Journal of Constructional Steel Research 121 (2016) 216–228

Contents lists available at ScienceDirect

Journal of Constructional Steel Research

Experimental study on the rotation capacity of cold-formed steel beams

Mohammed H. Serror ⁎, Emad M. Hassan, Sherif A. Mourad
Department of structural engineering, Faculty of engineering, Cairo University, Egypt

a r t i c l e i n f o a b s t r a c t

Article history: The rotation capacity of cold-formed steel (CFS) beams has been evaluated through experimental investigation.
Received 23 December 2015 Studies on different structural levels have been performed. At the element level, different profile slenderness ra-
Received in revised form 26 January 2016 tios have been considered, and different section shapes have been investigated by increasing the number of
Accepted 5 February 2016
flange bends: C-section and curved-section, which represents an infinite number of flange bends. At the connec-
Available online 22 February 2016
tion level, a web bolted moment resistant type of connection using through plate has been adopted. In web bolted
Keywords:
connections without out-of-plane stiffeners, premature web buckling results in early loss of strength. Hence, out-
Experimental study of-plane stiffeners have been examined to delay web and flange buckling and to produce relatively high moment
Cold-formed steel beam strength and ductility. The experimental results have been compared with numerical results obtained by the au-
Rotation capacity thors in another paper work. The results revealed that increasing the number of flange bends will not in all cases
Slenderness ratio enhance the behavior. Meanwhile, the use of out-of-plane stiffeners can increase the seismic energy dissipation,
Section shape the moment strength and the ductility, when compared with the case without stiffeners.
Out-of-plane stiffeners © 2016 Elsevier Ltd. All rights reserved.

1. Introduction
The use of CFS sections as main structural elements in building con-
structions is mainly limited to roof and side purlins, mezzanine floors,
storage racks and stud-wall frames with low seismic energy dissipation
capacity [1–6]. Generally, moment-resisting frames with CFS sections
cannot create plastic hinges in CFS beams; and in turn, cannot produce
adequate ductility for high seismic resistance [7–10]. The previous stud-
ies on behavior of CFS elements and associated beam-column connec-
tions in moment-resisting frames under monotonic and cyclic loading
are relatively limited [11–15]. It has been reported that the ductility is
mainly provided by the connection (yielding the material around the
bolt holes) while the beams and columns remain elastic [4–6]. Keerthan
and Mahendran [16] performed experimental study for the effect of
flange straps on the shear capacity and shear behavior of lipped C-
section. It has been reported that the use of flange straps enhances the
specimens shear and bending strength with 9 to 20% of its original
value. Chen et al. [17] measured the specimen capacity, modes of failure,
and local deformation for CFS lipped C-sections under web crippling
load. It has been reported that the section capacity increases with the in-
crease of bearing plate stiffness. Bending tests have been performed to
simply supported CFS beams with C-section, lapped Z-section, and
coupled back-to-back channel section, under four-point and three-
point loadings [18–20]. Yang et al. [14] investigated the energy dissipa-
tion and failure mechanism of CFS hollow rectangular columns and
reported an effective range for the wall length-to-width ratio. Padilla
⁎ Corresponding author.
E-mail address: serror@eng.cu.edu.eg (M.H. Serror).
et al. [21] studied the energy dissipation of CFS beams with C-section
under both monotonic and cyclic loading conditions. It has been report-
ed that the amount of energy dissipation is dependent on buckling
modes. Wang and Zhang [22] investigated the effect of edge stiffeners
on the flexural behavior of CFS beams with C-section. The experimental
results revealed that the edge stiffeners greatly increase the flexural
strength and affect the buckling mode. In addition, the beam flexural re-
sistance in case of non-pure bending is higher than that of pure bending
case. Bagheri et al. [12,23] performed full-scale experimental tests for
CFS structures under gravity and cyclic loading. Particular details for
the beam-to-column connection have been investigated to resist seis-
mic loads. It has been reported that the specimens could achieve a flex-
ural strength of 60% the plastic moment in association with a rotation
angle of 0.04 rad. Mirghaderi et al. [24] presented a new I-beam to
box-column connection by a vertical plate passing through the column.
It has been reported that the through plate connection can effectively
transfer the moment from the beam to the column with a rotation
angle of 0.04 rad. Experimental investigations have been performed
for specimens with different connection configurations, to evaluate
the strength and stiffness of bolted connections between CFS elements
[25,26]. Four modes of failure have been reported. Two modes have
been observed in the CFS section: due to bearing around the bolt hole;
and due to flexure. The former mode of failure could achieve a bending
strength up to 50% of the plastic moment. The other two modes of fail-
ure have been observed in the gusset plate: due to lateral torsion buck-
ling; and due to flexure. The later mode could achieve a bending
strength up to 75% of the plastic moment. Apparently, the bearing fail-
ure mode should be avoided as it reduces the strength of the connection.
Chung and Lau [27] studied the behavior of CFS in moment connections

