ARTICLE IN PRESS

Thin-Walled Structures 47 (2009) 190– 201

Contents lists available at ScienceDirect

Thin-Walled Structures
journal homepage: www.elsevier.com/locate/tws

Buckling capacities of axially loaded, cold-formed, built-up C-channels

Jessica Whittle, Chris Ramseyer 
University of Oklahoma, 202 W. Boyd, Room 334, Oklahoma City, OK 73019, USA

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

Article history: Cold-formed, built-up members are common compression elements in cold-formed steel joists, and
Received 12 July 2007 these built-up members are susceptible to unique buckling behaviors. Built-up member design is
Received in revised form addressed in Section C4.5 of the American Iron and Steel Institute 2001 Specification. Over 150
19 May 2008
experimental compression tests on closed-section, built-up members formed of intermediately welded
Accepted 24 May 2008
Available online 29 August 2008
c-channels were conducted, and these experimental values were compared to theoretical buckling
capacities based on the Section C4.5 modified slenderness ratio. Use of the modified slenderness ratio
Keywords: was exceedingly conservative. Capacities based on the unmodified slenderness ratio and C4.5 fastener
Built-up sections and spacing provisions were consistently conservative.
Cold-formed compression members
Published by Elsevier Ltd.
Modified effective length
Column shear

1. Introduction
As the demand for light-weight steel structures continues to
rise, efficient and accurate design of cold-formed steel elements is
essential. One frequently used cold-formed steel member is a
built-up member, formed by two or more attached steel elements.
Because of built-up members’ unique characteristic to buckle
under load, either as one single member or two individual
members, a specific provision for these members exists. The
design of cold-formed steel, built-up compression members is
addressed in Section D1.2 of the 2007 edition of the American Iron
and Steel Institute (AISI) North American Specification for the Design
of Cold Formed Steel Structural Members [1]. It is the purpose of this
research to verify the accuracy of Section D1.2 of the Specification
for various, intermediately welded, closed-section built-up mem-
bers in pure axial compression. Limited research has been
conducted in the area of cold-formed, built-up members, and
even less investigation exists in the area of closed-section, built-
up members with welded attachments.
2. Literature review
The 2007 AISI Specification Section D1.2 [1] specifies that built-
up members be designed with a modified slenderness ratio if
shear forces are induced between the weld or screw connectors.
The section also introduces a minimum fastener strength
requirement and a fastener spacing requirement for built-up
 Corresponding author. Tel.: +1 405 325 1415; fax: +1 405 325 4217.
E-mail address: ramseyer@ou.edu (C. Ramseyer).
members [1]. Refer to Eqs. (1) and (2) for the modified slenderness
ratio and intermediate fastener spacing provision as stated in AISI
Specification D1.2:
  sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
  2
KL KL 2 a
¼ þ . (1)
r m r o ri
AISI D1.2 equation (Eq. D1.2-1)—modified slenderness ratio [1].
 
a KL
p0:5 (2)
ri r o
AISI C4.5 fastener spacing provision [2].
where (KL/r)o is the overall slenderness ratio of entire section
about the built-up member axis; a is the intermediate fastener or
spot weld spacing; ri is the minimum radius of gyration of the full
unreduced cross-sectional area of an individual shape in a built-
up member; K is the effective length factor; and L is the unbraced
member length.
However, the modified slenderness ratio is heavily adopted
from the American Iron and Steel Construction (AISC) code for
hot-rolled, built-up members [2]. This is inconsistent with the fact
that the behavior of hot-rolled steel members is considerably
different than the behavior of cold-formed members, which are
often controlled by local or distortional buckling [3].
The research performed on hot-rolled built-up members is the
foundation of built-up member research. Zandonini’s [4] research
of stitched built-up struts led to the addition of the slenderness
modification equation into the 1986 AISC Specification, and in
1991 Aslani and Goel [5] verified the modified slenderness ratio
analytically and experimentally for hot-rolled members, thereby
verifying the AISC built-up member design method. Research in
1998 on bolted double-angle compression members by Sherman

0263-8231/$ - see front matter Published by Elsevier Ltd.

doi:10.1016/j.tws.2008.05.014
 ARTICLE IN PRESS

J. Whittle, C. Ramseyer / Thin-Walled Structures 47 (2009) 190–201 191

and Yura [6] concluded that preventing shear slip in the end
connections of built-up members was a critical factor for strength.
More recently, Lue et al.’s [7] research of hot-rolled, axially loaded,
built-up, double-channel columns reaffirmed the AISC column
design strengths with the modified slenderness ratio to be
generally conservative. Table 1 presents the research achieve-
ments in built-up compression members over the last 22 years,
including research in both hot-rolled steel and cold-formed steel
built-up members.
Not until 2005 was the modified slenderness ratio for cold-
formed built-up members experimentally investigated. Stone and
LaBoube’s [8] research of cold-formed, built-up I-sections con-
structed from steel studs supported AISI Specification Section D1.2
as conservative. They concluded that the slenderness modification
was on average conservative for thin members and exceedingly
conservative for thick members (greater than 0.89 mm (0.035 in.)).
The members tested were wide, thin c-channels intermediately
connected back to back with screws to model a typical, cold-
formed, I-shaped wall stud.
Additional work of cold-formed built-up c-channels was
conducted by Brueggen and Ramseyer [3] on smaller c-channels
in open- and closed-sections with intermediate, welded
stitch attachments. Their research on the built-up stub
columns concluded that the AISI design methods to be con-
servative for compact members but often unconservative for
‘‘members with slender elements.’’ Brueggen and Ramseyer [9]
recommended that additional tests be performed to determine
the effects of length and location (double- or single-sided),

