BES 2012 E-Proceeding PDF

Cold-formed Steel Innovations
Green & Viable Structural Solutions

Editor
Lau Hieng Ho

Curtin University,
Sarawak Malaysia

Proceedings of the 2012 Borneo Engineering Symposium on Cold-formed
Steel Innovations
16 July, Kuching, Sarawak, Malaysia
Published by

Curtin University Sarawak Malaysia
CDT 250
98000 Miri
Sarawak Malaysia

Phone: 60 85 443939 Fax: 60 85 443837 Email: enquiries@curtin.edu.my

Green & Viable Solutions
Editor Lau Hieng Ho

Copyright 2012 CURTIN UNIVERSITY SARAWAK MALAYSIA and editor.
All rights reserved

This book/CD-ROM, or parts thereof, may not be reproduced in any form or by any means,
electronic or mechanical, including photocopying, recording or any information storage and
retrieval system now known or to be invented, without written permission from the Publisher or
the Editor.

Disclaimer
No responsibility is assumed by the Publisher or the Editor for any injury and/or damage to
persons or property as a matter of products liability, negligence or otherwise, or from any use or
operation of any methods, products or ideas contained in the material herein. Contents, used in
the papers are how they are submitted by the contributors. Whilst every attempt made to
ensure that all aspects of the papers are uniform in style, the Publisher or the Editor or the BES
Organisers will not be responsible whatsoever for the accuracy, correctness or representations
of any statements or documents presented in the papers.

ISBN 978-983-44176-4-2

Cover design and production Wee Yee Boon.
Cover page photograph Charles Kung

Printed in Malaysia

iii

TABLE OF CONTENTS

Preface iv
Acknowledgements v

LECTURES

Strengthening of Cold-formed Ferritic Stainless Steel Tubular Sections using 1
CFRP Plate
S. M. Zahurul Islam and B. Young

Stressed Skin Effects on Cold-formed Steel Portal Frames with Semi-rigid 12
Joints
A. M. Wrzesien, J. B. P. Lim and R. M. Lawson

A Review of Cold-formed Steel Built-up Columns Research at Curtin 43
University
H. H. Lau and T. C. H. Ting

Sustainability of Cold-formed Steel Portal Frames in Malaysia 69
T. E. McGrath, R. P. D. Johnston, S. Nanukuttan, J. B. P. Lim, M. Soutsos
and C. C. Mei

A Case Study of Greener Building Approach Cold-Formed Steel is a 82
Green and Viable Building Solution
C. C. Mei and P. K. Chum

iv
PREFACE

This book contains all five papers presented at the Borneo Borneo Engineering
Symposium on Cold-formed Steel Innovations (BES 2012) held on 16
th
July 2012 in
Kuching, Sarawak, Malaysia.

The symposium is organised in collaboration with The Institution of Engineers Malaysia
(IEM), Sarawak Branch; Swinburne University of Technology, Sarawak Campus; and
Malaysia Cold-formed Steel Institute (MyCSI).and sponsored by BlueScope Steel (M)
Sdn. Bhd. and EcoSteel Sdn. Bhd. BlueScope Steel Malaysia

The papers cover recent innovations in the application of cold-formed steel as a
structural material, sustainability of cold-formed steel and other technical advances in
cold-formed steel construction.

Special thanks to Sharon Yee Jia Huey, Adeline Wong Ling Ying, Wee Yee Boon and
Florence Boon Ai Hui

Editor

Lau Hieng Ho

v

ACKNOWLEDGEMENTS

Sponsorship was provided by

BlueScope Steel (Malaysia) Sdn Bhd
Curtin University Sarawak Malaysia
EcoSteel Sdn Bhd

Supported by

The institution of Engineers Malaysia Sarawak Branch
Malaysia Cold-formed Steel Institute (MyCSI)
Swinburne University of Technology Sarawak campus

Edited by Lau Hieng Ho
Copyright 2012 Curtin University Sarawak Malaysia
ISBN 978-983-44176-4-2
STRENGTHENING OF COLD-FORMED FERRITIC STAINLESS
STEEL TUBULAR SECTIONS USING CFRP PLATE

S. M. ZAHURUL ISLAM
1
and B. YOUNG
1

1
Department of Civil Engineering, The University of Hong Kong, Hong Kong

ABSTRACT

Cold-formed stainless steel tubular structural members experienced web crippling failure due to
localize concentrated loads or reactions are investigated. A series of tests on carbon fibre-reinforced
polymer (FRP) strengthening of cold-formed stainless steel tubular structural members subjected to
End-Two-Flange and Interior-Two-Flange loading conditions is presented. A total of 20 web crippling
tests were conducted. The investigation was mainly focused on the effects of different surface
treatment and different widths of CFRP plate strengthening against web crippling. Two different
surface treatments were considered. Three different widths of CFRP plate strengthening were
considered. The test specimens consisted of cold-formed ferritic stainless steel EN 1.4003 square and
rectangular hollow sections. Most of the strengthened specimens were failed by debonding of CFRP
plates from the stainless steel tubes. There are mainly four different failure modes observed in the tests,
namely the adhesion, combination of adhesion and cohesion, interlaminar failure of CFRP plate, and
combination of adhesion and interlaminar failure of CFRP plate. The failure loads, failure modes and
the load-web deformation behaviour of the ferritic stainless steel sections are presented in this study. It
was found that the web crippling capacity of cold-formed ferritic stainless steel tubular sections may
increase up to 51% using CFRP strengthening. It is found that CFRP strengthening against web
crippling to be effective when the CFRP width has the same width as the bearing length.

INTRODUCTION

Cold-formed stainless steel tubular structural members are being increasingly used in architectural and
structural application because of the desirable features on corrosion resistance, durability, easy
maintenance, fire resistance, pleasing appearance and recyclability of the material. Cold-formed stainless
steel tubular structural members may experience web crippling failure due to the high local intensity of
concentrated loads or reactions. Fibre-reinforced polymer (FRP) is an advanced material which is
increasingly being used for strengthening and repair of existing metal structures. Externally bonded CFRP
can be used to strengthen the web crippling capacity of stainless steel tubular structural members.

Cold-formed Steel Innovations 2

Previous research on strengthening of metal structures is mainly focused on carbon steel members as
summarized by Zhao and Zhang [1]. Experimental investigation on web crippling strengthening of
rectangular carbon steel tubes and light steel beam using FRP have been conducted by Zhao et al. [2],
Fernando et al. [3], and Zhao and Al-Mahaidi [4]. However, stress-strain behaviour of carbon steel and
stainless steel is quite different. Stainless steel materials have lower proportional limits than carbon steel
which may affect the buckling and web crippling behaviour of structural tubular members. Zhou and
Young [5-6] conducted a series of tests on cold-formed stainless steel tubes subjected to web crippling. The
test specimens were not strengthened by CFRP. Therefore, it is novel to strengthen stainless steel tubular
structural members using CFRP subjected to web crippling, and the strengthening only applied to a
localize area under the concentrated load.

The effectiveness of strengthening directly depends on the properties of adhesive and FRP. Debonding
failure is a critical issue for the strengthening of stainless steel tubular members. The cohesion, adhesion
and combined of these two failure modes generally occur in FRP strengthening of carbon steel tubular
members [3]. The properties of the adhesive fully control the cohesion failure, while adhesion failure
depends on the surface characteristics of the adherents including the texture, roughness and chemical
composition of the surface [7]. Different failure modes for FRP strengthened steel structures have been
reported [1-3, 7-8]. The investigations on the bonding behaviour of FRP to steel joints have been also
conducted [9-12]. However, little research has been reported on FRP to stainless steel bonding. Up-to-date,
extensive research has been conducted on FRP-strengthened concrete structures and carbon steel structures,
but little research on FRP-strengthened stainless steel structures. It should be noted that interaction of shear
stress and normal stress is experienced in FRP strengthening of members failed by web crippling [3]. It is
not easy to find an analytical model for strengthening of tubular members subjected to web crippling.
Therefore, experimental investigation is needed for a better understanding on such structural behaviour.

The purpose of this paper is to investigate the effects of different surface treatments and different widths of
CFRP plate strengthening of stainless steel tubular sections against web crippling failure. A series of tests
on CFRP strengthening of stainless steel tubular structural members subjected to End-Two-Flange and
Interior-Two-Flange loading conditions is conducted. The load-web deformation behaviour of the stainless
steel sections is presented. The failure loads and different failure modes are also presented in this study.

MATERIAL PROPERTIES

The material properties of the ferritic stainless steel tube specimens were determined by tensile coupon
tests. The flat tensile coupons were taken from the centre of the face at 90
o
angle from the weld for all
ferritic stainless steel tubes in the longitudinal direction. The tensile coupons were prepared and tested
according to the American [13] and Australian [14] standards for the tensile testing of metals using
12.5 mm wide coupons of gauge length 50 mm. The coupons were tested in a MTS displacement
controlled testing machine. Two strain gauges and a calibrated extensometer of 50 mm gauge length
were used to measure the longitudinal strain. A data acquisition system was used to record the load and
strain at regular intervals during the tests. The static load was obtained by pausing the applied
straining for 1.5 minutes near the 0.2% tensile proof stress and the ultimate tensile strength. This
allowed the stress relaxation associated with plastic straining to take place. The material properties
obtained from the tensile coupon tests are summarized in Table 1, which includes the static 0.2% tensile
proof stress (
0.2
), static tensile strength (
u
), initial Youngs modulus (E
o
), exponent of Ramberg-Osgood
3 Cold-formed Steel Innovations

expression (n), and elongation after fracture (
f
) based on a gauge length of 50 mm. The measured
stressstrain curves obtained from the tensile coupon tests were used to determine the parameter n in
the Ramberg-Osgood expression [15].

TABLE1:
MEASURED MATERIAL PROPERTIES OF FERRITIC STAINLESS STEEL SECTIONS OBTAINED
FROM TENSILE COUPON TESTS
Test Specimen
0.2

(MPa)
u

(MPa)
E
o

(GPa)
n
f

(%)
F50x50x4 504 514 202.0 6.4 11.9
F120x40x3 426 459 203.5 6.2 21.5

In the first stage of the tests, two different adhesives of Araldite 2015 (E) and Araldite 420 (F), and the
high modulus CFRP Sika CarboDur H514 laminate plate (f) were used to investigate the effect of
different surface treatment. In the second stage of the tests, the high modulus CFRP Sika CarboDur
H514 laminate plate and adhesives of Araldite 2015 (E) were used for the strengthening of the ferritic
stainless steel tubular sections. It was shown that the high modulus CFRP Sika CarboDur H514
laminate plate and adhesive Araldite 2015 provided the best strengthening performance for ferritic
stainless steel tubular sections subjected to web crippling [16]. The high modulus CFRP laminate plate
had the nominal modulus of elasticity of 300 GPa, ultimate tensile strength of 1500 MPa, tensile strain
at fracture of 0.45% and thickness of 1.4 mm. This high modulus CFRP Sika CarboDur H514 laminate
plate was symbolized as f, which is the same symbol used in [16]. Furthermore, adhesive Araldite
2015 was used for the strengthening of the ferritic stainless steel tubular sections in this study. The
adhesive Araldite 2015 had the measured tensile strength of 19.7MPa, modulus of elasticity of 1.8GPa
and tensile strain at fracture of 3.3% based on a gauge length of 50 mm coupon test. The coupon test
of the adhesive Araldite 420 and Araldite 2015 is detailed in [17].

TEST SPECIMENS AND LABELLING

A series of tests was conducted on strengthened ferritic stainless steel tubular members using CFRP to
increase the web crippling capacity. The ferritic stainless steel type EN 1.4003 test specimens were
used in this study. Twenty specimens were tested under the End-Two-Flange (ETF) and
Interior-Two-Flange (ITF) loading conditions. The specimen lengths L=N+1.5d and L=N+3d for ETF
and ITF loading conditions respectively were used, where N is the bearing length 50 mm and d is the
overall web depth. The measured dimensions of the test specimens subjected to End-Two-Flange
(ETF) loading are shown in Table 2 using the nomenclature defined in Figure 1. These specimens
included rectangular hollow sections (RHS) and square hollow sections (SHS) strengthened with
CFRP. Definition of symbols of the ferritic stainless steel hollow section specimens are shown in
Figure 1. The specimen lengths were determined according to the ASCE Specification [18].

Figure1. Definition of symbols of ferritic stainless steel test specimen for web crippling tests

In this study, firstly, two different surface treatments using electric sander (S) and grinder (G) were
considered for CFRP strengthened ferritic stainless steel sections. One layer of CFRP plate of 50 mm
is attached on the outer surface of both sides of the webs at the end of the specimen. Secondly, a study
was also conducted on behaviour of stainless steel members using different widths of CFRP plate
strengthening against web crippling. Three different widths of FRP plate were N, N+0.75d and
N+1.5d for ETF as well as N, N+1.5d and N+3d for ITF loading against web crippling have been
investigated. The measured dimensions of the different widths of CFRP plate strengthened test
specimens are presented in Table 3. It should be noted that in the investigation of different widths of
CFRP plate, one layer of Sika CarboDur H514 CFRP and using adhesive Araldite 2015 were attached
on both side of the webs. The length of strengthening is identical to the length of the bearing (N)
which is 50 mm. The CFRP plates of height (h=d-2t-2r
i
) that equal to the flat portion of the web of the
specimen were used. The web crippling behaviour is also depends on the web slenderness
(depth-to-thickness) ratio.

The specimens were labelled such that the material types, nominal dimensions of the specimen,
loading condition, type and number of CFRP layer, type of adhesive and type of surface treatment can
be identified, which is shown in Table 2. For example, the label F120x40x3-ETF-f1-E-S defines the
following specimen: The first letter indicates the material type of the specimen, where F refers to
ferritic stainless steel. The following symbols are the nominal dimensions of the specimen in mm,
where 120403 (dbt) means the web depth = 120 mm, flange width = 40 mm and thickness = 3
mm. The following three letters indicate the loading condition of End-Two-Flange (ETF). The
following two letters indicate the type and number of carbon fibre-reinforced polymer (CFRP) layer,
where f1 means Sika CarboDur H514 of one layer but the reference test specimen without CFRP is
represented as 0. The following letter indicates the type of adhesive, where E = Araldite 2015 as
shown in Table 2. The following letter indicates surface treatment, where S = sander surface treatment.
If a test was repeated, then -R indicates the repeat test.

Firstly, some of the tests were conducted on different surface treatment using sander (S) and grinder
(G). Generally, the surface treatment by sander seems suitable for ferritic stainless steel tubular
sections. The surface treatment by sander was then used for the preparation of strengthened specimens
to investigate the effects of different widths of CFRP plate against web crippling. In Table 3, the
method of surface treatment is not shown in the specimen labels, and the specimens were treated by
sander. The CFRP strengthened width of all specimens was 50 mm which is the same length as the
bearing length N as shown in Table 2. In addition, the effect of different widths of CFRP plate
strengthening was also investigated. In Table 3, the width of CFRP strengthening is represented by the
parenthesis followed by the type and number of CFRP layer. For example,
F120x40x3-ETF-f1(N+0.75d)-E, where (N+0.75d) means the CFRP width is equal to N+0.75d.


TABLE 2
MEASURED DIMENSIONS AND TEST RESULTS OF CFRP STRENGTHENED F120X40X3 SPECIMENS
USING DIFFERENT SURFACE TREATMENT SUBJECTED TO END-TWO-FLANGE LOADING
Specimen
d
(mm)
b
(mm)
t
(mm)
r
i
L
(mm)
h/t P
u

(kN)
P
u
/P
u0
Failure
mode
F120x40x3-ETF-0 120.0 40.2 2.831 4.0 230.0 37.6 20.6 1.00 W
F120x40x3-ETF-f1-E-G 120.0 40.0 2.828 4.0 230.1 37.6 25.6 1.24 A
F120x40x3-ETF-f1-F-G 120.1 40.1 2.789 4.0 230.4 38.2 26.2 1.27 A
F120x40x3-ETF-f1-E-S 120.0 40.1 2.796 4.0 230.2 38.1 31.2 1.51 A+C
F120x40x3-ETF-f1-F-S 120.0 40.0 2.790 4.0 229.5 38.1 25.7 1.25 I
Note: A= adhesion, C = Cohesion, A+C = combination (A+ C), I = Interlaminar CFRP failure, W= Web crippling.

SPECIMEN PREPARATION

The effectiveness of the strengthening is mostly dependent upon the quality of the bond between
stainless steel and adhesive interface. The bond strength between stainless steel and adhesive is one of
the important issues in the CFRP strengthening of stainless steel structures.

TABLE 3
MEASURED DIMENSIONS AND TEST RESULTS OF DIFFERENT WIDTHS OF FRP PLATE
STRENGTHENED FERRITIC STAINLESS STEEL TUBULAR SECTIONS SUBJECTED TO
END-TWO-FLANGE LOADING AND INTERIOR-TWO-FLANGE LOADING CONDITIONS
Specimen
d
(mm)
b
(mm)
t
(mm)
r
i
L
(mm)
h/t P
u

(kN)
P
u
/P
u0
Failure
mode
F50x50x4-ETF-0 50.3 50.2 3.848 4.0 125.4 9.0 39.4 1.00 W
F50x50x4-ETF-f1(N)-E 50.1 50.2 3.842 4.0 125.9 9.0 43.8 1.11 A
F50x50x4-ETF-f1(N+0.75d)-E 50.0 50.0 3.843 4.0 126.0 8.9 44.3 1.12 A+C
F50x50x4-ETF-f1(N+1.5d)-E 50.0 50.1 3.831 4.0 125.2 9.0 44.6 1.13 A+I
F50x50x4-ITF-0 50.0 50.1 3.832 4.0 200.8 9.0 76.1 1.00 W
F50x50x3-ITF-f1(N)-E 50.1 50.0 3.829 4.0 200.5 9.0 77.7 1.02 A
F50x50x4-ITF-f1(N+1.5d)-E 50.0 50.1 3.840 4.0 200.3 8.9 78.3 1.03 A+C
F50x50x4-ITF-f1(N+3d)-E 50.0 50.1 3.832 4.0 200.6 9.0 78.5 1.03 A+C
F120x40x3-ETF-0 120.0 40.2 2.831 4.0 230.0 37.6 20.6 1.00 W
F120x40x3-ETF-f1(N)-E 120.0 40.2 2.796 4.0 229.9 38.1 31.1 1.51 A+C
F120x40x3-ETF-f1(N+0.75d)-E 120.0 39.9 2.823 4.0 229.8 37.7 31.8 1.54 A+C
F120x40x3-ETF-f1(N+1.5d)-E 120.0 39.9 2.822 4.0 230.3 37.7 32.0 1.55 A+C
F120x40x3-ITF-0 119.8 40.2 2.814 4.0 410.3 37.7 45.4 1.00 W
F120x40x3-ITF-f1(N)-E 120.0 39.9 2.788 4.0 410.9 38.2 49.3 1.09 A
F120x40x3-ITF-f1(N+1.5d)-E 120.1 40.0 2.816 4.0 410.8 37.8 49.9 1.10 A+C
F120x40x3-ITF-f1(N+3d)-E 120.0 40.0 2.810 4.0 410.1 37.9 50.4 1.11 A+C
Note: A= adhesion, C = Cohesion, A+C = combined, I = Interlaminar FRP failure, W= Web crippling.


Figure 2. Ferritic stainless steel sections treated by electric sander and grinder

The strength of the adhesive bond is directly proportional to the quality of the surfaces to which it is
mated [19]. Generally, grinding, sand paper or sand blasting surface treatments are used of metal
surface treatment for FRP strengthening [2-3, 7, 16-17, 19-20]. In this study, two different surface
preparation techniques of electric sander and electric grinder surface treatments were used to find the
effective surface treatment method for CFRP strengthened ferritic stainless steel sections. An electric
grinder with 11000 rpm 710w or a BOSCH (PSS-23) sander was used around 2-3 minutes for
treatment of each surface of ferritic stainless steel sections. The silicon carbide (80Cw-medium grit)
water proof electro coated abrasive/sand paper was used in the electric sander. After surface treatment,
the specimens were blown by air-gun to remove all rust from the treated surfaces. Figure 2 shows the
surfaces of the specimens treated by the two different methods.

The adhesive was then applied uniformly on the CFRP plate and the ferritic stainless steel tube
surfaces for the required bonding length. The excess adhesive and air were removed using a ribbed
roller and applied on the CFRP plate in the direction of the fibre with a small amount of force. The
thickness of the adhesive layer was maintained uniform. The adhesive thickness for most of the
specimens strengthened by CFRP plates was found to be around 1.0 mm. CFRP plates were bonded to
the webs only at the end and middle of the specimens for ETF and ITF loading, respectively. Figure 3
shows the strengthened test specimens using CFRP plate width N, N+0.75d and N+1.5d respectively.
The fibre direction was along the web direction. The test specimens were tested after 7 days of curing
at room temperature.

Figure 3. Test specimens of different widths CFRP plate strengthening
CFRP Width N
CFRP Width
N + 0.75d

N
CFRP Width
N + 1.5d

N
Treated stainless steel
surface by sanding
Treated stainless steel
surface by grinding

WEB CRIPPLING TESTS

The web crippling tests of ferritic stainless steel sections were carried out under the two loading
conditions specified in the ASCE Specification [18]. The specimens were tested under the
End-Two-Flange (ETF) and Interior-Two-Flange (ITF) loading conditions. The photograph of the test
setup and failure mode under ETF loading condition is shown in Figures 4 and 5, respectively. Two
identical bearing plates with half round of the same width were positioned at the end and middle of the
specimens for ETF and ITF loading conditions, respectively. Hinge supports were simulated by two
half rounds. The specimen was seated between the two bearing plates during the test. The thickness of
each bearing plate was 50 mm. The bearing length (N) was chosen at 50 mm for all tests. The flanges
of the specimen were not bolted to the bearing plates. Hence, the flanges are unrestrained. Four
transducers (LVDTs) were used to record the web deformations of the specimen as shown in Figure 4.
The web deformations of the specimen were obtained by the average readings of the four transducers
measured between the two bearing plates. A servo-controlled hydraulic testing machine was used to
apply a concentrated compressive force to the test specimen. Displacement control was used to drive
the hydraulic actuator at a constant speed of 0.3 mm/min.

