Aluminum Alloy Tubular Columns-Part II: Parametric Study and Design Using Direct Strength Method

ARTICLE IN PRESS
Thin-Walled Structures 44 (2006) 969985

www.elsevier.com/locate/tws
Aluminum alloy tubular columnsPart II: Parametric study and design

using direct strength method
Ji-Hua Zhu, Ben Young
Department of Civil Engineering, The University of Hong Kong, Pokfulam Road, Hong Kong
Received 23 February 2006; received in revised form 11 August 2006; accepted 18 August 2006
Available online 18 October 2006
Abstract
A parametric study of aluminum alloy columns of square and rectangular hollow sections was performed using nite element analysis
(FEA). The columns were compressed between xed ends. The parametric study included 120 columns with and without transverse welds
at the ends of the columns. An accurate and reliable nite element model was used for the parametric study. Design approaches for
aluminum alloy tubular columns with and without transverse welds were proposed. Column strengths predicted by the FEA were
compared with the design strengths calculated using the current American, Australian/New Zealand and European specications for
aluminum structures. In addition, the direct strength method (DSM), which was developed for cold-formed carbon steel members, was
used in this study for aluminum alloy columns. The design strengths calculated using the DSM were compared with the numerical results.
Furthermore, design rules modied from the DSM were proposed. It is shown that the proposed design rules accurately predicted the
ultimate strengths of aluminum welded and non-welded columns. The reliability of the current and proposed design rules was evaluated
using reliability analysis.
r 2006 Elsevier Ltd. All rights reserved.
Keywords: Aluminum alloys; Buckling; Column; Design; Finite element analysis; Heat-affected zone; Parametric study; Transverse welds
1. Introduction
Finite element analysis (FEA) has been widely used in
structural design. Compared with physical experiments,
FEA is relatively inexpensive and time efcient, especially
when a parametric study of cross-section geometry is
involved. In addition, FEA is more convenient for
investigation involving geometric imperfections of structural members, whereas this could be difcult to investigate
through physical tests. Although FEA is a useful and
powerful tool for structural analysis and design, it is
important to obtain an accurate and reliable nite element
model (FEM) prior to a parametric study of FEA to be
carried out.
Aluminum members are being used increasingly in
structural applications. The current American Aluminum
Design Manual [1], Australian/New Zealand Standard [2]
and European Code [3] for aluminum structures provide
Corresponding author. Tel.: +852 2859 2674; fax: +852 2559 5337.
E-mail address: young@hku.hk (B. Young).

0263-8231/$ - see front matter r 2006 Elsevier Ltd. All rights reserved.
doi:10.1016/j.tws.2006.08.012
design rules for compression members. The design rules in

the three specications follow the element approach that
separate the cross-section into elements for design. One
disadvantage of the element approach is that the computation procedure becomes more tedious for members having
complex cross-sections. This disadvantage can be overcome
by a new design method called the direct strength method
(DSM) that was developed by Schafer and Pekoz [4]. The
DSM was developed for cold-formed steel structures. The
test data used in the development of column design for
DSM were based on concentrically loaded pin-ended coldformed steel column for certain cross-sections and geometric limits [57]. The DSM has been adopted by the
North American Specication (NAS) [8,9] for cold-formed
steel structures. In this study, the DSM was used for the
design of aluminum alloy columns.
There are some advantages in using aluminum as a
structural material, such as high strength-to-weight ratio,
lightness, corrosion resistance and ease of production.
However, one disadvantage is that heat-treated aluminum
alloys could suffer loss of strength in a localized region
ARTICLE IN PRESS
970
J.-H. Zhu, B. Young / Thin-Walled Structures 44 (2006) 969985
Nomenclature
A
B
COV
DL
E
FEA
FEM
Fm
fy
fy-nw
fy-w
H
kc
L
LL
le
Mm
PAA
gross cross-section area

