Arab J Sci Eng (2017) 42:4157–4168

DOI 10.1007/s13369-017-2571-6

RESEARCH ARTICLE - CIVIL ENGINEERING

Evaluation on Effectiveness of Cold-Formed Steel Column

with Various Types of Edge Stiffener
P. Manikandan1 · G. Aruna1 · S. Balaji2 · S. Sukumar3 · M. Sivakumar4

Received: 7 July 2016 / Accepted: 8 May 2017 / Published online: 31 May 2017
© King Fahd University of Petroleum & Minerals 2017

Abstract In this study, four series of cold formed steel chan- Keywords Cold-formed steel · Distortional buckling · Edge
nel columns with different types of edge stiffeners is analyzed stiffener · Flexural buckling · Slenderness ratio
under pinned end conditions subjected to axial compres-
sion. The selected cross-sectional profiles are met with the
interactive failure modes. The cross-sectional dimensions are List of symbols
satisfied the pre-qualified section properties of cold-formed
steel structures. Test results are simulated with finite element A Cross-sectional area
analysis software using ANSYS, and good agreement has Pcre Critical elastic column buckling load in flexural
been achieved. After the validation of the proposed numer- buckling/flexural–torsional/torsional
ical model, the effect of variation of thickness, yield stress Pcrl Critical elastic local column buckling
and overall slenderness in the strength and behaviour of the Pcrd Critical elastic distortional column buckling
section are carried out. Interaction of distortional and flex- load
ural buckling is the governing failure modes of the column d Depth of the channel section
section. All the findings of FEA are compared with the Direct Pd, DSM Design strength calculated from the direct
Strength Specification for cold-formed steel structures, and strength method
the strength of the member is also determined by using the PEXP Experimental ultimate load
modified equations from the literature. Based on the results, rmin Least radius of gyration
a new design equation is proposed, and also it is validated by Pn, DSM Nominal strength calculated from the direct
the results available from the literature. strength method
Pn, ANB Nominal strength calculated from the design
rules in Anbarasu et al. [23]
B P. Manikandan Pnl Nominal strength for local buckling
lp_mani@[Link]
Pnd Nominal strength for distortional buckling
S. Balaji Pne Nominal strength for flexural–torsional buck-
[Link]@[Link]
ling
S. Sukumar Pn, Pro Proposed design rules for the direct strength
sukumar_237@[Link]
method
1 Centre for Structural Engineering research, Department d1 Size of the edge stiffener
of Civil Engineering, Sona College of Technology, Salem, d2 Size of the return lip
Tamil Nadu, India t Thickness of the channel section
2 Kongu Engineering College, Erode Dt, Tamil Nadu, India PANSYS Ultimate load calculated from the FE analysis
3 Department of Civil Engineering, Paavai Group of Institution, fy Yield stress of the material
Namakkal Dt, Tamil Nadu, India E Young’s modulus
4 Vishwa Infrastructures and Services Pvt. Ltd., Ahmadabad, fU Ultimate stress of the material
India λ Overall slenderness ratio of the column