Table 1
Research in built-up compression members

Researcher Steel type Topic Conclusions Year

published

Sukumar et al. [10] Cold-formed Buckling behavior of I-shaped built-up AISI design strengths found slightly conservative 2006
members formed with angles or equal to the actual strength. Developed design
curve for built-up columns undergoing
distortional or local-distortional buckling

Stone and LaBoube [8] Cold-formed Behavior of cold-formed steel built-up I- AISI Section D1.2 (slenderness modification) is 2005
sections (studs) formed with c-channels conservative on average for thin members and
and screw attachments exceedingly for thick members

Brueggen and Ramseyer [3] Cold-formed Buckling of closed and open- built-up AISI Section D1.2 on average conservative for 2005
sections with channels and compact sections but potentially unconservative
intermediately welded attachments for members with slender elements

Lue et al. [7] Hot-rolled Axial capacity of built-up columns with AISC column design strengths with the modified 2004
double channel sections (weld and bolt slenderness ratio found conservative
attachments)

Sherman and Yura [6] Hot-rolled Bolted double-angle compression Preventing shear slip in the end connection is a 1998
members critical factor for strength, and five intermediate
connections will achieve full elastic buckling of
the member

Aslani and Goel [5] Hot-rolled Analytical and experimental verification Testing verified the AISC slenderness modification 1991
of the modified slenderness ratio ratio and built-up member design method

Zandonini [4] Hot-rolled Stitched built-up struts Developments supported the addition of the 1985
slenderness modification equation into the 1986
AISC Specification

Fig. 1. Open I-shaped built-up latticed columns [10].

 ARTICLE IN PRESS
192 J. Whittle, C. Ramseyer / Thin-Walled Structures 47 (2009) 190–201
spacing, and number of weld attachments on the behavior of
welded built-up members.
In contrast, Sukumar et al.’s [10] investigation of the buckling
behavior of open, I-shaped built-up members attached as latticed
columns and formed of cold-formed, angle sections concluded
that the AISI design strengths were on average slightly conserva-
tive or equivalent to the actual strengths. Fig. 1 displays how the
built-up members were assembled with angle sections and lacing
bars. A design curve for the maximum strength of cold-formed
built-up columns experiencing distortional buckling or mixed
local-distortional buckling was also presented.
3. Experimental testing
3.1. Test objective and description
An experimental investigation of over 150 cold-formed, built-
up c-channels loaded in compression was conducted at the
University of Oklahoma’s Fears Structural Engineering Laboratory.
All specimens were tested in pure axial compression with pinned
end conditions. The primary objective was to find maximum
buckling capacities for a variety of built-up members to determine
the accuracy of the AISI 2007 Specification’s design methods for
built-up members. In addition, the effects of built-up member
characteristics, such as member thickness, member geometry,
column length, and location and number of intermediate
weld attachments were explored to obtain a broad-range of
Fig. 2. Single c-channel.
Fig. 3. Built-up member.
25.4 mm (1 in)
midpoint stitch weld
50.8 mm (2 in)
end weld
Fig. 4. Complete built-up member with midpoint attachment.
experimental data that could more thoroughly represent built-up
c-channel buckling behaviors.
Three repetitions of each member type were tested to obtain
reliable average buckling values, and single c-channel tests served
as a baseline measure for each built-up member type. A total of
153 experimental tests were conducted and 55 specimen types
investigated. Eight of the 55 specimen types had fewer than three
repeated specimens tested. The experimental testing focused on
closed-section built-up members, as this configuration provides
exceptional torsional resistance, and limited research has been
performed in this built-up member area. Not all of the built-up
members tested met the fastener spacing provisions present in
AISI Specification Section D1.2. The purpose of experimentally
testing built-up members that did and did not meet the provision
was to better relay the accuracy of Section D1.2 for properly
designed, built-up compression members. In addition, the type of
built-up members were chosen to provide experimental verifica-
tion of specific built-up section behavior for the research sponsor,
who was interested in constructability limitations and not
necessarily interested in the D1.2 limitations.
3.2. Specimen properties
The built-up members were created with two, lipped c-
channels welded together at the top and bottom with 50.8 mm
(2 in.) long welds and at intermediate locations along the member
with 25.4 mm (1 in.) long welds to form a closed box-section. All
welds were approximately 4.76 mm (0.1875 in.) thick. Figs. 2 and 3
display a single c-channel and a closed, built-up member,
respectively. Fig. 4 illustrates a complete built-up member with
a typical midpoint intermediate weld. All intermediate welds
were equally spaced along the length of the member.
The thicknesses and geometries of the c-channels were
variables of the experiment, chosen to reflect the section proper-
ties of common, built-up c-channel web members used in cold-
formed trusses. Lipped c-channels with nominal thicknesses of
1.626 mm (0.064 in.), 2.032 mm (0.080 in.), and 2.54 mm
(0.100 in.) were investigated. Different c-channel geometries,
specifically the flange and web lengths of 41.275 mm (1.625 in.)
and 66.675 mm (2.625 in.), were tested. All c-channels were
square-shaped. Fig. 5 indicates the typical channel geometries.
The number and location of intermediate weld attachments
and length of the member represented the diversity of
built-up members. Midpoint, third-point, and sixth-point welds
were all tested in addition to welding on a single side of the
member or both sides of the members. The efficient, single-sided
welding has recently become more prevalent as robotic fabrica-
tion processes for built-up members have grown in popularity.
Member lengths of 1803 mm (71 in.) and 1397 mm (55 in.) were
also researched.
3.3. Test set-up
The test rig was constructed based on typical, column
compression apparatuses. Two greased swivel pivots modeled
50.8 mm (2 in) end
weld
 ARTICLE IN PRESS