overall width of SHS and RHS
coefcient of variation
dead load
Youngs modulus
nite element analysis
nite element model
mean value of fabrication factor
material yield strength
non-welded material yield strength
welded material yield strength
overall depth of SHS and RHS
coefcient in the AS/NZS Standard
length of specimen
live load
column effective length
mean value of material factor
unfactored design strength for American Aluminum Design Manual
PAS/NZS unfactored design strength for Australian/New
Zealand Standard
Pcre
critical elastic buckling load in exural buckling,
p2EA/(le/r)2
Pcrl
critical elastic local column buckling load
PDSM column design strength calculated using the
direct strength method
PDSM-NW non-welded column design strength calculated using the modied direct strength
method
PDSM-W1 welded column design strength calculated
using the modied direct strength method (rst
approach)
when welding is involved, and this is known as heataffected zone (HAZ) softening. Previous research [10,11]
indicated that welds have signicant effect on column
strength. The test program presented by Zhu and Young
[12] showed that transverse welds at the ends of the
columns reduce the column strength for nearly 45%. The
current American Aluminum Design Manual [1], Australian/New Zealand Standard [2] and European Code [3] for
aluminum structures provide design rules for structural
members containing transverse welds.
An accurate and reliable non-linear FEM for aluminum
columns with and without transverse welds at the ends of
the columns has been developed by Zhu and Young [13].
The purpose of this paper is rstly to investigate the
behavior and design of aluminum columns using a
parametric study of FEA. The developed FEM [13] is used
for a parametric study of cross-section geometries.
Secondly, the design rules for aluminum welded and nonwelded columns are proposed based on the DSM. The
column strengths predicted by the FEA were compared
PDSM-W2 welded column design strength calculated

using the modied direct strength method
(second approach)
PEC9
unfactored design strength for Eurocode 9
PExp
experimental ultimate load of column
PFEA ultimate load predicted by FEA for parametric
study
Pm
mean value of tested-to-predicted load ratio
Pne
nominal axial strength for exural buckling
Pnl
nominal axial strength for local buckling
Pu
column strength
Py
yield strength of the section (fy A)
Py-nw yield strength of the section calculated using the
non-welded material properties (fy-nw A)
Py-w
yield strength of the section calculated using the
welded material properties (fy-w A)
r
radius of gyration of gross cross-section about
the minor y-axis of buckling
t
thickness of section
VF
coefcient of variation of fabrication factor
VM
coefcient of variation of material factor
VP
coefcient of variation of tested-to-predicted
load ratio
b
reliability index
lc
non-dimensional
slenderness
buckp p
forpexural
ling ( Py =Pcre ; Pynw =Pcre or Pyw =Pcre )

ll
non-dimensional slendernesspfor
interaction
of
local and exural buckling ( Pne =Pcrl )

local buckling coefcient specied in the EC9
rc
Code
rhaz
heat-affected zone (HAZ) softening factor specied in the EC9 Code
f
resistance factor
s0.2
static 0.2% proof stress.
with the design strengths calculated using the American

Aluminum Design Manual (AA), Australian/New Zealand
Standard (AS/NZS) and European Code (EC9) for
aluminum structures, as well as the direct strength method
(DSM) and proposed design rules. In addition, reliability
analysis was performed to assess the reliability of these
design rules.
2. Summary of test program
The test program presented by Zhu and Young [12]
provided experimental ultimate loads and failure modes of
aluminum alloy square and rectangular hollow sections
(RHS) compressed between xed ends. The test specimens
were fabricated by extrusion using 6063-T5 and 6061-T6
heat-treated aluminum alloys. The test program included
29 columns with both ends transversely welded to
aluminum end plates (welded columns), and 12 columns
without welding of end plates (non-welded columns).
The non-welded and welded material properties of the
ARTICLE IN PRESS
specimens were determined by longitudinal tensile coupon