123
4158 Arab J Sci Eng (2017) 42:4157–4168

1 Introduction plain lipped channel columns. The effective width concept

is tedious for complex cross section. Hence, a simple new
Applications of cold-formed steel structures are increasing design methodology (direct strength method) was developed
all over the world in building construction and automobile by Schafer and Peköz [15]. Schafer [16,17] also gave the
industry. The open cold-formed steel sections are failed by guidelines for geometric limits and validity of the developed
either local, distortional or Euler buckling. However, the formula for cross sections, including simple lipped chan-
interactive failure is more complex for designers. A simple nel with and without web stiffeners, Z section, hat section,
lipped channel column with edge stiffener is more familiar and rack upright section. Continuation of this research, Yan
with the construction industry. From the literature Wang et al. and Young [18] conducted a study on the behaviour of cold-
[1], it is found that the edge stiffeners can reduce effectively formed steel columns with complex edge stiffener and also
the rotational capacity of the flanges and improve the capac- evaluated the effectiveness of the DSM specifications.
ity of members. Though many studies are pertaining simple In this study, a series of cold-formed steel channel column
lipped channel column with upright and complex edge stiff- with different types of edge stiffeners is analysed under axial
ener, the results of channels with complex edge stiffener are compression. The selected cross-sectional profile is met with
scattered. Hence, in this study, the effect of cross-sectional the interactive failure modes. Test results are then simulated
dimension, variation of yield stress and overall slenderness by finite element analysis using ANSYS software package,
ratios in strength and behaviour of channel columns with and good conformity has been achieved. Following the val-
complex edge stiffener are discussed. idation, effects of variation of thickness, overall slenderness
A brief review of literature on the strength and behaviour and yield stress are discussed. All the findings from FEA are
of cold-formed steel channel column with different types compared with the Direct Strength Specification (DSM) for
of edge stiffeners which is related to this research work is cold-formed steel structures. A suitable design recommen-
presented here. A lot of experimental investigations have dation is stated to DSM for the cold-formed steel channel
been undertaken on the behaviour of lipped channel column column with complex edge stiffener. The efficiency of the
with simple edge stiffeners by Batista [2]. They devel- proposed model is verified against the test results available
oped the buckling design curve for stiffened and unstiffened in the literature [18].
long-channel column based on the Brazilian and European
specification for steel construction. Young and Rasmussen
[3] were taking up the study of the design lipped chan- 2 Experimental Program
nel column under fixed-end condition. Young and Hancock
[4] conducted a experimental study on the behaviour of 2.1 Test Program
cold-formed steel channel section with inward and outward
inclined edge stiffeners under axial compression. The similar A total of six specimens with four types of sections are tested
study was numerically verified by Young [5]. It was observed under axial compressive loading, including a simple channel
that North American Specifications provide unconservative column (Fig. 1a, called SC section), a simple channel column
results for the section with outward stiffener. To evaluate with inclined edge stiffener (Fig. 1b, called SC-I section), a
more accurate results for the local buckling strength of stub simple channel column with upright edge stiffener (Fig. 1c,
lipped channel column, a new design equation was suggested called SC-U section) and simple channel column with com-
by Anil Kumar and Kalyanaraman [6]. Test programmes on plex edge stiffener (Fig. 1d, called SC-C section).
plain channel subjected to local and coupled local–flexural
instabilities were discussed by Ungermann et al. [7]. Very 2.2 Test Specimens
recently, Ananthi et al. [8] did an experimental and numerical
investigation of strength and behaviour of unlipped channel A selected cross-sectional profile with defined nomenclature
columns with fixed ends. The North American Specifica- is presented in Fig. 1, and details of section geometries are
tions for cold-formed steel structures were provided slightly presented in Table 1. The geometrical sizes below are defined
unconservative results; however, for predicting the accurate by surface outline. The entire test specimens are having same
results, a modification factor was proposed. cross-sectional area with the nominal height of the web (d)
As per the DSM specifications, the design strength of is 130 mm, the overall nominal width of the flange (B) is 85
the column is the least of flexural, local and distortional mm, size of the lip (d1) is 15 mm and size of the rear lip (d2)
buckling. To improve the design guidelines of DSM spec- is 15 mm. The nominal width of the flange (b) for sections
ification for interactive failure modes, Young and Yan [9], SC-I and SC-U is 70 mm, whereas the nominal width of the
Becque and Rasmussen [10], Silvestre et al. [11], Dinis et flange (b) for section SC-C is 55 mm. In these tests, three
al. [12], Silvestre et al. [13], Loughlan et al. [14] discussed overall slenderness ratios (λ) are selected. The first column
the local/distortional and flexural–buckling interactions of of Table 1 is the specimen identifications, which is defined

123
Arab J Sci Eng (2017) 42:4157–4168 4159

Fig. 1 Cross-sectional
geometries

Table 1 Details of section geometries

Specimen ID Section dimension (mm) λ = L/r min
d b1 d1 d2 t

SC-λ-F270 130 85 – – 1.6 50

SC-I-λ-F270 130 70 15 – 1.6 50
SC-U-λ-F270 130 70 15 – 1.6 30
SC-U-λ-F270 130 70 15 – 1.6 40
SC-U-λ-F270 130 70 15 – 1.6 50
SC-C1-λ-F270 130 55 15 15 1.6 50
SC-C2-λ-F270 50 50 16 14 1.6 20–120
SC-C2-λ-F350 50 50 16 14 1.6 20–120
SC-C2-λ-F550 50 50 16 14 1.6 20–120
SC-C3-λ-F270 58 42 13 17 1.6 20–120 Fig. 3 Sample of the coupons

a sharp yield, the yield stress was determined by the 0.2%

offset method. Material properties are listed in Table 2.