Figure 4. Test setup of End-Two-Flange loading
condition
Figure 5. Failure mode under ETF loading

TEST RESULTS AND DISCUSSIONS

It is shown that debonding failure is a critical issue for CFRP strengthening of ferritic stainless steel
tubular members. The reference tests without the CFRP plate strengthening were failed by web
crippling. The failure modes of each test specimen are listed in Table 2. The failure mode of specimen
F120x40x3 strengthened by CFRP plate is shown in Figure 5 under ETF loading condition. The webs
of the specimens were buckled outward. Four main failure modes were observed in the tests, namely
the adhesion, combination of adhesion and cohesion, interlaminar failure of CFRP plate, and
combination of adhesion and interlaminar failure of CFRP plate. The adhesion failure was found at
physical interface between the adhesive and the adherents, whereas the cohesion failure is fully control
by the adhesive properties. The interlaminar CFRP failure was found when localized adhesive bonding
strength is strong between CFRP and ferritic stainless steel tubular surface, whereas the adhesive is not
able to mobilize stress distributions.
LVDT
Half Round
CFRP
Bearing Plate
Steel Specimen
Half Round

Figure 6. Comparison of different surface treatment effects on load-web deformation behaviour of CFRP
strengthened F120x40x3 specimens

Figure 7. Comparison of different widths of CFRP plate strengthening effects on load-web deformation
behaviour of F120x40x3-ETF-E specimens
0
5
10
15
20
25
30
35
0 1 2 3 4 5 6
Web deformation (mm)

L
o
a
d

(
k
N
)
F120x40x3-ETF-0
F120x40x3-ETF-f1-E-G
F120x40x3-ETF-f1-F-G
F120x40x3-ETF-f1-E-S
F120x40x3-ETF-f1-F-S
F0
E-S
F-G
F-S
E-G
0
5
10
15
20
25
30
35
0 1 2 3 4 5 6
L
o
a
d

(
k
N
)
F120x40x3-ETF-0
F120x40x3-ETF-f1(N)
F120x40x3-ETF-f1(N+0.75d)
F120x40x3-ETF-f1(N+1.5d)
N + 1.5d
N + 0.75d
F0
N

It may also caused by low bonding strength in the fibre of CFRP plate. The initial cracking is normally
started at the end of the CFRP plate that experienced high interfacial stresses developed in the region.
Debonding process is propagated gradually towards the mid-height of the webs after the appearance of
the first crack at the end of the plates. Figures 6-8 show the graphical presentations of load-web
deformation behaviour of CFRP strengthened ferritic stainless steel specimens by using different
widths of CFRP plate. Most of the specimens were failed due to debonding. The initial cracking
propagated to a certain limit then the CFRP debonded suddenly and the load dropped immediately.

The experimental ultimate web crippling loads per web with CFRP (P
u
) and without CFRP (P
u0
) are
presented in Tables 2 and 3. The specimens without strengthening of CFRP were tested as reference
tests and these specimens were labeled using a suffix of -0 as shown in Tables 2 and 3 as well as
labeled as F0 in Figures 6-8. The experimental ultimate web crippling loads per web with CFRP (P
u
)
and without CFRP (P
u0
) are presented in Tables 2 and 3. In the first stage of the tests, two different
surface treatments of electric sander (S) and electric grinder (G) were considered on the section
120x40x3. Two different adhesives of Araldite 2015 (E) and Araldite 420 (F), and the high modulus
CFRP Sika CarboDur H514 laminate plate (f) were used. Table 2 and Figure 6 show that the sander
surface treatment seems suitable for ferritic stainless steel tubular sections in terms of peak load
enhancement.

Figure 8. Comparison of different widths of FRP plate strengthening effects on load-web deformation behaviour
of F120x40x3-ITF-E specimens

A series of tests was also performed on ferritic stainless steel sections to investigate the effects of
different widths of CFRP plate strengthening against web crippling. Three different widths of CFRP
plate strengthening were considered. The different widths were N, N+0.75d and N+1.5d for
End-Two-Flange loading as well as N, N+1.5d and N+3d for Interior-Two-Flange loading, where N is
the bearing length which is 50 mm and d is the depth of the sections. The investigation was conducted
on two different ferritic stainless steel sections 50x50x4 and 120x40x3. The adhesive Araldite 2015 (E)
0
10
20
30
40
50
60
0 1 2 3 4 5 6
L
o
a
d

(
k
N
)
F120x40x3-ITF-0
F120x40x3-ITF-f1(N)
F120x40x3-ITF-f1(N+1.5d)
F120x40x3-ITF-f1(N+3d)
N + 3d
N + 1.5d
N
F0

and the high modulus Sika CarboDur H514 laminate plate (f) were used to strengthen the stainless
steel sections, since they provided better performance compared to other adhesives and FRPs. The test
results of different widths of CFRP plate strengthened ferritic stainless steel hollow sections subjected
to ETF and ITF loading conditions are shown in Table 3. The load-web deformation behaviour of
these sections are plotted in Figures 7 and 8 for ETF and ITF loading conditions, respectively. The
web crippling strengths increased by 1.8% and 2.9% for CFRP width increases from N to N+1.5d for
stainless steel sections 50x50x4 and 120x40x3 subjected to ETF loading, respectively as shown in
Table 3. The web crippling strengths increased by 1.0% and 2.2% for CFRP width increases from N to
N+3d for stainless steel sections 50x50x4 and 120x40x3 subjected to ITF loading, respectively. It is
found that the web crippling capacity of the ferritic stainless steel tubular sections increased slightly
by increasing the width of the CFRP plate from 50 mm onwards. The increases of CFRP width do not
provide much improvement on strengthening of the tested ferritic stainless steel sections. Therefore,
the width of the strengthening CFRP plate of 50 mm seems adequate for the ferritic stainless steel
tubular sections. The CFRP strengthening against web crippling is found to be effective when the
CFRP width having the same width as the bearing length (N).

CONCLUSIONS

A test program on carbon fibre-reinforced polymer (CFRP) strengthening of ferritic stainless steel
tubular structural members subjected to web crippling has been presented. A series of tests was
performed to investigate the effects of different surface treatment, and different widths of CFRP plate
for the strengthening of stainless steel tubular sections against web crippling failure. Two different
surface treatments using sander and grinder were considered. The sander surface treatment seems
suitable for ferritic stainless steel tubular sections in terms of peak load enhancement. The failure
loads, failure modes and the load-web deformation behaviour of the ferritic stainless steel sections
have been reported. The influence of different widths of CFRP plate strengthening against web
crippling subject to End-Two-Flange and Interior-Two-Flange loading conditions was also investigated.
It is shown that the increases of CFRP width do not provide much improvement on the strengthening
of ferritic stainless steel sections. Hence, the CFRP strengthening against web crippling is found to be
effective when the CFRP width having the same width as the bearing length for ferritic stainless steel
tubular sections.

REFERENCES

[1] Zhao X.L, Zhang L.: State-of-the-art review on FRP strengthened steel structures. Engineering
Structures, Vol.29, pp.1808-1823 (2007).
[2] Zhao X. L., Fernando D., Al-Mahaidi R.: CFRP strengthened RHS subjected to transverse end
bearing force, Engineering Structures, Vol. 28(11), pp.1555-1565 (2006).
[3] Fernando D., Yu T., Teng J. G., and Zhao X. L.: CFRP strengthening of rectangular steel tubes
subjected to end bearing loads: Effect of adhesive properties and finite element modelling, Thin-
Walled Structures, Vol.47(10), pp. 1020-1028 (2009).
[4] Zhao X. L., Al-Mahaidi R.: Web buckling of lightSteel beams strengthened with CFRP
subjected to end bearing forces, Thin-Walled Structures, Vol. 47(10), pp. 1029-1036 (2009).
[5] Zhou F. and Young B. Cold-formed stainless steel sections subjected to web crippling. Journal
of Structural Engineering, ASCE; Vol.132(1),pp.13414 (2006) .

[6] Zhou F. and Young B. Experimental and numerical investigations of cold-formed stainless steel
tubular sections subjected to concentrated bearing load. Journal of Constructional Steel
Research Vol.63(11), pp.14521466 (2007)
[7] Schnerch D.: Strengthening of steel structures with high modulus carbon fiber reinforced
polymer (CFRP) materials. PhD dissertation, North Carolina State University, Raleigh (NC),
(2005).
[8] Silvestre N., Young B. and Camotim D. Non-linear behaviour and load-carrying capacity of
CFRP-strengthened lipped channel steel columns. Engineering Structures, Vol.30(10),
pp.2613-2630 (2008).
[9] Xia S.H, and Teng J.G. Behaviour of FRP-to-steel bonded joints, Proceedings of the
International Symposium on Bond Behaviour of FRP in Structures, Hong Kong, China, 2005,
pp. 411418.
[10] El Damatty, A.A., and Abushagur M. Testing and modeling of shear and peel behavior for
bonded steel/FRP connections. Thin-Walled Structures, Vol. 41(11), pp. 987-1003 (2003).
[11] Dawood M. and Rizkalla S. Bond and splice behaviour of high modulus CFRP materials bonded
to steel structures. Third International Conference on FRP Composites in Civil Engineering,
Miami, Florida, USA. 2006,pp.705-708.
[12] Fawzia S., Zhao X.L., and Al-Mahaidi, R. Bondslip models for double strap joints strengthened
by CFRP. Composite Structures , Vol. 92(9), pp.2137-2145(2010).
[13] ASTM. Standard test methods for tension testing of metallic materials, E 8M-97. West
Conshohocken: American Society for Testing and Materials; 1997
[14] AS. Metallic materials Tensile testing at ambient temperature. Australian Standard, AS
1391-2007. Sydney, Australia: Standards Association of Australia; 2007
[15] Ramberg W. and Osgood W.R. Description of stressstrain curves by three parameters.
Technical note No. 902, National Advisory Committee for Aeronautics; 1943.
[16] Islam S. M. Z. and Young B. Effects of different adhesive and FRP on strengthening of
stainless steel tubular structural members , Proceedings of the 13th International Symposium on
Tubular Structures, Hong Kong, China,2010.pp.273-280.
[17] Islam S. M. Z. and Young B. FRP Strengthened aluminium tubular sections subjected to web
crippling, Thin-Walled Structures, Vol. 49(11):1392-1403 (2011).
[18] ASCE. Specification for the design of cold-formed stainless steel structural members. SEI/ASCE
8-02; Reston, VA: American Society of Civil Engineers; 2002.
[19] Fawzia S., Al-Mahaidi R., Zhao, X.L. and Rizkalla S. Strengthening of circular hollow steel
tubular sections using high modulus CFRP sheets. Construction and Building Materials,
Vol.21(4), pp. 839-845(2007).
[20] Pantelides C.P., Nadauld J. and Cercone L. Repair of cracked aluminum overhead sign
structures with glass fiber reinforced polymer composites, Journal of Composites for
Construction, ASCE, Vol. 7(2), pp. 118-126 (2003).
ISBN 978-983-44176-4-2

STRESSED SKIN EFFECTS ON COLD-FORMED STEEL PORTAL
FRAMES WITH SEMI-RIGID JOINTS

A.M. Wrzesien
1
, J.B.P. Lim
2
and R.M. Lawson
3

1
Department of Civil Engineering and Environmental Engineering, University of Strathclyde,
Glasgow, UK
2
SPACE, Queens University, Belfast, UK
3
Department of Civil Engineering, University of Surry, Guildford, UK

KEYWORDS

Cold-formed steel, portal frames, stressed skin action, semi-rigid joint, rotational stiffness

ABSTRACT

This paper describes full-scale portal frame tests, both with and without sheeting, conducted on a cold-
formed steel portal frame having semi-rigid joints of different stiffness. For such frames, bolts are
commonly used not only through the web of the sections, but also through the flanges. The tests cover
loading both the downward and horizontal direction. As expected, the effect of the sheeting on frame
deflections is small for loading in the downward direction, but the effect of sheeting on deflection
under horizontal loading has a large effect. From the results of these tests, it is apparent that stressed
skin design can be used effectively to minimise horizontal deflections on cold-formed steel portal
frames, when suspect to wind action.

INTRODUCTION

Previous studies have demonstrated the viability of using cold-formed sections to construct portal
frames having spans of up to 12m (Baigent and Hancock (1982), Kirk (1986), De Vos and Van
Rensburg (1997), Mills and LaBoube (2004). The key feature of all these studies was eaves and apex
joints that could be designed to function as being effectively rigid in bending. Such joints, however,
were either expensive to manufacture or difficult to erect on site. A review of these systems, including
detail of the joints, can be found in Wrzesien et al. (2011)

- 13 - Cold-formed Steel Innovations

In recent years, a number of researchers have concentrated on quantifying the rotational stiffness of
the eaves and apex joints (Mkelinen and Kankaanp (1996), Dubina et al. (2004), Kwon et al.
(2006)). Increasing the rotational stiffness of the joints will be beneficial for the overall response of
the frame, particularly in terms of reducing frame serviceability deflections. However, research
focussing on the beneficial effect of stressed skin action will also have a significant benefit by
reducing frame deflections, albeit only in the horizontal wind load case.

Lim and Nethercot (2004) and Chung et al. (2005) independently conducted research on determining
the rotational stiffness of cold-formed portal frame joints and applying the measured rotational
stiffness to frame analysis and design. The key features were the use of the largest cold-formed steel
channel sections that can currently be rolled for the column and rafter members, and the use of bolted
moment-resisting connections, formed through brackets, for the eaves and apex joints (see Figure 1).
Such an arrangement for the joints was chosen to ensure that the brackets remained easy to
manufacture and the joints easy to assemble on site. However, using the Eurocode 3 classification for
the non-dimensional stiffness of the joints, K
j
, in which a value of zero corresponds to a pinned joint
and a value of 25 corresponds to a rigid joint, the joints tested had values of K
j
ranging between 2.5
and 5. As the joints are therefore relatively flexible, serviceability deflections needed to be determined
and taken into account in frame design. Interestingly, however, the redistribution of bending moment
as a result of the semi-rigidity of the joints resulted in a semi-rigid cold-formed steel portal frame
being able to carry higher load than an equivalent rigid jointed cold-formed steel portal frame.

(a) Eaves joint (b) Apex joint

Figure 1 Details of the joints adopted for the cold-formed steel portal frame after Lim and Nethercot
(2003)

This paper instead focuses on the effect of stressed skin action on frames having flexible joints.
Stressed skin action is a well established phenomenon, based on taking into account the inherent
strength and stiffness of the metal cladding in overall frame behaviour. It was demonstrated by
extensive research that stressed skin action can reduce or eliminate the need for wind bracing, reduce
sway deflections under horizontal forces, and reduce the spread of the frame under vertical load.
Design recommendations was first presented in the Manual of stressed skin diaphragm design by
Davies and Bryan (1982), which was then incorporated into BS 5950-9 (1994), and the ECCS (1995).

Stressed skin design was originally investigated for traditional hot-rolled steel portal frames. In recent
years, for frames having spans of around 12 m, the entire frame is often built from cold-formed steel
members. These types of structures are often very flexible, and usually suffer from extensive sway
Cold-formed Steel Innovations - 14 -

deflections. In such cases, implementing the stressed skin action in their analysis can offer greater
benefits than for hot-rolled steel frames and permits load redistribution between adjacent frames.

Roof systems, are also consistently evolving. The typical connection detail for purlin to rafter
connection includes C or Z purlins connected to the rafters through a web cleat (see Figure 2). Such an
arrangement has relatively low stiffness against shear deformation. On the other hand, the use of
modern top-hat shaped purlins can simplify the connection detail, improve purlin to rafter
connection flexibility, and therefore increase the diaphragm stiffness.

Figure 2 Different purlin to rafter connection details

This paper uses a combination of full-scale tests and finite element beam idealisation to assess the
beneficial effect of stressed skin action of frames with various types of joints.

Full-scale panel tests are used to determine the stressed skin diaphragm stiffness of the sheeting. Full-
scale joints test are then used to determine the rotational stiffness of three different bolt-group
arrangements for the joints. The rotational stiffness of the base connection is also established
experimentally. Unlike the joints described by Lim and Nethercot (2003), where the bolts are only
used in the web of the channel sections, the bolt-group arrangements included bolts in the flanges as
well as in the web. Finally, full-scale frame tests are described covering both load applied in the
vertical direction as well as the horizontal direction. For the case of load applied in the horizontal
direction, frames are tested with and without sheeting. The initial displacement caused by members
misalignment and bolt-holes tolerance is also considered.

ROOF PANEL TESTS

Test Arrangement

Unlike typical Z or C type purlins recognized by current design code, modern hat-shaped purlins were
investigated in this study. The difference in both types of purlin to rafter connections is presented in
Figure 2.

Typical cantilever test arrangement was used (according to BS 5950-9) in order to investigate strength
and stiffness of different roof panels (see Figure 3). Testing program included 19 full-scale tests with
sheeting panels of different geometry and structural form, such as single skin, double skin with
insulation or rigid sandwich panels. The test results were reported by Wrzesien et al. (2009).


Figure 3 Plan view of the cantilever test arrangement

According to the information provided by the manufacturer the most popular type of cladding panel
denoted as AS 30/0.7 mm was chosen for further frame tests. Only this type of cladding will be
presented in this paper (see Figure 4). It was rolled from steel coil of 0.65mm net thickness, with 280
N/mm average yield strength and ultimate tensile strength of 378 N/mm. Self-drilling self-tapping
screws were used, as follows:
- 5.5 mm diameter with washer and seal for fixing the cladding to the purlins
- 6.3 mm diameter with and without washer for seam and purlin to rafter connection
respectively.

In every case, the same hat-shaped purlins were used, rolled from high tensile steel S550GD of
average measured yield strength of 635 N/mm and 1mm thickness.

Figure 4 Roof cladding profile AS 30/07

The panels displacement was measured by linear displacement transducers and overall deflection was
calculated from the formula:
=
1

2
[(a/b)(
3

4
)] (1)
where:
1,,4
- displacement of the four corners (see Figure 3)
a and b dimensions of the panel.

Theoretical analysis

The ultimate strength and flexibility of investigated cantilever diaphragms were calculated according
to design rules published by Davies & Bryan (1982).
The design sheer strength of individual fixings was calculated to BS 5950: Part 5, as follows:
P
bs
= 2.1dt
net
p
y_nom
(2)

where:
d - fixing diameter
t
net
- minimum net thickness of connected material excluding coating.

Conservative values for the flexibility of different types of fixing were assumed according to Davies
and Bryan (1982) and are presented in Table 1.

TABLE 1
MECHANICAL PROPERTIES ASSUMED IN CALCULATIONS
Fixing type
Strength
(kN/mm sheet thickness)
Flexibility
(mm/kN)
5.5 mm diameter screws
(sheet fixings)
3.2 0.25
(seam and shear connector
fixings)
3.7 0.30
(purlin to rafter connection)
7.3 0.50

Test results

In order to establish experimental strength and stiffness of the roof panel, a testing procedure similar
to this described in BS 5950-9 was used. The panel was subject to 4 load-unload cycles and load
versus deflection relationship was recorded each time. First stage was bedding down load when the
panel was subject to 7 kN load. Then non-destructive test (acceptance test) was carried out in order to
establish structural performance of the panel. The residual deflection in this test was 19.4% of the
maximum recorded and it was not exceeding standard 20% requirement. Later, equivalent of the
standard strength test was carried out for the load of approximately 40% higher than acceptance test.
In this test the residual strength was 15% of the maximum recorded. The last test was the failure test
in which the panel reached an ultimate test load of 28.2 kN, as shown in Figure 5. The mode of failure
for this panel was the combination of end sheet to purlin connection failure and seam failure, as
presented in Figure 5.

Based on failure test the design shear capacity and design shear flexibility of the panel (according to
Section 11.4 BS 5050-9) was calculated from:
V
des
= 0.9 (p
y_nom
/ p
y_avr
) V
ult
(3)
where:
p
y_nom

- nominal yield strength
p
y_avr
- average yield strength obtain from coupon test
V
ult
- ultimate test load


Figure 5 Load-deflection curve for AS 30/07 cladding profile

c
des
= / 0.6 V
des
(4)
where:
mean deflection of the panel at the load of 0.6 V
des

The results are presented in Table 1 and compared against shear strength and stiffness calculated from
Davies and Bryan (1982).

TABLE 2
COMPARISON OF EXPERIMENTAL AND THEORETICAL RESULTS FOR SHEAR ON ROOF PANEL
Cladding type

Experimental Theoretical
Ultimate test
load F
ult
(kN)
Design capacity
V
des

(kN)
Design flexibility
c
des

(mm/kN)
Design
Strength

(kN)
Design
flexibility
(mm/kN)
AS 30/0.7
28.2 22.7 0.53 10.5 0.66

JOINT TESTS

Details of tests

In this Section, tests to determine the strength and stiffness of three different types of bolt-group
arrangements were presented. The full-scale tests, described in the next section of this paper, will use
combinations of these joints for the eaves, apex and base joints. Figure 6 summarises the geometry of
each type of joint. As can be seen, combinations of bolts in the web and flange are considered along
with the bolt-group length.

