250 – 1.95 – 700 – A1 120.60 40.35 15.05 320.0
____________________________________________________________________
Table 4.8: Measured Dimensions of Type B Specimens
Specimen Web Flange Length

(mm) (mm) (mm)
G550 – 0.95 – 100 – B1 25.00 25.00 150
G550 – 0.95 – 200 – B1 25.05 25.00 150
G550 – 0.95 – 300 – B1 25.0 24.95 150
G550 – 0.95 – 300 – B2 24.95 24.90 150
G550 – 0.95 – 400 – B1 25.00 25.05 150
G550 – 0.95 – 400 – B2 25.05 24.95 150
G550 – 0.95 – 500 – B1 24.90 25.00 150
G550 – 0.95 – 600 – B1 25.00 25.05 150
G550 – 0.95 – 600 – B2 25.00 24.95 150
G550 – 0.95 – 700 – B1 25.00 25.05 150
G250 – 0.95 – 200 – B1 25.15 24.90 150
G250 – 0.95 – 200 – B2 25.00 25.00 150
G250 – 0.95 – 300 – B1 24.95 25.15 150
G250 – 0.95 – 300 – B2 25.00 25.20 150
G250 – 0.95 – 300 – B3 25.05 25.00 150
G250 – 0.95 – 400 – B1 25.05 24.95 150
G250 – 0.95 – 400 – B2 24.95 24.95 150
G250 – 0.95 – 500 – B1 24.85 24.90 150
G250 – 0.95 – 600 – B1 25.05 25.05 150
G250 – 0.95 – 700 – B1 25.00 25.10 150
G450 – 1.9 – 200 – B1 50.15 49.70 260
G450 – 1.9 – 300 – B1 50.05 50.05 260
G450 – 1.9 – 400 – B1 50.05 49.90 260
G450 – 1.9 – 400 – B2 50.00 50.10 260
G450 – 1.9 – 500 – B1 50.00 50.00 260
G450 – 1.9 – 500 – B2 50.35 49.95 260
G450 – 1.9 – 600 – B1 50.15 49.95 260
G450 – 1.9 – 700 – B1 50.00 50.00 260
G250 – 1.95 – 200 – B1 50.10 50.00 260
G250 – 1.95 – 300 – B1 50.50 50.00 260
G250 – 1.95 – 400 – B1 50.20 50.05 260
G250 – 1.95 – 400 – B2 50.15 50.00 260
G250 – 1.95 – 500 – B1 50.25 50.05 260
G250 – 1.95 – 500 – B2 50.05 49.95 260
G250 – 1.95 – 600 – B1 50.00 50.15 260
G250 – 1.95 – 700 – B1 50.15 50.10 260
____________________________________________________________________
4.1.2.2 Test Set-up and Procedure
Electric Furnace (Small)

The small electric furnace in the QUT Structural Laboratory was used to heat the
specimens (see Figure 4.10). Four glow bars fitted to the furnace heated the
specimens while a micro-computer based temperature control system controls the
pre-set temperature and heating rate. A maximum heating rate of up to 30oC/min can
be achieved with this furnace. However, it depends on the mass of steel and air
inside the furnace. Four glow bars emit radiative heating to the furnace contents.
Therefore the temperature at the centre of glow bars was slightly less than the pre-set
temperature. Hence the furnace temperature was set slightly higher than the required
temperature and a thermometer was located at the centre to monitor the specimen
temperature. In this way, the specimen temperature was measured with less than 1%
error at higher temperatures above 400oC and 2% error below 300oC. In most of the
cases, the specimen temperature during the loading phase was exactly the targeted
temperature. Two holes of 43 mm diameter were provided at the top and bottom of
the furnace to insert loading shafts. An observation window located on one side of
the furnace was used to observe the test specimen and its behaviour during loading.
Special openings on the furnace were used to insert an LVDT and a thermocouple to
the specimen. They were used to measure the out-of-plane deflection (buckling) and
the temperature of the specimen. There is another thermocouple permanently fixed to
the furnace. This thermocouple monitors the temperature inside the furnace and
communicates with the microcomputer system to control the furnace temperature.
Loading Arrangement
A loading arrangement was made using 253MA stainless steel, which gives better
performance at elevated temperatures up to 1100 oC. One end of a 40 mm diameter
circular bar of grade 253MA stainless steel was fitted to the 25 mm thick and 130
mm diameter plate of same grade stainless steel. Other end of the rod was threaded
to a length of 100 mm and was screwed into a rigid steel plate. A similar
arrangement was made to the top loading shaft. Bottom shaft was simply located on
the flat base of Tinius Olsen Testing Machine. Top loading shaft was fixed to the
cross head of the machine through a rigid plate. Both shafts were tightened using
locknuts. Special care was taken to align the loading shafts to avoid eccentric
____________________________________________________________________
loading. For this purpose, three 10 mm holes were drilled in both the top and bottom
loading plates at the same locations and three specially made steel rods of 10 mm
were fitted to the top loading plate. The top loading shaft with three rods was then
lowered so that these rods moved through the holes. The bottom loading shaft was
given some fine adjustments until the three small rods could move within the holes
without any friction (see Figure 4.11).
Thermo
couple to
Top rigid monitor the
plate and temperature
loading shaft inside the
furnace
Observation Micro
window computer
temperature
control
system
Bottom rigid Two phase

plate electricity
control
system
(a) Furnace and Loading Set-up
Figure 4.10: Electrical Furnace and its Attachments
____________________________________________________________________
(b) Temperature Control System (c) Glow Bars
(d) Hole to Insert LVDT (e) Hole to Insert Probe of Thermometer
Figure 4.10: Electrical Furnace and its Attachments
Figure 4.11: Aligning of Loading Shafts
____________________________________________________________________
To assess the accuracy of the loading set-up, four Type A specimens of 1.9 mm
G450 and 1.95 mm G250, which gave the highest buckling capacity at the ambient
temperature, were tested using this loading set-up attached to the furnace. They were
also tested without the loading set-up and the furnace. Comparison of their results
ensured the accuracy of the loading set-up (see Table 4.9). Axial shortening of the
specimen was taken by measuring the vertical movement of the machine cross-head.
In the case of higher axial load, axial shortening of the loading shafts need to be
avoided. Since the temperature of the loading shafts was not uniform at elevated
temperatures due to the covering of them by insulation, axial shortening of the shaft
can not be accurately estimated using the mechanical properties of stainless steel.
Therefore only the ultimate load was taken as the results. This effect on the lower
thickness specimens was insignificant.
Table 4.9: Comparison of Ultimate Loads using the Two Loading Set-ups
With loading shaft

Without loading shaft
arrangement inside the
Specimen arrangement (kN)
furnace (kN)
P1 P2 P3 P1 P2
G450 – 1.9 – 20 – A 150.67 153.23 149.99 148.57 149.22
G250 – 1.9 – 20 – A 105.55 100.88 99.71 100.30 101.57
Test Procedure
The test specimen was first placed between the loading plates. The LVDT was
inserted inside the furnace so that it was in contact with the plate element that
buckled. Location of the LVDT was marked and the relevant distances were
measured and noted. The furnace was programmed to the required temperature prior
to switching it on. Test specimen was heated to the required temperature and then
allowed another 20 minutes to ensure a uniform temperature throughout the
specimen prior to loading. During the heating phase the axial load on the specimen
was carefully monitored. Since the thermal expansion of the specimen and loading
shafts induce a compression load, the machine cross-head was raised very slowly so
that an initial load of 0.1kN was maintained. This initial load was applied to prevent
the lateral movement of the specimen when the LVDT was inserted to measure the
out-of-plane displacement. An appropriate load cell in the testing machine was
____________________________________________________________________
selected according to the expected failure load. It is important to avoid any loading
eccentricity to minimise the effect of unwanted moments. Therefore the loading axis
was made to coincide with the geometric centre of the section. The geometry of each
section was drawn on the plate of the bottom loading shaft so that the geometric
centre was located exactly at the centre. The specimen was then placed on the
corresponding geometry. Coating of the specimen burns at temperatures greater than
500oC. Therefore for the tests with temperatures greater than 500oC, the specimen
coating was removed by immersing the specimens in diluted hydrochloric acid for
about 45 minutes. These specimens were also used to measure their base metal
thicknesses. Exact temperature of the specimen was taken by using a portable
thermocouple. Probe of this thermocouple was inserted into the furnace until it was
in contact with the specimen. Finally the load was applied by lowering the upper
loading shaft that was attached to the Tinius cross-head. Figure 4.12 shows typical
failure patterns of Type A and B specimens at elevated temperatures.
(a)Failure Pattern of G550-0.95-700-A (Local Web Buckling)
(a) Failure Pattern of G550-0.95-700-B ( Local Flange Buckling)
Figure 4.12: Observed Local Buckling Failure at Elevated Temperatures
4.1.2.3 Results of Local Buckling Tests at Elevated Temperatures
Both axial compression load versus axial shortening and axial compression load
versus out of plane displacement were recorded and plotted. The maximum ultimate
____________________________________________________________________
load was noted from the display unit of the Testing Machine. In this chapter, only the
axial compression load versus axial shortening curves at elevated temperatures were
included for 0.95 mm thickness. This is because the axial shortening of the loading
shafts could not be ignored for the 1.9 mm thick specimens with a higher failure
load. However, a suitable correction could be made to eliminate this problem. Tables
4.10 and 4.11 give the ultimate loads of Type A and Type B specimens at elevated
temperatures. Figure 4.13 shows typical load-deflection curves for Type B specimen
at 300oC while Appendix B gives the other load-deflection curves.
30.00
25.00
20.00
Axial Load (kN)
15.00
10.00
5.00
0.00
0.00 0.50 1.00 1.50 2.00 2.50
(a) Axial Compression Load versus Axial Shortening

30
25
20
Axial Load (kN)
15
10
0
-4 -3 -2 -1 0
Out of Plane Deflection (mm)
(b) Axial Compression Load versus Out-of-Plane Deflection (flange)
Figure 4.13: Typical Load-Deflection Curves for G550-0.95-300-B

____________________________________________________________________
Table 4.10: Ultimate Loads of Type B Specimens at Elevated Temperatures
Ultimate load (kN)

Specimen
P1 P2 P3 Average
G550 – 0.95 – 200 – B 26.20 26.51 26.36
G550 – 0.95 – 300 – B 25.91 25.91
G550 – 0.95 – 400 – B 22.08 22.08
G550 – 0.95 – 500 – B 14.32 14.32
G550 – 0.95 – 600 – B 4.15 4.15
G550 – 0.95 – 700 – B 2.02 2.02
G250 – 0.95 – 200 – B 15.20 16.93 16.07
G250 – 0.95 – 300 – B 10.47 10.47 11.15 10.70
G250 – 0.95 – 400 – B 8.89 9.61 9.25
G250 – 0.95 – 500 – B 7.85 7.85
G250 – 0.95 – 600 – B 3.64 3.64
G250 – 0.95 – 700 – B 2.22 2.22
G450 – 1.9 – 200 – B 86.39 86.39
G450 – 1.9 – 300 – B 86.10 86.10
G450 – 1.9 – 400 – B 58.34 58.63 58.49
G450 – 1.9 – 500 – B 31.99 39.512 35.75
G450 – 1.9 – 600 – B 14.00 14.00
G450 – 1.9 – 700 – B 8.06 8.06
G250 – 1.95 – 200 – B 69.60 69.60
G250 – 1.95 – 300 – B 42.74 42.74
G250 – 1.95 – 400 – B 35.62 37.44 36.53
G250 – 1.95 – 500 – B 23.80 21.59 22.69
G250 – 1.95 – 600 – B 15.10 15.10
G250 – 1.95 – 700 – B 9.24 9.24
____________________________________________________________________
Table 4.11: Ultimate Loads of Type A Specimens at Elevated Temperatures
Ultimate load (kN)

Specimen
P1 P2 P3 Average
G550 – 0.95 – 100 – A 53.47 53.47
G550 – 0.95 – 200 – A 52.12 52.12
G550 – 0.95 – 300 – A 46.84 48.81 47.83
G550 – 0.95 – 400 – A 36.58 36.28 36.43
G550 – 0.95 – 500 – A 21.36 21.36
G550 – 0.95 – 600 – A 6.66 6.44 6.55
G550 – 0.95 – 700 – A 4.00 4.00
G250 – 0.95 – 200 – A 37.31 35.87 36.59
G250 – 0.95 – 300 – A 30.27 28.90 31.99 30.39
G250 – 0.95 – 400 – A 24.20 23.70 23.95
G250 – 0.95 – 500 – A 15.80 14.41 15.11
G250 – 0.95 – 600 – A 9.30 9.30
G250 – 0.95 – 700 – A 5.45 5.45
G250 – 1.9 – 300 – A 146.18 146.18
G250 – 1.9 – 500 – A 68.63 71.562 70.10
G250 – 1.9 – 600 – A 23.65 23.65
G250 – 1.9 – 700 – A 11.64 11.64
G250 – 1.95 – 300 – A 80.92 80.92
G250 – 1.95 – 400 – A 61.60 61.60
G250 – 1.95 – 500 – A 37.58 37.58
G250 – 1.95 – 600 – A 23.00 23.00
G250 – 1.95 – 700 – A 14.86 14.86
4.2 Comparison of Test Results with Predictions from the Current

Design Standards
4.2.1 Design Standards
Many cold-formed steel design codes, AS/NZS 4600 (SA, 2005), Eurocode 3: Part
1-3 (ECS, 1996), the North American Specification (NAS, 2003) and BS 5950 Part 5
(BSI, 1998), give design guidelines for cold-formed steel short columns subjected to
local buckling at ambient temperature. However, AS/NZS 4600 (SA, 2005),
Eurocode 3: Part 1-3 (ECS, 1996) and the North American Specification (NAS,
2003) give identical guidelines for local buckling. Therefore our test results were
compared with the predictions of AS/NZS 4600 (SA, 2005) and BS 5950 Part 5
____________________________________________________________________
(BSI, 1998), which recommend the use of the well known effective with method for
local buckling. Recently Schafer (2001) proposed a new method called the Direct
Strength Method (DSM) that eliminates the tedious calculations of effective with of
each plate element in the section. This method is now recommended as an alternative
method by AS/NZS 4600 and NAS. The major advantage of this method is its ability
to calculate the capacity of members subjected to local buckling interacting with
other buckling modes. However, the direct strength method needs accurate elastic
buckling loads based on a rational buckling analysis to calculate the ultimate
capacity. Although the above design codes were originally developed for calculating
the local buckling capacities at ambient temperature conditions, the same design
equations were used in this research to calculate the local buckling capacities at
elevated temperatures, simply by using the appropriately reduced mechanical
properties. The reduced mechanical properties for 0.95 mm thick cold-formed steel
were taken from Ranawaka and Mahendran (2009) while those for 1.9 mm and 1.95
mm thick cold-formed steels were taken from Dolamune Kankanamge (2009).
Eurocode 3 Part 1.2, AS/NZS 4600 and BS 5950 Part 5 use the effective width
method to find the local buckling capacity. However, the fire code, Eurocode 3 Part
1.2, recommends the use of effective width at ambient temperature with the yield
stress at the elevated temperature. Therefore Eurocode 3 Part 1.2 guidelines were
used with appropriately reduced yield stresses. Design recommendations of other
design standards were also used with reduced mechanical properties at elevated
temperatures. The direct strength method needs the relevant elastic buckling load
from a rational buckling analysis. Therefore finite element analyses were carried out
using ABAQUS to find the elastic buckling load of tested specimens. The specimen
length was taken as 2 or 3 times the buckling half wave length and hence interaction
with the flexural or flexural-torsional buckling modes was not observed. In addition
to the short length of specimens, the section dimensions were selected so that local
buckling of web and flange elements dominated the behaviour of test specimens.
Tables 4.12 and 4.13 compare the ultimate loads (capacity) from the tests and
various design standards. Sample calculations for each of these methods are given in
Appendix A. The mean and coefficient of variation (COV) for the ratios of test
results to the predicted results were also calculated based on the thicknesses, grades
and section types for each design standard.
____________________________________________________________________
Table 4.12: Comparison of Ultimate Loads from Tests and Design Standards
for Type A Sections
Average Test/Predicted ultimate load

Specimen Temp. test load
type (oC) (kN) EC3 P1.2 AS4600 BS 5950 P5 DSM
20 53.95 0.99 0.99 1.02 1.10
100 53.47 1.00 0.99 1.02 1.10
200 52.12 0.99 1.03 1.06 1.15
300 47.83 0.98 1.05 1.08 1.16
400 36.43 0.93 0.99 1.02 1.09
G550-
500 21.36 0.87 0.85 0.88 0.95
0.95-A
600 6.55 1.27 0.98 0.98 0.91
700 4.00 0.94 0.75 0.76 0.72
Mean 0.995 0.953 0.976 1.021
COV 0.121 0.104 0.108 0.149
0.837 0.820 0.836 0.823
20 39.09 1.02 1.02 1.05 1.08
200 36.59 1.15 1.16 1.19 1.22
300 30.39 1.32 1.26 1.30 1.32
400 23.95 1.43 1.35 1.38 1.37
G250- 500 15.11 1.26 1.17 1.19 1.17
0.95-A 600 9.30 1.16 1.07 1.08 1.05
700 5.45 1.15 1.05 1.06 1.02
Mean 1.212 1.153 1.178 1.177
COV 0.110 0.104 0.108 0.116
1.036 0.994 1.010 0.999
20 151.30 0.95 0.95 0.98 0.99
300 146.18 0.97 1.01 1.05 1.11
500 70.10 1.13 1.11 1.13 1.12
G450- 600 23.65 1.35 1.15 1.15 1.01
1.90-A 700 11.64 1.05 0.91 0.91 0.81
Mean 1.171 1.131 1.141 1.062
COV 0.091 0.091 0.091 0.094
1.012 0.978 0.986 0.914
20 102.04 1.05 1.05 1.07 1.02
300 76.92 1.21 1.19 1.20 1.14
400 61.60 1.34 1.29 1.30 1.21
500 37.58 1.15 1.09 1.10 1.02
G250-
600 23.00 1.07 1.01 1.01 0.93
1.95-A
700 14.86 1.22 1.15 1.15 1.06
Mean 1.089 1.028 1.044 1.007
COV 0.147 0.099 0.097 0.124
0.858 0.879 0.895 0.827
Note:- Test results of G550-0.95-A-700 appear to be not reliable as it is less than the
capacity of G250-0.95-A-700.
____________________________________________________________________
Table 4.13: Comparison of Ultimate Loads from Tests and Design Standards
for Type B Sections
Average Test/Predicted ultimate load

Specimen Temp.
test load EC3 BS5950
type (oC) AS 4600 DSM
(kN) P1.2 P5
20 26.50 1.06 1.06 1.12 0.94
200 26.36 1.08 1.13 1.19 1.01
300 25.91 1.15 1.23 1.30 1.10
400 22.08 1.22 1.30 1.37 1.16
G550- 500 14.32 1.27 1.24 1.32 1.11
0.95-B 600 4.15 1.75 1.28 1.31 1.01
700 2.02 1.03 0.80 0.83 0.65
Mean 1.223 1.147 1.206 0.998
COV 0.204 0.154 0.155 0.171
0.881 0.910 0.955 0.768
20 16.83 1.09 1.09 1.15 0.93
200 16.07 1.19 1.19 1.28 1.01
300 10.70 1.13 1.05 1.11 0.88
400 9.25 1.30 1.22 1.28 1.01
G250- 500 7.85 1.54 1.42 1.48 1.16
0.95-B 600 3.64 1.06 0.96 1.05 0.78
700 2.22 1.09 0.98 1.01 0.79
Mean 1.200 1.131 1.194 0.936
COV 0.140 0.141 0.135 0.144
0.981 0.923 0.984 0.760
20 96.89 1.11 1.11 1.19 1.00
200 86.39 1.00 1.06 1.13 0.95
300 86.10 1.05 1.16 1.23 1.04
400 58.49 0.94 1.01 1.08 0.93
G450- 500 35.75 1.00 1.05 1.12 0.91
1.90-B 600 14.00 1.48 1.24 1.29 0.93
700 8.06 1.40 1.24 1.31 0.88
Mean 1.141 1.125 1.193 0.947
COV 0.184 0.081 0.073 0.058
0.850 0.990 1.059 0.852
20 66.37 1.22 1.22 1.28 1.02
200 69.6 1.41 1.43 1.50 1.16
300 42.74 1.20 1.18 1.23 0.92
400 36.53 1.42 1.36 1.41 1.11
G250- 500 22.69 1.24 1.16 1.21 0.95
1.95-B 600 15.10 1.25 1.16 1.20 0.94
700 9.24 1.36 1.26 1.30 1.01
Mean 1.299 1.252 1.306 1.015
COV 0.071 0.082 0.086 0.090
1.157 1.103 1.145 0.886
Note:- Test results of G550-0.95-B-700 appear to be not reliable as it is less than the
capacity of G250-0.95-B-700.
____________________________________________________________________
Figures 4.14 to 4.20 show the comparison of test results for Type A and Type B
specimens with predicted capacities based on the design standards considered here.
These comparisons were plotted at ambient temperatures and different elevated
temperatures. Most of the results are on the safe side particularly at lower
temperatures while only a few results are on the unsafe side at higher elevated
temperatures.
160
140
SAFE
120
UNSAFE
100
Test Results (kN)
80
60
AS 4600-A BS 5950.5-A
40 DSM-A EC3 1.2-A
AS 4600-B BS 5950.5-B
20
DSM-B EC3 1.2-B
0
0 20 40 60 80 100 120 140 160
Predicted Capacities (kN)
Figure 4.14: Comparison of Test and Predicted Ultimate Loads at Ambient

Temperature
____________________________________________________________________
100
90
80 SAFE
70 UNSAFE
Test Results (kN)
60
50
40 AS 4600-A BS 5950.5-A
30 DSM-A EC3 1.2-A
20 AS 4600-B BS 5950.5-B
10 DSM-B EC3 1.2-B
0
0 10 20 30 40 50 60 70 80 90 100
Figure 4.15: Comparison of Test and Predicted Ultimate Loads at 200oC
150
125 SAFE
UNSAFE
100
Test Results (kN)
75
AS 4600-A BS 5950.5-A
50 DSM-A EC3 1.2-A
AS 4600-B BS5950-5
25
DSM-B EC3 1.2-B
0
0 25 50 75 100 125 150
Predicted capacities (kN)
____________________________________________________________________
80
70
60
SAFE
50
Test Results (kN)
UNSAFE
40
30 AS 4600-A BS 5950.5-A
DSM-A EC3 1.2-A

20
AS 4600-B BS 5950.5-B
10
DSM-B EC3 1.2-B
0
0 10 20 30 40 50 60 70 80
80
70
SAFE
60
UNSAFE
50
Test Results (kN)
40
AS 4600-A BS 5950.5-A
30
DSM-A EC3 1.2-A
20 AS 4600-B BS 5950.5-B
10 DSM-B EC3 1.2-B
0
0 10 20 30 40 50 60 70 80
____________________________________________________________________
30
25
SAFE
20
Test Results (kN)
UNSAFE
15
AS 4600-A BS 5950.5-A
10
DSM-A EC3 1.2-A
5 AS 4600-B BS 5950.5-B
DSM-B EC3 1.2-B

0
0 5 10 15 20 25 30
20
18
16
SAFE
14
UNSAFE
Test Results (kN)
12
10
8 AS 4600-A BS 5950.5-A
6 DSM-A EC3 1.2-A
4
AS 4600-B BS 5950.5-B
2
DSM-B EC3 1.2-B
0
0 2 4 6 8 10 12 14 16 18 20
____________________________________________________________________
4.2.2 Discussion of Results and Comparison
Eurocode 3 Part 1.2 Predictions

The mean values for Type A specimens are 0.995, 1.212, 1.171 and 1.089 while the
corresponding coefficients of variation are 0.121, 0.110, 0.091 and 0.147 (see Table
4.12). These mean values are close to one or greater than one while most of the
results are on the safe side. Therefore Eurocode 3 Part 1.2 can be used to predict the
local buckling capacities of cold-formed steel sections with stiffened elements at
elevated temperatures. However, the results are very conservative for higher
temperatures.
The mean values for Type B specimens are 1.223, 1.200, 1.141 and 1.299 while the
4.13). The mean values are always greater than one with most of the results on the
safe side. This implies that Eurocode 3 Part 1.2 method can be used to predict the
local buckling capacities of unstiffened elements at elevated temperature. However,
the predictions are not economical compared to the other design methods.
Eurocode 3 Part 1.2 predictions for higher grade steel specimens at 600oC are highly
conservative. The ratios of test over predictions are 1.27, 1.35, 1.75, and 1.48 for
specimens of G550-0.95-A, G450-1.9-A, G550-0.95-B and G450-1.9-B,
respectively. These higher values are due to the lesser capacity predictions. Eurocode
3 Part 1.2 assumes that the effective width is constant at elevated temperatures and is
equal to the effective width at ambient temperature. However, the effective width is
a function of the ratio of (E/fy)0.5 for a particular section. Since this ratio is very high
at 600oC, the effective width is also higher than that of ambient temperature (see
Figure 4.21). Therefore the capacities at 600oC predicted by European fire standard
are conservative in particular for higher strength steel grades.
____________________________________________________________________
1.2
1.0
0.8
0.5
(E/Fy) x1000
0.6
0.5
G550-0.95
0.4
G250-0.95
G250-1.95
0.2
G450-1.95
0.0
0 100 200 300 400 500 600 700 800
o
Temperature ( C)
Figure 4.21: Variation of (E/fy)0.5 with Temperature for Different Grades and
Thickness of Cold-formed Steel
AS/NZS 4600 Predictions

Tables 4.12 and 4.13 present the comparison of ultimate loads from the tests and
design standards at ambient and elevated temperatures. According to AS/NZS 4600
the mean values for Type A specimens are 0.953, 1.153, 1.131 and 1.028 while the
corresponding coefficients of variation are 0.104, 0.104, 0.091 and 0.099. Except for
G550-0.95 specimens, other mean values are greater than 1 and the coefficients of
variation are small for all cases. These figures imply that AS/NZS 4600 ambient
temperature guidelines can be safely used to predict the local buckling capacities of
cold-formed steel compression members with stiffened elements at elevated
temperatures.
corresponding coefficients of variations are 0.143, 0.131, 0.081 and 0.082. The mean
values are always higher than 1. This implies that the AS/NSZ 4600 design rules can
be safely used to predict the capacities of cold-formed steel columns with unstiffened
elements that are subjected to local buckling failures at elevated temperatures.
However, the predictions are more conservative than in the case of Type A
specimens with stiffened elements. Test to predicted ratios for different temperatures
____________________________________________________________________
in tables 4.12 and 4.13 appear to indicate that the rapid rise in (E/fy) ratio for high
strength steels beyond 500oC (see Figure 4.21) is influencing the comparisons, ie.
load ratios increase rapidly. This was also observed for predictions using BS5950
Part 5.
BS 5950 Predictions
Effective widths calculated from BS 5950 Part 5 are slightly less than those
calculated from other design standards such as AS/NZS 4600 (SA, 2005), Eurocode
3: Part 1-3 (ECS, 1996) and the North American Specification (NAS, 2003).
Therefore the predicted ultimate loads of BS 5950 Part 5 are less than those
calculated from other design standards. Mean values for Type A specimens are
0.976, 1.178, 1.141 and 1.044 while the corresponding coefficients of variations are
0.108, 0.108, 0.091 and 0.097 (see Table 4.12). Except for G550-0.95-A specimens,
the BS 5950 Part 5 predictions are more conservative. However, the predictions for
G550-0.95-A are more accurate and much closer to test results, and the associated
errors are less than 5% for most of the cases.
4.13 ). The mean values are always higher than one. This implies that the BS 5950
Part 5 design rules can be safely used to predict the capacities of cold-formed steel
columns with unstiffened elements subjected to local buckling at elevated
temperatures. However, the predictions are very conservative as observed with
AS/NZS 4600. It appears that using the same effective width equations for both
stiffened and unstiffened elements produces conservative predictions for unstiffened
elements. The k values for unstiffened elements are either 0.425 or 0.43. These k
values are based on an element that is simply supported on three sides with one free
edge. However, the presence of web element leads to a higher k value for the flanges
of unlipped channels (Yu, 2000). Since the effective width method does not consider
this effect, the code predictions are generally conservative.
Direct Strength Method

Mean values for Type A specimens are 1.021, 1.177, 1.062 and 1.007 while the
corresponding coefficients of variation are 0.149, 0.116, 0.094, and 0.124 (see Table
____________________________________________________________________
4.12). Except G250-0.95-A specimens, the direct strength method predictions are
accurate. However, the predictions for G250-0.95-A are more conservative.
Therefore the direct strength method can be used to predict the local buckling
capacities of cold-formed steel sections with stiffened elements at elevated
temperatures. In most cases, the direct strength method predictions are slightly lower
than the other design standards at lower temperatures up to 500oC. However, this
trend can not be seen with G250-0.95-A specimens. In this study, web elements of
lipped channels and the flange elements of unlipped channels are made slander while
all other elements are compact. Since the direct strength method discourages the
highly slender elements, the direct strength method may not be accurate enough to
predict the local buckling capacity.
The mean values for Type B specimens are 0.998, 0.936, 0.947and 1.015 while the
4.13). The mean values are always less than one. This implies that the direct strength
method predictions at elevated temperatures are slightly unsafe for cold-formed steel
members with unstiffened elements. The direct strength method predictions are
higher than those obtained from other design standards. The direct strength method
considers the elastic buckling load of the entire section instead of using the
conservative buckling coefficient values of individual elements (k = 0.43, 4) as with
the effective width based design method. This leads to higher column capacities and
thus more accurate predictions than the other design standards.
4.3 Capacity Reduction Factor
The American cold-formed steel structures code (AISI, 2007) recommends a

statistical model to determine the capacity reduction factors. This model accounts for
the variations in material, fabrication and the loading effects. The capacity reduction
factor is given by Equation 4.1.
− β 0 V m 2 +V f 2 + C p V p 2 + V q 2
Φ = 1.52 M m Fm Pm e 4.1
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
0 = 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
Vp and Pm values have to be determined from experiments or analyses. In this

investigation test ultimate loads were considered. Hence Vp and Pm are the mean and
coefficient of variation of the ratio of ultimate load from test to ultimate load
predicted by design standards. The substitution of all the above values leads to the
following equation.
− 2.5 0.0566 + C pV p 2
Φ = 1.672 Pm e 4.2
Equation 4.2 was used to determine the capacity reduction factors for the values
obtained from the AS/NZS 4600, BS 5950 Part 5, Eurocode 3 Part 1.2 and DSM.
The capacity reduction factors obtained for different grades of steels and thicknesses
are shown in Tables 4.12 and 4.13. AS/NZS 4600 recommends 0.85 for the
compression members. Most of the factors are greater than this limit. Therefore all
the design standards are reasonably accurate for the design of cold-formed steel
compression members subject to local buckling at elevated temperatures.
____________________________________________________________________
4.4 Conclusions
Local buckling behaviour and strength of cold-formed steel members was

investigated using tests at ambient and elevated temperatures. Cold-formed steel
sections with thinner (0.95 mm) and comparatively thicker (1.90 mm and 1.95 mm)
steels were selected with both high and low strength grades. Most commonly used
section types of lipped and plain channel sections were selected so that local
buckling was dominant in the web elements of lipped channels and flange elements
of unlipped channels. These buckling failures were ensured by preliminary
investigations carried out using CUFSM buckling analyses. Out of 27 ambient
temperature tests, 23 tests were carried out directly using the Tinius Olsen test
machine while 4 tests were conducted inside the furnace with loading shafts. A total
of 64 tests was carried out inside the furnace at elevated temperatures. The results
were compared with the predictions of different design standards. Due to the
identical design guidelines for local buckling in AS/NZS 4600 (SA, 2005), Eurocode
3: Part 1-3 (ECS, 1996) and North American Specification (NAS, 2003), only
AS/NZS 4600 (SA, 2005), BS 5950 Part 5 (BSI, 1998) and the direct strength
method recommendations were used in the capacity calculations. All these design
standards were developed for ambient temperature conditions. In this research, the
same guidelines were also used at elevated temperatures with appropriately reduced
mechanical properties. A comparative study undertaken in this research showed that
ambient temperature design guidelines can be safely used to predict the local
buckling capacities at elevated temperatures. Even the rapid relative variation of the
elastic modulus and yield stress at elevated temperatures beyond 500oC does not
appear to have affected the above observations. The direct strength method
predictions were more closer to the test results. However, it is slightly
unconservative for sections with unstiffened elements.
All of our observations are indirectly compared by the capacity reduction factors
calculated fro the chosen design methods in this study. The comparative study
showed that the elevated temperature guidelines of Eurocode 3 Part 1.2 (ECS, 2005)
gave conservative predictions than the other ambient temperature design guidelines.
Its predictions were very conservative at elevated temperatures beyond 500oC,
particularly for higher strength steels. Since Eurocode 3 Part 1.2 design method does
____________________________________________________________________
not account for the relative variation of elastic modulus and yield stress at elevated
temperatures, a significant drop of yield stress around 600oC gave reduced buckling
capacities. Other design methods (AS/NZS 4600, BS 5950 Part 5 and the direct
strength method) allow for this variation and hence a better correlation with test
results. However, since they were originally developed for steels with E/fy ratios at
ambient temperature, there is a sudden reduction in their predicted ultimate loads at
600oC (small compared with Eurocode 3 Part 1.2).
Eurocode 3 Part 1.2 limits the maximum temperature to 350oC for cold-formed steel
members. The ultimate load results from this research showed that cold-formed steel
members were able to retain more than 60% of ambient temperature strength at
350oC. This implies that the limit of 350oC is too conservative.
____________________________________________________________________
5 Development and Validation of Finite
Element Models of Cold-formed Steel
Compression Members Subject to Local
Buckling
Many numerical softwares are available to predict the structural behaviour of thin-
walled structural members subjected to different actions. Most of them use Finite
Element Analysis (FEA) and Finite Strip Analysis (FSA) techniques. These
numerical analyses predict accurate results provided the loading and boundary
conditions and relevant mechanical properties are modelled correctly. The major
advantage of these analyses is the elimination of extensive physical testing.
In this research, ABAQUS standard Version 6.7 was used in the finite element
analyses of cold-formed steel compression members subject to local buckling.
Convergence studies were carried out to find the suitable element size that was able
to simulate the behaviour observed in the experimental tests. MSC/PATRAN
software was used to create the input file as a pre-processor and the results were
observed by using it as a post-processor. Elastic buckling analyses (bifurcation
analyses) were first carried out to find the critical buckling load and associated
buckling modes of cold-formed steel compression members. Following this,
nonlinear analyses were carried out to obtain the ultimate load. Appropriate
imperfections were included to the most probable buckling modes to initiate the
nonlinear analyses. In most cases, the buckling mode corresponding to the lowest
Eigen value gave the critical buckling modes. Residual stresses were also included in
the nonlinear analyses.
Two kinds of finite element models, namely experimental and ideal models, were
developed. Experimental models were first developed to simulate the local buckling
tests and were validated using the load-deflection curves, deflected shape and
ultimate loads obtained from the tests reported in Chapter 4. Elevated temperature
tests were simulated by using the appropriate reduced mechanical properties. Ideal
models were then developed by simply changing the geometry and boundary
conditions to reflect ideal local buckling conditions. This chapter presents the details
_________________________________________________________________
of the finite element models developed to investigate the local buckling behaviour of
cold-formed steel compression members and their validation using the experimental
test results.
5.1 Development of Experimental Finite Element Models
5.1.1 Selection of Element Types
Shell type elements were selected for the simulations because cold-formed steel
members are made from thin steel sheets. Numerous shell element types are
available in the finite element simulation with ABAQUS. They are S4, S4R, S4R5,
S8R5 and S9R5. Element type S4 is a fully integrated, general-purpose shell element
available in ABAQUS standard. This element is recommended for accurate
simulation. However, S4 element needs more processing time and disk space.
Element types S4R5, S8R5 and S9R5 are thin shell elements with 4, 8 and 9 nodes,
respectively. These elements have reduced integration with five degrees of freedom.
All these element types were considered in the simulation of G550–0.95–20–A
specimen in order to select the most suitable element type. Both elastic buckling
analyses and nonlinear analyses were carried out with each element type. Suitable
geometric imperfections were applied to the lowest Eigen mode while residual
stresses were also included in the nonlinear analyses. Axial load versus axial
shortening curves for each element type were obtained. These curves were compared
with that obtained from the experimental test in Figure 5.1. it can be seen that all the
element types provided accurate simulations of the chosen test specimen up to the
ultimate load level. Higher order element types need more memory and processing
time while results are of the same order compared with four node element types.
Therefore selection had to be made from 4 nodes element types. S4 element type
needs more memory and processing time than the other element types with four
nodes. But the local buckling models are small and therefore the difference in
processing time is not significant. Memory requirement is not a critical issue for
small models because the latest computers are equipped with larger memory
capacities. Therefore S4 element type was selected in this research because it is a
fully integrated element with all degrees of freedom.
_________________________________________________________________
60
50
40
30
Test S4R5
20
S4R S4
S8R5 S9R5
10
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
Figure 5.1: Comparison of FEA Simulations Using Different Shell Element

Types with Test Results of G550–0.95–20–A
5.1.2 Selection of Element Size
Element size of the model is also an important factor in the finite element analyses.
Finer mesh gives more accurate results. However, it needs more memory and
processing time. Hence a very fine mesh does not give the most economical
simulation. Therefore a convergence analysis was carried out for G550–0.95–20–A
specimen with various mesh sizes from 1 mm x 1 mm to 9 mm x 9 mm in order to
find the optimum size of the finite element mesh. Figure 5.2 compares the effect of
element size on the elastic buckling and ultimate loads of G550-0.95-20-A specimen.
Analyses showed that the variation of ultimate load and elastic buckling loads
corresponding to the element sizes less than 3 mm x 3 mm mesh is negligible.
Therefore the element size was selected as 3 mm x 3 mm for the local buckling
simulations in this research.
_________________________________________________________________
60
Load (kN)
30
Ultimate Compression Load
Elastic Buckling Load
Optimum Element
Size
0
0 1 2 3 4 5 6 7 8 9 10
Element Size (mm)
Figure 5.2: Effect of Element Size on Elastic Buckling and Ultimate Loads of
G550-0.95-20-A Specimen
5.1.3 Loading and Boundary Conditions
In finite element analyses, it is important to use correct boundary conditions to

obtain accurate simulations. The boundary conditions should reflect the end supports
used in the experimental tests. In the experimental tests, both ends of the columns
were fixed against the rotation while the top end was allowed to move axially. These
end conditions allowed the specimens to fail symmetrically about the plane
perpendicular to the axis of the column at mid-height. Due to the symmetry of the
test specimen and loading (see Figure 5.3), it is economical to simulate one half of
the columns in the analyses. Therefore only one half of the column length was
simulated.
_________________________________________________________________
End support
Full column model Half column model
Figure 5.3: Experimental Models Used in Finite Element Analyses
In the experimental tests, an axial compression load was applied by lowering the
cross-head of the testing machine on to the upper end of the specimen while keeping
the lower end stationary on the base. Therefore only axial translation was allowed to
the nodes at upper end of the finite element model. These boundary conditions were
applied to the independent node of the rigid fixed MPC (Multi Point Constraint)
located at the upper end of the model. Rigid fixed MPC consists of an independent
node and any number of dependent nodes. Dependent nodes are connected to the
independent node using rigid beams and all six structural degrees of freedoms are
rigidly attached to each other. In this model, the independent node was located at the
geometric centre of the cross-section. Since the element size was selected as 3 mm x
3 mm, the distance between the dependent nodes was small. Hence this MPC acted
as a rigid surface that was rigidly connected to the upper end of the columns.
Out of plane translation was allowed at mid-height of the column model to simulate
the out of plane displacements of the buckling waves of the half column models. In
addition to the out of plane displacement, rotation about Z axis was also allowed at
mid-height even though the column was unlikely to twist. These boundary
conditions were applied to the perimeter nodes at the end surface of the model. The
load was applied to the independent node at the geometric centre of the rigid fixed
_________________________________________________________________
MPC. Due to small element size the distance between the dependent nodes was also
small. Therefore upper end (see Figure 5.4) acted as rigid plate on top of the column.
Half column model Boundary conditions applied to the MPC

at the upper end
Boundary condition applied to the nodes at the middle
(a) Type A Specimen
Figure 5.4 Loading and Boundary Conditions for Experimental Model
_________________________________________________________________
Half column model Boundary conditions applied
to the MPC at the upper end
Boundary condition applied to the nodes at the middle
(b) Type B Specimen
Figure 5.4: Loading and Boundary Conditions for Experimental Model
Imperfections of light gauge cold-formed steel members play a vital role in finite
element analyses. For the nonlinear analyses, it is important to include the
imperfections to initiate the appropriate buckling deformations. The value of
imperfection compared to the thickness is higher in cold-formed steel members due
to their small thicknesses. Therefore the ultimate load is significantly affected by the
imperfections. There are two types of imperfection, namely local imperfection and
global imperfection, both of which can be observed with cold-formed steel members.
Local imperfections affect the local or distortional buckling (section) capacities
while global imperfections mostly affect the overall buckling (member) capacity.
_________________________________________________________________
Local imperfections are expressed in terms of section dimensions and thickness
while global imperfections are expressed in terms of the member length.
Many researchers used the measured imperfections in their finite element

simulations (Ranawaka, 2006, Sivakumaran and Abdel-Rahman, 1998, Dubina and
Ungureanu, 2002). Some researchers used the section thickness as the value of
imperfection for the lipped channels made of thin cold-formed steels (Feng et al.,
2003, Chen and Young, 2006a). Initial imperfection of stiffened elements subjected
to local buckling is usually taken as a function of the web width. For example,
Outinen and Myllymaki (1995) and Feng et al. (2004) used h/200, where h is the
width of the stiffened element. Schafer and Pekoz (1998) recommended the initial
imperfection of web elements of lipped channel members to betaken as a function of
the width of web (w), 0.006w, or an exponential function based on the thickness (t)
of the web as 6te-2t. Many researchers including Schafer and Pekoz (1998) consider
the thickness as the imperfection for plain channels in the case of thinner sections
(Lee, 2004, Ranawaka, 2006, Schafer and Pekoz, 1998). After detailed
considerations of all the above, the imperfection models proposed by Schafer and
Pekoz (1998) for lipped and plain channels were used because they gave closer
predictions to the experimental ultimate loads.
Two types of residual stresses, namely flexural residual stress and membrane
residual stress, are present in steel members. The membrane residual stress is
negligibly small in cold-formed steel members as it does not involve any heat
treating or welding in the manufacturing process. However, the flexural residual
stress is significant as it involves folding of plate elements in the manufacturing
process. It is higher in the corners of the section due to folding. Schafer and Pekoz
(1998) proposed suitable flexural residual stress models for both roll-formed sections
and press-braked sections (see Figure 5.5) as a percentage of yield stress. They were
modified and used in this research. The residual stress models of Schafer and Pekoz
(1998) have higher residual stresses at the corners. But the press-braked specimens
made for this research have sharp corners, and the corner radius was negligibly
small. Therefore the corner regions with higher residual stresses due to press braking
_________________________________________________________________
can be neglected. Flexural residual stresses in the flanges were also assumed to be
the same as that of web for lipped channels because the amount of cold working
involved is considered to be the same in the manufacturing process. Figure 5.6
shows the proposed residual stress model used in this research.
Figure 5.5: Residual Stress Models Proposed by Schafer and Pekoz (1998)
17%
8%
8%
17%
17%
8%
8%
17%
Type A Type B
Figure 5.6: Proposed Residual Stress Models for this Research
The residual stresses reduce with increasing temperature (Lee, 2004). Therefore
reduced residual stresses were used in the analyses. Appropriate reduction factors for
residual stresses were proposed by Lee (2004) as given by Equation 5.1.
_________________________________________________________________
α = 1.0181 − 0.00128T (5.1)
Where,
α is the residual stress reduction factor
T is the temperature in oC and 20 ≤T ≤ 800
The residual stresses were modelled using the ABAQUS*INITIAL CONDITIONS

option with TYPE=STRESS, USER. The user defined initial stresses were created
using the SIGINI, FORTRAN user subroutine. It was necessary to ensure that the
coordinate system for stress components was correctly defined to produce the
residual stresses on the required axis. This was achieved by defining the variables of
SIGNI for each axis. For this purpose, the global coordinate system was used to
define the stress values on each element. This subroutine defines the variation of the
residual stress through the thickness. Five interpolation points were used to vary the
residual stress through the thickness. Figure 5.7 shows the developed residual stress
model of Type A and Type B specimens with 0.95mm thickness.
(a) Type A Section
Figure 5.7: Residual Stress Model for the 0.95mm G550 Cold-formed Steel
Section at Ambient Temperature
_________________________________________________________________
(b) Type B Specimen
Figure 5.7: Residual Stress Model for the 0.95mm G550 Cold-formed Steel
Section at Ambient Temperature
5.1.6 Mechanical Property Model
Mechanical property is an important parameter in finite element analyses.