tests. Initial overall and local geometric imperfections were
measured prior to testing. The specimens were tested
between xed ends at various column lengths ranged from
300 to 3000 mm. The column tests were performed using
displacement control. The details of the test program were
reported in Zhu and Young [12].
3. Finite element modeling
The nite element program ABAQUS [14] version 6.5
was used in the analysis for the simulation of aluminum
alloy xed-ended columns tested by Zhu and Young [12].
An accurate and reliable non-linear FEM for aluminum
welded and non-welded columns has been presented by
Zhu and Young [13]. The development of the FEM is
detailed in Zhu and Young [13]. In the FEM, the measured
cross-section dimensions, material properties and initial
geometric imperfections of the test specimens were
modeled. The xed-ended boundary condition was modeled by restraining all the degrees of freedom of the nodes
at both ends of the column, except for the translational
degree of freedom in the axial direction at one end of the
column. The nodes other than the two ends were free to
translate and rotate in any directions. The material nonlinearity was included in the FEM by specifying the true
values of stresses and strains. The plasticity of the material
was simulated by a mathematical model, known as the
incremental plasticity model, in which the true stresses and
true plastic strains were calculated in accordance with
ABAQUS [14]. The geometric imperfections were included
in the FEM by using the eigenvalue analyses. The
displacement control loading method was used in the
FEA that was identical to the loading method used in the
column tests. The S4R general-purpose shell elements were
used in the FEM. The size of the nite element mesh of
10 10 mm2 (length width) was used in the modeling of
the non-welded columns. The welded columns were
modeled by dividing the columns into different portions
along the column length. Therefore, the HAZ softening at
both ends of the columns were simulated. The size of the
nite element mesh for the welded columns is detailed in
Zhu and Young [13].
Table 1
Cross-section dimensions of the series for parametric study
Series
Type of
material
Depth
H (mm)
Width B
(mm)
Thickness t
(mm)
(H 2t=t
T5-A-NW
T5-B-NW
T5-C-NW
T5-D-NW
T5-E-NW
T5-F-NW
6063-T5
6063-T5
6063-T5
6063-T5
6063-T5
6063-T5
60
60
120
120
180
180
60
60
60
60
60
60
1.5
3.0
1.5
3.0
1.5
3.0
38.0
18.0
78.0
38.0
118.0
58.0
T6-A-NW
T6-B-NW
T6-C-NW
T6-D-NW
T6-E-NW
T6-F-NW
6061-T6
6061-T6
6061-T6
6061-T6
6061-T6
6061-T6
60
60
120
120
180
180
60
60
60
60
60
60
1.5
3.0
1.5
3.0
1.5
3.0
38.0
18.0
78.0
38.0
118.0
58.0
T5-A-W
T5-B-W
T5-C-W
T5-D-W
T5-E-W
T5-F-W
6063-T5
6063-T5
6063-T5
6063-T5
6063-T5
6063-T5
60
60
120
120
180
180
60
60
60
60
60
60
1.5
3.0
1.5
3.0
1.5
3.0
38.0
18.0
78.0
38.0
118.0
58.0
T6-A-W
T6-B-W
T6-C-W
T6-D-W
T6-E-W
T6-F-W
6061-T6
6061-T6
6061-T6
6061-T6
6061-T6
6061-T6
60
60
120
120
180
180
60
60
60
60
60
60
1.5
3.0
1.5
3.0
1.5
3.0
38.0
18.0
78.0
38.0
118.0
58.0
Note: 1 in. 25.4 mm, NW non-welded column series, W welded

column series.
For example, the label T5-A-NW-L500 denes the

following specimen:

4. Parametric study
The FEM closely predicted the experimental ultimate
loads and failure modes of the tested aluminum columns as
presented by Zhu and Young [13]. Hence, the model was
used for an extensive parametric study. The parametric
study included 120 specimens that consisted of 24 series, as
shown in Table 1. Each series contained 5 specimens with
column lengths of 500, 1200, 2000, 2700 and 3500 mm. The
specimens were labelled such that the type of aluminum
alloy, cross-section dimensions, welding condition and
column length could be identied, as shown in Tables 25.
971
The rst letter indicates the type of material of the

specimen, where T5 refers to the aluminum alloy
6063-T5, and T6 refers to the aluminum alloy 6061-T6.
The second part of the label indicates the cross-section
shape of the specimen, where A refers to a square
hollow section (SHS) with nominal cross-section dimension of 60 60 1.5 mm2. Table 1 shows the crosssection dimensions of each series using the nomenclature
dened in Fig. 1.
The following part of the label NW indicates the
welding condition of the specimen, where the letter
NW refers to the non-welded column, and the letter
W refers to the welded column.
The last part of the label L500 indicates the length of
the column, where the letter L refers to the column
length and the following digits are the nominal length of
the specimen in millimeters (500 mm).
The material properties of the specimens of 6063-T5

