Proceedings of the

8th International Conference on

Coupled Instabilities in Metal Structures
Lodz University of Technology,Poland, July 12-14, 2021

Interaction of stiffened and unstiffened element buckling modes in CFS

plain channel compression members
K. C. Kalam Aswathy1 and M. V. Anil Kumar2

Abstract
Plain channel compression members are composed of stiffened (web) and unstiffened (flanges) plate
elements. Effective width method (EWM) for the design of cold-formed steel (CFS) compression
members accounts for unstiffened and stiffened plate elements through plate buckling coefficient, k of
0.425 and 4.0, respectively in the elastic local buckling stress (fcrl) calculation. The more recent direct
strength method (DSM) recommends the local buckling strength equations for the design of plain
channel compression members. A lipped stub channel compression member undergoes interaction of
local buckling initiated by the stiffened elements and distortional buckling initiated by the partially
stiffened flange-lip assembly. Based on finite element analysis (FEA) results, it is demonstrated that such
an interaction between buckling modes initiated by the stiffened web and unstiffened flanges is possible
in plain channel compression members also. The behavior of plain channels can be represented as an
interaction of buckling of the unstiffened flange and stiffened web elements, which is equivalent to
local-distortional interaction in lip stiffened channels. The deformation plots are also provided to
understand these mode interactions. The present study focuses on stub columns and does not consider
the interaction of global buckling mode.

1. Introduction
CFS members are manufactured by press-breaking or cold-rolling thin steel sheets to the required cross-
sectional shapes at room temperature. Plate elements in CFS sections having one and both the
longitudinal edges supported are classified as unstiffened and stiffened elements, respectively (Fig. 1a).
The edge stiffened flanges are classified as stiffened or partially stiffened depending on the rigidity of
the edge stiffener (Fig. 1b). Stub CFS lip stiffened channel compression members may undergo failure
modes such as yielding of the material, local buckling of the plate elements, distortional buckling of the
cross-section in addition to interaction between these modes as shown in Fig. 1c. The out-of-plane
deformation of the plate elements due to the large plate width to thickness ratio (b/t or h/t) with no
translation of plate juncture is classified as local buckling. Buckling mode involving the rotation of the
flange-lip assembly about the flange-web juncture is classified as distortional buckling which occurs due
to insufficient rigidity of the edge stiffener to restrain the movement of the flange-lip juncture (Fig. 1c).
The results presented in this study are on plain channel compression members which may be considered
as a limiting case of lip-stiffened channels when lip depth (d) tends to zero.

1
PhD Scholar, Indian Institute of Technology Palakkad, Kerala, India-678623. Email:[email protected]
2
Assistant Professor, Indian Institute of Technology Palakkad, Kerala, India-678623. Email:[email protected]

Electronic copy available at: https://ssrn.com/abstract=3889448

 c) Local and distortional modes of buckling
c) Local and distortional
Figure 1: Types of platebuckling
elementsmodes
and buckling modes in CFS sections

The two commonly used design methods for CFS plain channel compression members are the effective
width method (EWM) and direct strength method (DSM) [AISI 2016; AS/NZS 2018]. EWM accounts for
the stiffened and unstiffened elements using plate buckling coefficients, k = 4.0 and 0.425, respectively
in the elastic local buckling stress, fcrl calculation (Eq. 1). The normalized ultimate strength (Pul/Py) of the
section is the summation of Pul/Py of individual plate elements of the cross-section which may be
computed using Eq. 2 where λl = √(fy/fcrl), is the nondimensional slenderness of the individual plate
elements. DSM has semi-empirical equations to compute the strength of the member corresponding to
different buckling modes based on the corresponding elastic buckling stresses or non dimensional
slenderness values of the whole member. Kumar and Kalyanaraman (2010) indicated that DSM local
buckling (referred to as DSM-LB here after) strength equation (Eq. 3) can be used for plain channel
compression members. NAS (2016) and AS/NZS (2018) recommends the DSM-LB equation for the
calculation of strength of plain channel compression members.