Laboratory tests were conducted on 6 joints, as well as back-to-back continuous member. Details of
the laboratory test set-up are shown in Figure 7. As can be seen, each joint test comprises two
identical bolt-group arrangements, one on either side of the vertical axis of symmetry, that are tested
under four-point bending. For all joint tests, the total length of the test specimen is 3 m and the
distance from the end support to the load point is 1 m. To prevent lateral-torsional buckling, lateral
restraints were provided at the supports, load points and at the mid-span.


The nominal thickness of each bracket was 3 mm, and the nominal diameters of the bolts and bolt-
holes were 16 mm and 18 mm, respectively; all the bolts used were M16 Grade 8.8 and had fully
threaded bolt-shanks.

(a) Joint type A bolts only in web

(b) Joint type B bolts in web and top flange

(c) Joint type C base connection

Figure 6 Details of the test joints

Figure 7 Details of the joint test arrangement

The dimensions corresponding with Figure 6 are presented in Table 3 for each type of joint. Joint type
A is similar to this investigated by Chung and Lau (1999) and Lim and Nethercot (2003), with bolt
only in the web of the section. The test results from joint type A will be used as benchmark for clear
comparison between joints type B, in which bolts are also present in one of the flange. Such
arrangement allows the bracket to be press break without necessity of welding. It also increases
buckling capacity of the bracket by introducing additional stiffener.

TABLE 3
SUMMARY OF THE GEOMETRY OF DIFFERENT JOINTS
Test/
/Type/
downward or
uplift
Bolt
array
web
Bolts
flange
a
bg

(mm)
b
bg
(mm)
a
b
(mm)
a
eb
b
ebb
b
ebt
(mm)
a
es
b
es
(mm)
T1/A/d
&
T2/A/u
3x3 N/A 160 80 530 35 59 26 35 36
T3/B/d
&
T4/B/u
3x3 3 160 80 530 35 59 39 35 36
T5/B/d 3x3 3 280 80 760 35 59 39 35 36
T6/C/d 3x3 N/A 160 80 250 35 36 36 35 36

For the joint type B two different lengths of the bolt-group were investigate in order to introduce
different rotational stiffness values. Joints type A and type B were tested under downward and uplift
loading in order to compare experimental values of strength and stiffness in both directions of load.

TABLE 4
SUMMARY OF COMPONENT GEOMETRY
Test
designation
Average component dimensions (mm)
Section

Bracket
Type A,C Type B

D B C t
cor
D
b
B
b
t
cor,b

Nominal 152.0 64.0 20.0 1.96 165.0/178.0/250 65.0 2.96
T0 152.3 64.1 19.9 1.97 N/A
T1/A/d 152.7 64.6 20.0 2.00 165.4 65.8 2.98
T2/A/u 152.3 64.9 20.1 1.98 165.0 66.8 2.98
T3/B/d 152.2 64.6 20.3 1.98 179.4 65.4 2.98
T4/B/u 152.6 64.4 20.2 1.98 179.9 65.8 2.98
T5/B/d 152.7 65.2 19.9 2.01 179.6 65.4 2.98
T6/C 152.7 65.3 19.9 2.03 251.3 60.1 3.01

In all tests, the measured dimensions of the back-to-back channel-sections as well as connection
bracket are shown in Table 4. Tensile coupons were collected from each channel section and
mechanical properties of steel were established. The average yield strength as well as average ultimate
tensile strength is presented in Table 5.

The bolts in each test were only finger-tightened without pre-stress being applied. Load cycles to
remove the slack from joint were not conducted as any bolt-hole elongation so caused would not be
recoverable. However, before each test, the bracket was pulled as far as possible in the direction of the
applied load; at this position the displacement transducers used to measure deflections were set to
zero. Bedding down change of angle due to slip and bolt-holes tolerance was also measured before
each test by digital inclinometer.

Analytical methods of predicting ultimate strength of tested joints

Two main design criterions have to be considered when designing cold-formed steel joints. First of all,
bearing capacity of the thin plate can not be reached and local buckling of the plate under concentrated
forces from bolts must be prevented.

Although design codes offer guidance on calculating bearing capacity there is no guidance regarding
analysis method for preventing buckling mode of failure. The following analytical methods were
reported in the literature:
1) Lim and Nethercot (2003) conducted extensive parametric study based on finite element shell
models of the joint,
2) Chung and Lau (1999) proposed design method based on using a quadratic interaction
equation for combination of shear and bending,
(M
Ed
/M
c,Rd
)
2
+ (V
Ed
/V
c,Rd
)
2
1 (5)
3) Dubina et al. (2008) proposed design method based on interaction equation for combination of
bending and web crippling
(M
Ed
/M
c,Rd
)

+ (F
Ed
/V
w,Rd
)

1.25 (6)

The strength of each tested joint was therefore examined according to BS EN 1993-1-3 (2006) method
2 and 3; was converted to maximum forced based on design resistance of the joint and is presented in
Table 5. Letter B denotes buckling resistance of critical component,

BSF and BWC denote bending and shear force and bending and web crippling respectively and T
represent maximum test force. In test T0 the resistance of the continuous back-to-back member was
calculated using effective width method. Standard effective width method was also used to calculate
buckling capacity of the vertical leg of the back-to-back angle bracket in test T1/A/d under pure
bending. Internal forces distribution was based on elastic design when the centre of the rotation of the
bolt-group coincides with its centre of gravity. The contribution of the flange bolts was neglected in
the analysis thus only web bolts were considered for joint type A and B as shown in Figure 8.


TABLE 5
EXPERIMENTAL RESULTS VERSUS DESIGN RESISTANCE CALCULATED TO EC3
Test
designation
f
yb
f
u
F
Rd,B
F
Rd,BSF
F
Rd,BWC
F
T
=F
Rd
/F
T

N/mm
2
N/mm
2

kN kN kN kN
T0 509 547 50.12 N/A N/A 43.70 1.15

T1/A/d
(s
s
=230mm)
377 486 7.80 N/A N/A 19.76 0.39

T2/A/u
(s
s
=230mm)
509 547 50.72 35.11 27.24 33.36 0.82
T3/B/d
(s
s
=230mm)
509 545 50.72 35.07 27.22 36.33 0.75
T4/B/u
(s
s
=230mm)
507 542 50.64 35.01 27.14 32.46 0.84
T5/B/d
(s
s
=350)
501 541 51.36 42.36 37.29 40.61 0.92


Figure 8 Force distribution within the joint considering single channel section

Analytical methods of predicting the rotational stiffness of tested joints

TABLE 6
ROTATIONAL STIFFNESS OF JOINTS
Test
designation
l
b
(mm)
a
bg
(mm)
b
bg

(mm)
k
j

(kN/mm)
S
j,ini
(kNm/rad)
S
j,ini,exp
(kNm/rad)
T1/A/d 145 160 80 10.48 503 638
T2/A/u 145 160 80 10.41 500 616
T3/B/d 145 160 80 10.42 500 (385) 601
T4/B/u 145 160 80 10.43 501 (171) 591
T5/B/d 205 280 80 10.54 1340 (547) 1229
( ) bedding down stiffness of the joint

The rotational stiffness of the joint is calculated based on linear stiffness of single screw lap joint. The
centres of the rotation for the bolt-group is assumed to lay at the centre of gravity and the flange bolts
contribution to stiffness have been neglected. Since the experimental values for linear stiffness of
single bolt lap joint were not available the analytical equation developed by Zadanfarrokh and Bryan
(1992) was used instead.
k
j
=5n (10/t
1
+ 10/t
2
-2) 10
-3
(mm/kN) (7)

where:
t
1
, t
2
thicknesses of the sheet of steel (t
1
and t
2
8mm)
n - flexibility factor

Theoretical values of rotational stiffness for each test are presented in Table 6 and compared against
rotational stiffness obtained from the experiments.

Beam idealisation of the test arrangement

A simple beam model was employed to predict a deflection of cold-formed steel beam connected
through bolted joint (see Figure 9). The flexural stiffness of the members as well as the bracket was
modelled assuming gross section properties. The position of the centre of rotation was established and
the rotational stiffness of the bolt-group was modelled as a rotational spring element.

Figure 9 Beam idealisation of test arrangement

The same calculation model was also used to calculate experimental rotational stiffness of the bolt-
group based on total deflection
t
. In order to eliminate the portion of deflection due to flexure of the
member (
f
), a continuous beam was also tested in order to validate theoretical strength and stiffness
of the cold-formed steel member. The rotational stiffness of the joint was represented as a bi-linear
moment-rotation relationship. The assumption was made that during bedding down stage until all the
bolts are in contact with steel plate, the stiffness is close to zero. The maximum slip rotation
s
was
calculated based on the length of the bolt-group and the assumption that bolt can freely move from the
centre of the bolt hole to its edge. For standard 2 mm clearance holes in section as well as bracket
(joint type A) maximum rotation
s
can reach 0.0224 rad in both directions. In case of joints type B,
the contribution of the stiffener was recognised and free rotation
s
was reduced by half.

Test results

As could be seen in Table 5, the design resistance of continues member calculated to EC3 was 15%
higher than obtained in the test. The back-to-back channel beam failed due to distortional buckling.
The difference between theoretical and experimental flexural stiffness was also observed, as could be
seen in Figure 10. In further modelling, the second moment of inertia was therefore reduced by the
factor of 0.772.

In case of test T1/A/d, the design resistance of the back-to-back angle bracket under pure bending
according to BS EN 1993-1-3 (2006) was only 40% of failure load. As expected, the bracket of the
tested joint failed prematurely due to local buckling of the unstiffened plate.

In case of all other tests, the failure occurred in the channel section due to local buckling of the web
followed by flange deformation. The uplift test for joint type A and B with the same bolt-group length
presented very similar failure load. This supports the assumption that the flange bolts contribute very
little to overall resistance of the joint.

By increasing the length of the bolt-group from 160 mm to 280 mm a 12% increase in failure load was
observed, but none of joints however could be considered as full-strength joints. The longest bolt-
group joint resisted a bending moment corresponding to 93% of continuous beam bending capacity.

Due to the fact that combine bending and shear force interaction equation (BSF) was not always
producing conservative design resistance, combine bending and web crippling (BWC) interaction
equation was used for further theoretical analysis.

Combined bending and web crippling design criterion produced more conservative results for shorter
bolt-group but produced very close agreement with experimental data for 280mm long bolt-group.

As could be observed in Figure 10 and Table 6, good agreement between theoretical and experimental
stiffness was observed. In Figure 10, the theoretical and experimental strength and stiffness of the
continuous member was superimposed onto a plot for comparison. The C-shape bracket in the joint
type B prevented the test assembly from initial sagging due to self-weight. It is evident that flange
bolts in joint type B contributed significantly to the stiffness of the joint in the bedding down stage,
once all the slip had occurred, the stiffness of joint type A and B was very similar.

a) Joint type A (a
bg
=160mm)

b) Joint type B (a
bg
=160mm)


b) Joint type B (a
bg
=280mm)

Figure 10 Experimental versus analytical moment-rotation relationship

The experimental strength and stiffness of the base connection as shown in Figure 6c was also
established. Intentionally, the connection was formed through very flexible 3mm thick back-to-back
angle cleats to act close to a perfect pin. The test results confirmed that rotational stiffness of the base
connection was 17 kNm/rad thus was assumed as zero in further analysis. The test was stopped when
the joint reached bending moment of 2 kNm due to very large deflection.

FRAME BEHAVIOUR

Full-scale tests

Three series of full-scale frame tests, designated Frames A(160), B(160) and B(280), were conducted
in order to investigate the interaction between joint stiffness and stressed skin action. Figure 11 shows
details of the general arrangement of the test setup. The frame is of 6 m span, 3 m height, and 10
o

pitch. As can be seen, the test setup comprises two frames, with the gable frame being restrained and
load only being applied to the adjacent test frame. The stiffness of the gable restraints is similar to the
stiffness of the gable cladding panel calculated to BS 5950-9. The gable frame is also restrained in
out-of-plane direction by the set of ties. Reaction force in each tie is measured through strain gauge
load cell and displacements of key nodes are measured by linear displacement transducers as shown in
Figure 11.

The geometry of the roof panels was kept similar to this presented in previous section of this paper, so
experimental data obtained in panel test can be used to model stiffness of the roof structure in clad
frame test. The same purlins, cladding, fixing details, etc. were used for clad frame test as for
component panel test, presented previously. In the clad frame test sheeting profiles were only placed
on the roof. The side walls and gable frame were not clad in order to measure reaction forces. Purlins
and side rails were installed in order to provide lateral-torsional restrains to column and rafters. The
test arrangements for bare and clad frame are presented in Figure 12.


a) Vertical load

b) Horizontal load

Figure 11 General arrangement of full-scale test frame


Two general load cases were considered:
1. Vertical load case (see Fig. 12a)
2. Horizontal load case (see Fig 12b)

a) Bare frame (vertical load)

b) Clad frame (horizontal load)

Figure 12 Front view of full-scale test frame

Figure 13 shows details of the joints. As can be seen, the eaves and apex joints are formed through
bolts in both the webs and flanges of the sections. For all joints, each joint comprise of a 3 x 3 array of
bolts in the web. The tests differed in the presence of flange bolts and bolt-group length.

In the case of Frame A, the brackets had no stiffeners along the edges. The purpose of this was to
allow the bolt-group to rotate freely without any contact with flanges. The back-to-back angle bracket,
as had been shown in component joint test, was not sufficiently thick to resist buckling, but it was still
used as a benchmark study.

In the case of Frame B, the brackets had stiffeners along the edges with an additional three flange
bolts. As can be seen, all the joints have a bolt-group width of 80 mm and the length differs from 160
mm to 280 mm, consistent with the bolt-group used in the joint tests. Similar to the joint tests, fully
threaded M16 bolts were used, with a 2 mm holes clearance.

For all the tests back-to-back channels were used for columns and rafters with average dimensions as
shown in Table 4. The average section properties for all members and brackets along with the average
mechanical properties of steel are presented in Table 7. The properties of steel for full-scale test
components were assumed the same as in joint and roof panel tests since all the components were
manufactured from the same batch of steel.

The flexural stiffness of back-to-back channel beam was reduced by the factor of 0.772 as obtained for
the component joint test. The stiffness of the bracket is not uniform along the length, and therefore
Table 7 presents geometrical properties at the beginning and at the mid-point of the bracket.

TABLE 7
SECTION PROPERTIES OF THE FULL-SCALE TEST COMPONENTS (BASED ON NOMINAL
DIMENSIONS) AND AVERAGE MECHANICAL PROPERTIES OF STEEL
Section type A
1
A
2
I
y,1
I
y,2
f
ya
f
ua

mm2 cm4 cm4 N/mm N/mm
Main members
(Back-to-back
lipped channel )
1191 429.7
331.7*
508 545
Eaves bracket type
160A
1325 2592

389.7

3749.7

370 476
Apex bracket
type 160A
1325 1579

389.7

728.6

Eaves bracket type
160B
1765 2458

822.4 5771.8

370 476
Apex bracket type
160B
1765 2020

822.4

1377.1

Eaves bracket type
280B
1765

3505

822.4

9097.6 370 476
Apex bracket type
280B
1765

2138

822.4

1697.2

* Values reduced by 0.772 factor


Type A (160)

Type B (160)

(a) Eaves joint details

Type A (160)

Type B (160)

(b) Apex joint

Type C (160)

(c) Column base

Figure 13 Details of eaves, apex and column base joints


Test programme

A total of nine frames were tested, which are summarised in Table 8. Within each series, three frames
were tested:
1. Vertical load without sheeting (stressed skin action has little effect on vertical load)
2. Horizontal load without sheeting
3. Horizontal load with sheeting

TABLE 8
SUMMARY OF FULL-SCALE FRAME TESTS
Test designation Bracket
stiffeners
Flange bolts Load direction Sheeting
Frame 1A (160) No N/A Vertical No
Frame 2A (160) No N/A Horizontal No
Frame 3A (160) No N/A Horizontal Yes
Frame 4B (160) Yes Yes Vertical No
Frame 5B (160) Yes Yes Horizontal No
Frame 6B (160) Yes Yes Horizontal Yes
Frame 7B (280)
Yes Yes Vertical No
Frame 8B (280) Yes Yes Horizontal No
Frame 9B (280) Yes Yes Horizontal Yes

For each set of frames, the vertical test was conducted without sheeting. The horizontal tests were
conducted both with and without sheeting, so that the effect of stressed skin could be determined.
Throughout the preliminary tests, it was observed that four screws purlin to rafter connection was able
to carry significant transverse bending. For bare frame horizontal tests purlins were fixed only through
one screw (pinned joint).

Load cycles to remove the initial bolt slip from the frame were not conducted as any bolt-hole
elongation so caused would not be recoverable. All the bolts were lightly tied with the spanner to
ensure that very little friction is present within the joint. The frame was assembled then the slope of
the roof as well as uprightness of columns were measured by digital inclinometer; at this position the
transducers used to measure deflection were set to zero. The values of initial sagging of the rafter as
well as horizontal displacement of the top of the column under self-weight were calculated based on
the angle difference between perfect geometry and measured one. The data is presented in Table 9 and
was letter superimposed into the experimental results.

As could be seen from Table 9, the initial deformation recorder due to slack was as much as 100 mm
rafter deflection for frame with joint type A. Frames with joints type B presented significantly better
behaviour and little initial deformation was recorded.

The same load was applied to each jack until the failure of the structure. The load was applied in steps
of approximately 0.5 kN. At the end of each load step, readings were taken.


TABLE 9
INITIAL CHANGES OF GEOMETRIES OF THE TESTED FRAMES UNDER SELF-WEIGHT
Test designation
(bolt-group length)
F
v,sw
sw
6A,sw
2A,sw

(kN) (deg) (mm) (mm)
Frame 1A (160) 2.30 8.45 81.4 19.2
Frame 2A (160) 1.12 8.10 99.9 23.1
Frame 3A (160) 1.72 8.65 70.9 16.8
Frame 4B (160) 2.36 9.60 21.0 5.2
Frame 5B (160) 1.18 9.85 7.9 1.4
Frame 6B (160) 1.78 9.65 18.3 4.5
Frame 7B (280)
2.48 10.00 0.0 0.0
Frame 8B (280) 1.27 10.00 0.0 0.0
Frame 9B (280) 1.87 9.85 7.9 2.1

Beam idealisation of the full-scale tests

Four different analysis models for modelling bare frame were proposed according to literature review
and are shown in Figure 14. The rotational stiffness of the joints was modelled as non-linear rotational
spring as recommended by Lim and Nethercot (2002) or Chung et al. (2008). The fundamental
difference between presented frame models and those found in the literature was that the free rotation
of the joint due to initial slip was incorporated and vertical as well as horizontal load case was
considered.

Component joint tests have confirmed that the joints bending capacity should be considered as
critical design criterion thus interaction between bending and web crippling as proposed by Dubina et
al. (2007) and Dubina and Ungureanu (2008) was used to calculate joints bending capacity. The
effect of axial forces on the overall capacity of the columns was neglected in the theoretical analysis.
The assumption was also made that purlins provide sufficient lateral restraints so global instability is
not considered.

In case of Model 1, joints are considered as rigid and the design resistance of the joint is calculated
from equation 6 and is presented in Table 10. The same analytical design resistance is used in Model 2
and Model 2_1. The rotational stiffness of the bolt-group is calculated according to equation 7. The
free rotation of the bolt-group due to initial slip is also calculated and assumed as close to zero (5% of
joints rotational stiffness value was used for easy model convergence). The finite length of the eaves
and apex connection bracket was also modelled. The connection brackets were modelled as tapered
sections of geometrical properties as described in Table 7. In previous studies, the same value of
rotational stiffness has been used for modelling eaves and apex joint (see Model 2). In Model 2 this
assumption was challenged by changing the values of rotational stiffness proportionally between
eaves and apex joint as described in Table 10.

In Model 3, values of free joints rotation due to initial bolts slip as well as joint stiffness and strength
were taken from component tests of relevant joints as described in Section 3.

The clad frame structural model is presented in Figure 14. The stiffness and the shear strength of the
cladding are modelled as bi-linear spring according to theoretical values presented in Table 2. In

Model 3, the experimental non-linear load-deflection relationship is implemented into the model
according to Figure 5.

(a) Bare frame

(b) Clad frame

Figure 14 Structural models used to simulate full-scale tests

As can be seen in Figure 11, a set of reaction ties was also present at the apex of the gable frame. The
load cells attached to these ties have recorded very little load which was finally neglected in the clad
frame model.

The in-plane and out-of-plane interaction forces at the eaves level were measured along with
respective displacement for each full-scale test. These forces would be also induced in real structure
and would depend on such parameter as in-pane stiffness of the internal and gable frame, shear
stiffness of the roof cladding and stiffness of side wall panels. Measured values of stiffness for each
tie member were incorporated into the models as bi-linear springs (see Table 12).

All the models were solved using non-linear analysis in the Autodesk Robot Structural Analysis
Professional software.