http://dx.doi.org/10.1016/j.jcsr.2016.02.005
0143-974X/© 2016 Elsevier Ltd. All rights reserved.
 M.H. Serror et al. / Journal of Constructional Steel Research 121 (2016) 216–228 217

Fig. 1. Experiment setup — support system, loading system, specimen components and instrumentation.
218 M.H. Serror et al. / Journal of Constructional Steel Research 121 (2016) 216–228

Fig. 2. Experiment setup — laboratory snapshots.

 M.H. Serror et al. / Journal of Constructional Steel Research 121 (2016) 216–228 219

resisting frame buildings. Both beam and connection elements have

been examined and the rotation capacity has been evaluated. A beam-
to-column connection has been adopted by employing a bolted web
to through plate connection and assuming rigid support for the column,
as shown in Fig. 1. This assumption is in conformance with the numer-
ical modeling and study that have been performed by the authors [29],
for the sake of verification and comparison between the numerical and
experimental results.

2. Experiment setup

Fig. 3. Cyclic loading as per AISC seismic provisions [9].
through testing specimens of two back-to-back channels connected to a
column using through plate. It has been reported that the hunched
through plate is more effective in moment resistance compared with
other shapes, it also improves the overall connection resistance.
Zahurul-Islam and Young [28] tested CFS hollow square section (HSS)
strengthened with fiber-reinforced polymer (FRP) under web crippling.
It has been reported that the FRP increases the strength of the CFS HSS
with a range of 4% to 76% of the un-strengthened section, according to
the loading conditions.
Hence, if CFS beams are used as main dissipative elements in earth-
quake resistant frame buildings, their ductility needs to be significantly
improved by delaying local buckling and allowing development of
large plastic deformations. This can be achieved by: optimizing the
shapes of CFS sections; using pertinent connection details, such as gusset
or through plates; and employing stiffeners in the beam-column connec-
tions [4–6,11,12].
This paper presents experimental investigation that underpins the
usage of CFS sections as energy dissipative elements for moment-
Figs. 1 and 2 illustrate the experiment setup that includes four main
components, namely: supporting system, loading system, specimen and
instrumentation.
The supporting system (see Figs. 1a, b and 2) consists of:
(1) supporting column that is of hot-rolled I-beam, which is rigidly con-
nected to the floor through anchor bolts, and (2) supporting lateral sys-
tem to avoid premature global instability of the loaded specimen, where
foam layer has been employed at the interface with the specimen to re-
duce the anticipated friction. This is in accordance with the literature
provisions [9], which require that both flanges of the beam shall be lat-
erally braced near plastic hinges and regions with concentrated forces.
The loading system consists of automated hydraulic jack with capac-
ity of 900 kN that is connected to a reaction frame from one side and to
the specimen from the other side through pinned connections (see Fig.
1a). A load cell has been attached to the loading system and has been
calibrated before each test. Both monotonic and cyclic loadings have
been applied to the beams. The cyclic loading, however, has been ap-
plied in accordance with the protocol given in section S6.2 of AISC Seis-
mic Provisions [9], as shown in Fig. 3. Meanwhile, the rotation angle (θ)
has been replaced with the displacement (Δ), where [θ = Δ/L′] and (L′)
is the effective length of the beam to the end of the through plate as
shown in Fig. 1c.
The specimen has been established for 2 m long cantilever CFS
beams that have been selected to represent a 4 m beam span in a later-
ally loaded moment resisting frame building. A web bolted to through-

Table 1
Monotonic loading specimens.

