TECHNICAL PAPER

Journal of the South African

Buckling of short
Institution of Civil Engineering
cold‑formed lipped
channels in compression
Vol 56 No 2, August 2014, Pages 46–53, Paper 1044

PROF MORGAN DUNDU is the Head of School of

Civil Engineering and the Built Environment at
M Dundu
the University of Johannesburg. He holds a BSc
(Eng) degree from the University of Zimbabwe,
and an MSc and PhD in Structural Engineering
from the University of the Witwatersrand. He is a This paper presents an experimental investigation of short cold-formed lipped channel
member of the American Society of Civil columns compressed between pinned ends. The short columns are subjected to pure axial
Engineers (ASCE), an editorial board member of
compressive loading. Twelve column specimens are tested and the columns are categorised
the Journal of Steel Construction: Design and Research, a member of the
European Convention of Constructional Steelwork (ECCS), a member of the
into three groups, depending on the length and thickness. The buckling modes of failure that
International Association of Bridge and Structural Engineering (IABSE), and a occurred include local buckling and distortional buckling. A comparison of the experimental
member of the Standards Association of South Africa (SASA). results with the loads predicted by the South African standard for the design of cold-formed
Contact details: steelwork (SANS 10162-2) shows that the code is not conservative enough to cater for these
School of Civil Engineering & the Built Environment columns.
University of Johannesburg
PO Box 524
Auckland Park
INTRODUCTION BACKGROUND
2006
South Africa Cold-formed lipped short columns are The study of the buckling failure modes for
T: +27 11 559 3815 vulnerable to local or distortional buckling short cold-formed lipped sections dates to
F: +27 11 559 3815 or both, because of the high width-to- the 1940s. Initial work on these members was
E: [email protected] thickness ratios of the web and/or flange. performed by Winter (1940; 1943). A sum-
Local buckling occurs when the line junc- mary of this work, including all the work done
tion of both corners does not change, with in the USA, is given by Winter (1949). In the
the adjoining lip, flange and web elements UK, studies on thin-walled columns are sum-
failing by plate flexure alone (Figure 1(a)). marised by Chilver (1951; 1953) and Harvey
Distortional buckling occurs in thin-walled (1953). Chilver and Harvey’s elastic plate
channel sections when the lip-stiffened buckling solution was based on the work of
flange of the section rotates about the Timoshenko and Gere (1936). The work was
flange–web junction. This usually occurs extended by Lundquist and Stowell (1943),
if the lip stiffener does not have enough to enable stability methods to be applied in
stiffness to prevent the flange from rotating. practice. It is worth noting that both Chiver
The rotation can cause the flange to either and Harvey’s methods of calculating the local
move outward or inward depending on buckling stress included the interaction of
the nature of the load, supporting system elements, which is common practice today.
or imperfections. At ultimate failure, both To ensure that local buckling (and not distor-
rotations can be accompanied by the bend- tional buckling) occurs in lipped channels,
ing of the web (Figures 1(b) and 1(c)). The Chiver suggested that the lip has to be stiff
wave length of this mode of failure is in enough; however, no criteria for achieving this
most cases between that of local and overall were given. This recognised that distortional
member buckling, which makes it a practi- buckling was a potential problem. The “effec-
cal member length. tive width” aspect of the solution was based

(a) (b) (c)

Local buckling Outward distortional buckling Inward distortional buckling
Keywords: cold-formed lipped channels, short columns,
compressive resistance, local buckling, distortional buckling Figure 1 Local and short half-wavelength distortional buckling modes