TABLE 10
NON-LINEAR ROTATIONAL SPRING MODELS FOR 160MM LONG BOLT-GROUP
Spring
position
Model Moment
sign
s
S
j,s
j
S
j
M
j
rad kNm/rad rad kNm/rad kNm
Bare frame 160A
Eaves
and apex
1 symmetry 0 rigid 0 rigid 13.16
Eaves
and apex
2 symmetry 0.0224 25 0.0479 494 13.16
Eaves 2_1 symmetry 0.0224 25 0.0351 988 13.16
Apex symmetry 0.0224 25 0.0734 247 13.16
Eaves
and apex
3 symmetry 0.0224 31 0.0479 627 16.68
Bare frame 160B
Eaves
and apex
Eaves
and apex
2 symmetry 0.0056 25 0.0320 494 13.16
Eaves 2_1 symmetry 0.0056 25 0.0188 988 13.16
Apex symmetry 0.0056 25 0.0583 247 13.16
Eaves
and apex
3 positive 0.0039 278 0.0293 546 16.23
negative 0.0039 278 0.0326 596 18.17


TABLE 11
NON-LINEAR ROTATIONAL SPRING MODELS FOR 280MM LONG BOLT-GROUP
Spring
position
Model Moment
sign
s
S
j,s
j
S
j
M
j
rad kNm/rad rad kNm/rad kNm
Bare frame 280B
Eaves
and apex
Eaves
and apex
2 symmetry 0.0034 65 0.0168 1309 17.7
Eaves 2_1 symmetry 0.0034 65 0.0101 2618 17.7
Apex symmetry 0.0034 65 0.0301 655 17.7
Eaves
and apex
3 positive 0.0069 547 0.0185 1229 18.15
negative 0.0069 547 0.0203 1229 20.31


TABLE 12
STIFFNESS OF THE BRACING MEMBER DESCRIBED AS A BI-LINEAR SPRING
Spring position k
1
D
1
k
2
kN/mm mm kN/mm
Clad frame 160A
3A
1.79 2.73 0.75
4A
1.30 3.33 1.46
3B
0.70 2.4 1.73
4B
1.79 1.63 1.93
Clad frame 160B
3A
1.57 3.19 0.69
4A
0.64 3.77 0.85
3B
1.13 0.84 1.39
4B
1.77 1.25 1.93
Clad frame 280B
3A
0.89 5.51 0.66
4A
0.96 5.07 0.94
3B
0.89 2.37 1.73
4B
1.65 2.03 2.00

Test results

Generally good agreement between analytical and experimental stiffness of tested frames, albeit
mainly under vertical load, was obtained (see Figure 15). Finite element beam analysis was also able
to capture initial deformation of the frame due to initial slip in the joints. As observed in the
component tests, the bending capacity calculated based on interaction between bending moment and
transverse force in the first line of bolts produced conservative results for the 160 mm long bolt-group.
The axial force in the column also has significant effect on web buckling capacity thus overall
capacity of the joint. When experimental moment-rotation relationship from pure bending test was
modelled in Model 3, higher frame capacity was always obtained. Rigid joint model predicted overall
capacity of the frame but significantly underestimated the frames vertical deflection.

All of the frames, regardless of the joint detail, failed in web buckling of the column section in the
vicinity of the bracket as shown in Figure 15. In case of Frame 1A (160), significant twist of the eaves
bracket was also observed as could be seen in Figure 15a.


In case of horizontal tests carried out on bare frames, the structural models proposed, base on the same
rotational stiffness value modelled at each bolt-group, significantly underestimated frames strength
and horizontal stiffness (see Figure 16). This fact was more evident in frames with more flexible
joints. As expected, Frame 2A (160) presented very little stiffness until all of the bolts were in contact
with the steel, which produced a very high sway deflection of 160 mm. Due to very large horizontal
displacement, tests 2A and 5B had to be stopped and measurement equipment rearranged. Four load
cycles in order to obtain the failure were carried out. Only first and last load cycle is presented in
Figure 16a and 16b. In case of longer bolt-group with flange bolt (see Figure 16c) the horizontal
displacement was significantly smaller, thus the failure load was obtained in first load cycle. As
expected, cycling of the load increased the joint stiffness significantly as seen in Figure 16a and 16b.

Interestingly according to test results the horizontal capacity of the frame was not affected by the
length of the bolt-group. This fact contradicts with the analytical method proposed in the literature in
which transverse force in the first line of bolts decreases when length of the bolt-group increases.

Similar to the vertical tests, all the joints failed by local web buckling of the column in the vicinity of
the bracket as presented in Figure 16.

Finally, tests on frames with roof cladding presented completely different load paths than those
applicable to bare frame, as could be seen in Figure 17. The amount of load attracted by each
structural component is therefore proportional to its stiffness. The total horizontal load applied to
internal frame was transferred through very strong and stiff cladding to the gable frame. In Frame 3A
(160), the building sustained a horizontal load 2.5 times higher that of bare frame. The maximum
horizontal displacement at the eaves of the loaded clad frames was also reduced by a factor of 0.21
comparing to this of the bare frame.

The structure failed in the combination of shear, bending and crushing of the top-hat purlins as shown
in Figure 17a. The gable rafter was also damaged during the test. The brackets and sections flanges
were bent upward. To stop this undesirable feature thin metal strap was fixed at the junction of both
brackets flanges as shown in Figure 17b.

It could be also concluded from Model 3 that the ultimate shear strength of the roof panel obtained
during component test was always higher obtained from the full-scale tests. This could be explained
by the fact that in the full-scale test the top-hate purlins are being twisted so one leg is crushed and the
other is pulled-out (see Figure 17c).

Generally, the true stiffness of the clad building was well predicted by finite element (FE) beam
models using representative stiffness of the members. As expected, since the strength of the clad
building was based on shear strength of the roof cladding, very conservative analytical value derived
in previous section, reduced significantly the resistance of the building obtained from the FE model.
By improving the standard method for predicting the shear strength of cladding panels, better
agreement in terms of overall resistance could be obtained.


Figure 15 Bare frame - total vertical load against apex deflection


Figure 16 Bare frame - total horizontal load against column displacement

Figure 17 Clad frame total horizontal load against column displacement


CONCLUSIONS

The importance of stressed skin design in case of flexible cold-formed steel frames has been
highlighted. Traditional 2D design concepts, based on bare frame analyses, are very conservative if
stressed skin action of the roof system is not included in the analyses. Considering components
presented in this paper, the shear stiffness of the roof panel is nearly 21 times higher than the stiffness
of the bare Frame 8B (280). This difference in stiffness results in a load redistribution between
internal and gable frame, with 23% of total horizontal load being transferred to internal frame and rest
being attracted to gable frame. Stressed skin design not only improves serviceability of the building
but also potentially offers lighter internal frames.

As was presented in this paper, typical semi-rigid 2D design reported in the literature is only accurate
for predicting buildings behaviour under vertical load. The same design concept however will
significantly over-estimate horizontal deflections resulting in heavier internal frames. The test results
for Frame 8B (280) show that horizontal stiffness of bare frame was 9.2 times lower for the clad
building. The research focusing on constructing very rigid and expensive joints is only economically
valid if vertical deflection controls the design. The stiffness of the joints, as shown in the paper, would
have much less effect on the overall horizontal deflection, if the behaviour is dominated by stressed
skin action of the roof system.

The feasibility of using current standards and state of the art research on cold-formed steel frames in
implementing stressed skin design was also examined. It was established that simple 3D finite element
beam models can represent the true behaviour of clad structure. Although the analytical methods
generally led to conservative design, significant improvements could be made, by improvements in
predicting the ultimate strength of joints and cladding panels.

Most cold-formed steel frames use bolted joints which suffer from the effect of slip due to bolt holes
tolerance which is usually neglected in design. This slip contributes largely to initial frame flexibility
and thus, the forces are attracted by the building envelope rather than primary frame. Simple rules are
presented in order to model this phenomenon so the failure of the cladding can be prevented.

REFERENCES

BAIGENT, A. H. & HANCOCK, G. J. 1982. The behaviour of portal frames composed of cold-
formed members. In: DEPARTMENT OF CIVIL ENGINEERING, U. O. S. (ed.) Thin-walled
structures - Recent technical advances and trends in design, research and construction.
Oxford: Elsevier Applied Science.
BS 5950-9 1994. Structural use of steelwork in building - Part 9: Code of practice for stressed skin
design. London: British Standards Institution.
BS EN 1993-1-3 2006. Eurocode 3 - Design of steel structures. Part 1-3: General rules -
Supplementary rules for cold-formed members and sheeting. Brussels: European Committee
for Standardization.
CHUNG, K. F. & LAU, L. 1999. Experimental investigation on bolted moment connections among
cold formed steel members. Engineering Structures 21, 898-911.
CHUNG, K. F., YU, W. K., HO, H. C. & WANG, A. J. Year. Advances in analysis and design of
cold-formed steel structures. In: Joint Structural Division Annual Seminar 2005, 2005. 67-87.
CHUNG, K. F., YU, W. K., HO, H. C. & WANG, A. J. 2008. Advances in analysis and design of
cold-formed steel structures. Advances in Structural Engineering, 11, 615-632.

DAVIES, J. M. & BRYAN, E. R. 1982. Manual of stressed skin diaphragm design, London, Granada.
DE VOS, G. P. & VAN RENSBURG, B. W. J. 1997. Lightweight cold-formed portal frames for
developing countries. Building and Environment, 32, 417-425.
DUBINA, D., STRATAN, A., CIUTINA, A., FULOP, L. & ZSOLT, N. Year. Monotonic and cyclic
performance on joints of cold formed steel portal frames. In: J., L., ed. Fourth International
Conference on Thin-Walled Structures., 2004 Loughborough, U. K., 381-388.
DUBINA, D., STRATAN, A. & NAGY, Z. Year. Full-scale testing of cold-formed steel pitched-roof
portal frames of back-to-back channel sections and bolted joints. In: BEALE, R. G., ed. 6th
International Conference on Steel and Aluminium Structures, 2007 Oxford, UK. 931-938.
DUBINA, D. & UNGUREANU, V. Year. Behaviour of multi-span purlins of bolted lapped cold-
formed z-sections. In: T., R. K. W., ed. Fifth International Conference on Coupled Instabilities
in Metal Structures., 2008 Sydney. 507-514.
DUBINA, D., UNGUREANU, V. & STRATAN, A. Year. Ultimate design capacity of pitch-roof
portal frames made by thin-walled cold-formed members. In: M., M., ed. Fifth International
Conference on Thin-Walled Structures. Recent Innovations and Developments., 2008
Brisbane, Australia. 387-394.
ECCS 1995. European recommendations for the application of metal sheeting acting as a diaphragm,
Brussels, European Convention for Constructional Steelwork.
KIRK, P. Year. Design of a cold formed section portal frame building system. In: 8th International
Specialty Conference on Cold-Formed Steel Structures, 1986 St. Louis, Missouri, USA. 295-
310.
KWON, Y. B., CHUNG, H. S. & KIM, G. D. 2006. Experiments of cold-formed steel connections and
portal frames. Journal of Structural Engineering, 132, 600-607.
LIM, J. B. P. & NETHERCOT, D. A. 2002. Design and development of a general cold-formed steel
portal framing system. The Structural Engineer, 80, 31-40.
LIM, J. B. P. & NETHERCOT, D. A. 2003. Ultimate strength of bolted moment-connections between
cold-formed steel members. Thin Walled Structures, 41, 1019-1039.
LIM, J. B. P. & NETHERCOT, D. A. 2004. Stiffness prediction for bolted moment-connections
between cold-formed steel members. Journal of Constructional Steel Research, 60, 85-107.
MKELINEN, P. & KANKAANP, J. Year. Structural design study on a light-gauge steel portal
frame with cold-formed sigma sections. In: 13th International Specialty Conference on Cold-
Formed Steel Structures, 1996 St. Louis, Missouri, USA. 349-371.
MILLS, J. & LABOUBE, R. 2004. Self-Drilling Screw Joints for Cold-Formed Channel Portal
Frames. Journal of Structural Engineering, 1799-1806.
WRZESIEN, A. M., LIM, J. B. P. & LAWSON, R. M. Year. The ultimate strength and stiffness of
modern roof systems with hat-shaped purlins. In: CHAN, S. L., ed. Sixth International
Conference on Advances in Steel Structures 2009 Hong Kong, China. 480-487.
WRZESIEN, A. M., LIM, J. B. P. & NETHERCOT, D. A. 2011. Optimum joint detail for a general
cold-formed steel portal frame. Advance in Structural Engineering, submitted
ZADANFARROKH, F. & BRYAN, E. R. Year. Testing and design of bolted connections in cold-
formed steel sections. In: 11th International Specialty Conference on Cold-Formed Steel
Structures, 1992 St. Louis, Missouri, USA. 625-662.

ISBN 978-983-44176-4-2

A REVIEW OF COLD-FORMED STEEL BUILT-UP COLUMNS
RESEARCH AT CURTIN UNIVERSITY

H. H. Lau
1
and T.C.H. Ting
1

1
Civil & Construction Engineering Department, Curtin University Sarawak Malaysia, CDT 250,
98009 Miri, Sarawak, Malaysia

KEYWORDS

Cold-formed Steel, Innovations, Built-up Section, Compression, stub column.

ABSTRACT

Cold-formed steel construction has grown from a simple imitation of wood construction to an
increasingly versatile construction system. Notable cold-formed steel innovations include built-up
sections, custom-curved members, light weight steel hybrid, and wall framing. The amount of
innovation in the products and related accessories for cold-formed steel has grown at an exponential
rate during the last decade. However, only a handful of universities include cold-formed steel design
in their undergraduate syllabus, although cold-formed steel is a good design material that is durable
and green for sustainable construction practices nowadays. This paper summarises some of the
cold-formed steel research carried out at Curtin University Sarawak Malaysia over the last five years
especially in cold-formed steel built-up section.

INTRODUCTION

Cold-formed steel sections are widely used in the construction industry due to their good structural
performance, versatility in application and high build-ability during construction. This is because

cold-formed steel has unique properties which benefit a wide variety of applications in the
construction industry. Significant technological advances in materials processing have led to the
development of complex cold-formed steels sections with excellent mechanical properties and
important progress has also been made in the improvement of construction tools in the cold-formed
steel section in construction applications. There is growing interest on practical applications of
cold-formed steel sections in building construction in South East Asian countries such as Malaysia
and Indonesia.

Cold-formed steel sections are mainly used as secondary structural members. However, it is now
possible to use cold formed steel sections as primary structural members in building construction. It is
a more economical design compared to hot-rolled steel in short span applications and has started to
replace other construction materials such as timber in Malaysia as a result of their superior strength to
weight ratio and ease of construction.

Cold-formed steel structural research programmes across the world have laid the ground for the
development of national and international specifications, codes and standards spanning both the design,
fabrication and erection processes in cold formed steel. Recommendations are made on research
activities aimed at overcoming obstacles to the wider use of cold-formed steel in construction. New
opportunities for cold-formed steel arising from the shift towards sustainable development are
reviewed, including its use in primary structures.

In Malaysia, the use of cold-formed steel in the construction industry is increasing, especially in the
construction of roof trusses. In order to decrease weight and span over a large area, an innovative
built-up section is introduced. This built-up section can be made up of two single C-channels
connected back-to-back using self-drilling screws (Figure 1). This type of section is relatively new.
Moreover, existing design procedures in the specifications for built-up section are adopted from
hot-rolled steel research. Thus, further research is necessary to study theperformance of this
cold-formed steel built-up back-to-back channels stub columns.

Figure 1. An example of Built-up Back-to-back Channels Section

Even though extensive research has been carried out to study single sections such as the C-channel,
limited research and literature are available for built-up sections. Built-up sections are doubly
symmetric sections and are different from singly symmetric sections by the fact that local buckling
does not cause a shift of its effective centroid. Three types of instability failure i.e. local, distortional

and global buckling are common for all cold-formed steel sections in compression. These complex
buckling behaviours lead to complex design problems. Currently, the design methods used are the
effective width method, the direct strength method, the finite strip method and the finite element
methods. Among the above methods, the finite element method is more economical and has been
widely used (Yang & Hancock, 2006).

This paper summarises the authors research work on the cold-formed built-up steel sections at Curtin
University Sarawak Malaysia over the last five years. Many of the works have been published in
various conferences. (Ting & Lau, 2011a), (Ting& Lau, 2011b), (Lau & Ting, 2010), (Lau & Ting,
2009)

LITERATURE REVIEW

Built-up sections are members which are composed of two or more structural sections connected
together mechanically by using self-drilling screws. Structural viability and installation requirements
in construction projects resulted in built-up sections being a highly used element in many low- and
mid-rise residential and commercial buildings. These built-up sections can be categorised to two
types, i.e. back-to-back C-channels (open sections) and boxed C-channels (closed sections). Built-up
back-to-back C-channels section are formed by connecting two identical simple C-lipped channels at
their webs with self-drilling screws whereas built-up box sections are formed by connecting two
identical simple C-lipped channels connected at their flanges with self-drilling screws as shown
inFigure 2.

(i) Built-up Back-to-back Section (ii) Built-up Box Section
Figure 2. Built-up Test Specimens

Researchers have noticed the advantages of utilizing built-up sections in steel structures (Yu 2000,
Beg et al. 2004, Niazi&Rondal 1990). Several studies have shown that the performance of a structure
is improved using built-up sections (Salem et al 2009, Young and Chen 2008 and Whittle and
Ramseyer 2009). These built-up sections are widely used in local projects in Malaysia to increase
the load bearing capacity of a cold-formed steel structure stretching over a vast span, commonly for
roof and wall systems, floor decking, framing of residential, industrial, commercial and agricultural
buildings. However, the potential of built-up section is limited without in-depth knowledge from
research and development works. Conservative design, such as assuming the strength of built-up
section as twice of the strength of an individual C-Channel, further restrict the commercial benefits of

these built-up sections.

An understanding of stress distribution is important in the built-up section elements because the
connected elements tend to act integrally and also individually. This can potentially lead to a change
in behaviour and also carrying capacity of the built-up sections. Moreover, box sections and
back-to-back cross sections behave differently. Most noticeably, the torsional-flexural buckling will
be greater due to the enhanced torsional stiffness of the closed cross-section shape in box section.

In order to further understand the characteristics, behaviour and design methods, various studies have
been conducted. While it is apparent that built-up sections have gained attention from academia and
industry, existing research work is still limited such asthe effect of slenderness ratio.

Research on the slenderness modification for built-up members began as early as 1952, when Bleich
(1952) proposed an analytical criterion to modify the overall slenderness ratio of battened columns.
The proposed equation was limited to battened columns with hinged-end conditions only. Later,
Zandoninis (1985) study on hot-rolled stitched built-up struts made a greater step as the 1986s AISC
Specification introduced the slenderness modification equation for built-up members based on his
research study. The modification takes into account the effects of connectors and their spacing on the
buckling mode of built-up members. The AISC Specification is the first AISC publication to introduce
the slenderness ratio formulae for built-up members. These formulae were later verified and improved
by Aslani and Goel (1991) with inclusion of the section separation ratio ().

As the amount of investigations increased, many design codes started to adopt the consideration of the
slenderness ratio. Thus more researchers were interested to further investigate the slenderness ratio
effects of the other design codes. In 2004, Lue et al.s (2004) research on the hot-rolled, axially
loaded, built-up, double-channel columns concluded that the AISC column design strengths with the
modied slenderness ratio were generally conservative. A more thorough research study was
conducted by Lue et al. (2006) as they made comparisons of the slenderness ratios from various
design codes including the AISC-ASD, the AISC-LRFD, the AS-4100 and the Canadian code CSA
S16-01. They found that the design strengths determined based on the slenderness ratio formulae of
the Australian Code (AS-4100) are more conservative. Liu et al. (2009) again concluded that the
Australian and Canadian design strengths are overall more conservative when compared to the current
AISC design because both designs do not utilise the use of a separation ratio ().

The slenderness modification was first investigated on cold-formed steel built-up sections by Stone
and LaBoube(2005). They investigated the effects of using the modification equation on perforated
built-up studs with different material thicknesses and concluded that the results with the slenderness
modication were conservative on built-up members. The results with slenderness modification were
on average conservative for thin members and highly conservative for thicker members (greater than
0.89 mm (0.035 in.)). Additional work was conducted by Brueggen and Ramseyer(2005) on
cold-formed channel sections in open- and closed-sections with intermediate, welded stitch
attachments. Their research on the built-up stub columns concluded that the AISI design methods
were conservative for compact members but not conservative for members with slender elements.
They also recommended for additional tests to be performed determining the effects of length and
location, spacing, and number of weld attachments on the behaviour of welded built-up members. In a

more recent study, Whittle and Ramseyers(2009) research on the closed-section, built-up members
concluded that the results were more conservative for longer and thicker built-up members with the
use of the modified slenderness ratio. Moreover, if the fastener and spacing requirements of the
provision C4.5 for built-up are followed, strength prediction using the modified slenderness ratio for
built-up members is marginally more conservative than using the unmodified slenderness ratio.

From literature review, many researchers have concluded that the modified slenderness ratio is
conservative for the designof cold-formed steel built-up members. Findings from the literature review
are summarised in Table 1.
TABLE 1
SUMMARY OF LITERATURE REVIEW ONTHE MODIFIED SLENDERNESS RATIO FOR THE DESIGN
OF BUILT-UP MEMBER
Year Author Findings Remarks
1952 Bleich

Modification for overall slenderness ratio proposed for
battened columns with hinged-end conditions only.
Limited to battened and
hinged condition
1985 Zandonini Slenderness modification taking into account of the
effects of connectors and their spacing on the buckling
mode of built-up members.
Connectors and their
spacing taken into account

1986 AISC

The first AISC publication introducing the slenderness
ratio formulae for built-up members.
Modified slenderness ratio
included in design code
1991 Aslani&Goel Include the section separation ratio () to
AISCsformulae and proposed a generalised version of
Bleichs analytical equation general end conditions.
Separation ratio
introduced

2003 Brueggen&Ramseyer

Recommended that additional tests be performed to
determine the effects of length and location, spacing, and
number of weld attachments on the behaviour of welded
built-up members.
Further investigation on
the effects of connectors
is required
2004 Lue et al. AISC column design strength with the
modifiedslenderness ratio is generally conservative.
Conservative
2005

Stone &LaBoube Slenderness modification was conservative on built-up
members.
Conservative
Brueggen&Ramseyer

For stub colums, AISI design methods is conservative for
compact members but unconservative for members with
slender elements.
Generally conservative

2006 Dung et al Australian Code (AS-4100) is more conservative.
Back-to-back snug-tight bolted sections, Canadian Code
(CSA S16-01) gives virtually the same design outcome as
the LRFD. Back-to-back welded sections, the CSA
S16-01 is less conservative.
Generally conservative

2009

Liu et al.

