You are on page 1of 14

Thin-Walled Structures 116 (2017) 180–193

Contents lists available at ScienceDirect

Thin-Walled Structures
journal homepage: www.elsevier.com/locate/tws

Full length article

Compression tests of built-up cold-formed steel hollow flange sections



MARK
Sivakumar Kesawan, Mahen Mahendran , Yomal Dias, Wen-Bin Zhao
Queensland University of Technology (QUT), Brisbane, Australia

A R T I C L E I N F O A B S T R A C T

Keywords: Cold-formed steel hollow flange sections are structurally efficient compared to the conventional open channel
Cold-formed steel structures sections. They can be of different forms including hollow flange I- and channel sections, which are generally
Built-up sections fabricated by first cold-forming a single sheet to form the desired shape and then connecting their flange and
Screw fastened web elements by continuous welding. Instead of welding, intermittently screw/rivet fastened or spot welded
Rivet fastened
connection can also be used. A detailed investigation is currently under way on the structural performance
Hollow flange sections
characteristics of such intermittently screw/rivet fastened steel hollow flange sections. This paper presents the
Direct strength method
Effective width method details of an experimental study performed to investigate the behaviour of built-up cold-formed steel hollow
flange sections in compression, where more than 45 stub columns were tested. Both hollow flange I- and channel
sections built-up using either single steel sheet or three steel elements were tested. Tests also included built-up
hollow flange channel sections with stiffened web elements. Test results including the failure loads, load versus
axial deformation plots and failure characteristics are presented in this paper. Test failure loads are compared
with the predictions based on both effective width and direct strength methods to verify their adequacy in
predicting the section compression capacity of built-up hollow flange sections. This paper has shown that
although hollow flange I-sections made of three elements with a 100 mm fastener spacing can perform similar to
continuously welded sections, hollow flange channel sections undergo premature failures even with a 50 mm
fastener spacing, and the reasons are discussed.

1. Introduction However, it was discontinued due to its complicated manufacturing


process involving simultaneous cold-forming and electric resistance
The usage of cold-formed steel structures in the building construc- welding [2].
tion industry is rapidly increasing due to their potential benefits LitesSteel Beam (LSB) is the recently invented hollow flange channel
including high strength to weight ratio, rapid constructability and ease (HFC) steel section developed by OneSteel Australian Tube Mills (OATM)
of transportability. To date, open Z-, unlipped or lipped channel using an improved cold-forming and dual electrical resistance welding
sections are the most commonly used cold-formed steel sections. The process (Fig. 1(b)). This section was extensively used in Australia since
open and thin-walled nature of these sections makes them suffer from 2002, and is on average 40% lighter than the traditional hot-rolled
local and distortional buckling effects, which ultimately reduce their structural sections of equivalent bending strength. Kesawan and Mahen-
ultimate capacities and moreover result in complex capacity calcula- dran [3–5] and Jatheeshan and Mahendran [6,7] demonstrated the
tions. Recently many attempts have been made to eliminate the improved fire performance of Light gauge Steel Frame (LSF) wall and
occurrence of such buckling modes and improve the load carrying floor systems when these cold-formed and fully welded HFC sections/
capacities of cold-formed steel members. The hollow flange section is LSBs were used as load-bearing studs and joists. However, LSB has also
one of the outcomes of such attempts. been discontinued recently, due to the changes in OATM's business
In early 1990s, a hollow flange section with two triangular flanges operations and expensive welding process. Instead of the continuous
(Fig. 1(a)), known as Hollow Flange Beam (HFB), was first introduced electric resistance welding used to connect the web and flange elements,
in Australia. This section has improved buckling capacities due to its other techniques such as spot welding or rivet/screw fastening can be
unique shape where its distortional buckling is eliminated by the adapted to form hollow flange sections. Fig. 2(a) and (b) show the new
torsionally rigid hollow flanges while its local buckling capacity is rivet/screw fastened or spot welded rectangular hollow flange I-sections
improved due to the absence of free edges and reduced web width. This (HFI) while Fig. 2(c) and (d) show the hollow flange channel (HFC)
cold-formed steel section is considered to have the advantages of both sections. However, the structural behaviour of these built-up hollow
hot-rolled I-section and conventional cold-formed C- and Z-sections [1]. flange sections has not been investigated adequately.


Corresponding author.

http://dx.doi.org/10.1016/j.tws.2017.03.004
Received 26 August 2016; Received in revised form 22 December 2016; Accepted 5 March 2017
Available online 28 March 2017
0263-8231/ © 2017 Elsevier Ltd. All rights reserved.
S. Kesawan et al. Thin-Walled Structures 116 (2017) 180–193

Outer flange From the introduction of cold-formed hollow flange sections in the
1990s, many research studies have been undertaken, but they were
Inner flange limited to their bending and shear studies [8–10]. Yi and Wilkinson
[11] and Kesawan and Mahendran [12] were the first to conduct an
extensive investigation on the compression capacity of cold-formed
Weld Web Weld
steel hollow flange sections. Their tests were also limited to welded
hollow flange channel sections (LSBs). Since reliable research data on
the structural performance of built-up hollow flange sections is not
available, detailed investigations are now being undertaken on their
structural behaviour at the Queensland University of Technology. As
part of these investigations, this study focuses on the section compres-
sion capacity of built-up hollow flange sections.
Fig. 1. Welded hollow flange sections, (a) Hollow Flange Beam (HFB), (b) LiteSteel Beam The built-up cold-formed HFI and HFC sections can be made using
(LSB).
either single steel sheet or three steel elements. The manufacturing
process of HFC sections using a single steel sheet is complex whereas
they can be easily made of three steel elements with their two hollow
flange elements connected to the web element through rivet/screw
fastening or spot welding. The significant advantage of the built-up
sections made of three steel elements is that the web and flange
elements can be of different steel grades and thicknesses (Fig. 2(b)
and (d)), for example, web element can be thicker than flange elements.
Furthermore, different steel grades can be used for web and flange
elements depending on the requirements. These enable the designers to
choose from a variety of options based on their requirements, thus
leading to an optimum design. However, it is critical to determine the
required rivet/screw spacing to minimise any loss in the section
capacity due to intermittent fastening. Although many studies
[13–19] have been undertaken in the past on the structural behaviour
of built up I- and box sections, no such attempts have been made on
built-up hollow flange sections.
To obtain reliable structural performance data and to develop
suitable design guidelines for built-up hollow flange sections in
compression, an experimental study was performed where more than
45 stub columns were tested. This paper presents the details of this
experimental study conducted to investigate the structural behaviour
and capacity of built-up hollow flange I- and channel section stub
columns subject to local buckling effects. The ultimate failure loads of
tested columns were compared with the predictions based on the
effective width and direct strength design methods in AS/NZS 4600
[20] in order to verify their applicability to built-up hollow flange
sections.
Past research studies [21–24] and AISI S100 [25] recommend
suitable connector spacing requirements for cover plated sections based
Fig. 2. Screw/Rivet fastened hollow flange sections, (a) HFI section made of single steel on three criteria, (1) the connector should be able to transmit the shear
plate, (b) HFI section made of three steel elements, (c) HFC section made of single steel forces between the connected parts, (2) connector spacing should be
plate, (d) HFC section made of three plate elements. higher than 1.16t E / fc to avoid column like buckling behaviour of the
*Note: Bf is the flange width, hf is the lip length, Hw is the web depth, tw is the web plates, and (3) connector spacing should be higher than a spacing of
thickness, tf is the flange thickness, Bl is the length of the outstanding element. three times the flat width of the narrowest unstiffened compression