46 Journal of the South African Institution of Civil Engineering  •  Volume 56  Number 2  August 2014
on work done by Von Kármán et al (1932) and Table 1 Dimensions of the columns
modifications by Winter (1947). It is interest-
L h b c t ri ry 3h 20r y
ing to note that, after almost six decades of Specimens
(mm) (mm) (mm) (mm) (mm) (mm) (mm) (mm) (mm)
thin-walled column research, the work used
500 300 75 20 2.0 3.00 25.7 900 514
today is still similar to Chilver’s work.
750 300 75 20 2.0 3.00 25.7 900 514
Distortional buckling was initially 300 × 75 × 20 × 2.0
recognised during the 1960s, under many 1 000 300 75 20 2.0 3.00 25.7 900 514
different names. Sharp (1966) analytically 1 250 300 75 20 2.0 3.00 25.7 900 514
approximated the distortional buckling stress
500 300 75 20 2.5 3.75 25.4 900 508
of aluminium lipped channels, and termed
it “overall” buckling. Dwight’s (1963) experi- 750 300 75 20 2.5 3.75 25.4 900 508
300 × 75 × 20 × 2.5
ments were used to verify Sharp’s results. 1 000 300 75 20 2.5 3.75 25.4 900 508
Using the folded-plate method, Goldberg et
1 250 300 75 20 2.5 3.75 25.4 900 508
al (1964) predicted the sectional distortion or
500 300 75 20 3.0 4.50 25.1 900 502
distortional buckling of open sections under
both compressive axial and bending load. 750 300 75 20 3.0 4.50 25.1 900 502
300 × 75 × 20 × 3.0
Wittrick (1968a; 1968b) discovered distortion- 1 000 300 75 20 3.0 4.50 25.1 900 502
al buckling modes (and called them torsional
1 250 300 75 20 3.0 4.50 25.1 900 502
modes) during a study of stiffened panels in
compression, using an exact stiffness method.
In the work of Desmond (1977) and Desmond the model, was developed. Further analysis studies by Prola and Camotim (2002),
et al (1981a; 19881b), which formed the basis and experiments were performed in which and Silvestre and Camotim (2004; 2006),
for the AISI (1996) specification on edge- distortional buckling was the failure mecha- based on the Generalised Beam Theory
stiffened elements, the term “stiffener” buck- nism (Lau & Hancock 1990). (popularly known as GBT), it was found
ling describes the distortional mode. It was In Japan, Hikosaka et al (1987), and that the distortional post-buckling strength
recognised that “stiffener” buckling occurred Takahashi (1988) conducted research to of a simple lipped channel made from
at a higher critical stress than local buckling. predict distortional buckling of thin-walled conventional carbon steel was higher if the
By providing a single empirical solution for members with polygonal cross section. flanges move towards each other (inwards)
the buckling coefficient k of an edge-stiffened Rasmussen and Hancock (1991) showed than if flanges move away from each other
element in either local or “stiffener” buckling, the importance of different end fixity on (outwards). However, for lipped channels
distortional buckling was treated as another distortional post-buckling behaviour. Kwon with intermediate stiffeners the post-
local mode, and not explicitly different from and Hancock (1992) performed experiments buckling strength was higher if the flanges
local plate buckling. on lipped channels with and without groove move away from each other, as observed
In order to increase the local buckling stiffeners in the web in which the distortional by Yang and Hancock (2004). Silvestre and
stress of slender webs of lipped channels, mode was unrestricted. These tests showed Camotim (2006) observed that for inward
Thomasson (1978) introduced small groove that interaction of distortional buckling with flange-lip movement, the stiffener develops
stiffeners. This eliminated local buckling, other modes is weak (supported by Young & high compression, which destabilises the
but created what Thomasson called a “local- Rasmussen 1998, 1999; Davies & Jiang 1998) web and is largely responsible for the less
torsional” buckling, which is distortional and that distortional buckling has lower post- stiff post-buckling response. Research about
buckling. To eliminate “local-torsional” buck- buckling capacity than local buckling (Schafer short cold-formed lipped channel sections is
ling, closely spaced braces were attached from 1997). Eurocode 3 (EN 3 1996) provided a still being undertaken to better understand
one lip to another. Once again the local mode method for predicting the distortional buck- their behaviour (bucking failure modes) and
became dominant. A finite strip method ling of cold-formed lipped channels. strength characteristics.
was developed by Sridharan (1982) to study Yang and Hancock (2004) observed a The aims of this paper are to investigate
distortional post-buckling behaviour, which significant increase in post-buckling strength the behaviour (buckling-failure modes) and
he called “local-torsional” buckling. Using when the flanges of carbon steel sections the strength (load-carrying capacity) of
this method, a rapid increase in membrane with intermediate stiffeners moved away short cold-formed lipped channel sections.
stresses was found at the tips of the edge- from the centroid, compared to those speci- According to Ziemian (2010), a short column
stiffening lips. Mulligan (1983), and Mulligan mens which had inward movement. Lecce is defined as a member that is sufficiently
and Pekoz (1984), observed distortional and Rasmussen (2005; 2006) investigated short so as to preclude member buckling
buckling during a local buckling study, and as distortional-buckling cold-formed stainless when compressed, but sufficiently long to
suggested by Thomasson, also called the phe- steel sections with intermediate stiffeners. contain the same initial residual stress pat-
nomena “local torsional” buckling. To restrict Although the specimens were identical, dis- tern as a much longer member cut from the
these phenomena, Mulligan provided braces tortional buckling caused the flanges of some same stock. For cold-formed steel sections,
in a manner similar to Thomasson. The need specimens to move away from the geometric the stub-column length should be suffi-
to investigate the behaviour of cold-formed centroid of the section, whereas the flanges ciently long to exhibit local buckling, as well
steel storage rack columns at the University of of others moved in towards the centroid. as the effect of cold-forming on the column
Sydney led to work on distortional buckling Inward or outward deformations may have performance. The length of cold-formed stub
(Hancock 1985; Lau & Hancock 1988). A hand been caused by imperfections. Specimens columns should not be less than three times
method (Lau & Hancock 1987) for predicting that failed by outward movement of the the largest dimension of the cross section,
the elastic distortional buckling stress, which flanges maintained the ultimate load over a and no more than 20 times the least radius-
used similar classic analytical techniques to greater vertical displacement, thus produc- of-gyration. The cold-formed lipped sections
Sharp (1966) but included web instability in ing a more ductile failure. In the extensive tested are given in Table 1, and all do not