J. Whittle, C. Ramseyer / Thin-Walled Structures 47 (2009) 190–201 193

9.525 mm 47.625 mm

22.225 mm
9.525 mm

66.675 mm

41.275 mm

66.675 mm
41.275 mm

Fig. 5. C-channel geometries.

the pinned end conditions. The testing rig, in combination with a

hand pump, was designed so that an operator could slowly apply
compressive load on the upright member.
Buckling capacities, axial shortening, and lateral displacement
of the specimens during testing were the measured data from the
laboratory testing. A 445 kN (100 kip) load cell measured the
applied loads on the specimen. Strain gauges were strategically
placed at third points along the member’s length to characterize
the member’s lateral movement during testing while another
strain gauge recorded the axial shortening of the column under
loading. A data acquisition system was utilized to gather the raw
data, and the LabView program graphically displayed this data
during loading.
Initial testing determined that a safety restraining mechanism
was necessary to combat the instability of the test specimens. A
large steel sleeve was fabricated to enclose the bottom pivot to
prevent it from slipping and damage. Bolted-chain restraints were
fabricated and attached to each tested specimen to limit damage if
the test specimen became unstable. Figs. 6 and 7 display the test
rig with a specimen in place.

3.4. Test procedure

Prior to fabrication, the geometric properties of each specimen

were measured to guarantee that the c-channels met AISI material
tolerances, and coupon tests were performed for the three
material thicknesses to verify the material properties of the steel.
Due to the sensitivity of cold-formed member testing to geometric
and loading imperfections, it was essential that these measures
were taken to ensure that all the members tested met standard
size tolerances and thereby minimize the potential experimental
errors caused by geometric defects. The c-channel dimensions (lip
lengths and side lengths) and thicknesses were measured at both
ends of the all the c-channels using calipers and a micrometer,
and any c-channels that did not meet member tolerances were
discarded and not used for testing.
Fig. 6. Side view: test rig without safety chains.
Fabricating the specimens followed the process of welding the
channels at the top and bottom, cutting the channels to the
desired lengths, and welding the 25.4 mm (1 in.) long intermedi- and ensure a distributed loading gradient on the member. Finally,
ate attachments for each specimen type. In some cases, the ends safety chains were affixed to the top and bottom of each specimen
of the built-up member were milled to remove any imperfections with screws.
 ARTICLE IN PRESS

194 J. Whittle, C. Ramseyer / Thin-Walled Structures 47 (2009) 190–201

The built-up member specimen was centered in the test rig
and aligned so that the applied load would be axial. This step was
critical in minimizing loading imperfections that are a sensitivity
concern for experimental testing of cold-formed members.
The safety chains attached to the specimen were then bolted into
the frame behind the test rig to restrain the specimen from
possible instability at failure. Lastly, the lateral-displacement wire
potentiometers were magnetically attached to the specimen at
third points, and the LVDT monitoring axial shortening was
positioned directly below the base plate of the swivel pivot. At a
distance, the operator gradually applied the load on the specimen
using a hand pump while simultaneously viewing the applied
load and axial shortening of the specimen on the data acquisition
machine. The final buckling of the specimen was typically
apparent by the peak of the load versus axial shortening curve
or the sudden failure of the specimen under loading. In addition to
the data collected on the data acquisition machine, other forms of
documentation, such as photographs, a lab notebook, and often
the measured transverse displacements between the c-channels
were maintained throughout the course of testing.

4. Test results

The maximum buckling capacity Ptest was the largest tested

axial load the built-up member could withstand before failure.
The AISI 2007 Specification Section C4 column design method in
combination with the D1.2 built-up member slenderness mod-
ifications were followed to calculate the nominal loads of the
members, Pn. Torsional and flexural–torsional buckling modes
were considered as per the Specification, and the effective areas of

Load vs. Displacement

DW2 - double sided, two welds located at third points
3
1803 mm (71") long, 41.275 mm (1.625") geometry,
2.54 mm (0.100") thick members
250

200
Load (kN)

150

100

50

0
0 2 4 6 8 10 12
Displacement (mm)

Fig. 7. Front view: test rig with safety restraints to control specimen. Fig. 8. Typical load versus displacement chart for a built-up member type.

Table 2
Intermediate attachment orientations and descriptions

Welded attachment orientation Description

Single

Double

SW1

DW1

SW2

DW2

SW5

DW5

 ¼ 200 end welds,  ¼ 100 intermediate welds

 ARTICLE IN PRESS

J. Whittle, C. Ramseyer / Thin-Walled Structures 47 (2009) 190–201 195

Table 3
Axial capacities—maximum buckling loads and AISI nominal loads (1397 mm length) (55 in.)