alloy investigated in the parametric study are identical to
the material properties of Series N-R2 in the experimental
program for the welded and non-welded material, whereas
the material properties of the specimens of 6061-T6 alloy
ARTICLE IN PRESS
972
Table 2
Comparison of FEA and design strengths for non-welded columns of aluminum alloy 6063-T5
Specimen
T5-A-NW-L500
T5-A-NW-L1200
T5-A-NW-L2000
T5-A-NW-L2700
T5-A-NW-L3500
Mean, Pm
COV, VP
Reliability index, b
T5-B-NW-L500
T5-B-NW-L1200
T5-B-NW-L2000
T5-B-NW-L2700
T5-B-NW-L3500
Mean, Pm
COV, VP
T5-C-NW-L500
T5-C-NW-L1200
T5-C-NW-L2000
T5-C-NW-L2700
T5-C-NW-L3500
Mean, Pm
COV, VP
T5-D-NW-L500
T5-D-NW-L1200
T5-D-NW-L2000
T5-D-NW-L2700
T5-D-NW-L3500
Mean, Pm
COV, VP
T5-E-NW-L500
T5-E-NW-L1200
T5-E-NW-L2000
T5-E-NW-L2700
T5-E-NW-L3500
Mean, Pm
COV, VP
T5-F-NW-L500
T5-F-NW-L1200
T5-F-NW-L2000
T5-F-NW-L2700
T5-F-NW-L3500
Mean, Pm
COV, VP
FEA
Comparison
PFEA (kN)
PFEA/PAA
PFEA/PAS/NZS
PFEA/PEC9
PFEA/PDSM
PFEA/PDSM-NW
48.3
48.1
47.5
46.6
38.4
125.4
120.7
113.6
94.7
75.4
63.2
62.7
61.9
56.0
40.5
164.0
155.3
152.1
125.5
114.6
62.1
61.3
58.3
51.8
41.7
178.9
174.2
162.8
146.2
124.9
0.92
0.91
0.90
0.92
0.88
0.91
0.019
2.57
0.97
0.98
1.04
0.97
0.93
0.98
0.039
2.83
1.21
1.20
1.25
1.38
1.19
1.25
0.063
3.67
0.95
0.90
0.89
0.82
0.85
0.88
0.057
2.29
1.20
1.18
1.49
1.62
1.55
1.41
0.146
3.21
1.05
1.02
0.95
0.91
0.92
0.97
0.061
2.66
0.92
0.91
0.90
0.92
0.88
0.91
0.019
2.35
1.09
1.05
1.04
0.97
0.93
1.02
0.060
2.65
1.21
1.20
1.25
1.38
1.19
1.25
0.063
3.47
1.01
0.95
0.93
0.82
0.85
0.91
0.084
2.06
1.20
1.18
1.49
1.62
1.55
1.41
0.146
3.05
1.11
1.08
1.01
0.91
0.92
1.01
0.088
2.42
0.95
1.00
1.07
1.19
1.25
1.09
0.116
2.29
0.99
1.01
1.07
1.09
1.20
1.07
0.078
2.51
1.16
1.19
1.24
1.20
0.98
1.16
0.089
2.72
0.99
0.99
1.05
0.98
1.13
1.03
0.062
2.45
1.12
1.13
1.12
1.04
0.90
1.06
0.090
2.40
1.02
1.04
1.03
1.00
1.00
1.02
0.016
2.61
0.91
0.93
0.96
0.96
1.01
0.96
0.039
2.73
0.98
1.01
1.02
1.09
1.04
1.03
0.061
3.04
1.18
1.21
1.29
1.28
1.06
1.20
0.074
3.43
0.96
0.94
1.00
0.99
1.03
0.98
0.036
2.86
1.15
1.17
1.19
1.15
1.06
1.14
0.045
3.46
1.01
1.02
1.02
1.01
0.99
1.01
0.010
3.06
0.90
0.93
0.96
0.98
1.05
0.96
0.111
2.65
0.98
1.01
1.02
1.09
1.04
1.03
0.039
3.04
1.05
1.08
1.16
1.17
1.00
1.09
0.068
3.09
0.96
0.95
1.02
1.02
1.03
1.00
0.037
2.92
0.95
0.97
1.00
0.98
0.92
0.96
0.033
2.78
0.95
0.96
0.98
0.98
0.98
0.97
0.012
2.88
Note: 1 kip 4.45 kN.
are identical to the material properties of Series H-R2 in

the experimental program for the welded and non-welded
material, as detailed in Zhu and Young [12,13]. The local
imperfection magnitude was 16% of the section thickness
which is equal to the mean value of the measured local
imperfection magnitudes of the tested specimens. The
overall imperfection magnitude was 1/2000 of the column
length. The size of the nite element mesh was kept at
10 10 mm2 (length width) for the non-welded columns.

