Journal of Constructional Steel Research 67 (2011) 1212–1224

Contents lists available at ScienceDirect

Journal of Constructional Steel Research

journal homepage: www.elsevier.com/locate/jcsr

Compression tests of longitudinally stiffened plates undergoing distortional

buckling
Young Bong Kwon ∗ , Ho Sang Park
Department of Civil and Environmental Engineering, Yeungnam University, Gyongsan, Republic of Korea

article info abstract

Article history: This paper describes a series of compression tests performed on longitudinally stiffened plates fabricated
Received 8 September 2010 from a mild steel plate of thickness of 4.0 mm with nominal yield stress of 235.0 MPa. The stiffened plates
Accepted 23 February 2011 with longitudinal stiffeners of a range of rigidities were tested to failure. The ultimate strengths and
performances of the longitudinally stiffened plates in compression undergoing distortional buckling or
Keywords: interaction between local and distortional buckling were investigated experimentally and theoretically.
Stiffened plates The compression tests indicated that the critical buckling mode was dependent mainly on the rigidity
Longitudinal stiffeners of the longitudinal stiffeners and the width-to-thickness ratio of the sub-panels. A noticeable interaction
Local buckling
between local and distortional buckling was also observed for some stiffened plates. A significant post-
Distortional buckling
Buckling interaction
buckling strength reserve was shown for those sections with distortional buckling and for those sections
Direct strength method showing interaction between local and distortional buckling. A limiting strength curve for distortional
buckling of longitudinally stiffened plates was studied. Simple design strength formulas in the direct
strength method are proposed to account for the distortional buckling and the interaction between local
and distortional buckling of longitudinally stiffened plates. The strength curves were compared with the
test and FE results conducted. The adequacy of the strength curve was confirmed. A set of conclusions on
the buckling behavior of longitudinally stiffened plates was drawn from the experimental studies.
© 2011 Elsevier Ltd. All rights reserved.

1. Introduction
Stiffened steel plates with longitudinal stiffeners are used
widely in steel structures such as steel box girders, box columns
and pylons for cable bridges. When compressed or bent, the
longitudinally stiffened plate may buckle in either local buckling
mode or distortional buckling mode [1]. The critical buckling
modes of the longitudinally stiffened plates shown in Fig. 1(a)
and (b) depend mainly on the in-plane flexural rigidity of the
longitudinal stiffener and the width-to-thickness ratio of the sub-
panels. The aspect ratio (l/w ) of the sub-panels is also an important
factor determining the buckling stress and mode of longitudinally
stiffened plates. According to the relative magnitude of the two
basic buckling stresses, the basic buckling modes may interact with
each other as shown in Fig. 1(c).
As the in-plane flexural rigidity of the longitudinal stiffeners
increases beyond the threshold limit, the basic buckling mode
changes from distortional buckling mode to local buckling mode
as shown in Fig. 2. Up to the limiting value, the buckling stress
∗ Corresponding author. Tel.: +82 11 802 2418; fax: +82 53 810 4622.
E-mail address: ybkwon@ynu.ac.kr (Y.B. Kwon).
0143-974X/$ – see front matter © 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.jcsr.2011.02.015
increases with the increase of the stiffener rigidity. However, the
buckling stress remains constant after the limit.
The design strength formulas for elastic local buckling, inelastic
local buckling and yielding of a material for longitudinally stiffened
box flanges is provided in current specifications such as ASSHTO
LRFD specifications (2007) [2]. The minimum value of the in-
plane flexural rigidity of a longitudinal stiffener to prevent
distortional buckling is provided. However, referring to the test
results [1], longitudinally stiffened plates satisfying the limit
of a longitudinal stiffener may undergo distortional buckling.
If longitudinally stiffened flanges undergo distortional buckling,
the strength equation based on local buckling cannot be used
effectively. Moreover, if local buckling and distortional buckling
of stiffened plates interact, the ultimate strength of stiffened
plates cannot be predicted properly by the current specifications.
However, Eurocode3 [3] has the provisions to account for both local
and distortional buckling of stiffened plates based on both column
and orthotropic plate theories.
Distortional buckling was first studied by Hancock (1978,
1985) [4,5]. The design strength of cold-formed steel sections
for the distortional mode was proposed by Kwon and Hancock
(1992) [6] and adopted by the AS/NZS (1996) [7]. The provision
for distortional buckling is also provided in NAS (AISI, 2001) [8].
The interaction between local buckling and distortional buckling of
 Y.B. Kwon, H.S. Park / Journal of Constructional Steel Research 67 (2011) 1212–1224 1213

2. Geometries and material properties of test sections

Notation
2.1. Section geometries
A Cross sectional area
b Width of stiffened plate
bf 1 Flange width of edge stiffener A series of compression tests was performed on mild steel
bf 2 Flange width of intermediate stiffener longitudinally stiffened plates of thickness 4.0 mm with nominal
Cc Slenderness limit for inelastic buckling yield and ultimate stresses 235 MPa and 400 MPa, respectively. The
d0 Initial imperfection in a distortional buckling mode plates were fabricated by continuous fillet welding to an effective
Fcrd Elastic distortional buckling stress width on both sides of the stiffener-plate joint. The size of the
Fcrl Elastic local buckling stress fillet welds was determined as 4.0 mm according to the AISC
Fkh Distortional buckling limiting stress specifications [13].
Fmax Ultimate strength determined in tests The finite strip method program BAP [14] was used for the
Fnl Local buckling limiting stress iterative buckling analyses to determine the optimum dimensions
Fy Nominal yield stress of the test sections. The program BAP is a buckling analysis
h1 Height of edge stiffener program, which can account for the inelasticity of the material
h2 Height of intermediate stiffener and the residual stress distribution. It assumes the hinged end
Is Inertia moment of longitudinal stiffener boundary conditions to impose a single half-wavelength on the
k Buckling coefficient of plate member analyzed. The sections tested were proportioned to fail
l Length of stiffened plate in the following categories: distortional buckling and the mixed
n Number of longitudinal stiffeners mode buckling between local buckling and distortional buckling.
Pkh Compression member design strength (=Fkh × A) The geometries of the test sections are shown in Fig. 3 and
Pnld Design limiting strength (=Fnld × A) summarized in Table 1. The width-to-thickness ratios of the sub-
R1 Slenderness limit for yielding panels ranged from 60 to 120. In fact, the provision for the
R2 Slenderness limit for inelastic buckling minimum moment of inertia of longitudinal stiffeners given by
tb Thickness of stiffened plate AASHTO specifications [2] to prevent distortional buckling has
tf 1 Flange thickness of edge stiffener been shown to be sometimes too conservative [15]. Therefore,
tf 2 Flange thickness of intermediate stiffener to ensure that the stiffened plates undergo distortional buckling
tw1 Web thickness of edge stiffener or show interaction between local and distortional buckling, the
tw2 Web thickness of intermediate stiffener inertia moment of the longitudinal stiffeners was set to have
w Width of sub-panel a value smaller than the values of Eq. (1), which was recently
α Aspect ratio (=l/w ) proposed as an equation for calculating the minimum moment of
ρ Reduction factor for plate buckling inertia of longitudinal stiffeners by Yoo et al. (2001) [15].
χ Reduction factor for column buckling √
Is = 0.3α 2 ntb3 w (1)
λd Distortionalbuckling slenderness ratio factor (=
Fy /Fcrd or Py /Pcrd ) where n = number of longitudinal stiffeners; w = width of the sub-

λl Local
 buckling slenderness ratio factor of sub-panel panel; tb = thickness of the stiffened plate; α = l/w aspect ratio
(= Fy /Fcrl ) of the sub-panel; l = length of the stiffened plate.