Mechanical properties are characterized by elastic modulus, stress-strain relationship
and Poisson’s ratio. There are two stress-strain models (see Figure 5.8), namely
elastic-perfect plastic and strain hardening models used in finite element analyses.
The first model assumes a constant yield stress in the plastic range whereas the
second model includes strain hardening in the inelastic range. Mechanical properties
are taken from the tensile coupon tests. The stress-strain curves from the tensile
coupon tests show the presence of strain hardening behaviour. Therefore accurate
results could be obtained by using the more accurate strain hardening model. The use
of strain hardening model becomes more important for the elevated temperature
simulations because the stress-strain curves at elevated temperatures do not show a
clear elastic region.
_________________________________________________________________
400
300
Stress (MPa)
200
Elastic-Perfect Plastic Model
100 Strain Hardening Model
0
0 0.002 0.004 0.006 0.008 0.01
Strain
Figure 5.8: Stress-Strain Models
The use of elastic-perfect plastic model is simple. Some researchers use this model
for the ambient temperature simulations. In order to compare the mechanical
property models, local buckling behaviour of G550-0.95-20-A and G550-0.95-500-A
specimens was simulated based on elastic-perfect plastic and strain hardening
models. Since a clear yielding region was not observed in the stress-strain curves,
0.2% proof stress was considered as the yield stress. Tensile coupon tests give the
nominal stress-strain curves. ABAQUS require stress strain relationship in terms of
true stress and logarithmic plastic strain. True stress and logarithmic plastic strain
values of selected points on a stress-strain curve were obtained from Equations 5.2
(a) and (b) using the nominal stress and strain values of those points.
σ true = σ nom (1 + ε nom ) (5.2 a)
σ true
ε ln pl = ln (1 + ε nom ) − (5.2 b)
E
Elastic buckling analyses were carried out first, which were followed by non-linear
analyses based on the two mechanical property models. S4 element type was used in
the simulations. Geometric imperfections and residual stresses were also included in
the analyses. The axial compression load versus axial shortening curves obtained
_________________________________________________________________
from these analyses were compared with the load-deflection curves from the
experimental test (see Figure 5.9). Both mechanical property models produced a
good agreement with the load-shortening curve from the ambient temperature test. In
the case of elevated temperature simulations at 500oC, the use of strain hardening
model gave a better agreement with the test result. In this case (500oC) experimental
test curve includes the shortening of the test specimen and the top and bottom
loading shafts. Since the mechanical properties of loading shafts are not available at
elevated temperatures, a suitable correction to the test curve was not included. If this
correction was made, the test curve would agree closely with the simulated curves
based on the strain hardening mechanical property model. Therefore it can be
concluded that strain hardening mechanical property model must be used to simulate
the local buckling behaviour of cold-formed steel columns, particularly at elevated
temperatures.
60
50
40
30
20 Test
Elastic Perfect Plastic
10 Strain Hardening
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
(a) At Ambient Temperature
Figure 5.9: Load-Shortening Curves of G550-0.95-A Specimen for Different

Mechanical Property Models
_________________________________________________________________
30
25
20
15
10 TEST
Elastic Perfect Plastic

5
Strain Hardening
0
0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40
(b) At 500oC
Figure 5.9: Load-Shortening Curves of G550-0.95-A Specimen for Different

Mechanical Property Models
Appropriate mechanical properties for G550-0.95 mm and G250-0.95 mm thick

cold-formed steels were obtained from Ranawaka and Mahendran (2009) while those
for G450-1.90 mm and G250-1.95 mm thick cold-formed steels were obtained from
Dolamune Kankanamge (2009). These researchers developed suitable formulae for
yield strength and elastic modulus reduction factors at elevated temperatures. The
predictive equations proposed by Ranawaka and Mahendran (2009) for the reduced
mechanical properties at elevated temperatures are given in Equations 5.3 to 5.5.
They were developed based on the tensile coupon tests carried out on cold-formed
steels up to 0.95 mm thickness.
Yield strength reduction factors for G550 -0.95 mm thick cold-formed steels,
f yT
= −0.00016T + 1.0003 20 o C ≤ T ≤ 200 o C (5.3 a)
f y 20
f yT
= 0.97 −
(T − 200)1.81 200 o C < T < 600 o C (5.3 b)
f y 20 58500
_________________________________________________________________
f yT
= −0.00037T + 0.3363 600 o C ≤ T ≤ 800 o C (5.3 c)
f y 20
f yT
= −0.0007T + 1.014 20 o C ≤ T ≤ 200 o C (5.4 a)
f y 20
f yT
= 3 .7 −
(T − 74)0.15 200 o C < T ≤ 800 o C (5.4 b)
f y 20 0.736
Elasticity modulus reduction factors for G550 and G250 -0.95 mm thick cold-
formed steels,
ET
=1 20 o C ≤ T ≤ 100 o C (5.5 a)
E 20
ET
= −0.0013T + 1.1297 100 o C < T ≤ 800 o C (5.5 b)
E 20
The predictive equations proposed by Dolamune Kankanamge (2009) for the

reduction mechanical properties at elevated temperatures are given by Equations 5.6
to 5.8. They were developed based on the tensile coupon tests carried out on G450-
1.90 mm, G450-1.50 mm, G250-1.5 mm and G250-1.95 mm cold-formed steels.
f y ,T
= 1−
(T − 20)
4.56
20 ≤ T < 300 o C (5.6 a)
10
f y , 20 1x10 T
f y ,T
= 0.95 −
(T − 300)
1.45
300 ≤ T < 600 o C (5.6 b)
f y , 20 7.76T
f y ,T
= −0.0004T + 0.35 600 ≤ T ≤ 800 o C (5.6 c)
f y , 20
f y ,T
= −0.0005T + 1.01 20 ≤ T ≤ 200 o C (5.7 a)
f y , 20
_________________________________________________________________
f y ,T
f y , 20
(
= 25 1.16 − T 0.022 ) 200 < T ≤ 800 o C (5.7 b)
Elasticity modulus reduction factors for G450-1.90 mm and G250 -1.95 mm

thick cold-formed steels,
ET
= −0.000833T + 1.0167 20 ≤ T ≤ 200 o C (5.8 a)
E 20
ET
= −0.00135T + 1.1201 200 < T ≤ 800 o C (5.8 b)
E 20
These equations for reduction factors need the mechanical properties at the ambient
temperature to calculate the reduced mechanical properties at elevated temperatures.
Ambient temperature mechanical properties of G550-0.95 mm and G250-0.95 mm
thick cold-formed steels were obtained from Ranawaka and Mahendran (2009) while
those for G450-1.90 mm and G250-1.95 mm were obtained from Dolamune
Kankanamge (2009). Table 5.1 gives the ambient temperature mechanical properties
for the selected steel thicknesses and strength grades. Equations 5.3 to 5.8 and Table
5.1 values were used to calculate the required mechanical properties at elevated
temperatures.
Table 5.1: Mechanical Properties of the Selected Cold-Formed Steels
Steel fy,20 (MPa) E20 (MPa)

G550-0.95 615 205000
G250-0.95 320 200000
G450-1.90 515 206000
G250-1.95 271 188000
The Ramberg Osgood stress-strain model and appropriate parameters proposed by

Ranawaka and Mahendran (2009) and Dolamune Kankanamge (2009) are given by
Equations 5.9 and 5.10. Ranawaka and Mahendran (2009) proposed 0.86 for β for
both high and low strength cold-formed steels while Dolamune Kankanamge (2009)
proposed the same value for high strength cold-formed steels, but 1.5 for low
strength cold-formed steels.
_________________________________________________________________
ηT
f f y ,T fT
εT = T + β (5.9)
ET ET f y ,T
For G550 steel

η T = −3.05 X 10 −7 T 3 + 0.0005T 2 − 0.2615T + 62.653 20 o C ≤ T ≤ 800 o C (5.10 a)
For G250 steel
η T = 0.000138T 2 − 0.085468T + 19.212 350 o C ≤ T ≤ 800 o C (5.10 b)
Elevated temperature mechanical properties (fyT and ET) calculated from Equations
5.3 to 5.8 were used in the above equations to derive the stress-strain curves at the
required temperatures for use in finite element analyses. The Ramberg Osgood
stress-strain model can not be used for the low strength cold-formed steels ar lower
temperatures because they show a yield pattern in the stress-strain curves. Therefore
the elastic-perfect plastic mechanical property model was used in the finite element
analyses of cold-formed steel columns at lower temperatures up to and including
300oC.
5.1.7 Analysis Methods
As the initial approach, finite strip analyses using CUFSM were carried out to
determine the elastic buckling loads and buckling half wave lengths of all the trial
sections. The selected specimens were analysed with using ABAQUS to determine
the elastic buckling and ultimate loads. MSC/PATRAN was used as a pre-processor
to create the input files and as a post-processor to read the results. Two types of
analyses, the bifurcation buckling analysis to determine the elastic buckling loads
and modes and the nonlinear analysis to determine the ultimate loads and
deformations, were carried out using ABAQUS. Nonlinear analyses were carried out
using the modified Riks method to find the ultimate compression load. In the
nonlinear analyses, the maximum load increment was controlled appropriately to
obtain smooth loading and accurate load factor at failure. Initial imperfections as
discussed in Section 5.1.5 were included into the appropriate eigen buckling mode
from the elastic buckling analyses. In order to improve the accuracy of the analyses,
residual stresses were also included in the nonlinear analyses.
_________________________________________________________________
5.2 Validation of Finite Element Models
It is important to validate the finite element model developed to simulate the local
buckling behaviour. For this purpose, a total of eight finite element models of the
tested specimens as discussed in Chapter 4 were simulated at ambient temperature
and different elevated temperatures. These eight models consisted of three
thicknesses (0.95 mm 1.90 mm and 1.95 mm), two grades (low and high strength
steels) and two section types (stiffened and unstiffened). They were validated using
the load-shortening curves, deflected shapes and ultimate loads.
5.2.1 Validation of FEA Models Using Load-Shortening Curves
Axial compression load versus axial shortening curves for the tests were obtained
from the experimental tests. Same curves ware also obtained from the finite element
analyses. The load-shortening curves from both methods showed good agreement.
Figures 5.10 (a) to (d) show the load-shortening curves obtained for some cases.
Therefore it can be concluded that developed models accurately simulate the local
buckling behaviour of cold-formed steel short columns.
30
25
20
15
TEST
10
FEA
0
0.00 0.40 0.80 1.20 1.60 2.00
(a) G250-1.95-600-A
Figure 5.10: Comparison of Load-Shortening Curves from FEA and

Experimental Tests
_________________________________________________________________
15
10
5
TEST
FEA
0
0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40
(b) G250-0.95-300-B
60
50
40
30
20 Test
FEA
10
0
0 0.2 0.4 0.6 0.8 1 1.2 1.4
(c) G550-0.95-20-A

Experimental Tests
_________________________________________________________________
150
100
TEST
50
FEA
0
0.00 0.50 1.00 1.50 2.00 2.50 3.00
(d) G450-1.90-300-A

Experimental Tests
5.2.2 Validation of FEA Models Using Ultimate Loads
Both the elastic buckling load and the ultimate load were obtained from the analyses.
However, as described in Chapter 4, the elastic buckling load could not be measured
accurately during testing. Therefore the ultimate load was used to validate the finite
element models.
In the case of Type A specimens, imperfection direction (sign) affects the ultimate
load. In finite element analyses, it is assumed that imperfection follows the pattern of
local buckling waves. Bifurcation buckling analyses give the elastic buckling loads
and associated buckling modes. Positive imperfection follows the direction of
corresponding eigen mode. The direction of negative imperfection is opposite to the
corresponding eigen mode. Therefore negative imperfection reverses the buckling
pattern of corresponding eigen mode. The ultimate load for the positive imperfection
case is higher than that for the negative imperfections. These differences can be seen
_________________________________________________________________
only with stiffened elements (Type A specimens). Tested specimens also showed the
failure mode with positive imperfection. Figure 5.11 shows the comparison of failure
modes of Type A specimen with the direction of imperfection. Table 5.2 shows the
comparison of the ultimate loads for Type A specimens from tests and finite element
analyses for both negative and positive imperfections. Table 5.3 summaries the mean
values and associated coefficients of variation for the ratio of ultimate load from test
to finite element analyses for both positive and negative imperfections of Type A
sections at different elevated temperatures. The ultimate loads for positive
imperfections show that they give more accurate predictions than the negative
imperfections. Test results of G250-0.95-A specimens showed higher results than all
other predictions. This may be due to the use of steel sheets from a different batch of
steel in the making of specimens.
(a) G550-0.95-700-A Negative Imperfection
(b) G550-0.95-700-A Positive Imperfection
Figure 5.11: Comparison of Failure Modes as a Function of Imperfection

Direction in a Tested Specimen
_________________________________________________________________
(c) Failed G550-0.95-700-A Specimen
(d) Failed G250-1.95-300-A Specimen
Figure 5.11: Comparison of Failure Modes as a Function of Imperfection

Direction in a Tested Specimen
_________________________________________________________________
Table 5.2: Finite Element Analysis Results for Type A Specimens
Test FEA Results (kN) Test/FEA

Specimen Temp Ultimate Load
Results Ultimate Load
Type (oC) EB
(kN) (+)Imp. (-) Imp. (+)Imp. (-) Imp.
20 53.95 31.66 56.80 55.30 0.95 0.976
100 53.47 31.66 55.70 54.30 0.96 0.985
200 52.12 27.54 52.70 51.10 0.99 1.020
300 47.83 23.42 46.60 45.60 1.03 1.049
G550- 400 36.43 19.31 36.90 36.80 0.99 0.990
0.95-A 500 21.36 15.19 24.80 24.20 0.86 0.883
600 6.55 11.07 6.36 6.24 1.03 1.050
700 4.00 6.95 5.10 4.99 0.78 0.802
Mean 0.948 0.969
COV 0.090 0.089
20 37.09 33.02 36.20 35.60 1.02 1.042
200 36.59 28.73 31.70 31.20 1.15 1.173
300 30.39 24.44 23.10 23.20 1.32 1.310
400 23.95 20.14 17.00 16.90 1.41 1.417
G250-
500 15.11 15.85 12.30 12.00 1.23 1.259
0.95-A
600 9.30 11.56 8.22 8.10 1.13 1.148
700 5.45 7.26 4.95 4.88 1.10 1.117
Mean 1.195 1.209
COV 0.111 0.105
20 151.30 152.50 155.00 150.00 0.98 1.009
300 146.18 109.05 138.00 133.00 1.06 1.099
500 70.10 67.88 60.70 58.70 1.15 1.194
G450-
600 23.65 47.29 21.10 20.60 1.12 1.148
1.90-A
700 11.64 26.70 12.20 11.90 0.95 0.978
Mean 1.053 1.086
COV 0.083 0.084
20 102.04 119.48 96.10 95.90 1.06 1.064
300 82.92 85.44 64.70 63.70 1.19 1.207
400 61.60 69.31 45.80 43.30 1.35 1.423
G250- 500 37.58 53.18 32.10 30.80 1.17 1.220
1.95-A 600 23.00 37.05 21.40 20.80 1.07 1.106
700 14.86 20.92 12.00 11.70 1.24 1.270
Mean 1.180 1.215
COV 0.090 0.105
Imp. =Imperfection EB=Elastic Buckling Load
_________________________________________________________________
Table 5.3: Effects of Positive and Negative Imperfections on the Ultimate Loads
of Type A Sections
Positive imperfection Negative imperfection

Specimen type
Mean COV Mean COV
G550-0.95-A 0.948 0.090 0.969 0.089
G250-0.95-A 1.195 0.111 1.209 0.105
G450-1.90-A 1.053 0.083 1.086 0.084
G250-1.95-A 1.180 0.090 1.215 0.105
Unlike Type A sections, the ultimate load of Type B sections is not sensitive to the
imperfection direction. In most cases it followed the failure pattern of positive
imperfection. Figure 5.12 compares the failure mode of a Type B specimen from
experimental tests and finite element analyses. However, it is sensitive to the
magnitude of imperfection. The imperfection model used in this research was
proposed by Schafer and Pekoz (1998). It proposes thickness (t) as the imperfection
for Type B specimens. To understand the imperfection sensitivity, two series of
analyses were undertaken assuming the imperfection values of 0.1t and 1t. Table 5.4
shows the comparison of results with these two imperfection values. It shows that an
imperfection of ‘t’ is accurate enough in most cases and the results are on the safe
side.
(a) Half Column Model (b)Failed Specimen
Figure 5.12: Comparison of Failure Modes from Finite Element Analyses and
Experimental Tests of Type B Specimens
_________________________________________________________________
Table 5.4: Finite Element Analysis Results for Type B Specimens
FEA Results (kN) Test/FEA

Test
Specimen Temp. Ultimate Load
Results
Type (oC) EB Imp.
(kN) Imp. t Imp. 0.1t Imp. t
0.1t
20 26.50 20.40 28.40 29.10 0.93 0.91
200 26.36 17.65 26.70 26.90 0.99 0.98
300 25.91 15.01 23.70 23.70 1.09 1.09
400 22.08 12.37 18.80 19.30 1.17 1.14
G550-
500 14.32 9.73 12.60 12.90 1.14 1.11
0.95-B
600 4.15 7.10 3.22 3.67 1.29 1.13
700 2.02 4.45 2.58 2.89 0.78 0.70
Mean 1.056 1.010
COV 0.160 0.161
20 16.83 19.80 16.60 18.10 1.01 0.93
200 16.07 17.22 14.60 15.70 1.10 1.02
300 10.70 14.65 10.80 12.10 0.99 0.88
400 9.25 12.08 8.20 8.42 1.13 1.10
G250-
500 7.85 9.50 5.83 6.07 1.35 1.29
0.95-B
600 3.64 6.93 3.81 4.28 0.96 0.85
700 2.22 4.35 2.29 2.59 0.97 0.86
Mean 1.072 0.991
COV 0.128 0.163
20 96.89 74.47 96.50 98.80 1.00 0.98
200 86.39 63.30 92.50 93.20 0.93 0.93
300 86.10 53.25 85.60 85.60 1.01 1.01
400 58.49 43.20 66.00 64.30 0.89 0.91
G450-
500 35.75 33.15 37.10 37.80 0.96 0.95
1.90-B
600 14.00 23.09 13.20 14.60 1.06 0.96
700 8.06 13.04 7.61 8.47 1.06 0.95
Mean 0.987 0.954
COV 0.065 0.034
20 66.37 75.79 58.20 65.20 1.14 1.02
200 69.60 70.75 52.80 58.70 1.32 1.19
300 42.74 64.42 38.60 44.10 1.11 0.97
400 36.53 43.97 28.20 28.80 1.30 1.27
G250-
500 22.69 33.38 19.60 20.80 1.16 1.09
1.95-B
600 15.10 23.50 13.00 14.50 1.16 1.04
700 9.24 13.27 7.25 8.25 1.27 1.12
Mean 1.208 1.099
COV 0.071 0.094
Imp. = Imperfection EB=Elastic Buckling Load
_________________________________________________________________
5.2.3 Validation of FEA Models Using Failure Modes
Accuracy of the developed models was finally assessed by comparing the failure
modes from the finite element analyses and test specimens. Figures 5.11 and 5.12
show the comparison of failure modes for Type A and Type B specimens. They
showed that the failure modes from FEA and experimental tests were in good
agreement. By considering the ultimate loads, load-shortening curves and failure
modes, it can be concluded that the developed finite element models for local
buckling are accurate. Hence these models can be used in the parametric study to
investigate the effects of different parameters such as width to thickness of plate
elements, end conditions, etc. on the local buckling behaviour of cold-formed steel
compression members at ambient and elevated temperatures.
5.3 Comparison of FEA Results with Code Predictions
In this study, experimental models were developed to simulate the tests carried out to
investigate the local buckling behaviour whereas ideal models were developed to
carryout the parametric study. To assess the accuracy of design rules, the code
predictions for the tested columns were calculated. The ultimate load obtained from
the finite element analyses were then compared with the code predictions. This kind
of comparison was made for the test results in Chapter 4. However, this comparison
is more accurate because of the finite element analyses are more reliable compared
to the test results. Tables 5.7 and 5.8 show the comparison of finite element results
with the code predictions. For each case the mean value, associated coefficient of
variation and the capacity reduction factors were estimated. These parameters were
then used to assess the accuracy of the code predictions for the tested specimens.
The ratios of finite element analyses results to the code predictions showed a good
consistency compared to the comparison of test results reported in Chapter 4. The
mean values of AS/NZS 4600 and BS 5950 Part 5 predictions for Type A are very
close to one while the associated coefficients of variation are small. Further, the
factors are also greater than the acceptable limit of 0.85 recommended by AS/NZS
_________________________________________________________________
4600. The direct strength method predictions are also reasonably accurate. However
its ratio of finite element analyses to code predictions decrease with the increasing
temperature. The direct strength method discourages the use of very slender web
elements in channel (Schafer 2002). In other word, the sections with very low elastic
buckling load. In this study, a very slender web element was included to observe the
local buckling in web. Therefore the direct strength method predictions may not be
accurate.
At elevated temperature the yield stress reduces the higher rate than the elastic
modulus. In the case of higher strength steels, the ratio of elastic modulus to yield
stress increase rapidly beyond 500oC. Therefore the slenderness of the plate elements
decreases beyond this temperature resulting a higher effective width. Since Eurocode
3 Part 1.2 recommends the use of ambient temperature effective area, the beneficial
effect of its increment is not accounted in the predictions. Therefore Eurocode 3 Part
1.2 predictions are conservative at higher temperatures, particularly for higher
strength steels beyond 500oC.
As observed in the Type A sections, AS/NZS 4600 and BS 5950 Part 5 predictions
showed a good agreement with the finite element analyses. In addition, the ratio of
finite element analyses to code predictions showed a good consistency with these
two codes. The direct strength method predictions for the Type B sections showed
reasonable predictions. But, the associated factors are less than the acceptable
limit of 0.85. Also there is a trend of decreasing the ratio of finite element analyses
to the direct strength method predictions with increasing temperature, particularly for
higher strength steels. Since the selected Type b sections contained a very slender
flange, the direct strength method may not be able to accurately predict the local
buckling capacity. As observed in Type A sections, Eurocode 3 Part 1.2 predictions
are very conservative, particularly for higher strength steel at elevated temperatures.
_________________________________________________________________
Table 5.5: Comparison Code Predictions with FEA Results for
Type A-Experimental Modelling
FEA FEA/Predicted ultimate load

Specimen Temp.
(kN) AS/NZS BS 5950 EC 3
type (oC) DSM
4600 P5 P1.2
20 56.80 1.04 1.07 1.16 1.04
100 55.70 1.03 1.06 1.15 1.04
200 52.70 1.04 1.07 1.15 1.00
300 46.60 1.02 1.05 1.13 0.95
400 36.90 1.00 1.03 1.11 0.94
G550-
500 24.80 0.99 1.02 1.10 1.01
0.95-A
600 6.36 0.95 0.95 0.88 1.24
700 5.10 0.96 0.97 0.94 1.20
Mean 1.004 1.028 1.078 1.052
COV 0.035 0.045 0.098 0.104
0.917 0.933 0.922 0.893
20 36.20 1.00 1.03 1.05 1.00
200 31.70 1.00 1.03 1.06 1.00
300 23.10 0.96 0.99 1.00 1.00
400 17.00 0.96 0.98 0.98 1.01
G250- 500 12.30 0.95 0.97 0.96 1.03
0.95-A 600 8.22 0.94 0.95 0.93 1.02
700 4.95 0.95 0.96 0.93 1.04
Mean 0.966 0.987 0.986 1.015
COV 0.025 0.031 0.056 0.016
0.886 0.902 0.885 0.934
20 155.00 0.98 1.00 1.01 0.98
300 138.00 0.96 1.00 1.05 0.92
500 60.70 0.96 0.98 0.97 0.98
G450- 600 21.10 1.03 1.02 0.90 1.20
1.90-A 700 12.20 0.96 0.95 0.85 1.10
Mean 0.975 0.991 0.956 1.034
COV 0.031 0.027 0.085 0.111
0.889 0.906 0.809 0.831
20 96.10 0.99 1.01 0.96 0.99
300 64.70 1.00 1.01 0.95 1.01
400 45.80 0.96 0.97 0.90 0.99
500 32.10 0.93 0.94 0.87 0.98
G250-
600 21.40 0.94 0.94 0.86 0.99
1.95-A
700 12.00 0.93 0.93 0.86 0.98
Mean 0.959 0.967 0.900 0.993
COV 0.031 0.036 0.051 0.012
0.876 0.880 0.809 0.914
_________________________________________________________________
Table 5.6: Comparison Code Predictions with FEA Results for
Type B-Experimental Modelling
FEA/Predicted ultimate load

Specimen Temp. FEA
AS/NZS BS 5950 EC3
type (oC) (kN) DSM
4600 P5 P1.2
20 28.40 1.13 1.20 0.99 1.13
200 26.70 1.14 1.21 1.00 1.10
300 23.70 1.13 1.19 0.99 1.05
400 18.80 1.11 1.17 0.97 1.04
G550- 500 12.60 1.09 1.16 0.96 1.12
0.95-B 600 3.22 0.99 1.02 0.77 1.36
700 2.58 1.02 1.06 0.81 1.32
Mean 1.087 1.144 0.926 1.159
COV 0.054 0.065 0.100 0.110
0.977 1.016 0.785 0.967
20 16.60 1.08 1.14 0.90 1.08
200 14.60 1.08 1.16 0.90 1.08
300 10.80 1.06 1.12 0.87 1.14
400 8.20 1.08 1.13 0.88 1.15
G250- 500 5.83 1.05 1.10 0.84 1.14
0.95-B 600 3.81 1.01 1.10 0.80 1.11
700 2.29 1.01 1.05 0.80 1.13
Mean 1.053 1.113 0.856 1.118
COV 0.031 0.033 0.051 0.026
0.963 1.017 0.772 1.025
20 96.50 1.11 1.19 1.00 1.11
200 92.50 1.14 1.21 1.02 1.07
300 85.60 1.15 1.22 1.03 1.05
400 66.00 1.14 1.22 1.05 1.06
G450- 500 37.10 1.09 1.17 0.94 1.04
1.90-B 600 13.20 1.17 1.21 0.88 1.39
700 7.61 1.17 1.23 0.83 1.32
Mean 1.138 1.208 0.963 1.149
COV 0.026 0.019 0.087 0.126
1.043 1.110 0.832 0.932
20 58.20 1.07 1.12 0.89 1.07
200 52.80 1.08 1.14 0.88 1.07
300 38.60 1.06 1.11 0.83 1.08
400 28.20 1.05 1.09 0.85 1.10
G250- 500 19.60 1.01 1.04 0.82 1.07
1.95-B 600 13.00 1.00 1.04 0.81 1.08
700 7.25 0.99 1.02 0.80 1.06
Mean 1.037 1.081 0.840 1.075
COV 0.037 0.044 0.045 0.010
0.945 0.980 0.761 0.991
_________________________________________________________________
5.4 Development of Ideal Model for Local Buckling Simulations
Ideal models are used to simulate the ideal local buckling situations. In the
experimental study stub columns failed in local buckling by initially forming a series
of local buckling half waves along the column length. These local buckling half
waves were not identical since the column ends were fixed (see Figure 5.11). They
are identical only when true pin-end supports are used. Therefore ideal models with
pinned end conditions were developed for use in the parametric study.
5.4.1 Boundary Conditions
In this study on boundary conditions, a quarter wave length model was selected
because it is the shortest model length to represent local buckling. A detail
investigation on model length is given in Section 5.3.2. In the ideal models based on
quarter wave length, free rotation was allowed at the edges of the upper end. Only
out of plane translation was allowed in addition to the rotation about the longitudinal
axis to represent the peak configuration of a sine wave. Rotation free boundary
condition was created by two methods, namely rigid body motion and explicit MPC
with rigid fixed MPC.
Method 1
A rigid surface was created using R3D4 elements on the top of the column. The
R3D4 element is a bilinear quadrilateral element with four-nodes. It has only
translational degrees of freedom. Therefore attached shell elements can freely rotate
at the common boundary of shell elements and rigid surface and thus buckling
rotations are unrestrained. The rigid surface is characterized by a reference node and
the load was applied to the reference node. This node was created on the rigid
surface and at the geometric centre of the section. Upper end boundary conditions
were also applied to the rigid reference node because the motion of the degrees of
freedom is constrained by the rigid body reference node. The bottom end was
unrestrained with respect to out of plane deflections (Ux and Uy) and rotation (Rz)
for buckling deformations. Figure 5.13 shows a quarter wave length model
developed with rigid surface at the top.
_________________________________________________________________
Rigid surface
made of R3D4
Elements Rigid body reference node
S4 shell elements
Figure 5.13: Quarter Wave Length Model Developed with Rigid Surface
Method 2
The second method was created using the explicit MPCs and a rigid fixed MPC.
Explicit MPC can be used to connect a dependent node to an independent node with
user defined degrees of freedom. The perimeter nodes were transformed and then the
transformed nodes were connected to the perimeter nodes using Explicit MPCs. All
the translational and rotational degrees of freedom about the axis of column were
connected using Explicit MPCs. Transformed nodes were selected as independent
nodes and perimeter nodes of top edge were selected as dependent nodes. As
rotational degrees of freedom about the axes along the top edges are unrestrained,
they can rotate to form local buckling waves. Independent nodes of explicit MPC
were connected to a node at the geometric centre using a rigid fixed MPC. The upper
edge of the explicit MPC was connected to the centre node with rigid beams using a
rigid fixed MPC. This behaves as a rigid surface. Axial load and boundary
conditions were applied to the centre node of the rigid fixed MPC. All degrees of
freedom except axial translation were restrained at the upper end of the column.
Bottom edges were unrestrained with respect to out of plane deflections (Ux and Uy)
_________________________________________________________________
and rotation (Rz) for buckling deformation. Figure 5.14 shows a quarter wave model
developed with MPCs to simulate a pin-end.
S4 shell elements
Rigid fixed MPC

Explicit MPCs
connected with Ux,
Uy, Uz and Rz
Figure 5.14: Quarter Wave Length Model Developed with MPCs
Analyses showed that both methods could be used to create rotation free boundary
conditions while applying the axial compression load at the geometric centroid. Both
methods predicted the same values for elastic buckling loads from bifurcation
analyses and ultimate loads from nonlinear analyses.
5.4.2 Model Length
The section capacity is given by the local buckling capacity and is independent of the
member length. Therefore local buckling can be modelled based on any number of
buckling half waves (see Figure 5.15). In the case of half length experimental model,
one end was kept fixed while the other end was allowed to translate out of plane.
This model does not represent pure local buckling waves because one end is fixed
and the length is not the exact multiple of half wave buckling lengths. As described
_________________________________________________________________
in Chapter 4, an additional 20 mm length was also included at both ends of test
specimens. Therefore boundary conditions should be changed to allow rotation at the
support. Shorter models give economical modelling as they save processing time and
memory requirements. Therefore it is important to find the shortest possible model
length for the ideal models. This can be achieved by considering the symmetry about
the plane through the peak of single half wave (see Figure 5.16). Therefore a quarter
wave length model was chosen to simulate the local buckling behaviour of cold-
formed steel compression members. However, its accuracy should be evaluated first.
Half wave length for

critical local buckling
Range of pure local

buckling
Figure 5.15: Buckling Plot for a Type A Section from CUFSM Analyses
Plane of symmetry
Figure 5.16: Development of Quarter Wave Model Considering Symmetry
_________________________________________________________________
The half wave lengths of the columns corresponding to the critical local buckling
mode were taken from the CUFSM analyses, which give elastic buckling load
factors and corresponding half wave lengths for different buckling modes. In order to
determine the accuracy of critical half wave length predictions, the G550-0.95-20-A
section was analysed using ABAQUS with varying model lengths. The CUFSM
analyses gave a local buckling half wave length of 50 mm. The ABAQUS quarter
wave models were then developed by varying the model length from 20 mm to 30
mm (25 + or – 5) at 1 mm intervals. Figure 5.17 shows the variation of the elastic
local buckling load against the quarter wave model length. The minimum buckling
load corresponds to a model length of 25 mm. Therefore the use of 50 mm as the half
wave length is accurate.
29.4
29.2
29.0
Elastic Buckling Load (kN)
28.8
28.6
28.4
28.2
28.0
27.8
27.6
20.0 22.0 24.0 26.0 28.0 30.0
Quarter Wave Length (mm)
Figure 5.17: Variation of Elastic Buckling Load with Model Length
In order to assess the accuracy of elastic buckling load predictions using the quarter
wave length model, elastic buckling loads were obtained for all the tested sections
using CUFSM and FEA. Table 5.7 compares the elastic buckling loads from both
methods. Elastic bucking loads from the quarter wave length models of Type A
sections were slightly less than the FEA predictions. Therefore a further study was
undertaken to understand the effect of model length on the buckling load predictions.
_________________________________________________________________
Table 5.7: Comparison of Elastic Buckling Loads from CUFSM and FEA for
Quarter Wave Length FEA Model
Section Elastic Buckling Load (kN)

CUFSM FEA
G550-0.95-20-A 29.73 27.83
G250-0.95-20-A 30.93 26.03
G450-1.90-20-A 145.28 136.01
G250-1.95-20-A 112.66 105.73
G550-0.95-20-B 16.91 16.87
G250-0.95-20-A 16.50 16.46
G450-1.90-20-A 67.62 66.61
G250-1.95-20-A 68.19 67.43
Accuracy of the ABAQUS quarter wave length model was assessed using the elastic
buckling loads predicted by the CUFSM analyses and the ultimate loads from
experimental tests. For this purpose, ideal models of all tested sections were
simulated with a quarter wave length, one half wave length and three half wave
lengths. For one half wave and three half waves length models, both ends were
modelled as pinned with one end prevented from translation along the longitudinal
axis. Even though the end support condition was fixed in the tests and a pinned
support was used in the finite element analyses, the ultimate loads were compared.
The comparison showed a closer prediction as local buckling capacity was less
affected by the end support conditions. The quarter wave length finite element model
predicted smaller elastic buckling loads than CUFSM, particularly for Type A
sections (see Tables 5.8 and 5.9). These analyses show that the model length has an
effect on the buckling load, in particular for Type A models. It appears that three half
wave length models produced the best agreement with CUFSM and test results.
_________________________________________________________________
Table 5.8: Comparison of Elastic Buckling and Ultimate Loads for Type A
Specimens with Different Model Lengths
Test EB Load CUFSM/FEA Test/FEA

Grade and Elastic buckling load Ultimate load
result (kN)
thickness
(kN) CUFSM QW HW 3HW QW HW 3HW
G550-0.95 53.95 29.73 1.068 1.068 1.021 0.946 0.946 0.992
G250-0.95 37.09 30.93 1.065 1.065 1.020 1.069 1.069 1.101
G450-1.90 151.30 145.28 1.068 1.068 1.023 0.946 0.952 0.982
G250-1.95 102.04 112.66 1.066 1.066 1.022 1.050 1.049 1.081
Mean 1.067 1.067 1.022 1.003 1.004 1.039
COV 0.001 0.001 0.001 0.057 0.055 0.050
Table 5.9: Comparison of Elastic Buckling and Ultimate Loads for Type B
Specimens with Different Model Lengths
Test EB Load CUFSM/FEA Test/FEA

Grade and (kN) Elastic buckling load
result ultimate load
thickness CUFSM
(kN) QW HW 3HW QW HW 3HW
G550-0.95 26.50 16.91 1.002 1.001 1.001 0.877 0.877 0.898
G250-0.95 16.83 16.50 1.002 1.001 1.001 0.990 0.990 1.008
G450-1.90 96.89 67.62 1.011 1.008 1.007 0.969 0.969 0.984
G250-1.95 66.37 68.19 1.011 1.010 1.007 1.125 1.125 1.064
Mean 1.007 1.005 1.004 0.990 0.990 0.988
COV 0.004 0.004 0.003 0.89 0.089 0.060
QW= Quarter Wave HW=Half Wave EB= Elastic Buckling
_________________________________________________________________
For further understanding of the effect of model length, G550-0.95-20-A specimen
was simulated with both ideal and experimental conditions and different model
lengths. A range of model lengths, from quarter wave to thirteen half waves, was
considered in this investigation. Three types of models were developed for this
investigation. The first type was the both ends pinned model with multiple half wave
lengths. The length of this model was an integral multiple of half wave length (ie. n x
HWL). The second type of model was also a pin-end model but with a quarter wave
at one end. It is an extended quarter wave length model with multiple half wave
lengths and symmetry conditions at one end. Hence its length is equal to 0.5nHWL,
where n is the number of half waves in the corresponding column. The third type of
model was the fixed-end model, which is similar to the experimental model. Since
the ends were fixed, the model or column length was an integral multiple of half
wave length plus 40 mm to avoid any end effect. Figure 5.18 shows these three types
of models.
Elastic buckling and nonlinear analyses were carried out for each type of models
while increasing the model length. Both elastic buckling and ultimate loads were
recorded for each model type. The results showed that shorter models do not predict
the results accurately. However, they converge to the elastic buckling load from the
CUFSM analyses and the ultimate load from the tests (see Figures 5.19 and 5.20) for
higher model lengths. Therefore a model length of three half waves is sufficiently
accurate to simulate the local buckling behaviour of ideal stub columns. Since the
curves for model types one and two overlap, it can be concluded that the extended
quarter wave buckling length model with n=3 (see Figure 5.18 b) can be used for the
local buckling simulation as it economises the processing time and memory
requirements.
_________________________________________________________________
Behaviour and Design of Cold-Formed Steel Compression Members at Elevated Temperatures 5-37
(a) Pin-end Model with Multiple Half Waves
(b) Extended Quarter Wave Model
(c) Experimental Model
Figure 5.18: Different Model Types Used for Model Length Study
_________________________________________________________________
35
30
Elastic Buckling Load (kN)

25
20
Pin End Full Length Model (Model Length = n x HWL)
Pin End Half Length Model (Model Length = 0.5 x n x HWL
15
Fixed End Full Length Model (ModelLength = n x HWL + 40 mm)
CUFSM
10 HWL = Half Wave Length
0
0 2 4 6 8 10 12
No. of Half Waves in Corresponding Column (n)
Figure 5.19: Elastic Buckling Load of Models with Varying Model Lengths
60
50
40
Ultimate Load (kN)
Pin End Full Length Model (Model Length = n x HWL)

30 Fixed End Full Length Model (Model Length = n x HWL + 40 mm)
Pin End Half Length Model (Model Length = 0.5 x n x HWL)
Test
20
HWL = Half Wave Length
10
0
0 2 4 6 8 10 12
No. of Half Waves in Corresponding Column (n)
Figure 5.20: Ultimate Load of Models with Varying Model Lengths
Analyses of G550-0.95-20-A specimens showed that the model length has an effect
on the buckling load although it is not significant. A model that represents a column
whose length less is than three half waves gives higher ultimate loads, but lower
elastic buckling loads (see Figure 5.19 and 5.20). A model that represents a column
of three half waves gave reasonably accurate results. Therefore the extended quarter
wave model with one half wave and one quarter wave was selected for the
parametric study of both Type A and Type B sections because it gives reasonably
_________________________________________________________________
accurate predictions of both elastic and ultimate loads with minimal processing time
and memory requirement.
5.5 Conclusions
Two kinds of finite element models, experimental and ideal finite element models,
were developed to simulate the local buckling behaviour of cold-formed steel lipped
and unlipped channels under axial compression load. They were developed to
validate the models using experimental results and for the use in parametric study,
respectively. During the validation process, the effects of imperfections, particularly
the direction and magnitude, were investigated using the available tests results
discussed in Chapter 4. The ultimate loads of lipped channels (Type A) with fixed
end conditions were affected by the magnitude and direction of imperfections.
Positive imperfection gave slightly higher ultimate loads than that of negative
imperfection. However, the maximum difference is less than 4%. Comparative study
with test results indicated that positive imperfections gave more accurate results.
However, the direction of imperfection did not affect the ultimate loads of unlipped
channels but reversed the buckling wave shown in the bifurcation buckling analyses.
Since the standard local buckling imperfection for a unstiffened flange of unlipped
columns is independent of flange width, a comparative study with test results was
carried our to verify the accuracy. Finally, the developed experimental finite element
models were validated using the ultimate loads, load-deflection curves and the
deflected shapes observed in the experimental tests. These models were then
converted to ideal modes by changing the geometry and boundary conditions. An
ideal model length of one half wave and one quarter wave was recommended after a
detailed investigation on model length.
_________________________________________________________________
6 Parametric Study and Numerical Analyses of
Local Buckling
A detailed parametric study was carried out to fully understand the effects of various
parameters that affect the local buckling behaviour of cold-formed steel columns at
ambient and elevated temperature and to review the accuracy of current design rules.
For this purpose, ideal finite element models were developed to simulate the local
buckling behaviour. Before developing the ideal models, eight experimental finite
element models were developed. These models were validated using 23 ambient
temperature tests and 65 elevated temperature tests reported in Chapter 4.
Experimental models were then modified to reflect the ideal conditions by simply
changing the geometric parameters and boundary conditions. Details of the ideal
models are given in Chapter 5. All the steel thicknesses and strength grades used in
the experimental study were selected for the parametric study. A range of width to
thickness ratios were selected while selecting the dimensions of the specimens. A
slender web in lipped channel and a slender flange in unlipped channel were selected
while keeping all other elements compact. As for the experimental study, lipped
channels were referred to as Type A sections whereas unlipped channels were
referred to as Type B sections.
The parametric study was first carried out at ambient temperature to assess the
accuracy of available design methods. It was then extended to elevated temperature
conditions by using the reduced mechanical properties. Analyses were carried out
from 100oC to 700oC at 100oC intervals up to 700oC. Elastic buckling and ultimate
loads for each section were obtained from the analyses. This chapter presents the
details of the parametric study and the results.
The results were compared with the ultimate compression capacity predictions from
the Australian/New Zealand standard AS/NZS 4600 (SA, 2005), the British Standard
BS 5950 Part 5 (BSI, 1998), Eurocode 3 Part 1.2 (2005) and the Direct Strength
Method (DSM). Appropriately reduced mechanical properties at elevated
temperatures were required in the finite element analyses and in the calculation of
ultimate loads from the design codes. They were obtained from Ranawaka and
Mahendran (2009) and Dolamune Kankanamge (2009).
_________________________________________________________________
6.1 Parametric Study at Ambient Temperature Conditions
The parametric study was first carried out at ambient temperature to asses the
accuracy of the current ambient temperature design rules in order to extend their use
to elevated temperatures. It was conducted separately for lipped and unlipped
channels in order to understand the effects of stiffened and unstiffened elements in
cold-formed steel sections.
6.1.1 Lipped Channel Sections (Type A)

Section dimensions were selected in order to investigate the dominant local buckling
in their web element. Therefore the chosen sections included slender web elements
and compact flanges and lips. The web width was varied to investigate the effect of
width to thickness ratio, while keeping the flange and lip sizes as constant. The
width to thickness ratio of web was varied from 30 to 80 for both low and high
strength steels. Following nomenclature was used to name Type A sections.
G550-0.95-30-20-7
Lip length
Flange width
Web width
Thickness
Grade
6.1.1.1 Comparison of FEA Results with Design Methods

The ultimate loads obtained from the finite element analyses were compared with the
corresponding loads calculated using the design rules of AS/NZS 4600, BS 5950
Part 5 and the direct strength method (DSM). Sample calculations for each of these
methods are given in Appendix C. Except the direct strength method, all the other
design methods are based on the effective width method.
Table 6.1 compares the ultimate loads from FEA and the above mentioned design
methods. The ratio of finite element analysis results to code predictions was
_________________________________________________________________
calculated for each specimen. This ratio was tabulated based on the thickness and
strength grade. The mean value and the associated coefficient of variation were then
calculated for each thickness and strength grade. The mean values of all the design
methods are close to one while the associated coefficients of variation are also small.
However, the mean value for AS/NZS 4600 and BS 5950 Part 5 are slightly less than
one while that for the direct strength method is more than one.
The ratios of finite element analysis results to the AS/NZS 4600 and BS 5950 Part 5
code predictions are almost the same when web width to thickness ratio was
increased. On the other hand the ratio increases for the direct strength method.
Figure 6.1 compares the direct strength method curve with the results of the
parametric study at ambient temperature in the form of Pult Py versus l ( P P ),
y cr
where Pult , Py and Pcr are the ultimate compression load, yield capacity of the section
and elastic buckling load, respectively. The results for low l values are below the
direct strength method curve. Hence these results are slightly unconservative. This
trend can be mostly observed with low strength steels and channels with a narrow
web. However, most of the results followed the direct strength method curve.
To investigate the effect of web slenderness, other elements of Type A section were
chosen as compact elements. A range of width to thickness ratios was selected from
30 to 150. For each width to thickness ratio, the ultimate loads from finite element
analyses were compared with the predictions of AS/NZS 4600, BS 5950 Part 5 and
the direct strength method (DSM). These comparisons were made for high and low
strength steel sections as shown in Figures 6.2 and 6.3. The width to thickness ratio
has a significant effect on the prediction of direct strength method for high strength
steel columns subjected to local buckling. Smaller values of width to thickness ratio
give slightly unsafe results while higher values give conservative results. It appears
that high strength steels gave safer results than low strength steels.
Finite element analyses results showed that increasing web element size increased
the local buckling capacity to a maximum and then the capacity decreased (see Table
6.1). The effective width method used in AS/NZS 4600 and BS 5950 Part 5 does not
lead to reduction in the local buckling capacity when the web width to thickness ratio
is increased. Hence their predictions become unsafe for larger web width to
_________________________________________________________________
thickness ratios. The use of more accurate buckling coefficients (instead of 4) is
likely to improve these predictions. On the other hand the direct strength method
capacity predictions increase first and then decrease with increasing web width to
thickness ratio. Therefore the direct strength method follows the trend of FEA
results. However, the ratio of FEA results to code predictions decrease slightly with
DSM at higher web width to thickness ratios. In other words, its rate of reduction of
local buckling capacity predictions is slightly greater than the actual rate at higher
web width to thickness ratios. In the case of smaller web width to thickness ratios,
DSM predictions are slightly unconservative, particularly for high strength steels.
However, AS/NZS 4600 and BS 5950 Part 5 predictions are still safe because the
decreasing trend of FEA results only commence with very high width to thickness
ratios. Schafer (2006) also highlighted the fact that the DSM can cope with the local
buckling capacity variations with increasing web width to thickness ratio. Further,
the DSM discourages the use of very slender web elements in lipped channel
sections with narrow flanges and lips (Schafer, 2002).
The mean values of all the design methods are close to one while associated
coefficient of variations are small. Since the parametric study considered pin-end
conditions, the local buckling capacities are slightly less compared to fixed-end
conditions. Hence the results of experimental finite element modelling showed a
good agreement with the predictions (see Tables 5.5 and 5.6). Further, the F factors
calculated and shown in Table 6.1 are also greater than 0.85. Therefore all the
ambient temperature design methods can be recommended for local buckling
capacity calculations of cold-formed steel compression members. As highlighted by
Schafer (2002), the use of larger web width to thickness ratios is not recommended
in lipped channels sections where local buckling is dominant.
_________________________________________________________________
Table 6.1: Comparison of FEA Results with Code Predictions for Type A
Sections at Ambient Temperature
FEA results
FEA/Code Prediction
Specimen AS/NZS BS5950
EB ULT DSM
4600 P5
G550-0.95-35-20-7 59.6 45.3 0.97 0.98 0.98
G550-0.95-45-20-7 41.4 45.7 0.96 1.04 0.99
G550-0.95-55-20-7 31.1 45.4 0.94 1.07 0.98
G550-0.95-65-20-7 24.7 46.2 0.95 1.12 1.00
G550-0.95-75-20-7 20.3 45.9 0.94 1.13 0.99
G550-0.95-100-20-7 13.6 43.8 0.89 1.11 0.94
G550-0.95-150-20-7 8.2 41.0 0.82 1.06 0.86
G450-1.9-60-40-15 300.7 153.0 0.96 0.92 0.96
G450-1.9-80-40-15 198.4 157.0 0.96 0.97 0.97
G450-1.9-100-40-15 143.9 159.0 0.95 1.03 0.98
G450-1.9-130-40-15 99.8 161.0 0.95 1.08 0.99
G450-1.9-160-40-15 75.3 161.0 0.94 1.11 0.99
G450-1.9-200-40-15 56.1 159.0 0.92 1.11 0.97
G450-1.9-300-40-15 33.5 152.0 0.87 1.08 0.91
Mean 0.934 1.058 0.965
COV 0.041 0.060 0.039
F 0.851 0.954 0.882
G250-0.95-35-20-7 58.1 25.6 0.98 0.95 0.97
G250-0.95-45-20-7 40.4 26.6 0.98 0.95 0.98
G250-0.95-55-20-7 30.4 27.0 0.98 0.99 0.99
G250-0.95-65-20-7 24.1 27.3 0.97 1.02 1.00
G250-0.95-75-20-7 19.8 27.6 0.97 1.04 1.02
G250-0.95-100-20-7 13.5 27.4 0.95 1.06 1.00
G250-0.95-150-20-7 7.8 25.2 0.86 1.00 0.91
G250-1.95-60-40-15 296.3 87.6 0.98 0.98 0.98
G250-1.95-80-40-15 195.5 93.2 0.97 0.93 0.97
G250-1.95-100-40-15 141.9 95.7 0.97 0.94 0.97
G250-1.95-130-40-15 98.3 98.0 0.97 0.99 0.99
G250-1.95-160-40-15 74.2 99.8 0.97 1.03 1.01
G250-1.95-200-40-15 101.0 0.97 1.05 1.02 0.97
G250-1.95-300-40-15 97.3 0.92 1.03 0.97 0.92
Mean 0.972 0.991 0.977
COV 0.043 0.040 0.032
F 0.886 0.904 0.896
EB:- Elastic Buckling Load ULT:- Ultimate Compression Load
_________________________________________________________________
1.2
0.8
PULT/PY
0.6
0.4 G550-0.95-20-A
G250-0.95-20-A
G450-1.9-20-A
0.2
G250-1.95-20-A
DSM
0
0 0.5 1 1.5 2 2.5 3 3.5 4
l
Figure 6.1: Comparison of FEA Results with DSM for Type A Specimens at
Ambient Temperature
1.00
FEA/Predicted Ultimate Load
G550-0.95 -20-A AS/NZS4600

G550-0.95-20-A DSM
G550-0.95-20-A-BS5950.5
G450-1.9-20-A AS/NZS4600
G450-1.9-20-A DSM
G450-1.9-20-A BS5950.5
0.00
0 20 40 60 80 100 120 140 160 180
B/T
Figure 6.2: Effect of Web Width to Thickness Ratio – High Strength Steels at
Ambient Temperature
_________________________________________________________________
1.00
FEA/Predicted Ultimate Load
G250-0.95-20-A AS/NZS 4600

G250-0.95-20-A BS5950.5
G250-0.95-20-A DSM
G250-1.95-20-A AS/NZS 4600
G250-1.95-20-A BS5950.5
G250-1.95-20-A DSM
0.00
0 20 40 60 80 100 120 140 160
B/T
Figure 6.3: Effect of Web Width to Thickness Ratio – Low Strength Steels at
Ambient Temperature
6.1.2 Unlipped Channel Sections (Type B)