Length (mm) (in.) Width (mm) (in.) Thickness (mm) (in.) Weld type Axial capacity (kN) (kip) Ptest/Pn

Average Ptest (kN) (kip) AISI Pn (kN) (kip)

1397 (55) 41.275 (1.625) 2.032 (0.080) Single 33.98 (7.64) 19.04 (4.28) 1.78
1397 (55) 41.275 (1.625) 2.032 (0.080) Double 104.1 (23.40) 63.52 (14.28) 1.64
1397 (55) 41.275 (1.625) 2.032 (0.080) SW 1 141.9 (31.89) 102.8 (23.11) 1.38
1397 (55) 41.275 (1.625) 2.032 (0.080) DW 1 175.3 (39.41) 97.46 (21.91) 1.80
1397 (55) 41.275 (1.625) 2.032 (0.080) SW 2 136.8 (30.75) 113.1 (25.42) 1.21
1397 (55) 41.275 (1.625) 2.032 (0.080) DW 2 174.4 (39.20) 107.3 (24.13) 1.62
1397 (55) 41.275 (1.625) 2.032 (0.080) SW 5 137.4 (30.88) 119.7 (26.91) 1.15
1397 (55) 41.275 (1.625) 2.032 (0.080) DW 5 164.4 (36.96) 115.6 (25.98) 1.42

Table 4
Axial capacities—maximum buckling loads and AISI nominal loads (1803 mm (71 in.) length, 41.275 mm (1.625 in.) width)

Length (mm) (in.) Width (mm) (in.) Thickness (mm) (in.) Weld type Axial capacity (kN) (kip) Ptest/Pn

Average Ptest (kN) (kip) AISI Pn (kN) (kip)

1803 (71) 41.275 (1.625) 1.626 (0.064) Single 14.99 (3.37) 9.56 (2.15) 1.57
1803 (71) 41.275 (1.625) 1.626 (0.064) Double 56.36 (12.67) 31.09 (6.99) 1.81
1803 (71) 41.275 (1.625) 1.626 (0.064) SW 1 76.51 (17.20) 52.27 (11.75) 1.46
1803 (71) 41.275 (1.625) 1.626 (0.064) DW 1 102.9 (23.13) 50.18 (11.28) 2.05
1803 (71) 41.275 (1.625) 1.626 (0.064) SW 2 66.28 (14.90) 59.78 (13.44) 1.11
1803 (71) 41.275 (1.625) 1.626 (0.064) DW 2 144.0 (32.37) 50.98 (11.46) 2.83
1803 (71) 41.275 (1.625) 1.626 (0.064) SW 5 77.98 (17.53) 65.48 (14.72) 1.19
1803 (71) 41.275 (1.625) 1.626 (0.064) DW 5 131.2 (29.49) 59.92 (13.47) 2.19

1803 (71) 41.275 (1.625) 2.032 (0.080) Single 24.11 (5.42) 14.28 (3.21) 1.69
1803 (71) 41.275 (1.625) 2.032 (0.080) SW 1 119.3 (26.82) 62.99 (14.16) 1.89
1803 (71) 41.275 (1.625) 2.032 (0.080) DW 1 143.2 (32.20) 62.99 (14.16) 2.27
1803 (71) 41.275 (1.625) 2.032 (0.080) SW 2 95.73 (21.52) 72.11 (16.21) 1.33
1803 (71) 41.275 (1.625) 2.032 (0.080) DW 2 146.4 (32.92) 72.11 (16.21) 2.03
1803 (71) 41.275 (1.625) 2.032 (0.080) SW 5 133.2 (29.95) 79.00 (17.76) 1.69
1803 (71) 41.275 (1.625) 2.032 (0.080) DW 5 142.6 (32.05) 79.00 (17.76) 1.81

1803 (71) 41.275 (1.625) 2.54 (0.100) Single 40.88 (9.19) 21.44 (4.82) 1.91
1803 (71) 41.275 (1.625) 2.54 (0.100) Double 116.6 (26.21) 44.35 (9.97) 2.63
1803 (71) 41.275 (1.625) 2.54 (0.100) SW 1 144.8 (32.54) 74.91 (16.84) 1.93
1803 (71) 41.275 (1.625) 2.54 (0.100) DW 1 206.2 (46.36) 74.91 (16.84) 2.75
1803 (71) 41.275 (1.625) 2.54 (0.100) SW 2 161.8 (36.38) 85.85 (19.30) 1.89
1803 (71) 41.275 (1.625) 2.54 (0.100) DW 2 221.6 (49.81) 85.85 (19.30) 2.58
1803 (71) 41.275 (1.625) 2.54 (0.100) SW 5 182.2 (40.95) 94.08 (21.15) 1.94
1803 (71) 41.275 (1.625) 2.54 (0.100) DW 5 223.7 (50.28) 94.08 (21.15) 2.38