2.2. Determination of length of specimens

cold-formed steel channel sections was investigated by Yang and
Hancock (2004) [9], and Kwon et al. (2009) [10]. The distortional
strength formulas [9,10] were used to predict the strength of
cold-formed channel sections undergoing interaction between
local buckling and distortional buckling. However, based on our
knowledge, distortional buckling of longitudinally stiffened plates
has not been extensively studied.
The aim of this paper is to investigate distortional buckling
and the interaction between local buckling and distortional
buckling of longitudinally stiffened plates experimentally and
theoretically. The direct strength method, which has been
recently adopted for the cold-formed steel sections by NAS
(2004) [11] and AS/NZS (2005) [12], is extended to account for
distortional buckling and the interaction between local buckling
and distortional buckling of longitudinally stiffened plates. A
series of compression tests was performed on welded stiffened
plates fabricated from mild steel plates of thickness 4.0 mm
with nominal yield stress 235 MPa to develop the direct strength
curves accounting for distortional buckling and the interaction
between local buckling and distortional buckling. Nonlinear FE
analyses of the sections tested have also been conducted to
compare their results with the test results. The predictions by
the proposed direct strength formulas were compared with those
by the current specifications [2,3]. The direct strength formulas
were proven accurate and efficient for predicting the ultimate
strength of longitudinally stiffened plates that experienced nearly
simultaneous local buckling and distortional buckling.
Distortional buckling and the interaction between local and
distortional buckling were the main focus of the compression
tests on the longitudinally stiffened plates. The overall length
of the longitudinally stiffened plates was an important factor
for determining the critical buckling mode. To determine the
optimum length for testing, it is important to predict the local and
distortional buckling stress accurately. The buckling stress versus
half-wavelengths obtained by using program BAP was used to
analyze the test sections so that the lengths could be determined
properly.
The inelastic buckling analysis should be executed for test
plates composed of sub-panels of intermediate width-to-thickness
ratios. Since the residual stress distribution had a significant effect
on the inelastic behavior of a section with small or intermediate
width-to-thickness ratios, it should be included in the inelastic
buckling analysis. The residual stress distribution model was
calibrated by iterative calculation of previous test results [1]. The
assumed residual stress distribution across the cross section of
the longitudinally stiffened plates in the inelastic buckling and
FE analyses is shown in Fig. 4. The magnitudes of the tensile and
compressive residual stresses were assumed as Fy and −0.2Fy ,
respectively. The residual stresses in the stiffeners and fringes of
the stiffened plate were assumed to have a minor effect on the
buckling strength and therefore, were neglected for simplicity in
the inelastic buckling analysis.
1214 Y.B. Kwon, H.S. Park / Journal of Constructional Steel Research 67 (2011) 1212–1224

(a) local buckling. (b) distortional buckling. (c) mixed mode buckling.

Fig. 1. Buckling modes of stiffened plates subjected to uniform compression.

Table 1
Test section geometries (cm).
Specimens n α w b tb h2 tw 2 l Is Eq. (1)

R-60-1 2 4.2 24.0 75.0 0.4 3.5 0.4 100.0 5.7 11.3
R-70-1 1 3.6 28.0 59.0 0.4 3.5 0.4 100.0 5.7 6.9
R-80-1 1 3.1 32.0 67.0 0.4 3.0 0.4 100.0 3.6 6.0
R-90-1 1 2.8 36.0 75.0 0.4 3.0 0.4 100.0 3.6 5.3
R-100-1 1 2.5 40.0 83.0 0.4 3.0 0.4 100.0 3.6 4.8
R-110-1 1 2.3 44.0 91.0 0.4 3.0 0.4 100.0 3.6 4.4
R-120-1 1 2.1 48.0 99.0 0.4 3.0 0.4 100.0 3.6 4.0
R-60-2 2 5.0 24.0 75.0 0.4 4.0 0.4 120.0 8.5 16.3
R-70-2 2 4.3 28.0 87.0 0.4 3.5 0.4 120.0 5.7 14.0
R-80-2 2 3.8 32.0 99.0 0.4 3.0 0.4 120.0 3.6 12.2
R-90-2 1 3.3 36.0 75.0 0.4 3.0 0.4 120.0 3.6 7.7
R-100-2 1 3.0 40.0 83.0 0.4 3.0 0.4 120.0 3.6 6.9
R-110-2 1 2.7 44.0 91.0 0.4 3.0 0.4 120.0 3.6 6.3
R-120-2 1 2.5 48.0 99.0 0.4 3.0 0.4 120.0 3.6 5.8
The dimensions bf 1 = 5.0 cm, h1 = 4.5 and tf 1 = 0.6 cm are constant but tw1 = 0.4 cm and tw1 = 1.5 cm for test series-1 and series-2, respectively.