Spot weld
spacing

10 mm

(a) (b)
Screws Spot-welds

Fig. 3. Screw fastened and spot welded 1PHFIs (a) Screw fastened, (b) Spot welded.
181
S. Kesawan et al. Thin-Walled Structures 116 (2017) 180–193

Fig. 4. Screw Fastened HFIs (a) 1PHFI, (b) 3PHFI.

Table 1
Ultimate failure loads of spot welded 1PHFIs.

Specimen Experimental failure load PEX (kN) AS/NZS 4600 predictions (kN) Comparisons

PEWM PDSM PDSM-mod PEX/PEWM PEX/PDSM PEX/PDSM-mod

180701PHFI12s100 168 167 126 151 1.01 1.33 1.11


180701PHFI12s200a 169 162 123 148 1.04 1.37 1.14
180701PHFI12s200b 165 163 124 149 1.01 1.33 1.11
180701PHFI12s400 160 161 124 148 0.99 1.29 1.08
180701PHFI16s100 227 219 191 211 1.04 1.19 1.08
180701PHFI16s200 225 217 191 210 1.04 1.18 1.07
180701PHFI16s400 215 219 191 211 0.98 1.13 1.02
180701PHFI20s100 298 278 265 276 1.07 1.12 1.08
180701PHFI20s200 296 272 263 273 1.09 1.13 1.08
180701PHFI20s400 291 272 263 272 1.07 1.11 1.07

Table 2
Ultimate failure loads of screw fastened 1PHFIs.

Specimen Experimental failure load PEX (kN) AS/NZS 4600 predictions (kN) Comparisons

PEWM PDSM PDSM-mod PEX/PEWM PEX/PDSM PEX/PDSM-mod

180701PHFI12s100-1 170 161 124 149 1.06 1.37 1.14


180701PHFI12s100-2 169 162 124 149 1.04 1.36 1.13
180701PHFI12s200-1 167 161 123 148 1.04 1.36 1.13
180701PHFI12s200-2 157 163 124 149 0.96 1.27 1.05
180701PHFI12s400 155 162 123 148 0.96 1.26 1.05
180701PHFI16s100 235 220 192 212 1.07 1.22 1.11
180701PHFI16s200 223 219 192 212 1.02 1.16 1.05
180701PHFI16s400 217 219 191 211 0.99 1.14 1.03
180701PHFI20s100 298 276 264 274 1.08 1.13 1.09
180701PHFI20s200 295 276 265 276 1.07 1.11 1.07
180701PHFI20s400 297 270 261 271 1.10 1.14 1.10

Table 3
Ultimate failure loads of screw fastened 3PHFIs.

Specimen Experimental failure load PEX (kN) AS/NZS 4600 predictions (kN) Comparisons

PEWM PDSM PDSM-mod PEX/PEWM PEX/PDSM PEX/PDSM-mod

160703PHFI1212s100 177 164 136 157 1.08 1.30 1.13


160703PHFI1212s200 170 167 138 158 1.02 1.23 1.08
160703PHFI1212s400 162 166 137 158 0.98 1.18 1.03
160703PHFI1220s100 194 191 172 188 1.02 1.13 1.03
160703PHFI1220s400 166 192 172 189 0.86 0.97 0.88
180703PHFI1616s100 227 227 196 217 1.00 1.16 1.05
180703PHFI1616s200 213 223 194 215 0.96 1.10 0.99
180703PHFI1616s400 197 227 196 217 0.87 1.01 0.91
180703PHFI1620s100 239 241 232 245 0.99 1.03 0.98
180703PHFI1620s200 251 240 232 245 1.05 1.08 1.02
180703PHFI1620s400 189 240 232 245 0.79 0.81 0.77

182
S. Kesawan et al. Thin-Walled Structures 116 (2017) 180–193

sections with larger connector spacing than given in AISI S100 [25].
Pin support Although the recommended spacing requirements in [21–25] are for
cover plated beam sections, the connector spacings used in the built-up
hollow flange sections are compared in this paper with the minimum
spacing specified in AISI S100 [25].

2. Experimental study of built-up hollow flange I-section stub


columns

Test 2.1. Hollow flange I-section specimens


specimen
Built-up hollow flange I-sections made using single steel sheet or
three steel elements (1PHFI and 3PHFI – Fig. 2(a) and (b)) were tested
in this study (32 tests). In these sections, the key parameters such as
Pin section dimensions, web to flange connection and its spacing were
support varied. The tested stub columns can be mainly categorised into two
groups,

1. Spot welded and screw fastened 1PHFI-sections


2. Screw fastened 3PHFI-sections
Fig. 5. Test set-up of HFI specimens.
In these stub columns subject to local buckling effects, the specimen
length was chosen as approximately three times the width of the widest
plate element. The 1PHFI specimens were first brake pressed from zinc
coated Grade G250 structural steel sheet in the form of hollow flange
section and then its web and flange elements were connected. For
3PHFI specimens, both flanges were separately brake pressed and then
connected to the web element. The built-up 1PHFI and 3PHFI members
are given the following designation: 180701P(3P)HFI1620s100, where
180 – section depth, 70 – flange width, 1P(3P)HFI – rectangular hollow
flange I-section made of one steel sheet or three steel elements, 16 – flange
thickness of 1.6 mm, 20 – web thickness of 2.0 mm, s100 – fastener spacing
of 100 mm.
The spot welding connecting the web and flange elements was
performed using spot-welding guns, where the spot-weld diameter used
was 5 mm. In screw fastened 1PHFI and 3PHFI sections 6 mm long
5 mm diameter hexagonal head self-drilling screws were used. The first
Fig. 6. Column failures (a) Specimen without end plate (180701PHFI12s200a), (b) connection – screw fastening or spot welding, was placed at a distance
Specimen with end plate (180701PHFI12s200b). of 10 mm from the specimen ends. Fig. 3(a) and (b) show the 1PHFI
specimens with screw fastened and spot welded connections, while
Fig. 4(a) and (b) show the typical screw fastened 1PHFI and 3PHFI
element. LaBoube et al. [24] showed that Yener's [21] recommenda- sections, respectively.
tions for connector spacing are unsafe for some cover plated sections, In 180701PHFI and 180703PHFI sections made of 2 mm thick
and concluded that AISI S100 [25] spacing requirement is adequate to plates, the flange width, flange height and lip length used were 70,
avoid column like buckling behaviour of plates. They proposed suitable 30 and 15 mm, respectively, and hence these elements are unlikely to
equations to determine the bending capacity of the cover plated undergo local buckling. However, the web element length was 120 mm,