The welded columns were modeled with 25 mm HAZ
extension at both ends of the columns. The nite element
mesh dimension was approximately 8.3 10 mm2 (length
width) for the HAZ regions at both ends of the welded
columns, and 10 10 mm2 (length width) for the main
body of the columns. The column strengths (PFEA) obtained
from the parametric study are shown in Tables 25.
ARTICLE IN PRESS
973
Table 3
Comparison of FEA and design strengths for non-welded columns of aluminum alloy 6061-T6
Specimen
T6-A-NW-L500
T6-A-NW-L1200
T6-A-NW-L2000
T6-A-NW-L2700
T6-A-NW-L3500
Mean, Pm
COV, VP
T6-B-NW-L500
T6-B-NW-L1200
T6-B-NW-L2000
T6-B-NW-L2700
T6-B-NW-L3500
Mean, Pm
COV, VP
T6-C-NW-L500
T6-C-NW-L1200
T6-C-NW-L2000
T6-C-NW-L2700
T6-C-NW-L3500
Mean, Pm
COV, VP
T6-D-NW-L500
T6-D-NW-L1200
T6-D-NW-L2000
T6-D-NW-L2700
T6-D-NW-L3500
Mean, Pm
COV, VP
T6-E-NW-L500
T6-E-NW-L1200
T6-E-NW-L2000
T6-E-NW-L2700
T6-E-NW-L3500
Mean, Pm
COV, VP
T6-F-NW-L500
T6-F-NW-L1200
T6-F-NW-L2000
T6-F-NW-L2700
T6-F-NW-L3500
Mean, Pm
COV, VP
FEA
Comparison
PFEA (kN)
PFEA/PAA
PFEA/PAS/NZS
PFEA/PEC9
PFEA/PDSM
PFEA/PDSM-NW
48.3
48.1
47.5
46.6
38.4
181.1
177.6
169.0
135.5
79.4
76.0
69.9
63.9
51.0
41.5
220.9
207.8
205.6
152.8
125.3
79.3
77.4
74.4
60.8
45.8
243.1
237.4
213.0
171.8
131.7
0.97
0.96
0.95
0.86
0.91
0.93
0.049
2.56
0.96
1.01
1.11
1.04
0.96
1.02
0.062
2.85
1.18
1.08
1.26
1.23
1.19
1.19
0.057
3.53
0.98
0.92
0.91
0.73
0.85
0.88
0.105
1.98
1.24
1.38
1.86
1.86
1.67
1.24
0.178
3.24
1.08
1.06
1.07
1.05
0.96
1.04
0.178
3.05
0.97
0.96
0.95
0.86
0.91
0.93
0.049
2.35
1.08
1.06
1.11
1.04
0.96
1.05
0.053
2.83
1.18
1.08
1.26
1.23
1.19
1.19
0.057
3.33
1.03
0.97
0.96
0.73
0.85
0.91
0.131
1.77
1.24
1.38
1.86
1.86
1.67
1.24
0.178
3.10
1.15
1.12
1.07
1.05
0.96
1.07
0.178
2.79
0.99
1.04
1.14
1.23
1.20
1.12
0.093
2.56
0.98
1.05
1.21
1.32
1.15
1.14
0.115
2.46
1.12
1.07
1.04
0.91
0.87
1.00
0.109
2.04
1.00
1.00
1.11
0.99
1.11
1.04
0.058
2.53
1.14
1.15
1.16
1.01
0.84
1.14
0.133
2.07
1.05
1.08
1.05
0.95
0.91
1.01
0.133
2.29
0.93
0.96
0.99
1.06
0.99
0.99
0.049
2.80
0.98
1.06
1.23
1.27
1.10
1.13
0.107
2.86
1.11
1.07
1.09
0.99
0.99
1.05
0.054
3.04
1.00
0.99
1.10
0.95
0.99
1.00
0.056
2.84
1.16
1.18
1.25
1.15
1.04
1.15
0.063
3.36
1.07
1.10
1.09
1.01
0.94
1.04
0.063
2.94
0.89
0.93
0.97
1.06
1.02
0.97
0.071
2.61
0.98
1.06
1.23
1.27
1.10
1.13
0.107
2.86
0.96
0.93
0.96
0.89
0.91
0.93
0.031
2.64
0.98
0.98
1.10
0.96
1.02
1.01
0.055
2.85
0.92
0.94
1.01
0.96
0.89
0.95
0.048
2.63
0.98
1.01
1.02
0.96
0.92
0.98
0.040
2.81
5. Design approaches
5.1. Current design rules for aluminum structures
The American Aluminum Design Manual (AA) [1],
Australian/New Zealand Standard (AS/NZS) [2] and