inelastic buckling analysis results in Fig. 5(a), the elastic or the

Fig. 2. Buckling stress versus longitudinal stiffener rigidity of stiffened plates.
Fig. 3. Cross section geometry of stiffened plate.
Fig. 4. Residual stress distribution.
The elastic and inelastic buckling stresses versus half-
wavelength curves of the R-60-2 and the R-90-2 specimens, sub-
jected to uniform compression, obtained using the program BAP
are illustrated in Fig. 5(a) and (b). For the R-60-2 where the width-
to-thickness ratio was smaller than 90, referring to the elastic and
inelastic critical stress for lengths longer than the length corre-
sponding to point B was determined as distortional buckling stress,
which was lower than the elastic or inelastic local buckling stresses
corresponding to point A. However, referring to the elastic and in-
elastic buckling stresses for lengths longer than the length corre-
sponding to point B in Fig. 5(b), the elastic critical buckling stress
of the R-90-2 was determined as the distortional buckling stress,
which is lower than the local buckling stress at half-wavelength
of approximate 230 mm (point A). However, the inelastic critical
buckling stress changed from distortional buckling to local buck-
ling at half-wavelength of 230 mm in the inelastic buckling analysis
where residual stress was included. This change was due to the fact
that the compressive residual stress in the flat area of the longitudi-
nally stiffened plates shown in Fig. 4 decreased the local buckling
stress more significantly than the distortional buckling stress. To
induce the final collapse of the test specimens mainly in the dis-
tortional buckling mode rather than the local buckling mode, the
overall lengths of the test specimens were determined as 1000 mm
for test series-1 and 1200 mm for test series-2, which were longer
than the length corresponding to the point B in Fig. 5(a) and
(b). For the R-60-2 of 1200 mm in length (point C) in Fig. 5(a),
the elastic or inelastic distortional buckling stress was the critical
buckling stress. Since the distortional buckling mode had a non-
negligible post-buckling strength reserve, an interaction between
local buckling and distortional buckling might occur. For the R-90-
2 of 1200 mm in length in Fig. 5(b), considering the inelastic buck-
ling stresses versus half-wavelength relations, the critical buckling
was a local buckling. However, since there was a significant post-
buckling strength reserve in the local buckling mode, there might
be an interaction between local buckling and distortional buckling
and failure can occur mainly in the distortional buckling mode dur-
ing testing. Therefore, considering the two cases mentioned above,
most of the test specimens were expected to fail mainly in the dis-
tortional buckling mode in the compression tests.
2.3. Material properties
The steel grade of the test sections was SM400 structural steel
to KS D3515 [16] (equivalent to ASTM A36 Steel), whose nominal
 Y.B. Kwon, H.S. Park / Journal of Constructional Steel Research 67 (2011) 1212–1224
(a) R-60-2.
Fig. 5. Elastic and inelastic buckling stress versus half-wavelength curves.
yield stress and ultimate tensile stress were 235 MPa and 400 MPa,
respectively. Tensile coupons were cut from the sub-panels of
the fabricated specimens and tested to determine the material
properties. Referring to the coupon test results, the stress–strain
relations displayed the typical characteristics of mild steel as
follows: a yield plateau, strain-hardening and large elongation of
approximate 24.0%. The average test yield stress of the material
obtained by the tensile coupon tests was 270.0 MPa, which was
higher than the nominal yield stress of 235.0 MPa by 14.9% and
average tensile stress was 430.0 MPa, which was higher than the
nominal strength by 7.5%. The test yield stress was used in the FE
analysis.
2.4. Geometric imperfections
Initial imperfections are mainly generated due to the weld-
ing process in attaching longitudinal stiffeners to the base plates.
Initial geometric imperfection data are essential for developing
representative initial conditions for numerical investigations. Im-
perfections were measured using calibrated laser displacement
sensors. The lasers were programmed to take readings approxi-
mately every 20 mm along the length of the specimen. Two sets
of readings were taken for each line. Imperfections were mea-
sured along the center line of the sub-panels and the tip of the
longitudinal stiffeners. Initial geometric imperfections are simply
classified as out-of-flatness for the sub-panel, camber deflection
and straightness of the stiffeners. Out-of-flatness is a localized de-
formation of sub-panels, the camber deflection measured along
the tip line of the longitudinal stiffeners means the out-of-plane
deflection of the stiffened plates and the straightness of the stiffen-
ers indicates the horizontal deflection of the tip line of the longitu-
dinal stiffeners. The camber deflections measured were relatively
larger than the other imperfections. The camber directly affects
the distortional buckling mode of longitudinally stiffened plates.
The out-of-flatness and camber deflections are summarized in Ta-
ble 2 but detailed imperfection data are presented in Kwon and
Park (2010) [17]. Referring to the measured initial imperfections in
Table 2, the out-of-flatness of the sub-panel was relatively smaller
than the out-of-plane cambers and therefore, might have a less sig-
nificant effect on buckling and ultimate strengths than the initial
camber deflections. The measured out-of-flatness ranged from 1.3
to 2.6 mm, which were smaller than the values of w/100 allowed
by ANSI/AASHTO/AWS [18]. The measured camber at the center
of longitudinal stiffener ranged from −1.6 to −5.5 mm, which
were generally larger than those specified in the AASHTO [2] and
KHBDS [19]. The AASHTO [2] and KHBDS [19] specify the tolerance
limit of camber as l/480 and l/1000, respectively. The negative sign
1215
(b) R-90-2.
of a camber deflection indicates the fact that the camber direction
was opposite to the attached side of the longitudinal stiffeners.
Initial imperfections should be included in FE analysis. How-
ever, the measured imperfections could not be included easily.
Therefore, the initial imperfection amplitude needed to be deter-
mined by an adequate formula. The assumed imperfection ampli-
tude in distortional buckling mode for FE analysis was based on
the Walker (1975) [20], which was proposed for cold-formed steel
sections, and calibrated to the measured data. The modified Walker
equation is given by
 0.5
Fy
d0 = 0.5t (2)
Fcrd
where Fcrd = critical elastic distortional buckling stress obtained by
FE analyses. The factor 0.5 was used instead of 0.3 in the Walker
equation. According to Eq. (2), a smaller camber imperfection is
calculated for a specimen with a large distortional buckling stress.
However, the measured camber was not proportional to the square
root of distortional buckling stress slenderness.
3. Finite element analysis
3.1. Buckling analysis of longitudinally stiffened plates
In order to predict the elastic local and distortional buckling
stresses of the stiffened plates tested in the fixed end boundary
conditions, a spline finite strip buckling analysis program SFSA [21]
was used. The lowest three elastic buckling stresses and modes
of the test specimens in the fixed end boundary conditions as
the test condition are summarized in Table 3. Generally, the
critical buckling stress (first buckling stress) of the most stiffened
plates, except R-70-1 and R-90-2, was local buckling. However,
distortional buckling mode or mixed mode between distortional
buckling and local buckling was found within the three lowest
buckling modes, except R-60-2 and R-70-2, where the fifth and
the fourth buckling mode were the first mixed buckling modes.
Generally, for the test specimens, the differences between the first
and the second buckling stress of the test specimens were very
small and those between the second and the third buckling ranged
approximately from 10% to 15% of first buckling stress. Even if the
lowest distortional or the mixed mode buckling stress was slightly
higher or lower than local buckling stress, since the local buckling
and distortional buckling modes generally have a significant post-
buckling strength reserve, it was expected that, regardless of
critical buckling mode, all the test sections would undergo buckling
interaction between local buckling and distortional buckling.
1216 Y.B. Kwon, H.S. Park / Journal of Constructional Steel Research 67 (2011) 1212–1224

Table 2
Measured initial imperfections (mm).
Specimens Out-of-flatness AWS (=w/100) Camber deflection KHBDS (=L/1000) AASHTO (=L/480) Eq. (2)

