International Journal of Steel Structures (2023) 23(2):503–520 Online ISSN 2093-6311

https://doi.org/10.1007/s13296-023-00708-4 Print ISSN 1598-2351

Flexural Integrity Between the Individual Channels of Built‑Up

Cold‑Formed Steel Beams
Mir Basar Suhaib1 · Manzoor Ahmad Tantray1

Received: 26 May 2022 / Accepted: 2 January 2023 / Published online: 7 January 2023
© Korean Society of Steel Construction 2023

Abstract
This paper investigates the influence of commonly used screw connections on the integrity of built-up beams formed by
connecting two lipped channel sections in a back-to-back configuration. An experimental investigation on the behaviour of
6 built-up CFS beams connected by self-drilling screws is presented in this paper. The screw spacing was varied along the
length as well as the depth of the beams. The test results indicate that the screw spacing has some influence on the capacity
of the built-up beams. An increase in screw spacing leads to an undesirable separation of channels and thus has a negative
impact on the composite action between the members. Finite element models were developed and after successful validation,
a parametric study involving 124 models was carried out, to investigate the impact of larger longitudinal screw spacings on
the integral behaviour of the beams. It was observed that the flexural capacity decreased by 11%, 14%, 17% and 19% when
the longitudinal distance between the screws increased from 50 to 1050 mm for beams with web depths 150 mm, 200 mm,
250 mm and 300 mm respectively. It can be concluded that flexural capacity is directly proportional to the decrease in the
screw spacing and increase in the number of screw rows. The results also indicate that the percentage decrease in strength is
greater for beams with larger web depths. An idealized cross section assuming complete composite action was simulated for
each web depth using FE modelling and the capacities of all the screw-connected “semi-rigid” beams were compared against
it. The capacity of screw connected built-up beams with web depths 150 mm, 200 mm, 250 mm and 300 mm was found
to be 0.71, 0.64, 0.61 and 0.59 times that of the idealized cross-section on average, respectively. The moment capacities of
screw-connected beams were also compared against the theoretical values obtained from AISI S-100. It was observed that
the design guidelines provided in AISI S-100 were unable to account for the reduction in the moment capacity due to the
screw connection in built-up members. AISI S-100 is conservative by as much as 34%, 42%, 46% and 48% for beams with
web depths 150 mm, 200 mm, 250 mm and 300 mm respectively.

Keywords Cold-formed steel · Back-to-back channels · Built-up beam · Flexural strength · Screw connection · Composite
action

1 Introduction required (Roy et al., 2021). Also, when CFS members are
required to be used as primary members in framed construc-
Cold-formed steel (CFS) built-up beams have gained wide tion, built-up, open or closed, doubly symmetric cross-sec-
recognition in the constructional steel industry. (Laím et al., tions are preferred (Selvaraj & Madhavan, 2019). These sec-
2013; Deepak & Shanthi, 2018; Ghannam, 2019; Selvaraj tions are formed by connecting individual, mono-symmetric
& Madhavan, 2019; Chen et al., 2020; Suhaib & Tantray, or point-symmetric profiles such as simple channel, lipped
2022a, b). These become necessary when a single monosym- channel, hat, U or Z sections, using suitable methods such
metric or point-symmetric member is insufficient to carry as bolting, riveting, welding or more commonly, using self-
the loads (Xu et al, 2009a) or when longer beam spans are driven screws. These connections when designed properly
prevent the individual members of a built-up member from
* Mir Basar Suhaib independent failures.
[email protected] Research on the premature failures of CFS built-up beams
due to localised buckling has received some attention in the
1
Department of Civil Engineering, National Institute last decade. (Laím et al., 2014; Wang & Young 2015, 2016a,
of Technology Srinagar, Srinagar, India

13
Vol.:(0123456789)
504 International Journal of Steel Structures (2023) 23(2):503–520

2016b, 2017; Laím et al. 2016; Laím & Rodrigues, 2018) investigated experimentally and numerically in this research.
However, studies on the integrity of the beams imparted Composite action in a built-up beam is considered to be a
by connections between the individual components is very function of the screw spacing. In the experimental phase, the
limited. The strength of a built-up beam depends not only effect of varying screws along the length and depth of the
on the member strength of individual components but also back-to-back connected built-up member, on the compos-
on the extent of the composite action between the various ite behaviour was studied. Accurate finite element models
components. A relatively high composite action can only be were developed which were further used to study composite
achieved if the fasteners are adequate. For example, using action in the beams with screws spaced closer and farther
large high strength bolts to connect two thin-walled sec- along the depth as well as the length of the beams. Finally,
tions is not only uneconomical, but stress concentrations an idealised cross section assuming complete composite
develop around the bolt holes due to the tendency of tear- action was simulated using FE modelling and the capacities
ing of thin-walled sheet due to large bolt forces (Georgieva of all the screw-connected “semi-rigid” beams were com-
et al., 2012). pared against it.
Screw connections, particularly self-tapping screws,
are most commonly used in CFS built-up members as they
are cost effective, offer ease of application and provide the 2 Experimentation
required level of rigidity. However, self-tapping screws
do not result in a fully integrated composite member. Xu 2.1 Specimens
et al. (2009a) investigated the influence of screw spacing on
the moment capacity of built-up box sections. The authors Six simply supported CFS built-up beam specimens were
reported that the capacity of the built-up section can be tested under pure bending. The specimens were prepared
considered to be equal to 90% of the combined capacity of from 1.2 mm thick steel sheets press braked into lipped
the two individual channels. Georgieva et al. (2012 , 2012a, channels. The base steel sheet was of grade CR1, conform-
b) studied the flexural behaviour of built-up beams formed ing to Indian Standard IS 513: 2008 (IS Code, 2008). Lipped
by connecting Z-sections by screws. It was found that the channels with 150 mm web depths, 50 mm flanges and
flexural rigidity was slightly above the total rigidity of two 15 mm lip lengths were investigated. The cross-section of
independent Z-sections but significantly below that of the the specimens with nominal dimensions are shown in Fig. 1
integrated section. Wang and Young (2018) investigated while Table 1 shows the actual dimensions. Self-drilling
cold-formed built-up open (I-section) and closed (box sec- screws of diameter 5 mm were used to connect the chan-
tion) beams under four-point bending with different screw nels back-to-back. Two different screw spacings, 150 mm
arrangements in the moment span. The beams had a screw and 275 mm along the length were investigated in one, two
spacing to web depth ratio ranging from 0.9 to 27.9 for or three-row arrangements along the depth of the beams as
open section and 0.5–25.6 for closed section beams. The shown in Fig. 2. The aim was to study how different screw
authors reported that in general the screw arrangements have spacings, in both the directions, affected the composite
very little influence on the moment capacities of built-up action of built-up beams.
beams. Further, it was also observed by Roy et al. (2021)
that the moment capacity increased by up to 15.6% in the 2.2 Specimen Labelling
beams when the screw spacing was decreased from 1900
to 475 mm. Although, these studies were extensive, but no In order to identify the specimens, a specific label was pro-
comparisons were made with the built-up sections assuming vided for each specimen. The label denoted the nominal web
full composite action. This limited the understanding of the depth of the beam, screw spacing along the span and the
effect of screw spacing on the flexural rigidity and structural number of rows of screws. For example, CFS150-S150-R1
integrity of the built-up members. Even the Direct Strength represents a cold-formed built-up beam with web depth
Method (AISI-American Iron & Steel Institute, 2016), rec- 150 mm (CFS150) and screw spacing equal to 150 mm along
ommends that the flexural strength and moment of inertia the length (S150) arranged in one row along the depth of
of the built-up sections be taken as the sum of capacities of the web (R1).
individual members. This recommendation considers par-
tial integrity imparted due to the point connections between 2.3 Material Testing
the members. Also, this plain assumption does not take into
consideration the composite action between the individual Tensile tests of standard coupons as specified in ASTM E8
members. (ASTM Standard E8/E8M-13a, 2013), were conducted. The
The composite action of the individual members of specimens were taken from both the web element as well
built-up CFS lipped channel beams under pure bending is as the flanges of the beams under investigation. Table 2