The Australian and Canadian design strengths are overall
more conservative because separation ratio is not
included.
Conservative

Whittle &
Ramseyer
The modified slenderness ratio is more conservative for
longer built-up members and thicker built-up members.
Conservative


Advantages of built-up Sections

Built-up sections offer advantages such as:
- Protection against web crippling (Yu 2000) the two channel sections stiffen and support each other
against web crippling at concentrated loads, thus, resulting in higher axial capacity.
- Provision of stiffness against high shear loads (Beg et al. 2004) for thin and deep sections, the two
channel sections stiffen and support each other against high shear loads.
- Increase in torsional stability (Niazi&Rondal 1990) the doubly-symmetric section has a coinciding
centroid and shear centre, thus increasing the torsional stability. A singly symmetrical cold-formed
steel section is weak laterally, and in torsion in particular.

These features of built-up sections offer advantages to the construction industry particularly in
spanning over a large area, carrying high loads while maintaining low self-weight, and reducing
construction time and cost.

Governing factors of the Strength of Built-up Columns

Since built-up sections are gaining importance in the industry, the factors that govern the strength and
behaviour of built-up sections are of interest. It is evident from Salem et al. (2004) that the buckling
capacityof a built-up section shall be a function of the gap in between the chord, spacing and stiffness
of connectors. The gap between the C-channels affects the slenderness of the column, thus, affecting
its strength. Spacing and stiffness of connectors along the members length influence the members
strength since these are the important factors that ensure that the two C-channels act integrally. This
shows that configuration of the built-up sections greatly affects the strength of the section.

Slenderness (i.e connector spacing) is one of the important governing factors of the strength of the
built-up section. It has been included within the specifications and codes by modified slenderness
ratio. In 1991, Aslani and Goel (1991) analytically and experimentally verified the use of the
slenderness modification equation in designing built-up members. Later, Brueggen and Ramseyer
(2003) identified that the modified slenderness ratio in the specificationis on average conservative for
compact sections but potentially not conservative for members with slender elements. More tests were
conducted by Stone and LaBoube (2005) who concluded that adopting slenderness modification in
strength prediction is on average conservative for thin members and overly conservative for thick
members. Later, Lue, Yen and Liu (2006) found that the AS4100 design code appears to result in
more conservative results then other standards/codes. More recently, Lue et al.s (2008) research
reaffirmed that using modified slenderness ratio to design for built-up sections is generally
conservative. Liu (2008) and his research team also concluded that conservative strength prediction is
obtained using modified slenderness if the required connector spacing (a/r) is met. Later, Whittle and
Ramseyer (2009) identified that the modified slenderness ratio is based on research adopted from hot
rolled steel. Thus, this makes strength prediction conservative.

THE DESIGN OF BUILT-UP COLUMNS

There were also developments on the methods used to predict the strength of built-up sections. Design

codes and specifications introduced the use of modified slenderness ratio to account for built-up
sections in Effective Width Method (EWM). Later, an alternative method, Direct Strength Method
(DSM) assisted by CUFSM program, is then introduced by Direct Strength Method Design Guide
(AISI 2006). The design guide uses finite strip analysis with restraints introduced at position where
screws occur.

Both North American Specification, (NAS, 2001 & 2002) and the Australia/New Zealand Standard
design code (AS/NZS 4600) (1998) adopt two design methods, namely Effective Width Method (EWM)
and Direct Strength Method (DSM).

The EWM considers each element of a cross section individually in its calculation. This method uses
reduced area (effective area), which involves tedious calculations, to account for the post-buckling
effect of cold-formed steel members. A compression member can basically be divided into four types of
elements, namely uniformly compressed stiffened elements, uniformly compressed stiffened elements
with an edge stiffener, uniformly compressed un-stiffened elements and uniformly compressed
elements with multiple intermediate stiffeners. Depending on which type of element is being considered,
the effective width (b
eff
) calculation will be affected. The effective area is calculated by multiplying the
effective width of each element to the thickness i.e.
t b A
eff e
=
compressive capacity of cold-formed
steel is then calculated by multiplying the effective area (A
e
) with the nominal compressive stress (F
n
) of
the cold-formed steel (NAS, 2002).

DSM has been developed by Schafer and Pekoz (1998). To overcome some limitations in EWM, this
method uses elastic buckling solutions and also takes into consideration of the interaction between
elements. There are manual calculation and software (such as CUFSM, and THINWALL programs)
available for calculating the elastic buckling solutions. However, one major drawback of the software is
that it only applicable for pin-ended conditions. If the fix-ended condition is investigated, then manual
calculations based on North American Specification are used to calculate the member axial capacity
(P
n
). This method determines the strength for local (P
crl
) and overall (P
cre
) interaction and distortional
(P
crd
) and overall (P
cre
) interaction and takes the lesser of the two as the compression strength (P
n
).

DSM does not require complex effective area calculations as in EWM. It provides a flexible design
procedure so that it simplifies the analysis of complex sections. It calculates the member strength
based on the members elastic buckling loads. The first DSM approach in this study (i.e. DSM I) uses
manual hand calculation from the design manual in determination of elastic buckling load.
Modifications on slenderness ratio (same as EWM) were introduced to calculate critical Euler
buckling stress (F
e
). For DSM II, finite strip analysis software, CUFSM, is used to determine P
crl
and
P
crd
. For both DSM I and DSM II, P
crl
and P
crd
are simply twice the single cross-section value in the
built-up cross section, thus, analysis was done by analysing a simple C-lipped channel. However,
P
y
/P
cre
differs because torsional-flexural mode is replaced by a separate torsion mode and a
strong-axis flexure mode. Due to difficulties to determine P
y
/P
cre
from CUFSM curve, hand
calculation methods were used. The finite strip analysis software CUFSM used in this research was
introduced by Schafer to predict the strength ratio. The first minima of the curve reveals load ratio for
local buckling where as the second minimum point shows load ratio for distortional buckling.


AISIs (2006) research study shows that buckling stress for local and distortional buckling are
unaffected by connecting two cross-sections together. However, in the built-up cross section, the weak
axis flexural mode is increased, and the torsional-flexural mode is replaced by a separate torsion mode
and a strong axis flexure mode (AISI 2006). Later, Megnounif, et al. (2007) proposed a new design
procedure for predicting the buckling strength of built-up sections based on EWM and DSM. Their
experiments on plain and lipped, built-up columns have shown that EWM is more accurate than DSM.
In another research, Young and Chen (2008) proposed three different analytical sections i.e. single
section; single section restrained at the flanges and double section with the flange thickness two times
the web thickness to predict the strength of built-up section. They concluded that the DSM, using a
single section, is conservative. Experimental results of the built-up section show an average
improvement of 19% compared to strength prediction of the member using individual member.

Built-up cold-formed steel columns are composed of two lipped channels inter-connected using a
series of screws. For most structures, the occurrence of a local collapse does not necessarily ensure
failure of the whole structure. However, for built-up members, failure of the screws results in an
extensive reduction in the overall member strength. Despite the vast availability of literature in the
general subject of columns, there is little focus on built-up columns. The American Iron and Steel
Institute 2001 Specification (AISI) (2001) and AS/NZS 4100 (1998) give no specific provision for the
design of built-up columns except stating that built-up members should be designed with a modified
slenderness ratio (Eqn. 1) if there are shear forces induced between the weld or screw connectors. The
provision C4.5 of AISI Specification also introduced a minimum fastener strength requirement and a
fastener spacing requirement (Eqn. 2) for built-up members. The fastener spacing requirement can
also be rewritten to determine the maximum longitudinal spacing of connections (Eqn. 3). These
requirements were rewritten in the form of equations with unified nomenclature as stated in provision
C4.5 of the AISI Specification as follows:
Modified slenderness ratio
2 2
|
|
.
|
\
|
+
|
|
.
|
\
|
=
|
.
|
\
|
yc
o
ybu m
r
s
r
KL
r
KL
(1)
Intermediate fastener spacing
o
ybu yc
r
KL
r
s
|
|
.
|
\
|
s 5 . 0 (2)
Maximum longitudinal spacing of connections
ybu
yc
r
Lr
s
2
max
= (3)

where (KL/r
ybu
)
o
is the overall slenderness ratio of the entire section about the built-up member axis; s
is the intermediate fastener or spot weld spacing; K is the effective length factor; and L is the
un-braced member length; s
max
is the maximum permissible longitudinal spacing of connectors; r
yc
is
the radius of gyration of one C-section about its centroidal axis parallel to the web; and r
ybu
is the
radius of gyration of the I-section about the axis perpendicular to the direction in which buckling
would occur for the given conditions of end support and intermediate bracing.

These equations are introduced to maintain the slenderness ratio of the individual C-section to be less
than or equal to half of the slenderness ratio for the overall compression member. This is to ensure that

the maximum spacing of the connectors is close enough to prevent buckling of individual C-sections,
if any one of the connectors may be loosened or ineffective. The buckling of individual C-sections
about their own axes parallel to the web is undesirable as it is at a load equal to or smaller than the
buckling loads of the overall member. With these requirements in provision C4.5 met, the built-up
member would function integrally as a single compression member rather than behaving as individual
C-sections. Despite being critical, the modied slenderness ratio is heavily adopted from the American
Iron and Steel Construction (AISC) code (1986) for hot-rolled, built-up members despite the fact that
the behaviour of hot-rolled steel members is very much different from the cold-formed steel members.
Thus, it is the purpose of this research to study the provision C4.5 of the Specification for cold-formed
steel built-up back-to-back channel columns.

APPLICATIONS OF BUILT-UP SECTIONS

This section presents several applications of cold-formed steel built-up sections in construction project
in Malaysia. These case studies presented here are projects taken up by EcoSteelSdn Bhd. Many local
projects in Malaysia require structural elements to span across vast areas. Thus, various parameters
such as carrying capacity, deflection, lateral stability, and constructability of the structure of the built
up sections become challenges faced by the local fabricators and designers. The sizes of the truss
elements constructed by singly symmetrical sections are increased to comply with design and
construction requirement. However, this leads to other on-site difficulties such as space limitation and
practicality in installation. Therefore, built-up sections are widely used nowadays. These back-to-back
built-up sections or boxed sections provide several advantages in construction such as:
- improved lateral stiffness;
- easier lifting for installation;
- simplified connection detailing of the web members to the chord members; and
- increased carrying capacity while maintaining lightweight.

Exposed Aesthetic Roof Truss

It is commonly perceived that cold formed steel structures cannot be left as exposed structures. It is
often covered up by ceilings or claddings. However, the roof truss in the new multi-purpose hall in the
Curtin University Sarawak Malaysia has been designed as exposed aesthetic roof structure. In this
case, the architect has worked closely with EcoSteelSdnBhd to carefully detail the truss connection,
joints and support in order the finish structures that can achieve the aesthetic requirements of the
building design. The built up sections used are C15025 which is 150mm in width and 2.5 mm
thickness for top and bottom chords. These gapped built-up sections are to accommodate the internal
web members of C10016 (Figure 3). The trusses support the double roof system with acoustic
insulations and spanned 32.5m with 3 degree of pitch (Figure 4). The structure was completed in
2009.


(a) Built-up Trusses (b) Built-up Section
Figure 3. Exposed Roof Structure (Photos courtesy of EcoSteelSdnBhd)

Figure 4. Multi-purpose Hall in Curtin University (Photo courtesy of EcoSteelSdnBhd)

Cone Shaped Roof Truss

The 3D cone-shaped roof structure is a feature roof over the Port Operation Building of the Integrated
Deep Sea Fishing Port at TanjungManis, Sarawak. The cone roof emulates the traditional ethnic
Melanau straw hat locally known as Terendak (Figure 5). The whole 3D cone-shaped roof truss is
pre-assembled on the ground and lifted by crane to the top of the Port Operation Building as shown in
Figure 6, which is about 27.0m above street level and 20.0m away from the edge of the

podium.Diameter of the cone roof is 18.8m with a 30 degree pitch.

Figure 7 shows the back to back sections used in this cone-shaped roof truss. The project was
completed back in 2006. A thorough discussion on design and construction of the roof structure was
previously published by Mei and his research team (2007).

(a) A Terendak (b) Pre-assembled on ground (c) Lifting
Figure 5. Three Dimensional Cone-shaped Roof of the Port Operation Building (Photo Courtesy of
EcoSteelSdnBhd)

Figure 6. Port Operation Building of the Integrated Deep Sea Fishing Port at TanjungManis(Photo Courtesy of
EcoSteelSdnBhd)

Figure 7. Built-up Back-to-back Section in Port Operation Building (Photo Courtesy of EcoSteelSdnBhd)


Conventional Roof Structure

This conventional roof structure spans 13.8m at 25 degrees pitch across the cafeteria of LobangBatu
High School, in Serian, Sarawak as shown in Figure 8. The truss structure in this project is made up of
built-up sections from C- channel sections with 75mm web width and 1.0mm thickness. These trusses
are then positioned at 1.2m centre-to-centre spacing. Back-to-back built-up sections are used in this
project to provide lateral stiffness to the pre-assembled built-up trusses during installation. With
sufficient lateral stiffness, the process of lifting, balancing and installation of the roof structure
becomes easier.

(a) Lifting the Roof Structure into Position (b) Installation of the Roof Structure
Figure 8. Roof Structure of cafeteria in SK LobangBatu, Serian, (Photo Courtesy of EcoSteelSdnBhd)

Octagonal Shaped Roof Structure

An octagonal-shaped roof structure shown in Figure 9 is designed for the Bandar Samariang High
School, Kuching, Sarawak. The roof structure spans across an 18.0m length while maintaining it
minimal weight. The structure was assembled on ground and hoisted into position using only one 20
tonnes capacity crane (Figure 10).

(a) Octagonal-shaped Roof Structure (b) Built-up Back-to-back Sections
forming the Octagonal Roof
Figure 9. Roof Structure of SMK Bandar Samariang (Photo Courtesy of EcoSteelSdnBhd)


Figure 10. Lifting Process at SMK Bandar Samariang (Photo Courtesy of EcoSteelSdnBhd)

Curved Roof Structure

The Sarawak International Medical Centre at Kota Samarahan, Sarawak, utilised a curved roof
structure to span across a length of 9.0m as shown in Figure 11 and Figure 12. Built-up sections are
used to build trusses that curve and span as required while maintaining a smaller depth and lighter
weight.

(a) Curved Roof Structure (b) Built-up Back-to-back Sections
forming the Curved Roof
Figure 11. Roof Structure of Sarawak International Medical Centre, Samarahan, Sarawak(Photo Courtesy of
EcoSteelSdnBhd)

Figure 12. Sarawak International Medical Centre, Samarahan, Sarawak (Photo Courtesy of EcoSteelSdnBhd)

Cantilevered Box Roof Structure

The construction of the Limbang Mosque, Sarawak used built-up box sections at 15 degree slope and
cantilever out from the lower roof of a mosque of the Mosque as shown in
Figure 13. In this project, small built-up box sections with depth of 75.0mm, and thickness of 1.0mm
are used as shown in Figure 14. This is because small C-channel sections were insufficient to provide
stiffness against high deflection. Thus, connecting two small C-channel sections to form a built-up
box section is adopted to improve the lateral stiffness. The box sections also improve serviceability of
the roof structure while spanning the same space as compared to single C-channel sections.

Figure 13. Roof Structure of Limbang Mosque, Sarawak (Photo Courtesy of EcoSteelSdnBhd)

Figure 14. Built-up Box Section used for Roof Structure of Limbang Mosque, Sarawak

Wall Frame Structure

Wall frame structures were constructed for MBO cinemas viewing halls at The Spring Shopping
Mall, Kuching, Sarawak. The built up sections used are C20025 which is 200mm width and 2.5 mm
thickness. They are spaced 400mm centre to centre. They span 15m vertically and are bolted to mild
steel angles at their ends.

Two cold formed steel built up section frames were constructed side by side for the adjacent viewing
halls (Figure 15&Figure 16). Thus each frame carries the wall cladding, which consists of 3 layers of
12mm gypsum boards, independently. The gypsum boards were fastened directly onto the built-up

sections and acts as lateral restraint to the critical flanges. This arrangement is also a form of insulation
against sound as required by the client.

(a) Double Frames (b) Wall Panel
Figure 15. Built-up Wall Frame (Photo Courtesy of EcoSteelSdnBhd)

(a) End connections (b) Connected to Gypsum board
Figure 16. Connections (Photo Courtesy of EcoSteelSdnBhd)

The main design challenges are the horizontal load, pressure differential and also imposed load.
Therefore, these built up sections were design as a parapet to resist any horizontal forces from the hand
rail or any other imposed load. They were designed to 0.25kPa of wind pressure and 0.3kN of point load
to meet the challenges.

FINITE ELEMENT MODELING

Finite element modelling has successfully been used in the past to predict the behaviour of cold-formed
steel structures. The capability of LUSAS in modelling built-up sections has been documented by many
publications. For built-up sections, correct modelling of screw connections and contacting surface
between the two lipped C-channels are the main issues.

Barriouset. al (2005) did a finite element analysis on shot peening i.e. one of the surface treatments for
metals. The analysis in this case is on contact between the shot and the plate. The author used the
slideline function in LUSAS to introduce a contact surface pair which consists of two slideline surfaces
called master and slave. Several convergence tests were done to control and avoid an excessive
penetration of a shot into a plate. In Omar et. als (2003) research study on concrete-filled steel tube
members, the contact and the friction between steel casing and concrete filling were modelled using
friction and smooth contact joint properties available in LUSAS.

Butterworth (1999) did a numerical study on beam to column bolted connections using LUSAS. In
order to reduce the models size and analysis time, tied slide lines were used to model the interface
between the end plate and the column flange. A similar research study on beam to column bolted
connections was carried out by Fuang (2007) and Chin (2008). They replicated the LUSAS model
from Butterworth by replacing simplified bolt model with actual model using volume geometry.
However, there were no clear explanations of the surface contact between the end plate and the
columns flange. Another research by Lee (2009)modelled the screws connecting plywood and
profiled steel sheet using 3 dimensional joint elements for beam with quadratic interpolation (JSH4).
In addition, push-out test on screws was carried out in Lees research study to better simulate the
actual joint condition. A normal tied mesh in LUSAS was used to tie the mesh so that the
displacement of the profiled steel follows the displacement of the plywood.

The finite element model was created using commercial software LUSAS v14.3 (FEA 2002) to
simulate the deformation curves and to predict the load carrying capacities for back-to-back channels
stub column for the research studies carried out at Curtin University Sarawak Malaysia.

The LUSAS model was subdivided into a discrete mesh size of mm mm 15 15 . Material non linearity
was modelled with von Mises yield criteria and plastic hardening. Steel properties for the specimens
were obtained from tensile coupon tests and subsequently used to define nonlinear material attributes
in the LUSAS finite element modelling. Load was applied as uniformly distributed loads at the loaded
end in automatic increments using the Total Langragrian geometric nonlinearity due to large
displacement and small strains. Figure 17shows the finite element model used.

Figure 17. LUSAS Model

Steel properties for the specimens were obtained from tensile coupon tests. The coupons were cut
from the centre of the web plate from the specimens that belong to the same batches as the column test
specimens to ensure that the tensile coupons represented the material properties of the tested column.
The coupon dimensions conformed to the Australian Standard AS1391 (1991) for the tensile testing of
metals using 12.5mm wide coupons of gauge length 50mm. A data acquisition system was used to
record the load and the readings of strain at regular intervals during the tests. Results obtained were
subsequently used to define nonlinear material attributes in LUSAS in the form of true stress versus
true strain curves.

The model was created using a linear 4-node quadrilaterial thick shell element (QTS4) shown in Figure
18. QTS4 is stiffened shell structures that can accommodate curved geometry which allows curves at
corners of the steel column. QTS4 formulation takes account of membrane, shear and flexural
deformations. QTS4 can be connected to beam elements to model connections, however, a simplified
model of 2mm strip was used.

Figure 18. Linear 4-nodeQuadrilaterial Thick Shell Element (QTS4)

Correct modelling of screw connections and the contacting surface between the two lipped C-channels
are keys to the modelling of a built-up section. LUSAS offers the slide line function to model contact
and impact problems. This function is often used to model the contact interaction between master and
slave features, such as the contact stiffness, friction coefficient, temperature dependency etc. In this
study, web of the lipped C-channels were in contact when fastened together in a back-to-back manner.
Surface-to-surface contact with no friction was defined between the outside surfaces of the two webs
with one of the surface defined as a master surface while the other as a slave surface (Figure 19). The
screw connections between the two channels were not explicitly modelled. A simplified model of a
2mm thin strip was used to connect two lipped C-channels together. This was to ensure that the channels
remain in contact with each other at discrete points over the entire load history and never separated.