400 mm 100 mm
200 mm

(a) (b) (c)


Fig. 7. Failure modes of 180701PHFI12s with varying screw spacing, (a) 400 mm, (b) 200 mm, (c) 100 mm.

183
S. Kesawan et al. Thin-Walled Structures 116 (2017) 180–193

Fig. 9. Comparison of load versus axial deformation curves of 3PHFI sections (a)
160703PHFI1212, (b) 160703PHFI.
Note: 1620s100 – flange and web thicknesses are 1.6 and 2.0 mm, respectively, and
intermittent fastener spacing is 100 mm.

Table 4
Effect of fastener types on the failure loads of 1PHFIs.

Specimen Pu-SP (kN) Pu-SC (kN) Pu-SP/P u-SC

180701PHFI12s100 168 170 0.99


180701PHFI12s200 167 162 1.03
180701PHFI12s400 160 155 1.03
180701PHFI16s100 237 235 1.01
180701PHFI16s200 225 223 1.01
180701PHFI16s400 215 210 1.02
180701PHFI20s100 298 298 1.00
180701PHFI20s200 296 295 1.00
180701PHFI20s400 291 297 0.98
Mean 1.01
COV 0.02

Note: Pu-SP, Pu-SC – ultimate failure load of spot welded and screw fastened 1PHFIs.

Fig. 8. Comparison of load versus axial deformation curves of 1PHFI sections, (a) Spot- spherical pin support with one load bearing plate was used at the top
welded 180701PHFI12, (b) Spot-welded 180701PHFI20, (c) Screw fastened end while a ball-bearing pin support was used at the bottom end. The
180701PHFI12. specimen was centrally placed in the loading set-up and then the load
was gradually applied at the top end. At failure, the column could not
thus it could undergo local buckling deformations. For 180701PHFI and sustain the applied load.
180703PHFI sections made of 1.2 mm thick plates, both web and flange One test was conducted for a spot welded 1PHFI column with rigid
local buckling were expected to occur. Details of the tested columns are plates attached to both ends in order to investigate the necessity of rigid
given in Tables 1–3. plates at the column ends. Two spot welded short columns
The steels used in the test series have a nominal yield strength of (180701PHFI12s200a - without end bearing plates and
250 MPa. Tensile tests were conducted on steel coupons cut from these 180701PHFI12s200b - with end bearing plates) were tested. Similar
sheets according to AS1391 [26]. Steels of 1.2, 1.6 and 2.0 mm elastic local buckling behaviour was observed in both of these tests, and
thicknesses had yield strengths of 310, 290 and 270 MPa, respectively, the ultimate failure loads (169 and 165 kN) were similar as evident
while the elastic modulus of all steels was found to be about from Table 1. The failure modes of columns shown in Fig. 6(a) and (b)
200,000 MPa. also agree well. The applied load versus axial deformation plots for both
columns are presented in Fig. 8(a), which agree well. Therefore, the
2.2. Test method and results remaining columns were tested without using the end plates.
The built-up 1PHFI stub columns were tested using the same
Fig. 5 shows the test set-up in the stub column tests. A semi- procedures mentioned above, and the results of the spot-welded and

184
S. Kesawan et al. Thin-Walled Structures 116 (2017) 180–193

Fig. 10. Local buckling failure modes of 1PHFI specimens, (a) Spot welded, (b) Screw fastened.

Table 5 screw fastened sections are summarised in Tables 1, 2. In these sections


Effect of fastener spacing on the failure loads 3PHFIs. local buckling waves developed in the web and flange elements after a
certain load, followed by the test column failure by yielding of their
Specimen Pu-s100 (kN) Pu-s400 (kN) Pu-s100/P u-s400
web and flange elements as shown in Fig. 7. Fig. 8(a)–(c) present the
160703PHFI1212 177 162 1.09 load versus axial deformation plots obtained from the tests. Their
160703PHFI1220 194 166 1.17 behaviour is typical for all the 1PHFI sections considered in this study.
180703PHFI1616 227 197 1.15 Fig. 8 shows that in general the elastic stiffness of both spot welded and
180703PHFI1620 239 189 1.27
screw fastened sections reduces with increased spacing.
Note: Pu-s100, Pu-s400 – ultimate failure loads of 3PHFIs with 100 and 400 mm screw Table 3 presents the results of the built-up hollow flange I-sections
spacing, respectively. made of three steel elements (3PHFI sections) with screw fastening

Fig. 11. HFC specimens, (a) 90HFC, (b) 150HFC, (c) 180 HFC, (d) Typical rivet spacing, (e) Fabricated specimens.

185
S. Kesawan et al. Thin-Walled Structures 116 (2017) 180–193

Instron 125
machine
62.5 60
Top loading 30
arm 45
15
280 155
45
46.5
30
Stub column
62.5
Pyrocrete

(b)

(a)
Fig. 12. Test set-up of HFC specimens (a) Stub column under load, (b) End Plate for 150 HFC sections.