13
International Journal of Steel Structures (2023) 23(2):503–520 505

Fig. 1  Nominal beam cross-

sectional dimensions with screw
row spacings

Table 1  Measured beam Specimen Channel Web hw (mm) Flanges Thickness (mm) Lip length Screw spacing
dimensions
bf 1(mm) bf 2(mm) 𝛿1(mm) (mm)

CFS150-S150-R1 a 150.1 49.7 50.1 1.18 15.1 151

b 149.8 49.6 49.9 1.19 14.9
CFS150-S150-R2 a 149.5 50.8 50.4 1.18 14.6 149.1
b 149.2 49.9 47.8 1.18 14.7
CFS150-S150-R3 a 149.6 47.9 48.3 1.19 15.2 148.4
b 149.3 47.4 48.9 1.18 14.8
CFS150-S275-R1 a 149.6 49.1 49.1 1.19 15.0 275.8
b 150.1 50.4 48.5 1.19 14.6
CFS150-S275-R2 a 147.2 48.2 49.2 1.2 14.9 275.9
b 146.9 49.0 50.2 1.17 15.1
CFS150-S275-R3 a 148.4 48.9 49.8 1.2 14.8 276.6
b 148.5 48.7 49.7 1.19 14.7

summarises the material properties while Fig. 3 presents the markings. Due to the limitation of the clamps holding the
stress–strain relationships of the specimens. Further, Fig. 4 LVDTs, it was not possible to measure the readings on the
presents the engineering and true stress–strain curves for top and bottom flanges simultaneously. As such, readings
one of the specimens. were first taken for the bottom flange and then the beam was
turned upside down to take the readings for the top flange.
2.4 Measurement of Geometric Imperfections The values were being recorded by a National Instruments
NI-USB 6002 8-channel data acquisition system. The maxi-
Cold-formed members have geometric defects as a result mum value of imperfections measured in web and flanges for
of their thin walls and the fabrication method involved. all the channel sections are given in Table 3.
These imperfections affect the capacities of thin-walled
structural steel members and hence were measured for all 2.5 Test Set‑Up
the specimens in the present investigation. A set-up simi-
lar to that used by Anapayan et al., (2011) was employed. Four-point bending tests were conducted on all the beam
The measurement apparatus consisted of a rigid frame as specimens. The beams were loaded with a hydraulic actua-
shown in Fig. 5, which was levelled to ensure the readings tor and a 100 kN load cell in a 500 kN loading rig. The
were reliable. The beam specimen was then placed on the test set-up is illustrated in Fig. 6. The beams were simply
frame and clamped in position. Each specimen had mark- supported on two solid half-rounds and laterally braced
ings at 100 mm intervals along the length, at the centre of at the supports. Hot rolled plates of thickness 12 mm and
the web and flanges. LVDTs with an accuracy of 0.01 mm width 50 mm, provided under the loads and the supports as
were employed to determine the deviations from straight shown in Fig. 2, prevented the crushing failure. The verti-
geometry on the flanges (𝛿 1 ) as well as the webs (𝛿2 ) at the cal displacements were measured using LVDTs as shown

13
506 International Journal of Steel Structures (2023) 23(2):503–520

Fig. 2  Screw spacing in beam

specimens and position of bear-
ing stiffeners in the specimens

in Fig. 6. The readings from the actuator and the LVDTs

were obtained at regular intervals using an 8-channel data
acquisition device.
Table 2  Material properties of CFS
2.6 Experimental Results
Coupon from Yield stress Ultimate stress Young’s
σy (MPa) σu (MPa) modulus E
(GPa) The effects of screw spacing as well as the number of rows
of screws on the moment capacity have been summarized in
CFS150-S150-R2 273.7 362 198.3
Table 4. The moment capacity of the simply supported beam
CFS150-S275-R1 270.9 354 198.8
CFS150-S275-R3 275.1 364 199.2

13
International Journal of Steel Structures (2023) 23(2):503–520 507

Fig. 5  Set-up to measure geometric imperfections in beams

Fig. 3  Stress–strain relationship for CFS coupons

test, the screws did not fail, demonstrating that their strength
was sufficient for fastening. However, it was observed that
the beam specimens with screws spaced at 275 mm along
the length, exhibited a tendency of separation of individual
channels (S), which was more evident in the specimen with
a single row of screws along the depth, as shown in Figs. 9
and 10. The web buckling failure (L) was different from the
separation failure (S), since in the former, webs of both the
channels buckled in unison and buckling occurred along one
direction of the centroidal plane of the beam, while as in the
later, the channels tend to move apart and hence the webs
buckled in both the directions of the centroidal plane. Fur-
ther, it was also observed that the gap between the individual
channels was wider within the moment span.
Table 4 shows that extending the longitudinal distance
between screws from 150 to 275 mm resulted in a 6.18 per-
cent, 3.54 percent, and 2.9 percent reduction in ultimate
moment capacity for one, two, and three rows of screws,
respectively. This means that the decrease in moment capac-
Fig. 4  Stress–strain relationship for CFS coupon CFS150-S150-R2 ity due to increased screw spacing is substantially greater for
with both the engineering and true stress–strain relationships a single row of screws. Figure 11a, b show the displacement
response of the specimens due to the applied moments.