R-60-1 1.5 2.4 −2.6 1.00 2.08 1.90

R-70-1 2.2 2.8 −3.5 1.00 2.08 2.11
R-80-1 2.5 3.2 −1.6 1.00 2.08 2.41
R-90-1 1.3 3.6 −3.3 1.00 2.08 2.71
R-100-1 1.6 3.0 −4.3 1.00 2.08 2.98
R-110-1 2.6 4.4 −5.1 1.00 2.08 3.22
R-120-1 2.1 4.8 −3.8 1.00 2.08 3.55
R-60-2 1.3 2.4 −2.2 1.20 2.50 1.94
R-70-2 2.6 2.8 −2.9 1.20 2.50 2.28
R-80-2 1.6 3.2 −5.5 1.20 2.50 2.65
R-90-2 1.3 3.6 −1.8 1.20 2.50 2.78
R-100-2 2.5 3.0 −4.6 1.20 2.50 3.08
R-110-2 2.5 4.4 −3.8 1.20 2.50 3.42
R-120-2 3.6 4.8 −5.3 1.20 2.50 3.72

Table 3
Elastic buckling stress of longitudinally stiffened plates.
Specimens 1st buckling 2nd buckling 3rd buckling
Stress (MPa) Mode Stress (MPa) Mode Stress (MPa) Mode

R-60-1 260.0 L(4) 262.2 L(5) 294.9 D(1) + L(3)

R-70-1 214.8 D(1) + L(3) 217.9 L(3) 241.6 L(3)
R-80-1 163.6 L(3) 165.5 L(3) 196.9 D(2) + L(4)
R-90-1 130.3 L(3) 135.6 D(1) 149.2 D(1) + L(3)
R-100-1 107.6 L(3) 108.4 L(3) 126.4 D(1) + L(3)
R-110-1 89.5 L(3) 91.8 L(3) 105.9 D(1) + L(3)
R-120-1 75.7 L(2) 79.9 L(3) 91.4 D(1) + L(3)
R-60-2 261.4 L(5) 262.9 L(6) a
326.9 D(1) + L(3)
b
R-70-2 193.1 L(4) 214.9 L(3) 206.6 D(1)
R-80-2 143.2 L(4) 146.7 D(1) + L(3) 147.8 D(1)
R-90-2 133.5 D(1) + L(4) 134.2 L(3) 153.4 L(3)
R-100-2 103.6 L(3) 110.4 D(1) + L(3) 124.9 D(1) + L(3)
R-110-2 90.6 L(3) 93.4 D(1) 102.7 D(1) + L(3)
R-120-2 77.4 L(3) 79.4 D(1) 86.6 D(1) + L(3)
a
5th buckling stress.
b
4th buckling stress.

3.2. FE analysis of longitudinally stiffened plates
The material and geometrical nonlinear analysis of the speci-
mens selected for the compression test was conducted using the
program LUSAS [22], to investigate the ultimate strength and the
structural behavior of the specimens. The 4-node thick shell ele-
ment (QTS4) was used for the numerical modeling of the stiffened
plate, and the loaded end was assumed as a fixed boundary condi-
tion with the vertical direction of the loaded end free to move to
allow vertical application of a load. The uniform displacement con-
trol technique was used at the loaded end to set the same boundary
conditions as the compression test. The bottom end of the stiffened
plates was assumed also as a fixed boundary condition as the test
condition. Since the average yield stress obtained from the tensile
tests of the coupons, which were cut from the sub-panels of the
stiffened plates, was approximately 270 MPa, the yield stress of
the material was taken as 270 MPa in the FE analysis of the spec-
imens, which was slightly higher than the nominal yield stress of
235 MPa. Young’s modulus was assumed to be 2.05 × 105 MPa
and Poisson’s ratio was taken as 0.3. The stress–strain relation of
the material was assumed elastic–perfectly plastic neglecting the
strain-hardening, and the von-Mises yield criterion was applied for
the plasticity theory of the material.
Initial imperfections and residual stresses significantly affect
the ultimate strength of stiffened plates. Initial geometric im-
perfections were included in the nonlinear FE analysis as an
amplification of the critical local buckling mode. The maximum
amplitude was assumed as w/100 (w : width of sub-panels) [18].
The equivalent preloads corresponding to the assumed imperfec-
tions in the distortional mode of the stiffened plates based on the
modified Walker equation in Eq. (2) were applied at the center of
the longitudinal stiffeners.
Typical deformed shapes for the R-60-1 with two longitudinal
stiffeners and the R-90-2 with a single stiffener obtained by FE
analysis are illustrated in Fig. 6. The interaction between local
buckling at short half-waves and distortional buckling at overall
length was clearly observed in the nonlinear FE analysis results.
The axial stress versus shortening relations of the stiffened plates
obtained by the nonlinear analysis is shown in Fig. 7. The FE
analysis results displayed a post-buckling strength before the
peak load in spite of significant buckling interaction. All the
stiffened plates showed a little or less post-buckling strength
reserve after local buckling. Referring to the buckling stress and
the ultimate stress in Fig. 7, the post-buckling strength reserve in
the local mode increased according to the increase of the width-
to-thickness ratio of the sub-panels. For the R-60-1 and the R-80-1
sections, numerical instability problem due to significant buckling
interaction occurred after peak load. The FE results are compared
with test results in the following sections in detail.
4. Compression tests
4.1. General
The longitudinally stiffened plates listed in Table 1 were
chosen for compression tests. A pseudostatic compression test was
conducted by using a 5000 kN UTM (Shimadzu Co., Ltd.) shown
in Fig. 8(a). Downward loading was applied at 0.01 mm/s by the
displacement control method. Vertical displacement was obtained
from the testing machine directly and horizontal displacements
 Y.B. Kwon, H.S. Park / Journal of Constructional Steel Research 67 (2011) 1212–1224 1217

Fig. 6. Deformed shapes of R-60-1 and R-90-2.

were fixed ends, where the deformation due to warping was also
restrained by the attached bearing plates of 20 mm in thickness.

4.2. Test results

After initial take up, the axial shortening of the section

Fig. 7. Axial stress versus shortening curves by FE analysis.
were measured by the displacement transducers, which were
attached at the center and at the quarter point of the longitudinal
stiffener and the sub-panel on the opposite side attached to the
longitudinal stiffener of test specimens as shown in Fig. 8(b). Since
local buckling occurred in several half-waves, strain gauges were
attached at the center and near ends of the back side of the sub-
panels where the peak of the local buckling wave was expected.
The boundary conditions at both ends of the test specimens
was linearly increased according to the load increment before
local or distortional buckling. As the load was increased, for
most test sections, local buckling occurred first as expected and
a significant post-buckling strength was observed before the
commencement of distortional buckling. Then, finally buckling
interaction between local and distortional buckling occurred, and
the test section collapsed mainly in the distortional mode as shown
in Fig. 9. However, for some sections such as R-60 and R-70,
distortional buckling occurred first, then local buckling followed
and interaction between them occurred before the peak load.
For the specimens such as R-100-1 and R-120-1, which were
composed of sub-panels with a large width-to-thickness ratio, only
the distortional buckling mode was observed in the final collapse
stage. Those sections failed mainly in the distortional mode rather
than the interaction mode. For all the test series-2 sections, the
buckled shape of the stiffened plates indicated mixed mode of
local buckling in three short half-waves and distortional buckling
in the section length half-wave. The buckled shapes could be
observed clearly in Figs. 9(d)–(i). This mixed mode was similar

(a) Test Set-up. (b) Locations of LVDTs.