Journal of the South African Institution of Civil Engineering  •  Volume 56  Number 2  August 2014 47
Table 2 Mechanical properties machine with an ultimate capacity of 100 kN
was used to test the coupons. The tensile load
Specimen Test No f y (MPa) f u (MPa) εu (%) E (MPa)
was applied at a low rate of 3 mm/min. A cali-
1 247.90 348.13 19.42 202 173 brated extensometer was used to measure the
2 238.64 341.19 22.44 198 634 strain. The material properties are shown in
300 × 75 × 20 × 2.0
3 241.83 345.33 19.24 201 492 Table 2. In this table fu is the 0.2% proof yield
strength, fy is the ultimate strength, E is the
Average 242.79 344.88 20.37 200 766
elastic modulus and εu is the strain at fracture
1 284.13 374.36 16.55 205 106 based on a gauge length of 50 mm. These
2 277.49 369.76 16.93 201 498 values were established from stress–strain
300 × 75 × 20 × 2.5
3 285.18 372.47 17.10 202 571 curves obtained from coupon tests. The 0.2%
yield strengths were used to calculate the
Average 282.26 372.20 16.86 203 058
code-predicted strengths of the columns.
1 254.91 344.44 21.26 206 749

2 268.05 360.85 19.63 208 032

300 × 75 × 20 × 3.0
3 253.84 345.97 20.16 204 143 TEST PROCEDURE
Twelve short columns were tested under
Average 258.94 350.42 20.35 206 308
pure axial compressive loading to establish
their compressive resistance strengths and
comply with the length provisions for a short MECHANICAL PROPERTIES failure modes. The specimens were delivered
column, as defined by Ziemian (2010). In this The material properties of the cold-formed by the manufacturer in lengths of 6 m. Each
table L is the length of the sub-column, h is column sections were determined from specimen was cut to a specified column
the depth, b is the width, c is the size of the coupon tests. Three longitudinal coupons length using an electric saw. The ends of the
lip, t is the thickness, ri is the internal radius were cut from the web to establish the yield columns were machined flat and the load was
and r y is the minimum radius of gyration of stress, ultimate stress, elastic modulus and the applied at the centroid (Figure 2(b)). As shown
the section. To cover a wider range of lengths elongation of the materials. The coupons were in the schematic experimental set-up, shown
it was decided to test four specimens of each prepared and tested according to the British in Figure 2(a), the load was applied through
section size. Standard, BS 18. A 1195 Instron testing two thick plates with grooves to accommodate

Load kN h

Y Y
600 × 140 × 30 mm steel plate
b
t
Centroid
Lipped channel section

c
600 × 140 × 30 mm steel plate

(a) Schematic experimental set-up (b) Loading position

(c) Top end condition (d) Bottom end condition

Figure 2 Typical experimental set-up

48 Journal of the South African Institution of Civil Engineering  •  Volume 56  Number 2  August 2014
Table 3 Test results
L fy N EX NY N EY NC N EX N eX N eX Buckling
Specimen
(mm) (MPa) (kN) (kN) (kN) (kN) NY N EY NC modes