Specimen ID CFS beam Through plate Steel grade Bolts diameter (Grade)

Beam section shape Dimensions (mm) Stiffeners configuration Dimensions (mm) (H2/H1 × B2/B1) tp (mm)

B01 C 200 × 85 × 20/2 CONFIG-0 550/200 × 400/160 8 ST-37/2 16 (10.9)

B02 C 200 × 85 × 20/2 CONFIG-4 550/200 × 400/160 8 ST-37/2 16 (10.9)
B03 C 300 × 100 × 25/3 CONFIG-0 650/350 × 450/175 10 ST-37/2 16 (10.9)
B04 C 300 × 100 × 25/3 CONFIG-4 650/350 × 450/175 10 ST-37/2 16 (10.9)
B05 Curved 200 × 40/2 CONFIG-0 550/200 × 400/160 8 ST-37/2 16 (10.9)
B06 Curved 300 × 50/3 CONFIG-0 650/350 × 450/175 10 ST-37/2 16 (10.9)

Fig. 4. Stiffeners configurations: (a) CONFIG-0 and (b) CONFIG-4.

220 M.H. Serror et al. / Journal of Constructional Steel Research 121 (2016) 216–228

Table 2
Cyclic loading specimens.

Specimen ID CFS beam Through plate Steel grade Bolt diameter (Grade)

Beam section shape Dimensions (mm) Stiffeners configuration Dimensions (mm) (H2/H1 × B2/B1) tp (mm)

B07 C 300 × 100 × 25/3 CONFIG-0 650/350 × 450/175 10 ST-37/2 16 (10.9)

B08 C 300 × 100 × 25/3 CONFIG-4 650/350 × 450/175 10 ST-37/2 16 (10.9)
B09 Curved 300 × 50/3 CONFIG-0 650/350 × 450/175 10 ST-37/2 16 (10.9)
B10 Curved 300 × 50/3 CONFIG-4 650/350 × 450/175 10 ST-37/2 16 (10.9)

plate beam-to-column connection has been adopted where the through
plate has been welded to the supporting column that was made up of
hot-rolled I-beam instead of CFS column, as shown in Fig. 1b. This as-
sumption is consistent with the literature [30], where it has been re-
ported that in this type of connection, if designed properly, the
column remain elastic without large deformations or local buckling.
Large deformation and local buckling only take place in the beams. Fig.
1c illustrates the specimen model along with the definition of beam
maximum moment (M = FL′) and rotation angle (θ), where the plastic
hinge region of the beam is assumed to be at the end of the through

Fig. 5. The deformations observed in specimens B01 and B02.

 M.H. Serror et al. / Journal of Constructional Steel Research 121 (2016) 216–228 221

Table 3 installed to measure the strain in the through plate. The strain gauge
Mechanical proprieties of specimens' material. SG-3 has been installed in the beam at location of last row of bolts
Coupon Specimen Yield stress Ultimate strength Elongation connecting the beam to the through plate. SG-3 measures the strain in
no. ID (N/mm2) (N/mm2) (%) the beam flange. The strain gauges SG-2 and SG-4 have been installed
1 B02 290 386 35 in the beam at a distance of 200 mm away from the last row of bolts
2 B03 239 321 18 connecting the beam to the though plate. SG-2 and SG-4 measure the
3 B05 265 342 22 strain in the beam web and flange elements, respectively. Fig. 1e also
4 B06 290 377 23
shows the arrangement of six LVDTs (LVDT-1 to LVDT-6) along the
5 B10 282 376 22
specimen. The LVDTs are of 100 to 200 mm strokes length to be able
to measure the maximum deflection of the specimen, which has been
plate. Fig. 1d illustrates the adopted CFS beam section shapes and estimated to be 96 mm according to the AISC seismic provision [9].
through plate, which comply with those adopted in the numerical LVDT-1 has been installed at the point of load application to measure
study performed by the authors [29]. the maximum deflection of the beam. LVDT-2 has been installed at the
The instrumentation includes strain gauges (SG) and linear variable level of the lateral support to measure the beam displacement at this lo-
displacement transducers (LVDT) which have been installed to measure cation. LVDT-3 and LVDT-4 have been installed at the levels of first and
the internal strain and specimen displacement, respectively. Fig. 1e last row of connection bolts, respectively. LVDT-5 and LVDT-6 have been
shows the arrangement of four strain gauges (SG-1 to SG-4) along the installed at the head plate connecting the through plate to the
specimen to measure the axial strains due to monotonic loading. Electri- supporting column. A data acquisition system of 40-port data logger
cal strain gauges have been employed with 6-mm length, 120-Ohm re- has been used to retrieve the data from load cell, strain gauges and
sistance and calibration at 30 °c. One strain gauge (SG-1) has been LVDTs.