connections. Four 3PHFI sections were considered where in one of them the screw spacing was increased to 400 mm, i.e., the difference was
(180703PHFI1620s100), web and flange elements were of different about 9% when the fastener spacing was increased from 100 to 400 mm
thicknesses (1.2 and 2 mm). The load versus axial deformation plots of in 180701PHFI12 columns. For 180701PHFI20 sections the ultimate
3PHFI columns are shown in Fig. 9, which reveal that the elastic capacities were about the same, irrespective of the screw spacing. This
stiffness of 160703PHFI1212 sections reduces with increased screw was because of the thicker web and flange elements. Based on the
spacing, while that of 160703PHFI sections is about the same irrespec- results, either 100 or 200 mm spacing can be used in the built-up 1PHFI
tive of screw spacing. columns. However, this study conservatively recommends a screw or
spot weld spacing of 100 mm.
2.3. Discussion of test results Unlike the ultimate capacities of built-up 1PHFI sections, there are
noticeable drops in the section capacity of built-up 3PHFI sections when
In this section the effects of web-flange connectivity and its spacing the screw spacing is increased from 100 mm to 200 mm, which is about
on the capacity of 1PHFI and 3PHFI section columns are discussed. 6% for 180703PHFI12 and 180703PHFI16 sections. Table 5 shows the
section capacity ratio of the 3PHFI sections with 400 mm screw spacing
2.3.1. Effects of connection types in comparison to those with 100 mm spacing. The ratios are higher for
Two types of connections, spot welding and screw fastening, were 3PHFI sections with 400 mm spacing, with a difference of more than
used to connect the web and flange elements of 1PHFI sections. Table 4 10% in most cases. The maximum difference is about 27% in
compares the failure loads of spot welded and screw fastened 1PHFI 180703PHFI1620 section columns. These results demonstrate that the
section columns, which shows that the differences between the failure reduction in compression capacity due to increasing screw spacing is
loads are very small. Table 4 also presents the mean value of the Pu-sp higher for 3PHFI sections compared to 1PHFI sections. This is because
(ultimate capacity of spot welded 1PHFI) to Pu-sc (ultimate capacity of 1PHFI section is made of single steel sheet where the web is connected
screw fastened 1PHFI), which is 1.01, and the coefficient of variation, to the flange from one side as the other side is similar to fully welded
which is 0.02. This indicates that both fastening types can provide the conditions (Fig. 2(a)). However, in 3PHFI sections, the web needs to be
same compression capacity of built-up 1PHFI sections. Importantly, the connected to flanges on both sides (Fig. 2(b)), and hence this connec-
fasteners did not fail during the tests. These conclude that the spot tion should be more rigid than that in 1PHFI section as the web element
welds and screw fasteners used are adequate to provide the required has a higher tendency to behave separately from flanges. If the screw
connectivity between the web and flange elements. Fig. 10 shows the spacing increases, the web element would tend to behave like an
local buckling failure modes of spot welded and screw fastened element unsupported along both edges, and the compression capacity
180701PHFI sections, which are very similar. will reduce. Therefore, a connection spacing of 100 mm or less is
recommended for 3PHFI sections. Furthermore, it is yet to be verified
2.3.2. Effects of screw spacing whether even 100 mm spacing provides a restraint condition equivalent
The connection spacing could affect the buckling behaviour and to continuously welded connections (Section 2.4.1).
capacity of the built-up 1PHFI sections. Hence connection spacing was
varied (100, 200 and 400 mm) with an aim to find the optimum 2.4. Comparison of ultimate capacities from experiments and steel design
spacing. From Tables 1–3 and the discussion in the last section on spot standards
welded or screw fastened 1PHFI sections, the small difference in the
ultimate capacities of spot welded or screw fastened 1PHFI sections was 2.4.1. Effective width design method
when screw/spot welding spacings of 100 and 200 mm were used. Kesawan and Mahendran [12] conducted a detailed study on the
However, in most of the cases the capacity significantly reduced when section compression capacity of welded hollow flange channel sections

186
S. Kesawan et al. Thin-Walled Structures 116 (2017) 180–193

is adequate to provide the required edge restraints. The calculation


procedure is the same in Eurocode 3 Part 1.3 [27] and AISI S100 [25].
The ultimate capacities of the tested columns were calculated using
the effective width method based on the measured centre line dimen-
sions. Since the elements are thin-walled in nature, the small corner
radius was neglected in the design calculations. The web element was
assumed to be fully supported along both longitudinal edges. The web
length was taken as the distance between the web to flange connec-
tions, with a uniform base metal thickness in the effective area
calculations. Furthermore, the length of the outstand element (lip)
was taken as the length from the lip-flange junction to the lip end.
Table 1 compares the experimental ultimate capacities with the
design rule predictions for spot welded 1PHFI sections. The design rule
predictions are slightly conservative for 1PHFI sections with 100 and
200 mm connection spacing while they are slightly unconservative for
1PHFI sections with 400 mm connection spacing.
This observation was the same for screw fastened 1PHFI sections
(Table 2). For 3PHFI sections, the predictions are on the safe side for
100 mm screw spacing (Table 3). In the case of 200 mm screw spacing,
the predictions are slightly conservative for 16703PHFI12 and
160703PHFI20 sections while they are unconservative for
160703PHFI16 sections. The design rule predictions are unconservative
when 400 mm spacing is used, with a difference of more than 12%.
Based on these observations it can be concluded that the effective width
method can be successfully used to predict the section compression
capacity of built-up 1PHFI members with a fastener spacing equal to or
less than 200 mm, and built-up 3PHFI members with a fastener spacing
equal to or less than 100 mm. Since the effective width method
successfully predicts the section compression capacity of the built-up
3PHFI members with 100 mm fastener spacing, it can also be concluded
that 100 mm fastener spacing is adequate for these sections.

2.4.2. Direct strength method


The direct strength method (DSM) has been increasingly used in the
design of cold-formed steel members due its simplicity in dealing with
complex buckling problems. The DSM design rules are given in
Appendix E of AS/NZS4600 [20] and AISI S100 [25]. DSM considers
the global, local and distortional buckling instabilities, and their
interactions and post-buckling capacities [28,29]. To date this method
has been verified to predict the structural capacities of various cold-
formed steel sections including lipped channel sections with and
without stiffeners and Z-sections [29], but its applicability is still under
investigations for hollow flange sections. Kesawan and Mahendran [12]
investigated their suitability for welded hollow flange channel (HFC)
sections and concluded that DSM equations cannot be used in the
present form as they are too conservative for cold-formed and fully
welded HFC sections with uniform yield strengths. Hence they proposed
modifications to the equations, which were limited to welded HFC
sections of uniform thickness and yield strength. However, there is a
need to investigate whether the modified DSM equations can be applied
to screw fastened HFI/HFC sections. The DSM equations in Appendix E
of AS/NZS4600 [20] and the modified DSM equations [12] are
summarised in Appendix A of this paper.
The local buckling capacities of the HFI sections were obtained from
the buckling analyses using ABAQUS. For this purpose it was assumed
Fig. 13. Comparison of load versus axial deformation curves for HFC sections (a) that the web-flange connection is continuous along the length. Since
HFC12G30012G300, (b) HFC12G300115G500, (c) HFC12G300190G450, (d) these are very short columns, the Pne is assumed to be equal to Py (see
HFC12G300055G300, (e) HFC12G30012G300w. Appendix A). Tables 1–3 provide the DSM predictions and their
comparisons with the experimental results. The comparisons show that
(LSBs), and showed that their capacities can be successfully predicted there are large variations between the test results and DSM predictions,
using the effective width method (EWM) given in AS/NZS 4600 [20] with the predictions being on the conservative side. These tables also
and Eurocode 3 Part 1.3 [27]. The following section evaluates whether compare the predictions using the modified DSM equations [12] with
these design rules are suitable to predict the section compression the test results, and their agreements are improved.
capacity of built-up 1PHFI and 3PHFI sections. Furthermore by Based on the AISI S100 [25] spacing requirements, to avoid column
comparing the AS/NZS4600 design rule predictions of 1PHFI and like buckling behaviour, the minimum connector spacings for the 1.2,
3PHFI sections, it can be verified whether the connection spacing used 1.6 and 2.0 mm thick web plates are 45, 63 and 81 mm, respectively,