Type B sections were selected so that local buckling is dominant in their flanges.
Therefore slender flange elements were selected with a compact web element. To
investigate the effect of flange width to thickness ratio, different flange sizes were
selected while keeping the web element as compact. The width to thickness ratio of
flange was varied from 20 to 50 for both low and high strength steels. Following
nomenclature was used to name Type B sections. The ultimate loads from finite
element analyses were compared with AS/NZS 4600, BS5950 Part 5, and the direct
strength method.
G550-0.95-20-10
Flange width
Web width
Thickness
Grade
_________________________________________________________________
Similar to lipped channels, the ratios of FEA to code predictions were calculated for
the design methods of AS/NZS 4600, BS 5950 Part 5, and the direct strength method
(DSM). These ratios were tabulated based on the thickness and strength grades (see
Table 6.2). The mean values, the associated coefficients of variation and the capacity
reduction factors for the FEA to code predictions ratio were calculated for each case.
Results showed that there are significant variations among the different design
methods. Both AS/NZS 4600 and BS 5950 Part 5 predictions are conservative.
However, larger flanges gave more conservative results (see Figures 6.5 and 6.6).
The possible reason of more conservative results for the effective width method is
the use of lower k value for effective width calculations. The k values for unstiffened
elements are either 0.425 or 0.43. These k values are based on an element that is
simply supported on three sides and one edge is free. However, when the resisting
effect of the web is considered, higher k value exists in the flanges of unlipped
channels (Yu, 2000). The effective width method does not consider this effect and
hence code predictions are conservative. Since the direct strength method does not
involve such a conservative approach, results are not so conservative as for AS/NZS
4600 and BS 5950 Part 5.
The DSM predictions are conservative for high strength steel members while slightly
unconservative for low strength steel sections with smaller flanges. Figure 6.4 shows
the comparison of results with the DSM. In the case of lipped channels, the DSM
discourages the use of a very slender web element (Schafer, 2002). In other words, it
discourages the members where the elastic buckling load is comparatively low
compared to the yield capacity of the section. Having a very slender flange also
creates the same situation of having a lower elastic buckling load. However, the
ratios of FEA ultimate loads to predictions are greater than those for higher flange
width to thickness ratios. In the case of lower flange width to thickness ratios, the
DSM is not accurate, particularly for lower strength steels. Hence the ultimate load
predictions of the DSM for unlipped channels with very low slender flanges are not
reliable. It may be the reason for lower capacity reduction factors (F) for the DSM.
However, the F factors are greater than 0.85 for all the design methods and hence
they are acceptable for ambient temperature design.
_________________________________________________________________
Table 6.2: Comparison of FEA Results with Code Predictions for Type B
Sections at Ambient Temperature
FEA results FEA/Code Prediction
Specimen AS/NZS BS5950
EB ULT DSM
4600 P5
G550-0.95-20-20 20.9 25.1 1.10 1.00 1.17
G550-0.95-20-30 13.2 29.5 1.27 1.15 1.36
G550-0.95-20-40 9.7 31.1 1.32 1.18 1.42
G550-0.95-20-50 7.7 31.7 1.34 1.17 1.42
G450-1.9-40-40 83.2 84.4 1.06 0.95 1.12
G450-1.9-40-60 52.7 100.0 1.23 1.09 1.32
G450-1.9-40-80 38.7 109.0 1.32 1.16 1.42
G450-1.9-40-100 30.8 111.0 1.34 1.14 1.43
Mean 1.248 1.105 1.333
COV 0.089 0.078 0.092
F 1.082 0.972 1.150
G250-0.95-20-20 20.4 14.1 1.04 0.88 1.08
G250-0.95-20-30 9.5 19.1 1.33 1.11 1.44
G250-0.95-20-40 7.5 20.7 1.43 1.16 1.54
G250-0.95-20-50 7.5 20.7 1.43 1.16 1.54
G250-1.95-40-40 82.4 49.6 1.01 0.85 1.04
G250-1.95-40-60 52.3 59.4 1.16 0.97 1.25
G250-1.95-40-80 38.4 67.8 1.30 1.07 1.41
G250-1.95-40-100 30.5 73.2 1.39 1.12 1.51
Mean 1.261 1.040 1.351
COV 0.135 0.119 0.150
F 1.017 0.861 1.058
_________________________________________________________________
1.2
0.8
PULT/Py
0.6
0.4
G550-0.95-20-B G250-0.95-20-B
0.2 G450-1.9-20-B G250-1.95-20-B
DSM
0
0 0.5 1 1.5 2 2.5 3 3.5 4
l
Figure 6.4: Comparison of FEA results with DSM for Type B Specimens at
Ambient Temperature
1.00
FEA/Predict
G250-0.95-20-B AS/NZS 4600

G250-0.95-20-B BS 5950.5
G250-0.95-20-B DSM
G250-1.95-20-B AS/NZS 4600
G250-1.95-20-B BS 5950.5
G250-1.95-20-B DSM
0.00
0 10 20 30 40 50 60
B/T
Figure 6.5: Effect of Flange Width to Thickness Ratio – High Strength Steels at
Ambient Temperature
_________________________________________________________________
1.00
FEA/Predicted Load
G550-0.95-20-B AS/NZS 4600

G550-0.95-20-B BS5950.5
G550-0.95-20-B DSM
G450-1.9-20-B AS/NZS 4600
G450-1.9-20-B BS 5950.5
G450-1.9-20-B DSM
0.00
0 10 20 30 B/T 40 50 60
Figure 6.6: Effect of Flange Width to Thickness Ratio – Low Strength Steels at
Ambient Temperature
6.2 Parametric Study at Elevated Temperature Conditions

A series of stub columns was analysed at elevated temperatures from 100oC to 700oC
at 100oC intervals. In the case of unlipped channels, all the members analysed at
ambient temperature were considered at elevated temperature. But in the case of
lipped channels the specimens with large web width to thickness ratios were not
included because the current ambient temperature design methods are not reliable for
very slender web elements. Elevated temperature simulations were carried out using
reduced mechanical properties. Both elastic buckling and ultimate loads from finite
element analyses (FEA) were recorded at each temperature. Cold-formed steel
design codes give suitable design rules to calculate the local buckling capacities at
ambient temperature. They were also used to calculate the local buckling capacities
at elevated temperatures by using the reduced mechanical properties at elevated
temperatures. In addition to the design standards considered in the ambient
temperature capacity calculations, Eurocode 3 Part 1.2 (fire code) was also used for
the capacity predictions. The ultimate load results of the parametric study and local
buckling capacities calculated using the design standards were tabulated according to
the type of section, grade of steel, thickness and temperatures. Similar to the ambient
_________________________________________________________________
temperature case investigation, the investigations on lipped channels (Type A) and
unlipped channels (Type B) were carried out separately.
6.2.1 Lipped Channel Sections (Type A)

Comparative results for Type A sections were tabulated separately at each
temperature. This allows us to investigate the effects of the relative variation of
elastic modulus and yield stress with increasing temperature. Tables 6.3 to 6.9
compare the results of FEA and code predictions. They provide the ratio of FEA to
code predictions, mean values of the predictions based on the thickness and strength
grade and associated coefficients of variation. In addition, capacity reduction factors
(F) were also calculated at each temperature based on the strength class. Figures 6.8
to 6.11 show the variation of mean values with temperature.
The ratios of FEA ultimate load to code predictions are slightly less than one.
However, this trend was not observed in the experimental modelling (see Tables 5.5
and 5.6). The reason is the smaller ultimate load from the ideal model due to pin-end
conditions. It was observed that the fixed-ended columns gave slightly higher
capacities than the pin-ended columns (see Figure 5.20 when n=3).
Since the elevated temperature simulations were not considered for sections with
larger web elements, FEA to code predictions show similar values with increasing
web size at any temperature (except for DSM) and the mean values can represent all
the sections of that case. Therefore the variation of mean values with temperature
was plotted in Figures 6.8 to 6.11 to understand the effect of temperature. The mean
values for the direct strength method are greater than one for high strength steels for
lower temperatures. In the case of lower strength steels, direct strength method
predictions are slightly unconservative compared to the high strength steels. But
these results are still in the acceptable range.
The mean values for AS/NZS 4600 and BS 5950 Part 5 are of the same order for all
the steel grades and thicknesses with increasing temperature. This pattern can be
seen only with the predictions using the effective width method. But the mean
values of the direct strength method remain the same order only up to 500oC for
higher strength steels and beyond 500oC, the mean values of high strength steels
_________________________________________________________________
decrease. The yield stress (fy) of higher strength steels decreases rapidly at a faster
rate than for modulus of elasticity (E) beyond 500oC as seen in Figure 6.7 (higher
E/fy ratios at 600oC). Therefore yield capacity of the section decreases rapidly while
elastic buckling load decreases at the same rate as lower elevated temperatures. This
causes a significant difference between the elastic buckling load and the yield
capacity. Therefore the slenderness of the section at higher temperatures (beyond
500oC), particularly for higher strength steels, is relatively small compared to that at
ambient temperature. It was observed that in the ambient temperature investigation
the results of lower slenderness ratio were slightly unconservative for DSM. Due to
these changes with higher E/fy ratios, the direct strength method in its current format
is struggling to cope with the fire design of compression members made of steels
with higher E/fy ratios.
1.4
1.2
1.0
(E/Fy)x1000
0.8
0.6
G550-0.95
0.4 G250-0.95
G250-1.95
0.2 G450-1.95
0.0
0 100 200 300 400 500 600 700 800
o
Temperature ( C)
Figure 6.7 Variation of E/fy ratio with temperature
The direct strength method formula for local buckling is a function of both elastic
buckling load and yield capacity of stub column. Due to the contribution of relatively
higher elastic buckling load in comparison to yield load, the DSM predictions are
slightly higher beyond 500oC. This is the opposite case of having very slender web
element in lipped channels. The direct strength method discourages very slender web
elements (Schafer, 2002). Further, very slender web elements decrease the elastic
buckling load and form a considerable difference between elastic buckling load and
_________________________________________________________________
yield capacity and hence lower capacity predictions. Similar observation is made at
elevated temperatures due to the significant drop in yield stress and yield load at
elevated temperature compared to the elastic buckling load. In other words, the
relative slenderness is high in the case of elevated temperatures or very slender webs.
For higher E/fy ratios, the elastic buckling load is relatively higher than the yield
capacity of stub columns. This causes DSM predictions slightly unconservative
because it is based on both the combination of elastic buckling load and the yield
capacity of the section. Reflecting this fact, the ratio of FEA to DSM predictions are
slightly lower for lower strength steels than that of higher strength steels even at
lower elevated temperatures, particularly below 500oC. Beyond 500oC, the E/fy ratio
increases significantly for higher strength steels and it is of the same order for low
strength steels (see Figure 6.7). Hence the ratios of FEA to the DSM predictions are
of the same order for the both low and high strength steels beyond 500oC.
The mean values for Eurocode 3 Part 1.2 method for low strength steels are of the
same order whereas the mean values for high strength steels have considerable
variation beyond 500oC. Since it recommends the use of ambient temperature
effective area for elevated temperature design, the increase of effective area of
higher strength steels at higher temperatures due to the higher E/fy ratio is not
considered. Hence the predicted capacity is relatively small compared to the FEA
results and hence more conservative predictions for higher strength steels at elevated
temperatures (see Figure 6.11).
Although the code predictions are slightly unsafe for columns with stiffened
elements, the capacity reduction factors (F) are greater than 0.85. Therefore it can
be concluded that ambient temperature design rules can be used in the design of
short cold-formed steel columns with stiffened elements subject to local buckling.
Further, the design rules based on the effective width method is more suitable for
elevated temperature conditions as it can cope with different E/fy ratios.
Eurocode 3 Part 1.2 predictions are reasonably accurate for lower temperatures while
being very conservative for higher temperatures. Since it recommends the use of
ambient temperature effective area at elevated temperature design, increases in
effective area due to higher E/fy ratios at elevated temperatures make its predictions
conservative, particularly for higher strength steels. Further, its limit of 350oC on the
_________________________________________________________________
maximum temperature is uneconomical. Ranby (2009) also highlights the fact that
the maximum temperature limitation is uneconomical.
Sections at 100oC
Specimen AS/NZS BS5950 EC3
EB ULT DSM
4600 P5 P1.2
G550-0.95-35-20-7 59.6 44.4 0.96 0.97 0.97 0.96
G550-0.95-45-20-7 41.4 44.9 0.95 1.03 0.98 0.96
G550-0.95-55-20-7 31.1 45.4 0.95 1.08 0.99 0.96
G550-0.95-65-20-7 24.7 45.4 0.95 1.11 0.99 0.95
G550-0.95-75-20-7 20.3 45.2 0.94 1.12 0.99 0.94
G450-1.9-60-40-15 280.7 152.0 0.96 0.91 0.96 0.95
G450-1.9-80-40-15 185.1 156.0 0.96 0.99 0.97 0.95
G450-1.9-100-40-15 134.3 158.0 0.95 1.04 0.99 0.95
G450-1.9-130-40-15 93.2 159.0 0.95 1.09 0.99 0.94
G450-1.9-160-40-15 70.3 160.0 0.94 1.13 1.00 0.93
Mean 0.951 1.047 0.983 0.949
COV 0.008 0.069 0.013 0.010
F 0.877 0.933 0.906 0.875
G250-0.95-35-20-7 58.1 22.4 0.90 0.88 0.90 0.90
G250-0.95-45-20-7 40.4 25.2 0.98 0.93 0.98 0.98
G250-0.95-55-20-7 30.4 25.7 0.98 0.98 0.99 0.98
G250-0.95-65-20-7 24.1 26.0 0.97 1.01 1.00 0.98
G250-0.95-75-20-7 19.8 26.3 0.97 1.03 1.01 0.98
G250-1.95-60-40-15 276.6 84.1 0.98 0.98 0.98 0.98
G250-1.95-80-40-15 182.5 89.3 0.97 0.93 0.97 0.97
G250-1.95-100-40-15 132.4 91.7 0.97 0.94 0.98 0.97
G250-1.95-130-40-15 91.8 93.8 0.97 1.00 1.00 0.97
G250-1.95-160-40-15 69.3 95.4 0.97 1.03 1.01 0.97
Mean 0.966 0.971 0.982 0.968
COV 0.024 0.051 0.032 0.025
F 0.887 0.879 0.899 0.889
_________________________________________________________________
Sections at 200oC
EB ULT DSM
4600 P5 P1.2
G550-0.95-35-20-7 51.8 42.5 0.95 0.98 0.97 0.94
G550-0.95-45-20-7 36.0 42.8 0.94 1.04 0.98 0.93
G550-0.95-55-20-7 27.1 43.3 0.94 1.09 0.99 0.93
G550-0.95-65-20-7 21.5 43.2 0.93 1.12 0.99 0.92
G550-0.95-75-20-7 17.7 42.9 0.92 1.13 0.98 0.90
G450-1.9-60-40-15 255.7 147.0 0.95 0.91 0.95 0.93
G450-1.9-80-40-15 168.6 151.0 0.95 0.99 0.97 0.93
G450-1.9-100-40-15 122.4 152.0 0.94 1.04 0.98 0.92
G450-1.9-130-40-15 84.8 154.0 0.94 1.10 0.99 0.92
G450-1.9-160-40-15 64.0 153.0 0.93 1.12 0.98 0.90
Mean 0.939 1.052 0.978 0.922
COV 0.011 0.070 0.013 0.014
F 0.865 0.937 0.901 0.849
G250-0.95-35-20-7 50.6 22.4 0.98 0.95 0.97 0.98
G250-0.95-45-20-7 35.1 23.2 0.98 0.94 0.98 0.98
G250-0.95-55-20-7 26.4 23.7 0.98 0.99 1.00 0.98
G250-0.95-65-20-7 21.0 23.9 0.97 1.02 1.01 0.97
G250-0.95-75-20-7 17.2 24.1 0.97 1.04 1.01 0.97
G250-1.95-60-40-15 251.9 79.8 0.97 0.97 0.98 0.97
G250-1.95-80-40-15 166.2 84.6 0.98 0.92 0.97 0.97
G250-1.95-100-40-15 120.6 86.8 0.98 0.95 0.98 0.97
G250-1.95-130-40-15 83.6 88.6 0.97 1.00 1.00 0.96
G250-1.95-160-40-15 63.1 90.0 0.97 1.04 1.02 0.96
Mean 0.975 0.982 0.992 0.971
COV 0.005 0.043 0.018 0.008
F 0.899 0.893 0.913 0.895
_________________________________________________________________
Sections at 300oC
EB ULT DSM
4600 P5 P1.2
G550-0.95-35-20-7 44.1 38.0 0.93 0.97 0.95 0.90
G550-0.95-45-20-7 30.6 38.2 0.92 1.03 0.96 0.89
G550-0.95-55-20-7 23.0 38.5 0.92 1.08 0.97 0.89
G550-0.95-65-20-7 18.3 38.3 0.91 1.10 0.96 0.88
G550-0.95-75-20-7 15.1 37.8 0.89 1.11 0.94 0.86
G450-1.9-60-40-15 215.0 138.0 0.95 0.93 0.96 0.91
G450-1.9-80-40-15 141.8 140.0 0.94 1.00 0.96 0.90
G450-1.9-100-40-15 102.9 141.0 0.93 1.06 0.97 0.89
G450-1.9-130-40-15 71.4 142.0 0.93 1.11 0.98 0.88
G450-1.9-160-40-15 53.9 141.0 0.91 1.13 0.97 0.87
Mean 0.923 1.052 0.962 0.887
COV 0.018 0.064 0.012 0.017
F 0.849 0.942 0.886 0.816
G250-0.95-35-20-7 43.0 16.5 0.97 0.96 0.97 0.99
G250-0.95-45-20-7 29.9 17.2 0.97 0.91 0.97 1.00
G250-0.95-55-20-7 22.5 17.6 0.97 0.96 0.99 1.00
G250-0.95-65-20-7 17.8 17.9 0.98 0.99 1.00 1.00
G250-0.95-75-20-7 14.7 18.1 0.98 1.02 1.01 1.00
G250-1.95-60-40-15 211.9 57.7 0.98 0.98 0.98 0.98
G250-1.95-80-40-15 139.8 61.7 0.97 0.94 0.97 0.98
G250-1.95-100-40-15 101.5 63.8 0.97 0.93 0.97 0.98
G250-1.95-130-40-15 70.3 65.3 0.97 0.98 0.99 0.98
G250-1.95-160-40-15 53.1 66.7 0.97 1.01 1.01 0.99
Mean 0.973 0.968 0.986 0.990
COV 0.005 0.036 0.017 0.010
F 0.897 0.884 0.908 0.913
_________________________________________________________________
Sections at 400oC
EB ULT DSM
4600 P5 P1.2
G550-0.95-35-20-7 36.3 30.4 0.92 0.96 0.95 0.90
G550-0.95-45-20-7 29.2 30.6 0.91 0.97 0.95 0.89
G550-0.95-55-20-7 19.0 31.0 0.92 1.07 0.96 0.89
G550-0.95-65-20-7 15.1 30.9 0.91 1.10 0.96 0.88
G550-0.95-75-20-7 12.4 30.6 0.89 1.11 0.95 0.87
G450-1.9-60-40-15 174.5 103.0 0.95 0.92 0.96 0.93
G450-1.9-80-40-15 115.1 104.0 0.94 0.99 0.96 0.91
G450-1.9-100-40-15 83.5 105.0 0.93 1.04 0.97 0.91
G450-1.9-130-40-15 57.9 106.0 0.93 1.09 0.98 0.90
G450-1.9-160-40-15 43.7 106.0 0.92 1.12 0.97 0.89
Mean 0.922 1.037 0.961 0.897
COV 0.018 0.069 0.010 0.019
F 0.848 0.924 0.886 0.825
G250-0.95-35-20-7 35.5 13.2 1.06 1.05 1.07 1.09
G250-0.95-45-20-7 24.6 13.1 1.00 0.94 1.00 1.04
G250-0.95-55-20-7 18.5 13.3 0.99 0.96 1.00 1.04
G250-0.95-65-20-7 14.7 13.3 0.98 0.98 1.00 1.03
G250-0.95-75-20-7 12.1 13.4 0.97 1.00 1.01 1.02
G250-1.95-60-40-15 171.9 45.6 0.99 0.99 0.99 0.99
G250-1.95-80-40-15 113.4 45.5 0.91 0.88 0.91 0.92
G250-1.95-100-40-15 82.3 45.4 0.88 0.83 0.88 0.89
G250-1.95-130-40-15 57.0 46.4 0.87 0.87 0.89 0.89
G250-1.95-160-40-15 43.1 46.6 0.86 0.89 0.90 0.88
Mean 0.951 0.939 0.965 0.979
COV 0.070 0.074 0.067 0.079
F 0.847 0.833 0.862 0.864
_________________________________________________________________
Sections at 500oC
EB ULT DSM
4600 P5 P1.2
G550-0.95-35-20-7 28.6 19.8 0.94 0.93 0.94 0.94
G550-0.95-45-20-7 19.9 20.0 0.93 0.99 0.95 0.93
G550-0.95-55-20-7 14.9 20.3 0.93 1.04 0.97 0.94
G550-0.95-65-20-7 11.9 20.4 0.92 1.07 0.97 0.93
G550-0.95-75-20-7 9.8 20.3 0.91 1.09 0.96 0.92
G450-1.9-60-40-15 133.9 59.9 0.95 0.92 0.95 0.96
G450-1.9-80-40-15 88.3 61.1 0.94 0.93 0.94 0.95
G450-1.9-100-40-15 64.1 62.0 0.93 0.98 0.95 0.95
G450-1.9-130-40-15 44.4 63.4 0.94 1.04 0.98 0.96
G450-1.9-160-40-15 33.5 63.9 0.94 1.07 0.98 0.96
Mean 0.933 1.006 0.959 0.944
COV 0.012 0.064 0.016 0.015
F 0.860 0.901 0.883 0.869
G250-0.95-35-20-7 27.9 8.9 0.99 0.99 1.00 1.02
G250-0.95-45-20-7 19.4 9.0 0.95 0.91 0.95 1.01
G250-0.95-55-20-7 14.6 9.3 0.95 0.91 0.96 1.01
G250-0.95-65-20-7 11.6 9.4 0.95 0.93 0.97 1.01
G250-0.95-75-20-7 9.5 9.6 0.95 0.96 0.98 1.02
G250-1.95-60-40-15 131.9 30.8 1.02 1.02 1.02 1.02
G250-1.95-80-40-15 87.1 31.2 0.93 0.92 0.94 0.96
G250-1.95-100-40-15 63.2 31.6 0.91 0.84 0.91 0.95
G250-1.95-130-40-15 43.8 32.7 0.91 0.89 0.93 0.96
G250-1.95-160-40-15 33.0 33.3 0.91 0.92 0.94 0.96
Mean 0.947 0.929 0.960 0.992
COV 0.038 0.055 0.035 0.030
F 0.864 0.838 0.878 0.909
_________________________________________________________________
Sections at 600oC
EB ULT DSM
4600 P5 P1.2
G550-0.95-35-20-7 20.8 4.8 0.97 0.97 0.97 1.08
G550-0.95-45-20-7 14.5 5.0 0.92 0.91 0.93 1.10
G550-0.95-55-20-7 10.9 5.2 0.92 0.86 0.92 1.14
G550-0.95-65-20-7 8.6 5.3 0.93 0.87 0.93 1.16
G550-0.95-75-20-7 7.1 5.4 0.93 0.90 0.95 1.18
G450-1.9-60-40-15 93.3 17.7 0.96 0.96 0.96 1.00
G450-1.9-80-40-15 61.5 19.3 0.94 0.94 0.95 1.06
G450-1.9-100-40-15 44.6 19.8 0.92 0.87 0.92 1.07
G450-1.9-130-40-15 30.9 20.4 0.91 0.87 0.92 1.09
G450-1.9-160-40-15 23.4 21.0 0.92 0.91 0.95 1.11
Mean 0.932 0.906 0.940 1.099
COV 0.021 0.044 0.019 0.048
F 0.857 0.824 0.865 0.997
G250-0.95-35-20-7 20.4 5.7 0.95 0.95 0.96 0.98
G250-0.95-45-20-7 14.1 6.0 0.93 0.89 0.93 0.99
G250-0.95-55-20-7 10.6 6.2 0.93 0.87 0.93 1.00
G250-0.95-65-20-7 8.4 6.3 0.93 0.91 0.94 1.01
G250-0.95-75-20-7 6.9 6.4 0.94 0.93 0.96 1.02
G250-1.95-60-40-15 91.9 19.7 0.99 0.99 0.99 0.99
G250-1.95-80-40-15 60.6 20.4 0.92 0.92 0.93 0.96
G250-1.95-100-40-15 44.0 20.9 0.91 0.85 0.91 0.96
G250-1.95-130-40-15 30.5 21.9 0.92 0.89 0.93 0.98
G250-1.95-160-40-15 23.0 22.4 0.92 0.93 0.95 0.98
Mean 0.934 0.913 0.943 0.987
COV 0.024 0.044 0.024 0.020
F 0.858 0.830 0.866 0.908
_________________________________________________________________
Sections at 700oC
EB ULT DSM
4600 P5 P1.2
G550-0.95-35-20-7 13.1 3.9 0.95 0.95 0.96 1.06
G550-0.95-45-20-7 9.1 4.0 0.93 0.89 0.93 1.08
G550-0.95-55-20-7 6.8 4.2 0.94 0.89 0.94 1.10
G550-0.95-65-20-7 5.4 4.2 0.94 0.92 0.95 1.12
G550-0.95-75-20-7 4.5 4.3 0.94 0.95 0.97 1.13
G450-1.9-60-40-15 52.7 11.6 1.00 1.00 1.00 1.04
G450-1.9-80-40-15 34.7 12.1 0.94 0.93 0.95 1.05
G450-1.9-100-40-15 25.2 12.4 0.93 0.86 0.92 1.06
G450-1.9-130-40-15 17.5 13.0 0.94 0.91 0.95 1.10
G450-1.9-160-40-15 13.2 13.3 0.94 0.95 0.97 1.11
Mean 0.945 0.925 0.954 1.085
COV 0.021 0.043 0.024 0.029
F 0.869 0.842 0.876 0.995
G250-0.95-35-20-7 12.8 3.4 0.96 0.96 0.96 0.99
G250-0.95-45-20-7 8.9 3.6 0.93 0.90 0.93 1.00
G250-0.95-55-20-7 6.7 3.7 0.94 0.87 0.93 1.02
G250-0.95-65-20-7 5.3 3.8 0.94 0.90 0.95 1.03
G250-0.95-75-20-7 4.4 3.9 0.94 0.93 0.97 1.04
G250-1.95-60-40-15 51.9 10.8 0.96 0.96 0.96 0.96
G250-1.95-80-40-15 34.2 11.3 0.90 0.90 0.91 0.94
G250-1.95-100-40-15 24.8 11.8 0.91 0.85 0.90 0.95
G250-1.95-130-40-15 17.2 12.3 0.91 0.89 0.93 0.97
G250-1.95-160-40-15 13.0 12.6 0.92 0.92 0.95 0.98
Mean 0.931 0.908 0.939 0.988
COV 0.022 0.039 0.024 0.035
F 0.856 0.828 0.862 0.903
_________________________________________________________________
1.000
Mean of FEA/Prediction
G550-0.95-A
G450-1.9-A
G250-0.95-A
G250-1.95-A
0.000
0 100 200 300 400 500 600 700 800
Temperature oC
Figure 6.8: Variations of Mean Values of Type A Section for AS/NZS 4600
1.000
G550-0.95-A
G450-1.9-A
G250-0.95-A
G250-1.95-A
0.000
0 100 200 300 400 500 600 700 800
Temperature oC
Figure 6.9: Variations of Mean Values of Type A Section for the Direct
Strength Method
_________________________________________________________________
1.000
G550-0.95-A
G450-1.9-A
G250-0.95-A
G250-1.95-A
0.000
0 100 200 300 400 500 600 700 800
Temperature oC
Figure 6.10: Variations of Mean Values of Type A Section for BS 5950 Part 5
1.000
G550-0.95-A
G450-1.9-A
G250-0.95-A
G250-1.95-A
0.000
0 100 200 300 400 500 600 700 800
Temperature oC
Figure 6.11: Variations of Mean Values of Type A Section for

Eurocode 3 Part 1.2
_________________________________________________________________
6.2.2 Unlipped Channel Sections (Type B)
The ratios of FEA to code predictions were calculated and tabulated based on the
thickness and strength grades and temperature. The mean values and associated
coefficients of variation for the ratios of FEA results to the predictions of design
standards were calculated at each temperature. As shown in Tables 6.10 to 6.16, the
mean values are always greater than one for predictions of AS/NZS 4600, BS 5950
Part 5 and Eurocode 3 Part 1.2. Hence, predictions are over-conservative. But the
coefficients of variation are significant. This implies that the FEA to code prediction
ratios are scattered. They are increasing with increasing width to thickness ratio of
flanges. The width to thickness ratio has a significant effect on the code predictions
for lipped channels. .
The mean predicted ratios for the direct strength method are closer to one whereas
the coefficients of variation are higher. They are less compared to the other design
standards considered here. However, the direct strength predictions for lower flange
width to thickness ratios are slightly unconservative. This trend is dominant for
lower strength steels. The trend of higher flange width to thickness ratios give higher
capacities as observed in all the predictive methods. Therefore the mean values were
compared at each temperature based on the thickness and strength grades to
investigate the effects of elevated temperatures (see Figures 6.12 to 6.15).
Except for Eurocode 3 Part 1.2 predictions, particularly for higher strength cold-
formed steel columns, the mean values decrease with increasing temperature.
However, they are still safe for all the code predictions. As reported in Chapter 4 and
in the ambient temperature investigation of unlipped channels, the use of smaller k
values for the elastic buckling load calculation of unstiffened elements lead to
uneconomical predictions. Since the direct strength method uses elastic buckling
loads from rational buckling analyses, it eliminates this shortcoming.
As reported under lipped channels, the ratios of finite element analyses to the direct
strength method predictions are comparatively small at lower elevated temperatures
for the low strength steels because of their higher E/fy ratios. Beyond 500oC, the
ratio of FEA and predictions from DSM decreases for higher strength steels and they
are of the same order for low strength steel columns reflecting the fact that the E/fy
_________________________________________________________________
ratios of both low and high strength steels are of the same order. Therefore, it seems
that the direct strength method can not cope with the different E/fy ratios due to the
deterioration of mechanical properties at different rates. Further, the direct strength
method predictions for the sections with smaller flanges (low slenderness) are
unconservative. This is more critical for low strength steels at elevated temperatures
as the slenderness reduces further with increasing temperature due to the slow
reduction of elastic modulus compared to yield stress. However, most of the mean
values are greater than one and the capacity reduction factors (F) are also about 0.85
as recommended by AS/NZS 4600.
Similar to the ambient temperature results, the predictions of AS/NZS 4600 and BS
5950 Part 5 are very conservative. With increasing temperature, the mean values of
FEA results to code predictions are slightly decreased. However, the mean values are
very high even at 700oC. Since the effective width method uses smaller k values
based on the assumption of simply supported connection at the web-flange junction,
the effective width method predictions are conservative for unlipped channels. The
mean values and associated capacity reduction factors (F) are greater than one.
Therefore AS/NZS 4600 and BS 5950 Part 5 can conservatively predict the local
buckling capacities of cold-formed steel columns at elevated temperatures.
_________________________________________________________________
Sections at 100oC

AS/NZS BS5950 EC3
Specimen EB ULT 4600 DSM P5 P1.2
G550-0.95-20-20 20.9 24.9 1.11 1.01 1.18 1.11
G550-0.95-20-30 13.2 29.1 1.26 1.15 1.36 1.27
G550-0.95-20-40 9.7 30.6 1.31 1.17 1.41 1.33
G550-0.95-20-50 7.7 31.2 1.33 1.16 1.42 1.36
G450-1.90-40-40 78.0 85.1 1.09 0.98 1.15 1.07
G450-1.90-40-60 49.5 99.9 1.25 1.12 1.34 1.23
G450-1.90-40-80 36.3 107.0 1.32 1.16 1.42 1.30
G450-1.90-40-100 28.8 109.0 1.33 1.15 1.42 1.37
Mean 1.250 1.113 1.338 1.255
COV 0.078 0.067 0.083 0.089
F 1.099 0.990 1.168 1.087
G250-0.95-20-20 20.4 13.4 1.03 0.87 1.07 1.04
G250-0.95-20-30 9.5 18.1 1.32 1.09 1.42 1.34
G250-0.95-20-40 7.5 19.8 1.42 1.15 1.54 1.45
G250-0.95-20-50 7.5 19.8 1.42 1.15 1.54 1.45
G250-1.95-40-40 76.9 47.5 1.02 0.85 1.05 1.01
G250-1.95-40-60 48.8 57.2 1.18 0.98 1.26 1.17
G250-1.95-40-80 35.8 64.8 1.31 1.08 1.42 1.30
G250-1.95-40-100 28.5 70.4 1.41 1.13 1.52 1.40
Mean 1.264 1.038 1.353 1.270
COV 0.133 0.118 0.150 0.139
F 1.023 0.861 1.060 1.016
_________________________________________________________________
Sections at 200oC

AS/NZS BS5950 EC3
G550-0.95-20-20 18.2 24.1 1.12 1.03 1.19 1.09
G550-0.95-20-30 11.5 27.7 1.26 1.16 1.35 1.23
G550-0.95-20-40 8.4 27.6 1.24 1.12 1.33 1.22
G550-0.95-20-50 6.7 29.1 1.30 1.15 1.38 1.29
G450-1.90-40-40 71.1 82.6 1.08 0.99 1.15 1.05
G450-1.90-40-60 45.0 96.6 1.24 1.12 1.33 1.20
G450-1.90-40-80 33.1 102.0 1.30 1.15 1.39 1.25
G450-1.90-40-100 26.3 103.0 1.30 1.13 1.38 1.31
Mean 1.230 1.106 1.313 1.205
COV 0.069 0.056 0.070 0.076
F 1.092 0.995 1.164 1.062
G250-0.95-20-20 17.7 12.3 1.03 0.88 1.07 1.03
G250-0.95-20-30 8.2 16.7 1.33 1.11 1.44 1.33
G250-0.95-20-40 6.5 18.1 1.43 1.16 1.54 1.43
G250-0.95-20-50 6.5 18.1 1.43 1.16 1.54 1.43
G250-1.95-40-40 70.1 44.9 1.02 0.86 1.06 1.00
G250-1.95-40-60 44.4 54.0 1.18 0.99 1.27 1.16
G250-1.95-40-80 32.6 60.9 1.31 1.08 1.42 1.28
G250-1.95-40-100 25.9 66.5 1.41 1.14 1.53 1.39
Mean 1.268 1.048 1.359 1.256
COV 0.135 0.117 0.149 0.138
F 1.022 0.870 1.067 1.006
_________________________________________________________________
Sections at 300oC

AS/NZS BS5950 EC3
G550-0.95-20-20 15.4 21.6 1.10 1.03 1.18 1.05
G550-0.95-20-30 9.8 24.5 1.22 1.14 1.32 1.17
G550-0.95-20-40 7.2 25.1 1.24 1.14 1.33 1.20
G550-0.95-20-50 5.7 25.4 1.25 1.12 1.33 1.21
G450-1.90-40-40 59.8 76.1 1.07 0.99 1.15 1.01
G450-1.90-40-60 37.9 89.7 1.24 1.14 1.33 1.16
G450-1.90-40-80 27.8 92.7 1.27 1.15 1.36 1.18
G450-1.90-40-100 22.1 92.8 1.26 1.12 1.34 1.23
Mean 1.206 1.104 1.293 1.151
COV 0.064 0.054 0.062 0.068
F 1.077 0.994 1.156 1.023
G250-0.95-20-20 15.1 9.1 1.01 0.85 1.05 1.05
G250-0.95-20-30 7.0 12.5 1.31 1.08 1.42 1.37
G250-0.95-20-40 5.6 13.6 1.41 1.13 1.53 1.47
G250-0.95-20-50 5.6 13.6 1.41 1.13 1.53 1.47
G250-1.95-40-40 58.9 32.8 1.00 0.83 1.02 1.02
G250-1.95-40-60 37.4 39.5 1.15 0.96 1.23 1.18
G250-1.95-40-80 27.5 45.1 1.29 1.05 1.40 1.32
G250-1.95-40-100 21.8 49.2 1.39 1.11 1.51 1.42
Mean 1.246 1.018 1.336 1.288
COV 0.138 0.121 0.157 0.141
F 0.999 0.841 1.031 1.025
_________________________________________________________________
Sections at 400oC

AS/NZS BS5950 EC3
G550-0.95-20-20 12.7 17.4 1.10 1.02 1.18 1.06
G550-0.95-20-30 8.1 19.9 1.23 1.14 1.33 1.18
G550-0.95-20-40 5.9 20.7 1.27 1.16 1.36 1.23
G550-0.95-20-50 4.7 20.7 1.26 1.13 1.34 1.23
G450-1.90-40-40 48.5 57.5 1.08 0.99 1.15 1.04
G450-1.90-40-60 30.7 67.3 1.24 1.13 1.34 1.19
G450-1.90-40-80 22.6 70.3 1.28 1.15 1.38 1.23
G450-1.90-40-100 17.9 70.7 1.28 1.12 1.37 1.28
Mean 1.218 1.105 1.306 1.180
COV 0.066 0.057 0.068 0.073
F 1.084 0.992 1.161 1.043
G250-0.95-20-20 12.4 7.6 1.14 0.94 1.16 1.21
G250-0.95-20-30 5.8 9.3 1.30 1.06 1.41 1.40
G250-0.95-20-40 4.6 9.8 1.35 1.07 1.46 1.45
G250-0.95-20-50 4.6 9.8 1.35 1.07 1.46 1.45
G250-1.95-40-40 47.8 25.9 1.00 0.83 1.02 1.03
G250-1.95-40-60 30.3 30.0 1.11 0.92 1.18 1.14
G250-1.95-40-80 22.3 32.2 1.17 0.95 1.26 1.20
G250-1.95-40-100 17.7 34.3 1.23 0.97 1.33 1.27
Mean 1.206 0.976 1.285 1.269
COV 0.103 0.088 0.124 0.121
F 1.025 0.848 1.056 1.048
_________________________________________________________________
Sections at 500oC

AS/NZS BS5950 EC3
G550-0.95-20-20 10.0 11.4 1.10 0.99 1.17 1.11
G550-0.95-20-30 6.3 13.1 1.23 1.11 1.33 1.25
G550-0.95-20-40 4.7 13.9 1.29 1.14 1.39 1.33
G550-0.95-20-50 3.7 14.1 1.30 1.13 1.39 1.35
G450-1.90-40-40 37.2 33.8 1.06 0.93 1.11 1.09
G450-1.90-40-60 23.6 40.1 1.22 1.07 1.31 1.26
G450-1.90-40-80 17.3 43.7 1.31 1.13 1.41 1.36
G450-1.90-40-100 13.8 44.7 1.33 1.12 1.43 1.44
Mean 1.230 1.078 1.318 1.274
COV 0.082 0.072 0.089 0.096
F 1.076 0.954 1.142 1.093
G250-0.95-20-20 9.8 5.1 1.04 0.85 1.06 1.13
G250-0.95-20-30 4.5 6.6 1.26 1.01 1.37 1.39
G250-0.95-20-40 3.6 7.3 1.36 1.07 1.48 1.51
G250-0.95-20-50 3.6 7.3 1.36 1.07 1.48 1.51
G250-1.95-40-40 36.7 17.5 1.00 0.82 1.01 1.06
G250-1.95-40-60 22.3 20.6 1.12 0.92 1.19 1.20
G250-1.95-40-80 17.1 23.1 1.22 0.98 1.32 1.32
G250-1.95-40-100 13.6 24.8 1.29 1.01 1.41 1.40
Mean 1.206 0.966 1.290 1.315
COV 0.115 0.098 0.142 0.129
F 1.006 0.827 1.026 1.071
_________________________________________________________________
Sections at 600oC

AS/NZS BS5950 EC3
G550-0.95-20-20 7.3 2.7 0.94 0.81 0.94 1.25
G550-0.95-20-30 4.6 3.3 1.08 0.86 1.13 1.49
G550-0.95-20-40 3.4 3.8 1.20 0.94 1.29 1.71
G550-0.95-20-50 2.7 4.2 1.30 1.00 1.42 1.89
G450-1.90-40-40 25.9 11.7 1.06 0.90 1.06 1.33
G450-1.90-40-60 16.4 13.9 1.19 0.96 1.26 1.54
G450-1.90-40-80 12.1 13.6 1.13 0.90 1.23 1.49
G450-1.90-40-100 9.6 17.2 1.41 1.10 1.54 1.96
Mean 1.164 0.934 1.234 1.583
COV 0.126 0.096 0.156 0.160
F 0.953 0.802 0.955 1.216
G250-0.95-20-20 7.1 3.2 0.97 0.80 0.98 1.07
G250-0.95-20-30 3.3 4.4 1.23 0.98 1.33 1.39
G250-0.95-20-40 2.6 4.9 1.34 1.05 1.46 1.52
G250-0.95-20-50 2.6 4.9 1.34 1.05 1.46 1.52
G250-1.95-40-40 25.6 11.2 0.96 0.80 0.97 1.03
G250-1.95-40-60 16.2 13.5 1.10 0.89 1.16 1.20
G250-1.95-40-80 11.9 14.8 1.17 0.94 1.27 1.29
G250-1.95-40-100 9.5 16.7 1.30 1.02 1.42 1.43
Mean 1.176 0.941 1.256 1.306
COV 0.131 0.109 0.160 0.147
F 0.954 0.792 0.964 1.029
_________________________________________________________________
Sections at 700oC

AS/NZS BS5950 EC3
G550-0.95-20-20 4.6 2.2 0.97 0.79 0.99 1.22
G550-0.95-20-30 2.9 2.6 1.12 0.91 1.19 1.44
G550-0.95-20-40 2.1 3.0 1.25 1.00 1.35 1.64
G550-0.95-20-50 1.7 3.3 1.34 1.05 1.46 1.80
G450-1.90-40-40 14.6 6.6 0.97 0.81 0.98 1.19
G450-1.90-40-60 9.3 7.3 1.02 0.83 1.08 1.28
G450-1.90-40-80 6.8 9.0 1.23 0.98 1.33 1.56
G450-1.90-40-100 5.4 9.5 1.27 1.00 1.39 1.71
Mean 1.146 0.921 1.221 1.480
COV 0.127 0.109 0.154 0.158
F 0.936 0.776 0.950 1.142
G250-0.95-20-20 4.5 1.9 0.95 0.80 0.96 1.07
G250-0.95-20-30 2.1 2.7 1.22 0.97 1.32 1.41
G250-0.95-20-40 1.7 2.9 1.33 1.04 1.45 1.54
G250-0.95-20-50 1.7 2.9 1.33 1.04 1.45 1.54
G250-1.95-40-40 14.4 6.2 0.94 0.78 0.95 1.01
G250-1.95-40-60 9.5 7.5 1.07 0.86 1.14 1.17
G250-1.95-40-80 6.7 8.5 1.19 0.95 1.29 1.30
G250-1.95-40-100 5.3 9.3 1.29 1.01 1.40 1.42
Mean 1.165 0.931 1.245 1.308
COV 0.137 0.113 0.165 0.157
F 0.934 0.780 0.946 1.010
_________________________________________________________________
1.000
G550-0.95-B
G450-1.9-B
G250-0.95-B
G250-1.95-B
0.000
0 100 200 300 400 500 600 700 800
o
Temperature C
Figure 6.12: Variations of Mean Values of Type B Section for AS/NZS 4600
1.000
G550-0.95-B
G450-1.9-B
G250-0.95-B
G250-1.95-B
0.000
0 100 200 300 400 500 600 700 800
Temperature oC
Figure 6.13: Variations of Mean Values of Type B Section for BS 5950 Part 5
_________________________________________________________________
1.000
G550-0.95-B
G450-1.9-B
G250-0.95-B
G250-1.95-B
0.000
0 100 200 300 400 500 600 700 800
Temperature oC
Figure 6.14: Variations of Mean Values of Type B Section for Direct Strength
Method
1.000
G550-0.95-B
G450-1.9-B
G250-0.95-B
G250-1.95-B
0.000
0 100 200 300 400 500 600 700 800
o
Temperature C
Figure 6.15: Variations of Mean Values of Type B Section for Eurocode 3 Part
1.2 Method
_________________________________________________________________
6.3 Conclusions
A total of 44 ambient temperature finite element simulations were carried out to
assess the accuracy of ambient temperature design rules and their limitations. It was
found that the effective width method is not accurate for lipped channels with very
high web width to thickness ratios. In the case of unlipped channels, all the design
codes (AS/NZS 4600, Eurocode 3 Part 1.2, BS5950 Part 5 and DSM) predict
conservative results while the direct strength method predictions are within the
acceptable limit. However, its predictions for sections with smaller flanges,
particularly for low strength steels are slightly unconservative.
A total of 252 elevated temperature finite element simulations were carried out to
investigate the local buckling behaviour of cold-formed steel compression members.
These analyses were carried out on lipped and unlipped channels. Simulations were
carried out at ambient and elevated temperatures up to 700oC at 100oC intervals.
Simulated FEA results were compared with the local buckling capacity predictions
of AS/NZS 4600, BS 5950 Part 5, Eurocode 3 Part 1.2, and the direct strength
method using the reduced mechanical properties.
The effective width method recommended by AS/NZS 4600 and BS 5950 Part 5
predict conservative results. Since the ideal model uses pure pinned-end conditions,
FEA results are slightly less than those corresponding to fixed-end conditions. This
causes lower values of FEA to code predictions, particularly for lipped channels.
However, they are still within acceptable limits. The direct strength method
predictions for the lipped channels made of high strength steel are slightly
conservative than the other design standards. This trend is more dominant for the
sections with larger web element because of lower local buckling capacity prediction
for the members with larger web elements than the smaller web elements. The direct
strength method prediction is a function of both elastic buckling load and yield
capacity. Since the elastic buckling load reduces with increasing web slenderness,
DSM predictions for lipped channels reduce with increasing web slenderness.
However, it is not the case for unlipped channels even when the flange slenderness is
increased.
Since Eurocode 3 Part 1.2 assumes the effective area calculated at ambient
temperature for all temperatures, the effective area increase at elevated temperatures
_________________________________________________________________
due to the slow reduction of elastic modulus compared to the yield stress is not
included in the calculations. Therefore the predictions are conservative for the
temperatures beyond 500oC, particularly for high strength steels.
In relation to the parametric study of unlipped channels, DSM predictions for high
strength steel sections made with smaller flanges were slightly unconservative. This
is more dominant for low strength steels and at higher temperatures. All other design
standards considered in this study gave conservative results. Code predictions were
high for sections with slender flanges. Further, the direct strength method in its
current format is struggling to cope with the fire design of cold-formed steel
compression members made of steels with higher E/fy ratios at elevated temperatures
Based on the above discussions, it can be concluded that ambient temperature design
guidelines can be used for local buckling capacity prediction of cold-formed steel
columns at elevated temperatures. Predictions are reasonably accurate for lipped
channels while they are conservative for unlipped channels. Further, AS/NZS 4600
design method, which is similar to Eurocode 3 Part 1.3 and North American
Specification, and BS 5950 Part 5 design method are more suitable and conservative
for the prediction of local buckling capacities at elevated temperatures.
_________________________________________________________________
7 Experimental Study of Cold-formed Steel
Compression Members Subject to Flexural-
torsional Buckling Behaviour
Cold-formed steel slender columns are generally subjected to flexural and flexural-
torsional buckling modes. Flexural-torsional buckling is the interaction of flexural
and torsional buckling. Flexural and flexural-torsional buckling modes are more
common in open cold-formed steel column sections. Design standards provide
guidelines for columns subject to flexural and flexural-torsional buckling modes at
ambient temperature. However, there are no specific design guidelines for elevated
temperature conditions. However, Eurocode 3 Part 1.2 (ECS, 2005) recommends the
same design guidelines, which were originally developed for Class 1, 2 and 3
sections representing hot-rolled steel sections.
In order to investigate the flexural-torsional buckling behaviour, an experimental

study was carried out at ambient and elevated temperatures. Suitable test sections
and thickness were selected based on the standard sections, thicknesses and grades
used in structural and architectural applications and available literature. Test section
dimensions and specimen lengths were selected based on a number of preliminary
analyses using a finite strip analyses program CUFSM and a finite element analyses
program ABAQUS so that flexural-torsional buckling govern the member behaviour.
In the case of elevated temperature conditions, reduced mechanical properties taken
from Ranawaka and Mahendran (2009) and Dolamune Kankanamge (2009) were
used in the analyses. Preliminary analyses showed that the member lengths of 1600
mm or higher gave flexural-torsional buckling with the selected sections. Therefore
two specimen lengths, 1800 mm and 2800 mm were selected for the tests. A furnace
consisting of three 1000 mm long segments was built for this research. The first test
series consisted of 2800 mm columns while the second test series consisted of 1800
mm columns. Tests on these two were carried out at ambient and elevated
temperatures up to 700oC at 100oC intervals (20oC, 200oC, 300oC, 400oC, 500oC,
600oC, 700oC), giving a total of 39 tests. Experimental results were used to
investigate the accuracy of code predictions and to validate the finite element model
that simulates flexural and flexural-torsional buckling behaviour of cold-formed steel
_________________________________________________________________
Behaviour and Design of Cold-formed Steel Compression Members at Elevated Temperatures.
7-1
columns. This chapter presents the details of experimental study of slender cold-
formed steel compression members and the results.
7.1 Test Specimens
Light gauge cold-formed steel thicknesses vary from 0.42 mm to 3 mm and are
available in two strength grades, namely high strength (G450, G500 and G550) and
low strength (G200, G250 and G300). Therefore three grades and thicknesses, G550-
0.95 mm, G450-1.90 mm and G250-1.95 mm, were selected to represent the light
gauge cold-formed steel domain. The most common cold-formed steel column
section is the lipped channel section. Initial analyses of lipped-sections were carried
out using the CUFSM. This program can only analyse pin ended columns. However,
it was decided to adopt fixed ends for the tests because perfect pin-end conditions are
difficult to model in experiments. It becomes more difficult at elevated temperatures.
However, CUFSM output could still be used by assuming that the effective length is
0.5 times the actual length. In the analyses of 2800 mm long columns, CUFSM
results were obtained for the 1400 mm long column. Figure 7.1 shows a buckling
plot of lipped channel sections as obtained from CUFSM. Lengths of the specimens
were selected avoiding local or distortional buckling regions. The selected sections
from CUFSM analyses were further analysed using ABAQUS finite element
program to ensure the occurrence of desired buckling modes. An independent check
was carried out using the AS/NZS 4600 (SA, 2005) design rules at ambient
temperature. In both test series, the same cross section was used with changing
length. The nominal dimensions of the selected test specimens are given in Table
7.1.
Table 7.1: Nominal Dimensions of Test Specimens
Nominal Section dimensions Length of