2.4 Test Set-up

The 400-kN capacity hydraulic ram is used for applying

the axial compressive load. For applying uniform load and
support conditions over the entire cross section, the 8-mm
thickness end plates are milled flat and welded on both ends
of the specimens. All the columns are tested under pin-ended
boundary condition. At each end of the column, rubber gas-
kets are positioned to facilitate the pinned boundary condition
at the supports [19,20]. A specimen is tested with a self-
Fig. 2 Stress–strain curve supported loading frame, and loads are applied by using the
calibrated hydraulic ram. The ultimate load is measured with
the help of load cell, and LVDT is used to measure the axial
as follows: SC, SC-I, SC-U and SC-C referred to the four shortening of the cross section. The test set-up is displayed
types of cross sections, respectively; Letter λ and the number in Fig. 4.
followed by referring to the slenderness of the specimen; the
last letter F and the number followed by referring to the yield
stress of the material. 3 Finite Element Modelling

2.3 Material Properties In order to examine the strength and structural behaviour of
the proposed column section over the different slenderness
The tensile coupon tests are carried out as per IS 1608-2005 ranges, a finite element model is established by using a com-
(Part-1). Stress–strain curve and coupons are shown in Figs. 2 mercial finite element software package ANSYS. The FE
and 3 respectively. Because the material tests did not display model is based on the centre line dimensions of the cross

123
4160 Arab J Sci Eng (2017) 42:4157–4168

Table 2 Details of material

S. no. Specimen ID E (MPa) f y (MPa) f U (MPa) % Elongation
properties
1 SC-λ50- F270 2.01 268 375 19
2 SC-I-λ50- F270 2.20 269 374 20
3 SC-U-λ30- F270 2.06 271 393 16
4 SC-U-λ40- F270 2.05 277 416 25
5 SC-U-λ50- F270 2.20 274 384 21
6 SC-C1-λ50- F270 2.11 269 385 24

section, and the highly efficient four-node shell element 181

is employed to generate the FE model. Through the conver-
gence study, a uniform mesh size of 10 × 10 mm is used in
the model (aspect ratios nearly equal to 1.0), and appropriate
mesh size of 100 mm2 is chosen for this study.
The influence of rounded corners with internal radius
r ≤ 5t and r ≤ 0.15w (r ≤ 5 × 1.6 = 8 mm, 0.15 × 85 =
12.75 mm) on section properties can be neglected based on
ENV (2006). All the specimens are fabricated using the lat-
est computerized cutting and sharp folding press-braking
machine. The sharp cutting edges provide right angular edge
only. The residual stresses are not measured because both the
flexural and membrane residual stresses tend to be insignifi-
cant compared with the yield stress in specimens made using
the press-braking process [18]. Based on that, the effect of
residual stress and cold-forming process has been ignored in
the FE model.
The end conditions of the columns are pin ended [at loaded
end Ux = Uy = Rz = 0 and unloaded end Ux = Uy = Uz
= Rz = 0, [21]]. To ensure the axial loading condition, all
Fig. 4 Test set-up the support reaction and load are applied at the CG of the

Fig. 5 Details of finite element

model

123
Arab J Sci Eng (2017) 42:4157–4168 4161

Fig. 6 Comparison of failure modes a SC-λ50- F270, b SC-I-λ50- F270

Table 3 Comparison of test and

S. no. Specimen ID Ultimate load (kN) PEXP /PANSYS Mode of failures
ANSYS results
PEXP PANSYS

1 SC SC-λ50-F270 24.32 25.63 0.95 L

2 SC-I SC-I-λ50-F270 61.90 71.16 0.87 L+D
3 SC-U1 SC-U-λ30-F270 70.60 72.30 0.98 L
4 SC-U2 SC-U-λ40-F270 67.80 71.37 0.95 L
5 SC-U3 SC-U-λ50-F270 64.70 71.36 0.91 L+D
6 SC-C SC-C1-λ50-F270 70.67 72.49 0.97 L+FB
Mean 0.94
SD 0.04
L local buckling, D distortional buckling, FT flexural–torsional buckling, FB flexural buckling

section. The CG of the section is formed as master node modes from the eigenvalue buckling analysis, is applied in
(MN) of the FE model at the loaded (MN1) and unloaded the nonlinear analysis model. Finally, the nonlinear analysis
(MN2) as shown in Fig. 5. For the uniform distribution of is performed to obtain the ultimate load and failure modes
load over the entire cross section, a rigid region is formed via of the sections by incorporating both geometric and material
master node at the loaded (MN1) and unloaded end (MN2) imperfections.
of the model as shown in Fig. 5. The preliminary geometrical
imperfection shape plays a critical function in the nonlinear
analysis of cold-formed steel column, as its preference may 4 Results and Discussion
adjust noticeably the corresponding buckling behaviour and
ultimate strength. For an example, typical failure modes of the columns are
First, the eigenvalue buckling analysis is performed using shown in Fig. 6. Similar results are obtained for all the
the geometry of the perfect member, to identify the possible columns, and results are tabulated in Table 3. All the columns
buckling modes. Following this, nonlinear analysis is per- are having an equal cross-sectional area. There are three dif-
formed on the same model, considering both the material ferent kinds of failure modes observed, i.e. local, interactive
and geometric nonlinearities. The material behaviour of the local and distortional buckling and interactive local and over-
section is described by a bilinear stress–strain curve. The all buckling. Columns SC, SC-U1 and SC-U2 sections fail
geometric imperfections are included in the FEM by using by local buckling.
a eigenvalue buckling analysis. The imperfections shape Local and distortional interactive buckling is observed for
corresponding to the first buckling mode is applied using the SC-I and SC-U3 section. The distortional deformation
the imperfection option in ANSYS, with a scaling factor mostly occurred at the mid-height of the columns. For SC sec-
of maximum amplitude of the measured imperfection in tion fails by local buckling, i.e. several buckling half waves
case of Schafer and Pekoz [22]. The maximum amplitude appeared along the length of the specimen (Fig. 6a). For SC-
of imperfections of 1 time the thickness for the local and I (Fig. 6b) and SC-U3 sections, there is inward distortional
distortional buckling and L/1000 for the overall buckling buckling occurred at the mid-length of the specimen. For