187
S. Kesawan et al. Thin-Walled Structures 116 (2017) 180–193

Local buckling of
complete web
element

(a) (b) (c) (d)


Fig. 14. Failure progression of HFC12G300115G500 column (a) Initial condition, (b) At ultimate load, (c) and (d) Post-ultimate load.

Table 6
Ultimate failure loads of rivet-fastened HFCs.

Specimen Experimental failure AS/NZS 4600 Comparison


load PEX (kN) prediction
PEWM PEX/PEWM

180HFC12G30012G300 112 173 0.65


150HFC12G30012G300 84 172 0.49
90HFC12G30012G300 85 164 0.52
180HFC12G300115G500 107 181 0.59
150HFC12G300115G500 117 180 0.65
90HFC12G300115G500 106 173 0.61
180HFC12G300190G450 133 236 0.56
150HFC12G300190G450 113 233 0.48
90HFC12G300190G450 104 201 0.52
180HFC12G300055G300 82 148 0.55
150HFC12G300055G300 94 148 0.64
90HFC12G300055G300 96 147 0.65
Fig. 15. Comparison of the failure modes of Rivet Fastened and Welded HFC columns (a) 180HFC12G30012G300w 164 173 0.95
Rivet Fastened 150HFC12G30012G300, (b) Welded HFC [11]. 150HFC12G30012G300w 162 172 0.94

while they should also be limited to 45 mm based on the unstiffened 3. Experimental study of built-up hollow flange channel section
plate width (Bl in Fig. 2(a) and (b)). However, although larger stub columns
connector spacings (100–400 mm) were used in the tested intermit-
tently screw fastened or spot welded HFI sections, they were still able to 3.1. Hollow flange channel section specimens
behave like a fully welded member. This is because unlike cover plated
hat sections, the plate element is sandwiched between the two This section presents the experimental study conducted to investi-
unstiffened elements of the hollow flanges (Figs. 2(a) and 4), and thus gate the behaviour of built-up HFC stub columns. For the HFC sections,
allowing larger connector spacing to be used. rivet fastening technique was used to connect the web and flange
Overall this experimental study and the comparisons of their results elements instead of screw fastening due to its several advantages. The
with design rule predictions have shown that the behaviour of inter- HFC sections considered in this study were made of three elements, two
mittently spot welded and screw fastened 1PHFI or 3PHFI sections is rectangular hollow flange elements and one web element. As discussed
similar to continuously welded HFI sections if the web to flange earlier, these built-up sections can be made of flange and web elements
connections are provided at a suitable spacing of about 100 mm or of varying steel grades and thicknesses, and this option was also
less. In these built-up I-sections, the maximum plate slenderness explored in this experimental study.
⎛ ⎞ The fabrication process of the test sections commenced with the
⎜b / t fy /250 ⎟ of the web element and the unstiffened element of the
⎝ ⎠ formation of two rectangular flanges of 51×20 mm from 1.2 mm G300
flange are 100 and 14, respectively. The recommendation of 100 mm cold-formed sheets (Fig. 11(a)–(e)). These flange sizes were restricted
connector spacing is applicable to HFI sections with their plate elements by the available fabrication facilities. In all the tested stub columns, the
having less than or equal to the above slenderness limits. Further flange sizes, column length (230 mm) and the steel grade were the same
investigations are needed to determine the required connector spacing while the depth, thickness and grade of the web element were varied.
for sections outside these limits. Furthermore, it is not known whether The required web plates were cut from the sheet and then connected to
the 100 mm connector spacing recommendation will be equally applic- the flanges using 6 mm diameter self-pierce rivets. Self-pierce riveting is
able to the intermittently screw fastened HFC sections (Fig. 2(c) and a cold-joining process where two or more sheets are connected by
(d)). Hence an experimental study of built-up hollow flange channel driving a rivet through the top sheet and upsetting the rivet, under the
sections was also undertaken and its details are given in the next influence of a die, into the lower sheet without piercing it. Importantly
section. this is a non-thermal process hence it does not induce membrane
residual stress in the specimens.
The cross-sectional depth of the built-up HFC member with rivet

188
S. Kesawan et al. Thin-Walled Structures 116 (2017) 180–193

180 mm

150 mm

90 mm

Local buckling of
complete web element

Rivet
failure

Fig. 16. Failure modes of HFC columns. (a) HFC12G30012G300 (b) 180HFC12G300190G450 columns.

sections are shown in Fig. 11(e), and are identified by the following
designation: 150HFC12G50016G350, where 150 – section depth, HFC –
hollow flange channel section made of three elements (3P was not used since
all the tested columns were made of three steel elements), 12 – flange
thickness of 1.2 mm, G500 – steel grade of flange element, 16 – web
thickness of 1.6 mm and G350 – steel grade of web element.
Tensile tests were also conducted on specimens cut from the 1.2 mm
G300, 1.15 mm G550, 0.55 mm G300 and 1.9 mm G450 steel sheets,
and their average yield strengths were found to be 352, 585, 373 and
527 MPa, respectively, while their elastic modulus was about
200,000 MPa.