Electronic copy available at: https://ssrn.com/abstract=3889448

In DSM, although the distortional buckling equations are also recommended for the design of lipped
channels (Fig. 1b) with partially stiffened flanges (small values of d), the plain channels which can be
considered as limiting case of lipped channel having d = 0 (Fig. 1a) are designed using only local buckling
equations. Therefore there is some inconsistency in the current specifications (Schafer and Adany 2005;
Bambach 2009). Aswathy and Kumar (2020), demonstrated that both elastic buckling stress and ultimate
strength of unstiffened plates can be accurately represented using distortional buckling equations. The
modified DSM equation (denoted as MDSM-DB) proposed by Kumar and Kalyanaraman (2014) (Eq. 4)
based on test and FEA results of sections that are prone to pure distortional buckling (i.e., no local or
global buckling) can be extended for the calculation of unstiffened elements also. Here λd [= √(fy/fcrd)] is
the distortional buckling slenderness which is a function of the elastic distortional buckling stress, fcrd.
Hence the work presented in the paper is an attempt to represent the behavior of stub plain channel
compression member as local-distortional interaction and extend the interaction equation (IE)
developed for lipped channels to plain channels also.

Kumar and Kalyanaraman (2018) had proposed an interaction equation (Eq. 5) which can account for all
the buckling mode interactions. Here, Pul/Py, Pud/Py and Pue/Py are the normalized ultimate strengths in
local, distortional and global modes of buckling, respectively. The local buckling strength is calculated
using Eq. 6 (Kumar and Kalyanaraman 2012), which is a function of web height to flange width ratio, h/b
in addition to nondimensional local buckling slenderness, λl. The effect of h/b ratio is accounted for
using α1 and β and Pul,max/Py represents the upper bound of this equation. The distortional buckling
strength (Pud/Py) is computed using Eq. 4.

The elastic buckling and ultimate strength behavior of stiffened and unstiffened elements can be
accurately represented using local buckling equations (Bambach 2009) and distortional buckling
equations (Aswathy and Kumar, 2020), respectively. This work aims at understanding the interaction of
buckling modes initiated by stiffened and unstiffened elements in plain channels which may lead to the
plain channel behavior as a case equivalent to interaction of local and distortional buckling modes in
lipped channels. Finite element analysis (FEA) is used to understand the behavior of plain channels and
understand the influence of the mode interactions on the ultimate strength.

2. Finite Element Modeling

The plain channel members with fixed end boundary conditions were modeled in finite element
package, ABAQUS using four noded shell elements with reduced integration (S4R) through center line of
the plate elements. The load and boundary conditions were applied at the centroid of the cross-section
which was connected to the nodes along the end cross-section using stiff beam elements. Initially, an

Electronic copy available at: https://ssrn.com/abstract=3889448

eigenvalue buckling analysis was done to get the elastic buckling loads and corresponding mode shapes.
Non linear analysis using Newton Raphson iteration scheme was performed considering geometric and
material non linearity. Mode shapes from eigenvalue analysis was used to seed geometric imperfections
using appropriate scaling factors.

2.1 Validation of FE Model

The finite element (FE) model was validated by simulating a few test results available in literature
[Young and Rasmussen (1998) and Mulligan and Pekoz (1987)]. The dimensions of the specimens
modeled (b, h, t as shown in Fig. 1) and yield stress (fy) are given in Table 1. The ratio of the ultimate
loads from test to that from FEA (Pu-test/Pu-FEA) are also presented in Table-1. The mean, μ and standard
deviations,  of Pu-test/Pu-FEA indicate the accuracy of the FE model.

Table 1: Comparison of FEA and test results

Specimen b h t L fy λ Pu-test Pu-FEA Pu-test/Pu-FEA
(mm) (mm) (mm) (mm) (MPa) (kN) (kN) (kN)
SC 60x30a 41.48 78.38 1.219 253.49 226 1.30 32.93 31.19 1.06
SC 180x60a 76.30 221.28 1.219 634.49 226 2.99 37.83 36.77 1.03
P36F1000b 35.53 95.43 1.48 1000.2 550 1.73 59.00 64.72 0.91
P36F2500b 35.64 95.83 1.48 2499.4 550 1.86 32.80 34.39 0.95
Mean, µ 0.99
Standard deviation, σ 0.07
a – Young and Rasmussen (1998); b – Mulligan and Pekoz (1987)

3. Interaction of Buckling Modes in Plain Channels

The displacement plots from FEA are used in this section to demonstrate that the buckling of plain
channels may be initiated by buckling of either the unstiffened flanges or the stiffened web. Thus the
behavior of these members may be considered as a combination of the stiffened and unstiffened
element buckling modes. This section aims at understanding the buckling mode interactions in plain
channel compression members and its effect on the ultimate strength.