Fig. 6. The deformations observed in specimens B03 and B04.

222 M.H. Serror et al. / Journal of Constructional Steel Research 121 (2016) 216–228
Table 1 lists the characteristics of six specimens (B01 to B06) which
have been tested under monotonic loading. Two specimens of C-section
have been tested versus two specimens of curved-section with the same
beam height and thickness to investigate the effect of section shape.
Furthermore, two specimens have been tested in association with out-
of-plane stiffeners to report the stiffeners effect. It is worth noting that
the authors have studied numerically the effect of four different stiff-
eners configurations [29], namely: CONFIG-1 to CONFIG-4. The configu-
ration CONFIG-4, as shown in Fig. 4, has been adopted in this
experimental study, since it represents the most efficient configuration
[29]. On the other hand, Table 2 lists the characteristics of four speci-
mens (B07 to B10) which have been tested under cyclic loading. Simi-
larly, beam section shape and stiffeners are of interest in specimens'
selection. It is worth noting that all bolts used are of friction-type M16
grade 10.9 and tightened with a torque complying with the code re-
quirements [11].
Five coupons have been extracted from the test specimens, where
uniaxial tensile test has been performed to determine the mechanical
properties of specimens' material. The obtained material properties
have been listed in Table 3.
The numerical study [29] has been performed through nonlinear in-
elastic post-buckling analysis, using ANSYS program [31]. The boundary
Fig. 7. The deformations observed in specimens B05 and B06.
conditions, loading points and constraints have been simulated in accor-
dance with the experimental setup shown in Fig. 1. The numerical
model basic parameters are: element type: 4-node shell element
(shell181), mesh size: 20 mm × 20 mm selected upon mesh sensitivity
analysis, and material: bi-linear stress–strain behavior with Fy =
240 MPa (yield stress), Fu = 360 MPa (tensile stress), E = 210 GPa
(elasticity modulus), Es = E/100 (hardening modulus) and ʋ = 0.33
(Poisson's ratio). The bolts are modeled by coupling the nodes lying at
the bolt centerline in all degrees of freedom. Meanwhile, the connector
between the beam back-to-back elements is modeled using rigid beam
element (MPC184). The details of the numerical study have been pre-
sented in another paper work [29].
3. Results and discussion
3.1. Monotonic loading results
Figs. 5 to 7 show the deformations in specimens B01 to B06, where a
comparison between results of the experimental tests and those of the
corresponding numerical models [29] has been illustrated. Apparently,
the experimental results are in good match with those of the numerical
predictions, in terms of location and mode of instability and location of
 M.H. Serror et al. / Journal of Constructional Steel Research 121 (2016) 216–228 223

maximum stresses. It is evident also that the instability and the maxi-
mum stresses take place at the end of the through plate where the
loads are transferred from the beam flanges to the web at the connec-
tion bolts. Furthermore, it is sound that the utilization of out-of-plane
stiffeners greatly delays the premature buckling and reduces the maxi-
mum stresses.
Figs. 8 to 10 show the recorded M-θ curves and strain gauges read-
ings for specimens B01 to B06. Specimens B01 and B03, at rotation
angle of 0.01 rad, undergo premature local buckling in both the
compression flange and the web, at the end of the through plate. Be-
yond 0.01 rad, distortional buckling takes place with significant loss of
strength. On the other hand, specimens B02 and B04 did not experience
instability until a rotation angle of 0.03 rad, where loss of strength has
been observed then. The delay of premature buckling is attributed to
the contribution of the out-of-plane stiffeners. The enhancement in
strength, initial stiffness and ductility due to the employment of out-
of-plane stiffeners (CONFIG-4) reaches up to 30%, 15% and 115%, respec-
tively, compared with the un-stiffened specimens (CONFIG-0).