Fig. 8. Compression test configurations.

1218 Y.B. Kwon, H.S. Park / Journal of Constructional Steel Research 67 (2011) 1212–1224

(a) R-60-1. (b) R-90-1. (c) R-100-1.

(d) R-60-2. (e) R-80-2. (f) R-90-2.

(g) R-100-2. (h) R-110-2. (i) R-120-2.

Fig. 9. Failure mode of test specimens.

 Y.B. Kwon, H.S. Park / Journal of Constructional Steel Research 67 (2011) 1212–1224 1219

(a) Test series-1. (b) Test series-2.

Fig. 10. Axial load versus shortening curves.

(a) Center of sub-panel at mid-height (DT2). (b) Longitudinal stiffener at mid-height (DT3).

(c) Longitudinal stiffener at quarter-height (DT4).

Fig. 11. Axial load versus lateral displacement curves.

to that obtained by the numerical analysis, as shown in Fig. 6.
Referring to Fig. 9, it could be concluded that the final failure mode
of the stiffened plates with longitudinal stiffeners was mainly the
distortional buckling mode or the interaction mode between local
buckling and distortional buckling.
Curves of the axial load versus shortening or horizontal
displacement are shown in Figs. 10 and 11, respectively. The
axial shortening of the test sections in Fig. 10 was measured
by the testing machine directly. The displacement in Fig. 11
was the horizontal displacement measured by the LVDTs, which
was located at the center of a longitudinal stiffener, at the
center of sub-panel and at the quarter point of the longitudinal
stiffeners of the specimens. As shown in Figs. 10 and 11, the
longitudinally stiffened plates displayed a post-buckling behavior
and gradual unloading occurred after the ultimate load. Referring
to Fig. 10, as the load was increased, the longitudinally stiffened
plates shortened elastically and rounded gradually after local
or distortional buckling on the load versus shortening curves.
For most sections of which the width-to-thickness ratio of sub-
panels was larger than 90, at or near the theoretical elastic local
buckling load, local buckling occurred first, as expected, within
the range of approximate 10%–15% of local buckling stress, and
a significant post-local-buckling strength reserve was observed
before distortional buckling. However, the R-60, R-70 and R-80
1220 Y.B. Kwon, H.S. Park / Journal of Constructional Steel Research 67 (2011) 1212–1224
sections seemed to show inelastic local buckling and for some
sections, distortional buckling was observed first and local
buckling followed.
For all of the sections, distortional buckling of one half-wave
was dominant. However, due to the fixed boundary condition, the
R-70-2 and the R-120-2 showed a different trend in displacements
of DT3 and DT4. After the peak load, the direction of displacements
was reversed, as shown in Figs. 11(a)–(c). For the R-90-2, the
directions of displacements of DT2 and DT3 were different
at the beginning, but after distortional buckling occurred, the
displacements were increased in the same direction. This indicated
that buckling interaction had occurred. For all of the sections for
test series-2, a significant interaction between local buckling and
distortional buckling was observed before the maximum load was
reached. After the peak load, there was no sudden drop of load, but
the horizontal displacement continued to increase rapidly as the
load was decreased slowly.
4.3. Comparison of experimental and FE results
Both FE analysis and test results showed more or less significant
interaction between local buckling and distortional buckling
according to the section geometry. While some test sections such
as R-100-1, R-110-1 and R-120-1 showed only the distortional
buckling deformations, numerical results of them showed a slight
interaction between local buckling mode and distortional buckling
mode. This was due to the imposed initial imperfections in the local
buckling mode. The FE and experimental results are compared in
Table 4. Elastic local and distortional buckling stresses obtained by
the SFSA were included for comparison. Referring to the ultimate
stresses in Table 4, the test and analysis results are quite reliable.
The ultimate stresses of the specimens with small width-to-
thickness ratio were lower than the elastic distortional buckling
stress but those of the specimens with large width-to-thickness
ratio were higher than the elastic distortional buckling stress. The
latter result is due to the fact that the post-buckling strength
reserve for the plates of large width-to- thickness ratio was much
greater than that for the sections of low width-to-thickness ratio
in either local or distortional buckling mode. The ratios of the test
results to the FE results ranged from 0.79 to 0.96 with an average
value of 0.90. The difference in the ultimate stress between the
test results and the FE results might be resulted from the fact that
the initial imperfections and residual stresses assumed in the FE
analysis were not the same as the measured values.
5. Design strength
5.1. Current design strength curves for stiffened plates
AASHTO LRFD specifications (2007) [2] provide the resistance
of stiffened plates in compression for three distinct regions based
on the width-to-thickness ratio of the sub-panel accounting for
local buckling of the sub-panels. For slender plates, the theoretical
elastic local buckling stress formula is given by
kπ 2 E
 2
t counted for by the current specifications except Eurocode3 [3].
Fcr = . (3)
12(1 − ν 2 ) w Therefore, in order to predict an accurate ultimate strength of stiff-
For stocky plates, full yielding can be achieved. In the transition
zone between elastic local buckling and yielding of the material,
the transition curve is defined by a sine curve, which is simplified
by
     
R2 − wt
Fy
1
 π kE
Fcr = Fy 1 − 1 − sin  
2 2 R2 − R1
where R1 and R2 are the slenderness limits for yielding and elastic
buckling, respectively. The k value in Eq. (4) accounts for the
stiffness of the longitudinal stiffeners and ranges from 1 to 4.
However, the specifications provide the k values for the case of one
or two longitudinal stiffeners in practical use.
KHBDS (2005) [19] provides similar local buckling strength
formulas to AASHTO LRFD specifications [2], which predicts more
conservative strengths than the AASHTO LRFD specifications [2].
The strength formulas are given by
Fcr = Fy (λl ≤ 0.5) (5a)
Fcr = (1.5 − λl )Fy (0.5 < λl ≤ 1.0) (5b)
0.5Fy
Fcr = (λl > 1.0) (5c)
λ2l
where

w Fy 12(1 − ν 2 )
λl = . (5d)
t kπ 2 E
Choi and Yoo (2005) [1,24] found that the AASHTO provisions [2]
give unconservative strengths in the inelastic range and recently
proposed a simple local buckling strength curve in the inelastic
range based on the CRC column buckling strength curve as
1  w 2 w
[ ] 
Fcr = Fy 1 − ≤ Cc (6a)
2Cc2 t t
where