123
4162 Arab J Sci Eng (2017) 42:4157–4168

Fig. 7 Load–deflection curve a SC, b SC-I, c SC-U3, d SC-C

SC-C section fails by interaction between local and over- 5 Parametric Studies
all buckling, the local buckling half wave is observed on
the flanges and maximum deformation appeared at the mid- Totally, the 56 models are analysed using the developed FE
height of the column, with the presence of complex stiffener model. In this parametric study, the effect of variations of
on the flanges. thickness, yield stress and overall slenderness are analysed,
The ultimate strength of the columns SC, SC-I, SC-U3 and and results are discussed in the subsequent sections.
SC-C is 24.32, 61.90, 64.70 and 70.67 kN respectively. It is
indicated that edge stiffeners have increased the torsional 5.1 Effect of Thickness Variation
rigidity of the section and improve the structural perfor-
mances from local buckling to flexural buckling. The strength In this parametric study, all the columns have a slender-
of the section can be enhanced by adding an edge stiffener ness ratio is 50 and with a yield stresses of 270 N/mm2 . The
to the flange in the section SC. However, the optimized, effect of variation of thickness and the comparison of results
simple channel column with complex edge stiffener (SC-C) between ultimate load obtained from FE analysis (PANSYS )
shows the ultimate capacity is more than 66% when com- and direct strength method (Pn, DSM ) are presented in Table 4.
pared to simple channel column (SC). Typical load versus In this study, three plate thicknesses (1.6, 2 and 3 mm) are
axial displacement curves are illustrated in Fig. 7. Compar- considered. From this parametric study, it is observed that the
isons of results between experiment (PEXP ) and FE analysis thickness of the section has a significant role in the strength
(PANSYS ) are presented in Table 3. The average and standard of the section. Rapidly, load carrying capacity of the section
deviation of PEXP /PANSYS are 0.94 and 0.04 respectively. increases with increases in the thickness of the section.
The results indicate that the proposed FE model predicts the
ultimate strength of the column with a high degree of pre- 5.2 Effect of Overall Slenderness Variation
cision and reliability. From the results, it is concluded that
the developed FE model is more suitable for the parametric From the test results, it is observed that the simple chan-
analysis. nel column with complex edge stiffener (SC-C) provides a
better performance compared to all other sections. Hence,

123
Arab J Sci Eng (2017) 42:4157–4168 4163

Table 4 Effect of thickness variation

S. no. Specimen ID Section dimension (mm) Ultimate load (kN) PANSYS /Pn, DSM
d b1 d1 d2 t PANSYS Pn, DSM

1 SC-λ50-F270-T1.6 130 85 – – 1.6 25.63 30.02 0.85

2 SC-λ50-F270-T2.0 130 85 – – 2 48.83 49.30 0.99
3 SC-λ50-F270-T3.0 130 85 – – 3 107.2 169.01 0.63
4 SC-I-λ50-F270-T1.6 130 70 15 – 1.6 71.16 78.07 0.91
5 SC-I-λ50-F270-T2.0 130 70 15 – 2 92.15 112.68 0.82
6 SC-I-λ50-F270-T3.0 130 70 15 – 3 163.46 199.26 0.82
7 SC-U-λ50-F270-T1.6 130 70 15 – 1.6 71.36 90.92 0.78
8 SC-U-λ50-F270-T2.0 130 70 15 – 2 107.74 125.1 0.86
9 SC-U-λ50-F270-T3.0 130 70 15 – 3 175.77 207.2 0.85
10 SC-C1-λ50-F270-T1.6 130 55 15 15 1.6 72.49 94.35 0.77
11 SC-C1-λ50-F270-T2.0 130 55 15 15 2 99.75 138.93 0.72
12 SC-C1-λ50-F270-T3.0 130 55 15 15 3 155.78 204.21 0.76
Mean 0.81
SD 0.09