Fig. 8. The moment-rotation behavior of specimens B01 to B06, in comparison with the FE model results.
224 M.H. Serror et al. / Journal of Constructional Steel Research 121 (2016) 216–228

Fig. 8 shows the recorded M-θ curves for specimens B01 to B06 in
comparison with the results of the corresponding numerical models
[29]. It is evident that the tested specimens B01, B03, B05 and B06
could achieve a strength ranging from 0.60 Mp up to 0.75 Mp at a rota-
tion angle of 0.01 rad without employing out-of-plane stiffeners, while
exceeding the slenderness ratio limits of the design codes [7–11]. On
the other hand, the tested specimens B02 and B04, employing out-of-
plane stiffeners with CONFIG-4, could achieve a strength ranging from
0.80 Mp up to 0.98 Mp at a rotation angle of 0.03 rad. The differences be-
tween the experimental results and the numerical predictions can be
attributed to: imperfections; differences between the material proper-
ties of the test specimen and the numerical model; actual dimensions
of the test specimens; and bolted connection slip. These factors were
not taken into account by the numerical analysis at this stage. Therefore,
they can affect the buckling mode shape, the bending-rotation behavior
and in particular the strength degradation.
Fig. 9 shows the readings of strain gauge SG-1 that reports the in-
duced strain in the through plate. It is evident that the strain level did
not exceed the yield limit; and in turn, yielding did not take place in
the through plate. This reveals that the failure shall always take place

Fig. 9. The readings of strain gauges SG-1 and SG2 for specimens B01 to B06.
 M.H. Serror et al. / Journal of Constructional Steel Research 121 (2016) 216–228 225

in the beam element rather than in the connection. This is matching the
target criteria of seismic design codes [7–11], where plastic hinge is
formed in the CFS beam in front of the through plate away from the con-
nection, enabling a ductile behavior without sudden failure. Upon plas-
tic hinge formation in the CFS beam, the strain level in the through plate
starts to decrease, as shown in the figures.
Fig. 9 shows also the readings of strain gauge SG-2 that reports the
induced strain in the beam web. When comparing the readings of spec-
imens B01 and B03 (Fig. 9a and c) with those of B05 and B06 (Fig. 9e and
f), the recorded strains in specimens B05 and B06 are higher than those
induced in specimens B01 and B03. However, these web strains are due
to premature local buckling, below the yield strain level. This reveals
that the webs of B05 and B06 did experience a local buckling level
that is higher than that occurred in the webs of B01 and B03, particularly
in the rotation range of 0.01 to 0.03 rad. It is also evident that the web
stain is minimal in specimens B02 and B04 (Fig. 9b and d), where out-
of-plane stiffeners are employed.
Fig. 10 shows the readings of strain gauge SG-3 (at the plastic hinge
location) in comparison with the readings of SG-4 (at 200 mm away
from the plastic hinge location). The readings are almost identical