500 242.79 119.31 230.65 138.02 136.32 0.52 0.86 0.88 LB

750 242.79 114.08 230.65 138.02 133.60 0.49 0.83 0.85 LB

300 × 75 × 20 × 2.0
1 000 242.79 109.86 230.65 138.02 129.85 0.48 0.80 0.85 LB

1 250 242.79 114.96 230.65 138.02 125.19 0.50 0.83 0.92 DB

500 282.26 159.79 333.07 210.96 207.01 0.48 0.76 0.77 LB

750 282.26 166.97 333.07 210.96 202.18 0.50 0.79 0.83 LB

300 × 75 × 20 × 2.5
1 000 282.26 182.36 333.07 210.96 195.61 0.55 0.86 0.93 LB

1 250 282.26 164.73 333.07 210.96 187.50 0.49 0.78 0.88 DB

500 258.94 156.14 362.52 251.44 247.16 0.43 0.62 0.63 LB

750 258.94 161.37 362.52 251.44 241.92 0.45 0.64 0.67 LB

300 × 75 × 20 × 3.0
1 000 258.94 178.18 362.52 251.44 234.77 0.49 0.71 0.76 LB

1 250 258.94 146.37 362.52 251.44 225.90 0.40 0.58 0.65 DB

LB = local buckling, DB = distortional buckling

a spherical ball, so as to simulate a pin. Thick stub columns. According to this classifica- columns are actually short, is the fact that
plates ensured the uniform spread of the load tion, the length of hot-rolled stub columns the effective yield resistance and the buckling
throughout the entire cross-sectional area should not be less than (2d + 250 mm) or resistance are almost the same.
of the column. Figure 2(c) and Figure 2(d) 3d, whichever is smaller, and not greater A comparison of the test results and
show the top and bottom end conditions of than 20ry or 5d, whichever is larger, where calculated un-factored resistances shows
the set-up. A Moog testing machine with an d is the depth of the shape and ry is radius- that the SANS 10162-2 specification and its
ultimate capacity of 2 000 kN was used for of-gyration about the minor axis (Ziemian related specification (AS/NZS 4600) over-
the experimental test. To allow gradual failure 2010). This classification is wide enough to estimate the compressive strengths of short
of the columns, the load was applied at a rate accommodate these sections. cold-formed lipped channel columns. The
of 2 mm/min. The testing machine measured Table 3 also shows the experimental standards are based on Winter’s (1947) post-
the axial compressive load and displacement. buckling resistance to be almost 50% of buckling equation, and all specimens tested
the gross-sectional yield resistance for the did not show post-buckling behaviour. On
300x75x20x2.0 and 300x75x20x2.5 speci- average the test loads for the 300x75x20x2.0,
TEST RESULTS mens. This percentage reduces to about 45% 300x75x20x2.5 and 300x75x20x3.0 specimens
The experimental compressive resistance for the 300x75x20x3.0 specimens. This huge are 85%, 83% and 66% of the effective squash
(N EX), gross yield compressive resistance decrease in resistance clearly shows the or buckling loads predicted by SANS 10162-2
(N Y = Agf y), effective yield compressive importance of considering local and distor- respectively. In the 300x75x20x3, the average
resistance (N EY = Aef y) and the buckling tional buckling in such members. It is well difference between the test results and the
compressive resistance (NC = Aefn) are given known that the main effect of local buckling code predicted strength is 34%. The largest
in Table 3. The effective area (Ae) is deter- is to cause a redistribution of the longitudinal cause of this low resistance is local buckling
mined using the effective width method, stress so that the greatest portion of the load and distortional buckling. Local buckling
where the channel lips area, the flanges area is carried by the corners and its adjacent occurred at the end of the column in the
and the web area of the column are analysed areas. This situation increases the compressed 500 mm, 750 mm, and 1 000 mm tests and
separately. N EY and NC are calculated using stresses in and around the plate junctions, may have been initiated by increased cross-
the stress conditions of f = f y and f = f u, and high bending or flexure stresses in the section distortions and initial geometric
respectively, where fn is buckling critical flat portions, resulting in capacities that are imperfections caused by the cutting process
stress. The predicted resistance values are well below the squash load of the section. The of these sections. Although no measurements
obtained using the South African code, effect is also seen in the corresponding effec- of these imperfections are given in this paper,
SANS 10162-2. This code is based on the tive yield and the buckling resistances. The the author noticed that the flange-lip junction
Australian code AS/NZS 4600. difference in column strength between the moved outward even before the sections were
Table 3 shows that there is not much sections tested is as a result of the different tested. Lam et al (2006) has shown that the
difference between the experimental com- thickness, yield strengths and the magnitude compressive capacity and modes of failure
pressive resistances of the chosen lengths of of the residual stresses. A clear trend can of lipped stub channels are very sensitive to
each specimen. This implies that all lengths be seen here: the ratio of the experimental the magnitude and distribution of the initial
of columns can be classified as short. If this strength to the predicted strength decreases geometric imperfections. In this investigation,
is the case then it is not appropriate to use with increase in thickness of the specimen. the compressive strength of ten fixed-ended
the length suggested by Ziemian (2010) for Residual stresses tend to affect thicker steel stub columns, manufactured from
cold formed sections. In order to classify sections and are responsible for the lower 150x65x13x1.6 mm and 100x50x10x1.6 mm
the length of these sections it is proposed to strength ratio of the 300x75x20x3.0 speci- cold-formed lipped channels, were found to be
adopt the length classification for hot-rolled mens. Another confirmation that the tested 75% and 77% of the strengths predicted by the