Fig. 8 Failure
modes—SC-C2-λ-F270 series

a detailed parametric study is carried out on the behaviour

of simple channel columns with complex edge stiffener. The
dimensions of the selected cross section are met with the
square (SC-C2) and rectangular (SC-C3) shapes. Two sets of
cross-sectional dimensions with three yield stress of 270, 350
and 550 N/mm2 and 12 overall slenderness variations gave
44 cross sections. The section geometry and dimensions are
shown in Fig. 1d and Table 1 respectively. For each cross
section, slenderness ratio (λ) varies from 20 to 120 with an
increment of 10. All the specimens are failed by interactive
failure mode.
As an illustration, the deformed shape at failure load and
effect of yield stress variation of SC-C2-20 series of column Fig. 9 Load–deflection curve for yield stress variation of SC-C2-20
series
are shown in Figs. 8 and 9 respectively. Similarly, Fig. 10
illustrates the load versus axial deformation curves for SC-C2
series of columns. For the SC-C2 column series, slender-
ness ratio is less than 30, the specimens failed by combining mode. The slenderness ratio is less than 50, all the specimens
local (L), distortional (D) and flexural (F) mode, whereas the failed by combining local (L) and distortional (D) buckling,
slenderness ratio is greater than 30, the specimens failed by whereas the slenderness ratio is greater than 50, the speci-
combining distortional (D) and flexural (F) mode. However, mens failed by combining local (L) and flexural (F) mode.
the SC-C3 series specimens were failed by interactive failure From this parametric study, it is clearly observed that the
type of cross section, the magnitude of yield stress and over-

123
4164 Arab J Sci Eng (2017) 42:4157–4168

Fig. 10 Load–deflection curve for a SC-C2-λ -F270, b SC-C2-λ -F350, c SC-C2-λ -F550

Table 5 Comparison of results

S. no. Specimen ID Ultimate load (kN) PANSYS /Pn, DSM PANSYS /Pd, DSM PANSYS /Pn, ANB PANSYS /Pn, Pro Mode of failures
PANSYS Pn, DSM

1 SC-C2-20-F270 87.28 81.69 1.07 1.26 1.28 1.10 L+D+FT

2 SC-C2-30-F270 68.45 78.31 0.87 1.03 1.05 0.90 L+D+FT
3 SC-C2-40-F270 65.92 73.86 0.89 1.05 1.07 0.92 D+FT
4 SC-C2-50-F270 60.18 68.56 0.88 1.03 1.05 0.91 D+FT
5 SC-C2-60-F270 53.05 62.34 0.85 1.00 1.02 0.88 D+FT
6 SC-C2-70-F270 45.73 55.78 0.82 0.96 0.98 0.85 D+FT
7 SC-C2-80-F270 39.8 50.14 0.79 0.93 0.95 0.82 D+FT
8 SC-C2-90-F270 34.46 44.25 0.78 0.92 0.93 0.80 D+FT
9 SC-C2-100-F270 29.64 38.00 0.78 0.92 0.93 0.81 D+FT
10 SC-C2-110-F270 25.49 32.14 0.79 0.93 0.95 0.82 D+FT
11 SC-C2-120-F270 22.09 27.15 0.81 0.96 0.97 0.84 D+FT
12 SC-C2-20-F350 105.86 106.56 0.99 1.17 1.19 1.03 L+D+FT
13 SC-C2-30-F350 101.61 101.88 1.00 1.17 1.19 1.03 L+D+FT
14 SC-C2-40-F350 99.16 93.94 1.06 1.24 1.26 1.09 D+FT
15 SC-C2-50-F350 94.00 85.59 1.10 1.29 1.31 1.13 D+FT
16 SC-C2-60-F350 83.06 76.03 1.09 1.29 1.31 1.13 D+FT
17 SC-C2-70-F350 68.78 66.17 1.04 1.22 1.24 1.07 D+FT
18 SC-C2-80-F350 53.29 57.94 0.92 1.08 1.10 0.95 D+FT
19 SC-C2-90-F350 47.52 49.63 0.96 1.13 1.15 0.99 D+FT
20 SC-C2-100- F350 39.00 41.00 0.95 1.12 1.14 0.98 D+FT
21 SC-C2-110-F350 32.45 33.98 0.95 1.12 1.14 0.99 D+FT
22 SC-C2-120-F350 27.381 28.67 0.96 1.12 1.14 0.99 D+FT