specimens with two-point loads was calculated using the

formula-
( )
3 Finite Element Study
Pult
MEXP = a (1) Abaqus CAE version 6.14 was used to perform finite ele-
2
ment analysis on all test specimens (ABAQUS CAE User’s
In the above formula Pult refers to the experimental ulti- Manual version6.14, 2014). The simulation of thin-walled
mate load and ‘a’ refers to the distance between the load and members involves geometric and material non-linearities
the support as has been shown in Fig. 7. and Abaqus has been found to handle these non-linearities
Table 4 summarizes the failure moments as well as the with reasonable accuracy (Schafer et al.,2010). Abaqus has
failure modes of all the tested beam specimens. In general, been used successfully by (Foraboschi, 2019, 2020) to model
all the beams failed in a combination of web buckling and and analyse steel structures.
flange distortion (L + F) as is shown in Fig. 8. During the

13
508 International Journal of Steel Structures (2023) 23(2):503–520

Table 3  Measured geometric Specimen Channel Maximum geometric imperfection Maximum geo-
imperfections metric imperfec-
tion
𝛿1(mm) 𝛿2(mm)
Top Flange Bottom Flange

CFS150-S150-R1 a 0.67 0.64 0.38

b 0.66 0.65 0.40
CFS150-S150-R2 a 0.52 0.62 0.49
b 0.56 0.59 0.45
CFS150-S150-R3 a 0.68 0.57 0.51
b 0.59 0.56 0.51
CFS150-S275-R1 a 0.61 0.65 0.36
b 0.54 0.64 0.39
CFS150-S275-R2 a 0.60 0.57 0.51
b 0.58 0.62 0.49
CFS150-S275-R3 a 0.51 0.59 0.59
b 0.50 0.58 0.56
Mean 0.59 0.60 0.47

Fig. 6  Test setup with instru-

mentations

Table 4  Experimental and finite Specimen Exp. moment capacity, FEA moment capacity, Comparison Failure
element results MEXP (kNm) MFEA (kNm) MEXP∕M
FEA

CFS150-S150-R1 5.18 5.41 0.96 L+F

CFS150-S150-R2 5.37 5.56 0.97 L+F
CFS150-S150-R3 5.58 5.77 0.97 L+F
CFS150-S275-R1 4.86 5.11 0.95 L+F+S
CFS150-S275-R2 5.18 5.29 0.98 L+F+S
CFS150-S275-R3 5.42 5.49 0.99 L+F
Mean 0.97
COV 0.015

13
International Journal of Steel Structures (2023) 23(2):503–520 509

Fig. 7  Schematic representation of the beam specimen

3.1 Geometry and Material Model

The beams were modelled based on the linear dimensions

provided in Fig. 1. Only half geometry of the beams was
Fig. 9  Separation of channels in CFS150-S275-R2
modelled as symmetry was assumed about the mid-span. The
cold-formed steel was defined as an elastic–plastic material
with a significant plastic zone. The isotropic yielding and
back-to-back channels. Friction between the two channels
plastic hardening were modelled using von Mises criteria.
was ignored while defining the tangential behaviour and nor-
True stress-stress relationship, as shown in Fig. 4, was used
mal behaviour was assumed to be hard. To prevent contact
to introduce non-linear material properties determined from
instabilities, a small sliding formulation was adopted.
the coupon testing. To convert engineering stress and strain
The screw connection between the two channels was not
to true stress and strain, the following equations were used:
physically modelled. Instead, a simplified screw connection
𝜎T = 𝜎(1 + 𝜀) (2) using MPC beam connectors was employed, as shown in
Fig. 12.
𝜀T = ln (1 + 𝜀) (3)
3.3 Element Type and Meshing
where 𝜎T and 𝜀T are the true stress and strain and 𝜎 and 𝜀 are
the engineering stress and strain respectively. The four-noded shell element S4R, was used to discretize
cold-formed beams, due to its capability to handle non-linear
3.2 Contact Modelling problems with large strains. This element has been previ-
ously used in a number of studies (Keerthan & Mahendran,
The "surface to surface" interaction with master–slave for- 2013; Islam et al., 2019; Wang & Young, 2018). The sensi-
mulation was used to describe contact between the two tivity of the solution to the mesh size was studied and based

Fig. 8  Buckling failure in

CFS150-S275-R2 specimen

13
510 International Journal of Steel Structures (2023) 23(2):503–520

Moment (kNm)
4

1 CFS150-S150-R1
CFS150-S150-R2
CFS150-S150-R3
0
0 5 10 15 20
Mid-Span Deflection (mm)
(a)

5
Moment (kNm)

Fig. 10  Separation of channels in CFS150-S275-R1

1 CFS150-S275-R1
CFS150-S275-R2
CFS150-S275-R3
on this, a square mesh of size 10 mm was adopted in all the 0
FE models. The same mesh size was chosen for the valida- 0 5 10 15 20

tion models as well as for the FE models used for parametric Mid-Span Deflection (mm)
study. (b)

3.4 Loading and Boundary Conditions Fig. 11  Moment capacity vs mid-span displacement curves. a
Moment-displacement relation for specimens with 150 mm screw
spacing. b Moment-displacement relation for specimens with
A simplified technique was adopted to simulate the load
275 mm screw spacing
transfer plates. A region 50 mm x web height (150 mm), as
shown in Fig. 12, was artificially reinforced by increasing
the elastic modulus by a factor of 10. All the nodes, within was achieved by setting ZSYMM = 0 (Uz = URx = URy = 0)
this region, were kinematically coupled in all degrees of to all the nodes at the mid span of the beam.
freedom to a node set. Two node sets were defined, one at
the loading point and the other at the support as shown in 3.5 Initial Geomteric Imperfections
Fig. 12. The experimental support conditions were simulated
by restraining the translational DOFs in all the directions In the thin-walled structures with large plate height to
and rotational DOF along the beam length. In the Abaqus thickness ratio, the impact of initial imperfections can be
CAE, this was achieved by setting Ux = Uy = Uz = URz = 0 significant and hence cannot be ignored. These geometric
at the support reference node set. The load was applied as imperfections were introduced in to the FE models using the
the total experimental displacement (−Uy) at the node set *IMPERFECTION function. The magnitude of geometric
representing the loading point. Also, mid-span symmetry imperfections was taken as 0.6, which is approximately the
maximum mean value of measurements provided in Table 3.