Journal of the South African Institution of Civil Engineering  •  Volume 56  Number 2  August 2014 49
 (a) 500 mm (b) 750 mm

(c) 1 000 mm (d) 1 250 mm

Figure 3 Bucking failure modes

British standard (BS 5950-5) respectively. This eccentricity is the distance from the effec- might be the size of the lip. Specimens with a
reduced resistance was influenced by cross- tive centroid to the geometric centroid. The smaller effective lip size resist less load than
section distortions and initial geometrical effective centroid is determined from the those with a larger lip size (Kwon & Hancock
imperfections, caused by the cutting process. effective cross section, calculated from the 1992). Kwon and Hancock (1992) also found
In a previous study, Wang et al (2006) showed effective widths of each plate. the maximum stresses in small-sized lipped
that, although the imperfections occurred in Another possible cause of the substantial channel sections to range from 25% to 34%
all the flat portions of the section, they were low ratios of test versus predicted strengths of the yield stress.
significant in the flanges. Cutting had little or
no effect at the web-flange corner. 140
The effect of cross-section distortions
and initial geometrical imperfections 120
was made worse by the pinned boundary
conditions chosen in this set-up. In singly
100
symmetric cross-sections with pinned ends,
Rhodes and Harvey (1977), and Young and
80
Load (kN)

Rasmussen (1998; 1999) have shown that the

redistribution of longitudinal stresses caused
by local buckling also produces a shift of the 60

line of action of the internal force or effective

centroid. The shift induces overall bending, 40
which significantly reduces the column
strength. In this paper the load was applied 20
through the geometric centroid, whereas the
specifications assume concentric loading to 0
be at the centroid of the effective cross sec- 0 1 2 3 4 5 6 7 8 9
tion. This means that the buckled member Displacement (mm)
now behaves as a beam-column, in which the 500 mm 750 mm 1 000 mm 1 250 mm
applied moment is calculated as a product of
the axial force and its eccentricity, where the Figure 4 Load-displacement graphs of 300 × 75 × 20 × 2 short columns

50 Journal of the South African Institution of Civil Engineering  •  Volume 56  Number 2  August 2014
FAILURE MODES 210
Two buckling modes of failure were observed
from the experimental tests, namely local 180
and distortional buckling. Local buckling
(LB) occurs when parts or portions of the 150
section elements yield before the actual
yielding of the member is achieved, and was 120

Load (kN)
observed in very short columns (typically in
lengths of 500 mm to 1 000 mm) (Figures 90
9(a) and 9(b)). The local buckling failure
mode of these columns was more excessive 60
at the ends of the columns, and caused the
web to buckle outwards towards the shear 30
centre of the members. This failure mode
is more distinct in thinner sections and the 0
buckle of these sections seems to cover a 0 1 2 3 4 5 6 7 8 9 10 11 12
larger area of the web than thicker sections. Displacement (mm)
As indicated above, the process of cutting 500 mm 750 mm 1 000 mm 1 250 mm
cold-formed steel sections to form a stub
column may have led to cross-section distor- Figure 5 Load-displacement graphs of 300 × 75 × 20 × 2.5 short columns
tion and initial geometric imperfection.
Distortional buckling (DB), also known as 210
“stiffener buckling”, is a mode characterised
by rotation of the flange at the flange/web 180
junction in members with edge-stiffened ele-
ments (Schafer & Pekoz 1998). Distortional
150
buckling was observed in 1 250 mm length
columns. Figure 3(d) shows evidence of dis-
Load (kN)