Fig. 10. The readings of strain gauges SG-3 and SG4 for specimens B01 to B06.
226 M.H. Serror et al. / Journal of Constructional Steel Research 121 (2016) 216–228
until a rotation angle that is associated with the onset of instability and
loss of strength. Then a sound bifurcation takes place, where the strain
level at the plastic hinge location drastically increases much more
than that at location away from the plastic hinge. When comparing
the readings of strain gauge SG-3 of specimens B01 and B03 (Fig. 10a
and c) with those of B05 and B06 (Fig. 10e and f), the recorded strains
in specimens B05 and B06 are much higher than those induced in spec-
imens B01 and B03. The flanges of B01 and B03 did experience prema-
ture local buckling and could not reach the yield strain level.
Contrarily, the flanges of B05 and B06 could undergo significant yield-
ing. It is also evident that the flange stain in specimens B02 and B04
(Fig. 10b and d), where out-of-plane stiffeners are employed, did not
exceed the yield limit until a rotation angle of more than 0.02 rad,
then yielding started to emerge in associating with the rotation in-
crease. It is worth noting that the slope of the rotation-strain curve
after yielding is more steep for specimens B05 and B06 compared
with that of specimens B02 and B04, due to the stabilizing effect of the
out-of-plane stiffeners.
The aforementioned observations of strain gauges SG-2, SG-3 and
SG-4 are evident in the deformation snapshots shown in Figs. 5a, 6a,
7a and b for specimens B01, B03, B05 and B06, respectively. These ob-
servations are matching with the findings of the numerical study [29],
where increasing the number of flange bends delays the premature
buckling of the flange; however, it inversely affects the premature
Fig. 11. The deformations observed in specimens B07 to B10.
bucking of the web. Therefore, by increasing the angle between the
web and the stiffening flange the effectiveness of web stiffening de-
creases. Since the adopted beam-to-column connection is “web bolted
to through plate”, the loads/stresses are transferred from the beam
flanges to the web at the connection. Consequently, the web stiffening
has more influence on the beam behavior compared with the flange
stiffening. This is the essence behind the superior behavior of the C-
section over the curved-sections within the range of the studied speci-
mens. This superior behavior is evident by comparing M-θ curves of
specimens B01 and B03 (Fig. 8a and c) with those of B05 and B06 (Fig.
8e and f), where the initial stiffness and strength of the C-section are
higher than those of the curved one.
3.2. Cyclic loading results
Specimens B07, B08 and B09 are identical to B03, B04 and B06, re-
spectively, however subjected to cyclic loading. Meanwhile, specimen
B10 is identical to B09, however stiffened with CONFIG-4.
Fig. 11 shows the deformations in specimens B07 to B10. Apparently,
the instability and the maximum deformations take place at the end of
the through plate where the loads are transferred from the beam flanges
to the web at the connection. Moreover, it is evident that the utilization
of out-of-plane stiffeners greatly delayed the premature buckling.
 M.H. Serror et al. / Journal of Constructional Steel Research 121 (2016) 216–228
Fig. 12 shows the recorded hysteretic M-θ curves as resulting from
the experimental tests in comparison with those resulting from the nu-
merical models [29], for specimens B07 to B10. It is evident that the ex-
perimental results are in good match with the numerical predictions,
where the differences are attributed to the aforementioned factors: im-
perfections; material properties; actual dimensions of test specimens;
and slip of bolts, in addition to the cyclic loading effects.
Specimens B07 and B09 could achieve a strength of 0.80 Mp and
0.65 Mp at a rotation angle of 0.01 rad and 0.02 rad, respectively. Be-
yond this rotation angle, premature local buckling occurs and loss of
strength takes place. In specimen B07 (C-section), the onset of buckling
occurs at the flanges and then followed with the web buckling. In spec-
imen B09 (curved section), however, the onset of buckling occurs at the
web and then followed with the flanges buckling. This is consistent with
the monotonic loading results, and attributed to the fact that the flanges
are weaker in the C-section compared with the curved section (infinite
number of bends); meanwhile, the web is stronger in the C-section
compared with the curved section (larger angle between flange and
web). It is worth noting that the weakness of the web in the curved sec-
tion was behind the observed strength reduction, where loading is
transferred from flanges to web at the connection. On the other hand,
the delay of flange buckling in the curved section enabled specimen
B09 to undergo larger rotation (more ductile) compared with the C-