13
International Journal of Steel Structures (2023) 23(2):503–520 511

Fig. 12  Typical finite element

model of a built-up beam Sffened surface at
loading point
Sffened surface at
loading point

Screws

Local Web Buckling

and Flange Distor on

Fig. 14  Typical separation of channels captured by finite element

Fig. 13  (L + F) failure observed in finite element model model

This value of imperfections was used along with the low- step, specified experimental displacement was applied, to
est buckling mode to incorporate buckling features into the carry forward the bending process. Further, to achieve
model. To model the shape of buckling, the nonlinear geom- convergence, automatic stabilization was employed. The
etry (*NLGEOM) was used in the analysis. default damping factor of 0.0002 and the maximum ratio
of stabilization to strain energy of 0.05 were used. From
3.6 Analyses the test results in Table 4, it was found that the FE results
were acceptable even with the incorporation of adaptive
A two-step process is commonly employed to analyze stabilization algorithm.
cold-formed steel structures (Wang & Young, 2018) (Sia-
haan et al., 2018). The first step is an elastic linear per-
turbation analysis and the second a non-linear analysis. 3.7 Validation of Finite Element Model
The linear perturbation analysis was carried out initially to
obtain the eigen vectors of interest. The first eigen vector The moment capacities obtained from the finite element
(buckling mode) is considered critical and was factored analysis were compared to the test results to validate the
to the magnitude of geometric imperfections in the non- FE model. Table 4 provides the comparison for all the
linear analysis. The non-linear post-buckling analysis was beam specimens. Between experimental and finite element
carried out using the General Static method. To overcome data, there is a 0.015 coefficient of variance. This indicates
the convergence issues due to instabilities in contact defi- that the moment capacities obtained from the FE analysis
nitions, two General Static steps as proposed by Wang bear a close co-relation with the experimental moment
and Young (2016a) were used. In the first step, a small capacities.
displacement, of the order Uy = − 1, was applied at the Also, as can be observed from Figs. 13 and 14, the finite
loading reference node set. The purpose of this step was to element model was able to predict the failure modes with
initiate contact between the components. In the subsequent reasonable accuracy. A comparison of moment-deflection

13
512 International Journal of Steel Structures (2023) 23(2):503–520

4 Parametric Study

4.1 Screw Spacing

The validated finite element model was used to study the

influence of spacing screws longitudnally in the range
from 50 to 1050 mm, on the moment capacity and com-
posite action of the beam specimens. These screw spacings
were studied for three different rows of screws comprising
a total of 124 models covering four web depths (h) of CFS
beams. The screw row arrangements have been shown in
Fig. 17. The labelling scheme of the finite element speci-
mens was consistent with those of the experimental tests.
Local web buckling and flange distortion (L + F) failure
was observed in all the specimens where the screws were
placed closer to each other. However, as the screw spac-
ing increased, local web buckling and flange distortion was
Fig. 15  Experimental vs FEM moment-deflection response for speci- accompanied by separation of individual channels (S), irre-
men CFS150-S150-R2 spective of the number of rows of screws. In general, it was
observed that (L + F) was the dominant failure mode for all
the web depths when the screw spacing was 150 mm or
less, while as (S) was the dominant failure mode when the
screw spacing was greater than 450 mm. Between 150 and
450 mm, the failures occurred such that there was no par-
ticular dominant mode. It is evident that for the screw spac-
ing of 1050 mm, the separation is quite significant. Further, a
single row of screws increased the tendency of channel sepa-
ration across all web depths. Also, the effect of longitudinal
spacing of screws on the moment capacity of the built-up
channel beams has been summarized in Tables 5, 6,7 and8,
Fig. 18a–d. The flexural capacity decreased by 11%, 14%,
17% and 19% when the screw spacing increased from 50
to 1050 mm for beams with web depths 150 mm, 200 mm,
250 mm and 300 mm respectively for 1 row of screws. This
indicates that the percentage decrease in strength is greater
for beams with larger web depths, 300 mm in the present
investigation. A similar decreasing trend was observed for
2-Row and 3-Row screw configurations as well. Further, it
Fig. 16  Experimental vs FEM moment-deflection response for speci- was observed that percentage increase in the moment capac-
men CFS150-S275-R3 ity was higher when the number of rows was increased for
screws with large longitudinal spacing. For example, for
a spacing of 1050 mm the moment capacity increased by
response obtained from FE analysis was compared with the 14.5% when the number of rows increased from one to three,
test results as shown in Figs. 15 and 16. Overall, the test and while for a spacing of 50 mm, the increase in moment capac-
finite element results correlated well and on this basis, the ity was 4% in beams with web depth of 300 mm.
FE model was considered appropriate for further parametric
investigation. 4.2 Ideal Section with Full Composite Action

In addition to studying the influence of screw spacing, an

idealised cross section assuming complete composite action
was simulated using FE model and the capacities of all the
screw-connected “semi-rigid” beams was compared against

13
International Journal of Steel Structures (2023) 23(2):503–520 513

Fig. 17  Screw row arrangement

for parametric study

Table 5  Finite element Specimen Moment capac- Moment capac- Moment capacity, Comparison Failure
investigation into the effect of ity, M0 (kNm) ity, Mi (kNm) ­MNAS (kNm)
screw spacing on CFS beams Mi∕M
0
Mi∕M
NAS
with web depth 150 mm
CFS150-CCA​ 7.63 – 7.46 – – L+F
CFS150-S50-R1 – 5.51 7.46 0.72 0.74 L+F
CFS150-S50-R2 – 5.68 7.46 0.74 0.76 L+F
CFS150-S50-R3 – 5.86 7.46 0.77 0.79 L+F
CFS150-S75-R1 – 5.49 7.46 0.72 0.74 L+F
CFS150-S75-R2 – 5.66 7.46 0.74 0.76 L+F
CFS150-S75-R3 – 5.85 7.46 0.77 0.79 L+F
CFS150-S150-R1 – 5.41 7.46 0.71 0.73 L+F
CFS150-S150-R2 – 5.56 7.46 0.73 0.75 L+F
CFS150-S150-R3 – 5.77 7.46 0.76 0.77 L+F
CFS150-S275-R1 – 5.27 7.46 0.69 0.71 L+F+S
CFS150-S275-R2 – 5.49 7.46 0.72 0.74 L+F+S
CFS150-S275-R3 – 5.69 7.46 0.75 0.77 L+F
CFS150-S300-R1 – 5.25 7.46 0.69 0.70 L+F+S
CFS150-S300-R2 – 5.47 7.46 0.72 0.73 L+F+S
CFS150-S300-R3 – 5.69 7.46 0.75 0.76 L+F
CFS150-S450-R1 – 5.17 7.46 0.68 0.69 L+F+S
CFS150-S450-R2 – 5.33 7.46 0.70 0.71 L+F+S
CFS150-S450-R3 – 5.56 7.46 0.73 0.75 L+F+S
CFS150-S600-R1 – 5.08 7.46 0.66 0.68 L+F+S
CFS150-S600-R2 – 5.25 7.46 0.69 0.70 L+F+S
CFS150-S600-R3 – 5.45 7.46 0.71 0.73 L+F+S
CFS150-S750-R1 – 4.99 7.46 0.65 0.67 L+F+S
CFS150-S750-R2 – 5.18 7.46 0.68 0.69 L+F+S
CFS150-S750-R3 – 5.41 7.46 0.71 0.73 L+F+S
CFS150-S900-R1 – 4.94 7.46 0.65 0.66 L+F+S
CFS150-S900-R2 – 5.12 7.46 0.67 0.69 L+F+S
CFS150-S900-R3 – 5.36 7.46 0.70 0.72 L+F+S
CFS150-S1050-R1 – 4.89 7.46 0.64 0.66 L+F+S
CFS150-S1050-R2 – 5.10 7.46 0.67 0.68 L+F+S
CFS150-S1050-R3 – 5.31 7.46 0.70 0.71 L+F+S
Mean 0.71 0.72