Steel (mm) specimens (mm)
thickness
Grade
(mm) Web Flange Lip Series 1 Series 2
0.95 550 55 35 9 1800 2800
1.95 250 75 50 15 1800 2800
1.90 450 75 50 15 1800 2800
_________________________________________________________________
7-2
Range of pure global buckling
Figure 7.1: Buckling Plot of Lipped Channels
Due to the coating of cold-formed steel, the base metal thickness was also measured.
For this purpose, steel specimens are usually immersed in diluted hydrochloric acid
to remove the coating (see Chapter 3). In this test series of long columns, the average
base metal thicknesses obtained for local buckling tests reported in Chapter 4, were
used because the variation of the base metal thickness was considered small. Table
7.2 shows the base metal thickness of selected cold-formed steels.
Table 7.2: Base Metal Thickness of Selected Cold-formed Steels
Steel grade and Average base metal

nominal thickness thickness (mm)
G550 – 0.95 0.95
G450 – 1.90 1.88
G250 – 1.95 1.95
All the test specimens were labeled according to the grade, thickness, test
temperature and length of specimen and specimen number as shown next.
_________________________________________________________________
7-3
G250 – 1.95 – 1800 – 20
Test temperature
Specimen length
Nominal thickness of specimens
Grade of steel
External dimensions of all the specimens were measured using a vernier caliper (to
measure cross-sectional dimensions) and steel tape (to measure length) prior to the
tests. Tables 7.3 and 7.4 give the measured cross-sectional dimensions of 39 tested
specimens.
Table 7.3: Measured External Dimensions of 1800 Test Series Specimens
Specimen Web Flange Lip Length

(mm) (mm) (mm) (mm)
G550 – 0.95 – 1800 – 20 54.94 34.88 8.00 1740
G550 – 0.95 – 1800 – 200 54.85 34.81 7.93 1740
G550 – 0.95 – 1800 – 300 55.02 34.78 7.67 1740
G550 – 0.95 – 1800 – 400 54.97 34.53 7.98 1740
G550 – 0.95 – 1800 – 600 54.73 34.76 7.63 1740
G550 – 0.95 – 1800 – 700 55.19 34.81 7.78 1740
G250 – 1.95 – 1800 – 20 74.82 50.06 14.87 1740
G250 – 1.95 – 1800 – 200 74.79 49.81 15.11 1740
G250 – 1.95 – 1800 – 300 74.79 49.93 15.07 1740
G250 – 1.95 – 1800 – 400 74.92 50.19 14.75 1740
G250 – 1.95 – 1800 – 500 74.93 49.97 15.32 1740
G250 – 1.95 – 1800 – 600 75.39 50.04 15.03 1740
G250 – 1.95 – 1800 – 700 74.79 49.71 14.86 1740
G450 – 1.90 – 1800 – 20 74.67 49.94 14.51 1740
G450 – 1.90 – 1800 – 200 75.43 49.83 15.06 1740
G450 – 1.90 – 1800 – 300 74.83 49.92 14.83 1740
G450 – 1.90 – 1800 – 400 74.59 49.65 14.73 1740
G450 – 1.90 – 1800 – 500 74.96 49.78 14.82 1740
G450 – 1.90 – 1800 – 600 74.93 49.81 14.86 1740
G450 – 1.90 – 1800 – 700 74.74 49.86 14.84 1740
_________________________________________________________________
7-4
Table 7.4: Measured External Dimensions of 2800 Test Series Specimens
Specimen Web Flange Lip Length

(mm) (mm) (mm) (mm)
G550 – 0.95 – 2800 – 20 54.82 34.78 7.82 2820
G550 – 0.95 – 2800 – 200 54.83 34.84 7.67 2820
G550 – 0.95 – 2800 – 400 54.75 34.63 7.81 2820
G550 – 0.95 – 2800 – 500 55.00 34.73 7.81 2820
G550 – 0.95 – 2800 – 600 54.76 34.81 7.68 2820
G550 – 0.95 – 2800 – 700 54.92 34.88 7.57 2820
G250 – 1.95 – 2800 – 20 74.97 49.88 14.69 2820
G250 – 1.95 – 2800 – 200 74.83 49.89 15.06 2820
G250 – 1.95 – 2800 – 300 75.02 49.83 14.91 2820
G250 – 1.95 – 2800 – 400 74.87 49.79 14.88 2820
G250 – 1.95 – 2800 – 500 74.65 49.77 14.89 2820
G250 – 1.95 – 2800 – 600 75.06 49.71 14.63 2820
G250 – 1.95 – 2800 – 700 74.58 49.61 15.02 2820
G250 – 1.95 – 2800 – 600R* 74.84 49.89 14.91 2820
G250 – 1.95 – 2800 – 700 R* 74.96 49.59 14.67 2820
G450 –1.90 – 2800 – 20 74. 27 49.78 14.92 2820
G450 – 1.90 – 2800 – 200 74.59 49.72 15.11 2820
G450 – 1.90 – 2800 – 400 74.94 49.93 14.81 2820
G450 – 1.90 – 2800 – 500 74.83 49.87 14.69 2820
G450 – 1.90 – 2800 – 600 74.61 49.84 14.91 2820
G450 – 1.90 – 2800 – 700 74.73 49.96 14.84 2820
G450 – 1.90 – 2800 – 700R* 74.68 49.83 14.81 2820
Note:- R* Indicates repeated test
_________________________________________________________________
7-5
7.2 Ambient and Elevated Temperature Test Set-up
Unlike the local buckling tests reported in Chapter 4, both ambient and elevated
temperature tests of long columns were carried out inside the furnace. There was no
standard testing machine in the QUT structural laboratory to test long columns.
Therefore a special test set-up was designed and built to test long columns of
different heights inside the furnace. The test set-up consists of a reaction frame,
furnace, control system of the furnace, loading set-up and hydraulic loading system.
Reaction Frame
The reaction frame consists of a 2.4 m long 900WB282 cross-head mounted between
two 5.0 m long 310UC137 columns (see Figure 7.2). These two columns were
rigidly attached to the strong floor to transfer the load, and were braced in both
directions at the top and mid height.
Electric Furnace and Control System

A new electrical furnace was designed to fulfill the requirements of testing cold-
formed steel columns at different heights and at different temperatures up to 1000oC.
The furnace consists of three segments of 1 m height and hence it can be assembled
as a 3 m, 2 m or 1 m height furnace according to the specimen height (see Figures
7.2 and 7.3). This furnace is heated using the heating coils attached to three sides of
its wall but not on its doors (see Figure 7.4). The heating of each segment can be
controlled separately. Each segment is connected to a Eurotherm control system and
has a thermo couple located at the centre (see Figure 7.5). Temperature of each
segment was measured by the thermocouple and accordingly Eurotherm controls the
power supply to these segments to maintain constant temperature. Appropriate
measurements were taken to minimize the hazards. To avoid electric shock, all the
doors are provided with door switches. The furnace does not energize until the doors
are fully closed. Since the maximum safe working temperature of this furnace is
1000oC, its control system (see Figure 7.6) is programmed to shutdown the furnace
automatically when it exceeds the safe working temperature. Figures 7.7 and 7.8
show the overviews of electric furnace with a test column located inside it.
_________________________________________________________________
7-6
Cross-head
900WB282
Three segment
furnace
Bracings
Eurotherm
temperature
control system
Figure 7.2: Three Segment Furnace and Temperature Control System
External
thermometer for
independent
temperature
measurements
(a) Door switch (b) Furnace and its attachments
Figure 7.3: Two Segment Furnace
_________________________________________________________________
7-7
Coils
Door on
walls
(a) View from top of a furnace segment (b) View with door open
Figure 7.4: Heating Coil Arrangement of the Furnace
Thermocouples
Separate power
supply to the
segments
Figure 7.5: Location of Thermocouples
_________________________________________________________________
7-8
Eurotherms for
each segment
Over
temperature
indicators
Power outlets
for each
segment
Isolator for the
heating elements
Figure 7.6: Control Panel of the Furnace
Figure 7.7: Two Segment Furnace with a Test Column
_________________________________________________________________
7-9
Figure 7.8: Three Segment Furnace with a Test Column
Loading Arrangement
The loading arrangement consists of two loading shafts at the top and bottom and a
hydraulic loading system. The hydraulic loading system consists of a hydraulic
pump, a pressure transducer and a hydraulic jack. The parts of top and bottom
loading shafts which are to be located inside the furnace were made with 253MA
stainless steel to achieve a better performance at elevated temperatures up to 1100
o
C. Loading shafts were made of 70 mm diameter 253MA stainless steel rods to
minimise the effects of loading shaft deformations. The upper loading shaft was
inserted through the opening provided in the furnace and then fixed at the top of the
furnace. It was then connected to the cross head through a 50 mm thick plate
attached to the cross-head and tightened using a locknut. The other end of the upper
loading shaft located inside the furnace was rigidly fixed to a 150 mm diameter and
25 mm thick 253MA stainless steel plate (see Figure 7.9). This plate was used to fix
the specimen to the loading shaft. Three 10 mm diameter holes were provided in the
plate for the specimen fixing.
_________________________________________________________________
7-10
Cross-head
Upper loading
shaft Stainless steel
plate to fix
specimens
50 mm thick
steel plate to
connect upper
loading shaft
Locknut
Figure 7.9: Loading Arrangement at the Upper End of the Furnace
At the bottom, a special loading arrangement was made to fit under the furnace (see
Figure 7.10). A hydraulic jack was used to apply an axial compression load to the
specimen. A 70 mm diameter shaft and a 150 mm diameter and 25 mm thick plate
made of 253MA stainless steel, as used for the upper end, were used at the lower end
to fix the specimen. The loading shaft was located inside a guide tube which was
connected to the levelling base through three additional supports. The levelling base
was made of a 40 mm thick and 500 mm diameter circular steel plate to avoid any
deflection. This levelling base plate was mounted on three 25 mm threaded bars. By
screwing the nuts on each of the levelling bars, the base plate could be levelled so
that the guide tube is vertical. The bottom loading shaft was located on top of the
hydraulic jack, which was positioned at the centre of the levelling base. The
hydraulic jack was then connected to a hydraulic pump through a pressure transducer
(see Figure 7.11). The pressure transducer measured the pressure in the hydraulic
_________________________________________________________________
7-11
jack, which was then converted into an axial load. Since the mechanical properties of
cold-formed steels drop significantly at elevated temperatures above 500oC, expected
failure loads have a significant variation. Therefore using the same hydraulic jack to
measure all the failure loads at ambient and elevated temperatures is not accurate.
Three hydraulic jacks with different capacities (5, 10 and 20 ton) were used with a
single pressure transducer (see Figure 7.12 (a)). All the jacks were calibrated with
this pressure transducer to measure the induced force with different hydraulic
pressures. Since 5 and 10 ton jacks are unstable due to small contact areas, jack
bases were made for them to stabilise them during loading (see Figure 7.12 (b)).
Guidance tube
Guidance tube
supports (3) Bottom loading
shaft
Hydraulic jack LVDT to

and jack base measure axial
shortening
Levelling base
Levelling bars (3)
Figure 7.10: Loading Arrangement at the Bottom End of the Furnace
_________________________________________________________________
7-12
Pressure transducer
Hydraulic pump
Figure 7.11: Hydraulic Pump and Pressure Transducer
20 10 5
(a) Jacks (20, 10 & 5 ton) (b) Jack Bases for 10 & 5 ton Jacks
Figure 7.12: Jacks and Jack Bases Used for the Tests
In the case of compression tests with long columns subjected to global buckling, end
conditions are very important. The length of specimens required can be reduced by
using pin-ends. However, a pure pin-end needs more sophisticated machining and
good lubricants and is more difficult to achieve at elevated temperatures. Since it is
difficult to achieve a pure pin-end condition inside the furnace, all the tests were
carried out using a fixed-end condition. Fixed end conditions need long specimens.
Therefore tests were limited to 2800 mm and 1800 mm. Three segments of the
furnace were used for longer columns while two segments were used for shorter
columns. To achieve fixed-end support, special end plates were made to fit into the
end of the specimens (see Figure 7.13). A groove of 12 mm deep and 10 mm wide of
the shape of specimen cross-section was made on 15 mm thick circular steel plate.
Geometric centre of the cross-section of the group was made to coincide with the
centre of the plate. A Rectangular Hollow Section (RHS) of 2 mm thick and 15 mm
height was then welded to the plate while surrounding the grove to enhance the grip
_________________________________________________________________
7-13
to the end. The specimen was placed on the grove and grove was filled with 165
procreate coil grout mixed with water. This grout gets fully hardened within 24
hours. The end plates of the specimen were then bolted to the end plate of the
loading shafts.
End plates
Ceramic fibre
insulations
Figure 7.13: End Plates used for Making Fixed End Support
7.3 Test Procedure
Cold-formed steel channel columns were loaded inside the furnace to find the
ultimate compression capacity both at ambient and elevated temperatures. The test
program was divided into two series, 1800 mm long columns and 2800 mm long
columns, based on the length of test specimens. In the case of the longer column
series (2800 mm long columns), all three furnace segments were assembled together
while for the shorter column series (1800 mm long columns) only two furnace
segments were assembled.
_________________________________________________________________
7-14
Specimens were fixed to the end plates by grouting and then at least 24 hours was
allowed to ensure sufficient hardening. This specimen was then placed between the
two loading shafts and bolted to form fixed ends. Since steel is a good heat
transferring media, it was needed to prevent the heat loss from specimen through 70
mm diameter loading shafts. Therefore ceramic fiber packing was placed between
the end plate and the loading shaft for the elevated temperature tests (see Figure
7.13). This insulation packing reduces the heat loss from the specimen. Due to the
heat loss from either ends, it took little longer to settle the temperature at the ends of
the specimen than at the middle. Therefore thermocouples were attached to either
ends at closer intervals and then the temperature of the specimen was monitored
particularly for tests at higher temperatures.
All the doors were closed tightly before energizing the power of the furnace. All the
doors were fitted with door switches for safety reasons. The furnace does not
energize until doors are fully closed. After closing the doors, the furnace was
connected to the three phase power supply and temperature of all segments was set
to the target temperature. Once the temperature of the specimen reached the target
temperature, the specimen was allowed another 10 minutes to ensure uniform
temperature throughout the column before loading. During the heating phase, special
care was taken to monitor the load applied to the specimen due to thermal expansion
effects. Once a steady state was reached, load was applied slowly using the hydraulic
jack until failure.
Out of plane deflections of the specimens were needed at the mid height in either
direction to validate the finite element models. However, it was not possible to
employ LVDTs inside the furnace. Therefore stainless steel cables were used to
connect LVDTs, which were kept outside the furnace and the specimen to measure
deflections (see Figure 7.14). For this purpose of measuring the deflection, 10 mm
diameter special openings have been provided on the furnace at mid height. Axial
shortening was measured using a 50 mm LVDT attached to the bottom loading shaft
(see Figure 7.10). All the data channels from pressure transducer and LVDTs were
connected to the data logger and the data logger plotted the load-deflection curves.
Ambient temperature tests were also carried out inside the furnace while keeping the
doors open.
_________________________________________________________________
7-15
LVDT
LVDT
cable
Opening for
Stainless steel inserting cable
cable
Figure 7.14: Measurement of Out-of-Plane Displacement
7.4 Measurements of Geometric Imperfections
Imperfection is also an important measurement, particularly for the finite element

simulations. Therefore imperfections of all the specimens were measured along the
specimens on all the surfaces except lips. Since the expected buckling mode was
global buckling, only the global imperfection was measured. The measuring table
fitted with a laser distance measuring instrument was used to measure the
imperfections of the specimens (see Figure 7.15). Each specimen was placed on the
lower support and then leveled using the legs of the measuring table. The laser head
was then focused to obtain the accurate distance from the face of the specimen. The
first two readings were taken from either ends, and if they were different, further
leveling was made until both ends of the specimen were at the same elevation. Once
the specimen was perfectly leveled (ie. both ends give the same reading), the
readings at one end was made zero. The moving head with the laser sensor was then
moved parallel to the specimen while recording the imperfections. At every 100 mm
intervals, imperfection readings on flanges and web along the specimens were taken
separately. This laser sensor was connected to a data logger and all the imperfection
readings were plotted against the specimen length (see Figure 7.16). Tables 7.5 and
7.6 show the measured imperfections. In most cases, the maximum imperfections
were observed to be on the web. The measured imperfections are significantly less
than the tolerance value of L/1000 recommended by AS 4100 (SA, 1998). The mean
imperfection values of G550-0.95 showed higher values than the other two
_________________________________________________________________
7-16
thicknesses. Since the measured imperfections were less L/1000, the test ultimate
compression capacities may exceed code predictions.
Moving head fitted with

laser head
Specimen
(a) Measuring Table (b) Laser Head
Figure 7.15: Imperfection Measurement Device
0.40
0.20 Length (mm)
0.00
0 500 1000 1500 2000 2500 3000
.
-0.20
-0.40
Imperfection (mm)
-0.60
-0.80
F1-2 F1-
-1.00
W1
-1.20 W2
F1-1 F1 - 2
W3
W-1 W -2
-1.40 F1-2 F1-1
W-3 F2-1
-1.60 F2-2
Figure 7.16: Variation of Imperfection along Specimen G450-1.90-2800-20
_________________________________________________________________
7-17
Table 7.5: Measured Maximum Imperfections of 1800 Test Series Specimens
Specimen Maximum Imperfection

Location (mm) As a ratio of L
G550 – 0.95 – 1800 – 20 W 0.70 L/2485

G550 – 0.95 – 1800 – 200 W 0.86 L/2023
G550 – 0.95 – 1800 – 300 W 0.64 L/2718
G550 – 0.95 – 1800 – 400 W 0.69 L/2521
G550 – 0.95 – 1800 – 600 F 0.24 L/7250
G550 – 0.95 – 1800 – 700 W 0.48 L/3625
Mean L/3437
G250 – 1.95 – 1800 – 20 W 0.68 L/2558
G250 – 1.95 – 1800 – 200 W 0.43 L/4046
G250 – 1.95 – 1800 – 300 W 0.46 L/3782
G250 – 1.95 – 1800 – 400 W 0.58 L/3000
G250 – 1.95 – 1800 – 500 F 0.48 L/3625
G250 – 1.95 – 1800 – 600 F 0.32 L/5437
G250 – 1.95 – 1800 – 700 W 0.38 L/4578
Mean L/3860
G450 – 1.90 – 1800 – 20 W 0.59 L/2949
G450 – 1.90 – 1800 – 200 W 0.89 L/1955
G450 – 1.90 – 1800 – 300 W 0.58 L/3000
G450 – 1.90 – 1800 – 400 W 0.40 L/4350
G450 – 1.90 – 1800 – 500 F 0.44 L/3954
G450 – 1.90 – 1800 – 600 F 0.35 L/4971
G450 – 1.90 – 1800 – 700 W 0.45 L/3866
Mean L/3577
Note:- W= Web, F=Flange, L=Length of the specimen
_________________________________________________________________
7-18
Table 7.6: Measured Maximum Imperfections of 2800 Test Series Specimens
Specimen Maximum Imperfection

Location (mm) As a ratio of L
G550 – 0.95 – 2800 – 20 W 1.95 L/1446

G550 – 0.95 – 2800 – 200 W 1.94 L/1453
G550 – 0.95 – 2800 – 400 W 2.25 L/1253
G550 – 0.95 – 2800 – 500 W 3.08 L/915
G550 – 0.95 – 2800 – 600 W 1.92 L/1468
G550 – 0.95 – 2800 – 700 W 1.48 L/1905
L/1406
G250 – 1.95 – 2800 – 20 W 0.69 L/4086
G250 – 1.95 – 2800 – 200 W 0.74 L/3810
G250 – 1.95 – 2800 – 300 W 0.55 L/5127
G250 – 1.95 – 2800 – 400 F 0.66 L/4272
G250 – 1.95 – 2800 – 500 W 0.58 L/4862
G250 – 1.95 – 2800 – 600R* W 0.58 L/4862
G250 – 1.95 – 2800 – 700R* F 0.39 L/7230
L/4892
G450 –1.90 – 2800 – 20 W 1.36 L/2073
G450 – 1.90 – 2800 – 200 W 1.10 L/2563
G450 – 1.90 – 2800 – 400 W 1.52 L/1855
G450 – 1.90 – 2800 – 500 W 1.18 L/2389
G450 – 1.90 – 2800 – 600 W 0.95 L/2968
G450 – 1.90 – 2800 – 700R* W 0.55 L/5127
L/2829
Note:- W= Web, F=Flange, L=Length of the specimen
7.5 Observations and Results
In the early stage of the long specimen test series, it was observed that the test results
showed higher capacities than the expected load for the tests at elevated
temperatures. The preliminary suspect was the temperature distribution along the
_________________________________________________________________
7-19
specimen. Therefore an elevated temperature test was carried out to obtain the
temperature profile along the specimen. Thermocouples were attached to the
specimen at closer intervals along the specimen (see Figure 7.17). The furnace was
then set to different elevated temperatures and energized. A considerable
temperature difference was observed along the specimen during the heating phase.
The difference in the temperature along the specimen length reduced with time.
However, settlement time was significantly higher for the test temperatures at and
above 500oC. For the tests below 500oC, temperature of the specimens became
uniform within few minutes. Further investigation revealed that 70 mm loading
shafts conducted the heat from both ends and thus caused lower temperatures at the
ends of the specimen than the middle, particularly at the beginning of the heating
phase. Thereafter ceramic fiber insulation packing was used between the plate on the
loading shaft and the end plate of the specimen (see Figure 7.13). In addition to that,
the specimen temperature was monitored, particularly at the ends at closer intervals
by attaching thermocouples. Once a uniform temperature was ensured in the
specimen, the load was applied. Affected tests due to non-uniform temperatures at
the ends were repeated, and the repeated test results were used in the numerical
study.
Figure 7.17: Measurement of Temperature Profile along the Specimen
As expected, almost all the specimens failed in flexural-torsional buckling (see

Figures 7.18 to 7.20). However, only the G550-1800-300 specimen showed
interaction of flexural-torsional buckling with local buckling (see Figure 7.21). Since
_________________________________________________________________
7-20
the elevated temperature tests were carried out inside the furnace, the deflected shape
could not be observed during the test. However, the failure pattern of ambient
temperature tests was observed because they were carried out with the doors open.
The failure pattern of specimens tested at elevated temperatures was determined
from the load-deflection curves. Larger deformations, particularly the out-of-plane
deflection and twisting, were observed with the tests at higher temperatures. Larger
deformations were due to reduced elastic modulii at elevated temperatures and thus
lower flexural and torsional rigidities. The data logger was programmed to draw
axial load versus axial shortening and axial load versus out of plane deflection
curves. From these curves, the maximum load was obtained as the failure load
(ultimate load). Figures 7.22 to 7.25 show typical load-deflection curves obtained for
the specimens at ambient and elevated temperatures for the two series of tests while
Tables 7.7 and 7.8 give the ultimate loads. Appendix D gives all the load-deflection
curves obtained for the flexural-torsional buckling tests.
At 20oC At 500oC
(a) Failure Mode of Shorter Column Series 2 (1800 mm)
Figure 7.18 Flexural-torsional Buckling of G250-1.95 mm Specimens
_________________________________________________________________
7-21
At 20oC At 600oC
(b) Failure Mode of Longer Column Series 1 (2800 mm)
Figure 7.18 Flexural-torsional Buckling of G250-1.95 mm Specimens
_________________________________________________________________
7-22
At 20oC At 600oC
(a) Failure Mode of Shorter Column Series 2 (1800 mm)
At 20oC At 600oC
Figure 7.19: Flexural-torsional Buckling of G450-1.90 mm Specimens
_________________________________________________________________
7-23
At 20oC At 600oC
(a)Failure Mode of Shorter Column Series 1 (1800 mm)
At 20oC At 600oC
Figure 7.20: Flexural-torsional Buckling of G550-0.95 mm Specimens
_________________________________________________________________
7-24
Figure 7.21: Interaction of Flexural-torsional Buckling with Local Buckling
_________________________________________________________________
7-25
70
60
AxialCompression Load (kN)
50
40
30
20
10
0
-2.00 0.00 2.00 4.00 6.00 8.00 10.00 12.00 14.00 16.00 18.00
(a) Axial Load Versus Axial Shortening
70
60
50
40
30
Flange
20
Web
10
0
-60.0 -40.0 -20.0 0.0 20.0 40.0 60.0
Out-of-Plane Deflection(mm)
(b) Axial Load Versus Out-of-plane Deflection
Figure 7.22: Load-Deflection Curves for Series 1 Specimen at Ambient

Temperature (G450-1.90-2800-20)
_________________________________________________________________
7-26
12
10
0
0.00 1.00 2.00 3.00 4.00 5.00 6.00
12
10
8
Flange 4
Web
2
0
-20.0 -15.0 -10.0 -5.0 0.0 5.0
Figure 7.23: Load-Deflection Curves for Series 1 Specimen at Elevated

Temperature-700oC (G450-1.90-2800-700)
_________________________________________________________________
7-27
120
100
80
60
40
20
0
0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00
120
100
80
60
40 Web
Flange
20
0
-5.00 0.00 5.00 10.00 15.00 20.00 25.00 30.00
Out-of-Plane Deflectioon (mm)
Figure 7.24: Load-Deflection Curves for Series 2 Specimen at Ambient

Temperature (G450-1.90-1800-20)
_________________________________________________________________
7-28
14
12
10
0
0.00 2.00 4.00 6.00 8.00 10.00 12.00 14.00
14.00
12.00
10.00 Axial Compression Load (kN)
8.00
6.00
web
4.00
Flange
2.00
0.00
-25.00 -20.00 -15.00 -10.00 -5.00 0.00 5.00
Figure 7.25: Load-Deflection Curves for Series 2 Specimen at Elevated

Temperature-700oC (G450-1.90-1800-700)
_________________________________________________________________
7-29
Table 7.7: Ultimate Compression Loads Obtained for the Shorter Column
Series
Ultimate Compression
Specimen
Load (kN)
G550 – 0.95 – 1800 – 20 24.72
G550 – 0.95 – 1800 – 200 22.91
G550 – 0.95 – 1800 – 300 21.33
G550 – 0.95 – 1800 – 400 20.41
G550 – 0.95 – 1800 – 600 6.49
G550 – 0.95 – 1800 – 700 3.50
G250 – 1.95 – 1800 – 20 87.94
G250 – 1.95 – 1800 – 200 83.76
G250 – 1.95 – 1800 – 300 56.07
G250 – 1.95 – 1800 – 400 45.37
G250 – 1.95 – 1800 – 500 32.20
G250 – 1.95 – 1800 – 600 23.03
G250 – 1.95 – 1800 – 700 12.60
G450 – 1.90 – 1800 – 20 120.42
G450 – 1.90 – 1800 – 200 105.99
G450 – 1.90 – 1800 – 300 83.20
G450 – 1.90 – 1800 – 400 73.43
G450 – 1.90 – 1800 – 500 46.35
G450 – 1.90 – 1800 – 600 17.86
G450 – 1.90 – 1800 – 700 11.45
_________________________________________________________________
7-30
Table 7.8: Ultimate Compression Loads Obtained for the Longer Column
Series
Specimen Ultimate Compression

Load (kN)
G550 – 0.95 – 2800 – 20 15.85
G550 – 0.95 – 2800 – 200 12.86
G550 – 0.95 – 2800 – 400 9.09
G550 – 0.95 – 2800 – 500 6.34
G550 – 0.95 – 2800 – 600 4.02
G550 – 0.95 – 2800 – 700 2.44
G250 – 1.95 – 2800 – 20 54.07
G250 – 1.95 – 2800 – 300 48.81
G250 – 1.95 – 2800 – 300 40.87
G250 – 1.95 – 2800 – 400 34.70
G250 – 1.95 – 2800 – 500 29.93
G250 – 1.95 – 2800 – 600 19.35/17.43R*
G250 – 1.95 – 2800 – 700 12.83/8.20R*
G450 –1.90 – 2800 – 20 61.30
G450 – 1.90 – 2800 – 200 50.64
G450 – 1.90 – 2800 – 400 43.11
G450 – 1.90 – 2800 – 500 29.58
G450 – 1.90 – 2800 – 600 17.16
G450 – 1.90 – 2800 – 700 10.27/9.67R*
Note:- R* Indicates repeated test
7.6 Comparison of Test Results with Current Design Standards
Eurocode 3: Part 1.2 (ECS, 2005) provides fire design guidelines for cold-formed
steel compression members. Eurocode 3: Part 1.2 recommendations were originally
developed for Class 1, 2 and 3 sections. However, it recommends the same
guidelines for Class 4 sections, which represent cold-formed steel sections. Most
importantly, it recommends the use of effective area calculated using ambient
temperature mechanical properties, for elevated temperature conditions. In addition
to this, cold-formed steel design codes, AS/NZS 4600 (SA, 2005), Eurocode 3: Part
1.3 (ECS, 1996), the North American Specification (NAS, 2003), BS 5950 Part 5
(BSI, 1998) and the Direct Strength Method provide design guidelines for the design
of compression members at ambient temperature. At present, designers simply use
_________________________________________________________________
7-31
the same guidelines for fire design with appropriately reduced mechanical properties.
Therefore ambient temperature design rules are also considered in this research.
However, AS/NZS 4600 (SA, 2005) and the North American Specification (NAS,
2003) provide identical guidelines. Therefore AS/NZS 4600 (SA, 2005), Eurocode 3:
Part 1-3 (ECS, 1996) and BS 5950 Part 5 (BSI, 1998) were selected from the
ambient temperature design codes. Except BS 5950 Part 5, all other design standards
give the same equation for the calculation of elastic torsional buckling load. BS 5950
Part 5 includes a factor of two in the equation for elastic torsional buckling that gives
higher values. Therefore in this study, BS5950 Part 5 predictions are based on the
modified equations that are similar to the other design standards.
Recently Schafer (2001) proposed a new method called the Direct Strength Method
(DSM), which can be used to predict the member capacity even with the interaction
of two or more buckling modes. The DSM formulae are similar to that of AS/NZS
4600 for flexural and flexural-torsional buckling. However, the direct strength
method uses the effective area based on yield stress while AS/NZS 4600 uses it
based on the stress corresponding to flexural-torsional buckling (fn). The difference
in the ultimate load from both methods is insignificant. However, DSM was also
considered in this study. Essential mechanical properties at elevated temperatures
were obtained from research carried out previously at QUT. Reduced mechanical
properties for 0.95 mm thick cold-formed steel were taken from Ranawaka and
Mahendran (2009) while those for 1.9 mm and 1.95 mm thick cold-formed steels
were taken from Dolamune Kankanamge (2009). Sample calculations for each
design standard are given in Appendix E.
All the test results are compared with the code predictions in Tables 7.9 to 7.14.
They show the comparison of test results with the code predictions at different
temperatures. In most cases code predictions were conservative. However, test to
prediction ratios were significantly higher except for the predictions of BS 5950 Part
5. This higher conservatism may be due to the following reasons. Beyond 400oC, it
was difficult to achieve a uniform temperature along the full length of test columns.
The temperature at the ends was less than the temperature at the middle. This non-
uniform temperature distribution introduced additional restraints to buckling because
of higher mechanical properties at the ends. Therefore test results beyond 400oC
_________________________________________________________________
7-32
were slightly higher than the predicted values. The other important factor was the
smaller imperfections of the tested specimens. Design standards assume a nominal
imperfection of Length L /1000. However, the measurement of imperfections
showed that the measured geometric imperfections are significantly lower than the
design standard value of L/1000. Therefore, test ultimate loads were higher than
code predictions. In addition, the associated capacity reduction factors ( ) were
calculated for each case. The capacity reduction factors are also greater than 0.85
and in most cases it is greater than 1.0. Therefore the code predictions are
conservative.
Figure 7.26 shows the comparison of test results with the AS/NZS 4600 design
curve. AS/NZS 4600 and DSM predictions show conservative predictions for both
long and short columns. Their predictions for high strength steels are more
conservative than those for low strength steel columns. The elastic buckling loads of
long columns are sometime higher than their ultimate compression loads. This is
more dominant for higher strength steel columns. Since AS/NZS 4600 and DSM
limit the column capacity to 87.7% of elastic buckling load, the code predictions are
very conservative, particularly for long columns made of high strength steels.
1.2
1
AS/NZS 4600
G250-1.95-1800
0.8 G250-1.95-2800
G450-1.90-1800
G450-1.90-2800
0.6 G550-0.95-1800
Pu/Ps
G550-0.95-2800
0.4
0.2
0
0 0.5 1 1.5 2 2.5 3 3.5 4
(Py/Pcr)0.5
Figure 7.26: Comparison of Test Results with AS/NZS 4600 Design Curve
_________________________________________________________________
7-33
Among the ambient temperature design standards, modified BS 5950 Part 5
predictions provide more accurate results while Eurocode 3 Part 1.3 predicts more
conservative capacities. However, Eurocode 3 Part 1.2 predictions gave overly
conservative predictions and they are greater than those of ambient temperature
design standards. Figure 7.27 compares the column design curves of Eurocode 3 Part
1.3 and Part 1.2. Eurocode 3 Part 1.2 design curves are below the Eurocode 3 Part
1.3 curve. Therefore Eurocode 3 Part 1.2 predictions are more conservative than
Eurocode 3 Part 1.3 predictions at ambient temperature. In the case of Eurocode 3
Part 1.2, imperfection factor is a function of ambient temperature yield stress. In
most cases, its design curves are below that of Eurocode 3 Part 1.3 and hence the
capacity predictions based on Eurocode 3 Part 1.2 are too conservative. Unlike the
predictions of other design standards, Eurocode 3 Part 1.2 predictions for low
strength steel columns are of the same order for high strength steel columns because
of the conservative design curve due to lower yield stresses.
1.2
EC3 P1.3 (Buckling Curve 'b")

0.8
EC3 P1.2 G250-1.95
Pu/Py
0.6 EC3 P1.2 G450-1.90
EC3 P1.2 G550-0.95

0.4
0.2
0
0 0.5 1 1.5 2 2.5 3 3.5 4
Figure 7.27: Design Curves of Eurocode 3 Part 1.3 and Part 1.2
_________________________________________________________________
7-34
Table 7.9: Comparison of Test Results with Code Prediction for the G250-1.95-2800 Test Series
Code Predictions(kN) Test/ Code Prediction
Temp. Test
o
C (kN) AS/NZS EC3 BS 5950 EC3 AS/NZS EC3 BS 5950 EC3
DSM DSM
4600 P1.3 P5 P1.2 4600 P1.3 P5 P1.2
20 54.07 49.37 49.37 42.49 48.66 34.48 1.10 1.10 1.27 1.11 1.57
200 48.81 42.60 42.60 36.65 42.07 30.00 1.15 1.15 1.33 1.16 1.63
300 40.87 34.43 34.43 29.74 33.99 23.92 1.19 1.19 1.37 1.20 1.71
400 34.70 26.85 26.85 23.41 26.61 18.62 1.29 1.29 1.48 1.30 1.86
500 29.93 19.94 19.94 17.51 19.78 13.81 1.50 1.50 1.71 1.51 2.17
600 17.43 13.52 13.52 11.95 13.44 9.38 1.29 1.29 1.46 1.30 1.86
700 8.20 7.65 7.65 6.75 7.60 5.31 1.07 1.07 1.21 1.08 1.55
Mean 1.226 1.226 1.406 1.238 1.518
COV 0.121 0.121 0.117 0.148 0.170
1.003 1.003 1.159 0.961 0.112
_________________________________________________________________________________________________________________
Behaviour and Design of Cold-formed Steel Compression Members at Elevated temperatures
7-35
Temp. Test
o
DSM DSM
4600 P1.3 P5 P1.2 4600 P1.3 P5 P1.2
20 87.94 76.17 76.17 70.35 81.54 54.95 1.15 1.15 1.25 1.08 1.60
200 83.76 67.85 67.85 62.02 72.80 48.64 1.23 1.23 1.35 1.15 1.72
300 56.07 51.30 51.30 47.95 54.73 37.33 1.09 1.09 1.17 1.02 1.50
400 45.37 38.34 38.34 36.39 40.71 28.26 1.18 1.18 1.25 1.11 1.61
500 32.20 27.70 27.70 26.45 29.16 20.54 1.16 1.16 1.22 1.10 1.57
600 23.03 18.44 18.44 17.66 19.31 13.75 1.25 1.25 1.30 1.19 1.67
700 12.60 10.44 10.44 10.00 10.94 7.78 1.21 1.21 1.26 1.15 1.62
Mean 1.183 1.183 1.257 1.117 1.613
COV 0.045 0.045 0.046 0.049 0.071
1.072 1.072 1.138 1.009 1.467
_________________________________________________________________________________________________________________
7-36
Temp. Test
o
DSM DSM
4600 P1.3 P5 P1.2 4600 P1.3 P5 P1.2
20 61.30 53.50 53.50 48.94 55.65 45.16 1.15 1.15 1.25 1.10 1.36
200 50.64 45.47 45.47 42.28 47.94 39.51 1.11 1.11 1.20 1.06 1.28
400 43.47 31.03 31.03 28.93 32.78 27.08 1.40 1.40 1.50 1.33 1.60
500 29.58 23.81 23.81 21.38 24.41 19.54 1.24 1.24 1.38 1.21 1.51
600 17.16 13.23 13.23 11.86 13.15 9.96 1.30 1.30 1.45 1.30 1.72
700 9.67 7.90 7.90 6.99 7.85 5.93 1.22 1.22 1.38 1.23 1.63
Mean 1.237 1.237 1.361 1.205 1.518
COV 0.084 0.084 0.085 0.090 0.112
1.066 1.066 1.171 1.028 1.273
_________________________________________________________________________________________________________________
7-37

Temp. Test
o
DSM DSM
4600 P1.3 P5 P1.2 4600 P1.3 P5 P1.2
20 120.42 110.62 107.43 94.28 115.30 82.89 1.09 1.12 1.28 1.04 1.45
200 105.99 100.12 96.58 83.59 102.84 75.04 1.06 1.10 1.27 1.03 1.41
300 83.20 87.85 84.35 72.52 89.26 66.21 0.95 0.99 1.15 0.93 1.26
400 73.43 68.95 66.45 57.42 70.65 51.71 1.06 1.11 1.28 1.04 1.42
500 46.35 46.17 45.14 39.97 48.68 34.73 1.00 1.03 1.16 0.95 1.33
600 17.86 17.38 17.38 16.76 17.95 13.69 1.03 1.03 1.07 1.00 1.30
700 11.45 10.72 10.72 10.28 11.18 8.40 1.07 1.07 1.11 1.02 1.36
Mean 1.037 1.062 1.187 1.003 1.363
COV 0.047 0.047 0.074 0.044 0.051
0.938 0.961 1.044 0.909 1.233
_________________________________________________________________________________________________________________
7-38
Temp. Test
o
DSM DSM
4600 P1.3 P5 P1.2 4600 P1.3 P5 P1.2
20 15.85 8.98 8.98 8.83 9.81 8.67 1.77 1.77 1.80 1.62 1.83
200 12.86 7.81 7.81 7.72 8.57 7.63 1.65 1.65 1.67 1.50 1.69
400 9.09 5.47 5.47 5.43 6.02 5.38 1.66 1.66 1.67 1.51 1.69
500 6.34 4.30 4.30 4.22 4.69 4.12 1.47 1.47 1.50 1.35 1.54
600 4.02 3.12 3.12 2.69 2.95 2.29 1.29 1.29 1.50 1.36 1.75
700 2.44 1.97 1.97 1.77 1.96 1.57 1.24 1.24 1.38 1.24 1.55
Mean 1.513 1.512 1.586 1.431 1.674
COV 0.143 0.142 0.096 0.094 0.068
1.164 1.166 1.340 1.213 1.489
_________________________________________________________________________________________________________________
7-39
Test
(kN) AS/NZS EC3 BS 5950 EC3 AS/NZS EC3 BS 5950 EC3
DSM DSM
4600 P1.3 P5 P1.2 4600 P1.3 P5 P1.2
20 24.72 21.16 20.11 18.64 21.91 17.76 1.17 1.23 1.33 1.13 1.39
200 22.91 18.41 17.49 16.42 19.26 15.83 1.24 1.31 1.39 1.19 1.45
300 21.33 15.65 14.87 14.10 16.50 13.71 1.36 1.43 1.51 1.29 1.56
400 20.41 12.90 12.26 11.59 13.57 11.24 1.58 1.66 1.76 1.50 1.82
600 6.49 5.17 5.12 4.62 5.11 3.45 1.26 1.27 1.40 1.27 1.88
700 3.50 3.80 3.70 3.27 3.73 2.58 0.92 0.95 1.07 0.94 1.36
Mean 1.256 1.309 1.412 1.221 1.572
COV 0.174 0.181 0.160 0.154 0.142
0.898 0.920 1.045 0.916 1.255
_________________________________________________________________________________________________________________
7-40
7.7 Conclusion
Flexural-torsional buckling behaviour and strength of cold-formed steel compression

members were investigated using full scale tests at ambient and elevated
temperatures. Three cold-formed steel thicknesses, G550-0.95 mm, G450-1.9 mm and
G250-1.95 mm, were selected to represent light gauge cold-formed steels. The most
commonly used section type of lipped channel section was selected with two lengths,
2800 mm and 1800 mm, so that flexural-torsional buckling was dominant during the
tests.
An electrical furnace was designed and built for the elevated temperature testing of
columns with various heights. This furnace consists of three furnace segments, each
can be controlled independently. Since a standard test rig is not available at the QUT
structural laboratory to accommodate long columns, a reaction frame and a loading
set-up was developed. This test set-up can be used within the furnace with variable
heights.
A total of 39 tests was carried out at ambient and elevated temperatures in the range
of 20 to 700oC to investigate flexural-torsional buckling behaviour of cold-formed
steel compression members and to validate finite element models. Experimental
results were compared with the predictions from different cold-formed steel design
standards. The comparative study undertaken in this research showed that ambient
temperature design standard guidelines can be safely used to predict the flexural-
torsional buckling capacities at elevated temperatures when appropriately reduced
mechanical properties are used. However, they appear to be too conservative. This is
due to the small member imperfections in the test column specimens than the design
standard value of L/1000, and the observed non-uniform temperature distribution at
the ends of test specimens at higher temperatures beyond 400oC.
Eurocode 3 Part 1.2 (ECS, 2005) gave overly conservative predictions than the
ambient temperature design standards. Its predictions were too conservative compare
to the Eurocode 3 Part 1.3, which provides similar design methods. Since the
imperfection factor used in this design code is a function of ambient temperature yield
stress, its design curves locate slightly lower that of Eurocode 3 Part 1.3, and hence
_________________________________________________________________
7-41
the capacity predictions are more conservative. Eurocode 3 Part 1.2 is therefore
considered unsuitable for the design of cold-formed steel compression members
subject to flexural-torsional buckling.
The predictions from the direct strength method, AS/NZS 4600 and Eurocode 3 Part
1.3 were found to be conservative. The predictions from the first two are closer to test
results than Eurocode 3 Part 1.3 predictions. Eurocode 3 Part 1.3 recommends the use
of buckling curve “b” for channel sections. Since its predictions are too conservative,
it appears that the buckling curve ‘b” is not suitable for channel sections. The
buckling curves “a” or “ao”, which are located above the buckling curve “b” may
predict more accurate capacities than buckling curve “b”. BS 5950 Part 5 predictions
were more close to the test results. Considering these facts, it can be concluded that
BS 5950 Part 5, AS/NZS 4600 and DSM based design rules are more suitable to
predict the flexural-torsional buckling capacities of cold-formed steel columns at
ambient and elevated temperatures. However, Eurocode 3 Part 1.3 can also be used
with improved accuracy if buckling curve “ao” is used.
The comparative study showed that most test results are higher than the predictions,
particularly at elevated temperatures. This may be due to the observed low
imperfections and non-uniform temperature distribution along the length of test
specimens. More accurate conclusive remarks and possible modifications can be
drawn from a parametric study based on finite element analyses since ultimate loads
are obtained under a more controlled environment than the tests. Such a detailed
parametric study is presented in the next chapter.
_________________________________________________________________
7-42
8 Development and Validation of Finite
Element Models of Cold-formed Steel
Columns Subject to Flexural-Torsional
Buckling
Finite element modelling is an important tool in structural engineering to reduce the

time and physical resources required for actual tests. Correct boundary conditions
and accurate mechanical properties are the important parameters in the finite element
analyses. In this research, ABAQUS standard Version 6.7 (HKS, 2002) was used in
the finite element analyses of cold-formed steel columns subject to flexural-torsional
buckling. Input files of finite element analyses are not user-friendly and are also
difficult to follow. To avoid these difficulties, pre-processors and post-processors are
available. In this research, MSC/PATRAN software was used to create the input file
as a pre-processor and the results were observed by using it as a post-processor.
Elastic buckling analyses (bifurcation analyses) were first carried out to determine
the critical buckling load and associated buckling modes. Following this, nonlinear
analyses were carried out to obtain the ultimate load. Appropriate geometric
imperfections were included to the most probable buckling modes to initiate the
appropriate buckling in the nonlinear analyses. In most cases, the buckling mode
corresponding to the lowest Eigen value gave the critical buckling modes. This was
flexural-torsional buckling as observed in the experimental study (Chapter 7).
Residual stress was also included in the nonlinear analyses. Convergence studies
were carried out to find the suitable element size that simulates a more closer
behaviour to the actual tests.
Mechanical properties are also important in the finite element analyses of

compression members at elevated temperatures. Therefore measured mechanical
properties were obtained for 1.9 mm and 1.95 mm thick cold-formed steels from
Dolamune Kankanamge (2009) while the mechanical properties of 0.95 mm thick
cold-formed steel were taken from Ranawaka and Mahendran, (2009) for
temperatures in the range of 20 to 800oC (see Chapter 3). The results of finite
element analyses were compared with test results reported in Chapter 7 for long
cold-formed steel columns. This chapter describes the development and validation of
_______________________________________________________________
finite element models to simulate the flexural-torsional buckling behaviour of cold-
formed steel compression members.
8.1 Development of Experimental Finite Element Models
Experimental finite element models were developed to simulate the tested cold-
formed steel lipped channel column specimens. They were G550-0.95-1800, G550-
0.95-2800, G450-1.90-1800, G450-1.90-2800, G250-1.95-1800 and G250-1.95-2800
series at temperatures in the range of 20 to 700oC. The measured dimensions of
lipped channel specimens including their thickness were used in the analyses (see
Chapter 7). Suitable element type and size were selected considering the accuracy
and economy of simulations. Appropriate boundary conditions were applied to
simulate the observed flexural-torsional buckling behaviour in tests. Imperfections
and residual stresses were also included in the analyses. Elevated temperature tests
were simulated using appropriately reduced mechanical properties.
8.1.1 Selection of Element Types
Thin cold-formed steel plates can be represented with shell elements such as S4,
S4R, S4R5, S8R5 and S9R5. Element type S4 is a fully integrated, general-purpose
shell element available in ABAQUS standard. This element is recommended for
accurate simulations. However, S4 element needs more processing time and disk
space compared to S4R5 element. Element types S4R5, S8R5 and S9R5 are thin
shell elements with 4, 8 and 9 nodes, respectively. They have reduced integration
with five degrees of freedom. These element types were considered in the simulation
of G450–1.90–2800-20 specimen in order to select the most suitable element type.
However, comparative analyses were limited to eight node elements since higher
order element (S9R5) need more processing time and memory requirements.
Nonlinear analyses, followed by bifurcation buckling analyses, were carried out with
each element type while considering imperfections and residual stresses. Ultimate
compression loads obtained from finite element analyses were compared with test
results in Table 8.1. This comparison showed that all the element types give similar
ultimate compression loads. Therefore S4 element type was selected for flexural-
_______________________________________________________________
torsional buckling simulations because it is a fully integrated, general-purpose shell
element available in ABAQUS standard version.
Table 8.1: Comparison of Test and FEA Ultimate Loads of G450-1.90-2800-20

for Different Element Types
Type S4R5 S4R S4 S8R5
TEST/FEA 0.959 0.955 0.949 0.933
8.1.2 Selection of Element Size
Element size of the model is also an important factor in the finite element analyses.
Finer mesh gives more accurate results. However, it needs more memory and
processing time. Hence a finer mesh does not give the most economical simulation.
Therefore a convergence analysis was carried out for G450-1.90-2800-20 specimen
with various mesh sizes from 15 mm x 15 mm to 3 mm x 3 mm to find the optimum
size of the finite element mesh. Since this convergence study was carried out to find
a suitable mesh size, residual stress was not included to simplify the analyses.
Therefore these results were not compared with the test results. As shown in Figure
8.1, 5 mm x 5 mm mesh size gives optimum results. Processing time and memory
requirements were also acceptable with this mesh size.
_______________________________________________________________
65.0
Ultimate Compression Load (kN)
64.5
64.0
63.5
63.0 Optimum Mesh Size
62.5
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
Mesh Size (mm x mm)
Figure 8.1: Ultimate Load of G450-1.90-2800-20 with Different Mesh Sizes
8.1.3 Loading and Boundary Conditions
Accurate boundary condition is the key factor for accurate simulations. The
boundary conditions should reflect the end supports used in the tests. Both ends of
the test columns were fixed against rotations and translations except that the bottom
end was allowed to move axially. These end conditions allowed the specimens to fail
symmetrically about the plane perpendicular to the axis of the column at mid height.
Due to the symmetry conditions of the test specimen and loading (see Figure 8.2), it
is economical to simulate one half of the test columns in the analyses. Therefore only
one half of the column length was simulated with appropriate boundary conditions as
shown in Figure 8.2.
_______________________________________________________________
Symmetric plane boundary condition
to allow out-of-plane deflection and
twist rotation at mid height
Member length L
L/2
Fixed end supports
Compression load
Figure 8.2: Idealisation of Half Column Model and Boundary Conditions
The axial load was applied to the test columns by a hydraulic jack via the bottom
loading shaft. Therefore axial translation was allowed at the bottom end while
keeping the top end fixed. To simulate this, only axial translation at the bottom end
was allowed in the finite element model. These boundary conditions were applied to
the independent node of the rigid fixed MPC (Multi Point Constraint) located at the
bottom end of the model. Rigid fixed MPC consists of an independent node and any
number of dependent nodes. Dependent nodes are connected to the independent node
using rigid beams with all six structural degrees of freedoms rigidly attached to each
other. In this model, independent node was located at the geometric centre of the
cross-section. All perimeter nodes at the bottom surface of the column were
considered as the dependent nodes. Since the element size was selected as 5 mm x 5
mm, the distance between the dependent nodes is small. Hence this MPC acts as a
rigid surface that is rigidly connected to the bottom end of the columns. In the tests,
_______________________________________________________________
end plates were rigidly attached to the columns by grouting. Therefore a rigid fixed
MPC accurately simulates the end support used for the tests.
Load and
12456
boundary
conditions
applied
345 boundary conditions through MPC
at the middle of column at the end
Typical finite element

mesh used in column
Figure 8.3: Loading and Boundary Conditions Applied to the Half length
Column Model
At the middle of the model, out of plane displacements were allowed while
restricting the axial displacement. In addition, axial rotation was allowed at the
middle of the model.
_______________________________________________________________
Imperfections of light gauge cold-formed steel members have a considerable

influence on their ultimate loads. Nonlinear analyses need imperfections to initiate
the appropriate buckling deformations. For the nonlinear analyses, imperfection is
introduced to the appropriate buckling mode obtained form bifurcation buckling
analyses. Appropriate eigen mode can be selected by comparing the failure mode
from experimental tests. In most cases, the first eigen mode gives the appropriate
elastic buckling mode. Larger imperfections reduce the load carrying capacity of the
members. Two kinds of imperfection, namely local imperfection and global
imperfection, can be observed with cold-formed steel members. Local imperfections
affect the local or distortional buckling capacities while global imperfections affect
the member capacity. Local imperfections are expressed in terms of section
dimensions and thickness while global imperfections are expressed in terms of
member length.
Many researchers used measured imperfections in their finite element simulations