section of specimen B07.
Specimens B08 and B10 could achieve a strength of 0.98 Mp and
1.20 Mp at a rotation angle of 0.04 rad and 0.03 rad, respectively,
while employing out-of-plane stiffeners with CONFIG-4. Beyond this ro-
tation angle, instability and loss of strength have been observed to
onset. The employment of out-of-plane stiffeners was behind the
Fig. 12. The hysteretic behavior of specimens B07 to B10, in comparison with the FE model results.
227
enhancement in the initial stiffness, strength and ductility of the speci-
mens. This is evident when comparing specimen B09 with B10, where
the later could achieve 190% higher strength. Hence, the behavior of
specimens B08 and B10 can be classified as “SMRF” and “rigid connec-
tion” according to AISC and Eurocode-3, respectively [7,8].
The amount of energy dissipation (area confined within the hyster-
esis loop) equal to 5.42 × 105 N.m.radian and 10.93 × 105 N.m.radian for
specimens B07 and B08, respectively. Meanwhile, the amount of energy
dissipation equal to 8.35 × 105 N.m.radian and 10.03 × 105 N.m.radian
for specimens B09 and B10, respectively. Hence, the enhancement in
energy dissipation due to the utilization of out-of-plane stiffeners
(CONFIG-4) equal to 120% and 25% for specimens with C-section and
curved section, respectively. Accordingly, the beam with C-section is
in need for stiffeners in order to enhance the ductility and energy dissi-
pation. Meanwhile, the beam with curved section can afford adequate
ductility and energy dissipation; however, it is in need for stiffeners in
order to enhance the strength.
4. Summary and conclusions
In this study, an experimental investigation has been presented to
underpin the usage of cold-formed steel (CFS) beams as energy dissipa-
tive elements, where the rotation capacity has been evaluated. At the el-
ement level, different profile slenderness ratios have been considered,
and different section shapes have been investigated by increasing the
number of flange bends: C-section and curved section. In addition,
out-of-plane stiffeners have been employed. At the connection level, a
web bolted to through plate beam-to-column connection has been
adopted. Both monotonic and cyclic loading procedures have been
228 M.H. Serror et al. / Journal of Constructional Steel Research 121 (2016) 216–228
applied. The experimental investigation has been performed in parallel
with a numerical investigation by the authors (presented in another
paper work), where results of both investigations have been compared
and verified. The main conclusions, within the range of the studied
parameters, can be summarized as follows:
1. Specimens without stiffeners could achieve a strength up to 0.75 Mp
at a rotation angle up to 0.01 rad. Meanwhile, specimens with stiff-
eners could achieve a strength up to 1.20 Mp at a rotation angle up
to 0.04 rad. Hence, the utilization of stiffeners greatly enhances
strength, initial stiffness, ductility, and energy dissipation of the CFS
beam.
2. Increasing the number of flange bends (curved flange) delays the
premature buckling of the flange; however, it inversely affects the
premature bucking of the web. Since the adopted connection is
“web bolted to through plate”, the web buckling has more influence
on the beam behavior compared with the flange buckling, where the
loads/stresses are transferred from the beam flanges to the web at
the connection. Therefore, the CFS beams with C-section did show
a superior behavior over the beams with curved section.
3. The beam with C-section is in need for stiffeners in order to enhance
the ductility and energy dissipation. Meanwhile, the beam with
curved section can afford adequate ductility and energy dissipation;
however, it is in need for stiffeners in order to enhance the strength.
4. The strain level in the through plate did not exceed the yield limit.
Meanwhile, the plastic hinge is formed in the CFS beam in front of
the through plate, where instability and maximum stresses take
place. This reveals that the failure shall always take place in the
beam element rather than in the connection, enabling a ductile be-
havior without sudden failure.
5. The experimental results are in good match with those of the numer-
ical study. Meanwhile, the reported differences can be attributed to:
imperfections; differences between the material properties of the
test specimen and the numerical model; actual dimensions of the
test specimens; bolted connection slip; and cyclic loading effects.
Acknowledgment
The authors would like to express their gratitude to Cairo University,
Faculty of Engineering, for funding the experimental testing.
References
[1] R.M. Lawson, New developments in use of cold formed steel in buildings, Constr.
Build. Mater. 6 (1992) 201–204.
[2] D. Dubina, Behavior and performance of cold-formed steel-framed houses under
seismic action, J. Constr. Steel Res. 64 (2008) 896–913.