3.1 Mode interactions

To identify whether there is interaction between buckling modes initiated by stiffened and unstiffened
elements in plain channels, the out-of-plane deformation of the web along the length is plotted using
the results obtained from ABAQUS at load step corresponding to ultimate load. The plain channel
specimens with h= 100mm, b= 100mm, L= 250mm and fy= 250MPa are used for this plot. Figure 2a
presents the out-of-plane deformation normalized with respect to the maximum deformation (U/Umax)
of the web. When the buckling is initiated by the web (stiffened element), the buckling mode
corresponds to multiple half waves of wavelength approximately equal to the width of the web plate
and is denoted as stiffened element mode in Fig. 2a. Similarly, when it is initiated by the flanges
(unstiffened elements), the buckle half waves are much longer and denoted as unstiffened element
mode Fig. 2a.

Electronic copy available at: https://ssrn.com/abstract=3889448

 (b) (c)

Figure 2: Out-of-plane deformation of web (a) normalized displacement plot; displacement contours of web
when: (b) web does not buckle and (c) web buckles

Along with the normalized out of plane deformation (U/Umax) in the web plate corresponding to
stiffened (local) and unstiffened (distortional) buckling modes, web deformation in load step
corresponding to the ultimate load from FEA model of two stub plain channel members are also plotted
in Fig. 2. The h/t ratios of the web of these two specimens are chosen such that in one case the web
buckles [h/t > 1.901√(E/fy); 79.1 > 55.1] along with the flange and in the other case it does not [h/t <
1.901√(E/fy); 31.5 < 55.1]. When the web does not buckle, the deformation plot corresponds to the
unstiffened element mode. When the web buckles, it is observed that the deformation plot is a
combination of the stiffened and unstiffened modes (Fig. 2a). The corresponding contour plots from
ABAQUS for these cases are included in Fig. 2b and 2c, respectively. This confirms that there can be a
combination of stiffened and unstiffened modes of buckling in plain channels which is equivalent to a
local distortional interaction phenomena in lipped channels.

3.2 Ultimate strength

To understand the influence of the interaction of buckling modes on the ultimate strength, plain
channels with a constant non dimensional slenderness, λ [=√(fy/fcr); where fcr is the critical elastic
buckling stress obtained from eigenvalue analysis] and varying h/b ratios were analyzed to determine
the ultimate load. Stub plain channels having b= 100mm, L= 250mm, and h/b varying from 0.50 to 3.0
were modeled. The constant λ value was achieved by adjusting fy and t. The ultimate strength
normalized with yield load (Pu/Py) obtained from FEA and the theoretical predictions using EWM, DSM
and interaction equation (IE) (Eqs. 2, 3 and 5 represents as EWM, DSM-LB and IE, respectively) are also
plotted in Fig. 3. In IE Pud/Py and Pul/Py are computed using Eq. 4 and 6 respectively. In this work Pue/Py
=1, as stub members are used.

Electronic copy available at: https://ssrn.com/abstract=3889448

It is observed that there is a decrease in ultimate strength up to 20.5 % with increase in h/b ratio due to
interaction of stiffened and unstiffened buckling modes compared to the cases where there is no
interaction. The DSM-LB equation is found to give comparable predictions of the ultimate strength
whereas EWM gives conservative predictions. But since both the methods does not account for
interaction of buckling modes, the variation of strength with respect to h/b is not accurately
represented. The buckling interaction equation (IE) can trace this trend for plain channel compression
members and shows slightly better accuracy.