(Ranawaka, 2006, Sivakumaran and Abdel-Rahman, 1998, Dubina and Ungureanu,
2002). Some researchers used the section thickness as the value of local imperfection
for the lipped channels made of thin cold-formed steels (Feng et al., 2003, Chen and
Young, 2006a). Therefore global imperfections of all the specimens were measured
in this research. Since the expected buckling mode was flexural-torsional buckling,
only global imperfections were measured as reported in Chapter 7. The maximum
global imperfections (expressed as a ratio of column length in Tables 7.5 and 7.6)
were used in the nonlinear analysers based on the critical flexural-torsional buckling
mode.
Residual stress model is the same as for the local buckling simulations reported in
Section 5.1.5 in Chapter 5.
_______________________________________________________________
8.1.6 Mechanical Property Model
In this research on flexural-torsional buckling failures the elastic-perfect plastic

model can also be used in the simulation. However, higher accuracy could be
obtained with strain hardening mechanical property models because the measured
stress-strain curves showed significant nonlinearities at elevated temperatures.
Therefore the strain hardening model was selected in the simulations. The measured
mechanical properties and the use of Ramberg-Osgood model are explained in
Section 5.1.6 of Chapter 5. The stress-strain curves for steel at various temperatures
in this research were based on the Ramberg-Osgood model with appropriate
coefficients given in Section 5.6.1.
8.1.7 Analysis Methods
As an initial approach, finite strip analyses using CUFSM were carried out to
determine the elastic buckling loads and buckling half wave lengths of all trial
sections. The selected specimens were then analysed using ABAQUS to determine
the elastic buckling and ultimate loads. MSC/PATRAN was used as a pre-processor
to create the input file and as a post-processor to read the results. Two types of
analyses, the bifurcation buckling analysis to determine the elastic buckling loads
and modes and the nonlinear analysis to determine the ultimate loads and
deformations, were carried out using ABAQUS. Nonlinear analyses were carried out
using the modified Riks method to find the ultimate compression load. In the
nonlinear analyses, the maximum load increment was controlled appropriately to
obtain smooth loading and accurate load factors. Measured imperfections were
included into the appropriate eigen buckling mode from the bifurcation buckling
analyses. To improve the accuracy of the analyses, residual stress was also included
in the nonlinear analyses.
8.2 Validation of Finite Element Models
It is important to validate the developed finite element model to determine whether it

can simulate the desired buckling and ultimate strength behaviour of cold-formed
_______________________________________________________________
steel columns. Usually test results are used for validation purposes. In this research,
the ultimate compression load, load-deflection curves and deflected shape from finite
element analyses were compared with corresponding test results reported in Chapter
7. Elastic buckling load could also be used to validate the finite element model.
However, it could not be estimated accurately from our tests because the failure of
specimens could not be observed during elevated temperature tests conducted inside
the furnace.
8.2.1 Validation of FEA Model using Load-Deflection Curves
Three deflection measurements were recorded in the tests. They are the out-of-plane
deflections of web and flange elements and the axial shortening of test column
specimens. These deflections could also be obtained from finite element simulations.
Accurate modelling should produce identical load-deflection curves from test and
finite element analyses. Figures 8.4 to 8.12 compare the load-deflection curves from
finite element analyses and tests for some selected cases. Since the load was applied
by manual operation of a hydraulic pump, the load-deflection curves from the tests
are not smooth as in finite element analyses. Axial shortening of the tests was
measured at the bottom end of the bottom loading shaft (see Figure 7.8). Therefore
axial shortening in test curves includes the shortening of loading shafts. In addition,
it also includes the compaction of insulation packing kept between the end plates and
plates on loading shafts. Therefore axial shortening in test curves is slightly higher
than the finite element analysis results. Further, out-of-plane deflections were
measured using 800 mm LVDTs. It can involve a maximum error of 2% (equivalent
to 1.6 mm) in the out-of-plane deflections. Having considered these facts, it can be
concluded that the load-deflection curves showed a reasonably good agreement
between the finite element analyses and tests. Therefore it can be concluded that the
developed finite element model can simulate the flexural torsional buckling
behaviour of lipped channel steel columns.
_______________________________________________________________
70
60
50
TEST
40 FEA
30
20
10
0
0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00
(a) Axial Load versus Axial Shortening
70
60
50
40
30 TEST-Flange
TEST-Web
20
FEA-Web
FEA-Flange
10
0
-60.0 -40.0 -20.0 0.0 20.0 40.0 60.0
(b) Axial Load versus Out-of-Plane Deflection
Figure 8.4: Comparison of Load-Deflection Curves from Tests and Finite

Element Analyses for G450-1.90-2800-20 Specimens
_______________________________________________________________
12
10
4
TEST
FEA
2
0
0.00 1.00 2.00 3.00 4.00 5.00 6.00
12
10
8
TEST-Flange 4
TEST-Web
FEA-Web 2
FEA-Flange
0
-20.0 -15.0 -10.0 -5.0 0.0 5.0

_______________________________________________________________
60
50
40
30
20
TEST
10 FEA
0
-1.00 0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00
60

50
40
30
TEST-Flange TEST-Web 20
FEA-Web FEA-Flange
10
0
-15.0 -13.0 -11.0 -9.0 -7.0 -5.0 -3.0 -1.0 1.0

_______________________________________________________________
45
40

35
30
25
20
15
TEST
10
FEA
5
0
-1.00 0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00
45
40
35

30
25
TEST-Flange 20
TEST-Web 15
FEA-Web
10
FEA-Flange
5
0
-20.0 -15.0 -10.0 -5.0 0.0 5.0

_______________________________________________________________
20
18
16
14
12
TEST
10 FEA
0
-1.00 0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00
20
18
16
14
12
10
8
TEST-Flange
6 TEST-Web
4 FEA-Web
FEA-Flange
2
0
-20.0 -15.0 -10.0 -5.0 0.0 5.0 10.0 15.0 20.0

_______________________________________________________________
10
8
7
4
TEST
3
2 FEA
0
-2.00 0.00 2.00 4.00 6.00 8.00 10.00 12.00 14.00
10
7
5
TEST-Flange
4
TEST-Web
3
FEA-Web
2 FEA-Flange
0
-60.0 -40.0 -20.0 0.0 20.0 40.0 60.0

_______________________________________________________________
140
Axial Compression Load (kN) 120
100
80
60
40 TEST
FEA
20
0
-2.00 0.00 2.00 4.00 6.00 8.00 10.00
140
120
100
80
60 TEST-Flange
TEST-Web
40
FEA-Web
20 FEA-Flange
0
-5.0 0.0 5.0 10.0 15.0 20.0

_______________________________________________________________
14
12
4
TEST
2 FEA
0
-1.00 0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00
14
12
10 Axial Compression Load (kN)
6
TEST-Flange
TEST-Web 4
FEA-Web
FEA-Flange 2
0
-20.0 -15.0 -10.0 -5.0 0.0 5.0

_______________________________________________________________
100
90
80
70
60
50
40
30
TEST
20
FEA
10
0
-1.00 0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00
100
90
80
70
60
50
TEST-Flange
40
TEST-Web
30 FEA-Web
20 FEA-Flange
10
0
-5.0 0.0 5.0 10.0 15.0 20.0

_______________________________________________________________
8.2.2 Validation of FEA Model Using Deflected Shape
Since the elevated temperature tests were carried out inside the furnace, visual
observation during the test could be only seen for the ambient temperature tests.
Video record of test and the failure animation from the finite element analyses of
ambient temperature test showed a very good agreement. Due to limited
instrumentation, video records of tests are only available for ambient temperature
tests. In this section, only the final deflected shapes of tested specimen and that from
the analyses are compared. This method allows comparing both ambient and
elevated temperatures tests. As shown in Figure 8.13 failure modes from FEA model
and test are similar. Figures 7.13 to 7.14 show the flexural-torsional buckling
observed for each thickness and length at ambient and elevated temperatures. They
all failed in flexural-torsional buckling.
(a) G450-1.90-1800-20
Figure 8.13: Comparison of Failure Modes for Validation Purpose

_______________________________________________________________
(b) 450-1.90-1800-600
_______________________________________________________________
(c) G450-1.90-2800-700

_______________________________________________________________
(d) G450-1.90-2800-700

_______________________________________________________________
8.2.3 Validation of FEA model Using Ultimate Loads
The developed finite element model was finally validated by comparing the ultimate
loads obtained from the tests and analyses. Elastic buckling and ultimate loads were
recorded for each simulation. Elastic buckling load was not measured during the test
since visual observations could not be made in the elevated temperature tests. But it
was possible to observe the buckling behaviour in the ambient temperature tests.
However, elastic buckling load from visual observations is not accurate as it depends
on the observer’s judgement. Therefore only the ultimate loads were compared. To
ensure that developed model works for different strength grades, thickness and
lengths, comparison was made separately based on thickness, strength grade and
lengths (see Tables 8.2 to 8.7). Mean values of Test/FEA ultimate load ratio for each
case are close to one while associated coefficients of variation are also small in most
cases. Therefore the developed finite element model can be used to simulate the
buckling behaviour of cold-formed steel long columns subjected to flexural-torsional
buckling effects.
Table: 8.2: Comparison of Ultimate Loads for G450-1.90-2800 Series
Temperature Ultimate Load (kN)

o TEST/FEA
C Test FEA
20 61.30 63.80 0.961
200 50.64 56.40 0.898
400 43.47 38.20 1.138
500 29.58 27.20 1.088
600 17.16 15.50 1.107
700 9.67 9.48 1.020
Mean 1.035
COV 0.090
_______________________________________________________________
Table 8.3: Comparison of Ultimate Loads for G250-1.95-2800 Series

o TEST/FEA
C Test FEA
20 54.07 56.3 0.960
200 48.81 48.5 1.006
300 40.87 40.4 1.012
400 34.70 26.8 1.295
500 29.93 21.4 1.398
600 17.43 15.9 1.096
700 8.20 9.28 0.884
Mean 1.093
COV 0.171

o TEST/FEA
C Test FEA
20 15.85 15.40 1.029
200 12.86 13.70 0.939
400 9.09 9.57 0.950
500 6.34 7.14 0.888
600 4.02 3.37 1.193
700 2.44 2.31 1.056
Mean 1.009
COV 0.108
_______________________________________________________________

o TEST/FEA
C Test FEA
20 120.42 129.00 0.933
200 105.99 110.00 0.964
300 83.20 96.40 0.863
400 73.43 78.00 0.941
500 46.35 53.40 0.868
600 17.86 20.50 0.871
700 11.45 11.30 1.013
Mean 0.922
COV 0.062

o TEST/FEA
C Test FEA
20 87.94 90.7 0.970
200 83.76 84.3 0.994
300 56.07 62.4 0.899
400 45.37 41.1 1.107
500 32.20 26.6 1.210
600 23.03 19 1.212
700 12.60 10.9 1.156
Mean 1.078
COV 0.116
_______________________________________________________________

o TEST/FEA
C Test FEA
20 24.72 25.40 0.973
200 22.91 22.50 1.018
300 21.33 19.30 1.105
400 20.41 15.80 1.292
600 6.49 6.17 1.052
700 3.50 4.56 0.768
Mean 1.035
COV 0.166
8.2.4 Development of Ideal Finite Element Models
Ideal finite element models are used in a parametric study under standard conditions.
In this research, the experimental finite element model described in the earlier
section was used as the ideal model with changes to the geometric parameters. The
boundary conditions and loading method remain the same in the ideal model. Instead
of the measured thicknesses and imperfections, nominal thicknesses and standard
imperfection values (L/1000) were used. Reduced mechanical properties and true
stress-strain curves based on Ramberg-Osgood model at elevated temperatures were
obtained from Dolamune Kankanamge (2009) and Ranawaka and Mahendran
(2009). Other essential geometric parameters such as section size, specimen length
etc. are selected based on range of parameters investigated.
8.3 Conclusion
This chapter has described a finite element model developed to simulate the flexural-
torsional buckling behaviour of cold-formed steel lipped channel columns under
axial compression load. This experimental finite element model included the
following features: S4 type elements, measured geometric dimensions and
_______________________________________________________________
imperfections, residual stresses, reduced mechanical properties and true stress-strain
curves based on Ramberg-Osgood model based on Ranawaka and Mahendran
(2009) and Dolamune Kankanamge (2009), appropriate loading and boundary
conditions as in experiments, and a half column length model due to symmetry
conditions.
The bucking and failure modes, load-deflection curves and ultimate loads for G550-
0.95, G450-1.90 and G250-1.95 cold-formed steel columns of 1800 mm and 2800
mm length and fixed ends at temperatures in the range of 20 to 700oC agreed
reasonably well with corresponding test results reported in Chapter 7. Hence it can
be concluded that the developed finite element models can be used to predict the
flexural-torsional buckling behaviour of cold-formed steel columns. Although the
developed experimental finite element model was validated only for the flexural-
torsional buckling of cold-formed steel columns at ambient and elevated
temperatures, it can also be used to predict their flexural buckling behaviour
accurately. An ideal model finite element model needed for the parametric study is
proposed based on this model.
_______________________________________________________________
9 Parametric Study of Flexural-Torsional
Buckling Behaviour of Cold-formed Steel
Compression Members
A detailed parametric study was carried out to understand the effects of different
parameters that affect the flexural-torsional buckling behaviour of cold-formed steel
compression members at ambient and elevated temperatures. For this purpose, a
series of finite element analyses was carried out using the ideal finite element model
of a fixed ended cold-formed steel compression member described in Chapter 8. The
accuracy of the current ambient temperature design rules for flexural-torsional
buckling of cold-formed steel compression members was assessed. The acceptability
of using the same design rules for fire design by employing appropriately reduced
mechanical properties was also investigated. In addition to the ambient temperature
design standards, elevated temperature design method given in Eurocode 3 Part 1.2
(2005) was also considered in the analyses. In this research, the selected lipped
channel sections did not show flexural buckling, for which the section should have
either narrow flanges or the member should be long. In the case of experimental
study, the specimen length had to be limited to 2800 mm while concentrically
loading lipped channel section with narrow flanges was somewhat difficult. Further,
lipped channel sections with narrow flanges are commonly used as flexural members
and not as compression members. Therefore this research was limited to flexural-
torsional buckling. This chapter presents the details of parametric study and the
results.
9.1 Details of Finite Element Model used in the Parametric Study
An ideal finite element model was used in the parametric study. Ideal and
experimental models have the same loading and boundary conditions as described in
Chapter 8. The only difference is the use of nominal values for the dimensions and
imperfections instead of measured values. Details of experimental and ideal models
are given in Chapter 8. As described in Chapter 5, reduced mechanical properties of
modulus of elastic and yield stress were obtained from Ranawaka and Mahendran
_________________________________________________________________
(2009) and Dolamune Kankanamge (2009). Poisson’s ratio was assumed to be 0.3 at
both ambient and elevated temperatures as was done by other researchers (Chen and
Young, 2007, Ranby, 1999, Outinen and Myllymaki, 1995). A nominal standard
geometric imperfection of L/1000 was used in the analyses. The stress-strain
relationship was obtained based on the reduced mechanical properties and Ramberg-
Osgood model as given in Ranawaka and Mahendran (2009) and Dolamune
Kankanamge (2009). Residual stresses were obtained from the proposed model of
Schafer and Pekoz (1998). Reduced residual stresses based on Lee (2004) were used
for the elevated temperature simulations. Elastic buckling analyses were carried out
to find the elastic buckling load and corresponding buckling modes. Appropriate
geometric imperfections were then included in the corresponding buckling mode and
used in the nonlinear analyses to determine the ultimate failure loads of chosen fixed
ended cold-formed steel compression members of varying geometry and length.
9.2 Flexural-torsional Buckling Behaviour at Ambient

Temperature
The major objective of this chapter is to assess the suitability of ambient temperature
design rules for elevated temperature design. Therefore it is essential to verify the
accuracy of current ambient temperature design rules. For this purpose, a series of
ambient temperature analyses was carried out for all the selected thicknesses and
steel grades. The ultimate load results were then compared with ambient temperature
design methods of AS/NZS 4600, Eurocode 3 Part 1.3, the Direct Strength Method
(DSM) and BS 5950 Part 5. In addition, Eurocode 3 Part 1.2 design method, which is
recommended for the fire (elevated temperature) design, was also included in the
comparative study. The details and design formulae of these selected design
standards are given in Chapter 2.
9.2.1 Selection of Member Sizes for the Parametric Study
All the tested steel thicknesses and strength grades (G450-1.90 mm, G250-1.95 mm
and G550-0.95 mm) were selected for the parametric study because they are
commonly used for the cold-formed structural steel columns. Three cross-sections
_________________________________________________________________
were selected for each thickness to obtain a wide range of results. For each cross-
section a range of member lengths (1500 to 4000 mm) were selected to expand the
available results. Following nomenclature was used the members used in the
parametric study.
G250-1.95-75-50-15-2500
Length
Lip
Flange
Web
Thickness
Grade
9.2.2 Comparison of Ultimate Loads with Predictions from the

Current Design Rules at Ambient Temperature
A series of finite element analyses (FEA) was carried out at ambient temperature to
assess the accuracy of current design rules for ambient temperature design. Elastic
buckling load and buckling mode from bifurcation buckling analyses and the
ultimate axial compression load from nonlinear analyses were obtained for all the
selected members. Most of the selected compression members failed in flexural-
torsional buckling. Some shorter columns with a slender web element showed local
buckling. Some longer columns showed convergence problems. These columns were
ignored because their member slenderness values were significantly higher leading
to very small axial strength. From these results, only 39 specimens were found to fail
in flexural-torsional buckling. These ultimate load results were then tabulated based
on the steel grade and thickness and compared with the predictions of current design
standards. Predicted ultimate loads for the selected specimens were calculated using
the ambient temperature design methods of AS/NZS 4600, DSM, Eurocode 3 Part
1.3 and BS 5950 Part 5 and the elevated temperature design method of Eurocode 3
Part 1.2. The ultimate loads from these design codes were calculated using the same
procedure shown in Appendix E. In these calculations, loading was considered to be
concentric for fixed ended columns as recommended by Young and Rasmussen
(1998). The ratio of ultimate loads from FEA to the code prediction was calculated
_________________________________________________________________
for each case and tabulated (see Tables 9.1 to 9.3). These tables also include a
capacity reduction factor ( ) in each case, calculated based on the AISI (2007)
procedure with a reliability index of ( ) of 2.5 for cold-formed steel members. The
capacity reduction factors are somewhat higher than expected in this study based on
FEA as the corresponding coefficients of variation are small compared to
experimental studies. Appendix F shows sample calculations of ( ) factors.
9.2.2.1 AS/NZS 4600 and the Direct Strength Method
Comparative results showed that the DSM and AS/NZS 4600 predictions are more
accurate than the other design standards (see Tables 9.1 to 9.3). The predictions in
the case of low strength steels (G250-1.95) agreed well with the AS/NZS 4600 and
DSM predictions. However, their predictions were conservative for higher strength
steels ofG450-1.90 and G550-0.95. Figure 9.1 shows the graphical representation of
the results and AS/NZS 4600 design curve. Appendix G shows sample calculations
used in plotting this figure.
1.2
1
AS/NZS 4600
0.8 G250-1.95-20
G450-1.90-20
Pu/Ps
0.6 G550-0.95-20
0.4
0.2
0
0 0.5 1 1.5 2 2.5 3 3.5 4
(Py/Pcr)0.5
Figure 9.1: Comparison of FEA Results with AS/NZS 4600 Prediction
_________________________________________________________________
Table 9.1: Comparison of FEA Ultimate loads with Code Predictions for G250-1.95 Series at Ambient Temperature
FEA (kN) Code Predictions (kN) FEA/Code Predictions
Member AS/NZS EC3 BS EC3 AS/NZS EC3 BS 5950 EC3
EBL Ult DSM DSM
4600 P1.3 5950 P5 P1.2 4600 P1.3 P5 P1.2
G250-1.95-60-40-15-1500 125.8 75.0 67.0 67.0 62.9 72.8 49.2 1.12 1.12 1.19 1.03 1.52
G250-1.95-60-40-15-2000 75.4 60.1 54.8 54.8 48.8 57.2 38.5 1.10 1.10 1.23 1.05 1.56
G250-1.95-60-40-15-2500 51.6 46.2 43.5 43.5 37.7 43.3 30.6 1.06 1.06 1.23 1.07 1.51
G250-1.95-60-40-15-3000 38.6 39.9 34.0 34.0 30.0 33.8 25.0 1.10 1.10 1.25 1.11 1.50
G250-1.95-60-40-15-3500 30.9 30.7 26.9 26.9 24.7 27.4 21.0 1.14 1.14 1.24 1.12 1.46
G250-1.95-75-50-15-1500 195.5 95.1 86.9 86.5 81.7 92.1 64.6 1.09 1.10 1.16 1.03 1.47
G250-1.95-75-50-15-2000 117.1 82.6 74.1 74.1 67.0 78.5 52.4 1.12 1.12 1.23 1.05 1.58
G250-1.95-75-50-15-2500 78.6 65.8 61.1 61.1 53.2 62.0 42.3 1.08 1.08 1.24 1.06 1.56
G250-1.95-75-50-15-3000 57.2 52.3 49.2 49.2 42.4 48.4 34.6 1.06 1.06 1.23 1.08 1.51
G250-1.95-75-50-15-3500 44.2 43.1 38.9 38.9 34.5 38.7 28.8 1.11 1.11 1.25 1.11 1.50
G250-1.95-90-60-15-2000 169.2 104.0 94.5 91.3 85.1 97.2 66.6 1.10 1.14 1.22 1.07 1.56
G250-1.95-90-60-15-2500 114.2 88.3 80.9 79.4 70.6 82.3 55.3 1.09 1.11 1.25 1.07 1.60
G250-1.95-90-60-15-3000 82.7 71.6 67.5 67.5 57.5 66.4 45.9 1.06 1.06 1.25 1.08 1.56
G250-1.95-90-60-15-3500 62.7 58.6 55.1 55.1 47.1 53.5 38.4 1.06 1.06 1.24 1.10 1.52
G250-1.95-90-60-15-4000 49.8 49.6 44.2 44.2 39.1 43.8 32.6 1.12 1.12 1.27 1.13 1.52
Mean 1.094 1.099 1.233 1.077 1.529
COV 0.023 0.025 0.021 0.029 0.025
Φ 1.006 1.010 1.134 0.994 1.405
EBL=Elastic Buckling Load in kN Ult= Ultimate Load in kN
_________________________________________________________________________________________________________________
9-5
Table 9.2: Comparison of FEA Ultimate with Code Predictions for G450-1.90 Series at Ambient Temperature
FEA Code Predictions FEA/Code Predictions
EBL Ult DSM DSM
4600 P1.3 5950 P5 P1.2 4600 P1.3 P5 P1.2
G450-1.90-60-40-15-1500 133.8 109.0 99.7 99.7 87.0 107.8 76.5 1.09 1.09 1.25 1.01 0.96
G450-1.90-60-40-15-2000 80.1 74.3 70.1 70.1 60.5 71.4 54.7 1.06 1.06 1.23 1.04 1.01
G450-1.90-60-40-15-2500 54.7 55.4 48.2 48.2 44.2 50.4 40.8 1.15 1.15 1.25 1.10 1.07
G450-1.90-60-40-15-3000 40.7 44.8 35.9 35.9 34.2 38.1 32.0 1.25 1.25 1.31 1.17 1.14
G450-1.90-60-40-15-3500 32.2 37.8 28.4 28.4 27.7 30.5 26.2 1.33 1.33 1.37 1.24 1.21
G450-1.90-75-50-15-1500 207.8 146.0 136.8 130.1 118.8 144.5 103.4 1.07 1.12 1.23 1.01 0.95
G450-1.90-75-50-15-2000 124.5 108.0 103.5 101.5 87.2 105.4 77.4 1.04 1.06 1.24 1.02 0.98
G450-1.90-75-50-15-2500 83.4 79.9 73.8 73.8 64.3 74.8 58.4 1.08 1.08 1.24 1.07 1.02
G450-1.90-75-50-15-3000 60.6 63.8 53.5 53.5 49.2 55.8 45.5 1.19 1.19 1.30 1.14 1.10
G450-1.90-75-50-15-3500 46.7 54.4 41.2 41.2 39.2 43.6 36.7 1.32 1.32 1.39 1.25 1.20
G450-1.90-90-60-15-2000 182.1 145.0 140.9 130.4 114.6 139.4 100.1 1.03 1.11 1.26 1.04 0.98
G450-1.90-90-60-15-2500 122.7 110.0 107.6 102.7 88.2 105.3 78.6 1.02 1.07 1.25 1.04 1.00
G450-1.90-90-60-15-3000 88.3 86.4 78.8 77.7 68.5 79.3 62.2 1.10 1.11 1.26 1.09 1.04
G450-1.90-90-60-15-3500 67.1 72.3 59.7 59.7 54.5 61.7 50.2 1.21 1.21 1.33 1.17 1.12
G450-1.90-90-60-15-4000 53.2 63.3 47.3 47.3 44.6 49.7 41.6 1.34 1.34 1.42 1.27 1.22
Mean 1.153 1.168 1.289 1.111 1.417
COV 0.099 0.086 0.048 0.082 0.032
Φ 1.001 1.028 1.171 1.023 1.298
_________________________________________________________________________________________________________________
9-6
Table 9.3: Comparison of FEA Ultimate with Code Predictions for G550-0.95 Series at Ambient Temperature
FEA Results Code Predictions FEA/Code Predictions
Member BS BS
AS EC3 EC3 AS/NZS EC3 EC3
EBL Ult DSM 5950 DSM 5950
4600 P1.3 P1.2 4600 P1.3 P1.2
P5 P5
G550-0.95-40-25-7.5-1500 16.1 17.8 14.3 14.3 13.2 15.0 12.5 1.25 1.25 1.34 1.19 1.42
G550-0.95-40-25-7.5-2000 10.0 13.2 8.8 8.8 8.6 9.5 8.3 1.50 1.50 1.53 1.39 1.59
G550-0.95-40-25-7.5-2500 7.2 10.2 6.3 6.3 6.3 6.8 6.1 1.62 1.62 1.61 1.49 1.66
G550-0.95-55-35-7.5-1500 32.6 29.2 30.2 27.5 24.6 28.9 22.5 0.97 1.06 1.19 1.01 1.30
G550-0.95-55-35-7.5-2000 19.5 22.2 17.6 17.1 16.1 18.1 15.1 1.26 1.30 1.38 1.22 1.47
G550-0.95-55-35-7.5-2500 13.2 17.9 11.8 11.8 11.3 12.5 10.8 1.52 1.52 1.58 1.44 1.66
G550-0.95-55-35-7.5-3000 9.6 14.8 8.6 8.6 8.5 9.2 8.2 1.72 1.72 1.74 1.60 1.80
G550-0.95-70-45-7.5-2500 22.8 24.5 20.9 19.3 18.6 20.9 17.4 1.17 1.27 1.31 1.17 1.40
G550-0.95-70-45-7.5-3000 16.4 20.7 14.9 14.3 14.0 15.3 13.2 1.39 1.45 1.48 1.35 1.56
Mean 1.376 1.409 1.463 1.318 1.540
COV 0.172 0.146 0.117 0.140 0.102
Φ 0.990 1.020 1.168 1.210 1.297
_________________________________________________________________________________________________________________
9-7
The ratio of finite element analysis results to code prediction increases with
increasing strength of steel. This trend dominates in the case of higher strength
steels. This means that the effect of yield strength and the member slenderness are
not well reflected in the design standards. Analyses showed that ultimate flexural-
torsional buckling capacities are greater than the elastic buckling loads for longer
columns of higher strength steels. Figures 9.2 and 9.3 compare the finite element
analysis results of a selected section of low and high strength steels with AS/NZS
4600 design curve. Sample calculations for obtaining these curves are given in
Appendix G. In the case of low strength steels, flexural-torsional buckling load is
lower than the elastic buckling load. However, it is sometimes greater than the
elastic buckling load for higher strength steel columns. A previous research by Chen
and Young (2007b) also experienced higher values of FEA to code predictions which
is the case of higher ultimate loads than the elastic buckling loads. AS/NZS 4600
and DSM limits the flexural or flexural-torsional buckling capacity to 87.7% of the
elastic buckling load. Therefore the code predictions for longer high strength steel
columns are conservative. In addition, the capacity reduction factors ( ) are always
greater than one for AS/NZS 4600 predictions. Since the current factor is 0.85, the
code predictions appear to be quite conservative.
Ultimate Load (AS/NZS 4600)

PULT /PS
Ultimat Load (FEA)
Elastic Buckling (AS/NZS 4600)

or
PEB/Py
0
0.00 50.00 100.00 150.00 200.00 250.00 300.00
Le/ry
Figure 9.2: Comparison of FEA Ultimate Load of G250-1.95-60-40-15 Series

with AS/NZS 4600 Design Curves at Ambient Temperature
_________________________________________________________________
1
PULT/PS
Ultimate Load (AS/NZS 4600)
Ultimat Load (FEA)
Elastic Buckling (AS/NZS 4600)

or
PEB/Py
0
0.00 50.00 100.00 150.00 200.00 250.00 300.00
Le/ry
Figure 9.3: Comparison of FEA Results of G450-1.90-60-40-15 Series with

AS/NZS 4600 Design Curve at Ambient Temperature
9.2.2.2 Eurocode 3 Part 1.3
Eurocode 3 Part 1.3 appears to be the most conservative standard among the ambient
temperature design standards (see Tables 9.1 to 9.3). The mean values of FEA to
code predictions are greater than 1.0 while associated capacity reduction factors ( )
are also higher then 0.85. They are much greater than those of other ambient
temperature design standards. It recommends the buckling curve “b” for cold-formed
steel lipped channel sections. However, comparison of FEA results with the different
buckling curves given in Eurocode 3 Part 1.3 showed that buckling curve “b” is
conservative. Instead it appears that the results follow buckling curve “ao” (see
Figures 9.4 a to c). Sample calculations for obtaining these curves and plotting the
data points are given in Appendix G. As observed in AS/NZS 4600 and the DSM,
the ratio of FEA to code predictions increases with increasing length. Further,
predictions were very conservative for higher strength steels. The reason for the
conservative prediction is the higher non-dimensional slenderness of the long
columns of higher strength steels.
_________________________________________________________________
1.2
1 Buckling Curve ao
Buckling Curve a
0.8
Buckling Curve b
Pu/Ps
0.6
Buckling Curve c
0.4
FEA Results
0.2
0
0 0.5 1 1.5 2 2.5 3 3.5 4
(a) G250 Steel Specimens
1.2
1 Buckling Curve ao
Buckling Curve a
0.8
Buckling Curve b
Pu/Ps
0.6
Buckling Curve c
0.4
FEA Results
0.2
0
0 0.5 1 1.5 2 2.5 3 3.5 4
(b) G450 Steel Specimens
Figure 9.4: Comparison of FEA Ultimate Loads with

Eurocode 3 Part 1.3 Design Curves
_________________________________________________________________
1.2
Buckling Curve ao
1
Buckling Curve a
0.8 Buckling Curve b
Buckling Curve c
Pu/Ps
0.6
FEA Results
0.4
0.2
0
0 0.5 1 1.5 2 2.5 3 3.5 4
(c) G550 Steel Specimens

Eurocode 3 Part 1.3 Design Curves
Eurocode 3 Part 1.2 predictions are very conservative for the ambient temperature
design (see Tables 9.1 to 9.3). This is due to the use of higher imperfection factors
for elevated temperature design standard. Eurocode 3 Part 1.3 recommends the use
of 0.34 (buckling curve “b”) as imperfection factor whereas Eurocode 3 Part 1.2
recommends the use of an equation based on yield stress. Therefore the buckling
curve depends on the yield stress of steel. The buckling curves of higher strength
steels are lower than that of lower strength steels. The imperfection factors for G450-
1.90, G250-1.95 and G550-0.95 steels at ambient temperature are 0.44, 0.61 and
0.40, respectively. Since these imperfection factors are greater than the Eurocode 3
Part 1.3 recommended value of 0.34 for channel sections, all the ambient
temperature design curves of Eurocode 3 Part 1.2 are below the Eurocode 3 Part 1.3
design curve. Since Eurocode 3 Part 1.3 provided more conservative predictions than
_________________________________________________________________
all the other ambient temperature design standards for cold-formed steel compression
members, Eurocode 3 Part 1.2 predictions are the most conservative among the
considered design standards because its design curves located below the Eurocode 3
Part 1.3 design curve.
9.2.2.4 BS 5950 Part 5
BS 5950 Part 5 design rules give higher elastic torsional buckling loads (PT). This
higher torsional buckling loads leads to unconservative capacity predictions.
Therefore a modification was made to the elastic torsional buckling equation of BS
5950 Part 5 design method as given in Equation 9.1 so that it is the same as the other
design standards (see Chapter 2). Table 9.4 compares the predictions from the
current and modified design methods. The mean value of FEA ultimate loads to code
predictions with proposed equation is close to one while associated coefficient of
variations is small. Further the mean value for the predictions of current method is
considerably low while associated coefficient of variation is higher than that of
modified method. Therefore it is appear that the modified method is more accurate
than the current method. However, the predictions based on the modified BS 5950
Part 5 are still slightly unconservative for the short columns of low strength steels,
but not for high strength steels (see Tables 9.1 to 9.3).
1 π 2 EC w
PT = 2 GJ + (9.1)
ro L2E
_________________________________________________________________
Table 9.4: Comparison of Ultimate Loads from the Current and Modified
Design Methods of BS5950 Part 5
FEA/Code Prediction
Specimen
Modified Current
G250-1.95-60-40-15-1500 1.03 0.97
G250-1.95-60-40-15-2000 1.05 0.88
G250-1.95-60-40-15-2500 1.07 0.74
G250-1.95-60-40-15-3000 1.11 0.65
G250-1.95-60-40-15-3500 1.12 0.59
G250-1.95-75-50-15-1500 1.03 1.00
G250-1.95-75-50-15-2000 1.05 0.94
G250-1.95-75-50-15-2500 1.06 0.80
G250-1.95-75-50-15-3000 1.08 0.67
G250-1.95-75-50-15-3500 1.11 0.59
G250-1.95-90-60-15-2000 1.07 1.01
G250-1.95-90-60-15-2500 1.07 0.93
G250-1.95-90-60-15-3000 1.08 0.79
G250-1.95-90-60-15-3500 1.10 0.68
G250-1.95-90-60-15-4000 1.13 0.61
G450-1.90-60-40-15-1500 1.01 0.84
G450-1.90-60-40-15-2000 1.04 0.74
G450-1.90-60-40-15-2500 1.10 0.62
G450-1.90-60-40-15-3000 1.17 0.57
G450-1.90-60-40-15-3500 1.24 0.55
G450-1.90-75-50-15-1500 1.01 0.93
G450-1.90-75-50-15-2000 1.02 0.79
G450-1.90-75-50-15-2500 1.07 0.64
G450-1.90-75-50-15-3000 1.14 0.56
G450-1.90-75-50-15-3500 1.25 0.53
G450-1.90-90-60-15-2000 1.04 0.91
G450-1.90-90-60-15-2500 1.04 0.79
G450-1.90-90-60-15-3000 1.09 0.66
G450-1.90-90-60-15-3500 1.17 0.59
G450-1.90-90-60-15-4000 1.27 0.56
G550-0.95-40-25-7.5-1500 1.19 0.77
G550-0.95-40-25-7.5-2000 1.39 0.90
G550-0.95-40-25-7.5-2500 1.49 0.79
G550-0.95-55-35-7.5-1500 1.01 0.75
G550-0.95-55-35-7.5-2000 1.22 0.79
G550-0.95-55-35-7.5-2500 1.44 0.71
G550-0.95-55-35-7.5-3000 1.60 0.65
G550-0.95-70-45-7.5-2500 1.17 0.78
G550-0.95-70-45-7.5-3000 1.35 0.85
Mean 1.146 0.747
COV 0.123 0.185
1.057 0.689
_________________________________________________________________
9.3 Flexural-torsional Buckling Behaviour at Elevated
Temperatures
In this section, the ultimate loads of cold-formed steel compression members subject
to flexural-torsional buckling at elevated temperatures were compared with the
predictions of ambient and elevated temperature design standards. Elevated
temperatures selected for this study were 100oC to 700oC at 100oC intervals. All the
members selected for ambient temperature simulations were also selected for the
elevated temperature simulations. Some of the shorter columns showed local
buckling interaction with flexural-torsional buckling. Flexural buckling was not
observed in the selected range of column lengths. Longer columns of selected
sections experienced numerical convergence problems as the required length for
flexural buckling was very high. Since local buckling behaviour was already
investigated (see Chapters 4, 5 and 6), only flexural-torsional buckling failure results
were considered in this study. Same nomenclature used for ambient temperature
study was used to name the elevated temperature specimens. The ultimate loads from
FEA were tabulated and compared with the current ambient temperature design rules
with appropriately reduced mechanical properties at elevated temperatures.
9.3.1 Comparison of Ultimate Loads with Predictions from the