[3] H. Moghimi, H. Ronagh, Performance of light-gauge cold-formed steel strap- braced
stud walls subjected to cyclic loading, Eng. Struct. 31 (2009) 69–83.
[4] M. Casafont, A. Arnedo, F. Roure, A. Rodrıguez-Ferran, Experimental testing of joints
for seismic design of lightweight structures. Part1. Screwed joints in straps, Thin-
Walled Struct. 44 (2006) 197–210.
[5] M. Casafont, A. Arnedo, F. Roure, A. Rodrıguez-Ferran, Experimental testing of joints
for seismic design of lightweight structures. Part2: bolted joints in straps, Thin-
Walled Struct. 44 (2006) 677–691.
[6] M. Casafont, A. Arnedo, F. Roure, A. Rodrıguez-Ferran, Experimental testing of joints
for seismic design of lightweight structures. Part3: gussets, corner joints, x-braced
frames, Thin-Walled Struct. 45 (2007) 637–659.
[7] Eurocode 3: Design of Steel Structures: Part 1.1: General Rules and Rules for Build-
ings, EN 1993-1-1, 2005.
[8] ANSI/AISC 360-05, Specification for Structural Steel Buildings, American Institute of
Steel Construction, Illinois, 2005.
[9] ANSI/AISC 341-05, Seismic Provisions for Structural Steel Buildings, American Insti-
tute of Steel Construction (AISC), Illinois, 2005.
[10] AISI, S110. Standard for Seismic Design of Cold-Formed Steel Structural Systems -
Special Bolted Moment Frames, American Iron and Steel Institute, Washington,
DC, 2007.
[11] ECP-205 (ASD), Egyptian Code of Practice for Steel Construction and Bridges, Hous-
ing and Research Center, Giza. Egypt, 2011.
[12] A. Bagheri Sabbagh, R. Mirghaderi, M. Petkovski, K. Pilakoutas, An integrated thin-
walled steel skeleton structure (two full-scale tests), J. Constr. Steel Res. 66 (2010)
470–479.
[13] B. Calderoni, A. De Martino, A. Formisano, L. Fiorino, Cold formed steel beams under
monotonic and cyclic loading: experimental investigation, J. Constr. Steel Res. 65
(2009) 219–227.
[14] N. Yang, Y. Zhong, Q. Meng, H. Zhang, Hysteretic behaviors of cold-formed steel
beam-columns with hollow rectangular section: experimental and numerical simu-
lations, Thin-Walled Struct. 80 (2014) 217–230.
[15] C.-M. Uang, A. Sato, J.-K. Hong, K. Wood, Cyclic testing and modeling of cold-formed
steel special bolted moment frame connections, J. Struct. Eng. 136 (2010) 953–960.
[16] P. Keerthan, M. Mahendran, Experimental investigation and design of lipped chan-
nel beams in shear, Thin-Walled Struct. 86 (2015) 174–184.
[17] Y. Chen, X. Chen, C. Wang, Experimental and finite element analysis research on
cold-formed steel lipped channel beams under web crippling, Thin-Walled Struct.
87 (2015) 41–52.
[18] L. Wang, B. Young, Design of cold-formed steel channels with stiffened webs sub-
jected to bending, Thin-Walled Struct. 85 (2014) 81–92.
[19] C.H. Pham, A.F. Davis, B.R. Emmett, Numerical investigation of cold-formed lapped Z
purlins under combined bending and shear, J. Constr. Steel Res. 95 (2014) 116–125.
[20] A. Bagheri Sabbagh, M. Petkovski, K. Pilakoutas, R. Mirghaderi, Experimental work
on cold-formed steel elements for earthquake resilient moment frame buildings,
Eng. Struct. 42 (2012) 371–386.
[21] D. Padilla-Llano, C.D. Moen, M.R. Eatherton, Cyclic axial response and energy dissi-
pation of cold-formed steel framing members, Thin-Walled Struct. 78 (2014)
95–107.
[22] H. Wang, Y. Zhang, Experimental and numerical investigation on cold-formed steel
C-section flexural members, J. Constr. Steel Res. 65 (2009) 1225–1235.
[23] A. Bagheri Sabbagh, M. Petkovski, K. Pilakoutas, R. Mirghaderi, Cyclic behaviour of
bolted cold-formed steel moment connections: FE modelling including slip, J.
Constr. Steel Res. 80 (2013) 100–108.
[24] S.R. Mirghaderi, S. Torabian, F. Keshavarzi, I-beam to box — column connection by a
vertical plate passing through the column, Eng. Struct. 32 (2010) 2034–2048.
[25] M.F. Wong, K.F. Chung, Structural behavior of bolted moment connections in cold-
formed steel beam–column sub-frames, J. Constr. Steel Res. 58 (2002) 253–274.
[26] J.B.P. Lim, D. a Nethercot, Ultimate strength of bolted moment-connections between
cold-formed steel members, Thin-Walled Struct. 41 (2003) 1019–1039.
[27] K.F. Chung, L. Lau, Experimental investigation on bolted moment connections
among cold formed steel members, Eng. Struct. 21 (1999) 898–911.
[28] S.M. Zahurul-Islam, B. Young, FRP strengthening of lean duplex stainless steel hol-
low sections subjected to web crippling, Thin-Walled Struct. 85 (2014) 183–200.
[29] E.M. Hassan, M.H. Serror, S.A. Mourad, Behavior of cold-formed steel in moment-
resisting frames, M.Sc. thesis at department of structural engineering, Faculty of en-
gineering, Cairo University, 2016.
[30] A. Bagheri Sabbagh, M. Petkovski, K. Pilakoutas, R. Mirghaderi, Ductile moment-
resisting frames using cold-formed steel sections: An analytical investigation, J.
Constr. Steel Res. 67 (2011) 634–646.
[31] Swanson analysis system, ANSYS release 11.0. Houston, USA, 1998.