the built-up members were calculated using the Cold-Formed
Sections (CFS) software with applied loads equivalent to the
maximum recorded test loads. The 66.675 mm (2.625 in.) chan-
nels resulted in effective areas that differed from the gross areas
with the maximum experimental load applied. This was attrib-
uted to the wider channels’ susceptibility to local and distortional
buckling.
The descriptions specify the configuration of the member;
‘‘single’’ refers to a single c-channel, and ‘‘double’’ refers to a built-
up member attached only at the ends. The symbols SW and DW
indicated single-sided welds and double-sided welds, respec-
tively, while the numerals represent the number of intermediate
attachments on the sides welded. For example, DW5 represents a
built-up member intermediately attached with five welds on both
sides of the member, spaced at sixth-points along the specimen.
Table 2 displays the various member configurations. Charts were
created for each built-up member type indicating the relationship
between the applied compressive load and the axial shortening of
the member. Each built-up member type had three specimens
tested. Fig. 8 is a typical load versus displacement chart; each
curve represents one member test. Tables 3–5 include the
experimental axial capacities and AISI analytical axial capacities
based on the modified slenderness ratio.
5. Data analysis
5.1. Analysis of AISI specification D1.2—the modified slenderness
ratio
The test results of the 41.275 mm (1.625 in.) wide, 2.032 mm
(0.080 in.) thick, and 1397 mm (55 in.) long built-up members
were compared to the experimental buckling capacities of
Brueggen and Ramseyer [3], who performed tests on identical
members using a slightly different test rig. The members were
placed in a test frame with a loading beam at the top and a heavy
beam underneath providing the base reaction; a hydraulic
cylinder, load cell, and LVDT were affixed above the upright
member and a pivot at the base to ‘‘accommodate any out-of-true
of the specimen ends’’ [9]. Experimental results were on average
within 5% of Brueggen and Ramseyer’s buckling loads; the
maximum difference was 14%. This verified the acceptable
performance of the test set-up and supported the resulting
experimental buckling capacities.
The primary goal of the test results was to validate the
effectiveness of using the modified slenderness ratio in D1.2 to
calculate axial buckling capacities. Fig. 9 illustrates the effects of
built-up member properties, such as member thickness, width,
 ARTICLE IN PRESS

196 J. Whittle, C. Ramseyer / Thin-Walled Structures 47 (2009) 190–201

Table 5
Axial capacities—maximum buckling loads and AISI nominal loads (1803 mm (71 in.) length, 66.675 mm (2.625 in.) width)

Length (mm) (in.) Width (mm) (in.) Thickness (mm) (in.) Weld type Axial capacity (kN) (kip) Ptest/Pn

Average Ptest (kN) (kip) AISI Pn (kN) (kip)

1803 (71) 66.675 (2.625) 1.626 (0.064) Single 43.37 (9.75) 16.86 (3.79) 2.57
1803 (71) 66.675 (2.625) 1.626 (0.064) Double 128.4 (28.87) 82.34 (18.51) 1.56
1803 (71) 66.675 (2.625) 1.626 (0.064) SW 1 143.9 (32.34) 119.4 (26.84) 1.20
1803 (71) 66.675 (2.625) 1.626 (0.064) DW 1 171.8 (38.62) 109.3 (24.58) 1.57
1803 (71) 66.675 (2.625) 1.626 (0.064) SW 2 152.4 (34.26) 126.2 (28.36) 1.21
1803 (71) 66.675 (2.625) 1.626 (0.064) DW 2 177.5 (39.91) 118.7 (26.69) 1.50
1803 (71) 66.675 (2.625) 1.626 (0.064) SW 5 161.2 (36.23) 128.2 (28.83) 1.26
1803 (71) 66.675 (2.625) 1.626 (0.064) DW 5 176.6 (39.69) 124.7 (28.03) 1.42

1803 (71) 66.675 (2.625) 2.032 (0.080) Single 64.90 (14.59) 25.80 (5.80) 2.51
1803 (71) 66.675 (2.625) 2.032 (0.080) Double 185.9 (41.80) 112.3 (25.24) 1.66
1803 (71) 66.675 (2.625) 2.032 (0.080) SW 1 205.3 (46.15) 150.8 (33.91) 1.36
1803 (71) 66.675 (2.625) 2.032 (0.080) DW 1 225.8 (50.76) 144.1 (32.40) 1.57
1803 (71) 66.675 (2.625) 2.032 (0.080) SW 2 195.9 (44.05) 165.0 (37.09) 1.19
1803 (71) 66.675 (2.625) 2.032 (0.080) DW 2 229.1 (51.51) 152.8 (34.34) 1.50
1803 (71) 66.675 (2.625) 2.032 (0.080) SW 5 215.6 (48.46) 163.3 (36.72) 1.32
1803 (71) 66.675 (2.625) 2.032 (0.080) DW 5 240.2 (54.00) 155.7 (35.01) 1.54

1803 (71) 66.675 (2.625) 2.54 (0.100) Single 96.48 (21.69) 38.21 (8.59) 2.53
1803 (71) 66.675 (2.625) 2.54 (0.100) Double 280.9 (63.14) 135.9 (30.56) 2.07
1803 (71) 66.675 (2.625) 2.54 (0.100) SW 1 330.6 (74.31) 208.4 (46.86) 1.59
1803 (71) 66.675 (2.625) 2.54 (0.100) DW 1 409.2 (92.00) 189.9 (42.70) 2.15
1803 (71) 66.675 (2.625) 2.54 (0.100) SW 2 352.3 (79.19) 218.4 (49.10) 1.61
1803 (71) 66.675 (2.625) 2.54 (0.100) DW 2 434.2 (97.60) 200.9 (45.16) 2.16
1803 (71) 66.675 (2.625) 2.54 (0.100) SW 5 345.8 (77.74) 230.8 (51.88) 1.50
1803 (71) 66.675 (2.625) 2.54 (0.100) DW 5 384.6 (86.45) 220.3 (49.52) 1.75

Average Buckling Capacities of All Built-Up Members

Effect of Built-Up Member Properties on Strength, Ptest/Pn (based on {KL/r}m, {KL/r}o, and Test Values)
(based on {KL/r}m) 300
3.00
Average Buckling Load (kN)