Current Design Rules
Elastic buckling load, the buckling mode and the ultimate axial compression load
were obtained for all the selected members. From these FEA results, 273 specimens
failed in flexural-torsional buckling. All the ambient temperature design standards
were included in the comparative study (see Tables 9.5 to 9.25). In the case of
Eurocode 3 Part 1.2, a reduced yield stress was considered with the ambient
temperature effective area. For each temperature, the results were tabulated based on
the thickness and strength grades. In each case the mean values and associate
coefficients of variation (COV) of FEA to code predictions were estimated. Based on
the mean values, associated coefficients of variation and the number of specimens in
each case the capacity reduction factors ( ) were also calculated. This factor was
then used to assess the accuracy of the design methods. Table 9.26 shows the
summary of the comparative study.
_________________________________________________________________
Table 9.5: Comparison of FEA Ultimate Loads with Code Predictions for G250-1.95 Series at 100oC
Member AS/NZS EC3 BS 5950 EC3 AS/NZS EC3 BS 5950 EC3
EBL Ult DSM DSM
4600 P1.3 P5 P1.2 4600 P1.3 P5 P1.2
G250-1.95-60-40-15-1500 117.5 71.8 63.7 63.7 59.8 69.4 46.9 1.13 1.13 1.20 1.03 1.53
G250-1.95-60-40-15-2000 70.4 56.9 51.8 51.8 46.0 54.0 36.7 1.10 1.10 1.24 1.05 1.55
G250-1.95-60-40-15-2500 48.2 43.5 40.9 40.9 35.4 40.7 29.1 1.06 1.06 1.23 1.07 1.49
G250-1.95-60-40-15-3000 36.0 34.7 31.7 31.7 28.2 31.7 23.8 1.09 1.09 1.23 1.10 1.46
G250-1.95-60-40-15-3500 28.5 28.9 25.2 25.2 23.2 25.7 20.0 1.15 1.15 1.25 1.12 1.44
G250-1.95-75-50-15-1500 182.5 91.1 82.9 82.4 77.7 88.0 61.5 1.10 1.11 1.17 1.04 1.48
G250-1.95-75-50-15-2000 109.3 78.2 70.3 70.3 63.4 74.5 49.6 1.11 1.11 1.23 1.05 1.58
G250-1.95-75-50-15-2500 73.3 62.2 57.7 57.7 50.1 58.4 39.9 1.08 1.08 1.24 1.06 1.56
G250-1.95-75-50-15-3000 53.4 49.2 46.2 46.2 39.8 45.5 32.5 1.06 1.06 1.24 1.08 1.51
G250-1.95-75-50-15-3500 41.2 40.6 36.3 36.3 32.4 36.3 27.1 1.12 1.12 1.26 1.12 1.50
G250-1.95-90-60-15-2000 157.9 98.6 89.9 86.7 80.7 92.4 63.3 1.10 1.14 1.22 1.07 1.56
G250-1.95-90-60-15-2500 106.6 83.3 76.6 75.1 66.6 77.8 52.3 1.09 1.11 1.25 1.07 1.59
G250-1.95-90-60-15-3000 77.6 67.5 63.6 63.6 54.1 62.5 43.3 1.06 1.06 1.25 1.08 1.56
G250-1.95-90-60-15-3500 59.0 55.1 51.7 51.7 44.2 50.2 36.2 1.07 1.07 1.25 1.10 1.52
G250-1.95-90-60-15-4000 46.8 46.7 41.3 41.3 36.7 41.1 30.7 1.13 1.13 1.27 1.14 1.52
Mean 1.096 1.101 1.235 1.079 1.524
COV 0.025 0.026 0.019 0.029 0.029
Φ 1.007 1.011 1.137 0.990 1.398
_________________________________________________________________________________________________________________
9-15
Member AS EC3 BS EC3 AS/NZS EC3 BS 5950 EC3
EBL Ult DSM DSM
4600 P1.3 5950 P5 P1.2 4600 P1.3 P5 P1.2
G450-1.90-60-40-15-1500 124.9 104.0 96.1 96.1 83.2 103.2 73.5 1.08 1.08 1.25 1.01 1.41
G450-1.90-60-40-15-2000 74.7 70.3 66.0 66.0 57.2 67.4 52.0 1.07 1.07 1.23 1.04 1.35
G450-1.90-60-40-15-2500 51.0 52.7 45.0 45.0 41.6 47.3 38.6 1.17 1.17 1.27 1.11 1.37
G450-1.90-60-40-15-3000 38.0 42.7 33.5 33.5 32.1 35.8 30.2 1.28 1.28 1.33 1.19 1.42
G450-1.90-60-40-15-3500 30.1 36.0 26.5 26.5 26.0 28.6 24.6 1.36 1.36 1.39 1.26 1.46
G450-1.90-75-50-15-1500 194.0 142.0 133.1 126.0 114.0 140.1 100.2 1.07 1.13 1.25 1.01 1.42
G450-1.90-75-50-15-2000 116.2 103.0 98.8 96.6 82.7 100.0 74.1 1.04 1.07 1.25 1.03 1.39
G450-1.90-75-50-15-2500 77.9 75.8 68.9 68.9 60.6 70.5 55.5 1.10 1.10 1.25 1.08 1.37
G450-1.90-75-50-15-3000 56.6 60.8 49.9 49.9 46.3 52.4 43.0 1.22 1.22 1.31 1.16 1.41
G450-1.90-75-50-15-3500 43.6 51.8 38.4 38.4 36.8 40.9 34.6 1.35 1.35 1.41 1.27 1.50
G450-1.90-90-60-15-2000 170.0 136.0 135.9 125.2 109.4 133.6 96.7 1.00 1.09 1.24 1.02 1.41
G450-1.90-90-60-15-2500 114.5 104.0 101.8 97.0 83.5 99.6 75.1 1.02 1.07 1.25 1.04 1.38
G450-1.90-90-60-15-3000 82.5 82.1 73.6 72.6 64.5 74.6 59.1 1.12 1.13 1.27 1.10 1.39
G450-1.90-90-60-15-3500 62.3 69.2 55.7 55.7 51.2 57.9 47.5 1.24 1.24 1.35 1.19 1.46
G450-1.90-90-60-15-4000 49.6 60.8 44.1 44.1 41.8 46.6 39.2 1.38 1.38 1.45 1.31 1.55
Mean 1.166 1.182 1.299 1.122 1.419
COV 0.111 0.096 0.055 0.091 0.038
0.997 1.029 1.175 0.983 1.297
_________________________________________________________________________________________________________________
9-16
Member BS BS
AS/NZS EC3 EC3 AS/NZS EC3 EC3
4600 P1.3 P1.2 4600 P1.3 P1.2
P5 P5
G550-0.95-40-25-7.5-1500 16.1 17.6 14.3 14.3 13.2 15.0 12.5 1.23 1.23 1.33 1.17 1.41
G550-0.95-40-25-7.5-2000 10.0 13.1 8.8 8.8 8.6 9.5 8.3 1.48 1.48 1.52 1.38 1.58
G550-0.95-40-25-7.5-2500 7.2 10.1 6.3 6.3 6.3 6.8 6.1 1.60 1.60 1.60 1.48 1.65
G550-0.95-55-35-7.5-1500 32.6 29.2 30.2 27.5 24.5 28.9 22.4 0.97 1.06 1.19 1.01 1.30
G550-0.95-55-35-7.5-2000 19.5 22.0 17.6 17.1 16.0 18.1 15.1 1.25 1.29 1.37 1.21 1.46
G550-0.95-55-35-7.5-2500 13.2 17.9 11.8 11.8 11.3 12.5 10.8 1.52 1.52 1.58 1.44 1.66
G550-0.95-55-35-7.5-3000 9.6 14.8 8.6 8.6 8.5 9.2 8.2 1.72 1.72 1.74 1.60 1.81
G550-0.95-70-45-7.5-2500 22.8 24.4 20.9 19.3 18.6 20.8 17.4 1.17 1.27 1.31 1.17 1.40
G550-0.95-70-45-7.5-3000 16.4 20.7 14.9 14.3 13.9 15.3 13.2 1.39 1.45 1.49 1.35 1.57
Mean 1.370 1.402 1.458 1.313 1.537
COV 0.173 0.147 0.117 0.140 0.102
Φ 1.035 1.114 1.219 1.055 1.314
_________________________________________________________________________________________________________________
9-17
EBL Ult DSM DSM
4600 P1.3 5950 P5 P1.2 4600 P1.3 P5 P1.2
G250-1.95-60-40-15-1500 107.0 67.7 59.7 59.7 55.8 65.2 43.6 1.13 1.13 1.21 1.04 1.55
G250-1.95-60-40-15-2000 64.1 52.9 48.2 48.2 42.6 50.1 33.7 1.10 1.10 1.24 1.06 1.57
G250-1.95-60-40-15-2500 43.9 40.1 37.7 37.7 32.6 37.4 26.6 1.06 1.07 1.23 1.07 1.51
G250-1.95-60-40-15-3000 32.8 32.0 28.9 28.9 25.8 29.1 21.6 1.11 1.11 1.24 1.10 1.48
G250-1.95-60-40-15-3500 26.0 26.7 22.9 22.9 21.2 23.5 18.1 1.17 1.17 1.26 1.13 1.47
G250-1.95-75-50-15-1500 166.2 86.4 78.0 77.2 72.7 82.9 57.6 1.11 1.12 1.19 1.04 1.50
G250-1.95-75-50-15-2000 99.6 73.5 65.7 65.7 58.8 69.5 46.2 1.12 1.12 1.25 1.06 1.59
G250-1.95-75-50-15-2500 60.8 57.5 53.5 53.5 46.2 54.0 37.0 1.08 1.08 1.25 1.07 1.56
G250-1.95-75-50-15-3000 48.6 45.4 42.4 42.4 36.5 41.8 30.0 1.07 1.07 1.24 1.09 1.51
G250-1.95-75-50-15-3500 37.6 37.5 33.1 33.1 29.7 33.3 24.9 1.13 1.13 1.26 1.13 1.50
G250-1.95-90-60-15-2000 143.8 92.9 84.3 81.0 75.2 86.6 59.1 1.10 1.15 1.24 1.07 1.57
G250-1.95-90-60-15-2500 90.1 77.6 71.3 69.8 61.7 72.3 48.7 1.09 1.11 1.26 1.07 1.59
G250-1.95-90-60-15-3000 70.7 62.4 58.8 58.6 49.8 57.7 40.1 1.06 1.06 1.25 1.08 1.56
G250-1.95-90-60-15-3500 53.8 50.9 47.3 47.3 40.6 46.1 33.4 1.08 1.08 1.25 1.10 1.52
G250-1.95-90-60-15-4000 42.7 43.1 37.6 37.6 33.6 37.6 28.3 1.15 1.15 1.28 1.15 1.52
Mean 1.103 1.109 1.243 1.084 1.534
COV 0.029 0.030 0.018 0.030 0.025
Φ 1.012 1.017 1.144 0.994 1.409
_________________________________________________________________________________________________________________
9-18
Table 9.9: Comparison of FEA Ultimate Loads Results with Code Predictions for G450-1.90 Series at 200oC
Member AS/NZS EC3 BS EC3 AS/NZS EC3 BS EC3
EBL Ult DSM DSM
4600 P1.3 5950 P5 P1.2 4600 P1.3 5950 P5 P1.2
G450-1.90-60-40-15-1500 113.7 97.0 90.7 90.7 77.8 96.5 69.3 1.07 1.07 1.25 1.00 1.40
G450-1.90-60-40-15-2000 68.1 64.9 60.1 60.1 52.9 62.1 48.4 1.08 1.08 1.23 1.05 1.34
G450-1.90-60-40-15-2500 46.5 48.8 40.9 40.9 38.3 43.4 35.7 1.19 1.19 1.28 1.12 1.37
G450-1.90-60-40-15-3000 34.6 39.7 30.5 30.5 29.5 32.8 27.8 1.30 1.30 1.35 1.21 1.43
G450-1.90-60-40-15-3500 27.4 33.5 24.1 24.1 23.8 26.1 22.7 1.39 1.39 1.41 1.28 1.48
G450-1.90-75-50-15-1500 176.6 133.0 127.1 119.8 107.1 133.0 95.4 1.05 1.11 1.24 1.00 1.39
G450-1.90-75-50-15-2000 105.9 95.1 91.9 89.7 76.7 92.7 69.4 1.03 1.06 1.24 1.03 1.37
G450-1.90-75-50-15-2500 70.9 70.1 62.8 62.8 55.8 64.8 51.6 1.12 1.12 1.26 1.08 1.36
G450-1.90-75-50-15-3000 51.5 56.7 45.5 45.5 42.5 48.0 39.8 1.25 1.25 1.34 1.18 1.43
G450-1.90-75-50-15-3500 39.7 48.4 35.0 35.0 33.7 37.4 31.9 1.38 1.38 1.44 1.29 1.52
G450-1.90-90-60-15-2000 154.8 128.0 128.3 117.7 102.2 125.2 91.5 1.00 1.09 1.25 1.02 1.40
G450-1.90-90-60-15-2500 104.3 95.6 93.7 89.1 77.3 92.0 70.3 1.02 1.07 1.24 1.04 1.36
G450-1.90-90-60-15-3000 75.1 76.2 67.0 66.1 59.4 68.6 54.9 1.14 1.15 1.28 1.11 1.39
G450-1.90-90-60-15-3500 57.1 64.7 50.8 50.8 47.0 53.1 44.0 1.27 1.28 1.38 1.22 1.47
G450-1.90-90-60-15-4000 45.2 57.0 40.2 40.2 38.3 42.6 36.2 1.42 1.42 1.49 1.34 1.58
Mean 1.180 1.197 1.310 1.132 1.418
COV 0.123 0.108 0.063 0.101 0.047
Φ 0.993 1.028 1.178 0.980 1.290
_________________________________________________________________________________________________________________
9-19
Member BS BS
AS/NZS EC3 EC3 AS/NZS EC3 EC3
4600 P1.3 P1.2 4600 P1.3 P1.2
P5 P5
G550-0.95-40-25-7.5-1500 14.0 15.9 12.4 12.4 11.6 13.2 11.1 1.28 1.28 1.37 1.21 1.43
G550-0.95-40-25-7.5-2000 8.8 11.9 7.7 7.7 7.6 8.3 7.3 1.55 1.55 1.57 1.43 1.63
G550-0.95-40-25-7.5-2500 6.3 9.1 5.5 5.5 5.5 6.0 5.4 1.65 1.65 1.64 1.51 1.68
G550-0.95-55-35-7.5-1500 28.5 26.1 26.3 23.9 21.8 25.6 20.3 0.99 1.09 1.20 1.02 1.29
G550-0.95-55-35-7.5-2000 17.0 20.2 15.3 14.9 14.1 15.9 13.4 1.32 1.36 1.43 1.27 1.51
G550-0.95-55-35-7.5-2500 11.5 15.8 10.3 10.3 9.9 10.9 9.5 1.54 1.54 1.59 1.45 1.66
G550-0.95-55-35-7.5-3000 8.4 13.0 7.5 7.5 7.5 8.1 7.2 1.73 1.73 1.75 1.61 1.80
G550-0.95-70-45-7.5-2500 19.9 21.5 18.2 16.8 16.4 18.4 15.5 1.18 1.28 1.31 1.17 1.39
G550-0.95-70-45-7.5-3000 14.3 18.6 12.9 12.4 12.2 13.4 11.7 1.44 1.50 1.52 1.38 1.59
Mean 1.410 1.443 1.486 1.340 1.551
COV 0.168 0.142 0.116 0.140 0.104
Φ 1.076 1.157 1.244 1.079 1.321
_________________________________________________________________________________________________________________
9-20
Member ASNZS EC3 BS 5950 EC3 AS/NZS EC3 BS 5950 EC3
EBL Ult DSM DSM
4600 P1.3 P5 P1.2 4600 P1.3 P5 P1.2
G250-1.95-60-40-15-1500 90.0 50.7 45.0 45.0 42.6 48.6 33.4 1.13 1.13 1.19 1.04 1.52
G250-1.95-60-40-15-2000 54.0 41.7 37.5 37.5 33.7 39.2 26.5 1.11 1.11 1.24 1.06 1.58
G250-1.95-60-40-15-2500 36.9 32.3 30.3 30.3 26.3 30.2 21.2 1.06 1.07 1.23 1.07 1.52
G250-1.95-60-40-15-3000 27.6 27.9 24.2 24.2 21.1 23.7 17.4 1.15 1.15 1.32 1.18 1.60
G250-1.95-60-40-15-3500 21.9 21.6 19.3 19.3 17.5 19.4 14.7 1.12 1.12 1.24 1.12 1.47
G250-1.95-75-50-15-1500 139.8 63.8 58.1 58.1 55.2 61.3 43.6 1.10 1.10 1.16 1.04 1.46
G250-1.95-75-50-15-2000 83.8 56.4 50.2 50.2 46.0 53.1 35.7 1.12 1.12 1.23 1.06 1.58
G250-1.95-75-50-15-2500 56.2 45.6 42.1 42.1 37.0 42.8 29.2 1.08 1.08 1.23 1.06 1.56
G250-1.95-75-50-15-3000 40.9 36.7 34.5 34.5 29.7 33.9 24.0 1.06 1.06 1.24 1.08 1.53
G250-1.95-75-50-15-3500 31.6 30.2 27.8 27.8 24.3 27.3 20.1 1.09 1.09 1.24 1.11 1.50
G250-1.95-90-60-15-2000 121.0 70.2 63.6 61.8 57.9 65.4 45.2 1.10 1.14 1.21 1.07 1.55
G250-1.95-90-60-15-2500 81.7 60.9 55.1 54.4 48.7 56.3 37.9 1.11 1.12 1.25 1.08 1.61
G250-1.95-90-60-15-3000 59.5 50.1 46.7 46.7 40.1 46.2 31.7 1.07 1.07 1.25 1.09 1.58
G250-1.95-90-60-15-3500 45.2 41.1 38.7 38.7 33.0 37.5 26.7 1.06 1.06 1.24 1.10 1.54
G250-1.95-90-60-15-4000 35.9 34.7 31.6 31.6 27.6 30.9 22.8 1.10 1.10 1.26 1.12 1.52
Mean 1.098 1.101 1.235 1.086 1.542
COV 0.024 0.026 0.029 0.032 0.029
Φ 1.009 1.011 1.133 0.995 1.414
_________________________________________________________________________________________________________________
9-21
Member AS/NZS EC3 BS 5950 EC3 AS/NZS EC3 BS EC3
EBL Ult DSM DSM
4600 P1.3 P5 P1.2 4600 P1.3 5950 P5 P1.2
G450-1.90-60-40-15-1500 95.7 84.2 80.0 79.9 67.8 83.9 61.2 1.05 1.05 1.24 1.00 1.38
G450-1.90-60-40-15-2000 57.2 55.9 50.6 50.6 45.4 53.0 41.9 1.11 1.11 1.23 1.06 1.33
G450-1.90-60-40-15-2500 39.1 42.4 34.4 34.4 32.6 36.9 30.7 1.23 1.23 1.30 1.15 1.38
G450-1.90-60-40-15-3000 29.1 34.7 25.6 25.6 25.0 27.8 23.8 1.35 1.35 1.39 1.25 1.46
G450-1.90-60-40-15-3500 23.0 29.1 20.3 20.3 20.2 22.1 19.3 1.43 1.43 1.44 1.32 1.51
G450-1.90-75-50-15-1500 148.6 118.0 114.5 107.1 94.1 118.4 85.6 1.03 1.10 1.25 1.00 1.38
G450-1.90-75-50-15-2000 89.1 81.6 79.1 77.0 66.0 79.7 60.8 1.03 1.06 1.24 1.02 1.34
G450-1.90-75-50-15-2500 59.7 60.6 52.8 52.8 47.7 55.2 44.6 1.15 1.15 1.27 1.10 1.36
G450-1.90-75-50-15-3000 43.4 49.6 38.2 38.2 36.1 40.7 34.2 1.30 1.30 1.37 1.22 1.45
G450-1.90-75-50-15-3500 33.4 42.6 29.4 29.4 28.6 31.7 27.3 1.45 1.45 1.49 1.34 1.56
G450-1.90-90-60-15-2000 130.2 109.0 113.2 102.9 89.0 109.5 81.4 0.96 1.06 1.23 1.00 1.34
G450-1.90-90-60-15-2500 87.7 82.0 79.0 75.1 66.4 78.9 61.4 1.04 1.09 1.24 1.04 1.33
G450-1.90-90-60-15-3000 63.2 66.4 56.4 55.6 50.7 58.4 47.5 1.18 1.20 1.31 1.14 1.40
G450-1.90-90-60-15-3500 48.0 57.0 42.7 42.7 40.0 45.1 37.8 1.34 1.34 1.43 1.27 1.51
G450-1.90-90-60-15-4000 38.0 50.3 33.8 33.8 32.5 36.1 31.0 1.49 1.49 1.55 1.39 1.62
Mean 1.209 1.227 1.331 1.152 1.423
COV 0.144 0.126 0.079 0.119 0.064
Φ 0.986 1.028 1.180 0.976 1.279
_________________________________________________________________________________________________________________
9-22
EBL Ult DSM DSM
4600 P1.3 P5 P1.2 4600 P1.3 P5 P1.2
G550-0.95-40-25-7.5-1500 11.9 13.7 10.5 10.5 10.0 11.2 9.6 1.30 1.30 1.38 1.22 1.43
G550-0.95-40-25-7.5-2000 7.4 10.2 6.5 6.5 6.5 7.1 6.3 1.56 1.56 1.58 1.44 1.62
G550-0.95-40-25-7.5-2500 5.3 7.8 4.7 4.7 4.7 5.1 4.6 1.66 1.67 1.64 1.52 1.68
G550-0.95-55-35-7.5-1500 24.1 22.3 22.3 20.3 18.8 22.0 17.7 1.00 1.10 1.19 1.01 1.26
G550-0.95-55-35-7.5-2000 14.4 17.0 13.0 12.6 12.1 13.6 11.6 1.30 1.34 1.41 1.25 1.47
G550-0.95-55-35-7.5-2500 9.7 13.6 8.7 8.7 8.5 9.3 8.2 1.56 1.56 1.61 1.46 1.66
G550-0.95-55-35-7.5-3000 7.1 11.2 6.4 6.4 6.4 6.9 6.2 1.76 1.76 1.76 1.63 1.81
G550-0.95-70-45-7.5-2500 16.9 18.5 15.5 14.2 14.1 15.7 13.4 1.20 1.30 1.32 1.18 1.38
G550-0.95-70-45-7.5-3000 12.1 15.8 11.0 10.6 10.5 11.5 10.1 1.44 1.50 1.51 1.37 1.57
Mean 1.420 1.453 1.487 1.342 1.542
COV 0.171 0.144 0.121 0.144 0.111
Φ 1.077 1.161 1.235 1.072 1.301
_________________________________________________________________________________________________________________
9-23
EBL Ult DSM DSM
4600 P1.3 P5 P1.2 4600 P1.3 P5 P1.2
G250-1.95-60-40-15-1500 73.0 29.3 33.7 33.7 32.1 35.9 25.3 0.87 0.87 0.91 0.82 1.16
G250-1.95-60-40-15-2000 43.8 24.8 28.5 28.5 26.0 29.9 20.3 0.87 0.87 0.96 0.83 1.22
G250-1.95-60-40-15-2500 30.0 21.1 23.6 23.6 20.7 23.6 16.5 0.89 0.89 1.02 0.90 1.28
G250-1.95-60-40-15-3000 22.4 17.8 19.3 19.3 16.7 18.8 13.7 0.92 0.92 1.07 0.95 1.30
G250-1.95-60-40-15-3500 17.7 15.2 15.6 15.6 13.9 15.4 11.6 0.97 0.97 1.09 0.99 1.31
G250-1.95-75-50-15-1500 113.4 37.8 43.1 43.1 41.5 45.2 32.7 0.88 0.88 0.91 0.84 1.16
G250-1.95-75-50-15-2000 68.0 32.5 37.8 37.8 35.2 39.9 27.2 0.86 0.86 0.92 0.81 1.19
G250-1.95-75-50-15-2500 45.6 28.2 32.3 32.3 28.8 33.0 22.5 0.87 0.87 0.98 0.85 1.25
G250-1.95-75-50-15-3000 33.2 24.2 27.0 27.0 23.5 26.6 18.7 0.90 0.90 1.03 0.91 1.29
G250-1.95-75-50-15-3500 25.6 20.8 22.3 22.3 19.3 21.6 15.8 0.93 0.93 1.08 0.96 1.32
G250-1.95-90-60-15-2000 98.2 40.7 47.5 46.6 43.9 48.9 34.1 0.86 0.87 0.93 0.83 1.19
G250-1.95-90-60-15-2500 66.2 35.8 41.8 41.6 37.5 42.8 29.0 0.86 0.86 0.95 0.84 1.24
G250-1.95-90-60-15-3000 48.2 31.5 36.0 36.0 31.3 35.9 24.5 0.87 0.88 1.01 0.88 1.29
G250-1.95-90-60-15-3500 36.7 27.4 30.5 30.5 26.1 29.5 20.8 0.90 0.90 1.05 0.93 1.32
G250-1.95-90-60-15-4000 29.1 23.8 25.5 25.5 21.9 24.5 17.9 0.93 0.94 1.09 0.97 1.33
Mean 0.892 0.894 0.999 0.887 1.256
COV 0.038 0.037 0.066 0.069 0.047
Φ 0.815 0.817 0.896 0.794 1.142
_________________________________________________________________________________________________________________
9-24
EBL Ult DSM DSM
4600 P1.3 5950 P5 P1.2 4600 P1.3 P5 P1.2
G450-1.90-60-40-15-1500 77.7 65.4 62.6 61.9 53.5 66.3 47.8 1.05 1.05 1.22 0.99 1.37
G450-1.90-60-40-15-2000 46.4 44.0 41.0 40.7 36.2 42.5 33.2 1.07 1.07 1.21 1.04 1.32
G450-1.90-60-40-15-2500 31.7 33.2 27.9 27.8 26.2 29.7 24.5 1.19 1.19 1.27 1.12 1.36
G450-1.90-60-40-15-3000 23.6 27.0 20.8 20.7 20.1 22.4 19.0 1.30 1.30 1.34 1.21 1.42
G450-1.90-60-40-15-3500 18.7 22.8 16.5 16.4 16.3 17.9 15.5 1.39 1.39 1.40 1.28 1.47
G450-1.90-75-50-15-1500 120.5 90.2 88.0 82.8 73.8 91.9 66.0 1.03 1.09 1.22 0.98 1.37
G450-1.90-75-50-15-2000 72.2 64.4 63.1 61.5 52.6 63.6 47.8 1.02 1.05 1.23 1.01 1.35
G450-1.90-75-50-15-2500 48.4 47.5 42.8 42.8 38.2 44.3 35.4 1.11 1.11 1.24 1.07 1.34
G450-1.90-75-50-15-3000 35.2 38.6 31.0 31.0 29.1 32.8 27.3 1.24 1.24 1.33 1.18 1.42
G450-1.90-75-50-15-3500 27.1 33.0 23.9 23.9 23.0 25.6 21.8 1.38 1.38 1.43 1.29 1.51
G450-1.90-90-60-15-2000 105.6 86.2 88.4 81.1 70.2 86.2 63.2 0.97 1.06 1.23 1.00 1.36
G450-1.90-90-60-15-2500 71.2 65.0 64.1 60.9 53.0 63.0 48.3 1.01 1.07 1.23 1.03 1.34
G450-1.90-90-60-15-3000 51.2 51.7 45.7 45.1 40.7 46.9 37.7 1.13 1.15 1.27 1.10 1.37
G450-1.90-90-60-15-3500 38.9 44.1 34.6 34.6 32.2 36.3 30.1 1.27 1.27 1.37 1.21 1.46
G450-1.90-90-60-15-4000 30.9 38.9 27.4 27.4 26.2 29.1 24.8 1.42 1.42 1.48 1.34 1.57
Mean 1.172 1.189 1.299 1.123 1.403
COV 0.129 0.113 0.068 0.106 0.051
1.071 1.087 1.165 0.966 1.272
_________________________________________________________________________________________________________________
9-25
EBL Ult DSM DSM
4600 P1.3 P5 P1.2 4600 P1.3 5950 P5 P1.2
G550-0.95-40-25-7.5-1500 9.8 11.2 8.7 8.7 8.2 9.3 7.9 1.29 1.29 1.37 1.21 1.43
G550-0.95-40-25-7.5-2000 6.1 8.4 5.4 5.4 5.3 5.8 5.2 1.55 1.55 1.57 1.43 1.62
G550-0.95-40-25-7.5-2500 4.4 6.4 3.8 3.8 3.9 4.2 3.8 1.66 1.66 1.64 1.51 1.68
G550-0.95-55-35-7.5-1500 19.9 18.3 18.4 16.8 15.4 18.1 14.5 0.99 1.09 1.19 1.01 1.27
G550-0.95-55-35-7.5-2000 11.9 13.8 10.7 10.4 9.9 11.2 9.5 1.28 1.32 1.39 1.23 1.45
G550-0.95-55-35-7.5-2500 8.0 11.1 7.2 7.2 7.0 7.7 6.7 1.54 1.54 1.59 1.45 1.65
G550-0.95-55-35-7.5-3000 5.9 9.3 5.3 5.3 5.2 5.7 5.1 1.77 1.77 1.78 1.64 1.83
G550-0.95-70-45-7.5-2500 13.9 15.3 12.8 11.7 11.6 12.9 11.0 1.20 1.30 1.32 1.18 1.39
G550-0.95-70-45-7.5-3000 10.0 13.2 9.1 8.7 8.6 9.5 8.3 1.45 1.52 1.53 1.40 1.59
Mean 1.416 1.449 1.486 1.341 1.544
COV 0.173 0.147 0.123 0.146 0.112
Φ 1.070 1.151 1.230 1.068 1.301
_________________________________________________________________________________________________________________
9-26
EBL Ult DSM DSM
4600 P1.3 5950 P5 P1.2 4600 P1.3 P5 P1.2
G250-1.95-60-40-15-1500 56.0 22.4 24.2 24.2 23.2 25.7 18.3 0.92 0.92 0.96 0.87 1.22
G250-1.95-60-40-15-2000 33.6 19.5 20.8 20.8 19.1 21.7 14.9 0.94 0.94 1.02 0.90 1.30
G250-1.95-60-40-15-2500 23.0 16.0 17.5 17.5 15.4 17.5 12.2 0.91 0.91 1.04 0.92 1.31
G250-1.95-60-40-15-3000 17.7 14.0 14.5 14.5 12.6 14.1 10.2 0.96 0.96 1.11 1.00 1.37
G250-1.95-60-40-15-3500 13.6 11.9 12.0 12.0 10.5 11.6 8.7 0.99 0.99 1.13 1.03 1.37
G250-1.95-75-50-15-1500 87.0 28.5 30.9 30.9 30.0 32.2 23.6 0.92 0.92 0.95 0.89 1.21
G250-1.95-75-50-15-2000 52.1 25.3 27.4 27.4 25.8 28.8 19.9 0.92 0.92 0.98 0.88 1.27
G250-1.95-75-50-15-2500 35.0 22.2 23.7 23.7 21.4 24.3 16.6 0.94 0.94 1.04 0.92 1.34
G250-1.95-75-50-15-3000 25.5 19.1 20.1 20.1 17.6 19.8 13.9 0.95 0.95 1.09 0.96 1.37
G250-1.95-75-50-15-3500 19.7 16.3 16.8 16.8 14.6 16.2 11.8 0.97 0.97 1.12 1.00 1.38
G250-1.95-90-60-15-2000 75.3 31.3 34.2 33.8 31.9 35.1 24.7 0.91 0.93 0.98 0.89 1.27
G250-1.95-90-60-15-2500 58.2 28.1 30.4 30.4 27.6 31.2 21.2 0.92 0.93 1.02 0.90 1.33
G250-1.95-90-60-15-3000 37.0 24.8 26.5 26.5 23.3 26.5 18.1 0.94 0.94 1.06 0.94 1.37
G250-1.95-90-60-15-3500 28.5 21.6 22.7 22.7 19.6 22.1 15.5 0.95 0.95 1.10 0.98 1.40
G250-1.95-90-60-15-4000 22.3 18.6 19.3 19.3 16.5 18.4 13.4 0.97 0.97 1.13 1.01 1.39
Mean 0.941 0.942 1.049 0.938 1.327
COV 0.024 0.024 0.060 0.057 0.046
Φ 0.865 0.866 0.946 0.848 1.207
_________________________________________________________________________________________________________________
9-27
Member AS/NZ EC3 BS 5950 EC3 AS/NZS EC3 BS 5950 EC3
EBL Ult DSM DSM
S 4600 P1.3 P5 P1.2 4600 P1.3 P5 P1.2
G450-1.90-60-40-15-1500 59.6 43.7 41.4 41.4 36.8 45.3 32.0 1.05 1.05 1.19 0.97 1.36
G450-1.90-60-40-15-2000 35.6 31.4 30.4 30.4 26.2 31.0 23.4 1.03 1.03 1.20 1.01 1.34
G450-1.90-60-40-15-2500 24.3 23.5 21.4 21.4 19.3 22.1 17.7 1.10 1.10 1.22 1.06 1.33
G450-1.90-60-40-15-3000 18.1 18.8 16.0 16.0 15.0 16.8 13.9 1.18 1.18 1.25 1.12 1.35
G450-1.90-60-40-15-3500 14.3 15.8 12.6 12.6 12.2 13.4 11.4 1.25 1.25 1.30 1.18 1.38
G450-1.90-75-50-15-1500 92.5 57.5 56.0 53.7 49.7 59.4 42.6 1.03 1.07 1.16 0.97 1.35
G450-1.90-75-50-15-2000 55.4 45.2 43.8 43.2 37.4 45.1 32.7 1.03 1.05 1.21 1.00 1.38
G450-1.90-75-50-15-2500 37.1 33.9 32.7 32.7 28.0 32.6 25.1 1.04 1.04 1.21 1.04 1.35
G450-1.90-75-50-15-3000 27.0 26.8 23.8 23.8 21.5 24.5 19.7 1.13 1.13 1.24 1.09 1.36
G450-1.90-75-50-15-3500 20.8 22.5 18.3 18.3 17.2 19.2 16.0 1.23 1.23 1.31 1.17 1.41
G450-1.90-90-60-15-2000 81.1 59.4 58.6 54.6 48.7 58.6 41.6 1.01 1.09 1.22 1.01 1.43
G450-1.90-90-60-15-2500 54.6 46.7 46.2 44.3 38.1 45.4 33.3 1.01 1.06 1.23 1.03 1.40
G450-1.90-90-60-15-3000 39.3 36.6 35.1 34.6 29.9 34.7 26.7 1.04 1.06 1.23 1.06 1.37
G450-1.90-90-60-15-3500 29.9 30.2 26.6 26.6 23.9 27.1 21.7 1.14 1.14 1.26 1.11 1.39
G450-1.90-90-60-15-4000 23.7 26.2 21.0 21.0 19.6 21.9 18.1 1.24 1.25 1.34 1.20 1.45
Mean 1.101 1.113 1.237 1.068 1.378
COV 0.080 0.070 0.039 0.070 0.024
Φ 0.975 0.995 1.130 0.955 1.266
_________________________________________________________________________________________________________________
9-28
EBL Ult DSM DSM
4600 P1.3 P5 P1.2 4600 P1.3 5950 P5 P1.2
G550-0.95-40-25-7.5-1500 7.7 8.1 6.8 6.8 6.3 7.2 5.9 1.19 1.19 1.29 1.14 1.37
G550-0.95-40-25-7.5-2000 4.8 6.0 4.2 4.2 4.1 4.5 4.0 1.43 1.43 1.46 1.33 1.53
G550-0.95-40-25-7.5-2500 3.4 4.7 3.0 3.0 3.0 3.3 2.9 1.56 1.56 1.56 1.44 1.61
G550-0.95-55-35-7.5-1500 15.6 13.8 14.5 13.2 11.6 13.7 10.6 0.95 1.05 1.19 1.01 1.31
G550-0.95-55-35-7.5-2000 9.3 10.8 8.5 8.2 7.6 8.6 7.1 1.28 1.32 1.41 1.25 1.51
G550-0.95-55-35-7.5-2500 6.3 8.2 5.7 5.7 5.4 6.0 5.1 1.46 1.46 1.53 1.38 1.61
G550-0.95-55-35-7.5-3000 4.6 6.9 4.1 4.1 4.1 4.4 3.9 1.67 1.67 1.70 1.56 1.77
G550-0.95-70-45-7.5-2500 11.0 11.6 10.0 9.2 8.9 9.9 8.2 1.16 1.26 1.31 1.17 1.41
G550-0.95-70-45-7.5-3000 7.9 9.8 7.1 6.8 6.7 7.3 6.3 1.37 1.43 1.47 1.33 1.55
Mean 1.340 1.372 1.434 1.290 1.518
COV 0.165 0.140 0.108 0.131 0.093
Φ 1.029 1.104 1.215 1.053 1.314
_________________________________________________________________________________________________________________
9-29
EBL Ult DSM DSM
4600 P1.3 P5 P1.2 4600 P1.3 5950 P5 P1.2
G250-1.95-60-40-15-1500 39.0 16.3 16.1 16.1 15.5 17.0 12.3 1.01 1.01 1.05 0.96 1.33
G250-1.95-60-40-15-2000 23.4 14.2 14.0 14.0 12.9 14.6 10.1 1.02 1.02 1.10 0.98 1.41
G250-1.95-60-40-15-2500 16.0 12.0 11.9 11.9 10.5 11.9 8.3 1.01 1.01 1.14 1.01 1.44
G250-1.95-60-40-15-3000 12.0 10.0 9.9 9.9 8.6 9.6 7.0 1.01 1.01 1.16 1.04 1.44
G250-1.95-60-40-15-3500 9.5 8.5 8.3 8.3 7.2 8.0 6.0 1.02 1.02 1.17 1.06 1.42
G250-1.95-75-50-15-1500 60.6 19.9 20.5 20.5 19.9 21.2 15.7 0.97 0.97 1.00 0.94 1.27
G250-1.95-75-50-15-2000 36.3 18.3 18.3 18.3 17.3 19.1 13.3 1.00 1.00 1.06 0.96 1.37
G250-1.95-75-50-15-2500 24.4 16.2 15.9 15.9 14.5 16.3 11.2 1.02 1.02 1.12 0.99 1.45
G250-1.95-75-50-15-3000 17.7 13.8 13.7 13.7 12.0 13.5 9.4 1.01 1.01 1.15 1.02 1.46
G250-1.95-75-50-15-3500 13.7 11.7 11.6 11.6 10.0 11.1 8.1 1.01 1.01 1.17 1.05 1.45
G250-1.95-90-60-15-2000 52.5 22.3 22.8 22.6 21.4 23.3 16.5 0.98 0.99 1.04 0.96 1.35
G250-1.95-90-60-15-2500 35.4 20.4 20.4 20.4 18.7 20.9 14.3 1.00 1.00 1.09 0.98 1.43
G250-1.95-90-60-15-3000 25.8 18.1 17.9 17.9 15.9 17.9 12.2 1.01 1.01 1.14 1.01 1.48
G250-1.95-90-60-15-3500 19.6 15.6 15.5 15.5 13.4 15.0 10.5 1.01 1.01 1.16 1.04 1.48
G250-1.95-90-60-15-4000 15.6 13.4 13.2 13.2 11.4 12.6 9.1 1.01 1.01 1.18 1.06 1.47
Mean 1.006 1.006 1.116 1.003 1.416
COV 0.013 0.012 0.050 0.041 0.044
Φ 0.927 0.927 1.013 0.915 1.290
_________________________________________________________________________________________________________________
9-30
Member DS AS/NZS EC3 BS 5950 EC3 AS/NZ EC3 BS EC3
EBL Ult DSM
M 4600 P1.3 P5 P1.2 S 4600 P1.3 5950 P5 P1.2
G450-1.90-60-40-15-1500 41.5 16.4 15.3 15.3 14.8 15.9 12.6 1.07 1.07 1.11 1.03 1.30
G450-1.90-60-40-15-2000 24.8 14.4 13.5 13.5 12.7 14.0 10.7 1.06 1.06 1.14 1.03 1.34
G450-1.90-60-40-15-2500 17.0 12.3 11.7 11.7 10.5 11.7 9.0 1.05 1.05 1.17 1.05 1.36
G450-1.90-60-40-15-3000 12.6 10.3 10.0 10.0 8.8 9.7 7.7 1.03 1.03 1.18 1.06 1.35
G450-1.90-60-40-15-3500 10.0 8.8 8.5 8.5 7.4 8.1 6.6 1.02 1.03 1.18 1.08 1.33
G450-1.90-75-50-15-1500 74.8 20.5 19.4 19.4 18.9 19.8 15.4 1.06 1.06 1.09 1.03 1.34
G450-1.90-75-50-15-2000 38.6 18.6 17.5 17.5 16.7 18.1 13.6 1.06 1.06 1.11 1.03 1.37
G450-1.90-75-50-15-2500 25.9 16.5 15.6 15.6 14.4 15.9 11.8 1.06 1.06 1.15 1.04 1.40
G450-1.90-75-50-15-3000 14.5 13.2 13.6 13.6 12.1 13.4 10.2 0.97 0.97 1.09 0.98 1.30
G450-1.90-75-50-15-3500 14.5 12.1 11.7 11.7 10.2 11.2 8.8 1.03 1.03 1.19 1.08 1.38
G450-1.90-90-60-15-2000 56.5 22.8 21.7 21.6 20.6 22.0 15.6 1.05 1.05 1.11 1.04 1.46
G450-1.90-90-60-15-2500 38.0 20.8 19.7 19.7 18.3 20.0 14.1 1.06 1.06 1.13 1.04 1.48
G450-1.90-90-60-15-3000 27.4 18.5 17.6 17.6 15.9 17.6 12.5 1.05 1.05 1.16 1.05 1.48
G450-1.90-90-60-15-3500 20.8 16.0 15.5 15.5 13.7 15.1 11.0 1.03 1.03 1.17 1.06 1.45
G450-1.90-90-60-15-4000 16.5 13.8 13.6 13.6 11.7 12.9 9.7 1.02 1.02 1.18 1.07 1.42
Mean 1.042 1.042 1.143 1.045 1.383
COV 0.024 0.024 0.031 0.023 0.044
Φ 0.958 0.958 1.048 0.961 1.260
_________________________________________________________________________________________________________________
9-31
EBL Ult DSM DSM
4600 P1.3 P5 P1.2 4600 P1.3 5950 P5 P1.2
G550-0.95-40-25-7.5-1500 5.6 3.8 3.8 3.8 3.4 3.7 2.8 1.02 1.02 1.12 1.03 1.37
G550-0.95-40-25-7.5-2000 3.5 2.9 2.9 2.9 2.5 2.7 2.2 1.00 1.00 1.16 1.07 1.34
G550-0.95-40-25-7.5-2500 2.5 2.3 2.2 2.2 1.9 2.1 1.7 1.04 1.05 1.18 1.10 1.33
G550-0.95-55-35-7.5-1500 11.4 5.9 5.9 5.8 5.4 5.8 3.8 1.00 1.02 1.09 1.02 1.55
G550-0.95-55-35-7.5-2000 6.8 4.9 4.9 4.9 4.3 4.7 3.2 1.00 1.00 1.14 1.05 1.52
G550-0.95-55-35-7.5-2500 4.6 3.9 3.9 3.9 3.3 3.6 2.7 0.99 0.99 1.16 1.07 1.45
G550-0.95-55-35-7.5-3000 3.4 3.1 3.0 3.0 2.6 2.8 2.2 1.03 1.03 1.19 1.11 1.41
G550-0.95-70-45-7.5-2500 8.0 6.1 5.9 5.6 5.0 5.4 3.5 1.03 1.09 1.23 1.15 1.75
G550-0.95-70-45-7.5-3000 5.7 5.0 4.9 4.7 4.1 4.4 3.1 1.02 1.06 1.23 1.15 1.64
Mean 1.016 1.029 1.167 1.080 1.482
COV 0.018 0.031 0.039 0.044 0.097
Φ 0.935 0.942 1.064 0.981 1.276
_________________________________________________________________________________________________________________
9-32
Member AS/NZ EC3 BS 5950 EC3 AS/NZ EC3 BS 5950 EC3
EBL Ult DSM DSM
S 4600 P1.3 P5 P1.2 S 4600 P1.3 P5 P1.2
G250-1.95-60-40-15-1500 22.0 9.1 9.1 9.1 8.8 9.6 6.9 1.00 1.00 1.04 0.95 1.31
G250-1.95-60-40-15-2000 13.2 8.1 7.9 7.9 7.3 8.2 5.7 1.02 1.02 1.11 0.98 1.42
G250-1.95-60-40-15-2500 9.0 6.9 6.7 6.7 5.9 6.7 4.7 1.02 1.02 1.15 1.02 1.46
G250-1.95-60-40-15-3000 6.8 5.7 5.6 5.6 4.9 5.4 3.9 1.01 1.01 1.17 1.05 1.44
G250-1.95-60-40-15-3500 5.4 4.8 4.7 4.7 4.1 4.5 3.4 1.02 1.02 1.18 1.06 1.42
G250-1.95-75-50-15-1500 34.2 11.4 11.6 11.6 11.3 12.0 8.9 0.98 0.98 1.01 0.95 1.28
G250-1.95-75-50-15-2000 20.5 10.4 10.4 10.4 9.8 10.8 7.5 1.00 1.01 1.06 0.96 1.38
G250-1.95-75-50-15-2500 13.8 9.3 9.0 9.0 8.2 9.2 6.3 1.03 1.03 1.13 1.00 1.46
G250-1.95-75-50-15-3000 10.0 7.9 7.7 7.7 6.8 7.6 5.3 1.02 1.02 1.16 1.03 1.47
G250-1.95-75-50-15-3500 7.7 6.6 6.5 6.5 5.7 6.3 4.6 1.01 1.01 1.17 1.05 1.45
G250-1.95-90-60-15-2000 29.6 12.8 12.9 12.8 12.1 13.2 9.4 0.99 1.00 1.06 0.97 1.37
G250-1.95-90-60-15-2500 19.0 11.7 11.5 11.5 10.6 11.8 8.1 1.02 1.02 1.11 0.99 1.45
G250-1.95-90-60-15-3000 14.6 10.3 10.1 10.1 9.0 10.1 6.9 1.02 1.02 1.15 1.02 1.49
G250-1.95-90-60-15-3500 11.1 8.9 8.8 8.8 7.6 8.5 6.0 1.01 1.01 1.17 1.04 1.49
G250-1.95-90-60-15-4000 8.8 7.6 7.5 7.5 6.4 7.1 5.2 1.02 1.02 1.18 1.06 1.48
Mean 1.011 1.012 1.122 1.009 1.424
COV 0.012 0.012 0.050 0.041 0.044
Φ 0.932 0.933 1.018 0.921 1.297
_________________________________________________________________________________________________________________
9-33
Member DS AS/NZS EC3 BS 5950 EC3 AS/NZS EC3 BS 5950 EC3
EBL Ult DSM
M 4600 P1.3 P5 P1.2 4600 P1.3 P5 P1.2
G450-1.90-60-40-15-1500 23.4 9.6 9.5 9.5 9.1 9.9 7.7 1.01 1.01 1.05 0.97 1.24
G450-1.90-60-40-15-2000 14.0 8.5 8.2 8.2 7.6 8.6 6.5 1.03 1.03 1.11 0.99 1.31
G450-1.90-60-40-15-2500 9.9 7.2 7.0 7.0 6.2 7.0 5.4 1.03 1.03 1.16 1.03 1.34
G450-1.90-60-40-15-3000 7.1 6.0 5.9 5.9 5.1 5.7 4.5 1.02 1.02 1.17 1.05 1.33
G450-1.90-60-40-15-3500 5.6 5.0 4.9 4.9 4.3 4.7 3.8 1.02 1.02 1.18 1.07 1.31
G450-1.90-75-50-15-1500 36.4 11.9 12.0 12.0 11.7 12.4 9.5 0.99 0.99 1.02 0.96 1.25
G450-1.90-75-50-15-2000 21.8 11.0 10.8 10.8 10.2 11.2 8.3 1.02 1.02 1.08 0.98 1.33
G450-1.90-75-50-15-2500 14.6 9.7 9.4 9.4 8.6 9.6 7.1 1.03 1.03 1.13 1.01 1.37
G450-1.90-75-50-15-3000 10.6 8.3 8.1 8.1 7.1 8.0 6.0 1.02 1.02 1.16 1.03 1.37
G450-1.90-75-50-15-3500 8.2 7.0 6.9 6.9 5.9 6.6 5.2 1.02 1.02 1.17 1.06 1.35
G450-1.90-90-60-15-2000 31.9 13.4 13.4 13.2 12.6 13.6 9.6 1.00 1.01 1.07 0.98 1.39
G450-1.90-90-60-15-2500 21.5 12.3 12.0 12.0 11.0 12.3 8.6 1.02 1.02 1.12 1.00 1.43
G450-1.90-90-60-15-3000 15.5 10.9 10.6 10.6 9.4 10.6 7.5 1.03 1.03 1.16 1.03 1.45
G450-1.90-90-60-15-3500 11.8 9.3 9.2 9.2 8.0 9.0 6.6 1.01 1.01 1.17 1.04 1.42
G450-1.90-90-60-15-4000 9.3 8.0 7.9 7.9 6.8 7.5 5.7 1.01 1.01 1.18 1.01 1.32
Mean 1.018 1.018 1.127 1.014 1.348
COV 0.012 0.011 0.046 0.034 0.044
Φ 0.938 0.938 1.023 0.928 1.228
_________________________________________________________________________________________________________________
9-34
EBL Ult DSM DSM
4600 P1.3 P5 P1.2 4600 P1.3 P5 P1.2
G550-0.95-40-25-7.5-1500 3.5 2.8 2.7 2.7 2.4 2.7 2.0 1.03 1.03 1.18 1.04 1.38
G550-0.95-40-25-7.5-2000 2.2 2.0 1.9 1.9 1.7 1.9 1.5 1.04 1.04 1.19 1.08 1.34
G550-0.95-40-25-7.5-2500 1.6 1.6 1.4 1.4 1.3 1.4 1.2 1.14 1.14 1.23 1.14 1.35
G550-0.95-55-35-7.5-1500 7.1 4.6 4.5 4.3 3.9 4.4 2.9 1.02 1.07 1.16 1.05 1.57
G550-0.95-55-35-7.5-2000 4.3 3.5 3.5 3.4 3.0 3.3 2.4 1.00 1.02 1.19 1.06 1.49
G550-0.95-55-35-7.5-2500 2.9 2.7 2.6 2.6 2.2 2.4 1.9 1.03 1.03 1.20 1.09 1.42
G550-0.95-55-35-7.5-3000 2.1 2.2 1.9 1.9 1.7 1.9 1.5 1.14 1.14 1.26 1.16 1.43
G550-0.95-70-45-7.5-2500 5.0 4.3 4.2 3.9 3.4 3.8 2.6 1.01 1.08 1.25 1.13 1.63
G550-0.95-70-45-7.5-3000 3.6 3.4 3.3 3.1 2.7 3.0 2.2 1.05 1.09 1.26 1.16 1.56
Mean 1.051 1.072 1.213 1.102 1.464
COV 0.049 0.042 0.031 0.042 0.071
Φ 0.952 0.976 1.111 1.003 1.300
_________________________________________________________________________________________________________________
9-35
Table 9.26: Summary of the Comparative Analyses of FEA Ultimate Loads with Code Predictions
Temp. Grade and DSM AS/NZS 4600 EC3 Part 1.3 BS 5950 Part5 EC3 Part 1.2
(oC) Thickness Mean COV Φ Mean COV Φ Mean COV Φ Mean COV Φ Mean COV Φ
G250-1.95 1.094 0.023 1.006 1.099 0.025 1.010 1.233 0.021 1.134 1.08 0.03 0.99 1.529 0.025 1.405
G450-1.90 1.153 0.099 1.001 1.168 0.086 1.028 1.289 0.048 1.171 1.11 0.08 1.02 1.417 0.032 1.298
20 G550-0.95 1.376 0.172 1.042 1.409 0.146 1.121 1.463 0.117 1.223 1.32 0.14 1.21 1.540 0.102 1.297
G250-1.95 1.096 0.025 1.007 1.101 0.026 1.011 1.235 0.019 1.137 1.08 0.03 0.99 1.524 0.029 1.398
G450-1.90 1.166 0.111 0.997 1.182 0.096 1.029 1.299 0.055 1.175 1.12 0.09 0.98 1.419 0.038 1.297
100 G550-0.95 1.37 0.173 1.035 1.402 0.147 1.114 1.458 0.117 1.219 1.31 0.14 1.06 1.537 0.102 1.314
G250-1.95 1.103 0.029 1.012 1.109 0.03 1.017 1.243 0.018 1.144 1.08 0.03 0.99 1.534 0.025 1.409
G450-1.90 1.18 0.123 0.993 1.197 0.108 1.028 1.31 0.063 1.178 1.13 0.10 0.98 1.418 0.047 1.290
200 G550-0.95 1.41 0.168 1.076 1.443 0.142 1.157 1.486 0.116 1.244 1.34 0.14 1.08 1.551 0.104 1.321
G250-1.95 1.098 0.024 1.009 1.101 0.026 1.011 1.235 0.029 1.133 1.09 0.03 1.00 1.542 0.029 1.414
G450-1.90 1.209 0.144 0.986 1.227 0.126 1.028 1.331 0.079 1.180 1.15 0.12 0.98 1.423 0.064 1.279
300 G550-0.95 1.42 0.171 1.077 1.453 0.144 1.161 1.487 0.121 1.235 1.34 0.14 1.07 1.542 0.111 1.301
G250-1.95 0.892 0.038 0.815 0.894 0.037 0.817 0.999 0.066 0.896 0.89 0.07 0.79 1.256 0.047 1.142
G450-1.90 1.172 0.129 0.978 1.189 0.113 1.014 1.299 0.068 1.163 1.12 0.11 0.97 1.403 0.051 1.272
400 G550-0.95 1.416 0.173 1.070 1.449 0.147 1.151 1.486 0.123 1.230 1.34 0.15 1.07 1.544 0.112 1.301
G250-1.95 0.941 0.024 0.865 0.942 0.024 0.866 1.049 0.06 0.946 0.94 0.06 0.85 1.327 0.046 1.207
G450-1.90 1.101 0.08 0.975 1.113 0.07 0.995 1.237 0.039 1.130 1.07 0.07 0.95 1.378 0.024 1.266
500 G550-0.95 1.34 0.165 1.029 1.372 0.14 1.104 1.434 0.108 1.215 1.29 0.13 1.05 1.518 0.093 1.314
G250-1.95 1.006 0.013 0.927 1.006 0.012 0.927 1.116 0.05 1.013 1.00 0.04 0.92 1.416 0.044 1.290
G450-1.90 1.042 0.024 0.958 1.042 0.024 0.958 1.143 0.031 1.048 1.04 0.02 0.96 1.383 0.044 1.260
600 G550-0.95 1.016 0.018 0.935 1.029 0.031 0.942 1.167 0.039 1.064 1.08 0.04 0.98 1.482 0.097 1.276
G250-1.95 1.011 0.012 0.932 1.012 0.012 0.933 1.122 0.05 1.018 1.01 0.04 0.92 1.424 0.044 1.297
G450-1.90 1.018 0.012 0.938 1.018 0.011 0.938 1.127 0.046 1.025 1.01 0.03 0.93 1.348 0.044 1.228
700 G550-0.95 1.051 0.049 0.952 1.072 0.042 0.976 1.213 0.031 1.111 1.10 0.04 1.00 1.464 0.071 1.300
_________________________________________________________________________________________________________________
9-36
9.3.1.1 AS/NZS 4600 and the Direct Strength Method
Comparative results showed that the DSM and AS/NZS 4600 predictions are more
accurate than the other design standards (see Tables 9.5 to 9.26). As seen in Table
9.26 the mean value of FEA ultimate loads to code predictions is closer to 1.0 in
most cases except for G250-1.95 steel columns at 400oC and 500oC and G550-0.95
steel columns. Since AS 4600 and DSM predictions for flexural-torsional buckling
are very close to each other, only AS/NZS 4600 was considered in further analyses
and graphical representation.
Figure 9.5 (a) to (c) show the comparison of all the FEA results with AS/NZS 4600
predictions. Sample calculations used in plotting these figures are given in Appendix
G. As seen in Figure 9.5 and Tables 9.5 to 9.26, the predictions for low strength steel
(G250-1.90) at 400oC and 500oC were slightly below the design curve. Therefore
these results are slightly unconservative (see Figure 9.5 a). The possible reason is the
change of stress-strain behaviour beyond 300oC for low strength steels. Up to 300oC,
low strength steels show a clear yielding point in their stress-strain curves. Beyond
this temperature, their stress-strain relationship does not show clear yielding. Instead
it includes a significant nonlinear behaviour before yielding. However the lowest Φ
factor for G250 steel columns (0.82) was not considerably below the accepted value
of 0.85.
AS/NZS 4600 predictions are conservative for higher strength steel columns, in
particular for G550-0.95 columns. The mean values of predictions were about 1.4 at
temperatures up to 500oC. However, it decreased suddenly to 1.029 at 600oC and
then increased to 1.072 at 700oC. This is due to the rapid relative variation between
yield stress and elastic modulus of G550-0.95 steel at 600oC (see Figure 6.7). Same
observation could be made in the case other design codes. Since the AS/NZS 4600
code predictions are conservative at lower temperatures, the relative variation in
mechanical properties at 600oC does not make the AS/NZS 4600 predictions
unconservative at these higher temperatures.
Considering the above discussion, it is recommended that the current design rules
based on AS/NZS 4600 and DSM can be used to predict the ultimate loads of cold-
formed steel columns subject to flexural-torsional buckling at ambient and elevated
_________________________________________________________________
9-37
temperatures by using the appropriately reduced mechanical properties of steels at
1.2
AS/NZS 4600
1
100
200
0.8
300
400
Pu/Ps
0.6
500
600
0.4
700
0.2
0
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5
0.5
(Py/Pcr)
(a) G250 Steel Columns
1.2
AS/NZS 4600
1
100
200
0.8
300
400
Pu/Ps
0.6
500
600
0.4
700
0.2
0
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5
0.5
(Py/Pcr)
(b) G450 Steel Columns