123
Arab J Sci Eng (2017) 42:4157–4168 4165

Table 5 continued
S. no. Specimen ID Ultimate load (kN) PANSYS /Pn, DSM PANSYS /Pd, DSM PANSYS /Pn, ANB PANSYS /Pn, Pro Mode of failures
PANSYS Pn, DSM

23 SC-C2-20-F550 154.35 165.64 0.93 1.10 1.11 0.96 L+FT

24 SC-C2-30-F550 150.97 151.55 1.00 1.17 1.19 1.03 D+FT
25 SC-C2-40-F550 137.06 133.97 1.02 1.20 1.22 1.06 D+FT
26 SC-C2-50-F550 115.00 114.53 1.00 1.18 1.20 1.04 D+FT
27 SC-C2-60-F550 89.56 93.77 0.96 1.12 1.14 0.99 D+FT
28 SC-C2-70-F550 69.49 74.19 0.94 1.10 1.12 0.97 D+FT
29 SC-C2-80-F550 56.56 59.41 0.95 1.12 1.14 0.98 D+FT
30 SC-C2-90-F550 46.2 47.94 0.96 1.13 1.15 1.00 D+FT
31 SC-C2-100-F550 36.23 38.81 0.93 1.10 1.12 0.96 D+FT
32 SC-C2-110-F550 31.15 32.14 0.97 1.14 1.16 1.00 D+FT
33 SC-C2-120-F550 26.22 27.15 0.97 1.14 1.16 1.00 D+FT
34 SC-C3-20-F270 88.91 78.28 1.14 1.34 1.36 1.17 L+DB
35 SC-C3-30-F270 68.36 76.24 0.90 1.05 1.07 0.93 L+DB
36 SC-C3-40-F270 65.70 73.26 0.90 1.06 1.07 0.93 L+DB
37 SC-C3-50-F270 64.48 69.88 0.92 1.09 1.10 0.95 L+DB
38 SC-C3-60-F270 58.07 66.01 0.88 1.03 1.05 0.91 L+FB
39 SC-C3-70-F270 52.85 61.52 0.86 1.01 1.03 0.89 L+FB
40 SC-C3-80-F270 53.94 57.33 0.94 1.11 1.13 0.97 L+FB
41 SC-C3-90-F270 48.00 53.02 0.91 1.07 1.08 0.94 L+FTB
42 SC-C3-100-F270 46.67 47.86 0.98 1.15 1.17 1.01 L+FTB
43 SC-C3-110-F270 44.51 43.10 1.03 1.21 1.24 1.07 L+FTB
44 SC-C3-120-F270 42.35 38.74 1.09 1.29 1.31 1.13 L+FTB
Mean 0.94 1.11 1.13 0.97
SD 0.13 0.10 0.11 0.09
L local buckling, D distortional buckling, FT flexural–torsional buckling, FB flexural buckling

√
all slenderness ratio is significantly affecting the strength and where λ1 = Pne /Pcrl .
behaviour of the column. The nominal axial strength (Pnd ) for distortional buckling
is
⎧
6 Theoretical Investigations ⎨ Py for λd ≤ 0.561
Pnd =  0.6  0.6
Pcrd Pcrd
⎩ 1 − 0.25 Py Py Py for λd > 0.561
Based on the direct strength method (DSM), the nominal
capacity of members in axial compression (Pn, DSM ) shall be (3)
minimum of local buckling (Pnl ), distortional buckling (Pnd )
and flexural–torsional buckling (Pne ). 
where λd = Py /Pcrd .
 The critical elastic local load Pcrl and distortional buck-
(0.658) λc
2
for λc ≤ 1.5
Pne =   Py (1) ling load Pcrd are computed using linear elastic finite strip
0.877
λc2
Py for λc > 1.5 buckling analysis. Three dissimilar approaches are used to
estimate the ultimate strength of the simple channel column

where λc = Py /Pcre and Py = A f y .Py is the squash load. with complex edge stiffener. In the first DSM approach, the
The nominal axial strength (Pnl ) for local buckling is nominal ultimate strength of the section (Pn, DSM ) is equal
⎧ to the least of (Pne, Pnl, Pnd ). In the second DSM approach,
⎨ Pne for λ1 ≤ 0.776 the design ultimate strength of the section (Pd, DSM ) is equal
Pnl =  0.4  
to the 0.85 times the least of (Pne, Pnl, Pnd ). In the third
⎩ 1 − 0.15 PPcrlne
Pcrl
Pne Pne for λ1 > 0.776
approach, nominal resistance of the section is estimated by
(2) using a modification factor proposed by Pn, ANB [23], which