2π 2 E
Cc = . (6b)
3(1 − ν 2 )Fy
Eurocode3 (2006) [3] stipulates a different method from
AASHTO [1] and KHBDS [18], in which the resistance of stiff-
ened plates is based on the effective area. Eurocode3 [3] essen-
tially considers the individual stiffener strut separately based on
the column theory, which consists of a stiffener with the cor-
responding width of the flange plate, to have reduction factor
χc . It also considers the orthotropic plate theory to have a re-
duction factor ρ for the smeared equivalent orthotropic plate.
The resultant reduction factor of stiffened plates is interpolated
between ρ and χc . Therefore, Eurocode3 [3] can consider both
local buckling and distortional buckling of stiffened plates rea-
sonably. However, the computation procedures for the effec-
tive areas and reduction factors of stiffened plates are fairly
complex.
As mentioned above, the design provisions in current spec-
ifications can account appropriately for the elastic local buck-
ling and inelastic local buckling, which are dependent on the
width-to-thickness ratios. However, current specifications except
Eurocode3 [3] cannot account for the stiffness of longitudinal stiff-
eners in the design strength of stiffened plates. In other words, the
distortional buckling of longitudinally stiffened plates cannot be
considered by the current specifications. Furthermore, the inter-
action of local buckling and distortional buckling cannot be ac-
ened plates undergoing local buckling, distortional or their inter-
active buckling, a simple and designer-friendly method needs to be
developed.
5.2. Direct strength method proposed

  (4) The ultimate strength formula for cold-formed steel compres-
 sion members at the intermediate half-wavelength, which failed
 Y.B. Kwon, H.S. Park / Journal of Constructional Steel Research 67 (2011) 1212–1224 1221

Table 4
Comparison of FE and test results.
Specimens Max. stress Elastic distortional buckling stress (MPa) Elastic local buckling stress (MPa) Test/FE result
Tests (MPa) FE results (MPa)

R-60-1 170.2 189.1 294.9 260.0 0.90

R-70-1 153.6 159.8 214.8 217.9 0.96
R-80-1 142.5 163.5 196.9 163.6 0.86
R-90-1 139.5 154.1 135.6 130.3 0.91
R-100-1 130.6 140.4 126.4 107.6 0.93
R-110-1 119.6 132.6 105.9 89.5 0.91
R-120-1 113.9 128.4 91.4 75.7 0.89
R-60-2 195.4 214.7 326.9 261.4 0.91
R-70-2 171.6 192.5 206.6 193.1 0.89
R-80-2 159.1 180.8 147.8 143.2 0.88
R-90-2 175.3 186.5 133.5 134.2 0.94
R-100-2 164.7 179.1 110.4 103.6 0.92
R-110-2 158.5 170.4 93.4 90.6 0.93
R-120-2 136.0 151.1 79.4 77.4 0.91
Mean 0.91
SD 0.023

in the distortional mode, was first proposed by Kwon and Han-

cock (1992) [6]. The formulas have been a seed in the direct
strength method (DSM) developed by Schafer and Pekoz [25] to
account for distortional buckling. The DSM has been implemented
in cold-formed carbon steel codes including the NAS Appendix 1
(AISI 2004) [26] and AS/NZS 4600(2005) [27]. The distortional
buckling limiting stress curves were the empirical formulas based
on tests and the elastic distortional buckling stress calculated by
the rigorous FEM or FSM program, and they were used to predict
the ultimate strength of compression members undergoing distor-
tional buckling. The distortional buckling limiting stress formulas
are given by
Fkh = Fy for λd ≤ 0.561 (7a)
  0.6   0.6
Fcrd Fcrd
Fkh = Fy 1.0 − 0.25 for λd > 0.561 (7b)
Fy Fy
Fig. 12. Comparison of distortional buckling limiting stress curve and tests.
where

Fy The distortional buckling limiting stress formulas need mod-
λd = (7c) ification to predict the distortional buckling limiting strength of
Fcrd
welded longitudinally stiffened sections undergoing distortional
Fcrd : elastic distortional buckling stress buckling or interaction distortional buckling and local buckling.
Fy : nominal yield stress. The modified distortional buckling stress formulations are given
The elastic distortional buckling stresses in Eqs. (7a) and (7b) by
can be calculated by the program SFSM [7] or general FE
Fnd = Fy for λd ≤ 0.587 (8a)
programs. The predictions by the distortional strength equations  0.55   0.55
are compared with the test results in Fig. 12. Referring to the

Fcrd Fcrd
predicted strengths and test results in Fig. 12, the distortional Fnd = Fy 0.88 − 0.18
Fy Fy
strength formulas gave a conservative prediction of the ultimate
strength of some sections with or without slight interaction for λd > 0.587. (8b)
of local and distortional buckling. However, Eqs. (7a) and
(7b) predict unconservatively the distortional strength of the The distortional buckling limiting stress curves are compared with
specimens whose distortional buckling slenderness ratio factor test results in Fig. 13. Previous test results by Choi et al. [23] where
ranged from 0.9 to 1.3, in comparison with test ultimate the width-to-thickness ratios of the sub-panels ranged from 30
strengths. One of the main reasons for the unconservative to 60 and the inertia moment of the longitudinal stiffeners were
prediction by the strength formulas (7a), (7b) is that they larger than Eq. (1) are also included in Fig. 13 for comparison.
were developed based on the tests of cold-formed sections The comparison shows that the predictions of the ultimate
undergoing distortional buckling or slight interaction between strength of test specimens by Eqs. (8a) and (8b) are reasonable
local buckling and distortional buckling. Actually, local buckling
in comparison with the test results except for two specimens.
has an adverse effect on the ultimate strength of sections when
For those specimens whose distortional buckling slenderness is
local buckling and distortional buckling occur simultaneously or
nearly simultaneously and fail in a distortional buckling mode. near 1.0, the predictions of the ultimate strength by Eq. (8b) are
Another important reason is that the buckling behavior of welded slightly unconservative. However, referring to the comparison of
sections is affected significantly by the residual stresses resulting test results and distortional buckling limiting stress curve in Fig. 13,
from the welding process, while residual stresses are negligible for it can be concluded that Eqs. (8a) and (8b) can be effectively
cold-formed steel sections. The magnitude of initial imperfections used for the prediction of the ultimate strength of longitudinally
might be also an important reason for the unconservative stiffened plates undergoing distortional buckling that interact with
prediction. local buckling and failing in distortional buckling mode.
1222 Y.B. Kwon, H.S. Park / Journal of Constructional Steel Research 67 (2011) 1212–1224
Fig. 13. Comparison of tests and modified strength formulas.
Fig. 14. Comparison of local buckling limiting stress curve and tests.
For comparison, the local buckling limiting stress curve for
welded section columns proposed by Kwon et al. [28] is compared
with test results. The resulting limiting stress Fnl for local buckling
of welded section columns is given by
Fnl = Fy for λl ≤ 0.816
  0.5   0.5
Fcrl Fcrl
Fnl = 1 − 0.15 Fy for λl > 0.816
Fy Fy
where
λl = Fy /Fcrl