250
2.50
200

2.00
150
Ptest/Pn

1.50 100
P n based on (KL/r)m
1.00 50 P n based on (KL/r)o
P test
41.275 mm, 1397 mm long
0.50 41.275 mm, 1803 mm long
0
1.626 2.032 2.54
66.675 mm, 1803 mm long
0.00 Material Thickness (mm)
1.626 2.032 2.54
Material Thickness (mm) Fig. 10. Relationship between the using modified slenderness ratio, unmodified
slenderness ratio, and test values.
Fig. 9. Material thickness versus Ptest/Pn (based on modified slenderness ratio).

predictions based on the modified slenderness ratio. In addition,

and length, on the strength ratio of the member. The ratio of
Ptest/Pn indicates how conservative the AISI 2007 Specification is
for designing built-up columns using the modified slenderness
ratio in D1.2. There was a trend for the design of columns using
the average modified slenderness ratio to be more conservative
for thicker members than thinner members. This upward trend of
material thickness compared to the ratio of experimental capacity
to nominal capacity followed the pattern first mentioned in Stone
and LaBoube’s [8] research of built-up studs. This trend between
material thickness and degree of AISI capacities being conserva-
tive was also dually noted in Brueggen and Ramseyer’s [3]
research.
Comparison of the seven various built-up member configura-
tions indicated that built-up members with shorter widths (more
stubby sections) result in more conservative AISI nominal strength
the axial buckling capacities based on the modified slenderness
ratio were more conservative for longer columns than shorter
columns, as seen in the comparison between the 1803 mm (71 in.)
and 1397 mm (55 in.) specimens.
The relationship between calculating axial capacities based on
the modified slenderness ratio, based on the unmodified slender-
ness ratio, and the pure experimental test values is displayed in
Fig. 10. The average buckling capacities for each member thickness
were compared. The method using the modified slenderness ratio
was consistently conservative. It was 72.5% conservative on
average when compared to the experimental buckling capacities
for all built-up members tested, and it was 77% conservative on
average for all built-up members meeting the provisions of
Section D1.2. The percentage of being conservative increased as
member thickness increased. This conservative percentage was
 ARTICLE IN PRESS

J. Whittle, C. Ramseyer / Thin-Walled Structures 47 (2009) 190–201 197

based on both double- and single-sided attachments. The average
axial capacities based on the modified slenderness ratio of Section
D1.2 were 60% conservative on average for all 1.626 mm (0.064 in.)
thick members tested, 64% conservative on average for all
2.032 mm (0.080 in.) thick members tested, and 107% conservative
on average for all 2.54 mm (0.100 in.) thick member tested. There
were no Ptest/Pn values below 1.
Similarly, the axial capacities based on the unmodified
slenderness ratio were also very conservative on average. Use of
the unmodified slenderness ratio would represent ignoring
Section D1.2 for built-up members and designing a member
solely on the column design specification. Overall, the method
using the unmodified slenderness ratio was on average 46%
conservative for all built-up members tested, and it was 65%
conservative on average for all built-up members meeting the
provisions of Section D1.2. The degree of the design method based
on the unmodified slenderness ratio being conservative also
increased with the increased thickness of the member. The
average axial capacities using the unmodified slenderness ratio
were 34% conservative on average for all 1.626 mm (0.064 in.)
thick members tested, 44% conservative on average for all
2.032 mm (0.080 in.) thick members tested, and 72% conservative
on average for all 2.54 mm (0.100 in.) thick member tested.
The axial capacities based on the unmodified slenderness ratio
were unconservative for four specimens, three of the double
Effects of C4.5 Fastener Spacing Provision, "a", Ptest/Pn
(based on the unmodified slenderness ratio, {KL/r}o)
2.50
2.00
Ptest/Pn
c-channels with no intermediate connections (which do not meet
the D1.2 spacing provision) and one single-sided, third-point
welded specimen. Fig. 11 indicates the effectiveness of using the
D1.2 spacing provision (solid-colored data points ‘‘meet provi-
sion,’’ which is the restriction of the distance ‘‘a’’ between
intermediate connectors) in conjunction with using the method
based on the unmodified slenderness ratio. The four specimens
that were unconservative using the method based on the
unmodified slenderness ratio are examined in greater detail in
Fig. 12.
The four built-up members that are unconservative when
compared to the unmodified slenderness axial capacities may
have been more susceptible to distortional and local buckling
effects because of their very thin wall thicknesses of 1.625 mm
(0.064 in.) or, in the case of the 1397 mm (55 in.) long specimen
with 2.032 mm (0.080 in.) thick channels, a shorter column length
that is less likely to buckle in pure flexural buckling. The double
built-up members have no intermediate attachments, and these
members often buckled as separate c-channels. This may have led
to unequal loading between the two c-channels. A single c-
channel could experience a greater buckling load as an individual
column than each channel would have experienced if the built-up
column buckled as one, unified member. The single-sided, third-
point welded built-up member was very close to unity at 0.98;
however, the observed buckling mode during failure was pure
flexural buckling without obvious indications of distortional or
local buckling. Additional tests of this particular built-up member
type should be explored in the future to verify the results of this
specific single-sided, third-point welded specimen.
The average axial capacity based on the modified slenderness
ratio was approximately 26% more conservative on average than
the average axial capacity based on the unmodified slenderness
ratio for all members tested. However, if all the fastener and
spacing provisions for built-up members in D1.2 were followed

and the unmodified slenderness ratio used, the method based on

the unmodified slenderness ratio was then only 12% less
1.50

1.00

0.50
1.626 2.032 2.54
Material Thickness (mm)

Fig. 11. Effect of C4.5 spacing provision on Ptest/Pn (based on unmodified

slenderness ratio).