3.2. Test method

Since HFC sections are mono-symmetric, fixed end conditions were


simulated at both ends of the columns. A special arrangement was made
at the column ends in this test series. A hole which is larger than the
fabricated built-up HFC stub columns was made in a 8 mm thick steel
Fig. 17. Plate elements of HFC Sections. plate. This plate was then bolted to a 20 mm thick steel plate (Fig. 12).
Both ends of the test specimen were then placed inside the hole and
filled with Pyrocrete grouting as shown in Fig. 12. Thereafter, the test
fastening connection was varied by choosing different lengths for web specimen was placed in the machine and a compression load was
elements. In this study built-up HFC sections of three depths were gradually applied to it until failure. Fig. 12 shows the test set-up for
considered −90, 150 and 180 mm. Further, different thickness/grade HFC specimens. The load versus axial deformation plots of tested
web elements were also used: 1.2 mm G300, 1.15 mm G550, 0.55 mm columns are shown in Fig. 13(a)–(e).
G300 and 1.9 mm G450. The rivet fastening was used at 50 mm
(< 100 mm) spacing as shown in Fig. 11(d) while the first rivet was
placed at a distance of 15 mm from the column ends. The built-up HFC

189
S. Kesawan et al. Thin-Walled Structures 116 (2017) 180–193

6 4
8
Element B 2 3
7 1
Element A 14 mm
9 10 mm
Element C 10

11

10'
9'
7' 1'
2' 3'
(a) (b) 6' 8'
4'
Fig. 18. Hollow flange I- and channel sections made of three steel plates (a) 3PHFI, (b) 5'
3PHFC.

3.3. Test results and their comparison with design rule predictions Fig. 20. Rivet Fastened stiffened HFC section (a) HFC12G30012G300 stub column with
50 mm rivet spacing (b) Plate elements of stiffened channel section.
During the tests, local buckling waves were observed along the
specimen length. These buckling waves then started to confine to Table 7
one location as shown in Fig. 14. At this time, local buckling waves Ultimate failure loads of stiffened and unstiffened HFCs.
were also observed in the outstand element between the rivet
locations, i.e. flange and lip elements. Unlike the local web buckling Specimen Experimental AS/NZS Comparison
failure load 4600
observed in the welded hollow flange sections [12] shown in Fig. 15(b),
PEX (kN) prediction
local buckling occurred across the entire cross-section width through PEWM PEX / PEWM
buckling deformations of web, flange and lip elements between the
rivet locations. Fig. 14(a)–(d) show this failure mode including the Unstiffened150HFC12G300115G500 137 185 0.73
Stiffened150HFC12G300115G500 175 217 0.81
progression of local buckling, ultimate and post-ultimate failures. This
Unstiffened150HFC12G30012G300 110 176 0.63
failure mode occurred because the intermittent rivet fastening did not Stiffeed150HFC12G30012G300 146 192 0.76
provide full restraints similar to continuous welding. This kind of
buckling behaviour would have led to premature failures of test
columns. In all other tests, the failure behaviour was similar, except capacity of cold-formed and welded hollow flange channel sections with
for 180HFC12G30019G450 sections, where the rivets also failed before a reasonable degree of accuracy [12]. Hence if the experimental failure
reaching their ultimate loads. Figs. 14–16 show the failed load results agree well with the effective width based capacity predic-
HFC12G300115G500, HFC12G30012G300 and 180HFC12G30019- tions for the built-up HFC sections based on the assumption of a
G450 columns. continuous web-flange connection, it can be concluded that the inter-
The effective width method predicted the section compression mittent rivet fastening spacing used is adequate to attain their full section

Fig. 19. 180HFC12G30012G300W Columns (a) Before failure, (b) After failure.

190
S. Kesawan et al. Thin-Walled Structures 116 (2017) 180–193

compression capacity. The compression capacities of the tested built-up where w is the rivet spacing of 16.6 mm) were then tested, and their
HFC stub columns were calculated using the effective width method compression capacities are given in Table 6. These columns have higher
(EWM) in AS/NZS 4600 [20] based on the measured centre line capacities than those HFC sections with 50 mm rivet spacing, and the
dimensions. The different elements considered in the calculations are differences were more than 45%. Furthermore, importantly the test
shown in Fig. 17. In these elements 7, 7′, 8, 8′, 9 and 9′ were considered results matched very closely with the EWM design predictions.
as one edge supported outstand elements while other elements were Fig. 19(b) shows that local buckling failure did not occur across the
considered as both edges supported, and thus corresponding k factors of entire section width between the rivet locations. These observations
0.43 and 4 were used. Table 6 compares the AS/NZS 4600 predictions demonstrate that the built-up HFC sections with 16.6 mm rivet spacing
with the experimental results, which shows large discrepancies between have acted similar to continuously welded sections. However, the use of
them. As an example, for 150HFC12G300115G500 section, the predicted such a small rivet spacing is not desirable in practical applications.
experimental failure load is about 35% less than that of EWM prediction. As another solution to improve the compression capacity of inter-
This proves that the built-up HFC sections with even a small rivet spacing mittently rivet fastened HFC sections, the use of stiffened web elements
of 50 mm did not behave similar to continuously welded sections. As was attempted as shown in Fig. 20. In order to investigate the behaviour
evident from some cases in Table 6, differences of more than 40% can be of such web stiffened sections, an experimental study was performed for
seen between the test results and EWM based capacity predictions. two different sections. For comparison purposes, these sections with
Ropalin et al. [30] and Keerthan and Mahendran [31] performed both unstiffened and stiffened web elements were considered. Hence a
bending and shear tests of built-up HFC sections with 50 mm rivet total of four tests was conducted and the details are summarised in
spacing. Although they observed differences between the shear and Table 7. The same test method described in Section 3.2 was used.
bending capacities of welded and built-up sections, the differences were Table 7 presents the test failure capacity results while the failed
only about 10%. Element 5 (Fig. 17) is the major contributor for the specimens are shown in Figs. 21 and 22 with the latter showing the
shear capacity of the built-up HFC section, and if the flanges provide failure mode including the progression of local buckling, ultimate and
sufficient rigidity for Element 5, these sections would reach the shear post-ultimate failures. The section compression capacities were pre-
capacity closer to that of continuously welded section, and this was the dicted using the effective width method in AS/NZS 4600 based on
case in Keerthan and Mahendran [31]. In the bending tests, one of the continuous connection between elements and, are compared with test
flanges is in compression while the other in tension with web Element 5 failure capacities in Table 7. The HFC sections with stiffened web
subjected to bending action. However, in the built-up HFC columns, all elements gave higher ultimate loads than those with unstiffened web
the plate elements are in compression and are susceptible to local elements, but their failure modes again demonstrated the effects of
buckling. If the flange elements buckle, then their support to the web discontinuity along the web-flange-lip connection (Figs. 21 and 22).
element becomes questionable in the intermittently rivet fastened The use of web stiffeners prevented local buckling across the full section
sections. This might result in the web element behaving as an width between the rivet locations. However, local buckling failure
unsupported element along both edges. In the column tests the length could now be seen in the middle web element and then in the flange
of Element B in Fig. 18(b)) was less than the screw spacing, and would element. Due to this, the built-up HFC sections with web stiffened
have acted as one edge supported element and thus subjected to elements and 50 mm rivet spacing could not reach the ultimate load of
premature local buckling. This means adequate support was not identical sections with continuous welded connections. Table 7 proves
available to the web element along both of its longitudinal edges, this through the section compression capacity calculations using EWM
which led to premature failure of the built-up HFC columns. However, based on continuous connections between various plate elements as
this did not happen in the 3PHFI sections since the length of Element A shown in Fig. 20(b).
(Fig. 18(a)), which supports the web element along both longitudinal Based on the AISI S100 [25] requirements, the minimum connector
edges, is small and, thus these elements do not undergo premature local spacing for the 1.2 mm G300, 1.15 mm G550, 0.55 mm G300 and
buckling. From this lesson learnt it can be expected that if the built-up 2.0 mm G450 thick web plates to avoid column like buckling behaviour
sections should attain their full compression capacities, the screw/rivet are 43, 32, 19 and 55 mm, respectively, which should also be limited to
spacing should be reduced further. 60 mm based on the unstiffened plate width (Bl in Fig. 2(c) and (d)). A
To verify this, the rivet spacing was reduced from 50 mm to connector spacing of 50 mm for the above mentioned web plates is
16.6 mm as shown in Fig. 19(a) in the built-up HFC columns. These larger than the minimum spacing based on AISI S100 [25], except for
sections (180HFC12G30012G300w and 150HFC12G30012G300w, the 2.0 mm G450 thick web plate. But the 16.6 mm connector spacing