120
tortional buckling. In this case distortional
buckling is characterised by the closing up of
90
the two flanges. This mode of failure is prob-
ably unsymmetrical, because the section was
distorted during the cutting process (Lam et 60
al 2006).
30

LOAD DISPLACEMENT 0
The experimental load-displacement graphs 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14
for the three different sections are shown Displacement (mm)
in Figure 4. The graphs are linear-elastic 500 mm 750 mm 1 000 mm 1 250 mm
up to the maximum load and there is little
or no evidence of inelastic behaviour. For Figure 6 Load-displacement graphs of 300 × 75 × 20 × 3 short columns
the 300x75x20x2 specimens, the load–axial
shortening curves experience abrupt drops achieved between the different lengths. reduced the effectiveness of the lips of the
immediately after the ultimate loads, Columns with length ranging from 500 mm 300x75x20x2 mm and 300x75x20x2.5 mm
while for specimens 300x75x20x2.5 and to 1 000 mm failed by local buckling of the column sections.
300x75x20x3.0 the load–axial shortening web, whilst the 1 250 mm column length Local buckling in the web and the result-
curve follows a smooth decrease of loads failed by distortional buckling. The ultimate ing low strength of the lipped channels
at the post-ultimate range (Figures 4 and compressive resistance obtained from may have been caused by the increased
5). The abrupt drops correspond to the the test results ranged from 66% to 85% cross-section distortion and initial geometric
sudden collapse of the columns in the tests. of the compressive resistance, calculated imperfections caused by the cutting process
Generally, columns with larger thickness based on SANS 10162-2 and its related of these sections. Before the tests were
experience less abrupt drops in load immedi- specifications. This implies that the design- performed, it was noticed that the flange-
ately after the ultimate load, compared with predicted strengths by the SANS 10162-2 lip junction had moved outward. Lam et
thinner sections. specification generally overestimated (not al (2006) has shown that the compressive
conservatively) the design compressive capacity and modes of failure of lipped stub
strengths of the short cold-formed lipped channels are very sensitive to the magnitude
CONCLUSION channel columns. The predicted strengths and distribution of the initial geometric
Ziemian’s (2010) suggested length for for the 300x75x20x3 mm columns are much imperfection. Although the imperfections
short columns does not apply to the tested larger than the test strength. This huge occurred in all the flat portions of the sec-
sections; however, the sections can be difference is caused by the fact that the tion, they were significant in the flanges
classified as short, considering that there lips for these sections are all fully effective (Wang et al 2006). Cutting had little or no
is little or no variation in the test strength in resisting local buckling. Local buckling effect at the web-flange corner. Further,