Unconservative Tests with the Unmodified

Slenderness Ratio, Ptest/Pn (based on {KL/r}o)
1.00
41.275 mm, 1397 mm long, fails provision
SW2
0.98 41.275 mm, 1803 mm long, fails provision
66.675 mm, 1803 mm long, fails provision
0.96 41.275 mm, 1803 mm long, meets provision

0.94
Ptest/Pn

0.92
0.90
double
0.88
0.86
double
0.84
double
0.82
1.626 2.032 2.54 Front view: flexural-torsional Close-up of the specimen
Material Thickness (mm) buckling

Fig. 12. Unconservative tests with the unmodified slenderness ratio, Ptest/Pn. Fig. 13. Flexural–torsional buckling of single c-channel.
 ARTICLE IN PRESS

198 J. Whittle, C. Ramseyer / Thin-Walled Structures 47 (2009) 190–201

conservative than the average axial capacity based on the
modified slenderness ratio and unconservative for only one
specimen, the single-sided third-point welded specimen (SW2)
noted in Fig. 12. Fig. 10 displays that the difference between the
results using the unmodified or modified slenderness ratio stays
fairly constant.
5.2. Qualitative analysis of intermediate weld patterns and buckling
modes
conservative estimates of the axial capacities for built-up
members with attachments on both sides.
Single c-channels buckled consistently in the flexural–tor-
sional or torsional buckling mode. However, strong-axis buckling
of the c-channel became more pronounced with increased
material thickness. Mixed buckling modes involving distortional
buckling and other buckling modes occurred with the wide and
thin channels, 66.675 mm (2.625 in.) wide and 1.626 mm

Finally, the test results indicated general conclusions about

number and location of intermediate attachments and the
attachments’ effect on the maximum buckling capacities and
modes of buckling. The most robust attachment configuration was
difficult to assess because the strongest built-up section wavered
between the double-sided, sixth-point welds and third-point
welds (DW5 and DW2) for a given member geometry. Often, the
double-sided sixth-point welded and third-point welded built-up
members had similar buckling capacities. This would indicate that
an upper capacity limit is being reached with the double-sided
third-point (DW2) welds. The single-sided, sixth-point welded
built-up members (SW5) generally had the highest buckling
capacities for the single-sided members on average, which would
indicate that with single-sided welds, the third-point weld
locations (SW2) do not provide an upper capacity limit.
Angled view: prying of channels
There is no provision within the Specification to account for the
strength increase of double-sided welds compared to single-sided
welds. Tables 3–5 indicate that the built-up sections with double-
sided attachments consistently achieved higher buckling capa-
cities than their single-sided attached counterparts. Double-sided
welds provide better cross-section stability at the attachment
location and a very high single element torsional restraint. cross-section view at mid-span
Because double-sided attachments are not addressed in the
Specification, the modified slenderness ratio produced extremely Side view: flexural buckling

Fig. 15. Side view: flexural–torsional buckling of SW5, 41.275 mm (1.625 in.) wide
built-up member.

Back view: prying of c-channels

Close-up: opposing c-channels

cross-section view at mid-span

Front view: flexural-torsional
Side view: double member cross-section view at mid-span buckling
(separate channel flexural buckling)
Fig. 16. Front and back view: flexural–torsional buckling of SW5, 41.275 mm
Fig. 14. Double member buckling. (1.625 in.) wide member.
 ARTICLE IN PRESS

J. Whittle, C. Ramseyer / Thin-Walled Structures 47 (2009) 190–201 199

(0.064 in.) thick. Fig. 13 displays the flexural–torsional buckling of welds. These doubly attached members typically buckled sud-
a single c-channel. Double built-up members without intermedi- denly and resulted in sharp, angular buckling shapes. The angular
ate attachments buckled in various ways, sometimes as one large buckled shape appeared as a ‘‘hinge’’ in the column, and it was
unit and often with each c-channel buckling separately. Fig. 14 most likely an exaggerated form of the buckled shape in Fig. 17
displays a double built-up member with separate channel flexural with increased deformation around the hinge weld. Separate
buckling. member buckling and lateral movement of the channels was
The most common buckling mode for the single-side welded common.
built-up members was flexural buckling. Many of the specimens The wider 66.675 mm (2.625 in.) members with double-sided
buckled in a large, fluid half-sine wave with the c-channels attachments exhibited extremely sudden failures. Little lateral
prying apart at the back of the member due to the lack of translation or buckling of the column was noticeable during
fasteners at the back of the single-sided sections. The group of loading, and the sudden buckled shape occurred almost instanta-
images in Figs. 15 and 16 depict the flexural–torsional neously at failure. These members exhibited ‘‘snap through’’
buckling and prying apart of the channels in a typical single-side failure and also had a separate movement of the individual
welded member. The 41.275 mm (1.625 in.) wide sections gen- c-channels. They also displayed buckling at unique locations not
erally buckled gradually and had very definite flexural and
flexural–torsional buckling modes; the 66.675 mm (2.625 in.)
wide sections had less-definite buckling modes and typically
buckled suddenly with a mix of flexural and other buckling
modes, such as distortional or torsional buckling. Inelastic
deformations were more common in the vicinity of the welds in
the 66.675 mm (2.625 in.) members with single-side attachments
(Fig. 17).
Built-up sections with double-sided attachments buckled in
various ways, and the behaviors were more difficult to categorize
and predict than the behaviors of the members with single-side

Front view: lower section crippling Front view: close-up

Back view: prying of c-channels Fig. 18. Front view: crippling and separate channel buckling of DW1 member.