123
4166 Arab J Sci Eng (2017) 42:4157–4168

Fig. 11 Comparison of results a SC-C2-λ -F270 b SC-C2-λ -F350 c SC-C2-λ -F550 d SC-C3-λ -F270

is equal to the 0.836 times the least of (Pne, Pnl, Pnd ). How-
ever, to obtain exact resistance of the simple channel column
with complex edge stiffener, a new design expression is pro-
posed (Pn,Pro ). The comparison of results between the FEA
and various design standards is discussed in Table 5 and
Fig. 11.
Except for the SC-C3-λ-F550 series of column, the
direct strength method (Pn, DSM ) is producing unconservative
results. The strength obtained from the Pd, DSM and Pn, ANB
approach predicts unconservative results for the SC-C2-λ-
F350 and SC-C2-λ-F550 series of column. The mean and
standard deviation of PANSYS /Pn, DSM are 0.94 and 0.13, the
mean and standard deviation of PANSYS /Pd, DSM are 1.11 and
0.10 and the mean and standard deviation of PANSYS /Pn, ANB
are 1.13 and 0.11, respectively. Fig. 12 Relationship between ANSYS and DSM results
However to predict the accurate results, a new design
modification factor is proposed in this study. A linear regres-
PEXP /Pn, Pro are 1.01 and 0.02, respectively. From the test
sion analysis is conducted between PANSYS and Pn, DSM as
results, it is observed that the proposed design equation mod-
shown in Fig. 12. The relationship between the ultimate
erately predicts the strength of the simple channel column
capacity predicted by finite element analysis (PANSYS ) and
with complex edge stiffener.
the direct strength method (Pd, DSM ) is almost linear as
Pn, pro = 0.968* Pn, DSM . The mean and standard devia-
tion of PANSYS /Pn, pro are 0.97 and 0.09, respectively. The
proposed design equation is also verified by the results avail- 7 Summary and Conclusions
able from the literature (18), and the results are displayed in
Table 6. The corresponding mean and standard deviation of An experimental and numerical investigation of the structural
behaviour of pinned-ended simple cold-formed steel chan-

123
Arab J Sci Eng (2017) 42:4157–4168 4167

Table 6 Comparison of
Specimen ID as per literature Load (kN) PEXP /Pn, Pro
proposed model with test results
from literature [18] PEXP PDSM Pn, Pro

T1.5F120L0500 168.90 164.70 163.60 1.03

T1.5F120L3000 131.30 133.30 131.61 1.00
T1.5F120L3500 127.40 123.20 121.31 1.05
T1.9F120L0500 233.70 234.2 234.42 1.00
T1.9F120L1000 231.20 230.8 230.96 1.00
T1.9F120L1500 227.30 225.1 225.15 1.01
Mean 1.01
SD 0.02