Fig. 21. Failed Rivet Fastened stiffened HFC sections (a) HFC12G30012G300 column (b) HFC12G300115G500 column.

191
S. Kesawan et al. Thin-Walled Structures 116 (2017) 180–193

Fig. 22. Failure progression of stiffened HFC12G300115G500 column (a) Initial condition, (b) At ultimate load, (c) and (d) Post-ultimate load.

for 1.2 mm G300 steel web plate is less than the minimum spacing both types of built-up HFI sections. The effective width method in cold-
based on AISI S100 [25]. In this study, built-up HFC sections with formed steel design standards with an assumption of continuous
50 mm spacing did not behave like a fully welded section, but the connections between various plate elements was able to predict the
section with 16.6 mm spacing behaved liked a fully welded section. section compression capacities of HFI sections obtained from tests for
Hence the AISI S100 [25] guideline appears to be applicable to the 100 mm screw/spot weld spacing. This confirms that the behaviour of
built-up HFC sections made of three plate elements. Further investiga- HFI sections is similar to continuously welded sections with minimum
tions are needed since this study was based on a small number of tests detrimental effects caused by intermittent fastening along web to flange
of built-up HFC sections. connections. However, the current direct strength method based design
In summary, this research has shown that the effect of intermittently equations were unable to predict the section compression capacities.
rivet/screw fastening on the section compression capacity of HFC The predictions from the modified direct strength method in [12]
sections is quite significant (> 35%) in comparison with continuously matched reasonably well with the test results.
welded HFC sections, unless very small rivet/screw spacings are used. In contrast to HFI sections, the behaviour of HFC section stub
The use of stiffened web elements improved this situation, but the loss columns fabricated using three steel elements was significantly influ-
in the section compression capacity was still about 25%. Alternatively, enced by intermittent fastening with premature local buckling occur-
HFC sections fabricated using single steel sheet as shown in Fig. 2(c) ring across the full section width. The rivet fastened HFC sections were
can be used without such a large loss in compression capacity. unable to reach a section compression capacity that is closer to that of
However, intermittently rivet/screw fastened doubly symmetric HFI continuously welded section unless a very small rivet spacing
sections (single steel sheet or three steel elements) do not suffer from (< 20 mm) was used. The use of stiffened web elements also could
such detrimental effects in relation to their section compression not eliminate this problem. The use of the effective design method with
capacities. the assumption of continuous connections between plate elements did
not provide the failure loads from the tests, thus confirming the
4. Conclusions significant detrimental effects of intermittent fastening on the section
compression capacity of HFC sections. Therefore the use of HFC
This paper has described a detailed experimental investigation on sections fabricated using single steel sheet is recommended as it is
the section compression capacity of built-up cold-formed steel hollow unlikely to suffer from the above detrimental effects. Overall this study
flange sections. Hollow flange I- (HFI) and hollow flange channel (HFC) has improved the current knowledge on the section compression
sections were considered that were made of either single steel sheet or capacities of built-up cold-formed steel hollow flange sections.
three steel elements, and their web and flange elements were connected
by screw/rivet fastening or spot welding. Experimental results showed
that screw fastening and spot welding provided the same level of Acknowledgements
connectivity between web and flange elements. It was found that HFI
sections made using single steel sheet performed adequately with less The authors would like to thank QUT for providing the necessary
than 15% capacity reductions for larger screw/weld spacings such as laboratory testing and computing facilities, and QUT and Australian
400 mm, but not for those made using three steel elements. Therefore, a Research Council (Grant Number LP0669668) for providing the finan-
fastener spacing of 100 mm or less is conservatively recommended for cial support to conduct this research project.

Appendix A. Design rules

Direct Strength Method (DSM) in AS/NZS4600


2
For λc ≤ 1.5, Pne = (0.658λc ) Py (A1)

⎛ 0.877 ⎞
For λc > 1.5, Pne = ⎜ 2 ⎟ Py
⎝ λc ⎠ (A2)

λc = Py / Pcre (A3)

Py = Ag fy (A4)

192
S. Kesawan et al. Thin-Walled Structures 116 (2017) 180–193

Where; Ag is the gross sectional area, fy is the yield stress and Pcre is the minimum of the critical elastic column buckling load in flexural, torsional or
flexural-torsional buckling.
For local buckling,

Pne ⎡ ⎛ P ⎞0.4 ⎤ ⎛ P ⎞0.4


For > 0.776, Pnl = ⎢1 − 0.15 ⎜ crl ⎟ ⎥ ⎜ crl ⎟ Pne
Pcrl ⎢⎣ ⎝ Pne ⎠ ⎥⎦ ⎝ Pne ⎠ (A5)

Pne
For ≤ 0.776, Pnl = Pne
Pcrl (A6)
Where; Pnl is the local buckling capacity, Pcrl is the critical elastic local buckling load Modified DSM Equations [12]

Pne ⎡ ⎛ P ⎞0.24⎤ ⎛ P ⎞0.24


For > 0.71, Pnl = ⎢1 − 0.1 ⎜ crl ⎟ ⎥ ⎜ crl ⎟ Pne
Pcrl ⎢⎣ ⎝ Pne ⎠ ⎥⎦ ⎝ Pne ⎠ (A7)

Pne
For ≤ 0.71, Pnl = Pne
Pcrl (A8)

References Struct. 79 (2014) 202–217.