Figure 19. Contact Analysis using LUSAS

The specimens were modelled with fixed ends. End plates were constructed by connecting the lines
and nodes at the both ends of the column to form a surface. The boundary condition was then applied
on this end surface. In the model, the ends of the columns were fixed against all degrees of freedom to
simulate fixed end condition, except for the displacement at the loaded end where the specimen is
allowed to move in the direction of the applied load. The nodes other than the ends were free to
translate and rotate in any directions.Load was applied as uniformly distributed loads at one end in
automatic increments using the Total Langragrian geometric nonlinearity due to large displacement
and small strains . A unit load was applied at the top end of the specimen where displacements in the
direction of the load were freed.
Slave
surface
Master
Surface
2mm strip
Screw
connection

The behaviour of built-up columns composed of thin-walled lipped C-section members connected
back-to-back is investigated numerically using the finite elements method. Verification of this finite
elements model in predicting ultimate capacities and the corresponding results of the tested specimens
were carried out and available in previous authors publication (Ting & Lau, 2011). A parametric
study covering different material thicknesses and connector spacing was performed using these verified
results and finite element models generated by a commercially available finite element package
LUSAS version 14.4 . Thin-shell elements with four nodes and six degrees of freedom at each node (i.e.
QTS4) were used to model the built-up columns. Large displacement and small strains analyses have
been incorporated in the finite elements model using the Total Langragrian geometric nonlinearity.
Material nonlinearity was also modelled with von Mises yield criteria and plastic hardening. Steel
properties for the specimens were obtained from tensile coupon tests and subsequently used to define
the nonlinear material attributes in the LUSAS finite elements modelling. The columns were loaded
under uniform compression at the loaded end in automatic increments. The end conditions of the
columns were pin ended with the loaded end prevented from rotations at the z-axis, and displacements
in both x and y directions. At the other end, the unloaded end was prevented from displacement in all
three directions x, y, and z and also prevented from rotation at the z-axis.

EXPERIMENTAL RESULTS

Specimens
Five built-up back-to-back channels stub columns were made up by connecting two single C-channels
back-to-back at their webs. Dimensions of the specimens are tabulated in Table 2. Specimens were
labelled with BU where BU refers to Built-up. Dimension of flange, thickness and length are also
shown in the labelling with thickness denoted by T and length denoted by L.

TABLE 2
MEASURED DIMENSIONS OF BACK-TO-BACK CHANNELS STUB COLUMNS
Specimen
Web, A
(mm)
Flange, B
(mm)
Lip, C
(mm)
Thickness, t
(mm)
Radius, R
(mm)
Length, L
(mm)
Screw
Spacing, s
(mm)
BU75T12L250-1 76 18.0 9.5 1.2 1.5 252 49.5
BU75T12L250-2 76 18.0 9.5 1.2 1.5 250 49.5
BU75T12L250-3 75 18.5 9.0 1.2 1.5 253 50.0
BU75T12L250-4 75 18.5 9.0 1.2 1.5 252 50.3
BU75T12L250-5 75 18.0 9.5 1.2 1.5 250 50.0

The material properties of the specimens were determined by tensile coupon tests. The coupons were cut
from the centre of the web plate from the specimens that belong to the same batches as the column test
specimens. This is to ensure that the tensile coupons represent the material properties for tested column
test specimens. The coupon dimensions conformed to the Australian Standard AS1391 (1991) for the
tensile testing of metals using 12.5mm wide coupons of gauge length 50mm. A data acquisition system
was used to record the load and the readings of strain at regular intervals during the tests. Results are as
shown inTable 3.


TABLE 3
MATERIALS PROPERTIES OBTAINED FROM TENSILE COUPON TEST
Material Properties Nominal Coupon Test
Youngs Modulus (GPa) 200 205
Yield Stress (MPa) 550 550.5

Setup

Compression tests were conducted on 5 built-up back-to-back channels stub columns. Axial force was
applied to the specimens via the GOTECH, GT-7001-LC60 600kN capacity Universal Testing
Machine (UTM). The test rig and specimen setup is as shown inFigure 20. 12.5mm thick plates were
welded to the ends of the specimen to ensure a fixed rigid flat end. Load and shortening were recorded
using the data acquisition software of the testing machine. Specimens were placed with their centroid
at the marked loading point on the bottom bearing. Load cell at the top was then lowered until it
touches the top end plate of the specimen. Level was used to check for the straightness of the
specimen setup.

Figure 20. Schematic Test Setup for Column Compression Test

Table 4compares the ultimate loads predicted by finite element models to the results obtained from the
experiments and theoretical calculations. The study shows that finite element model predicts the stub
column capacity well. The average ratio of (P
test
/P
FEM
) is 0.97. The FEM model consistently
overestimates the test results because imperfection was not accounted in the model. The axial
capacity, P
n
obtained using the EWM, gave results which are close to the ultimate test strength
(P
test
/P
EWM
) of 1.05. The strength predictions by DSM are also close to the ultimate test strength
(P
test
/P
DSM
) of 1.14 but slightly more conservative than EWM. Results based on EWM are 5% higher
while results based on DSM are 12% higher for stub column. EWM better predicts the stub column
capacity because the EWM takes the effectiveness of elements into account. The effectiveness of
elements is important for stub columns because local buckling dominates the failure. Although
distortional buckling and buckling interaction are adequately considered by DSM, these are not the
major mode of failure in stub columns. These five back-to-back built-up stub columns also evaluated
and verified the finite element model.The result of BU75T12L250 is plotted in Figure 21.


TABLE 4
LOAD CARRYING CAPACITY OF BACK-TO-BACK CHANNELS STUB COLUMNS
Specimen P
test
(kN) P
FEM
(kN) P
EWM
(kN) P
DSM
(kN) P
test
/ P
FEM
P
test
/ P
EWM
P
test
/ P
DSM

BU75T12L250 1 122.5 124.62 115.89 106.65 0.98 1.06 1.15
BU75T12L250 2 118.5 125.88 115.93 106.68 0.94 1.02 1.11
BU75T12L250 3 121.9 125.69 115.86 106.60 0.97 1.05 1.14
BU75T12L250 4 121.8 125.64 115.87 106.60 0.97 1.05 1.14
BU75T12L250 5 125.5 124.72 115.82 106.73 1.01 1.08 1.18
Mean 0.97 1.05 1.14

Figure 21. Graph of Load vs Displacement for BU75T12L250-1

Failure Modes

Each model is 250 mm long and connected at the intervals of 50, 75, 100, 150 or 200 mm. The section
size used were simple C lipped channels with web width of 75mm, flange width of 20mm, lip width of
10mm, and thickness is 1.2mm. The model was set up in such a way that buckling of the specimens
would occur in yy axis only.

A parametric study was performed concentrating on the behaviour of built-up back-to-back channel stub
columns. The purpose of the study was to assess the behaviour of the built-up cold-formed steel
compression members and to determine if the present AISI design provision is appropriate for
cold-formed steel members. The present study highlighted two parameters identified in (Stone
&Laboube, 2005), (Brueggen&Ramseyer, 2003) and (Whittle &Ramseyer, 2009) that directly influence
the behaviours of the built-up columns. These parameters are identified as: (i) s/L which is defined as the
ratio of the clear distance between connectors-to-the overall length of the column; and (ii) t which is
defined as the material thickness. This study investigated a total of 12 models which were classified into
2 specimen groups. The built-up sections were made up of two channels connected together
back-to-back as illustrated in Figure 22.


Figure 22. Cross Sectional Geometry and the Configuration of the Investigated Built-up Members

From observation, local buckling at web occurred as an initial buckling mode. Later, the flanges
deflected when load was close to the ultimate load. After that, deflection in web and flanges increased
and accompanied by a sudden drop in load when the specimen failed. This behaviour indicated that the
failure of these stub columns resulted mainly from local buckling and significant distortional buckling
only occurred when near the ultimate load. When compared with LUSAS model, the buckling shapes
and behaviour are similar to those of the test (Figure 23). The LUSAS model predicted that at fix-fix end
condition, significant local buckling occurs in the web with less significant local buckling at the flange
and lip. The proposed model well predicts the behaviour of built-up stub column.

Figure 23. Tested Specimens versus Failure Modes from Finite Element Model

Effect of Thickness (t) on Column Strength

The finite elements model results, PFEM, for four different thicknesses (i.e. 1.0mm, 1.2mm, 1.55mm
and 1.95mm) were compared to the theoretical results calculated in accordance to the AISI specification.
Comparisons were also carried out for both modified and unmodified predicted axial capacity, Pn, as
determined by provision C4 of the AISI specification. Figure 4 and Figure 5 illustrate the comparison of
axial capacity predicted by the finite elements method, the Effective Width Method, and the Direct
Strength Method. Results from Figure 24 indicate that for the models with thicker material thickness,
the existing AISI design equations without using the modified slenderness ratio are conservative.
However, the capacity of the thinner material (1.0mm) was slightly overestimated. Similarly,
comparison was carried out with the modified slenderness ratio in Figure 25. The results obtained were
s
s
s
t
75mm
250mm
20mm
10mm

similar but with greater increment for the models with thicker material thickness. Comparison
betweenFigure 4 and Figure 5 shows that the strength prediction for thicker models is more conservative
when the modified slenderness ratio is used. For thinner models, the strength predictions by both were
very close. This suggests that the modification of the slenderness ratio is not necessary for the thicker
materials. For both unmodified and modified slenderness ratios, finite element method consistently
overestimates the axial strength predicted by the Direct Strength Method. The finite elements models
with the thicker material thickness also show an upward pattern when the modified slenderness ratio is
used and the strength predictions are also more conservative. This trend was also observed by Stone and
LaBoubes (2005) research on built-up studs. In more recent research studies, Whittle and Ramseyer
(2009) further confirmed that modified slenderness ratio is conservative for members with thicker
material thickness.

Figure 24. Comparison of Finite Element Results and
Calculated Results using Unmodified Slenderness
Ratio (KL/r)
bu

Figure 25. Comparison of Finite Element Results and
Calculated Results using Modified Slenderness Ratio
(KL/r)
m

Effect of connector Spacing(s) on Column Strength

The finite elements model results, PFEM, for various spacing (i.e. 50mm, 75mm, 100mm, 150mm and
200mm) were compared to the theoretical results calculated in accordance to the AISI specification.
Comparisons were made to both modified and unmodified predicted axial capacity, Pn, as determined
by provision C4 of the AISI specification. Figure 26 illustrates the failure modes of the stub columns at
various spacing.

Built-up members buckled in various ways, sometimes as one integral section and often individually
with each c-channel buckling separately. Figure 26 shows the buckling behaviour of the built-up
members with various spacing. The most common buckling modes for the built-up stub column were
local buckling and distortional buckling. The smaller spacing built-up stub columns (30mm, 50mm,
75mm), buckled with angular buckling shapes. The buckled shape appeared as a hinge around the
screw connection of the column when the deformation increased. The channel sections also moved
laterally as a built-up integral section. With connectors at 100mm spacing, the channel sections moved
together laterally at the upper part of the column, while the channels at the bottom part moved separately
with the channels prying apart. However, the hinge phenomena occurring at smaller spacing built-up
stub columns is still visible. This hinge separates the buckling behaviour into two parts with the upper

column buckling integrally and the bottom column buckling separately. Models with larger spacing
(150mm) buckled separately with the channel sections prying apart at mid length due to the lack of
fasteners along the length of the member. With the largest connector spacing (200mm), the prying apart
of the channel sections occurred at mid length of the member. Similar buckling behaviours were also
observed by Whittle and Ramseyer (2009) in their research on the built-up columns with different
lengths, and different cross section sizes.

(a) s = 30mm (b) s = 50mm (c) s = 75mm (d) s = 100mm (e) s =
150mm
(f) s =
200mm
Figure 26. Buckling Behaviour of Stub Columns at Various Spacing Predicted by LUSAS

Strength Predictions

Figure 27. Graph of P
n
/P
y
vs s/L for BU07512 Stub Column

Figure 27 shows the relationship between the normalized ratio of the spacing between connectors to the
overall length of the column (s/L), and the normalized ratio of the predicted load to the yield load of the
column (P
n
/P
y
) for different design methods. Figure 27 shows that connector spacing has a significant
effect on the strength of the built-up columns. For the same column length, as the spacing increased, the
corresponding ultimate load decreased. This is attributed to higher values of the modified slenderness
ratio which directly leads to a reduction in the load carrying capacity of the built-up column. The greater
the connector spacing, the C-channels starts to behave individually rather than as one integral section.
This is because the overall global buckling of the member is dominating the behaviour, while the effect
of the local buckling becomes minimal. These types of columns are more susceptible to distortional and

overall buckling effects. This is due to lack of intermediate connectors to hold the c-channels together as
an integral built-up column. Thus, explains the reduction of built-up column strength with increased
connector spacing. The average strength capacity calculated based on the modified slenderness ratio
was more conservative than the average strength capacity calculated based on the unmodified
slenderness ratio. However, if all the fastener and spacing provisions for built-up members in C4.5 were
met, the strength prediction without slenderness modification calculation was less conservative than the
prediction based on the modified slenderness ratio calculation. This indicates that the strength
prediction based on the unmodified slenderness ratio calculation were not conservative for built-up stub
columns which do not meet the C4.5 spacing provision. For built-up stub columns that met the C4.5
spacing provision, the finite element results were closer to the theoretical results predicted using
modified slenderness ratio.

CONCLUSIONS

Compression tests on five built-up back-to-back channels stub columns fabricated from high strength
G550 steel has been performed. All stub columns fail in combined local and distortional buckling
modes with local buckling as the dominant failure mode. It is observed that local buckling occurred
before distortional buckling, and the ultimate load is usually achieved shortly after distortional
buckling occurs. Comparing test and theoretical results, EWM from the North American Specification
gives a better prediction than DSM.

A finite element model was designed using LUSAS to study the load carrying capacity and buckling
behaviour of built-up stub column. The model was verified against experimental results on
back-to-back channels and they collate well. Test data of five back-to-back channels stub columns
also show that the failure of stub columns was resulted from local and distortional buckling. It was
observed that local buckling occurred before distortional buckling, and maximum deformation usually
occurred near the mid height of the stub columns. Predictions correlate well with the test results,
whereas, prediction by direct strength method is slightly conservative for cold-formed steel built-up
stub column compared to test results.

A finite elements method is also used to investigate the effect of the material thickness and the
connector spacing on the behaviour of the built up column. The finite elements model results were
compared to theoretical buckling capacities calculated based on the provision C4.5 for modied
slenderness ratio of AISI Specification. Based on the results presented, it is reasonable to draw the
conclusions that the average strength capacity calculated based on the modified slenderness ratio of
provision C4.5 is conservative for built-up members. Secondly, the modified slenderness ratio of the
provision C4.5 is not necessary for thicker members when computing axial capacity. Finally, the
capacity of built-up sections decreased with the increased spacing, s. For the models with larger
spacing, the modification ratio resulted in a more conservative buckling strength prediction. However,
further studies with different cross section sizes of channel sections and different spacing with
different member lengths should be studied analytically and experimentally to reaffirm the conclusion
drawn.


ACKNOWLEDGEMENT

Authors would like to thank EcoSteelSdnBhd for the laboratory specimens and permission to use the
photos and related information.

REFERENCE

American Institute of Steel Construction, AISC, (1986) Load and resistance factor design specifications for
structural steel buildings, 1st edition, American Institute of Steel Construction, Inc.; Chicago (IL).
American Iron and Steel Institute, AISI.(2006), Direct Strength Method design guide.American Iron and Steel
Institute, Washington, DC.
Aslani F. and Goel S.C. (1991) An Analytical Criterion for Buckling Strength of Built-up Compression
Members, Engineering Journal, AISC, 28(4), pp.159-168.
AS/NZS 1391 (1991):Methods of Tensile Testing of Metals, Australian Standard/ New Zealand
Standards:1391:1991, Standards Australia, Sydney, Australia.
AS/NZS 4600 (1998): Cold-formed Steel Structures, Australian Standard/New Zealand Standards:4600:1998,
Standards Australia, Sydney, Australia.
Barrios D.B., Angelo E., Goncalves E., (2005) Finite Element Shot Peening Simulation for Residual Stress
Analysis and Comparison with Experimental Results, MECOM 2005, VIII Congress on Computational
Mechanics, Buenos Aires, Argentina. Computational Mechanics Volume XXIV, No. 3.
Beg, D, Korelc, J, Skuber, P, Pavlovi, L, Moe, P, &Kristani, N. (2004), Lecture 9.1: Thin-walled members and
sheeting, from http://www.fgg.uni-lj.si/kmk/ESDEP/master/wg09/l0100.htm.
Bleich F. (1952) Buckling strength of metal structures, McGraw-Hill, New York, p. 167 C92.
Brueggen B, and Ramseyer C., (2005) Capacity of built-up cold-formed steel axial compression
members,Structural stability research council, Proceedings of the Annual Stability Conference, Rolla,
Missouri.
Brueggen B, and Ramseyer C., (2003) Cold formed steel joist member buckling capacity testing, Report for
star building systems, 2003, Oklahoma City. Norman, Oklahoma: Fears Structural Engineering
Laboratory.
Butterworth J., (1999) Finite Element Analysis of Structural Steelwork Beam to Column Bolted Connections,
Constructional Research Unit, School of Science & Technology, University of Teesside, UK.
Chin, H.K.H. (2008), Finite Element Investigation on the Strength of Semi Rigid Steel Connection, Bachelor
Thesis, UniversitiTeknologi Malaysia, Skudai, Johor DarulTakzim, Malaysia.
FEA Co., Ltd. (2002) Lusas Element Reference Manual & Users Manual (version 13.4), United Kingdom.
Fuang, S.L. (2007), Finite Element Analysis on Flush End Plate Beam to Column Bolted Steel Connection on
Major Axis using LUSAS 13.57 Software, Bachelor Thesis, UniversitiTeknologi Malaysia, Skudai, Johor
DarulTakzim, Malaysia.
Lau, H. H. and Ting, T. C. H. T, (2010) LOAD CARRYING CAPACITY OF HIGH STRENGTH
COLD-FORMED STEEL BUILT-UP BOX SECTIONS.Proceeding of the 4
th
International Conference
on Steel & Composite Structures, ICSCS, July 21-23, 2010, Sydney, Australia.pg494-499
Lau, H. H., Ting, T. C. H., F. E. Tang and C. C. Mei, (2010) Application of cold-formed steel built-up sections
in building construction. Proceeding of the 3
rd
International Symposium on Cold-formed Metal
Structures, July 26, 2010, Hong Kong, China.pg69-91

Lau, H.H., and Ting T.C.H., (2009) An Investigation of the Compressive Strength of Cold-formed Steel
Built-up I Sections.Proceedings of the Sixth International Conference on Advances in Steel Structures,
ICASS09, 16-19
th
December 2009, Hong Kong, China, pp. 441-449.
Lee C.A., (2009) Finite Element Analysis of Single Skin Plywood Steel Deck Dry Floor System, Bachelor
Thesis, UniversitiTeknologi Malaysia, Skudai, Johor DarulTakzim, Malaysia.
Lue D-M, Yen T, Liu J-L, and Hsu Y T., (2004) Experimental investigation for buckling strength of
double-channel columns, Proceedings of the annual stability conference, Long Beach, CA.
Lue D.M., Yen T., and Liu J.-L., (2006) Experimental investigation on built-up columns, Journal of
Constructional Steel Research, Vol. 62, 1325C1332.
Liu J-L, Lue D.M., and Lin C.H., (2009) Investigation on slenderness ratios of built-up compression
members, Journal of Constructional Steel Research 2009, Vol. 65, 237C248.
Megnounif, A, Djafour, M, Belarbi, A, &Kerdal, D. (2007), Strength buckling predictions of cold-formed steel
built-up columns, Structural engineering and mechanics, vol. 28, no. 4, pp. 443-60.
Mei, C.C., Lau, H.H. and Choo, A.H., Construction and detailing of cold formed steel space truss, Proceedings
of the 9
th
International Conference on Steel, Space and Composite Structures, Yantai, China, 2007.
NAS (2002): AISI Manual Cold-formed Steel Design, American Iron and Steel Institute, Washington DC.
NAS (2001): North American Specification for the Design of Cold-formed Steel Structural Members, American
Iron and Steel Institute, Washington DC (2001).
Niazi, M, &Rondal, J. (1990), Stability of built-up beams and columns with thin-walled members, Journal of
Constructional Steel Research, vol. 16, pp. 329 - 35.
Omar T.S., Butterworth J.W. & Lowe P.G., (2003), Development of a Finite Element Model for Concrete-filled
Steel Tube Members in Flexure, 10th International Colloquium on Structural & Geotechnical
Engineering, Ain Shams University, Cairo, Egypt.
Salem, A.H., El Aghoury, M.A., Hassan, S.K. and Amin, A.A. (2004), Post-buckling strength of battened columns
built from cold-formed lipped channels, Emirates Journal for Engineering Research, 9(2), pp.117-125.
Schafer, B.W., Pekz, T. (1998).Direct Strength Prediction of Cold-Formed Steel Members using Numerical
Elastic Buckling Solutions.Proceedings of the Fourteenth International Specialty Conference on
Cold-Formed Steel Structures. St. Louis, Missouri. 69-76
Stone T.A., and LaBoube R.A., (2005) Behaviour of Cold-formed Steel Built-up I-sections, Thin-Walled
Structures, Vol. 43, pp. 1805-1817.
Ting, T.C.H. and LAU, H.H., (2011) Compression Test on Cold-formed Steel Built-Up Back-to-Back
Channels Stub Columns. Advanced Materials Research, (Vol 201 203),.pg.2900-2903.
Ting, T. C. H. and Lau, H.H, (2011) A Numerical Investigation on Cold-formed Steel Built-up Back-to-back
Channel Stub Columns.Proceeding of the 7th International Conference on Steel & Aluminium
Structures, ICSAS, July 13-15, 2011, Kuching, Malaysia.pg440-445
Whittle, J. and Ramseyer, C. (2009), Buckling capacities of axially loaded, cold-formed, built-up C-channels,
Thin-Walled Structures, 47, pp.190-201.
Yang D., and Hancock G.J. (2006) Numerical Simulation of High-strength Steel Box-shaped Columns Failing
in Local and Overall Buckling Modes, Journal of Structural Engineering,132(4):5419.
Young, B. and Chen, J. (2008), Design of cold-formed built-up closed sections with intermediate stiffeners,
Journal of Structural Engineering, 134(5), pp.727-737.
Yu, WW. (2000), Cold formed steel design, 3rd edition, Wiley-IEEE, United States of America.
Zandonini R. (1985) Stability of compact built-up struts: experimental investigation and numerical
simulation,Construzionimetalliche, No. 4.