Fcrl : elastic local buckling stress.
The resulting limiting stress equations are achieved simply by
replacing nominal column strength Fne based on overall column
buckling by yield stress Fy . The predictions by the local limiting
stress equations are compared with the test results and the NAS
strength curve [26] where nominal column strength Fne is replaced
by Fy in Fig. 14. Referring to Fig. 14, the local limiting stress
curve predicted the ultimate strengths of longitudinally stiffened
plates tested whose local buckling slenderness is smaller than
approximately 1.4 unconservatively in comparison with the test
results. The NAS curve is more unconservative than Eqs, (9a) and
(9b) and gave unconservative predictions for most test specimens
except three specimens in test series-2.
Referring to the comparison between resulting limiting stress
curves and test results in Figs. 13 and 14, the modified distortional
buckling limiting stress curve can predict the ultimate strength
of test specimens undergoing local and distortional buckling
simultaneously more accurately than the local buckling limiting
stress curve. Therefore, it can be concluded that distortional
buckling rather than local buckling governs the ultimate strength
of longitudinally stiffened plates when local and distortional
buckling occur simultaneously or nearly simultaneously. However,
for the practical use of the local and distortional limiting stress in
design of longitudinally stiffened plates, the limits of the width-
to-thickness ratio and inertia moment of longitudinal stiffeners
should be provided more accurately.
5.3. Comparison with current specifications and tests
To compare the strengths predicted by the DSM and the current
specifications directly, the strength equations should be expressed
in terms of load. The ultimate strength of longitudinally stiffened
plates can be taken as the nominal strength (Pnd ), which is obtained
by multiplying the distortional limiting stress (Fnd ) in Eqs. (8a) and
(8b) by the gross section area (A). The alternative formulations of
Eqs. (8a) and (8b) in terms of load are given by
Pnd = Py for λd ≤ 0.587 (10a)
  0.55   0.55
Pcrd Pcrd
Pnd = Py 0.88 − 0.18
Py Py
for λd > 0.587 (10b)
where

Py
λd = (10c)
Pcrd
Pcrd : elastic distortional buckling load (= Fnd × A)
Py : nominal yield strength (= Fy × A).
Referring to the previous test results conducted by Choi et al.
[23], the longitudinally stiffened plates satisfying the width-to-
thickness ratio limits for the sub-panels and flexural rigidity
requirement for the longitudinal stiffeners specified in AASHTO
LRFD specifications, failed in an interaction buckling mode.
Therefore, direct comparison of the tests and the strengths
predicted by the DSM proposed and the current specifications
for the longitudinally stiffened plates composed of more slender
(9a) sub-panels and less rigid longitudinal stiffener than the minimum
values, respectively, is also worthwhile. The comparison between
test results and predictions obtained by the current specifications
(9b)
such as AASHTO LRFD specifications [2] and Eurocode3 [3], CRC
type strength curves proposed by Choi and Yoo [24] and the DSM
proposed in Eqs. (9a) and (9b) are compared in Table 5.
Referring to the predicted strengths and test results in Table 5,
(9c)
both proposed DSM strength curves and Eurocode3 seemed to
provide reasonably conservative predictions for the resistance
of the longitudinally stiffened plates in comparison with test
results except the R-60-1. For the R-60-1 sections, the DSM gave
slightly unconservative prediction but Eurocode 3 produced fairly
conservative strength in comparison with test result. The main
difference between DSM and Eurocode3 for the two test series
resulted from the fact that the edge stiffeners were assumed
to have yield stress for Eurocode3 but the limiting distortional
buckling stress for the DSM. Since the longitudinally stiffened
plates of the two test series had different edge stiffeners in
thickness and different ratio of the edge stiffener area to the gross
area of longitudinally stiffened plates might produce inconsistent
differences between the two test series. The DSM produced more
conservative predictions for the slender sections such as the
R-70-1, R-80-1, R-90-1, R-90-2, R-100-2, R-110-2 and R-120-2
than Eurocode3 but for the R-60-1, R-60-2, R-70-2 and R-80-
2 sections, vice versa. Since based on only the local buckling
stress, the AASHTO specifications and the CRC curves gave much
more conservative strengths for the longitudinally stiffened plates
 Y.B. Kwon, H.S. Park / Journal of Constructional Steel Research 67 (2011) 1212–1224 1223

Table 5
Comparison of tests, current specifications and DSM (kN).
Specimens w/t Tests DSM AASHTO CRC KHBDS Eurocode3

R-60-1 60.0 713.4 722.9 917.7 734.4 448.7 568.2

R-70-1 70.0 541.6 515.9 533.2 510.6 269.0 550.6
R-80-1 80.0 546.1 501.2 443.8 443.8 224.2 536.0
R-90-1 90.0 580.1 496.9 380.5 380.5 192.0 523.5
R-100-1 100.0 585.8 518.8 332.4 332.4 167.8 514.7
R-110-1 110.0 575.4 512.1 294.6 294.6 149.2 508.9
R-120-1 120.0 585.1 509.5 264.0 264.0 133.6 504.9
R-60-2 60.0 1029.8 949.3 1119.9 956.5 542.3 832.0
R-70-2 70.0 979.8 854.8 863.4 853.6 431.7 793.5
R-80-2 80.0 958.7 776.1 712.2 712.2 356.1 762.8
R-90-2 90.0 867.4 613.0 463.9 463.9 231.7 756.2
R-100-2 100.0 869.8 596.1 399.4 399.4 199.7 747.4
R-110-2 110.0 887.0 583.6 349.5 349.5 174.7 739.1
R-120-2 120.0 803.6 569.8 309.9 309.9 155.0 737.5