Figure 3: FEA results for plain channels a) λ = 2.0; b) λ = 3.0

The strength predictions using interaction equation (Eq. 5; denoted as IE), EWM (Eq. 2)and DSM-LB (Eq.
3) against 13 test results from literature [Young and Rasmussen (1998); Mulligan and Pekoz (1987) and
Talja (1990)] and 26 FEA results generated as part of this work are plotted in Fig. 4. The statistics of the
ratio of the ultimate strength from FEA or test and theoretical predictions (Pu-FEA,Test/Pu-Theory) are also
given in Table 2. The statistics of Pu-FEA,Test/Pu-Theory indicates that both IE and DSM-LB methods give
accurate strength predictions for plain channel compression members. There is a slight improvement in
standard deviation when the ultimate strength is calculated using IE. Thus indicates that plain channels
can also accurately be represented as a case of local distortional interaction phenomena and existing
interaction equations can be extended for plain channel compression members also.

Figure 4: Pu /Py-Theory vs. Pu /Py-FEA,Test

Electronic copy available at: https://ssrn.com/abstract=3889448

 Table 2: Statistics of Pu-FEA,Test/Pu-Theory
Statistical descriptors Pu-Theory
DSM-LB EWM IE
Number, n 39 39 39
Mean, µ 0.98 1.33 1.02
Standard deviation, σ 0.083 0.147 0.076
Maximum 1.11 1.59 1.12
Minimum 0.81 1.01 0.79

4. Conclusions
From the comparison of deformation plots obtained from FEA, it is concluded that there is an
interaction between the stiffened and unstiffened element buckling modes in plain channel
compression members, which is equivalent to local-distortional buckling interaction in lip stiffened
channels. Even though not very significant in magnitude, the interaction of buckling modes may lead to
reduction up to 20% in ultimate strength compared to the specimens where interaction is not present.
Hence the interaction equation (Eq. 5) which was originally developed for the lipped channels
considering the local and distortional buckling strengths are found to be appropriate for the strength of
plain channels also. The current DSM–LB (Eq. 3) equation is also found to give comparable strength
prediction of these sections. This work may lead to unifying current design procedure irrespective of the
type of plate elements (stiffened, partially stiffened or unstiffened) in CFS compression members.

References
ABAQUS. [Computer software]. ABAQUS standard user’s manual, Version 6.8. Dassault Systemes Simulia,
Providence, RI.
American Iron and Steel Institute (AISI) (2016). North American Specification for the design of cold-formed steel
structural members, S136-16, Washington, DC.
Australian/New Zealand Standard (AS/NZS) (2018). Cold-formed steel structures, AS/NZS 4600:2018, Sydney,
Australia.
Bambach, M. (2009). “Design of uniformly compressed edge-stiffened flanges and sections that contain them.”
Thin-Walled Struct., 47(3), 277–294.
Kalam Aswathy, K. C., & Anil Kumar, M. V. (2020). Unstiffened Elements as Limiting Case of Distortional Buckling of
Partially Stiffened Elements. Journal of Struct. Eng., 146(9), 04020171.
Kumar, M. A. and Kalyanaraman, V. (2010). “Evaluation of direct strength method for CFS compression members
without stiffeners.” J. Struct. Eng., 136(7), 879–885.
Kumar, M. A. and Kalyanaraman, V. (2012). “Design strength of locally buckling stub-lipped channel columns.” J.
Struct. Eng., 138(11), 1291-1299.
Kumar, M. A. and Kalyanaraman, V. (2014). “Distortional buckling of CFS stiffened lipped channel compression
members.” J. Struct. Eng., 140(12), 04014099.
Kumar, M. A. and Kalyanaraman, V. (2018). “Interaction of Local, Distortional, and Global Buckling in CFS Lipped
Channel Compression Members.” J. Struct. Eng., 144(2), 04017192.
Mulligan, G. P., & Pekoz, T. (1987). “Local buckling interaction in cold-formed columns.” J. Struct. Eng., 113(3), 604-
620.
Schafer, B. W. and Adany, S. (2005). “Understanding and classifying local, distortional and global buckling in open
thin-walled members.” Tech. Session and Mtg., Structural Stability Research Council. Montreal, Canada.
Talja, A. (1990), “Design of the buckling resistance of compressed HSS channels.” Research Note No. 1163,
Technical Research Centre of Finland.
Young, B., & Rasmussen, K. J. (1998). “Tests of fixed-ended plain channel columns.” J. Struct. Eng., 124(2), 131-139.

Electronic copy available at: https://ssrn.com/abstract=3889448