AS/NZS 4600 Design Curve at Elevated temperatures
_________________________________________________________________
9-38
1.2
AS/NZS 4600
1
100
200
0.8
300
400
Pu/Ps
0.6
500
600
0.4
700
0.2
0
0 0.5 1 1.5 2 2.5 3 3.5 4 4.5
0.5
(Py/Pcr)
(c) G550 Steel Columns
Figure 9.5: Comparison of FEA Ultimate Loads with AS/NZS 4600 Design
Curve at Elevated Temperature
The stress-strain curve of low strength steels has a significant nonlinear region than
in the high strength steels. Nonlinear behaviour of low strength steels at lower
elevated temperatures starts well before the yielding point (0.2% proof stress).
Figures 9.6 (a) to (d) compare the stress-strain curves of low and high strength steels
at 400oC, 500oC and 600oC. At 400oC more than 50% of the yield stress of G250-
1.95 is in the nonlinear region whereas that for G450-1.9 it is 23%. Usually flexural-
torsional buckling failures occur in the elastic region. Since low strength steel loses
their linear behaviour well below the 0.2% proof stress, failure occurs before it
reaches the 0.2% proof stress. Therefore the axial load carrying capacity is reduced
at 400oC for low strength steel columns. Further investigation was carried out
considering the stress-strain curves of low strength steel at 500oC and 600oC (see
Figures 9.5 (b) and (c)). It was found that the sizes of nonlinear regions below the
0.2% proof stress are 37% and 28% at 500oC and 600oC, respectively. The
associated mean values for G250-1.95 at these temperatures showed gradual increase
reflecting the improvement due to decreasing nonlinear region in the stress-strain
curves at these temperatures.
_________________________________________________________________
9-39
200
150
Stress (MPa)
53% of fy
100
50
0
0 0.0005 0.001 0.0015 0.002 0.0025 0.003 0.0035 0.004 0.0045 0.005
Strain
(a) G250-1.95 -400oC
100
37% of fy
75
Stress (MPa)
50
25
0
0 0.0005 0.001 0.0015 0.002 0.0025 0.003 0.0035 0.004 0.0045 0.005
Strain
(b) G250-1.95 -500oC
Figure 9.6: Comparison of Stress-strain Curves at Elevated Temperatures
_________________________________________________________________
9-40
60
28% of fy
Stress (MPa)
40
20
0
0 0.0005 0.001 0.0015 0.002 0.0025 0.003 0.0035 0.004 0.0045 0.005
Strain
(c) G250-1.95 -600oC
400
23% of fy
300
Stress (MPa)
200
100
0
0 0.001 0.002 0.003 0.004 0.005 0.006 0.007 0.008 0.009 0.01
Strain
(d) G450-1.90-400oC
Figure 9.6: Comparison of Stress-strain Curves at Elevated Temperatures
9.3.1.2 BS5950 Part 5
The modified BS5950 Part 5 also provides reasonably accurate predictions. The
comments made in relation to AS/NZS 4600 also apply to BS 5950 Part 5. However,
_________________________________________________________________
9-41
its predictions are slightly less than that of AS/NZS 4600. Its predictions are slightly
unconservative for low strength steels at 400oC and 500oC. The possible reason is
the significant nonlinearity in the stress-strain curve of lower strength steels as
explained in the previous section.
Eurocode 3 Part 1.3 predictions are conservative than the Australian and British
design standards. As seen in the comparison for the Australian and British design
standards, higher strength steels gave conservative results than lower strength steels.
Eurocode 3 Part 1.3 recommends buckling curve “b” for channel sections. However,
comparison of results showed that the buckling curve “b” produces more
conservative results than the buckling curve “ao” and “a”, particularly for high
strength steels (see Figures 9.7 a to c). Sample calculations for plotting these figures
are shown in Appendix G. In the case of low strength steels at 400oC and 500oC,
results are unconservative even for predictions with buckling curve “b” due to
significant nonlinearity in the stress-strain curve before yielding. For all the other
cases buckling curve “a” provides conservative predictions while buckling curve
“ao” produce more accurate results. However, some results are between these two
buckling curves. Considering the distribution of results, buckling curves given in
Tables 9.27 and 9.28 are recommended for lipped channel columns for fire design
based on the steel grade and temperatures.
Table 9.27: Proposed Buckling Curves from Eurocode 3 Part 1.3 for the Fire
Design of Low Strength Steel columns
Temperature oC 20 ≤ T ≤ 300 300 < T ≤ 500 500 < T ≤ 700

Buckling curve ao c b
Table 9.28: Proposed Buckling Curves from Eurocode 3 Part 1.3 for the Fire
Design of High Strength Steel columns
Temperature oC 20 ≤ T ≤ 500 500 < T ≤ 700

Buckling curve ao a
_________________________________________________________________
9-42
1.2
Buckling Curve ao
1 Buckling Curve a
Buckling Curve b
0.8 Buckling Curve c
100
200
Pu/Ps
0.6
300
400
0.4 500
600
0.2 700
0
0 0.5 1 1.5 2 2.5 3 3.5 4
(a) G250 Steel Columns
1.2
Buckling Curve ao
1
Buckling Curve a
Buckling Curve b
0.8 Buckling Curve c
100
Pu/Ps
0.6 200
300
400
0.4
500
600
0.2 700
0
0 0.5 1 1.5 2 2.5 3 3.5 4
(b) G450 Steel Columns
Figure 9.7: Comparison of FEA Ultimate loads with Eurocode 3 Part 1.3
Buckling Curves
_________________________________________________________________
9-43
1.2
Buckling Curve ao
1 Buckling Curve a
Buckling Curve b
Buckling Curve c
0.8
100
200
Pu/Py
0.6 300
400
0.4 500
600
0.2 700
0
0 0.5 1 1.5 2 2.5 3 3.5 4
(c) G550 Steel Columns
Figure 9.7: Comparison of FEA Ultimate loads with Eurocode 3 Part 1.3
Buckling Curves
Eurocode 3 Part 1.2 is the only available elevated temperature design standard which
provides design rules for fire design of cold-formed steel members. Its predictions
are overly conservative because imperfection factors based on ambient temperature
yield stress are higher than that of buckling curve “b” recommended by Eurocode 3
Part 1.3. Even though Eurocode 3 Part 1.2 limits the maximum design temperature to
350oC, parametric study results showed that cold-formed steel columns possess
significant load carrying capacity beyond 350oC. It seems that Eurocode 3 Part 1.2 is
a modified version of Eurocode 3 Part 1.3 with some limitations. The introduced
modifications and limitation make it uneconomical at any temperatures. Therefore
considering the results of individual cases and the summary reported in Tables 9.1 to
9.3 and 9.5 to 9.26, Eurocode 3 Part 1.2 is not recommended for the fire design of
cold-formed steel compression members since it provides overly conservative
capacity predictions.
_________________________________________________________________
9-44
9.4 Conclusion
Flexural-torsional buckling behaviour of cold-formed steel compression members

was investigated using 39 ambient temperature and 273 elevated temperature finite
element simulations. The ultimate load results were compared with available cold-
formed steel design standards at ambient and elevated temperatures. From the
comparative study of FEA ultimate load results and the code predictions, following
conclusive remarks were drawn.
• Flexural-torsional buckling behaviour and strength was influenced by the

level of nonlinearity in the stress-strain curves. At elevated temperatures,
stress-strain curves become more non-linear. In the case of elevated
temperature capacity predictions, all the design codes provided reasonably
accurate predictions except for low strength steel columns at 400 and 500oC.
This is caused by their significant nonlinear stress-strain behaviour before the
yield point.
• Eurocode 3 Part 1.2 predictions were very conservative. Among the design
standards considered in this research, elevated temperature predictions of
AS/NZS 4600 and the DSM showed a good agreement and a better
consistency. In this research, some of the columns experienced higher
ultimate loads than the elastic buckling loads, particularly for long columns
of higher strength steels. Since AS/NZS 4600 and DSM limit the column
capacity to 87.7% of the elastic buckling load, their predictions for long high
strength steel columns are slightly more conservative. Even though their
capacity reduction factors ( ) for low strength steel columns at 400 and
500oC are small (0.82), it is not considerably below the recommended limit
of 0.85. Therefore AS/NZS 4600 and DSM can be used to predict the
flexural-torsional buckling capacity of cold-formed steel compression
members at elevated temperatures with the use of appropriately reduced
mechanical properties.
• Eurocode 3 Part 1.3 design method was found to be the most conservative
among the ambient temperature design standards. This is mainly due to the
use of buckling curve “b”. Eurocode 3 Part 1.3 predictions with buckling
_________________________________________________________________
9-45
curve “ao” appear to be more accurate at ambient temperature. In the case of
elevated temperatures, bucking curves given in Tables 9.27 and 9.26 can be
used for elevated temperature capacity estimations.
• The predictions of BS 5950 Part 5 with its current equation for the elastic
torsional buckling load is unconservative and can not be recommended. The
predictions of modified BS 5950 Part 5 design method were found to be more
accurate.
• Eurocode 3 Part 1.2 design method cannot be recommended for fire design as
it predicts overly conservative capacities. Since it limits maximum
temperature to 350oC, its design method can not be used to severe fire
situation. Further, analyses results shoed that there are considerable
capacities remain beyond 350oC.
• Although this research is limited to flexural-torsional buckling, it is

considered that all the findings are likely to be equally applicable to flexural
buckling.
_________________________________________________________________
9-46
10 Conclusion and Recommendations
This thesis has described a detailed investigation on the local and flexural-torsional
buckling behaviour of light gauge cold-formed steel compression members at
ambient and elevated temperatures. Both experimental and finite element analyses
were undertaken to improve the knowledge and understanding of the local and
flexural-torsional buckling behaviour of cold-formed steel compression members
subjected to simulated fire conditions, and hence this has led to safer structural
design methods for fire conditions.
A derailed review on essential mechanical properties was first carried out on the
light gauge cold-formed steels to find the reduced mechanical properties and the
stress-strain relationship at elevated temperatures (Chapter 3). The most accurate
stress-strain models found in the literature were then validated using a series of
tensile coupon tests carried out at ambient and elevated temperatures so that they can
be used in the validation of finite element models, parametric study and numerical
investigations.
In the second phase of this research, a series of local buckling tests was carried out
on lipped and unlipped channels at ambient and elevated temperatures (Chapter 4).
Three thicknesses with low and high strength cold-formed steel stub columns were
chosen in the experimental investigation on local buckling behaviour. Temperature
range was selected up to 700oC at every 100oC intervals from 200oC in steady state
tests undertaken in a small electric furnace. Fixed-end condition was adopted for all
the tests.
In the third phase of this research, finite element models were developed to simulate
the local buckling behaviour using the ABAQUS finite element program (Chapter 5).
Developed finite element models were validated by comparing the load-deflection
curves, ultimate loads and deflected shapes from the tests and the same from the
finite element analyses at both ambient and elevated temperatures. Ideal models were
then developed by appropriately changing the geometry and the boundary
conditions. A detailed parametric study was carried out using the ideal models to
improve the knowledge and understanding of the local buckling behaviour and to
_________________________________________________________________
Behaviour and design of Cold-formed Steel Compression Members at Elevated Temperatures
10-1
identify the effects of geometric parameters on the local buckling behaviour at
ambient and elevated temperatures (Chapter 6). In the fourth phase of this research,
two series of flexural-torsional buckling test were carried out on cold-formed steel
lipped channel columns at ambient and elevated temperatures (Chapter 7). Both low
and high strength grades of relatively thicker light gauge steels were selected for
flexural-torsional buckling tests. A three segment electrical furnace was used to heat
the columns. The first test series contained 2800 mm long columns which required
the three segment furnace whereas the second series contained 1800 mm columns
which required the two segment furnace. Fixed end-condition was used in all the
tests. Steady state elevated temperature tests were carried out up to 700oC.
In the final phase of this research, finite element models were developed to simulate
the flexural-torsional buckling behaviour of cold-formed steel columns (Chapter 8).
Similar to local buckling investigation, finite element models were first validated
using the load-deflection curves, ultimate load and the deflected shape of the failed
specimens. A detailed parametric study was carried out using the validated finite
element models (Chapter 9) for a number of section sizes and a range of column
lengths in order to expand the range of parameters.
The most valuable outcomes obtained from this research are as follows:
• Significantly improved knowledge and understanding of the local and

flexural-torsional buckling behaviour of light gauge cold-formed steel
compression members at both ambient and elevated temperatures
• Assessment of the currently available elevated temperature design method for

local and flexural-torsional buckling behaviour of cold-formed steel
compression members.
• Assessment and recommendations on the use of current ambient temperature

design methods for elevated temperature design with reduced mechanical
properties.
• Effect of stress-strain relationship on cold-formed steel column capacity.
_________________________________________________________________
10-2
This research has shown that in general the use of current ambient temperature
design methods with reduced mechanical properties is suitable for the design of cold-
formed steel compression members at elevated temperatures then the current fire
design methods. Following important conclusions and recommendations have been
drawn based on specific topic investigated in this research project.
10.1 Mechanical Properties of Cold-formed Steels at Elevated

Temperatures
• Tensile coupon tests were carried out successfully at ambient and elevated
temperatures. Elevated temperature strains were measured accurately using
the latest non-contact strain measurement technique based on laser speckle
extensometer. Accurate stress-strain relationships were obtained up to 700oC.
• Yield stress of high strength steels does not reduce much until 300oC.
Thereafter, it is reduced at a higher rate up to 600oC beyond which it is
reduced at a slower rate. However, at 600oC, it has lost almost 90% of
ambient temperature strength. Unlike yield strength, elastic modulus showed
a linear reduction pattern. The ductility of cold-formed steels was noted to
increase with increasing temperature with significant increases beyond
500oC. However, the lowest ductility was observed at 200oC. A significant
increase in the ductility was seen for the temperatures beyond 500oC.
• The formulae proposed by Ranawaka and Mahendran (2009) and Dolamune

Kankanamge (2009) accurately predict the mechanical properties of thin
cold-formed steels at elevated temperatures.
• Modified Ramberg Osgood stress-strain model proposed by Ranawaka and

Mahendran (2008) and Dolamune Kankanamge (2009) can be used to obtain
accurate stress-strain relationships of thin cold-formed steels at elevated
temperatures.
_________________________________________________________________
10-3
10.2 Local Buckling Behaviour at Ambient and Elevated
Temperatures
• A total of 27 ambient temperature tests and 64 elevated temperature tests
were carried out successfully to investigate the local buckling behaviour of
short cold-formed steel columns.
• Developed experimental finite element models accurately predicted the local

buckling behaviour of tested stub columns with pin-end conditions.
• Ideal finite element models with pin-end conditions were developed to carry
out a detailed parametric study. Analyses showed that the length of ideal
model of the lipped channels has an effect on the elastic buckling and
ultimate loads. Even though a quarter wave length model can ideally
represent the local buckling phenomenon, lipped channel columns can not be
accurately modelled with quarter wave length models. However, it can be
used for unlipped channels. This research has shown that a model that
represents a column of three half waves sufficiently model the local buckling
behaviour of lipped channel columns.
• A total of 44 ambient and 252 elevated temperature simulations were

undertaken to investigate the local buckling behaviour of cold-formed steel
compression members using a detailed parametric study. Code predictions for
all the specimens used in the parametric study and the corresponding ratios of
finite element analysis results to code predictions were calculated as a part of
the assessment on the accuracy of current design rules.
• Increasing web slenderness led to higher ultimate loads of lipped channel

sections, which decreased after reaching a maximum limit. The current
design rules based on the effective width method (AS/NZS 4600, Eurocode 3
Part 1.3, Eurocode 3 Part 1.2, North American Standards (NAS) and BS5950
Part 5) was not able to predict this trend. However, this occurred at higher
web width to thickness ratios that are not commonly used in compression
members. The direct strength method (DSM) was able to follow the variation
of ultimate load for lipped channels with increasing web slenderness.
_________________________________________________________________
10-4
• Effective width method predictions for unlipped channels are more
conservative for lower temperatures. However, it is slightly unconservative at
higher temperatures for narrow sections with smaller flanges. Similar pattern
can be observed with DSM predictions. But it is unconservative for narrow
sections at higher temperatures.
• Mechanical properties deteriorate at different rates at elevated temperatures,

that is the ratio of elastic modulus to yield stress (E/fy) changes with
increasing temperature. It appears that E/fy ratio has an effect on the local
buckling capacity. Effective width method can cope with different E/fy ratios
while DSM appears to be struggling to cope with it.
• Eurocode 3 Part 1.2 predictions for lipped and unlipped channels sections are
of the same order as for AS/NZS 4600 predictions at lower temperatures. But
it is more conservative at higher temperatures as it uses ambient temperature
effective area at higher temperatures. Eurocode 3 Part 1.2 limitation on the
maximum temperature of 350oC is not economical. Test and finite element
analyses showed the significant axial compression capacity was still available
beyond 350oC.
• The predictions of BS5950 Part 5 are conservative for unlipped channels. Its
predictions for lipped channels are more accurate than AS/NZS 4600
predictions while its predictions for unlipped channels are conservative than
AS/NZS 4600 predictions.
10.3 Flexural-torsional Buckling Behaviour at Ambient and

• A total of 39 flexural-torsional buckling tests of 1800 and 2800 long columns
were successfully carried out at ambient and elevated temperatures.
• Test results showed closer agreements with the code predictions particularly
with AS/NZS 4600, modified BS 5950 Part 5 and DSM. Predictions for high
strength steel columns were slightly conservative than those for low strength
steel columns. The capacity reduction factors are also greater than 0.85.
Therefore AS/NZS 4600, modified BS 5950 Part 5 and DSM can be used to
_________________________________________________________________
10-5
predict the flexural-torsional buckling capacities at elevated temperatures.
Further, Eurocode 3 Part 1.3 and Part 1.2 predictions are more conservative
than the Australian/New Zealand and British standards.
• Developed finite element models accurately simulated the flexural-torsional

buckling behaviour of slender cold-formed steel compression members.
• A total of 39 ambient and 273 elevated temperature finite element

simulations were undertaken to investigate the flexural-torsional buckling
behaviour. Code predictions for all the specimens used in the parametric
study and the corresponding ratios of finite element analysis results to code
predictions were calculated as a part of the assessment on the accuracy of
current design rules.
• The stress-strain relationship of low strength steels show significant

nonlinearity before yield point at 400oC and was observed to affect the
ultimate capacity of columns. Hence the code predictions for low strength
steels are slightly unconservative at 400oC.
• AS/NZS 4600 and the DSM showed a good agreement and a better
consistency when compared with parametric study results. In this research,
some of the columns experienced higher ultimate loads than the elastic
buckling loads, particularly for the long columns made of higher strength
steels. Since AS/NZS 4600 and DSM limit the column capacity to 87.8% of
the elastic buckling load, their predictions for high strength steel long
columns are more conservative. Although their capacity reduction factors (F)
for low strength steel columns at 400 and 500oC are small (0.82), it is not
considerably below the recommended limit of 0.85. Therefore AS/NZS 4600
and DSM can be used to predict the flexural-torsional buckling capacity of
cold-formed steel compression members at elevated temperatures by using
the appropriately reduced mechanical properties.
• Eurocode 3 Part 1.3 design method was found to be the most conservative
design method among the ambient temperature design methods. Use of
buckling curve “b” makes it more conservative. Eurocode 3 Part 1.3
predictions with buckling curve “ao” appear to be more accurate at ambient
_________________________________________________________________
10-6
temperature. In the case of elevated temperatures, bucking curves given in
Tables 9.26 and 9.27 can be used for elevated temperature capacity
estimations.
• The predictions of BS 5950 Part 5 with its current formulae for flexural-
torsional buckling are unconservative and can not be recommended. The
equation for torsional buckling of BS5950 Part 5 was modified as follows so
that it is identical to that in other design standards. The predictions of
modified BS 5950 Part 5 method were found to be more accurate.
1 ⎛ π 2 EC w ⎞
PT = ⎜⎜ GJ + ⎟
ro2 ⎝ L2E ⎟⎠
• Eurocode 3 Part 1.2 design method can not be recommended for fire design
as it predicts overly conservative capacities. Eurocode 3 Part 1.2 limitation
on the maximum temperature of 350oC is not economical.
• Although this research is limited to flexural-torsional buckling, it is

considered that all the findings are likely to be equally applicable to flexural
buckling.
10.4 Future Research

To date, detailed research has been undertaken on the local, distortional and flexural-
torsional buckling behaviour of light-gauge cold-formed steel members at elevated
temperatures. However, interaction of the above buckling modes has not been
considered yet. Therefore investigation on the buckling interaction would complete
the knowledge and understanding on the buckling behaviour of light gauge cold-
formed steel compression at elevated temperatures.
Usually mechanical properties are obtained from the tensile coupon tests although
some researchers undertook compression coupon tests at ambient temperature. It was
observed that the ratio of test to finite element analyses increases with increasing
temperature. This may be due to the differences between the stress-strain
relationships for compression and tension at elevated temperature. Therefore
_________________________________________________________________
10-7
determination of mechanical properties from compression coupon tests at elevated
temperatures is recommended.
It was found that the significant nonlinearity in the stress-strain curve gave unsafe
results. Therefore, further research on the effect of nonlinearity in the stress-strain
curves is recommended.
This research was limited to lipped and unlipped channel sections. Further research
is needed to investigate the applicability of some of the important observation to
other cold-formed steel sections.
This research was limited to flexural-torsional buckling. Therefore a further research

is recommended for flexural buckling even though the same guidelines are
applicable.
_________________________________________________________________
10-8
Appendix A
Stress-strain Curves Obtained for G550-0.42 Specimens at Ambient and
800
600
400
Stress (MPa)
Laser Speckle Extensometer
Strain Gauge
200
0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5
Strain (%)
G550-0.42-20-1
800
600
400
Stress (MPa)
Laser Speckle Extensometer
Strain Gauge
200
0
0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5
Strain (%)
G550-0.42-20-2
_________________________________________________________________
A-1
800
600
400
Stress (MPa)
200
0
-0.5 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5
Strain (%)
G550-0.42-100-1
800
600
400
Stress (MPa)
200
0
-0.5 0.5 1.5 2.5 3.5 4.5 5.5
Strain (%)
G550-0.42-100-2
_________________________________________________________________
A-2
800
600
400
Stress (MPa)
200
0
-0.2 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4
Strain (%)
G550-0.42-200-1
800
600
400
Stress (MPa)
200
0
-1.0 0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10.0 11.0 12.0
Strain (%)
G550-0.42-300-1
_________________________________________________________________
A-3
600
400
Stress (MPa)
200
0
-1.0 0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10.0 11.0 12.0
Strain (%)
G550-0.42-400-1
600
400
Stress (MPa)
200
0
-1.0 0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10.0 11.0 12.0
Strain (%)
G550-0.42-400-2
_________________________________________________________________
A-4
400
200
Stress (MPa)
0
-1.0 0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10.0 11.0 12.0
Strain (%)
G550-0.42-500-1
400
200
Stress (MPa)
0
-1.0 0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0 9.0 10.0 11.0 12.0
Strain (%)
G550-0.42-500-2
_________________________________________________________________
A-5
100
80
60
Stress (MPa)
40
20
0
0.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0
Strain (%)
G550-0.42-600-1
100
80
60
Stress (MPa)
40
20
0
0.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0
Strain (%)
G550-0.42-600-2
_________________________________________________________________
A-6
60
40
Stress (MPa)
20
0
0.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0 40.0
Strain (%)
G550-0.42-700-1
40
20
Stress (MPa)
0
0.0 5.0 10.0 15.0 20.0 25.0 30.0
Strain (%)
G550-0.42-700-2
_________________________________________________________________
A-7
Appendix B
Load-Deflection Curves of Local Buckling Tests
20
20
15
15
Axial Load (kN)
10
Axial Load (kN)

10
Cross-head of Tinius
5 LVDT
5
0
0
0.00 0.50 1.00 -0.40 -0.35 -0.30 -0.25 -0.20 -0.15 -0.10 -0.05 0.00 0.05 0.10
Axial Shortening (mm) Out-of-Plan Deflection (mm)
G250-0.95-B-20-1
20
20
15
Axial Load (kN)
15
10
10
Axial Load (kN)

5 LVDT
5
0
0
0.00 0.50 1.00 -0.10 0.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70
G250-0.95-B-20-2
_________________________________________________________________________________________________________________
B-1
20 20
15
15
Axial Load (kN)
Axial Load (kN)

10
10
5
5
0
0
0.0 0.2 0.4 0.6 0.8 1.0 1.2
-0.70 -0.60 -0.50 -0.40 -0.30 -0.20 -0.10 0.00 0.10
Axial Shortening (mm) Out-of-Plane Deflection (mm)
G250-0.95-B-20-3
20
20
15
15
Axial Load (kN)
Axial Load (kN)

10
10
5 LVDT 5
0 0
0.00 0.50 1.00 1.50 -0.20 -0.10 0.00 0.10 0.20 0.30 0.40 0.50
G250-0.95-B-200-1
_________________________________________________________________________________________________________________
B-2
20
20
15
15
Axial Load (kN)

Axial Load (kN)
10
10
X head of Tinius
5 LVDT 5
0 0
-0.10 0.40 0.90 -0.10 -0.08 -0.06 -0.04 -0.02 0.00
G250-0.95-B-200-2
15
15
10
10
Axial Load (kN)
Axial Load (kN)

5 5
0 0
0.00 0.50 1.00 1.50 2.00 -0.50 -0.40 -0.30 -0.20 -0.10 0.00
G250-0.95-B-300-1
_________________________________________________________________________________________________________________
B-3
10 10
Axial Load (kN)
Axial Load (kN)

5 5
LVDT
0 0
0.00 0.50 1.00 1.50 -0.80 -0.70 -0.60 -0.50 -0.40 -0.30 -0.20 -0.10 0.00 0.10
G250-0.95-B-300-2
10
10
Axial Load (kN)
Axial Load (kN)

5 5
X head of Tinius
LVDT
0 0
0.00 0.50 1.00 1.50 2.00 -0.20 -0.10 0.00 0.10 0.20 0.30 0.40 0.50
G250-0.95-B-300-3
_________________________________________________________________________________________________________________
B-4
10
10
Axial Load (kN)

Axial Load (kN)
5 5
X head of Tinius
LVDT
0 0
0.00 0.50 1.00 -0.02 0.03 0.08 0.13 0.18
G250-0.95-B-400-1
10
10
Axial Load (kN)
Axial Load (kN)

5
5
X head of Tinius
LVDT
0 0
0.00 0.50 1.00 1.50 2.00 -1.20 -1.00 -0.80 -0.60 -0.40 -0.20 0.00
G250-0.95-B-400-2
_________________________________________________________________________________________________________________
B-5
10
10
Axial Load (kN)
Axial Load (kN)

5
0 0
0.00 0.50 -0.70 -0.60 -0.50 -0.40 -0.30 -0.20 -0.10 0.00 0.10
G250-0.95-B-500-1
5 5
4 4
Axial Load (kN)
3 3
Axial Load (kN)

2 2
1 1
0 0
0.00 0.50 1.00 1.50 -3.00 -2.50 -2.00 -1.50 -1.00 -0.50 0.00
G250-0.95-B-600-1
_________________________________________________________________________________________________________________
B-6
3 3
2 2
Axial Load (kN)
Axial Load (kN)

1 1
0 0
-0.10 0.40 0.90 1.40 -3.50 -3.00 -2.50 -2.00 -1.50 -1.00 -0.50 0.00
G250-0.95-B-700-1
30
30
25
25
20
Axial Load (kN)
20
Axial Load (kN)

15
15
10 Cross-head of Tinius 10
LVDT
5
5
0 0
0.00 0.50 1.00 1.50 -0.80 -0.70 -0.60 -0.50 -0.40 -0.30 -0.20 -0.10 0.00 0.10
G550-0.95-B-20-1
_________________________________________________________________________________________________________________
B-7
30
30
25
25
20
20
Axial Load (kN)

Axial Load (kN)
15 15
10 10
LVDT
5 5
0 0
0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 -1.60 -1.40 -1.20 -1.00 -0.80 -0.60 -0.40 -0.20 0.00 0.20
G550-0.95-B-20-2
30
30
25 25
20 20
Axial Load (kN)
Axial Load (kN)

15 15
LVDT
5 5
0 0
-0.50 0.00 0.50 1.00 1.50 2.00 2.50 -0.05 -0.03 -0.01 0.01 0.03 0.05 0.07 0.09 0.11 0.13 0.15
G550-0.95-B-20-3
_________________________________________________________________________________________________________________
B-8
30
30
25
25
20
20
Axial Load (kN)

Axial Load (kN)
15
15
10
10
5 5
0 0
0.00 0.50 1.00 1.50 2.00 -0.10 0.00 0.10 0.20 0.30 0.40 0.50
G550-0.95-B-200-1
30
30
25
25
20
Axial Load (kN)
20
15
15
Axial Load (kN)

10
10
5
5
0 0
0.00 0.50 1.00 1.50 2.00 -0.60 -0.50 -0.40 -0.30 -0.20 -0.10 0.00 0.10
G550-0.95-B-200-2
_________________________________________________________________________________________________________________
B-9
30 30
25 25
Axial Load (kN) 20 20
Axial Load (kN)

15 15
LVDT
5 5
0 0
0.00 0.50 1.00 1.50 2.00 2.50 -3.50 -3.00 -2.50 -2.00 -1.50 -1.00 -0.50 0.00
G550-0.95-B-300-1
25
20
Axial Load (kN)
15
10
LVDT
5
0
0.00 0.50 1.00 1.50 2.00
G550-0.95-B-400-1
_________________________________________________________________________________________________________________
B-10
15
15
10
10
Axial Load (kN)
Axial Load (kN)

5 5
LVDT
0 0
0.00 0.50 1.00 1.50 2.00 -2.00 -1.50 -1.00 -0.50 0.00
G550-0.95-B-500-1
5
4
Axial Load (kN)
Axial Load (kN)

2
Cross-head of Tinius 2
LVDT
1
1
0
0
-0.10 0.30 0.70 1.10 -2.00 -1.50 -1.00 -0.50 0.00
G550-0.95-B-600-1
_________________________________________________________________________________________________________________
B-11
2.5
2.0
1.8
2.0
1.6
1.4
Axial Load (kN)
1.5
1.2
Axial Load (kN)

1.0
1.0
0.8
0.6
0.5
0.4
0.2
0.0
0.0
-0.10 0.30 0.70 1.10 -2.00 -1.50 -1.00 -0.50 0.00
G550-0.95-B-700-1
_________________________________________________________________________________________________________________
B-12
80
80
70
70
60
60
50
Axial Load (kN)
50
Axial Load (kN)

40
40
30
30
20
20
10
10
0
0
0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 -2.00 -1.50 -1.00 -0.50 0.00 0.50
G250-1.95-B-20-1
70
70
60
60
50
50
Axial Load (kN)
Axial Load (kN)

40
40
30
30
20
20
10
10
0
0
0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60
-1.00 -0.80 -0.60 -0.40 -0.20 0.00
G250-1.95-B-20-2
_________________________________________________________________________________________________________________
B-13
80
80
70
70
60
60
50
Axial Load (kN)
50
Axial Load (kN)

40
40
30
30
20
20
10
10
0
0
0.00 0.20 0.40 0.60 0.80 1.00 1.20 -0.01 0.01 0.03 0.05 0.07 0.09
G250-1.95-B-20-3
45
45
40
40
35
35
30
30
Axial Load (kN)
25
25
Axial Load (kN)

20
20
15
15
10
10
5
5
0
0
0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50 4.00 4.50 -0.10 0.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80 0.90
G250-1.95-B-300-1
_________________________________________________________________________________________________________________
B-14
40 40
35 35
30 30
25 25
Axial Load (kN)
Axial Load (kN)

20 20
15 15
10 10
5 5
0 0
0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00 9.00 10.00 -0.20 0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40
G250-1.95-B-400-1
40
40
35
35
30
30
25 25
Axial Load (kN)
Axial Load (kN)

20 20
15 15
10 10
5 5
0 0
0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50 4.00 -0.50 0.00 0.50 1.00 1.50 2.00
G250-1.95-B-400-2
_________________________________________________________________________________________________________________
B-15
25
25
20
20
Axial Load (kN)

15
15
Axial Load (kN)

10
10
5
5
0
0
0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80
-0.20 0.00 0.20 0.40 0.60 0.80 1.00 1.20
G250-1.95-B-500-1
25
20
Axial Load (kN)
15
10
0
-0.50 0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50
G250-1.95-B-500-2
_________________________________________________________________________________________________________________
B-16
16
16
14
14
12
12
10
10
Axial Load (kN)
Axial Load (kN)

8
8
6
6
4 4
2 2
0 0
-0.20 0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80 -0.10 0.00 0.10 0.20 0.30 0.40 0.50 0.60
G250-1.95-B-600-1
10
7
Axial Load (kN)
0
-0.20 0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80
G250-1.95-B-700-1
_________________________________________________________________________________________________________________
B-17
100
100
90
90
80
80
70
Axial Load (kN) 70
60
60
50
50
Axial Load (kN)

40
40
30
30
20
20
10
10
0
0
0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80 2.00 -0.20 -0.10 0.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70
G450-1.90-B-20-1
100
100
90
90
80
80
70
70
60
Axial Load (kN)
60
50
50
Axial Load (kN)

40
40
30 30
20 20
10 10
0 0
0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80 -0.10 0.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80 0.90
G450-1.90-B-20-2
_________________________________________________________________________________________________________________
B-18
100
90
80
70
Axial Load (kN)

60
50
40
30
20
10
0
0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60
G450-1.90-B-200-1
100 100
90 90
80 80
70 70
Axial Load (kN)
60 60
50 50
Axial Load (kN)

40 40
30 30
20 20
10 10
0 0
0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80 -1.20 -1.00 -0.80 -0.60 -0.40 -0.20 0.00 0.20
G450-1.90-B-300-1
_________________________________________________________________________________________________________________
B-19
70
70
60
60
50
50
Axial Load (kN)

40
Axial Load (kN)
40
30
30
20
20
10
10
0 0
-0.50 0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50 -0.50 0.00 0.50 1.00 1.50 2.00 2.50
Axial Shortening (mm) Axial Shortening (mm)
G450-1.90-B-400-1 G450-1.90-B-400-2
35 45
40
30
35
25
30
Axial Load (kN)
Axial Load (kN)

20
25
15 20
15
10
10
5
5
0 0
-0.50 0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50 -0.50 0.00 0.50 1.00 1.50 2.00 2.50 3.00
G450-1.90-B-500-1 G450-1.90-B-500-2
_________________________________________________________________________________________________________________
B-20
9
16 8
14 7
12 6
Axial Load (kN)

10 5
Axial Load (kN)
8 4
6 3
4 2
2 1
0 0
-0.50 0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50 4.00 4.50 -0.50 0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50
G450-1.90-B-600-1 G450-1.90-B-600-1
60
50
40
Axial Load (kN)
30
20
10
0
0.00 0.40 0.80 1.20
G550-0.95-A-20-1
_________________________________________________________________________________________________________________
B-21
60
60
50
50
40
40
Axial Load (kN)
Axial Load (kN)

30
30
20 20
10 10
0 0
0.00 0.40 0.80 1.20 -0.10 0.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70 0.80 0.90
G550-0.95-A-20-2
60
60
50
50
40
40
Axial Load (kN)
Axial Load (kN)

30
30
20
20
10 10
0 0
0.00 0.40 0.80 1.20 -2.00 -1.80 -1.60 -1.40 -1.20 -1.00 -0.80 -0.60 -0.40 -0.20 0.00 0.20
G550-0.95-A-20-3
_________________________________________________________________________________________________________________
B-22
55
50
50
45
45
40
40
35
35
Axial Load (kN)

Axial Load (kN)
30
30
25
25
20 20
Cross-head of Tinius Cross-head of Tinius
15 15
LVDT LVDT
10 10
5 5
0 0
0.00 0.50 1.00 1.50 2.00 0.00 0.50 1.00 1.50 2.00
G550-0.95-A-100-1 G550-0.95-A-200-1
50
55
45 50
40 45
35 40
Axial Load (kN)
35
Axial Load (kN)

30
30
25
25
20
20
15 Cross-head of Tinius Cross-head of Tinius
15
LVDT LVDT
10
10
5 5
0 0
-0.10 0.40 0.90 1.40 1.90 -0.10 0.40 0.90 1.40 1.90
G550-0.95-A-300-1 G550-0.95-A-300-2
_________________________________________________________________________________________________________________
B-23
40
40
35
35
30
30
25
Axial Load (kN)
25
Axial Load (kN)

20
20
15 15
10 10
LVDT LVDT
5 5
0 0
-0.10 0.40 0.90 1.40 1.90 -0.10 0.40 0.90 1.40 1.90 2.40
G550-0.95-A-400-1 G550-0.95-A-400-2
25
10
20
Axial Load (kN)
15
Axial Load (kN)

5
10
LVDT
5
LVDT
0 0
-0.10 0.40 0.90 1.40 1.90 2.40 -0.10 0.40 0.90 1.40 1.90 2.40
G550-0.95-A-500-1 G550-0.95-A-600-1
_________________________________________________________________________________________________________________
B-24
10
40
35
30
Axial Load (kN)
25
Axial Load (kN)

5
20
15 LVDT
LVDT 10
0 0
-0.10 0.40 0.90 1.40 1.90 2.40 0.00 0.50 1.00 1.50
G550-0.95-A-700-1 G250-0.95-20-1
40
40
35
35
30
30
Axial Load (kN)

25
25
Axial Load (kN)
20
20
15
15
10 10
5 5
0 0
0.00 0.50 1.00 -0.60 -0.50 -0.40 -0.30 -0.20 -0.10 0.00 0.10
G250-0.95-A-20-2
_________________________________________________________________________________________________________________
B-25
40
40
35
35
30
30
25
Axial Load (kN)

25
Axial Load (kN)

20
20
15 Cross-head of Tinius
15 LVDT
10
10
5 5
0 0
0.00 0.50 1.00 0.00 0.50 1.00 1.50
G250-0.95-A-20-3 G250-0.95-A-200-1
40
35
35
30
30
25
25
Axial Load (kN)
Axial Load (kN)

20
20
Cross-head of Tinius 15 Cross-head of Tinius

15 LVDT LVDT
10
10
5 5
0 0
0.00 0.50 1.00 1.50 -0.1 0.4 0.9 1.4
G250-0.95-A-200-2 G250-0.95-A-300-1
_________________________________________________________________________________________________________________
B-26
30
25
20
Axial Load (kN)

15
LVDT
10
0
-0.10 0.40 0.90 1.40
G250-0.95-A-300-2
35
35
30
30
25
25
Axial Load (kN)
20
20
Axial Load (kN)

LVDT
10 10
5 5
0 0
-0.10 0.40 0.90 1.40 0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00
G250-0.95-A-300-3
_________________________________________________________________________________________________________________
B-27
30
25
25
20
20
Axial Load (kN)
Axial Load (kN)

15
15
LVDT 10
10 LVDT
5 5
0 0
-0.10 0.10 0.30 0.50 0.70 0.90 1.10 1.30 1.50 -0.10 0.10 0.30 0.50 0.70 0.90 1.10 1.30 1.50
G250-0.95-A-400-1 G250-0.95-A-400-2
20 20
15 15
Axial Load (kN)
Axial Load (kN)

10 10
LVDT LVDT
5 5
0 0
-0.10 0.10 0.30 0.50 0.70 0.90 1.10 1.30 1.50 -0.10 0.10 0.30 0.50 0.70 0.90 1.10 1.30 1.50
G250-0.95-A-500-1 G250-0.95-A-500-2
_________________________________________________________________________________________________________________
B-28
6
10
Axial Load (kN)

Axial Load (kN)
3
5
LVDT 2
1 LVDT
0 0
-0.10 0.10 0.30 0.50 0.70 0.90 1.10 1.30 1.50 -0.10 0.10 0.30 0.50 0.70 0.90 1.10 1.30 1.50
G250-0.95-A-600-1 G250-0.95-A-700-1
160
160
140
140
120
120
100
100
Axial Load (kN)
Axial Load (kN)

80
80
60
60
40
40
20 20
0 0
0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80 -4.00 -3.50 -3.00 -2.50 -2.00 -1.50 -1.00 -0.50 0.00 0.50
G450-1.9-A-20-1
_________________________________________________________________________________________________________________
B-29
180
180
160
160
140
140
120
120
Axial Load (kN)
100
100
Axial Load (kN)

80
80
60
60
40
40
20
20
0
0
0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80 -6.00 -5.00 -4.00 -3.00 -2.00 -1.00 0.00 1.00
G450-1.9-A-20-2
160
160
140
140
120
120
100
100
Axial Load (kN)
80
Axial Load (kN)

80
60
60
40
40
20
20
0
0
-0.20 0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80
-2.00 -1.50 -1.00 -0.50 0.00 0.50
G450-1.9-A-20-3
_________________________________________________________________________________________________________________
B-30
160
160
140
140
120
120
100
Axial Load (kN)
100
80
80
60
Axial Load (kN)

60
40
40
20
20
0
0
0.00 0.50 1.00 1.50 2.00 2.50 3.00
-0.10 0.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70
G450-1.9-A-300-1
80 80
70 70
60 60
50 50
Axial Load (kN)
40 40
Axial Load (kN)

30 30
20 20
10 10
0 0
0.00 0.50 1.00 1.50 2.00 2.50 -1.00 0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00
G450-1.9-A-500-1
_________________________________________________________________________________________________________________
B-31
80
70
60
50
Axial Load (kN)

40
30
20
10
0
0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50
G450-1.9-A-500-2
25
25
20
20
Axial Load (kN)
15
15
Axial Load (kN)

10
10
5
5
0
0
0.00 0.50 1.00 1.50 2.00 2.50
-0.50 0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50 4.00 4.50 5.00 5.50 6.00
G450-1.9-A-600-1
_________________________________________________________________________________________________________________
B-32
14
14
12
12
10
Axial Load (kN) 10
8
8
Axial Load (kN)

6
6
4
4
2
2
0
0
-0.20 0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80
-0.50 0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50 4.00 4.50 5.00 5.50 6.00
G450-1.9-A-700-1
120
120
100
100
80
Axial Load (kN)
80
60
60
Axial Load (kN)

40
40
20
20
0
0
-0.20 0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40
-5.00 -4.00 -3.00 -2.00 -1.00 0.00 1.00
G250-1.95-A-20-1
_________________________________________________________________________________________________________________
B-33
120
120
100
100
80
Axial Load (kN) 80
60
60
Axial Load (kN)

40
40
20
20
0
0
-0.20 0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 1.80
-4.00 -3.50 -3.00 -2.50 -2.00 -1.50 -1.00 -0.50 0.00 0.50
G250-1.95-A-20-2
120
120
100
100
80
80
Axial Load (kN)
Axial Load (kN)

60
60
40
40
20
20
0
0
-0.20 0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 -6.00 -5.00 -4.00 -3.00 -2.00 -1.00 0.00
G250-1.95-A-20-3
_________________________________________________________________________________________________________________
B-34
90 120
80
100
70
60 80
Axial Load (kN)
50
Axial Load (kN)

60
40
30 40
20
20
10
0 0
-0.20 0.30 0.80 1.30 1.80 2.30 2.80 3.30 3.80 4.30 4.80 -1.00 -0.50 0.00 0.50 1.00 1.50 2.00 2.50 3.00
G250-1.95-A-300-1
70
70
60
60
50
50
40
Axial Load (kN)
40
Axial Load (kN)

30
30
20
20
10
10
0
0
-0.20 0.30 0.80 1.30 1.80 2.30 2.80 3.30
-2.50 -2.00 -1.50 -1.00 -0.50 0.00 0.50
G250-1.95-A-400-1
_________________________________________________________________________________________________________________
B-35
40
40
35
35
30
30
25 25
Axial Load (kN)
20 20
Axial Load (kN)

15 15
10 10
5 5
0 0
-0.20 0.30 0.80 1.30 1.80 2.30 2.80 -1.00 0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00
G250-1.95-A-500-1
25 25
20 20
15 15
Axial Load (kN)
Axial Load (kN)

10 10
5 5
0 0
-0.20 0.30 0.80 1.30 1.80 -0.50 0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50 4.00
G250-1.95-A-600-1
_________________________________________________________________________________________________________________
B-36
16
16
14
14
12
12
10
10
Axial Load (kN)
8
8
Axial Load (kN)

6
6
4 4
2 2
0 0
-0.20 0.30 0.80 1.30 1.80 -0.50 0.00 0.50 1.00 1.50 2.00 2.50
G250-1.95-A-700-1
_________________________________________________________________________________________________________________
B-37
Appendix C
Sample Calculations for Local Buckling Capacity of Lipped Channel Cold-
formed Steel Column
Local buckling capacity calculations for G550-0.95-20-A specimen using
AS/NZS 4600 design method
t = 0.95 fy = 615 Nmm −2 E = 205000 Nmm −2 62
⎡ kπ 2 E ⎤⎛ t ⎞
2
=⎢ ⎥⎜ ⎟
f cr ( web )
(
⎣12 1 − υ
2
) ⎦⎝ b ⎠ 27
⎡ 4 × π 2 × 206000⎤⎛ 0.95 ⎞
2
=⎢ ⎥⎜ ⎟ = 174 Nmm −2
(
⎣ 12 1 − 0.3
2
)
⎦⎝ 62 ⎠ 9
2
⎛ 62 ⎞
f cr ( flange) = 174 × ⎜ ⎟ = 917.51 Nmm −2
⎝ 27 ⎠
2
⎛ 62 ⎞ ⎛ 0.43 ⎞
f cr (lip ) = 174 × ⎜ ⎟ × ⎜ ⎟ = 887.69 Nmm
−2
⎝ ⎠ ⎝
9 4 ⎠
⎡ f* ⎤
λ=⎢ ⎥ f y = 615 Nmm −2
⎢⎣ f cr ⎥⎦
615 615 615

λweb = = 1.880 λ flange = = 0.819 λlip = = 0.832
174 917.51 887.69
λweb = 1.880 > 0.673 λ flange = 0.819 > 0.673 λ lip = 0.832 > 0.673
⎡ 0.22 ⎤ ⎡ 0.22 ⎤ ⎡ 0.22 ⎤

⎢1 − 1.88 ⎥ ⎢1 − 0.819 ⎥ ⎢1 − 0.832 ⎥
ρ=⎣ ⎦ = 0.470 ρ=⎣ ⎦ = 0.893 ρ=⎣ ⎦ = 0.884
1.88 0.819 0.832
Effective width = ρb
Effective area = 0 .95 × (0 .470 × 62 + 0 .893 × 27 × 2 + 0 .884 × 9 × 2 )
= 88.60 mm2
Local buckling capacity = Aefffy
= 88.60 × 615 × 10 −3
= 54.5kN
Eurocode 3 Part 1.3 and North American Specifications (AISI) design methods are
identical to that of AS/NZS 4600
_______________________________________________________________
C-1
BS 5950 Part 5 design method
t = 0.95 fy = 615 Nmm −2 E = 205000 Nmm −2 62
2
⎛t⎞
pcr = 0.904 EK ⎜ ⎟
⎝b⎠
27
2
⎛ 0.95 ⎞
p cr ( web ) = 0.904 × 205000 × 4 × ⎜ ⎟ = 174.0 Nmm
−2
⎝ 62 ⎠
9
2
⎛ 62 ⎞
p cr ( flange ) = 174.0 × ⎜ ⎟ = 917.7 Nmm −2
⎝ 27 ⎠
2
0.425 ⎛ 27 ⎞
p cr (lip ) = 917.7 × × ⎜ ⎟ = 877.55 Nmm −2
4 ⎝ 9 ⎠
beff
f c / pcr ≤ 0.123 , =1
b
f c / pcr > 0.123 ,

beff
b
[ {
= 1 + 14 ( f c / p cr )
1/ 2
− 0.35 }]
4 − 0. 2
⎛ fc ⎞ 615.00
⎜⎜ ⎟⎟ = = 3.5344 > 0.123
⎝ cr ⎠ web
p 174
−0.2
⎧ ⎡⎛ 615 ⎞ 12 ⎤ ⎫⎪
4
⎛ beff ⎞ ⎪
⎜⎜ ⎟⎟ = ⎨1 + 14 ⎢⎜ ⎟ − 0.35⎥ ⎬ = 0.4187
⎝ b ⎠ web ⎪⎩ ⎢⎣⎝ 174 ⎠ ⎥⎦ ⎪
⎭
⎛ fc ⎞ 615.0
⎜⎜ ⎟⎟ = = 0.670 > 0.123
⎝ p cr ⎠ flange 917.7
−0.2
⎧ ⎡⎛ 615 ⎞ 12 ⎤ ⎫⎪
4
⎛ beff ⎞ ⎪
⎜⎜ ⎟⎟ = ⎨1 + 14 ⎢⎜ ⎟ − 0.35⎥ ⎬ = 0.9019
⎝ b ⎠ flange ⎪⎩ ⎢⎣⎝ 917.7 ⎠ ⎥⎦ ⎪
⎭
⎛ fc ⎞ 615.00
⎜⎜ ⎟⎟ = = 0.7008 > 0.123
⎝ p cr ⎠ lip 877.55
−0 . 2
⎧ ⎡⎛ 615 ⎞ 12 ⎤ ⎫⎪
4
⎛ beff ⎞ ⎪
⎜⎜ ⎟⎟ = ⎨1 + 14 ⎢⎜ ⎟ − 0.35⎥ ⎬ = 0.8902
⎝ b ⎠ lip ⎪⎩ ⎢⎣⎝ 877.55 ⎠ ⎥⎦ ⎪
⎭
Aeff = (62 × 0.4187 + 27 × 0.9019 × 2 + 9 × 0.8902 × 2 ) × 0.95 = 86 .16 mm 2
Local buckling capacity = Aeff f y = 86 .16 × 615 × 10 −3 = 52 .99 kN
_______________________________________________________________
C-2
Local buckling capacity calculations for G550-0.95-20-A specimen using the
Direct Strength Method
t = 0.95
fy = 615 Nmm −2
E = 205000 Nmm −2
62
27
9
Elastic buckling load = 31.66 kN (from FEA)
⎡ ⎛ Pcrl ⎞ ⎤⎛ Pcrl ⎞
0.4 0.4
P
Pn = ⎢1 − 0.15⎜ ⎟ ⎥⎜ ⎟ P for > 0.776 ,
⎢⎣ ⎝ P ⎠ ⎥⎦⎝ P ⎠ Pcrl
Else Pn = P = Ag f y
P = (62 + 27 × 2 + 9 × 2) × 0.95 × 615 × 10 −3 = 78.29 kN
P 78.29
= = 1.572 > 0.776
Pcrl 31.66
⎡ ⎛ 31.66 ⎞ ⎤⎛ 31.66 ⎞
0.4 0.4
Pn = ⎢1 − 0.15⎜ ⎟ ⎥⎜ ⎟ × 78.29 = 49 kN
⎢⎣ ⎝ 78.29 ⎠ ⎥⎦⎝ 78.29 ⎠
_______________________________________________________________
C-3
Eurocode 3 Part 1.2 design method 62
t = 0.95
fy = 615 Nmm −2 27
−2
E = 205000 Nmm
9
Aeff ( Ambient ) = 88 .64 mm 2
(from AS/NZS4600 calculations which are similar to
Eurocode3 part 1.3)
Pc = Aeff ( Ambient ) × f yT = 88 .64 × 615 × 10 −3 = 54 .5 kN
Capacity calculations for G550-0.95-300-A specimen