composed of more slender sub-panels than for the R-70-1 and R-
70-2 sections, but for the R-60-1 and R-60-2 sections, AASHTO gave
unconservative predictions. The KHBDS predicted the strengths
of all the test sections too conservatively. Consequently, it can
be concluded that the current design specifications such as
AASHTO and KHBDS based on the local buckling stress, cannot
predict the compressive resistance of longitudinally stiffened
plates showing distortional buckling or interaction of local
buckling and distortional buckling but Eurocode 3 and DSM
proposed can give reasonable predictions of the compressive
resistance of longitudinally stiffened plates with distortional
buckling or interaction between local buckling and distortional
buckling.
6. Conclusions
A series of compression tests on longitudinally stiffened plates
were conducted to study the ultimate strength and buckling
behavior of these plates. Stiffened plates with sub-panels of
various width-to-thickness ratios and longitudinal stiffeners of
various flexural rigidities made of mild steel SM400 (nominal
yield stress, Fy = 235 MPa) plates were tested to failure.
In the case of sections where local buckling was the critical
mode, local buckling interacted with distortional buckling before
peak of load and failure. However, for some sections where
distortional buckling was the dominant buckling mode, there
was no interaction between local and distortional buckling. The
design strength formula developed by Kwon and Hancock (1992)
for cold-formed steel sections was modified slightly to account
for the different characteristics of the longitudinally stiffened
plates. Simple strength formulas for the direct strength method
were proposed for the longitudinally stiffened plates undergoing
distortional buckling or interaction between local buckling and
distortional buckling. From the experimental study, the following
conclusions are drawn:
1. Longitudinally stiffened plates undergoing interaction between
local buckling and distortional buckling showed a significant
post-buckling strength reserve, regardless of the dominant
buckling mode, either distortional or mixed of local and
distortional buckling.
2. The interaction between local buckling and distortional buck-
ling slightly reduced the ultimate strength of the longitudinally
stiffened plates which failed in the distortional mode despite
the post-buckling strength reserve.
3. The Eurocode3 (2006) can give a reasonable prediction of
ultimate strength for longitudinally stiffened plates undergoing
distortional buckling or interaction between local buckling
and distortional buckling. However, AASHTO LRFD (2007), the
KHBDS (2005) and the CRC type curves (Choi and Yoo, 2005)
based on the local buckling stress cannot properly predict
the ultimate strength of stiffened plates with distortional
buckling.
4. The modified strength formulas for the direct strength method
can reasonably predict the ultimate strength of longitudinally
stiffened plates undergoing distortional buckling or interaction
between local buckling and distortional buckling, and failing
mainly in the distortional buckling mode.
Acknowledgements
This paper forms a part of the project entitled ‘‘Development of
Direct Strength Method for Welded Sections’’ being conducted in
the Department of Civil Engineering at the Yeungnam University.
This research was supported by the 2010 research fund of National
Research Foundation of Korea.
References
[1] Choi BH, Hwang MH, Yoon TY, Yoo CH. Experimental study of inelastic
Buckling strength and stiffness requirements for longitudinally stiffened
panels. Engineering Structures 2009;31:1141–53.
[2] AASHTO LRFD bridge design specifications. 4th ed. Washington (DC, USA):
American Association of State Highway and Transportation Officials, Inc.;
2007.
[3] European Committee for Standardisation (ECS). Eurocode 3: Design of steel
structures, part 1–5: Plated structural elements, Brussels, Belgium. 2006.
[4] Hancock GJ. Local, distortional and lateral buckling of I-beams. Journal of
Structural Division, ASCE 1978;104(11):1787–98.
[5] Hancock GJ. Distortional buckling of storage rack column. Journal of Structural
Division, ASCE 1985;111(12):2770–83.
[6] Kwon YB, Hancock GJ. Strength design curves for Thin-Walled sections
undergoing local and distortional Buckling. Journal of Structural Engineering,
ASCE 1992;118(7):1786–803.
[7] Standards Australia, Cold-Formed Steel Structures AS/NZS 4600: Sydney
(NSW, Australia); 1996.
[8] American Iron and Steel Institute (AISI), Design of cold-formed steel structural
members, Washington (DC, USA); 2001.
[9] Yang D, Hancock GJ. Compression tests of high strength steel channel columns
with interaction between local and distortional Buckling. Journal of Structural
Engineering, ASCE 2004;130(12):1954–63.
[10] Kwon YB, Kim BS, Hancock GJ. Compression tests of high strength cold-
formed steel channels with buckling interaction. Journal of Constructional
Steel Research 2009;65:278–89.
[11] Standards Australia, Cold-formed steel structures AS/NZS 4600: Sydney (NSW,
Australia); 2005.
[12] American Iron and Steel Institute (AISI). Appendix 1 design of cold-formed
steel structural members with the direct strength method, Washington (DC,
USA); 2004.
[13] American Iron and Steel Construction (AISC). Load and resistance factor design
specification for steel structural buildings, Chicago (Il, USA); 2005.
[14] Kwon YB. Buckling analysis program (BAP) (ver 2.0). Korea: Yeungnam
University; 2000.
[15] Yoo CH, Choi BH, Ford EM. Stiffness requirements for longitudinally stiffened
Box-girder flanges. Journal of Structural Engineering, ASCE 2001;127(6):
705–11.
[16] Korean Standards, KS D3515; Welded Structural Steel, 2005.
[17] Park HS. Master of science thesis, department of civil and environmental
engineering. Gyongsan (Korea): Yeungnam University; 2010.
1224 Y.B. Kwon, H.S. Park / Journal of Constructional Steel Research 67 (2011) 1212–1224
[18] ANSI/AASHTO/AWS D1.5: 2008 Bridge Welding Code. Washington (DC):
American Association of State Highway and Transportation Officials and
American Welding Society; 2008.
[19] Korean Association of Construction and Transportation Officials, Korean
Highway Bridge Design Specifications, Seoul (Korea); 2005.
[20] Walker AC. Design and analysis of cold-formed sections. London: International
Textbook Company Ltd; 1975.
[21] Law SCW, Hancock GJ. Buckling of thin Flat-Walled structures by a spline finite
strip method. Thin-Walled Structures 1986;4(4):269–94.
[22] FEA Co., Ltd., Lusas Element Reference Manual & User’s Manual (version
14.3–7), 2010.
[23] Choi BH, Hwang MO, Park CH. Experimental study on inelastic buckling
behaviour of stiffened plates in compression, proceedings, 3rd international
symposium on steel structures, Seoul (Korea); 2005, p. 89–96.
[24] Choi BH, Yoo CY. Strength of stiffened flanges in horizontally curved Box
Girders. Journal of Engineering Mechanics, ASCE 2005;131(2):167–76.
[25] Schafer BW, Pekoz T. Direct strength prediction of cold-formed steel members
using numerical elastic Buckling solutions. In: Shanmugan NE, Liew JYR,
Thevendran V, editors. Thin-Walled structures, research and development.
Elsevier; 1998. p. 137–44.
[26] American iron and steel institute, North American specifications for the design
of cold-formed steel structural members, Supplement No. 1, Washington (DC,
USA); 2004.
[27] Standards Australia, Cold-formed steel structures AS/NZS 4600: 2005, Sydney
(NSW, Australia); 2005.
[28] Kwon YB, Kim NK, Hancock GJ. Compression tests of Welded section columns
undergoing Buckling interaction. Journal of Constructional Steel Research
2007;63:1590–602.