Journal of the South African Institution of Civil Engineering  •  Volume 56  Number 2  August 2014 51
SANS 10162-2 and its related specifications Desmond, T P, Peköz, T & Winter, G 1981a. Edge University, Department of Structural Engineering,
(AISI S100-12 and AS/NZS 4600) are based stiffeners for thin-walled members. Journal of the Report No 83–1.
on Winter’s (1947) post-buckling equation. Structural Division, ASCE, 107(2): 329–353. Mulligan, G P & Pekoz, T 1984. Locally buckled thin-
The equation assumes that post-buckling Desmond, T P, Peköz, T & Winter, G 1981b. walled columns. Journal of the Structural Division,
strength will develop in the web and flange. Intermediate stiffeners for thin-walled members. ASCE, 110(11): 2635–2654.
No post-buckling behaviour was shown by Journal of the Structural Division, ASCE, Prola, L C & Camotim, D 2002. On the distortional
all the specimens tested in this paper. 107(4): 627–648. buckling of cold-formed lipped channel steel
The pinned boundary conditions chosen Dwight, J B 1963. Aluminum sections with lipped columns. Proceedings, Structural Stability Research
in this set-up did not make the situation any flanges and their resistance to local buckling. Council (SSRC) Annual Stability Conference,
better. It has been shown by Rhodes and Proceedings, Symposium on Aluminum in Seattle, WA, pp 571–590.
Harvey (1977), and Young and Rasmussen Structural Engineering, London. Rasmussen, K J R & Hancock, G J 1991. The flexural
(1998; 1999) that in singly symmetric EN 3 1996. Eurocode: Design of steel structures. Part behavior of thin-walled singly symmetric columns.
cross-sections with pinned ends, the redis- 1.3. General rules and rules for buildings. Brussels: Proceedings, International Conference on Steel and
tribution of longitudinal stresses caused by Committee for Standardization (CEN). Aluminum Structures, May, Singapore.
local buckling shifts the line of action of Goldberg, J E, Bogdanoff, J L & Glauz, W D 1964. Rhodes, J & Harvey, J M 1977. Interaction behaviour
the internal force or effective centroid. The Lateral and torsional buckling of thin-walled beams, of plain channel columns under concentric or
shift introduces a moment, which signifi- Proceedings of the International Association for Bridge eccentric loading. Proceedings, 2nd International
cantly reduces the column strength. Since and Structural Engineering (IABSE), 24: 91–100. Colloquium on the Stability of Steel Structures,
the load was applied at the geometric cen- Hancock, G J 1985. Distortional buckling of steel Liege, Belgium, pp 439–444.
troid, the eccentricity is the distance from storage rack columns. Journal of the Structural SANS 10162-2 2011. The structural use of steel. Part 2:
the effective centroid (determined using the Division, ASCE, 111(12): 2770–2783. Cold-formed steel structures. Pretoria: South African
effective width method) to the geometric Harvey, J M 1953. Structural strength of thin-walled Bureau of Standards.
centroid. channel sections. Engineering, 175: 291–293. Schafer, B W 1997. Cold-formed steel behavior and
Hikosaka, H, Takami, K & Maruyama, Y 1987. Analysis of design: Analytical and numerical modeling of
elastic distortional instability of thin-walled members elements and members with longitudinal stiffeners.
ACKNOWLEDGEMENTS with open polygonal cross-section. Proceedings of PhD Thesis, Ithaca, NY: Cornell University.
The author wishes to acknowledge Nhlanhla the Japan Society of Civil Engineering, Structural Schafer, B W & Pekoz, T 1998. Computational
Tshabalala, a final-year Civil Engineering Engineering/Earthquake Engineering, 4(1): 31–40. modelling of cold-formed steel: Characterizing
student, who assisted in carrying out the tests Kwon, Y B & Hancock, G J 1992. Tests of cold- geometric imperfections and residual stresses.
reported in this paper, and the University of formed channels with local and distortional Journal of Constructional Steel Research,
Johannesburg Research Committee (URC) for buckling. Journal of the Structural Division, ASCE, (47): 193–210.
sponsoring this research. 117(7): 1786–1803. Sharp, M L 1966. Longitudinal stiffeners for
Lam, S S E, Chung K F & Wang, X P 2006. Load- compression members. Journal of the Structural
carrying capacities of cold-formed steel cut stub Division, ASCE, 92(5): 187–211.
REFERENCES columns with lipped c-section. Thin-Walled Silvestre, N & Camotim, D 2004. Local-plate and
AISI (American Iron and Steel Institute) 1996. Structures, 44: 1077–1083. distortional postbuckling behaviour of cold-formed
Specification for the design of cold-formed steel Lau, S C W & Hancock, G J 1987. Distortional buckling steel lipped channel columns with intermediate
structural members. Washington DC: AISI. formulas for channel columns. Journal of the stiffeners. Proceedings, 17th International Specialty
AISI (American Iron and Steel Institute) 2012. S100-12 Structural Division, ASCE, 113(5): 1063–1078. Conference on Cold-formed Steel Structures,
2012. North American specification for the design of Lau, S C W & Hancock, G J 1988. Distortional University of Missouri-Rolla, Orlando, FL, pp 1–8.
cold-formed steel structural members. Washington buckling tests of cold-formed channel sections. Silvestre, N & Camotim, D 2006. Local-plate and
DC: AISI. Proceedings, 9th International Specialty Conference distortional postbuckling behaviour of cold-formed
AS/NZS 4600 2005. Cold-formed steel structures. on Cold-formed Steel Structures, St. Louis, steel lipped channel columns with intermediate
Sydney: Standards Australia. Missouri, pp 45–73. stiffeners. Journal of the Structural Division, ASCE,
BS 18 1987. British standard method for tensile testing Lau, S C W & Hancock, G J 1990. Inelastic buckling 132(4): 529–540.
of metals. London: British Standards Institution. of channel columns in the distortional mode. Thin- Sridharan, S 1982. A semi-analytical method for the
BS 5950-5 1998. Structural use of steelwork in building: Walled Structures, 10(1): 59–84. post-local torsional buckling analysis of prismatic
Code of practice for design of cold formed thin gauge Lecce, M & Rasmussen, K J R 2005. Experimental plate structures. International Journal of Numerical
sections. London: British Standards Institution. investigation of the distortional buckling of cold- Methods in Engineering, 18: 1685–1697.
Chilver, A H 1951. The behaviour of thin-walled formed stainless steel sections. Department of Civil Takahashi, K 1988. A new buckling mode of thin-walled
structural members in compression. Engineering, Engineering, Sydney: University of Sydney, Research columns with cross-sectional distortions In: Tooth,
172(4466): 281–282. Report No 844. A S & Spencer, J (Eds), Solid Mechanics, Elsevier,
Chilver, A H 1953. The stability and strength of Lecce, M & Rasmussen, K J R 2006. Distortional 25: 553–573.
thin-walled steel struts. The Engineer, 196(5089): buckling of cold-formed stainless steel sections: Thomasson, P 1978. Thin-walled c-shaped panels in
180–183. Finite element modelling and design. Journal of the axial compression. Stockholm, Sweden: Swedish
Davies, J M & Jiang, C 1998. Design for distortional Structural Division, ASCE, 132(4): 505–514. Council for Building Research.
buckling. Journal of Constructional Steel Research, Lundquist, E E & Stowel, E Z 1943. Principles of Timoshenko, S P & Gere, J M 1936. Theory of Elastic
46(1): 174–175. moment distribution applied to the stability Stability, 2nd ed. New York: McGraw-Hill.
Desmond, T P 1977. The behaviour and strength of structures composed of bars or plates. Yang, D & Hancock, G J 2004. Compression tests
of thin-walled compression members with NACA Advance Restricted Report L-326. of high-strength steel channel columns with
longitudinal stiffeners. Ithaca, NY: Cornell Mulligan, G P 1983. The influence of local buckling interaction between local and distortional
University, Department of Structural Engineering, on the structural behavior of singly symmetric buckling. Journal of the Structural Division, ASCE,
Report No 369. cold-formed steel columns. Ithaca, NY: Cornell 130(12): 1954–1963.