[16] M. Fadhluhartini, S.H.M. Mohd, M. Shahrin, Md.T. Mahmood, Ultimate load of
built-up cold-formed steel column, ARPN J. Eng. Appl. Sci. 9 (11) (2014)
[1] R.I. Dempsey, Structural behaviour and design of hollow flange beams, in: 2095–2101.
Proceedings of the Second National Structural Engineering Conference 1990, [17] I. Georgieva, L. Schueremans, L. Pyl, Composed columns from cold-formed steel Z-
Institution of Engineers, Australia, Adelaide, 1990, pp. 327–335. profiles. experiments and code-based predictions of the overall resistance, Eng.
[2] V. Doan, M. Mahendran, A Residual stress model for hollow flange beams, Research Struct. 37 (2014) 125–134.
Report 1996 Physical Infrastructure Centre, Queensland University of Technology [18] B. Young, J. Chen, Design of cold-formed steel built-up closed sections with
Brisbane, Australia, 1996. intermediate stiffeners, J. Struct. Eng. 134 (5) (2008) 727–737.
[3] S. Kesawan, M. Mahendran, Fire tests of load-bearing LSF walls made of hollow [19] T.A. Stone, R.A. LaBoube, Behavior of cold-formed steel built-up I-sections, Thin-
flange channel sections, J. Constr. Steel Res. 155 (2015) 191–205, http://dx.doi. Walled Struct. 43 (2005) 1805–1817.
org/10.1016/j.jcsr.2015.07.020. [20] SA Australia/New Zealand Standard AS/NZS 4600 Cold-Formed Steel Structures,
[4] S. Kesawan, M. Mahendran, Predicting the performance of LSF walls made of Sydney, Australia, 2005.
hollow flange sections in fire, Thin-Walled Struct. 98 (A) (2016) 111–126, http:// [21] M. Yener, Criteria for connection spacing in cold-formed steel, ASCE J. Struct. Eng.
dx.doi.org/10.1016/j.tws.2015.03.014. 110 (9) (1984) 2178–2195.
[5] S. Kesawan, M. Mahendran, Thermal performance of load-bearing walls made of [22] L.D. Luttrell, K. Balaji, Properties for cellular decks in negative bending, in:
cold-formed hollow flange channel sections in fire, Fire Mater. 40 (5) (2015) Proceedings of the 11th International Specialty Conference on Cold-Formed Steel
704–730, http://dx.doi.org/10.1002/fam.2337. Structures 1992, University of Missouri-Rolla MO, USA, 1992.
[6] V. Jatheeshan, M. Mahendran, Experimental study of LSF floors made of hollow [23] M.L. Jones, Spacing of Connections in Compression Elements for Cold-formed Steel
flange channel section joists under fire conditions, J. Struct. Eng. ASCE 142 (2015) Members (Master of Science Thesis), University of Missouri-Rolla, MO, USA, 1997.
(04015134-1-14). [24] R. LaBoube, W. Yu, M. Jones, Spacing of connections in compression flanges of
[7] V. Jatheeshan, M. Mahendran, Numerical study of LSF floors made of hollow flange built-up cold-formed steel beams, in: Proceedings of the 14th International
channels in fire, J. Constr. Steel Res. 115 (2015) 236–251. Specialty Conference on Cold-Formed Steel Structures, St. Louis, Missouri, USA,
[8] P. Keerthan, M. Mahendran, Experimental studies on the shear behaviour and 1998, pp. 563–578.
strength of LiteSteel beams, Eng. Struct. 32 (10) (2010) 3235–3247. [25] American Iron and Steel Institute (AISI S100), North American Specification for the
[9] T. Anapayan, M. Mahendran, D. Mahaarachchi, Section moment capacity tests of Design of Cold-formed Steel Structural Members, AISI, Washington, DC, USA, 2012.
LiteSteel beams, Thin-Walled Struct. 49 (4) (2011) 502–512. [26] Standards Australia (SA) AS 1391. Methods for the Tensile Testing of Metals at
[10] T. Anapayan, M. Mahendran, D. Mahaarachchi, Lateral distortional buckling tests of Elevated Temperatures, Sydney, Australia, 2007.
a new hollow flange channel beam, Thin-Walled Struct. 49 (1) (2010) 13–25. [27] EN 1993-1-3, Eurocode 3: Design of Steel Structures, Part 1-3: General Rules –
[11] Z.B. Yi, T. Wilkinson, Compression Capacity of Hollow Flange Channel Stub Supplementary Rules for Cold-formed Members and Sheeting, European Committee
Columns, Research Report 2007 School of Civil Engineering, University of Sydney for Standardization, Brussels, Belgium, 2006.
Australia, 2007. [28] B.W. Schafer, The direct strength method of cold-formed steel member design, J.
[12] S. Kesawan, M. Mahendran, Section compression capacity of high strength cold- Constr. Steel Res. 64 (7–8) (2008) 766–778.
formed hollow flange channels, J. Constr. Steel Res. 133 (2017) 202–213, http://dx. [29] B. Schafer, Designing cold-formed steel using the direct strength method, in:
doi.org/10.1016/j.jcsr.2017.02.015. Proceedings of 18th International Speciality Conference on Cold-formed Steel
[13] I. Georgieva, L. Schueremans, L. Vandewalle, L. Pyl, Design of built-up cold-formed Structures, 2006, pp. 475–489.
steel columns according to the direct strength method, Procedia Eng. 40 (2012) [30] S. Ropalin, M. Mahendran, P. Keerthan, Section moment capacity tests of rivet
119–124. fastened rectangular hollow flange channel beams, J. Constr. Steel Res. 125 (2014)
[14] H. Lau, T. Ting, An investigation of the comprehensive strength cold-formed steel 252–262.
built-up I section. in: Proceedings of the Sixth International Conference on [31] P. Keerthan, M. Mahendran, Shear tests of rivet fastened rectangular hollow flange
Advances in Steel Structures, Dec 16, Hong Kong, China, 2009, pp. 441–449. channel beams, J. Constr. Steel Res. 121 (2016) 330–340.
[15] Y. Li, S. Wang, Z. Shen, Ultimate load-carrying capacity of cold-formed thin-walled
columns with built-up box and I section under axial compression, Thin-Walled

193

You might also like