nel column with various types of edge stiffener is discussed. derness ranges of the column with grade of steel that is
Totally, six columns are tested experimentally, and the results equal to 550 N/mm2 .
are verified by using developed finite element modelling, and • The proposed design approach (Pn, pro ) estimates safe and
good correlation is achieved. The effect of variations of thick- reliable results in all the aspects of the column with the
ness, yield stress and overall slenderness are analysed by mean and standard deviation of 0.97 and 0.09, respec-
using the developed FE model. The strength of the sections tively.
obtained from the finite element analysis is compared with
the direct strength method specifications for the cold-formed
steel structures. The subsequent conclusions are drawn from References
these investigations.
1. Wang, C.; Zhang, Z.; Zhao, D.; Liu, Q.: Compression tests and
numerical analysis of web-stiffened channels with complex edge
• All the slenderness range of the column, global buckling stiffeners. J. Constr. Steel Res. 116, 29–39 (2016)
2. Batista, E.M.: Buckling curve for cold-formed compressed mem-
is observed in premature stage of analysing and followed bers. J. Constr. Steel Res. 28, 121–136 (1994)
by the local and distortional buckling is observed in the 3. Young, B.; Rasmussen, J.R.: Design of lipped channel columns. J.
later stage of testing. Struct. Eng. 124, 140–148 (1998)
• The current DSM (Pn, DSM ) approaches overestimated 4. Young, B.; Hancock, G.J.: Compression tests of channels with
inclined simple edge stiffeners. J. Struct. Eng. 129, 1403–1411
the strength of the section. It confirms the inadequacy of
(2003)
the current DSM specifications in predicting the ultimate 5. Young, B.: Design of channel columns with inclined edge stiffen-
strength of the column experiencing in overall buckling ers. J. Constr. Steel Res. 60, 183–197 (2004)
interaction. 6. Anil Kumar, M.V.; Kalyanaraman, V.: Design strength of locally
buckling stub-lipped channel columns. J. Struct. Eng. 138, 1291–
• In addition, it is found that the Pn, DSM and Pd, DSM
1299 (2012)
approaches provide overestimated results, whereas the 7. Ungermann, D.; Lubke, S.; Brune, B.: Tests and design approach
Pn, ANB and Pn, pro approaches predict the under- for plain channels in local and coupled local-flexural buckling
anticipated results in the column with grade of steel that based on eurocode3. Thin-Walled Struct. 81, 108–120 (2014)
8. Ananthi, G.B.; Palani, G.S.; Iyer, N.R.: Numerical and theoretical
is equal to 270 N/mm2 . However, Pn, ANB and Pn, pro
studies on cold-formed steel unlipped channels subjected to axial
approaches estimate to some extent unsafe and exhibit compression. Lat. Am. J Solids Struct. 12, 1–17 (2015)
the speckled results. 9. Young, B.; Yan, J.: Design of cold-formed steel channel columns
• Similarly, Pn, ANB and Pd, DSM approaches provide under- with complex edge stiffeners by direct strength method. J. Struct.
Eng. 130, 1756–1763 (2002)
estimated results for all the slenderness ranges of the 10. Becque, J.; Rasmussen, K.J.R.: Experimental investigation of
column, whereas Pn, DSM and Pn, pro approaches estimate local-overall interaction buckling of stainless steel lipped channel
slightly unsafe for the intermediate overall slenderness columns. J. Constr. Steel Res. 65, 1677–1684 (2009)
(λ) range of the column (30–90) with grade of steel that 11. Silvestre, N.; Camotim, D.; Dinis, P.B.: Direct strength prediction
of lipped channel columns experiencing local-plate/distortional
is equal to 350 N/mm2 . However, the proposed design interaction. Adv. Steel Const. 5, 49–71 (2009)
approach (Pn, pro ) estimates safe and reliable results in 12. Dinis, P.B.; Camotim, D.; Batista, E.M.; Santos, E.:
all slenderness ranges of the column. Local/distortional/global mode coupling in fixed lipped channel
• It is also observed that Pn, ANB and Pd, DSM approaches columns: behaviour and strength. Adv. Steel Const. 7, 113–130
(2011)
provide underestimated results for all the slenderness 13. Silvestre, N.; Camotim, D.; Dinis, P.B.: Post-buckling behaviour
ranges of the column, whereas Pn, DSM , and Pn, pro and direct strength design of lied channel columns experiencing
approaches estimate safe and reliable results in all slen- local/distortional interaction. J. Constr. Steel Res. 73, 12–30 (2012)

123
4168 Arab J Sci Eng (2017) 42:4157–4168

14. Loughlan, J.; Yidris, N.; Jones, K.: The failure of thin-walled lied 20. Manikandan, P.; Arun, N.: Behaviour of partially closed stiffened
channel compression members due to coupled local-distortional cold-formed steel compression member. Arab. J. Sci. Eng. 41,
interactions and material yielding. Thin-Walled Struct. 6, 14–21 3865–3875 (2016)
(2012) 21. Salem, A.H.; El Aghoury, M.A.; Hassan, S.K.; Amin, A.A.: Post-
15. Schafer, B.W.; Peköz, T.: Direct strength prediction of cold formed buckling strength of battened columns built from cold-formed
steel members using numerical elastic buckling solutions. In: lipped channels. Emir. J. Eng. Resc. 9(2), 117–125 (2004)
Proceedings of 14th International Specialty Conference on Cold- 22. Schafer, B.; Pekoz, T.: Computational modelling of cold-formed
Formed Steel Structures, Univ. of Missouri-Rolla, Rolla, pp, 69–76 steel: characterizing geometrical imperfections and residual
(1998) stresses. J. Constr. Steel Res. 47, 193–210 (1998)
16. Schafer, B.W.: Local, distortional, and Euler buckling of thin- 23. Anbarasu, M.; Sukumar, S.: Local/distortional/global buckling
walled columns. J. Struct. Eng. 128, 289–299 (2002a) mode interaction on thin walled lied channel columns. Lat. Am.
17. Schafer, B.W.: Progress on the direct strength method. In: Proceed- J. Solids Struct. 11, 1363–1375 (2014)
ings of 16th International Specialty Conference on Cold-Formed 24. ENV 1993-1-4:2006/AC.: European pre-standard design of steel
Steel Structures, Orlando, pp. 647–662 (2002b) structures-part 1-3:General rules—supplementary rules for cold
18. Yan, J.; Young, B.: Column tests of cold-formed steel channels formed thin gauge members and sheeting
with complex stiffeners. J. Struct. Eng. 128, 737–745 (2002)
19. Sukumar, S.; Parameswaran, P.; Jayagopal, L.S.: Local distortional
and Euler-buckling of thin walled built-up open sections under
compression. J. Struct. Eng. SERC India 32, 54–447 (2006)

123