Pc = Aeff ( Ambient ) × f y 300 = 88 .64 × 553 × 10 −3 = 49 .0 kN
_______________________________________________________________
C-4
Appendix D
Load-Deflection Curves of Flexural-torsional Buckling Tests
70 70

60 60
50 50
40 40
30 30
Flange
20 20
Web
10 10
0 0
0.00 2.00 4.00 6.00 8.00 10.00 12.00 -60.0 -40.0 -20.0 0.0 20.0 40.0 60.0
Axial Shortening (mm) Out-of-Plane Deflection(mm)
G450-1.90-2800-20
70
60
60
50
50
40

40
30
30
20
20
Flange
10 10 Web
0 0
-2.00 0.00 2.00 4.00 6.00 8.00 10.00 12.00 -80.0 -60.0 -40.0 -20.0 0.0 20.0 40.0 60.0 80.0 100.0
G450-1.90-2800-200
_________________________________________________________________________________________________________________
Behaviour and Design of Cold-formed Steel Compression Members at Elevated Temperatures D-1
40 40
35 35

30 30
25 25

20 20
15 15
Flange
10 10
Web
5 5
0 0
-2.00 0.00 2.00 4.00 6.00 8.00 10.00 12.00 14.00 -70.0 -60.0 -50.0 -40.0 -30.0 -20.0 -10.0 0.0 10.0 20.0
G450-1.90-2800-400
35
35
30
30
25 25

20 20
15 15
**Initial part of graph is not available Flange

10 10
due to an instrumentation error
Web
5 5
0 0
-1.00 0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 -60.0 -50.0 -40.0 -30.0 -20.0 -10.0 0.0 10.0 20.0 30.0
G450-1.90-2800-500
_________________________________________________________________________________________________________________
20 20
18
16

15
14

12
10 10
8 Flange
6 Web
5
4
0 0
-2.00 0.00 2.00 4.00 6.00 8.00 10.00 12.00 14.00 16.00 -35.0 -30.0 -25.0 -20.0 -15.0 -10.0 -5.0 0.0 5.0
G450-1.90-2800-600
12 12
10 10
8 8

6 6
Flange
4 4
Web
2 2
0 0
-2.00 0.00 2.00 4.00 6.00 8.00 10.00 12.00 -40.0 -20.0 0.0 20.0 40.0 60.0 80.0 100.0
G450-1.90-2800-700
_________________________________________________________________________________________________________________
12 12
10 10

8 8

6 6
4 Flange 4
Web
2 2
0 0
-1.00 0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00 9.00 10.00 -30.0 -25.0 -20.0 -15.0 -10.0 -5.0 0.0 5.0
G450-1.90-2800-700R*
60
60
50
50
40
40

30
30
20 Flange 20
Web
10 10
0 0
-1.00 0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00 -35.0 -30.0 -25.0 -20.0 -15.0 -10.0 -5.0 0.0 5.0
G250-1.95-2800-20
_________________________________________________________________________________________________________________
60 60
50 50
Axial Compression Load (kN) 40 40

30 30
Flange
20 20
Web
10 10
0 0
-2.00 0.00 2.00 4.00 6.00 8.00 10.00 12.00 -40.0 -20.0 0.0 20.0 40.0 60.0 80.0
G250-1.95-2800-200
45 45
40 40
35 35
30 30

25 25
20 20
15 Flange 15
10 Web 10
5 5
0 0
-2.00 0.00 2.00 4.00 6.00 8.00 10.00 12.00 -18.0 -16.0 -14.0 -12.0 -10.0 -8.0 -6.0 -4.0 -2.0 0.0 2.0
G250-1.95-2800-300
_________________________________________________________________________________________________________________
40
40
35
35
30
30

25
20

20
15
15
Flange
10 10
Web
5 5
0 0
-2.00 0.00 2.00 4.00 6.00 8.00 10.00 12.00 14.00 -70.0 -60.0 -50.0 -40.0 -30.0 -20.0 -10.0 0.0 10.0
G250-1.95-2800-400
35 35
30 30
25
25
20 20

15 15
10 Flange
10
Web
5 5
0 0
-5.00 0.00 5.00 10.00 15.00 20.00 25.00 -150.0 -100.0 -50.0 0.0 50.0 100.0 150.0
G250-1.95-2800-500
_________________________________________________________________________________________________________________
25
25
20
20

15
15

10
10
Flange
5
5
Web
0 0
-2.00 0.00 2.00 4.00 6.00 8.00 10.00 12.00 14.00 16.00 18.00 -80.0 -60.0 -40.0 -20.0 0.0 20.0 40.0
G250-1.95-2800-600
20 20
18 18
16 16
14 14
12 12

10 10
8 8
6 6
4 Web
4
2 2
0 0
-2.00 0.00 2.00 4.00 6.00 8.00 10.00 12.00 14.00 -35.0 -30.0 -25.0 -20.0 -15.0 -10.0 -5.0 0.0 5.0
G250-1.95-2800-600R
_________________________________________________________________________________________________________________
14
14
12
12
10
8
8

6
6
4
Flange 4
2 Web
2
0
0
-2.00 0.00 2.00 4.00 6.00 8.00 10.00 12.00 14.00 -30.0 -25.0 -20.0 -15.0 -10.0 -5.0 0.0 5.0
G250-1.95-2800-700
9
9
8
8
7
7
6
6
5
5

4
4
3
3
Flange
2 2
Web
1 1
0 0
-2.00 0.00 2.00 4.00 6.00 8.00 10.00 12.00 14.00 -40.0 -35.0 -30.0 -25.0 -20.0 -15.0 -10.0 -5.0 0.0 5.0
G250-1.95-2800-700R
_________________________________________________________________________________________________________________
18
18
16
16
14
14
12
12
10

10
8
8
6
6
Flange
4
4 Web
2
2
0 0
0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00 9.00 10.00 -60.0 -40.0 -20.0 0.0 20.0 40.0 60.0 80.0 100.0 120.0
G550-0.95-2800-20
14
14
12
12
10
10

8
6
6
4 Flange
4
Web
2 2
0 0
-2.00 0.00 2.00 4.00 6.00 8.00 10.00 12.00 14.00 16.00 -80.0 -60.0 -40.0 -20.0 0.0 20.0 40.0 60.0 80.0
G550-0.95-2800-200
_________________________________________________________________________________________________________________
10 10
9 9
8 8
7 7
6 6

5 5
4 4
3 3 Flange
2 2 Web
1 1
0 0
-2.00 0.00 2.00 4.00 6.00 8.00 10.00 12.00 14.00 16.00 18.00 -80.0 -60.0 -40.0 -20.0 0.0 20.0 40.0 60.0 80.0 100.0 120.0 140.0
G550-0.95-2800-400
8
10
7
6

4
5
Flange
3
Web
2
0 0
-2.00 0.00 2.00 4.00 6.00 8.00 10.00 12.00 14.00 -70.0 -60.0 -50.0 -40.0 -30.0 -20.0 -10.0 0.0 10.0 20.0 30.0
G550-0.95-2800-500
_________________________________________________________________________________________________________________
5

4
3
3

2
Flange 2
Web
1
1
0
0
-1.00 0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 -25.0 -20.0 -15.0 -10.0 -5.0 0.0 5.0
G550-0.95-2800-600
3 3
2 2

Flange
1 1
Web
0 0
-1.00 0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 -25.0 -20.0 -15.0 -10.0 -5.0 0.0 5.0
G550-0.95-2800-700
_________________________________________________________________________________________________________________
120
120
100
100

80
80
60
60
web
40
40 Flange
20
20
0 0
0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 -5.00 0.00 5.00 10.00 15.00 20.00 25.00 30.00
Axial Shortening (mm) Out-of-Plane Deflectioon (mm)
G450-1.90-1800-20
120 120
100 100
80 80

60 60
40 Flange 40
Web
20 20
0 0
-1.00 0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 -30.0 -25.0 -20.0 -15.0 -10.0 -5.0 0.0 5.0
G450-1.90-1800-200
_________________________________________________________________________________________________________________
90 90
80 80
70 70
60 60

50 50
40 40
30 30 Flange
20 20 Web
10 10
0 0
-1.00 0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00 9.00 10.00 -10.0 0.0 10.0 20.0 30.0 40.0 50.0 60.0
G450-1.90-1800-300
80
80
70
70
60
60
50
50

40
40
30
30
Flange
20 20 Web
10 10
0 0
-5.00 0.00 5.00 10.00 15.00 20.00 -80.0 -60.0 -40.0 -20.0 0.0 20.0 40.0 60.0 80.0 100.0 120.0
G450-1.90-1800-400
_________________________________________________________________________________________________________________
50
50
45
45
40
40

35

30
30
25 25
20 20
Flange
15 15
Web 10
10
5 5
0 0
-2.00 0.00 2.00 4.00 6.00 8.00 10.00 12.00 14.00 16.00 18.00 -45.0 -40.0 -35.0 -30.0 -25.0 -20.0 -15.0 -10.0 -5.0 0.0 5.0
G450-1.90-1800-500
20 20
18 18
16 16
14 14

12 12
10 10
8 8
Flange
6 6
Web
4 4
2 2
0 0
-2.00 0.00 2.00 4.00 6.00 8.00 10.00 12.00 14.00 16.00 -10.0 0.0 10.0 20.0 30.0 40.0 50.0 60.0 70.0 80.0 90.0
G450-1.90-1800-600
_________________________________________________________________________________________________________________
14.00
14
12.00
10 10.00

8 8.00
6 6.00
web
4 4.00
Flange
2 2.00
0 0.00
0.00 2.00 4.00 6.00 8.00 10.00 12.00 14.00
-25.00 -20.00 -15.00 -10.00 -5.00 0.00 5.00
G450-1.90-1800-700
100 100
90 90
80 80
70 70

60 60
50 50
40 40
Flange
30 30
Web
20 20
10 10
0 0
-1.00 0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 -5.0 0.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0 40.0 45.0
G250-1.95-1800-20
_________________________________________________________________________________________________________________
90 90
80 80
70 70

60 60

50 50
40 40
30 Flange
30
Web
20 20
10 10
0 0
-1.00 0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00 -10.0 -5.0 0.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0 40.0 45.0
G250-1.95-1800-200
60 60
50 50
40 40

30 30
20 Flange
20
Web
10 10
0 0
-1.00 0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 -5.0 0.0 5.0 10.0 15.0 20.0 25.0 30.0
G250-1.95-1800-300
_________________________________________________________________________________________________________________
60
60
50
50

40 40

30 30
Flange
20 20
Web
10 10
0 0
-1.00 0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00 -10.0 0.0 10.0 20.0 30.0 40.0 50.0 60.0
G250-1.95-1800-400
35 35
30 30
25
25

20 20
15 15
10 Flange 10
Web
5 5
0 0
-2.00 0.00 2.00 4.00 6.00 8.00 10.00 -40.0 -35.0 -30.0 -25.0 -20.0 -15.0 -10.0 -5.0 0.0 5.0 10.0
G250-1.95-1800-500
_________________________________________________________________________________________________________________
25
25
20
20
15

15
10
10
Flange
5 5 Web
0 0
-2.00 0.00 2.00 4.00 6.00 8.00 10.00 12.00 14.00 16.00 18.00 -10.0 0.0 10.0 20.0 30.0 40.0 50.0 60.0 70.0 80.0 90.0
G250-1.95-1800-600
14 14
12 12
10
10

8 8
6 6
4 Flange 4
Web
2 2
0 0
-2.00 0.00 2.00 4.00 6.00 8.00 10.00 12.00 14.00 16.00 18.00 20.00 -35.0 -30.0 -25.0 -20.0 -15.0 -10.0 -5.0 0.0 5.0
G250-1.95-1800-700
_________________________________________________________________________________________________________________
30
30
25
25
20 20

15 15
10 10 Flange
Web
5 5
0 0
0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 -10.0 0.0 10.0 20.0 30.0 40.0 50.0 60.0
G550-0.95-1800-20
25
25
20
20

15
15
10 10
Flange
Web
5 5
0 0
-2.00 0.00 2.00 4.00 6.00 8.00 10.00 12.00 -40.0 -20.0 0.0 20.0 40.0 60.0 80.0 100.0 120.0 140.0 160.0
G550-0.95-1800-200
_________________________________________________________________________________________________________________
25 25
20 20

15 15
10 10
Flange
5 Web
5
0 0
-1.00 0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00 -10.0 0.0 10.0 20.0 30.0 40.0 50.0 60.0
G550-0.95-1800-300
25
25
20
20
15

15
10 10
Flange
5 5
Web
0 0
-1.00 0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 -30.0 -20.0 -10.0 0.0 10.0 20.0 30.0 40.0 50.0 60.0 70.0
G550-0.95-1800-400
_________________________________________________________________________________________________________________
7
7
6
6
5
5

4
4
3 3
2 2
Flange
1 1 Web
0 0
0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00 -10.0 0.0 10.0 20.0 30.0 40.0 50.0
G550-0.95-1800-600
4 4
3 3

2 2
Flange
1 1
Web
0 0
0.00 1.00 2.00 3.00 4.00 5.00 6.00 7.00 8.00 9.00 -30.0 -25.0 -20.0 -15.0 -10.0 -5.0 0.0 5.0 10.0
G550-0.95-1800-700
_________________________________________________________________________________________________________________
Appendix E
Flexural-torsional Buckling Sample Calculations
AS/NZS 4600 for G450-1.90-2800-20 specimen
E=206000MPa, 48
fy=515MPa,
14
L = 2820 mm
73
348113.38 132947.8
rx = = 30.66 mm ry = = 18.95 mm
370.36 370.36
xo = 44.29 mm yo = 0
roi = 44.29 2 + 30.66 2 + 18.95 2 = 57.01 mm
π 2E π 2 206000
f ox = f ox = = 961.2 MPa
(le r )2 (2820 × 0.5 30.66)2
GJ ⎛ π 2 EI w ⎞
f oz = ⎜1 + ⎟
Aroi2 ⎜ GJlez2 ⎟
⎝ ⎠
79230.7 × 436.3 ⎛ π 2 × 206000 × 184876391.9 ⎞

f oz = ⎜ 1 + ⎟ = 185.19 MPa
370.36 × 57.012 ⎜⎝ 79230.7 × 436.3 × (2820 × 0.5)2 ⎟
⎠
f oxz =
1 ⎡
2 β ⎢⎣
( f ox + f oz ) − (( f ox
2
)
+ f oz ) − 4 β f ox f oz ⎤
⎥⎦
β = 1 − (44.29 57.01)2 = 0.4
f oxz =
1 ⎡
2 × 0.4 ⎢⎣
(961.2 + 185.19 ) − ((961.2 + 185.19) 2
)
− 4 × 0.4 × 961.2 × 185.19 ⎤
⎥⎦
= 164.7 MPa
fy 515
λc = = = 1.77 > 1.5
f oc 164.7
⎛ ⎞
f n = ⎜ 0.877 2 ⎟ f y λc > 1.5
⎝ λc ⎠
(
f n = 0.877
1.77 2
)× 515 = 144.45 MPa
Ae @ f n ,
_________________________________________________________________
Behaviour and Design of Cold-formed Steel Compression Members at Elevated Temperature
E-1
⎡ kπ 2 E ⎤⎛ t ⎞
2
f cr = ⎢ 2 ⎥⎜ ⎟
⎣12(1 − υ )⎦⎝ b ⎠
⎡ 4 × π 2 205000 ⎤⎛ 1.88 ⎞
2
=⎢ ⎥⎜ ⎟ = 493.94 MPa
( )
f cr ( web )
⎣ 12 1 − 0.3 ⎦⎝ 73 ⎠
2
⎡ f* ⎤ 144.45
λ=⎢ ⎥= = 0.541 ≤ 0.673 ∴ρ =1 be ( web ) = b = 73
⎣⎢ f cr ⎦⎥ 493.94
⎡ 4 × π 2 205000 ⎤⎛ 1.88 ⎞
2
f cr ( flange ) = ⎢ ⎥⎜ ⎟ = 1142.45 MPa

⎣ 12 (1 − 0.3 2
) ⎦ ⎝ 48 ⎠
⎡ f* ⎤ 144.45
λ=⎢ ⎥= = 0.356 ≤ 0.673 ∴ ρ = 1 be ( flange) = b = 48
⎢⎣ f cr ⎥⎦ 1142.45
⎡ 0.43 × π 2 205000 ⎤⎛ 1.88 ⎞

2
f cr (lip ) = ⎢ ⎥⎜ ⎟ = 1443.7 MPa

⎣ 12(1 − 0.3 ) ⎦⎝ 14 ⎠
2
⎡ f* ⎤ 144.45
λ=⎢ ⎥= = 0.316 ≤ 0.673
⎢⎣ f cr ⎥⎦ 1443.7
∴ ρ = 1 be ( flange) = b = 14
∴ Ae = Ag
Buckling capacity = Ae f n = 370.36 × 144.45 × 10 −3 = 53.5kN
The calculations for elevated temperatures are similar to the above and are simply
based on the reduced E and fy values at elevated temperatures.
_________________________________________________________________
E-2
BS 5950 Part 5 for G450-1.90-2800-20 specimen
48
π 2 EI x
PE =
L2E 14
π 2 × 206000 × 348113.38 × 10 −3
PEX = = 356 kN
(2820 2 ) 2 73
π 2 × 206000 × 132947 × 10 −3
PEY = = 135.95 kN
(
2820
2
)
2
PTF =
1
2β
[ {
(PEx + PT ) − (PEx + PT )2 − 4 βPEx PT }
1/ 2
]
Modified equation for torsional buckling
1 ⎛ π 2 EC w ⎞
PT = 2 ⎜ GJ + ⎟
⎜ L2E ⎟⎠
ro ⎝
⎛ ⎞
1 ⎜ π 2 × 206000 × 184876391.9 ⎟
PT = ⎜ 79230.7 × 436.3 + ⎟ = 68.59 kN
57.012 ⎜
⎝
2820
2
2
( ⎟
⎠
)
PTF =
1
(2 × 0.4 )
[ {
(356 + 68.59 ) − (356 + 68.59 )2 − 4 × 0.4 × 356 × 68.59 = 61.00 kN
1/ 2
} ]
Pcs + (1 + η )PE
φ=
2
Pcs = 171.23 kN (From Local Buckling Capacity)
PEY 135.95
For Le r > 20, η = 0.002(αLE r − 20 ) α = = = 1.493
PTF 61.00
((
η = 0.002 1.493 × 2820 2 18.95 − 20 = 0.182 ) )
171.23 + (1 + 0.182 ) × 61
φ= = 121.67
2
PE Pcs
Pc =
φ + φ 2 − PE Pcs
61.00 × 171.23
Pc = = 55.65 kN
121.67 + 121.67 2 − 61 × 171.23
_________________________________________________________________
E-3
Eurocode 3 Part 1.3 for G450-1.90-2800-20 specimen
σ cr ,TF = f oxz = 164.7 MPa , σ cr ,T = f oz = 185.19 MPa σ cr ( web) = 493.94 MPa ,
σ cr ( flange ) = 1142.45 MPa and σ cr (lip ) = 1443.7 MPa from AS/NZS 4600
calculations 48
14
73
(f σ cr ) = 515 493.94 = 1.021 > 0.673

_
λp = yb
⎛ ⎞
⎜ 0.22 _ ⎟
−
( )
⎜⎜ 1.0 ⎟⎟
λ ⎠ = 1.0 −
0.22
ρ web =⎝ 1.021 = 0.768
p
_
1.021
λp
(f σ cr ) = 515 1142.45 = 0.671 < 0.673 ρ flange = 1

_
λp = yb
(f σ cr ) = 515 1143.7 = 0.597 < 0.673 ρ lip = 1

_
λp = yb
Aeff = (73 × 0.768 + 48 × 2 + 14 × 2 ) × 1.88 = 338.5 mm 2
β A = (Aeff Ag ) = 338.5 370.36 = 0.914
(f σ cr ) β A = ⎜
⎛ 515 ⎞
_
λ= yb ⎟ × 0.914 = 1.691
⎝ 164.7 ⎠
⎡ ⎛_ ⎞ ⎤
[ ]
_
φ = 0.5⎢1 + α ⎜ λ − 0.2 ⎟ + λ2 ⎥ = 0.5 1 + 0.34(1.691 − 0.2) + 1.6912 = 2.183
⎣ ⎝ ⎠ ⎦
1
χ= = 0.281
(
2.183 + 2.183 2 − 1.6912 )0.5
−3
N b, Rd = χAeff f y γ M 1 = 0.281× 338.5 × 515 × 10 = 48.94 kN
1.0
_________________________________________________________________
E-4
Eurocode 3 Part 1.2 for G450-1.90 – 2800-400
Calculations at 400oC
48
14
73
_
Aeff @ 20oC = 338.57 mm 2 and λ @ 20 o C = 1.691 (From Eurocode 3 Part 1.3
calculation)
0.5
⎡ 357 ⎤
λθ = λ [k y ,θ k E ,θ ]0.5 = 1.691 × ⎢ 515 ⎥ = 1.848
⎢119500 206000 ⎥
⎣⎢ ⎦⎥
α = 0.65 235 f
yθ
α = 0.65 235 357 = 0.439
ϕθ =
1
2
[ 1
] [
1 + αλθ + λθ2 = 1 + 0.439 × 1.848 + 1.848 2 = 2.695
2
]
1 1
χ fi = = = 0.215
ϕθ + ϕθ2 − λθ2 2.695 + 2.695 2 − 1.8482
N b, fi ,t , Rd = χ fi Ak y ,θ f y θ γ M , fi = 0.215 × 338.57 × 357 × 10 −3 = 25.95 kN
_________________________________________________________________
E-5
Direct Strength Method for G450-1.90-2800-20 specimen 48
14
73
Elastic Buckling Load = 164.7 × 370.36 × 10 −3 = 61kN (From AS/NZS 4600

calculations)
Ny 370.36 × 515 × 10 −3
λc = = = 1.768
NC 61.00
λc > 1.5 ⎛ 0.877 ⎞

N CE = ⎜ 2 ⎟
( )
× 370.36 × 515 × 10 −3 = 53.5 kN
⎝ 1.768 ⎠
_________________________________________________________________
E-6
Appendix F
Sample calculation for capacity reduction factor (F)
Calculation with reference to Table 6.1
− β 0 Vm 2 +V f 2 + C pV p 2 +Vq 2
Φ = 1.52 M m Fm Pm e
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
− 2.5 0.0566 + C pV p 2
Φ = 1.672 Pm e
Number of specimens n = 14
m = Degree of freedom = n-1
Cp = Correction factor depending on the number of tests
⎛ 1 ⎞⎛ m ⎞ ⎛ 1 ⎞⎛ 13 ⎞
= ⎜1 + ⎟⎜ ⎟ = ⎜1 + ⎟⎜ ⎟ =1.266
⎝ n ⎠⎝ m − 2 ⎠ ⎝ 14 ⎠⎝ 13 − 2 ⎠
Pm = mean value of FEA to predicted load ratio=0.934
Vp = Coefficient of variation of FEA to predicted load ratio=0.041
0.0566+1.226×0.0412
Φ = 1.672 × 0.934e −2.5
Φ = 0.851
_________________________________________________________________
Behaviour and Design of Cold-formed Steel Compression Members at Elevated temperatures F-1
Appendix G
Sample Calculations for Plotting of Flexural-torsional Buckling Loads with
AS/NZS 4600 Design Curve
G250-1.95-60-40-15-1500 specimen at ambient temperature ( with reference to
Figure 9.1) 40
E=188000MPa,
15
fy=271MPa,
L = 1500 mm
60
t = 1.95 mm
Ag = 331 .5 mm 2
Pcr = 125 .84 kN From FEA
Pult = 75kN From FEA
Py = Ag f y = 331.5 × 271 × 10 −3 = 89.84kN
Py 89.84
λ= = = 0.84492
Pcr 125.84
f n = 201.96 MPa From AS/NZS 4600 calculation method
⎡ kπ 2 E ⎤⎛ t ⎞
2
f cr = ⎢ 2 ⎥⎜ ⎟
⎣12(1 − υ )⎦⎝ b ⎠
⎡ 4 × π 2 188000 ⎤⎛ 1.88 ⎞
2
=⎢ ⎥⎜ ⎟ = 717.9 MPa
( )
f cr (web )
⎣ 12 1 − 0.3 ⎦⎝ 60 ⎠
2
⎡ fn ⎤ 201.96
λ=⎢ ⎥= = 0.530 ≤ 0.673 ∴ρ =1 be ( web ) = b = 60
⎢⎣ f cr ⎥⎦ 717.9
∴ be ( flange) = b = 40
⎡ 0.43 × π 2 188000 ⎤⎛ 1.95 ⎞

2
=⎢ ⎥⎜ ⎟ = 1234 MPa
( )
f cr (lip )
⎣ 12 1 − 0.3 ⎦⎝ 15 ⎠
2
⎡ fn ⎤ 201.96
λ=⎢ ⎥= = 0.404 ≤ 0.673
⎢⎣ f cr ⎥⎦ 1234
∴ ρ = 1 be(lip ) = b = 15
∴ Aeff @ f n = Ag
_________________________________________________________________
G-1
Ps = 331.5 × 271 × 10 −3 = 89.84kN
Pu 75
= = 0.8348
Ps 89.84
⎛ Py Pu ⎞
Po int ≡ ⎜ , ⎟ ≡ (0.8449,0.8348)
⎜ Pcr Py ⎟
⎝ ⎠
Sample Calculations for Plotting of Flexural-torsional Buckling Loads with

Eurocode 3 Part 1.3 Design Curve
G250-1.95-60-40-15-1500 specimen at ambient temperature ( with reference to
Figure 9.4) 40
15
E=188000MPa,
fy=271MPa,
60
L = 1500 mm
t = 1.95 mm
Ag = 331 .5 mm 2
Pcr = 125 .84 kN From FEA
Pult = 75kN From FEA
Pcr 125.84 × 1000

σ cr = = = 379.61 MPa
Ag 331.5
⎡ kπ 2 E ⎤⎛ t ⎞
2
=⎢ ⎥⎜ ⎟
( )
f cr
⎣12 1 − υ ⎦⎝ b ⎠
2
⎡ 4 × π 2 188000 ⎤⎛ 1.88 ⎞
2
=⎢ ⎥⎜ ⎟ = 717.9 MPa
( )
f cr ( web )
⎣ 12 1 − 0.3 ⎦⎝ 60 ⎠
2
⎡ fy ⎤ 271
λ=⎢ ⎥= = 0.614 ≤ 0.673 ∴ρ =1 be ( web ) = b = 73
⎢⎣ f cr ⎥⎦ 717.9
∴ be ( flange) = b = 40
_________________________________________________________________
G-2
⎡ 0.43 × π 2 188000 ⎤⎛ 1.95 ⎞
2
f cr ( lip ) = ⎢ ⎥⎜ ⎟ = 1234 MPa

⎣ 12(1 − 0.3 ) ⎦⎝ 15 ⎠
2
⎡ fy ⎤ 271
λ=⎢ ⎥= = 0.468 ≤ 0.673
⎢⎣ f cr ⎥⎦ 1234
∴ ρ = 1 be ( flange) = b = 15
∴ Aeff @ f y = Ag
Py = 331.5 × 271 × 10 −3 = 89.84kN
Pu 75
= = 0.8348
Py 89.84
⎛ Py Pu ⎞
Po int ≡ ⎜⎜ , ⎟ ≡ (0.8449,0.8348)
⎟
⎝ Pcr Ps ⎠
_________________________________________________________________
G-3
Reference
Ali, F. (2001), Determining the Effective Length of Fixed End Steel Columns in Fire,
Journal of Applied Fire Science, Vol. 10, Issue 1: 41-44.
Ali, F.A., and O’Connor, D.J. (1996-1997), Calculation of Axial Forces Generated
in Restrained Pin Ended Steel Columns Subjected to High Temperatures, Journal of
Applied Fire Science, Vol. 6, Issue 4: pp. 383-394.
Ali, F. and O’Connor, D. (2001), Structural Performance of Rotationally Restrained

Steel Columns in Fire, Fire Safety Journal, Vol. 36: pp. 679-691.
Aasen, B. (1985), An Experimental Study on Steel Columns Behaviour at Elevated

Temperatures, The Norwegian Institute of Technology, Trondheim, Norway.
Batista, M. and Rodrigues, F.C. (1992). Residual Stress Measurements on Cold-

formed Profiles, Experimental Technique, Vol. 16, Issue 5, pp. 25–29.
British Standards Institution (BSI 1998), British Standard 5950: Structural Use of
Steelwork in Buildings, Part 5: Code of Practice for Design of Cold-formed Thin
Gauge Sections.
British Standards Institution (BSI 2005), British Standard 5950: Structural Use of
Steelwork in Buildings, Part 8: Code of Practice for Fire Resistant Design.
Buchanan, A.H. (2001), Structural Design for Fire Safety, John Wiley & Sons, New
York.
Buckled column and deformed beams at Broadgate, Retrieved 24-03-2009,

http://911research.wtc7.net/mirrors/guardian2/fire/SCI-4-1.jpg
_________________________________________________________________
Behaviour and Design of Cold-formed Steel Compression Members at Elevated temperatures R-1
Burgrees, I.W., Olawale, A.O. and Plank, R.J. (1992), Failure of Steel Columns in
Fire, Fire Safety Journal, Vol. 18, Issue 2: pp. 183-201.
C&Z Interchangable Purlin Machine, retrieved 24-03-2009,

http://www.hbfxywj.com/ENGLISH/info/147-1.htm
Chen, J. and Young, B. (2006a), Design Cold-formed Steel Lipped Channel Columns
at Elevated Temperatures, Engineering Structures, Vol. 29. pp 2445-2456.
Chen, J. and Young, B. (2006b), Corner Properties of Cold-Formed Steel Sections at

Elevated Temperatures, Thin-Walled Structures, Vol. 44, Issue 2, pp 216-223.
Chen, J. and Young, B. (2007a), Experimental investigation of cold-formed steel

material at elevated temperatures, Thin-Walled Structures, Volume 45, Issue 1, pp
96-110.
Chen, J. and Young, B. (2007b), Cold-formed steel lipped channel columns at

elevated temperatures, Engineering Structures, Volume 29, Issue 10, pp 2445-2456.
Cold formed sections, Retrieved 24-03-2009,

http://imgs.tootoo.com/ed/64/ed643ec92d855ca11721cb35a0be5f46_m.jpg
Cold-formed Steel Members in Trusses, Retrieved 24-03-2009,

http://www.marinoware.com/Images/project_desktop.jpg
Cold-Formed Structures, Retrieved 24-03-2009,

http://www.eng.mcmaster.ca/civil/faculty/sivakumaran/CFS-building.gif
Dhanalakshmi, M. and Shanmugam, N.E. (2001), Design for openings in cold-

formed steel channel stub columns, Thin-Walled Structures, Vol. 39, Issue 12, pp
961-981.
_________________________________________________________________
Dolamune Kankanamge (2009), Flexural Behaviour of Cold-formed Steel Beams at
Elevated Temperatures, PhD thesis, Queensland university of technology, Brisbane,
Austrlia (in Press).
Dubina, D. and Ungureanu, V. (2002), Effect of imperfections on numerical

simulation of instability behaviour of cold-formed steel members, Journal of Thin-
Walled Structures, Vol. 40, Issue 3, pp 239-262.
Eurocode 3 (2005) CEN ENV 1993-1-2, Design of steel structures, General rules,
Structural Fire Design, Brussels.
Eurocode 3: (1996) CEN ENV 1993-1-3 Design of Steel Structures, General rules-
Supplementary rules for Cold-formed Thin gauge Members and Sheeting, Brussels.
Fatal Gloucester fire destroys apartment building, synagogue, Retrieved 24-03-2009,

http://cache.boston.com/bonzai-fba/Globe_Photo/2007/12/15/1197726448_0520.jpg
Feng, M., Wang, Y.C. and Davies, J.M. (2003a), Structural behaviour of thin-walled
Short Steel Channel Columns at Elevated Temperatures. Part 1: Experiments,
Journal of Thin-Walled Structures, Vol. 41, Issue 6, pp543-570.
Feng, M., Wang, Y.C. and Davies, J.M. (2003b), Structural Behaviour of thin-walled
Short Steel Channel Columns at Elevated Temperatures. Part 2: Design calculations
and Numerical Analyses, Journal of Thin-Walled Structures, Vol. 41, Issue 6, pp
571-594.
Feng, M., Wang, Y.C. and Davies, J.M. (2004), A Numerical Imperfection Sensitivity
Study of Cold-formed Thin-walled Tubular Steel Columns at Uniform Elevated
Temperatures, Thin-Walled Structures, Vol. 42: pp. 533-555.
_________________________________________________________________
Franssen, J.M., Cooke, G.M.E. and Latham, D.J. (1995), Numerical Simulation of a
Full Scale Fire Test on a Loaded Steel Framework, Journal of Constructional Steel
Research, Vol. 35: pp. 377-408.
Franssen, J.M., Schleich, J.B., Cajot G.L. and Azpaizu, W. (1996), A Simple Model
for the Fire Resistance of Axially Loaded Members Compression with Experimental
Results, Journal of constructional research, Vol. 37, Issue 3: pp. 175-204.
Feng, M., Wang, Y.C. and Davies, J.M., (2003c), Axial Strength of Cold-formed
Thin-walled Steel Channels under Non-uniform Temperatures in Fire, Fire Safety
Journal, Vol., pp. 679–707.
Franssen, J.M., Cooke, G.M.E. and Latham, D.J. (1995), Numerical Simulation of a
Full Scale Fire Test on a Loaded Steel Framework, Journal of Constructional Steel
Franssen, J.M., Schleich, J.B., Cajot G.L. and Azpaizu, W. (1996), A Simple Model
for the Fire Resistance of Axially Loaded Members Compression with Experimental
Results, Journal of constructional research, Vol. 37, Issue 3: pp. 175-204.
Feng, M., Wang, Y.C. and Davies, J.M.. (2004), A Numerical Imperfection
Sensitivity Study of Cold-formed Thin-walled Tubular Steel Columns at Uniform
Elevated Temperatures, Thin-Walled Structures, Vol. 42, Issue 4, pp 533-555.
Gardner, L., and Baddoo, N.R. (2006), Fire Testing and Design of Stainless Steel
Structures, Journal of Constructional Steel Research, Vol. 62, pp 532-543.
Girlich, T.J., Collier, P.C.R. and Buchanan, A.H. (1996), Design of Light Steel-
framed Walls for fire resistance, Fire and Materials, Vol. 20, pp 79-96.
Hancock, G. J. (1998), Design of Cold-Formed Steel Structures, Australian Institute

of Steel Construction, 3rd Edition, Sydney, NSW, Australia.
_________________________________________________________________
Kaitila, O. (2002), Finite Element Modelling of Cold-formed Steel Members at High
Temperatures, Helsinki University of technology Laboratory of Steel Structures
Publications 24.
Kingsland KPT 30175 Press Brake, Retrieved 24-03-2009,

http://www.weldall.co.uk/Press_Brake.jpg
Klippstein, K.H. (1980), Strength of Cold-formed Studs Expose to Fire, American

Iron and Steel Institute, Washington DC.
Lee, J., Mahendran, M. and Makelainen, P. (2003), Prediction of Mechanical

Properties of Light Gauge Steels at Elevated Temperatures, Journal of Construction
Steel Research, Vol. 59, Issue 12, pp1517-1532.
Lee, J. (2004), Local Buckling Behaviour and Design of Cold-Formed Steel

Compression Members at Elevated Temperatures, PhD thesis, Queensland
University of Technology, Brisbane, Australia.
Liu, Y. and young, B. (2003), Buckling of Stainless Steel Square Hollow Section
Compression Members, Journal of Construction Steel Research, Vol. 59, Issue 2,
pp165-177.
Makelainen, P. and Miller, K. (1983), Mechanical Properties of Cold-formed

Galvanized Steel Z32 at Elevated temperatures, Helsinki University of technology,
Finland.
North American Specification for the Design of Cold-formed Steel Structural

Members, NAS (2001), American Iron and Steel Institute
Outinen, T.A. and Myllymaki, J. (1995), The Local Buckling of RHS Members at
Elevated Temperatures, VTT research notes 1672, Technical Research Centre of
Finland, ESPOO.
_________________________________________________________________
Outinen, J. Kaitila, O. and Makelainen, P. (2000), A Study for the Development of
the Design of Steel Structures in Fire Conditions, First International Workshop of
Structures in Fire, Copenhagen, Denmark, PP 267-281.
Outinen, J. and Makelainen, P. (2001a) Effect of High Temperature on Mechanical

Properties of Cold-formed Structural Steel, Tubular Structures IX, Proceeding of the
Ninth International Conference on Tubular Structures, Dusseldorf, Germany, pp.
439-444
Outinen, J. and Makelainen, P. (2004), Mechanical Properties of Structural Steel at

Elevated Temperatures and After Cooling Down, Fire and materials, Vol. 28, Issue
2-4, PP 237-251.
Popovic, D., Hancock, G.J. and Rasmussen, K.J.R. (1999), Axial compression Tests
on Cold-Formed Angles Loaded Parallel with a Leg, Journal of Structural
Engineering, Vol. 127, Issue 6, pp 600-607.
Popovic, D., Hancock, G.J. and Rasmussen, K.J.R. (2000), Compression Tests of
DuraGal Angles Loaded Parallel with a Leg, Research Report No. R799,
Department of Civil Engineering, The University of Sydney, AUSTRALIA
Popovic, D., Hancock, G.J. and Rasmussen, K.J.R. (2001), Compression Tests of
Cold-Formed Angles, Journal f Structural Engineering, Vol. 125, Issue. 5, pp 515-
523.
Ranawaka, T. (2006), Distortional Buckling Behaviour of Cold-Formed Steel

Compression Members at Elevated Temperatures, PhD thesis, Queensland university
of technology, Brisbane, Austrlia.
Ranawaka, T. and Mahendran, M. (2009), Experimental study of the mechanical

properties of light gauge cold-formed steels at elevated temperatures, Fire Safety
Journal, Volume 44, Issue 2, pp, Pages 219-229
_________________________________________________________________
Ranby, A. (1999), Structural Fire Design of Thin Walled Steel Sections, Report
188:5, Swedish Institute of Steel Construction and Division of steel Construction,
Lulea University of Technology.
Rodrigus, J.P.C., Neves, I.C. and Valente, J.C. (2000), Experimental Research on the
Critical Temperature of Compressed Steel Element with Restrained Thermal
Elongation, Fire Safety Journal, Vol. 35, Issue 2: pp 77-98.
Sidey, M.P. and Teague, D.P. (1998), Elevated Temperature Data for Structural
Grades of Galvanised Steel, British Steel (Welsh Laboratories) Report, UK.
Standards Australia (SA 1991), AS 1391-1991, Methods for Tensile Testing of

Metal, Sydney, Australia.
Standards Australia, (SA 1998), Australian Standard, AS/NZS 4100:1998, Steel

Structures, Sydney, Australia.
Standards Australia, (SA 2005), Australian/New Zealand Standard, AS/NZS

4600:2005, Cold-formed Steel Structures, Sydney, Australia.
Sivakumaran, K.S. and Abdul-Rahman, N. (1998), A finite Element Analysis Model

for the Behaviour of Cold-formed Steel Members, Thin-Walled Structures, Vol.
31, Issue 4, pp 305-324.
Schafer, B.W. (2001), Thin-walled Column Design Considering Local, Distortional

and Euler Buckling, Proceeding of the Structural Stability research council Annual
Stability Conference, Ft. Lauderdale, F1 419-438.
Schafer, B.W. (2002), Local, Distortional and Euler Buckling of Thin-walled

Columns, Journal of Structural Engineering, Vol. 128, Issue 3: pp. 289-299.
_________________________________________________________________
Schafer, B.W. (2006), Development and Progress in the Direct Strength Method of
Cold-formed steel Member Design, Innovative design of Steel Structures, pp. 43-61.
Schafer, B.W. and Pekoz, T. (1998), Computational Modelling of Cold-formed Steel:

Characterising Geometric Imperfections and Residual Stresses, Journal of
Constructional Steel Research, Vol. 47, pp 193-210.
EN 10002-5:1991. Metallic Materials. Tensile Testing. Part 5: Method of Testing at

Talamona, D., Franssen, J.M., Schleich, J.B., and Kruppa, J. (1997), Stability of Steel
Columns in case of Fire: Numerical modelling, Journal of Structural Engineering,
Vol. 123, Issue 6: pp. 713-720.
Von Karman, T., Sechler, E.E., and Donnell, L.H. (1932), The Strength of Thin
Plates in Compression, Transactions ASME, 54.
Walker, A.C. (1975), Design and Analysis of Cold-formed Section, Halsted Press,
New York.
Weng, C. C. and Pekoz, T. (1990), Residual Stresses in Cold-Formed Steel

Members, Journal of Structural Engineering, Vol. 116, Issue 6, pp 1611-1625.
Winter, G. (1947), Strength of Thin Steel Compression Flanges, Transactions ASCE,

Vol. 112, pp 527-554.
Wang, Y.C. (1997), The Effect of Frame Continuity on the Behaviour of Steel
Columns under Fire Conditions and Fire resistance Design Proposal, Journal of
Construction Steel Research, Vol.41, Issue 1, pp 93-111.
_________________________________________________________________
Wang, Y.C., and Davies, J.M. (2003a), Fire Tests of Non-sway Loaded and
Rotationally Restrained Steel Column Assemblies, Journal of Constructional Steel
Wang, Y.C., and Davies, J.M. (2003b), An Experimental Study of Non-sway Loaded
and Rotationally Restrained Steel Column Assemblies under Fire Conditions:
Analysis of Test Results and Design Calculations, Journal of Constructional Steel
Yan, J. and Young, B. (2002), Column Tests of Cold-Formed Steel Channels with
Complex Stiffeners, Journal of Construction Steel Research, Vol. 128, Issue 6, 737-
745.
Yang, D., Hancock. G. J., and Rasmussen, K.J.R. (2002), Compression Tests of
Cold-Reduced High Strength Steel Long Columns, Research Report No R816,
Department of Civil Engineering, University of Sydney, Australia.
Yang, K.C., Lee, H.H. and Chen, O. (2006a), Performance of Steel H Columns
Loaded under Uniform Temperature, Journal of Constructional Steel Research,
Vol. 62, Issue 3: pp. 262-270.
Yang, K.C., Lee, H.H. and Chen, O. (2006b), Experimental Study of Fire-Resistant
Steel H-columns at Elevated Temperature, Journal of Constructional Steel Research,
Vol. 62, Issue 6: pp. 544-553.
Young, B. (2004a), Design of Channel Columns with Inclined Edge Stiffeners,

Journal of Constructional Steel research, Vol. 60, Issue 2, pp 183-197.
Young, B. (2004b), Tests and Design of Fixed-ended Cold-formed Steel Plain Angle
Columns, Journal of Structural Engineering, Vol.130: pp. 1931-1940.
_________________________________________________________________
Young, B. (2005), Experimental Investigation of Cold-formed Steel Lipped Angle
Concentrically Loaded Compression Members, Journal of Structural Engineering,
Vol. 131, Issue 9: pp. 1390-1396.
Young, B. and Ellobody, E. (2005), Buckling Analyses of Cold-formed Steel Lipped

Angle Columns, Journal of Structural Engineering, Vol. 131, Issue 10, pp. 1570-
1579.
Young, B. and Ramussen, J.R. (1998a), Tests of Fixed-ended Plain Channel

Columns, Journal of Structural Engineering, Vol. 124, Issue 2, pp. 131-139.
Young, B. and Ramussen, J.R. (1998b), Design of Lipped Channel Columns, Journal
of Structural Engineering, Vol. 124, Issue 2, pp. 140-148.
Young, B. and Rasmussen, K.J.R. (1999), Behaviour of Cold-formed Singly

Symmetric Columns, Thin-Walled Structures, Vol. 33, Issue 2: pp. 83-102.
Yu, W. W. (2000), Cold-formed Steel Design, John Wiley & Sons, New York, Third
edition
_________________________________________________________________

Yasintha Heva Thesis PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Yasintha Heva Thesis PDF

Uploaded by

Copyright:

Available Formats

BEHAVIOUR AND DESIGN OF COLD-

FORMED STEEL COMPRESSION MEMBERS

YASINTHA BANDULA HEVA

Yasintha Bandula Heva

School of Urban Developments

A THESIS SUBMITTED TO THE SCHOOL OF URBAN DEVELOPMENTS

Light gauge cold-formed steel, local buckling, flexural-torsional buckling,

Behaviour and Design of Cold-formed Steel Compression Members at Elevated Temperatures iv

As the third phase of this research, flexural-torsional buckling behaviour was

Behaviour and Design of Cold-formed Steel Compression Members at Elevated Temperatures v

• Bandula Heva, D.Y.M. and Mahendran, M. (2007) Local Buckling Behaviour of

Behaviour and Design of Cold-formed Steel Compression Members at Elevated Temperatures vi

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

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

CHAPTER 3 Mechanical Properties of Cold-formed Steels at

CHAPTER 4 Experimental Study of Cold-formed Steel

CHAPTER 5 Development and Validation of Finite Element

CHAPTER 6 Parametric Study and Numerical Analyses of Local

CHAPTER 7 Experimental Study of Cold-formed Steel

Behaviour and Design of Cold-formed Steel Compression Members at Elevated Temperatures ix

CHAPTER 9 Parametric Study of Flexural-Torsional Buckling

CHAPTER 10 Conclusion and Recommendations................................. 10-1

Behaviour and Design of Cold-formed Steel Compression Members at Elevated Temperatures x

Appendix B Load-deflection Curves of Local Buckling Tests…………………B-1

Appendix C Sample Calculations for Local Buckling Capacities……………...C-1

Appendix D Load-deflection Curves of Flexural-torsional Buckling Tests……D-1

Appendix E Sample Calculations for Flexural-torsional Buckling capacities….E-1

Appendix F Sample Calculations for Capacity Reduction factor………………F-1

Appendix G Sample Calculations for plotting Flexural-torsional Buckling

Behaviour and Design of Cold-formed Steel Compression Members at Elevated Temperatures xi

Figure 1.1 Roll-forming Process ....................................................................... 1-2

Behaviour and Design of Cold-formed Steel Compression Members at Elevated Temperatures xv

Behaviour and Design of Cold-formed Steel Compression Members at Elevated Temperatures xx

A,Ag cross sectional area

First and foremost I offer my sincerest gratitude to my supervisor, Prof. Mahan

It is my pleasure to thank my fellow post-graduate students, friends for their support

Finally I would like to express thanks to my parents, my mother-in-law and my

Figure 1.1: Roll-forming Process

This Figure is not available online.

Figure 1.2: Press Braking Process (http://www.weldall.co.uk/Press_Brake.jpg)

This Figure is not available online.

Figure 1.3: Cold-formed Steel Sections

1.1.1 Buckling Failure

1.1.2 Low Fire Resistance

The structural behaviour of cold-formed steel compression members submitted to

(a) Deck/Cladding Applications

(b) Light Weight Floor Systems

(c) Residential Uses

The use of cold-formed steel in residential construction instead of timber has

This Figure is not available online.

Figure 1.4: Use of Cold-formed Steel Members in Residential Construction

There is an increasing trend of usage of cold-formed steel in truss systems in

This Figure is not available online.

Figure 1.5: Use of Cold-formed Steel Members in Trusses

(e) Steel Stud Wall Panel Systems

(f) Other Applications

1.3 Fire Safety in Buildings

This Figure is not available online.

Figure 1.6: Fire in a Steel Building (http://cache.boston.com/bonzai-

1.4 Problem Definition

To date, no detailed research that involves both experimental and numerical

1.5.1 Main Objective

1.5.2 Specific Objectives