52 Journal of the South African Institution of Civil Engineering  •  Volume 56  Number 2  August 2014
Young, B & Rasmussen, K J R 1998. Design of lipped Wang, X P, Lam, S S E & Chung, K F 2006. Cross- Winter, G 1949. Performance of compression plates as
channel columns. Journal of the Structural Division, section distortion due to cutting of cold-formed parts of structural members. Research, Engineering
ASCE, 124(2): 140–148. steel lipped C-section. Thin-Walled Structures, Structures Supplement, Colston Papers, London, 2: 179.
Young, B & Rasmussen, K J R 1999a. Behaviour of cold- 44: 271–280. Wittrick, W H 1968a. General sinusoidal stiffness
formed singly symmetric columns. Thin-walled Winter, G 1940. Tests on light studs of cold-formed matrices for buckling and vibration analyses of
Structures, 33: 83–102. steel. Ithaca, NY: Cornell University, 3rd Progress thin flat-walled structures. International Journal of
Young B & Rasmussen K J R 1999b. Shift of effective Report (unpublished). Mechanical Sciences, 10: 949–966.
centroid of channel columns. Journal of the Winter, G 1943. Tests on light studs of cold-formed Wittrick, W H 1968b. A unified approach to the initial
Structural Division, ASCE, 125(5): 524–531. steel. Ithaca, NY: Cornell University, 2nd Summary buckling of stiffened panels in compression. The
Von Kármán, T, Sechler, E E & Donnell, L H 1932. The Report (unpublished). Aeronautical Quarterly, 265–283.
strength of thin plates in compression. Transactions Winter, G 1947. Strength of thin steel compression Ziemian, R 2010. Guide to Stability Design Criteria for
of the ASME, 54: 53–57. flanges. Transactions of the ASCE, 112(2305): 1. Metal Structures, 6th ed. Hoboken, NJ: Wiley.

Journal of the South African Institution of Civil Engineering  •  Volume 56  Number 2  August 2014 53