ISBN 978-983-44176-4-2

SUSTAINABILITY OF COLD-FORMED STEEL PORTAL FRAMES
IN MALAYSIA

T. E. McGrath
1
, R. P. D. Johnston
1
, S. Nanukuttan
1
, J. B. P. Lim
1
, M. Soutsos
1
.and C. C. Mei
2

1
School of Planning, Architecture and Civil Engineering, David Keir Building, The Queens
University of Belfast, BT9 5AG, Northern Ireland
2
EcoSteel Sdn Bhd, Demak Laut Industrial Park, 93050 Kuching, Sarawak Malaysia

KEYWORDS

Cold-formed steel, hot-rolled steel, life cycle assessment, life cycle costing, CO
2
emissions

ABSTRACT

This paper attempts to quantify the sustainability advantages of cold-formed steel over conventional
hot-rolled steel for portal frames, using both life cycle assessment (LCA) and life cycle cost (LCC).
Portal frames having spans of 18 m, 24 m and 30 m are designed using both types of steel. If the 18 m
span frame is composed of cold-formed steel instead of conventional hot-rolled steel, it is shown that
it requires 33% less steel. On the other hand, for the 30 m span frame, the cold-formed steel frame is
4% heavier. LCA is primarily concerned with the carbon emissions required for the manufacture of
the steel. The carbon emissions associated with cold-formed steel and hot-rolled steel from different
sources are shown to vary depending on which methodology is used for calculating carbon emissions;
at worst cold-formed steel results in 30% more emissions than hot-rolled steel, for the same weight of
steel, whilst at best both cold-formed steel and hot-rolled steel result in the same emissions. Despite
these differences, when the secondary members and cladding are taken into account in the LCA,
differences in the different types of primary framing are shown to be almost negligible. The LCC is
concerned with not only the cost of the steel, but also with the labour costs and the cost of having a
crane on site; as cold-formed steel is lighter and can be more easily handled on-site. It is shown that
both costs are lower for all three frames and that the cold-formed steel frames of 18 m and 30 m span
are cheaper than the hot-rolled steel frames by 33% and 15%, respectively. The use of the life cycle
assessment has therefore helped quantity associated embodied carbon and energy with the cold-
formed and hot-rolled, with the differences between section types relatively negligible when
considered in context of the entire building with cladding, and the real differences between the two
types of steel due to the ease of erection on site.


INTRODUCTION

Portal frames are most commonly used in the agricultural and industrial sectors, due to their ability to
span long distances and provide open plan column free interiors. Whilst such buildings are normally
designed on the basis of their structural integrity, designers are now increasingly encouraged to think
in terms of life cycle assessment (LCA) and life cycle costs (LCC) when designing all types of
buildings, including portal frames. This change in attitude is evident in both in the UK and Australia.

This paper is concerned with a comparison between the life cycle assessment (LCA) and life cycle
cost (LCC) of cold-formed steel and hot-rolled steel portal frames in Malaysia. In order to undertake
such a comparison, the weight of steel required for each type of material was calculated; estimates
were made for the labour and machinery costs.

Traditionally, portal frames are constructed using hot-rolled steel sections for the primary / load-
bearing members (i.e. columns and rafters), with cold-formed steel being used for the secondary
elements (i.e. purlins, side rails and sheeting). However, for portal frames with more modest spans
(around 18 m) [1], the use of cold-formed steel sections for the primary columns and rafters is a viable
alternative, and buildings completely composed of cold-formed steel are possible [2].

The benefits of such cold-formed steel portal frames have widely been seen in Australia and New
Zealand, where more than half the portal frames constructed are cold-formed steel. With anticipated
design guidance and design software, the use of cold-formed steel for portal frame construction could
be expected to increase significantly over the next 5-10 years. With the climate of Malaysia being
similar to that of Australia and New Zealand, it could be expected that a similar market for cold-
formed steel portal frames could exist in Malaysia.

For frames of modest spans, cold-formed steel has many benefits over hot-rolled steel; high strength-
to-weight ratio of the steel, ease of transportation and ease of erection by manual semi-skilled labour
[3]. Savings on transportation can be achieved due to the inherent ability of cold-formed steel sections
to be efficiently stacked with minimised volume and weight compared to hot-rolled steel. In addition,
hot-rolled steel would require additional fabrication costs, with haunches and purlin cleats being
welded, and the sections requiring painting to prevent rusting. In comparison, cold-formed steel
sections have a pre-galvanised rust resistant surface.

Figure 1. Cold-formed steel portal frame (photograph courtesy of CSB)

In Australia, the Building Code of Australia (BCA) [4] has been one of the main drivers for the use of
cold-formed steel, with their principals being implemented by BlueScope Steel. In BlueScopes
technical booklets [5], the reduce-re-use-recycle system is explained in which it focuses on the
importance of dematerialisation, amongst others. Dematerialisation is a term defined as the
development of high-strength steel products so that the same function is achieved using fewer raw
materials. Such innovations maintain structural integrity and functionality, with lower material use.
With its inherent high strength-to-weight ratio, the use of cold-formed steel in portal frame building
design is a prime example of this.

In the UK, the government has set an ambitious and, legally binding target to reduce national
greenhouse gas emissions [6]. Drivers include The Target Zero programme of work, which provides
guidance on the design and construction of sustainable, low and zero carbon buildings in the UK [7].

As described above, LCA and LCC are two management methodologies that can be used to evaluate
the environmental impacts and costs of a building, respectively. LCA is concerned with environmental
impacts of material choice, starting from the extraction of the raw materials, manufacturing,
production, use or operational phase and finishing with the final disposal. This is known as from
cradle to grave [8]. On the other hand, LCC provides an economic comparison of the proposed
capital investments that are expected to reduce long-term operating costs of the building systems [9].
LCC can lead to more sustainable buildings by helping both the client and designer consider the
designs long-term life-span, as opposed to only considering the building costs at the point of
handover and payment. Both LCA and LCC should be taken into consideration by the client and
designer, in order to gain a better understanding of the whole lifespan variables of a building, and
make a more informed choice.

LIFE CYCLE ASSESSMENT

Life cycle assessment (LCA) is an assessment methodology that quantifies the environmental impacts
of a product, considering the entire life cycle, starting from the extraction of the raw materials,
manufacturing, use phase and eventually the final disposal. There are various types of LCA which
may examine different areas of a products life cycle. For example cradle to gate, examines the
extraction and manufacturing stages of a products life, whilst cradle to grave would include these
steps with the addition of transport, operation/maintenance over life and predictions for eventual
disposal.

The early 1990s saw the development of a formal LCA methodology with broad international
harmonisation, in a series of meetings arranged by the Society for Environmental Toxicology and
Chemistry (SETAC). These meetings brought about the international standardisation of LCA practices
through the ISO 14040 series with work still ongoing from UNEP-SETAC Life Cycle Initiative
Project, the European Commissions JRC Life Cycle Thinking Project and in CEN TC 350
Sustainability of Construction Works. Life cycle assessment, as defined under ISO 14044 [10],
consists of four interlinked phases which are listed and explained in Figure 2. The comprehensive
nature of LCA ensures that burden-shifting from one stage to another, or from one location to
another is prevented. The high recycling rate for steel at the end-of-life (in the UK figures of between
90-99% have been reported [11]) indicate that a cradle to grave system boundary should be selected
as discussed more thoroughly in [11]. However most peer-reviewed LCA figures relate only to cradle
to gate impacts as there is difficulty applying assumptions (reprocessing method, building service life
etc.) which are suitable for all steels over a life cycle. Some figures are presented for both cradle to

gate and cradle to grave system boundaries in Table 3, with the more conservative cradle to gate
figures adopted for modelling the results in this paper.

Figure 2. LCA stages (ISO 14040:2006 altered)

EMBODIED ENERGY AND EMBODIED CO
2eq

Embodied energy and embodied CO
2eq
[12] are two indicators of environmental impact that may be
assessed using life cycle assessment. They are the quantity of energy required or CO
2eq
released by all
of the activities associated with a production process, including the relative proportions consumed in
all activities upstream to the acquisition of natural resources and the share of energy/carbon used in
making equipment and in other supporting functions. The CO
2eq
indicator represents the Kyoto basket
of 6 greenhouse gases, converting methane, nitrous oxide, hydrofluorocarbons, perfluorocarbons and
sulphur hexafluoride to an equivalent CO
2
value using different multiplication factors.

Building materials used in the construction industry have an important environmental impact due to
the energy used and carbon dioxide emitted in the production process [13]. These impacts are often
ignored by designers, although a better choice of materials and construction methods along with
optimisation of the structure can significantly reduce the amount of energy embodied and CO
2eq
in a
building. Building materials which incorporate industrial and consumer wastes, e.g. fly-ash concrete,
can reduce both the depletion of natural resources and the pollution generated by the extraction of the
raw materials.

LIFE CYCLE COSTING

Life cycle costing is standardised under ISO15686:2008 [14] under part 5 of the Building and
Constructed Assets Service Life Planning series and is a methodology used to assess the cost of an

asset or part of an asset over a given life while fulfilling performance requirements. It is an analytical
tool that can be used to identify and measure potential savings due to varying items such as design,
material or energy choices in a building. For life cycle costing to be most effective it should look at
the cost of the building over the life time, not just the initial construction costs, with careful
consideration of the capital cost reducing maintenance, operation and end-of-life cost. Further
discussion of benefits and drawbacks of LCC practices are outlined in [15]. As this paper compares
cold-formed steel to hot-rolled steel portal frame construction, it will be limited to capital cost, with
operational and maintenance cost assumed to be the same for the predicted life cycles.

BUILDING DESCRIPTIONS

In this paper, three spans of portal frame are considered: 18 m, 24 m and 30 m. All three portal frames
have a pitch of 10, height-to-eaves of 4.5m, and a building length of 30 m. For the hot-rolled steel
portal frame, bay centers of 6 m were adopted. For the cold-formed steel portal frame, 6m bay centers
were also adopted, although for the 30 m span frame, 3.5 m bays were also investigated.

Table 1 below shows the estimated construction costs for all three spans, in both cold-formed steel and
hot-rolled steel. The costs shown are used later on in this paper for the calculation of LCC.

TABLE 1
BUILDING COSTING (PERS. COMM. MEI CHEE, ECOSTEEL)
18m 24m 30m
Units Cold-formed Hot-rolled Cold-formed Hot-rolled Cold-formed Hot-rolled
Material costs
Section RM/kg 3.4 3.3 3.4 3.3 3.4 3.3
Purlin RM/kg 3.4 3.4 3.4 3.4 3.4 3.4
Side Rails RM/kg 3.4 3.4 3.4 3.4 3.4 3.4
Cladding RM/m
2
40 40 40 40 40 40
Labour costs
Section RM/kg 2 2.7 2 2.7 2.7 2.7
Purlin RM/kg 2 2 2 2 2 2
Side Rails RM/kg 2 2 2 2 2 2
Cladding RM/m
2
12 12 12 12 12 12
Crane rental costs
Crane (Y/N) N/A 1 x med 1 x small 1 x med 2 x small 1 x large
Crane Cost RM/day 0 1200 700 1200 1400 1800
Crane Dur. Days 0 6 2 8 3 8
Cherry Picker RM/day 400 400 400 400 400 400
Duration Days 4 12 6 14 8 21

Structural Design

Figure 3 shows details of the frame analysis model for the cold-formed steel portal frame. Instead of
haunches, knee braces are used to provide the in-plane stability of the cold-formed steel portal frame
(see Figure 4). The yield stress of the cold-formed steel portal frame was assumed to be 450 N/mm
2
. In

the case of the hot-rolled steel portal frame, grade S275 steel was assumed.

In accordance with design practice in Malaysia, the following the Australian codes of practice were
used:
AS/NZS 1170.0:2002 Structural design actions. Part 0: General principles [16]
AS/NZS 1170.1:2002 Structural design actions. Part 1: Permanent, imposed and other actions
[17]
AS/NZS 1170.2:2002 Structural design actions. Part 2: Wind actions [18].

The following loads were used for the structural design:
Permanent load (G): 0.20 kN/m
2
(purlins, rails, cladding) and self-weight of the members
Imposed load (Q): 0.25 kN/m
2

A basic wind pressure q
u
of 0.656 kN/m
2
was assumed. All buildings were considered to have a
dominant opening, with two values for the internal pressure coefficient, C
pi
, being adopted, namely, -
0.3 and +0.6.

In accordance with AS 1170-0 (2002) [18], each frame was checked at the ultimate limit state for the
following three ultimate load combinations:
ULC1 = 1.2G + 1.5Q
ULC3 = 0.9G + WLC
ULC2 = 1.2G + WLC
It should be noted that ULC3 is used for the uplift wind load combination.

Figure 4 Comparison of haunches used in hot-rolled steel with knee and apex braces used in cold-formed steel

Structural Results

The section costs and weights for each frame are shown in Table 2. The overall weight is shown in
Figure 5.
Figure 3 Geometry of models

TABLE 2
SECTION SIZES USED FOR EACH SPAN (BASED ON 1 PRIMARY PORTAL)

Cold-Formed Steel Hot-Rolled Steel
Span Structure Section Weight (kg) Cost (RM) Structure Section Weight (kg) Cost (RM)
18m Column 2C40025 220 746 Column 254x146 UB 43 395 1302

Rafter 2C40025 458 1556 Rafter 254x146 UB 37 (H) 708 2337

Eaves Strut 2C20015 31 104

Apex Strut C25018 24 82

Total 732 2487

Total 1103 3638





Total 1147 3898

Total 1699 5608





Total 2310 7852

Total 2198 7252

Figure 5. Comparison of weight of primary frames (kg/m
2
)

From Figure 5, it can be seen for both the 18 m and 24 m span frames, the cold-formed steel frame is
lighter than the hot-rolled steel frame by approximately thirty percent. On the other hand, for the 30 m
span frame, the hot-rolled steel frame is slightly lighter than the cold-formed steel frame. However, it
should be noted that if the bay spacing of the 30 m span cold-formed steel frame is reduced from 6 m
to 3.5 m, the cold-formed steel frame becomes more efficient. Figure 6 and Figure 7 show the
buildings weights and cost, respectively. It is interesting to note that although engineers normally
focus on the weight of the main frame, that the cost of the secondary members (purlins, side rails and
cladding) are actually more than that of the main frame.

Figure 6. Weight breakdown of structural components (kg/m
2
)

Figure 7. Cost break down of structural components (RM/m
2
)

SUSTAINABILITY OF THE BUILDING

The life cycle assessment compares the cold-formed steel and hot-rolled steel portal frames using the
cradle to gate boundary, taking into account the environmental impact associated with the extraction
and manufacture of the steel.

Life Cycle Assessment - Material Embodied Energy & Embodied CO
2eq
There are a wide range of values reported in literature for cold-formed and hot-rolled steel sections
with a compilation of some of these values shown in Table 3. This wide range may be attributed to a
number of factors such as different system boundaries, varying recycling rates and data originating
from varying technologies in different locations. For example the World Steel Association [19] figures
appear much lower than the other reported figures as different system boundaries have been selected
with assumptions made about the end-of-life of the materials.


TABLE 3
EMBODIED CO
2EQ
AND EMBODIED ENERGY

Embodied CO2eq
kgCO2eq/kg
Embodied
Energy
MJ/kg
System
Boundary
SECTIONS
ICE Database [15]
Steel Section (ROW
1
Average Recycled
Content 35.5%) [15]
2.12 28.10
Cradle to gate
Coil sheet galvanised (ROW
1
Average
Recycled Content 35.5%)
2.12 29.50
Cradle to gate
Steel Section (UK Average Recycled
Content 59%)
1.53 21.50 Cradle to gate
Coil Sheet Galvanised (UK Average
Recycled Content 59%)
1.54 22.60 Cradle to gate
Yellishetty et al [19]
Basic oxygen furnace(PSP)
2
19.8-31.2 Cradle to gate
Electric arc furnace (PSP)
2
Open hearth furnaces (PSP)
2
Electric arc furnace (SSP)
3
World Steel Association [19]
Sections (HRS) 0.762 13.12 Cradle to grave
Purlins & Rails (CFS) 1.10 19.38 Cradle to grave
CLADDING [15]
Steel coil (galvanised) 2.12 29.50 Cradle to gate
Polyurethane rigid foam 4.26 101.50 Cradle to gate
(ROW)
1
Rest of world
(PSP)
2
Primary steel production
(SSP)
3
Secondary steel production

The embodied energy and CO
2eq
analysis is carried out using cradle to gate values of the embodied
energy and carbon of the material components from the Bath Inventory of Carbon & Energy (ICE)
Database [15] (see Table 3). The ICE data is based on World Steel Association (formerly known as
the International, Iron and Steel Institute). For steel the methodology for recycling is an important
component of determining the embodied impact assessment, as most steel will contain a small amount
of recycled material. There are three main methodologies for recycling, firstly, the recycled content
approach, which allocates the full benefit of material recycling to the input side of a product system
with no benefit of end of life recyclability. Secondly, the substitution method, where the creation of
recyclable material is allocated the full benefit of recycling at the end of life, and lastly, the 50:50
method, that is a combination of the two previous methodologies and allocates half of the benefits of
using recycled materials (start of life) and half of the benefits of creating recycled materials (end of
life recyclability) [15]. The ICE data uses the recycled content approach for the calculations of
embodied energy and carbon. The values of embodied energy and carbon for the materials obtained
take into consideration the three year average rest of world (covering non-EU consumption)
recycling rate of 35.5%.

It should be noted that these figures vary if secondary processes such as cutting, welding, painting,
transportation, were taken into consideration and are used to serve as indicator of performance.

The manufacture of cold-formed steel as described [21] requires the use of virgin materials with low
quantities of residual elements such as copper and steel which have an adverse effect on the steel
properties preventing it from being rolled into thin sections. Thus the recycled content, and use of
recycled scrap, must be limited. Given this limitation it is thought the difference between the
embodied energy and carbon of the cold-formed and hot-rolled sections should be smaller, with

averaging over the industry perhaps not penalising cold-formed steel for having lower recycled steel
content. To allow for this potential difference the Eco-invent database, a comprehensive Swiss-based
database which contains over 4,000 datasets was consulted. For a cradle to gate analysis it reports
scrap contents of 0.12 kg of iron scrap included in each 1 kg of steel (likely to represent cold-formed
steel), a lower recycled steel consumption than other steel sections which has a range of 0.2-1.1 kg of
iron scrap per kg of steel (which would be more suitable for hot-rolled) [21]. This lower scrap content
would indicate that the cold-form steel could potentially have a higher impact than reported in some
places. This is considered in the analysis with the cold form steel modelled with a 35.5% recycled
content rate as shown in Figure 8 & Figure 9 with an error bar used to show the increase in embodied
carbon and energy impact if the recycled content was reduced to, for example, 10%. Cladding material
was based around steel composite panel of 0.5 mm / 0.4 mm internal / external thick galvanised steel
panel with insulation of 40 mm thick.

Figure 8. Embodied energy (MJ/m
2
) of primary structure and cladding per m
2

Figure 9. Embodied CO
2eq
of primary structure and cladding per m
2


Life Cycle Costing: Capital Costing

The life cycle cost considers only the cost of the material, labour and machinery costs. The results are
summarized in Figure 10. As can be seen, when the cost of the crane and labour is considered, the
cold-formed steel portal frame is cheaper than the hot-rolled steel portal frame for all frame spans.

Figure 10. Life cycle cost of frames (RM/m
2
)

CONCLUSION

This paper has identified some of the potential cost advantages of cold-formed steel portal frames over
hot-rolled steel portal frames, both in terms of sustainability as well as cost. Similarly to Australia,
cold-formed steel portal frames are shown to be economic, relative to hot-rolled steel portal frames,
for spans of up to 30 m. Whilst for such a span, the weight of the primary framing is similar to that of
hot-rolled steel, reduced labour and crane hiring costs have resulted in the cold-formed steel portal
frame being 14% cheaper than hot-rolled steel. For the 18 m span building, the cold-formed steel
frame requires 33% less steel and the overall building is also 33% cheaper than hot-rolled steel,
primarily due to the reduced crane hiring cost.

The use of the life cycle assessment has help quantity associated embodied carbon and energy with the
cold-formed and hot-rolled, with the differences between section types relatively negligible when
considered in context of the entire building with cladding.