13
514 International Journal of Steel Structures (2023) 23(2):503–520

Table 6  Finite element Specimen Moment capac- Moment capac- Moment capacity, Comparison Failure
investigation into the effect of ity, M0 (kNm) ity, Mi (kNm) ­MNAS (kNm)
screw spacing on CFS beams Mi∕M
0
Mi∕M
NAS
with web depth 200 mm
CFS200-CCA​ 11.32 – 10.98 – – L+F
CFS200-S50-R1 – 7.33 10.98 0.65 0.67 L+F
CFS200-S50-R2 – 7.68 10.98 0.68 0.70 L+F
CFS200-S50-R3 – 7.88 10.98 0.70 0.72 L+F
CFS200-S75-R1 – 7.33 10.98 0.65 0.67 L+F
CFS200-S75-R2 – 7.66 10.98 0.68 0.70 L+F
CFS200-S75-R3 – 7.86 10.98 0.69 0.72 L+F
CFS200-S150-R1 – 7.26 10.98 0.64 0.66 L+F
CFS200-S150-R2 – 7.58 10.98 0.67 0.69 L+F
CFS200-S150-R3 – 7.74 10.98 0.68 0.70 L+F
CFS200-S275-R1 – 7.15 10.98 0.63 0.65 L+F+S
CFS200-S275-R2 – 7.47 10.98 0.66 0.68 L+F+S
CFS200-S275-R3 – 7.63 10.98 0.67 0.69 L+F
CFS200-S300-R1 – 7.12 10.98 0.63 0.65 L+F+S
CFS200-S300-R2 – 7.44 10.98 0.66 0.68 L+F+S
CFS200-S300-R3 – 7.62 10.98 0.67 0.69 L+F
CFS200-S450-R1 – 7.01 10.98 0.62 0.64 L+F+S
CFS200-S450-R2 – 7.32 10.98 0.65 0.67 L+F+S
CFS200-S450-R3 – 7.53 10.98 0.67 0.69 L+F+S
CFS200-S600-R1 – 6.87 10.98 0.61 0.63 L+F+S
CFS200-S600-R2 – 7.17 10.98 0.63 0.65 L+F+S
CFS200-S600-R3 – 7.48 10.98 0.66 0.68 L+F+S
CFS200-S750-R1 – 6.74 10.98 0.60 0.61 L+F+S
CFS200-S750-R2 – 7.08 10.98 0.63 0.64 L+F+S
CFS200-S750-R3 – 7.39 10.98 0.65 0.67 L+F+S
CFS200-S900-R1 – 6.56 10.98 0.58 0.60 L+F+S
CFS200-S900-R2 – 6.94 10.98 0.61 0.63 L+F+S
CFS200-S900-R3 – 7.22 10.98 0.64 0.66 L+F+S
CFS200-S1050-R1 – 6.32 10.98 0.56 0.58 L+F+S
CFS200-S1050-R2 – 6.79 10.98 0.60 0.62 L+F+S
CFS200-S1050-R3 – 7.04 10.98 0.62 0.64 L+F+S
Mean 0.64 0.66

it. The idealised cross-section can be imagined to have con- The moment capacities of the built-up beams with
tinuous welds at the web-flange junctions. Since the welds screw connections were compared with a cross-section
effectively cause the webs to deform in unison, the built-up assuming complete composite action, as summarized in
section should approximate the condition of full compos- Tables 5, 6,7 and8. The capacity of screw connected built-
ite action. The welded connection was simulated by using up beams with web depths 150 mm, 200 mm, 250 mm and
the TIE constraint with node to node discretization. This 300 mm was found to be 0.71, 0.64, 0.61 and 0.59 times
prevents the separation of individual channels and the web that of the idealized cross-section on average respectively.
plates behave as a single unit, thus imparting a high degree This reduction in the flexural capacity of the screw-con-
of flexural rigidity. This section was labelled as CFSxxx- nected built-up beams can be attributed to the complex
CCA and the ultimate capacity of this section was denoted interaction of local and global failures, the flexibility of
by M0 . The moment capacity of all other specimens was the web elements to move independent of each other and
denoted by Mi and was compared against that of the fully to the semi-rigid nature of the screw connections. Further,
integrated section. as seen in Fig. 19, the separation of channels at the web-
flange junction is completely inhibited, which is possibly