Side view: flexural-torsional buckling

(bulge around midpoint weld) Front view: distortion around
midpoint weld

cross-section view at Side view: separate Side view: member distortion

mid-span (back view) channel movement

Fig. 17. Buckling of SW1, 66.675 mm (2.625 in.) wide built-up member. Fig. 19. Side view: crippling and separate channel buckling of DW1 member.
 ARTICLE IN PRESS

200 J. Whittle, C. Ramseyer / Thin-Walled Structures 47 (2009) 190–201

1.626 mm (0.064”) 2.034 mm (0.080”) 2.54 mm (0.100”)

thick member thick member thick member

Close-up – 1.625 mm Close-up – 2.034 mm Close-up – 2.54 mm

(0.064”) thick member (0.080”) thick member (0.100”) thick member
Fig. 20. Isolated deformation of members with double-sided attachments.

necessarily at the mid-spans of the c-channels. Buckling often 6. Conclusions

appeared as a crippling of the two-separate c-channels. Figs. 18
and 19 display crippling of individual channels in one built-up
member.
Another common buckled shape with the 66.675 mm
(2.625 in.) wide members with double-sided attachments was a
mostly unbuckled, straight column exhibiting an isolated section
of crushing, bulging, and dimpling at failure. This area was
generally located at the top, bottom, or surrounding a weld on the
built-up member. Fig. 20 identifies the crushing and bulging of
built-up members with various material thicknesses. The thickest
built-up members produced very little buckling and may have had
distortional buckling around the welds.
Generally, the larger geometry, the 66.675 mm (2.625 in.)
wide members, had less definite buckling modes and more
uncertainty with possibly mixed and distortional buckling modes.
Instances of strong-axis buckling were more common with the
66.675 mm (2.625 in.) wide members and occurred more often in
the 2.54 mm (0.100 in.) thick members. In addition, it is important
to note that many members’ buckling modes could not be
identified clearly.
 The average axial capacity based on the modified slenderness
ratio of Section D1.2 is exceedingly conservative for built-up
members, in general, greater than 70% conservative. The unity
check range for all built-up members tested that the meet D1.2
provisions is 1.11–2.83. The value of 1.0 represents a legal built-
up member according to the AISI Specification.
 Use of the modified slenderness ratio is more conservative for
longer built-up members and thicker built-up sections.
 Using the modified slenderness ratio of Specification D1.2 for
built-up members is marginally more conservative than using
the standard column procedures of C4 with the unmodified
slenderness ratio, if the built-up member fastener and spacing
provisions of D1.2 are followed.
J The axial capacities determined using the unmodified
slenderness ratio are on average 12% less conservative than
the axial capacities calculated based on the modified
slenderness ratio for all built-up members tested that meet
D1.2 provisions.
 ARTICLE IN PRESS
J. Whittle, C. Ramseyer / Thin-Walled Structures 47 (2009) 190–201
J The axial capacities calculated using the unmodified
slenderness ratio are unconservative for only one specimen
meeting the provisions of D1.2 for built-up members. The
unity check for this specimen is 0.98.
 The built-up member fastener and spacing provisions of D1.2
are effective. These provisions in combination with the
unmodified slenderness ratio are very conservative for
closed-section, intermediately welded, built-up members.
The results range from a unity of 0.98–2.5, with values greater
than 1.0 representing conservatively designed members.
Acknowledgment
I would like to sincerely thank Star Building Systems for
providing the material and sponsoring this research.
References
[1] American Iron and Steel Institute, AISI. North American specification for the
design of cold-formed steel structural members. Washington, DC: American
Iron and Steel Institute; 2007.
201
[2] American Iron and Steel Institute, AISI. Commentary on North American
specification for the design of cold-formed steel structural members.
Washington, DC: American Iron and Steel Institute; 2007.
[3] Brueggen B, Ramseyer C. Capacity of built-up cold-formed steel axial
compression members. Structural stability research council (2005). In:
Proceedings of the annual stability conference, Rolla, Missouri, 2005.
[4] Zandonini R. Stability of compact built-up struts: experimental investigation
and numerical simulation. Construzioni metalliche, No. 4, 1985.
[5] Aslani F, Goel SC. An analytical criterion for buckling strength of built-up
compression members. Eng J 1991;28(4):159–68 (American Institute of Steel
Construction, Inc., Chicago, IL).
[6] Sherman DR, Yura JA. Bolted double angle compression members. J Construct
Steel Res 1998;46(1–3):470–1.
[7] Lue D-M, Yen T, Liu J-L, Hsu YT. Experimental investigation for buckling
strength of double-channel columns. Structural stability research council
(2004). In: Proceedings of the annual stability conference, Long Beach, CA,
2004.
[8] Stone TA, LaBoube RA. Behavior of cold-formed steel built-up I-sections. Thin
Wall Struct 2005;43:1805–17.
[9] Brueggen B, Ramseyer C. Cold formed steel joist member buckling capacity
testing. Report for star building systems, Oklahoma City. Norman, Oklahoma:
Fears Structural Engineering Laboratory; 2003.
[10] Sukumar S, Parameswaran P, Jayagopal LS. Local-, distortional-, and euler
buckling of thin walled built-up open cross-sections under compression. J
Struct Eng—Madras 2006;32(6):447–54.