REFERENCES

[1] Rahman, M., Lim, J. and Xu, Y., Effect of column base strength on steel portal frames in fire,
Proceedings of the Institution of Civil Engineers, Structures and Buildings, vol. 165, Issue SB1,
2011, pp.1-20.
[2] Phan, T.D. and Lim, J., Design optimization of cold-formed steel portal frames taking into
account the effect of building topology, Engineering Optimization, 2012.
[3] Lim, J. and Nethercot D., Design and development of a general cold-formed steel portal frame
system, The Structural Engineer, 5 Nov, 2002.
[4] Office of the Queensland Parliamentary Counsel (2008) Building Regulations 200, Reprinted as
in force on 18 July 2008, Reprint No. 2E. Office of the Queensland Parliamentary Council,
Brisbane, 56pp.
[5] BlueScope Steel, Technical Bulletin No.4, Version 2, 10 May, 2010
[6] Great Britain. Climate Change Act 2008: Elizabeth II. Chapter 27. (2008) London, The
Stationery Office.
[7] Target Zero Guidance on the design and construction of sustainable, low carbon warehouse
buildings, Report v2.0, August 2011.
[8] Ortiz, O., Castells, F. and Sonnemann, G., Operational energy in the life cycle of residential
dwellings: The experience of Spain and Columbia, Applied Energy, vol. 87, 2010, pp.673-680.
[9] Gurung, N. and Mahendran, M., Comparative life cycle costs for new steel portal frame
building systems, Building Research and Information, vol.30(1), 2002, pp.35-46.
[10] ISO 14044 Environmental Management Life Cycle Assessment Requirements and
guidelines, ISO Geneva, 2006.
[11] British Constructional Steelwork Association & TATA Steel (2012) The whole story From
cradle to grave. Available at:
http://www.tatasteelconstruction.com/en/sustainability/the_whole_story/ (Accessed
21/06/2012).
[12] Mari, T.S. and Kuppusamy, S., Consideration of embodied energy of building materials in
local construction, 4th International Conference on Built Environment in Developing
Countries, Malaysia, 2010, p910-921.
[13] Gluch, P. and Baumann, H., The life cycle costing (LCC) approach: a conceptual discussing of
its usefulness of environmental decision-making, Building and Environment, vol.39, 2004,
pp.571-580.
[14] ISO 15686 Buildings & constructed assets Service life planning Part5: Life cycle costing,
ISO Geneva, 2008.
[15] Hammond, G., Jones, C., Inventory of carbon and energy (ICE), Version 2.0, BSRIA BG,
UK, 2011.
[16] AS/NZS 1170.0:2002 Structural design actions. Part 0: General principles, 2002
[17] AS/NZS 1170.1:2002 Structural design actions. Part 1: Permanent, imposed and other actions,
2002
[18] AS/NZS 1170.2:2002 Structural design actions. Part 2: Wind actions, 2002
[19] Yellishetty, M. Mudd, G.M., Ranjith, P.G. & Tharumarajah, A. (2011) Environmental life-cycle
comparisons of steel production and recycling: sustainability issues, problems and prospects.

[20] World Steel Association (2012) Publications - Position papers - Life cycle assessment.
Available at: http://www.worldsteel.org/publications/position-papers/lca.html (Accessed
21/06/2012).
[21] Department of Environment & Climate Change (2011) Energy flows in the UK iron and steel
industry Special features, Digest of United Kingdom Energy Statistics. Available at:
http://www.decc.gov.uk/assets/decc/11/stats/publications/energy-trends/articles/2087-energy-
flows-iron-steel-trends-article.pdf (Accessed 25/06/2012)
[22] Eco-invent database V2.2. Swiss Centre for Life Cycle Inventories. Available at:
http://www.ecoinvent.ch/ (Accessed 21/06/2012).

ISBN 978-983-44176-4-2

A CASE STUDY OF GREENER BUILDING APPROACH -
COLD-FORMED STEEL IS A GREEN AND VIABLE BUILDING
SOLUTION

C.C. Mei
1
and P.K. Chum

2

1
EcoSteel Sdn Bhd, Demak Laut Industrial Park, 93050 Kuching, Sarawak Malaysia
2
BlueScope Steel (Malaysia) Sdn Bhd Bukit Kapar Road, 42200 Kapar, Selangor, Malaysia

KEYWORDS

Cold-formed steel, green, sustainable, reuse, reduce, recycle, light.

ABSTRACT

The resources for building material are finite. The quest for efficient use of material requires serious
and innovative use of material while not sacrificing essential performance quality of buildings.
Numerous guidelines for greener building are devised to achieve that and all based on the 3R
principles. In this paper, we are examining cold-formed steel as a green building material and how
cold-formed steel can be innovated and innovatively applied in various parts of the building so that it
utilises less material and maintain the performance requirements complying to the current design
codes and building regulation.

INTRODUCTION

Today, population growth and increasing consumer demands on depleting resources motivate building

professionals: developer, architects, engineers and contractors strive to minimise activities that deplete
natural resources by carefully selecting construction methods and materials.

In the quest to better plan and use resources, many rating tools and systems have been devised to
provide guidelines to measure sustainability of developments. These rating tools exist to serve
different needs of the country or region. And the primary need is to make all development greener and
sustainable.

Figure 1. Complex System of International Rating Tools (Reed, A et al, 2009)

In Malaysia, the green rating for any development is measured by Green Building Index (GBI). It is
the green rating tool in Malaysia for buildings to promote sustainability in the built environment and
raise awareness among Developers, Architects, Engineers, Planners, Designers, Contractors and the
Public about environmental issues.

The GBI rating tool provides an opportunity for developers and building owners to design and
construct green, sustainable buildings that can provide energy savings, water savings, a healthier
indoor environment, better connectivity to public transport and the adoption of recycling and greenery
for their projects and reduce our impact on the environment. GBI is developed specifically for the
Malaysian-tropical climate, environmental and developmental context, cultural and social needs.

Despite the numerous tools available in different parts of the world and in different countries and also
despite its complexity, all rating tools are typically developed based on the popular 3R principles, ie
Reduce, Reuse and Recycle. This paper demonstrates how cold-formed steel can fit into the principle
of 3R and is a green material to be adopted in any development and to examine the application of
cold-formed steel as green building materials in the Malaysian context generally.

DOES GREEN PAY OFF?

There is scepticism in going green as:
1. there is no perceived market need for green spaces
2. it is costly to go green.
3. There is no end user benefit in going green.

This is a general misconception as within a short span of 3 years of launching GBI in Malaysia, there
are over 26 million square feet of spaces attained at least GBI certification (GBI Malaysia, 2012).
Most of these spaces are commercial and residential. It demonstrates that the demand for green spaces
is real and that there is a real benefit of going green. It also demonstrates that the effort of going green
does not rest solely in the public sectors but the private developer is taking an active and leading role.
This scenario in Malaysia is an echo of the development in Singapore, Europe and America where
many of the private development strive to attain green certification. In Does Green Pay Off? paper,
Miller concludes that as of 2008, there is a demand for green spaces in America (Miller, 2008). This
clearly demonstrate acceptance and hence demand of green properties.

COLD-FORMED STEEL vs MILD STEEL

Steel is a green construction material compared to competing material such as concrete, aluminium
and glass. It does not consume more energy nor result in higher emission to construct a steel frame
building (Guggemos, 2005). In addition, it is the most recyclable construction material. Steel rarely
ended in landfills.

Steel is an alloy made by combining iron and other elements, such as carbon. Mild steel is also
commonly known as low-carbon steel. The typical yield strengths found in Malaysia are 275MPa and
355MPa and are mainly used as structural element.

Cold-Formed Steel (CFS) is the common term for steel products made by rolling or pressing steel thin
gauges sheet into other goods, at room temperature. Cold worked steel products are also commonly
found in all areas of manufacturing of durable goods like appliances or automobiles but the phrase
cold-form steel is commonly used by engineer to describe construction materials. The use of cold-
formed steel construction materials has become more and more popular since its initial introduction of
codified standards in 1946. In the construction industry both structural and non-structural elements are
created from thin gauges of sheet steel.

STEEL PRODUCTION PROCESS IS GETTING GREENER

Reduce in energy consumption


Steel production is energy intensive. The steel-making industry has adopted aggressive position to
deploying sophisticated energy management system to ensure efficient energy use and recovery
throughout the production process. This aggressive initiative halved energy consumption in steel
production since 1975 (World Steel Association, 2008). The quest to make steel production more
energy efficient is a self-fulfilling business process as energy does represent a major cost component
in steel products. To stay competitive as a viable construction material, production cost of steel has to
be kept low and reducing waste and energy consumption is the obvious choice.

Reducing waste - towards Zero Waste

Worldwide steel industry, through management and improvement in the production process, has
dramatically reduced the consumption of raw material in steel making. In the 70's, steel industry
required 144 kg of raw material to produce 100 kg of steel. In 2008, world steel industry produces 1.3
billion tons of steel by utilising 1.48 billion tons of raw materials representing industrial average of
13% losses in material, a far cry as compared to 44% losses almost 40 years ago (World Steel
Association, 2010).

BlueScope Steel has established the same philosophy of considering by-products, rather than waste.
Over the last two decades, numerous uses have been created for the 360 kilograms of by-products that
are produced per tonne of crude steel, which would once have gone to landfill, sewer or been
discharged to the atmosphere. As on going effort, every stage of the steelmaking process is
continuously reviewed to find areas where resource use can be minimised, or where resources can be
sourced in a more sustainable way (Bluescope Steel, 2012).

Down cycle waste material

Material efficiency is a measure of how efficiently a company uses raw materials to produce its
products and by-products in order to maximise the use of precious natural resources and minimise
waste. BlueScope Steel's innovative use of by-products and the implementation of effective waste
management systems mean that material use efficiencies are consistently greater than 95% (Bluescope
Steel, 2012). Recycled slag aggregates has been traditionally use to replace cement in concrete making
process. Addition of slag is known to improve concrete durability and reduction of heat of hydration
of concrete.

Recycling of emitted gas

All BlueScope facilities harness as much energy as is economically feasible from the gases generated
by on-site activities: gaseous by-products are cleaned, and then used to generate energy for the plant.
This not only utilises the by-products of the steelmaking process, but also contributes to resource
sustainability and greenhouse gas (GHG) emission reduction.

Independent on-site power generation means that there is less demand for power from the main grid -
which is likely to be generated using GHG intensive fossil fuels - and non renewable energy sources.
The by-products that are removed from the gases are also utilised: they may be sold as raw-material

inputs for other industries or returned to the steelmaking process.

80% per cent of the energy required to power the coke ovens at Port Kembla Steelworks is derived
from gases produced in the blast furnace. The other 20% of the energy required comes from gases
recovered from the coke ovens themselves.

The Western Sydney Service Centre was one of the first paintlines to have a regenerative thermal
oxidiser in the world. This system converts the gases generated by the paintline process to some of the
energy needed to power the plant. The co-generation plant at New Zealand Steel generates 60% of the
works electricity needs from gases produced in the multi hearth furnace and rotary kilns that were
once flared to the atmosphere (Bluescope Steel, 2012).

COLD-FORMED STEEL IN GREENER BUILT ENVIRONMENT

Steel is the most reused and recycled material in the world comparing to other material such as
aluminium, glass and paper. It is the easiest material to be segregated from other waste, simply by
using magnet. Statistic published by World Recycling Institute shows that nearly all the structural
steel has been recycled (Steel Recycling Institute, 2010). Steel can also be reused, i.e. the same steel
section can be reapplied in other part of building or another building. It rarely ended up into landfill,
unlike concrete and timber (Figure 1). In addition, there are attributes of cold-formed steel that
enhance the application of steel in general in a greener built environment.

Cold-formed steel reduces material use

Cold-formed steel available in the market are typically G450-G550, twice the strength of mild steel
sections or 55% higher strength as compared to G355 steel. It has the highest strength-to-weight ratio
of any structural building material. The higher steel strength results in lower material use and
therefore reducing the life cycle cost (Steel Framing Alliance, 2008).

With extensive research effort in University of Sydney, Roger and Hancock conclude that G550 steel
possesses adequate ductility and that the original equation by Dhalla and Winter was conservative
(Rogers & Ahncock, 1997). The research finding was later incorporated in North America and
Australian design standards enable commercial application of high strength cold-formed steel in
construction industry. This translate directly to material saving if use cold-formed steel in building
construction.

Roofing that was once manufactured at 0.55 mm thick, is today made from high-strength
COLORBOND steel 0.42 mm thick. Structural framing also benefited from that advancement, some
house framing that used to be 1.20-1.60 mm thick is only 0.60-0.80 mm thick today a saving of
33%. These figures were shown in Table 1 for comparison.

ISBN 978-983-44176-4-2

Figure 2. End of life scenario (BCSA, 2011)

ISBN 978-983-44176-4-2

TABLE 1
MATERIAL SAVING AS A RESULT OF HIGHER MATERIAL STRENGTH

Reduces construction waste

Application of cold-formed steel reduces construction waste generated on-site. Cold-formed steel can
be manufactured to exact length, with careful pre-planning, coordination and accuracy in
measurement. In UNIMAS Student Pavilions project, the entire roof was constructed using cold-
formed steel rested on cold-formed steel pipe sections. All structural elements were prefabricated and
pre-cut off site before sending it on site for fitting, generating negligible construction waste on and off
site. The total construction steel waste generated from this project amounted to 1% of the total amount
of steel used. Majority of that waste was attributed from off cuts of roofing and accessories steel
sheets. This waste was send back to the factory to be sorted and categorised to be reused or recycled.

Zincalume steel enhances life-span of steel

Zincalume steel improves the durability of steel in any given environment. Based on test in various
parts of the world, Zincalume steel with AZ150 coating out-perform galvanised steel of Z275. The
relative performance does change with the corrosiveness of the test environment and the exposure
conditions of the test site. Typically, it is found that attic spaces near to beach front with breaking surf
to be the most severe environment. In that condition, Zincalume (AZ150) steel outperformed
galvanised steel (Z275) by at abut 2 times.

For exposed conditions like roof, AZ150 outperformed Z275 by almost 4 times at severe marine
condition as shown in Figure 3 and Figure 4. This test figure agrees well with figure published by
BlueScope Steel in the Technical Bulletin.

Therefore, Zincalume steel improves durability and hence reduces the maintenance need of the
structures. It will also prolong the life-span of the structure. This will ultimately reduce the life-cycle
cost of the building built using the more durable material.

Element
(mm) (mm)
Roofing 0.55 0.42 24%
Framing 1.20 0.80 33%
pre G550
steel
post G550
steel
Material
Saving

Figure 3. Comparison of AZ150 to Z275 life span in roof structure application
(Photo and information courtesy of BlueScope Steel)

Figure 4. Ratio of Average Corrosion Rate of 55% Al-Zn vs Galvanised Coating
(GalvInfo Center, 2007)

Reuse Steel is extremely versatile

Mark Gorgolewski remarked the known fact and the current progressive philosophy that buildings
are huge reservoir of energy and building materials. That philosophy requires us to rethink how we
plan, design, build and demolish our buildings. He proposed that we need to view demolition as a start
of a new material cycle. Mark is proposing a philosophy that sees beyond cradle. To see that
happening, Mark also suggested a few issues that may need to be addressed. That mainly hinge on the
acceptability of the material to be reused (Gorgolewski, 2006).

In Singapore, the Building and Construction Authority has moved forward in a similar direction. The
Authority allows steel strutting material to be reused, with the guideline established for the
practitioners and published by Chiew, (Chiew, 2010).

Generally, cold-formed steel like other steel material can be reused. However, the reused material may
need to be sorted and classified before acceptance for reuse. Damaged element can also be reused with
member cut and reuse as shorter length member. It can also be down-cycle and reuse as secondary
structure or non structural elements.


These are some of the examples how cold-formed steel can be reused.

Figure 5. Off-cuts of sheet metals can be reuse as sun shades which replaces aluminium as a conventional
materials for shading.

Figure 6. Damaged or balance roofing material can be down-cycle and formed into various shapes and
reapplied as alternative material to fencing.

Figure 7. Manufacturing off-cuts, leftovers and coil balance can be down-cycled as secondary structural
members such as architectural sub-frames for architectural panel and fascia

Reduce direct heat load long overhang shading sun

Tropical sun generates heat to buildings and spaces. Sheds will be the most efficient way to reduce the
heat load to the building. Often, architects will design extensive cantilever as building feature as well
as sheds to shield the building against direct sun light. Cold-formed steel possess high strength-to-
weight and stiffness-to-weight ratios. Hence it can be suitable to design for structure of long
cantilever. The cantilever design has been used at the student pavilion at UNIMAS as well as the
Heart and Cancer Specialist Hospital at Samarahan, Kuching.

Reduce transportation and logistic cost

Higher material strength reduces material use. In addition to that, it will also contribute towards
reduction in transportation and logistic cost directly. The benefits are:
1. reduce trips to transport material to site as compared to mild steel structures needed for the
same building or roof areas
2. reducing the crane lifting movements
3. reducing the crane capacity as the weights are lighter or it can be easily bundled into
manageable weight
4. reducing the site storage areas and hence improve site organisation

This indirectly reduces CO
2
emission contributed by the construction activity. This contribution is
even more significant in places where materials are required to be transported over a long overland
journey, such as in Sarawak where goods need to be transported over a greater distance between
Kuching and other major towns.

Figure 8. Six metres overhang reduces direct heat load from sun and yet provide adequate lighting, Student
Pavilion at UNIMAS

Figure 9. Nine metres overhang to create a floating effect of roof


Figure 10. Surau SMK Bandar Samaiang the entire roof covering 18m span was lifted in one 40t crane
movement. The entire lifting process took 20 minutes

Figure 11. Feature cone roof for the port operation building at Tanjung Manis Port roof covering and trusses
of 20 m clear span were assemble on ground and lifted with one crane movement onto the roof top. The entire
lifting process took 35 minutes


I nnovation to improve member strength

Typical sections available for cold-formed steel are lipped C-channel and Z-channel. These sections
are of single symmetry or point symmetry. They are by itself weak in minor axis. This weakness can
be easily overcome by innovation in detailing and built-up of sections improving symmetry of the
single section. Such innovation has been applied successfully in many projects and the improvement
in lateral stiffness is apparent when lifting the truss member in place (Mei et al, 2006).

Figure 12. Back-to-back C sections with a gap to improvement structural strength in lateral torsional buckling

Such improvement in lateral stiffness has not been adequately reflected in the current design codes
and improvement can be made to the codes to further enhance the effectiveness of cold-formed steel
design. Improvement of built-up sections has been captured by researchers and one such research in
Curtin Sarawak concluded that the current provisions in codes for estimated buckling strength of built-
up members are conservative 9Lau et al, 2012).

The sections built-up have been successfully applied in many projects such as:
1. truss spanning 25.2 m over the main sanctuary in Trinity Methodist Church
2. truss spanning 30.0 m over the multi-purpose hall at Curtin University, Sarawak


Figure 13. Erected trusses at the Trinity Methodist Church, Kuching, Sarawak

Figure 14. Trusses over the Multipurpose Hall at the Curtin Unversity, Sarawak Malaysia


CONCLUSION

Steel and especially cold-formed steel is an extremely versatile and flexible construction material. It is
the most recycled construction material and its waste rarely ended up in land fill. The development of
high-strength cold-formed steel enables reduction in construction material used. In addition, the
advance Zincalume technology will also prolong the life-span of the structures and hence reducing the
need of maintenance and replacement. This will reduce the life-cycle cost of the building. High-
strength cold-formed will yield structure with lesser steel materials and hence reduces the embodied
energy as well as the CO
2
emission. Cold-formed steel will produce lighter structures and hence
reducing the CO
2
emission contributed by the construction activities as a result of smaller crane, less
crane movements and reduction in transportation. It is also found that within the provision of the
design standards, there are rooms available to further innovate the use of cold-formed steel making it a
greener construction material.

REFERENCES

Guggemos. A.A ., (2005) Comparison of Environmental Effects of Steel- and Concrete-Framed
Buildings, Journal of Infrastructural System, ASCE, June 2005, pp93-100].
Bluescope Steel (2012) Efficient Production Australia from
http://sustainability.bluescopesteel.com.au/efficient-production
British Constructional Steelwork Association, BCSA (2011), The whole Story From Cradle to
Grave, United Business Media (UBM)
Chiew,S.P. (2010) Extension of BC1: 2008 to Reusable Steel Strutting System
GalvInfo Center, (2007) GalvInfoNote 1.4:, 55% Aluminum-Zinc Alloy Coated Steel Sheet ,
January, International Zinc Association
GBI Malaysia (2012), Greening Malaysia, from http://www.greenbuildingindex.org/index.html.
Guggemos. A.A ., (2005) Comparison of Environmental Effects of Steel- and Concrete-Framed
Buildings, Journal of Infrastructural System, ASCE, June 2005, pp93-100].
Gorgolewski, M., (2006) The Implication of Reuse and Recycling for the Design of Steel Buildings,
Canadian Journal of Civil Engineering, 33.4 pp 489-496
Lau, HH. Ting, T.C.H., (2012)A Parametric Study on the Modified Slenderness Ratio for Cold-
formed Steel Built-up Back-to-back Channel Stub Column, in Proceedings of 7th International
Conference on Advances in Steel Structures, Nanjing, pp263-271]
Mei, C.C., Lau, H.H., Tawil, H., (2006) Long Clear Span of Cold-formed Roof Truss A Structural
Solution, In: S.P. CHIEW, ed. Proceedings of the 8th International Conference on Steel Space &
Composite Structures, 15-17 May 2006 Kuala Lumpur. Singapore: CI-Premier Pte Ltd, 335-342.
Miller. M, (2008) Does Green Pay Off?, Journal of Sustainable Real Estate, American Real Estate
Society.


Reed, A et al, (2009)International comparison of Sustainable Rating Tools, Journal of Sustainable
Real Estate American Real Estate Society.
Rogers C.A., Hancock G.J.(1997), Ductility of G550 Sheet Steel in Tension, Journal of Structural
Engineering, ASCE, Vol. 123, No. 12, pp1586-1594
Steel Framing Alliance (2008), Steel is the New Green Facts About Steel, from
Steel Recycling Institute (2010) 2010 Recycling Rate, from http://www.recycle-
steel.org/en/Recycling%20Resources/Steel%20Recycling%20Rates.aspx
World Steel Association (2008); . World Steel Fact Sheet, - Energy, worldsteel.org, Belgium
World Steel Association (2010); . World Steel Fact Sheet, - The Three Rs, worldsteel.org, Belgium

BES 2012 E-Proceeding 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

BES 2012 E-Proceeding PDF

Uploaded by

Copyright:

Available Formats

Cold-formed Steel Innovations

Green & Viable Structural Solutions

You might also like