13
International Journal of Steel Structures (2023) 23(2):503–520 515

Table 7  Finite element Specimen Moment capac- Moment capac- Moment capacity, Comparison Failure
investigation into the effect of ity, M0 (kNm) ity, Mi (kNm) ­MNAS (kNm)
screw spacing on CFS beams Mi∕M
0
Mi∕M
NAS
with web depth 250 mm
CFS250-CCA​ 15.00 – 14.66 – – L+F
CFS250-S50-R1 – 9.44 14.66 0.63 0.64 L+F
CFS250-S50-R2 – 9.60 14.66 0.64 0.65 L+F
CFS250-S50-R3 – 9.77 14.66 0.65 0.67 L+F
CFS250-S75-R1 – 9.43 14.66 0.63 0.64 L+F
CFS250-S75-R2 – 9.58 14.66 0.64 0.65 L+F
CFS250-S75-R3 – 9.76 14.66 0.65 0.67 L+F
CFS250-S150-R1 – 9.26 14.66 0.62 0.63 L+F
CFS250-S150-R2 – 9.48 14.66 0.63 0.65 L+F
CFS250-S150-R3 – 9.69 14.66 0.65 0.66 L+F
CFS250-S275-R1 – 9.07 14.66 0.60 0.62 L+F+S
CFS250-S275-R2 – 9.32 14.66 0.62 0.64 L+F+S
CFS250-S275-R3 – 9.58 14.66 0.64 0.65 L+F+S
CFS250-S300-R1 – 9.05 14.66 0.60 0.62 L+F+S
CFS250-S300-R2 – 9.28 14.66 0.62 0.63 L+F+S
CFS250-S300-R3 – 9.57 14.66 0.64 0.65 L+F+S
CFS250-S450-R1 – 8.87 14.66 0.59 0.61 L+F+S
CFS250-S450-R2 – 9.06 14.66 0.60 0.62 L+F+S
CFS250-S450-R3 – 9.34 14.66 0.62 0.64 L+F+S
CFS250-S600-R1 – 8.68 14.66 0.58 0.59 L+F+S
CFS250-S600-R2 – 8.92 14.66 0.59 0.61 L+F+S
CFS250-S600-R3 – 9.19 14.66 0.61 0.63 L+F+S
CFS250-S750-R1 – 8.42 14.66 0.56 0.57 L+F+S
CFS250-S750-R2 – 8.78 14.66 0.59 0.60 L+F+S
CFS250-S750-R3 – 9.01 14.66 0.60 0.61 L+F+S
CFS250-S900-R1 – 8.17 14.66 0.54 0.56 L+F+S
CFS250-S900-R2 – 8.61 14.66 0.57 0.59 L+F+S
CFS250-S900-R3 – 8.92 14.66 0.60 0.61 L+F+S
CFS250-S1050-R1 – 7.95 14.66 0.53 0.54 L+F+S
CFS250-S1050-R2 – 8.43 14.66 0.56 0.58 L+F+S
CFS250-S1050-R3 – 8.72 14.66 0.58 0.59 L+F+S
Mean 0.61 0.62

another reason for the enhanced flexural capacity of the 5 Design Rules
fully composite sections.
From the comparison, it can be concluded that the com- The moment capacities of the built-up beams were calcu-
posite action decreases in screw-connected beams as the lated using the AISI’s North American Specification- AISI
web depth increases. This conclusion is further concreted S-100 (AISI-American Iron & Steel Institute, 2016). AISI
by the fact that the greatest reduction in screw connected S-100 and other notable design guidelines don’t take into
built up beam with 150 mm web depth is about 36% (in consideration the screw connection between the mem-
the specimen CFS150-S1050-R1) while as in beam with bers of a built-up section. Hence, it should be noted that
web depth 300 mm, the reduction in moment capacity moment capacity of the beams, used in the present inves-
is as high as 50% (in the specimen CFS300-S1050-R1). tigation, as obtained from the design calculations, is not
influenced by the screw spacing.

13
516 International Journal of Steel Structures (2023) 23(2):503–520

Table 8  Finite element Specimen Moment capac- Moment capac- Moment capacity, Comparison Failure
investigation into the effect of ity, M0 (kNm) ity, Mi (kNm) ­MNAS (kNm)
screw spacing on CFS beams Mi∕M
0
Mi∕M
NAS
with web depth 300 mm
CFS300-CCA​ 18.51 - 18.07 – – L+F
CFS300-S50-R1 – 11.54 18.07 0.62 0.64 L+F+S
CFS300-S50-R2 – 11.67 18.07 0.63 0.65 L+F
CFS300-S50-R3 – 11.99 18.07 0.65 0.66 L+F
CFS300-S75-R1 – 11.46 18.07 0.62 0.63 L+F+S
CFS300-S75-R2 – 11.61 18.07 0.63 0.64 L+F
CFS300-S75-R3 – 11.98 18.07 0.65 0.66 L+F
CFS300-S150-R1 – 11.25 18.07 0.61 0.62 L+F+S
CFS300-S150-R2 – 11.49 18.07 0.62 0.64 L+F
CFS300-S150-R3 – 11.85 18.07 0.64 0.66 L+F
CFS300-S275-R1 – 10.98 18.07 0.59 0.61 L+F+S
CFS300-S275-R2 – 11.33 18.07 0.61 0.63 L+F+S
CFS300-S275-R3 – 11.76 18.07 0.64 0.65 L+F+S
CFS300-S300-R1 – 10.92 18.07 0.59 0.60 L+F+S
CFS300-S300-R2 – 11.29 18.07 0.61 0.62 L+F+S
CFS300-S300-R3 – 11.74 18.07 0.63 0.65 L+F+S
CFS300-S450-R1 – 10.45 18.07 0.56 0.58 L+F+S
CFS300-S450-R2 – 11.07 18.07 0.60 0.61 L+F+S
CFS300-S450-R3 – 11.63 18.07 0.63 0.64 L+F+S
CFS300-S600-R1 – 10.08 18.07 0.54 0.56 L+F+S
CFS300-S600-R2 – 10.81 18.07 0.58 0.60 L+F+S
CFS300-S600-R3 – 11.35 18.07 0.61 0.63 L+F+S
CFS300-S750-R1 – 9.65 18.07 0.52 0.53 L+F+S
CFS300-S750-R2 – 10.44 18.07 0.56 0.58 L+F+S
CFS300-S750-R3 – 11.18 18.07 0.60 0.62 L+F+S
CFS300-S900-R1 – 9.46 18.07 0.51 0.52 L+F+S
CFS300-S900-R2 – 10.21 18.07 0.55 0.57 L+F+S
CFS300-S900-R3 – 10.90 18.07 0.59 0.60 L+F+S
CFS300-S1050-R1 – 9.32 18.07 0.50 0.52 L+F+S
CFS300-S1050-R2 – 10.03 18.07 0.54 0.66 L+F+S
CFS300-S1050-R3 – 10.67 18.07 0.58 0.59 L+F+S
Mean 0.59 0.61

According to AISI specification, the unfactored design b = w for 𝜆 ≤ 0.673 (6)

strength (Mn), using the effective width approach, is calcu-
lated as follows: Or
Mn = Se × Fy (4) b = 𝜌w for 𝜆 > 0.673 (7)

√
Se = Ix y
/ � � f
cg (5) 1.052 w
𝜆= √ (8)
k t
√
E
where Fy is the nominal yield strength, Se is the elastic sec-
tion modulus relative to top fibre, ycg is the depth of neutral
axis with respect to the compression flange and Ix is second 1 − 0.22∕𝜆
𝜌= ≤1 (9)
moment of area of the effective section, determined by using 𝜆
a reduction factor, given by:

13
International Journal of Steel Structures (2023) 23(2):503–520 517

Fig. 18  Moment capacities of specimens from finite element para- web depth 200 mm. c Ultimate moment capacities of CFS beams
metric study. a Ultimate moment capacities of CFS beams with web with web depth 250 mm. d Ultimate moment capacities of CFS
depth 150 mm. b Ultimate moment capacities of CFS beams with beams with web depth 300 mm

where b = effective design width; w = width of compression account for the reduction in the moment capacity due to the
element; ρ = reduction factor; k = plate buckling co-efficient; screw connection in built-up members. AISI S-100 is con-
t = thickness of compression element; E = modulus of elas- servative by as much as 34%, 42%, 46% and 48% for beams
ticity; f = maximum compressive edge stress in the element. with web depths 150 mm, 200 mm, 250 mm and 300 mm
The moment capacities of the beams, as calculated using respectively. }Figure 20a–d graphically shows the reduction
AISI S-100 are referred to as ­MNAS and have been listed in ratio Mi∕MNAS plotted against the screw spacing/ web depth
{

Tables 5, 6,7 and8 for the beams having web depths 150 mm, ratio for the beam specimens. It can be clearly observed that
200 mm, 250 mm and 300 mm respectively. From Tables 5, an increase in the screw spacing to web depth ratio leads to
6,7 and8 it can be observed that ultimate moments calculated a decrease in the moment capacity of the screw-connected
from AISI S-100 are in agreement with the values obtained built-up beams.
from the finite element investigation, when the specimens
are considered as fully composite. This also implies that
the design guidelines provided in AISI S-100 are unable to

13
518 International Journal of Steel Structures (2023) 23(2):503–520

Fig. 19  Failure of integrated

section with full composite
action

6 Conclusions depths 150 mm, 200 mm, 250 mm and 300 mm respec-
tively. This also indicates that the percentage decrease
The composite action of built-up CFS I-beams is investi- in strength is greater for beams with larger web depths.
gated experimentally and numerically in this article. Com- 2. Web buckling and flange distortion (L + F) was the
posite action is considered to be a function of screw spacing. dominant failure mode for all the web depths when the
Hence screw spacing is varied along the span as well as the screw spacing was 150 mm or less, while as separation
depth of the beams and the influence of the screw connec- of channels (S) was the dominant failure mode when the
tions on the integrity of beams is evaluated experimentally. screw spacing was greater than 450 mm. Hence the com-
Finite element models are developed and validated against posite action between the members is negatively affected
the experimental findings. The validated finite element mod- due to larger screw spacing.
els are then employed in a detailed parametric study involv- 3. The tendency of separation was more in the specimens
ing 124 specimens. An idealised cross section assuming with a single row of screws. Also, the separation of indi-
complete composite action was simulated using FE model vidual channels was more evident within the pure bend-
and the capacities of all the screw-connected “semi-rigid” ing zone of the built-up beams.
beams was compared against it. The failure modes, moment 4. No failure was observed in the screws at the failure, indi-
capacities and reduction in the moment capacities are evalu- cating the strength of screws was adequate for connec-
ated and discussed. From the experimental and FEA find- tion.
ings, the following conclusions can be drawn: 5. The capacity of screw connected built-up beams with
web depths 150 mm, 200 mm, 250 mm and 300 mm
1. In general, the flexural capacity is directly proportional was found to be 0.71, 0.64, 0.61 and 0.59 times that of
to the decrease in the screw spacing and increase in the the idealized cross-section on average, respectively. This
number of screw rows. When the longitudinal distance reduction in the flexural capacity of the screw-connected
between the screws for the investigated beams was built-up beams can be attributed to the complex interac-
increased from 50 to 1050 mm, the flexural capacity tion of local and global failures, the flexibility of the web
reduced by 11%, 14%, 17% and 19% for beams with web elements to move independent of each other and to the
semi-rigid nature of the screw connections compared to
continuously welded elements.

13
International Journal of Steel Structures (2023) 23(2):503–520 519

Fig. 20  Reduction in the moment capacity of screw-connected 200 mm. c Reduction in moment capacity of beams with web depth
beams. a Reduction in moment capacity of beams with web depth 250 mm. d Reduction in moment capacity of beams with web depth
150 mm. b Reduction in moment capacity of beams with web depth 150 mm

6. Composite action decreases in screw-connected beams provided by the Government of India. Also, the authors thankfully
as the web depth increases. The greatest reduction in acknowledge the National Institute of Technology Srinagar, India to
provide the infrastructure to conduct the present research.
screw connected built up beam with 150 mm web depth
is about 36% (in the specimen CFS150-S1050-R1) while Declarations
as in beam with web depth 300 mm, the reduction in
moment capacity is as high as 50% (in the specimen Conflict of Interests The authors report that they have no interests to
CFS300-S1050-R1) compared to the respective fully declare, that could have influenced the research presented in this paper.
integrated beams.
7. The design guidelines provided in AISI S-100 are una-
ble to account for the reduction in the moment capacity References
due to the screw connection in built-up members. AISI
ABAQUS CAE User’s Manual version 6.14 (2014) ‘Dassault Systemes
S-100 is conservative by as much as 34%, 58%, 54%
Simulia Corp.’ USA.
and 52% for beams with web depths 150 mm, 200 mm, AISI-American Iron and Steel Institute (2016) ‘North American Speci-
250 mm and 300 mm respectively. fication for the Design of Cold-Formed Steel Structural Members
AISI S100–16’, Aisi, 8(1).
Anapayan, T., Mahendran, M., & Mahaarachchi, D. (2011). Lateral
Acknowledgements This project was funded by a research grant under distortional buckling tests of a new hollow flange channel beam.
Technical Education Quality Improvement Programme (TEQIP-III) Thin-Walled Structures, 49(1), 13–25. https://​doi.​org/​10.​1016/j.​
tws.​2010.​08.​003

13
520 International Journal of Steel Structures (2023) 23(2):503–520

Chen, B., et al. (2020). Moment capacity of cold-formed channel beams Schafer, B. W., Li, Z., & Moen, C. D. (2010). Computational mod-
with edge-stiffened web holes, un-stiffened web holes and plain eling of cold-formed steel. Thin-Walled Structures, 48(10–11),
webs. Thin-Walled Structures. https://​doi.​org/​10.​1016/j.​tws.​2020.​ 752–762. https://​doi.​org/​10.​1016/j.​tws.​2010.​04.​008
107070 Selvaraj, S., & Madhavan, M. (2019). Structural design of cold-formed
Deepak, M. S., & Shanthi, V. M. (2018). Distortional buckling-moment steel face-to-face connected built-up beams using direct strength
resistance capacity of hybrid double-i-box beams. Journal of method. Journal of Constructional Steel Research, 160, 613–628.
Structural Engineering, 144(9), 04018132. https://​doi.​org/​10.​ https://​doi.​org/​10.​1016/j.​jcsr.​2019.​05.​053
1061/​(asce)​st.​1943-​541x.​00021​00 Siahaan, R., Keerthan, P., & Mahendran, M. (2018). Lateral distor-
Foraboschi, P. (2019). Lateral load-carrying capacity of steel columns tional buckling of rivet fastened rectangular hollow flange channel
with fixed-roller end supports. Journal of Building Engineering. beams. Journal of Constructional Steel Research, 144, 295–309.
https://​doi.​org/​10.​1016/j.​jobe.​2019.​100879 https://​doi.​org/​10.​1016/j.​jcsr.​2018.​01.​003
Foraboschi, P. (2020). Predictive formulation for the ultimate combina- ASTM Standard E8/E8M-13a (2013) Standard Test Methods for Ten-
tions of axial force and bending moment attainable by steel mem- sion Testing of Metallic Materials. ASTM International, pp. 1–27.
bers. International Journal of Steel Structures, 20(2), 705–724. Doi: https://​doi.​org/​10.​1520/​E0008_​E0008M-​13A.
https://​doi.​org/​10.​1007/​s13296-​020-​00316-6 Suhaib, M. B., & Tantray, M. A. (2022a). ‘Behaviour of cold-formed
Georgieva, I., et al. (2012a). Experimental investigation of built-up steel tensile members strengthened with GFRP using different
double-Z members in bending and compression. Thin-Walled techniques. Innovative Infrastructure Solutions. https://d​ oi.o​ rg/1​ 0.​
Structures, 53, 48–57. https://​doi.​org/​10.​1016/j.​tws.​2011.​12.​017 1007/​s41062-​022-​00851-7
Georgieva, I., et al. (2012b). Numerical study of built-up double-Z Suhaib, M. B., & Tantray, M. A. (2022b). Behaviour of GFRP stiff-
members in bending and compression. Thin-Walled Structures, ened cold-formed steel built-up beams. International Journal
60, 85–97. https://​doi.​org/​10.​1016/j.​tws.​2012b.​07.​005 of Steel Structures, 22(4), 1042–1059. https://​doi.​org/​10.​1007/​
Georgieva, I., Schueremans, L., & Pyl, L. (2012). Composed columns s13296-​022-​00621-2
from cold-formed steel Z-profiles: Experiments and code-based Wang, L., & Young, B. (2015). Beam tests of cold-formed steel built-
predictions of the overall compression capacity. Engineering up sections with web perforations. Journal of Constructional Steel
Structures, 37, 125–134. https://d​ oi.o​ rg/1​ 0.1​ 016/j.e​ ngstr​ uct.2​ 011.​ Research, 115, 18–33. https://​doi.​org/​10.​1016/j.​jcsr.​2015.​08.​001
12.​017 Wang, L., & Young, B. (2016a). Behavior of cold-formed steel built-up
Ghannam, M. (2019). Bending moment capacity of cold-formed steel sections with intermediate stiffeners under bending I: Tests and
built-up beams. International Journal of Steel Structures, 19(2), numerical validation. Journal of Structural Engineering, 142(3),
660–671. https://​doi.​org/​10.​1007/​s13296-​018-​0155-2 04015150. https://​doi.​org/​10.​1061/​(asce)​st.​1943-​541x.​00014​28
Islam, S. M. Z., Cai, Y., & Young, B. (2019). Design of CFRP-strength- Wang, L., & Young, B. (2016b). Behavior of cold-formed steel built-up
ened stainless steel tubular sections subjected to web crippling. sections with intermediate stiffeners under bending. II: Paramet-
Journal of Constructional Steel Research, 159, 442–458. https://​ ric study and design. Journal of Structural Engineering, 142(3),
doi.​org/​10.​1016/j.​jcsr.​2019.​04.​043 04015151.
IS Code (2008) Cold Reduced Low Carbon Steel Sheet And Strip (Fifth Wang, L., & Young, B. (2017). Design of cold-formed steel built-up
Revision), Indian Standards. sections with web perforations subjected to bending. Thin-Walled
Keerthan, P., & Mahendran, M. (2013). Suitable stiffening systems Structures, 120, 458–469. https://​doi.​org/​10.​1016/j.​tws.​2017.​06.​
for LiteSteel beams with web openings subjected to shear. Jour- 016
nal of Constructional Steel Research, 80, 412–428. https://​doi.​ Wang, L., & Young, B. (2018). Behaviour and design of cold-formed
org/​10.​1016/j.​jcsr.​2012.​08.​004 steel built-up section beams with different screw arrangements.
Laím, L., et al. (2014). Experimental analysis on cold-formed steel Thin-Walled Structures, 131, 16–32. https://d​ oi.o​ rg/1​ 0.1​ 016/j.t​ ws.​
beams subjected to fire. Thin Walled Structures, 74, 104–117. 2018.​06.​022
https://​doi.​org/​10.​1016/j.​tws.​2013.​09.​006 Xu, L., Sultana, P., & Zhou, X. (2009a). Flexural strength of cold-
Laím, L., Paulo, J., & Rodrigues, C. (2016). Numerical analysis formed steel built-up box sections. Thin-Walled Structures, 47(6–
on axially-and-rotationally restrained cold-formed steel beams 7), 807–815. https://​doi.​org/​10.​1016/j.​tws.​2009a.​01.​005
subjected to fi re. Thin Walled Structures, 104, 1–16. https://​doi.​
org/​10.​1016/j.​tws.​2016.​03.​004 Publisher's Note Springer Nature remains neutral with regard to
Laím, L., & Rodrigues, J. P. C. (2018). Fire design methodologies jurisdictional claims in published maps and institutional affiliations.
for cold-formed steel beams made with open and closed cross-
sections. Engineering Structures, 171, 759–778. https://​doi.​org/​ Springer Nature or its licensor (e.g. a society or other partner) holds
10.​1016/j.​engst​r uct.​2018.​06.​030 exclusive rights to this article under a publishing agreement with the
Laím, L., Rodrigues, J. P. C., & da Silva, L. S. (2013). Experimen- author(s) or other rightsholder(s); author self-archiving of the accepted
tal and numerical analysis on the structural behaviour of cold- manuscript version of this article is solely governed by the terms of
formed steel beams. Thin-Walled Structures, 72, 1–13. https://​ such publishing agreement and applicable law.
doi.​org/​10.​1016/j.​tws.​2013.​06.​008
Roy, K., et al. (2021). Flexural behaviour of back-to-back built-up
cold-formed steel channel beams: Experiments and finite ele-
ment modelling. Structures, 29, 235–253. https://​doi.​org/​10.​
1016/j.​istruc.​2020.​10.​052

13