European Code (EC9) [3] for aluminum structures provide
design rules for aluminum columns with and without
transverse welds. The design rules in the AA Specication
for calculating the design strengths of non-welded aluminum columns are based on the Euler column strength. The
inelastic column curve, based on the tangent modulus, is
well approximated by a straight line using buckling
constants [11]. The buckling constants were obtained from
Tables 3.33 and 3.34 of Part I-B of the AA Specication.
Local buckling stress for the section as a whole was the
weighted average local buckling stress for the individual
elements, based on gross section properties, while the local
ARTICLE IN PRESS
974
Table 4
Comparison of FEA and design strengths for welded columns of aluminum alloy 6063-T5
Specimen
T5-A-W-L500
T5-A-W-L1200
T5-A-W-L2000
T5-A-W-L2700
T5-A-W-L3500
Mean, Pm
COV, VP
Reliability index,
T5-B-W-L500
T5-B-W-L1200
T5-B-W-L2000
T5-B-W-L2700
T5-B-W-L3500
Mean, Pm
COV, VP
Reliability index,
T5-C-W-L500
T5-C-W-L1200
T5-C-W-L2000
T5-C-W-L2700
T5-C-W-L3500
Mean, Pm
COV, VP
Reliability index,
T5-D-W-L500
T5-D-W-L1200
T5-D-W-L2000
T5-D-W-L2700
T5-D-W-L3500
Mean, Pm
COV, VP
Reliability index,
T5-E-W-L500
T5-E-W-L1200
T5-E-W-L2000
T5-E-W-L2700
T5-E-W-L3500
Mean, Pm
COV, VP
Reliability index,
T5-F-W-L500
T5-F-W-L1200
T5-F-W-L2000
T5-F-W-L2700
T5-F-W-L3500
Mean, Pm
COV, VP
Reliability index,
FEA
Comparison
PFEA (kN)
PFEA/PAA
PFEA/PAS/NZS
PFEA/PEC9
PFEA/PDSM
PFEA/PDSM-W1
PFEA/PDSM-W2
30.6
30.2
29.9
28.4
24.5
105.5
99.1
90.3
67.6
55.9
35.8
35.2
35.6
32.1
31.7
117.1
112.5
111.6
98.0
84.1
44.1
43.4
42.4
41.0
37.5
131.1
131.8
126.3
115.8
104.8
1.24
1.22
1.21
1.15
0.99
1.16
0.087
3.16
2.19
2.06
1.88
1.40
1.16
1.74
0.253
2.73
0.96
0.94
0.95
0.86
0.93
0.93
0.044
2.58
1.59
1.53
1.52
1.33
1.14
1.42
0.129
3.43
1.16
1.14
1.11
1.28
1.39
1.22
0.097
3.23
1.33
1.33
1.28
1.17
1.06
1.23
0.095
3.32
1.39
1.37
1.35
1.29
1.11
1.30
0.087
3.41
2.45
2.30
2.10
1.57
1.30
1.95
0.253
2.87
1.07
1.05
1.07
0.96
0.95
1.02
0.059
2.68
1.78
1.71
1.70
1.49
1.28
1.59
0.129
3.64
1.16
1.14
1.11
1.28
1.39
1.22
0.097
3.05
1.49
1.49
1.43
1.31
1.19
1.38
0.095
3.55
1.10
1.14
1.22
1.29
1.38
1.23
0.092
2.92
1.38
1.38
1.42
1.30
1.48
1.39
0.048
3.81
1.20
1.22
1.30
1.24
1.36
1.27
0.051
3.40
1.25
1.26
1.35
1.33
1.42
1.32
0.053
3.55
1.45
1.47
1.49
1.50
1.46
1.47
0.013
4.26
1.32
1.38
1.40
1.39
1.45
1.39
0.032
3.91
0.58
0.60
0.64
0.68
0.71
0.64
0.086
0.89
0.83
0.83
0.87
0.78
0.84
0.83
0.040
2.11
0.67
0.68
0.74
0.73
0.83
0.73
0.090
1.38
0.68
0.68
0.73
0.71
0.75
0.71
0.045
1.45
0.82
0.83
0.87
0.91
0.95
0.88
0.064
2.22
0.74
0.77
0.79
0.80
0.83
0.79
0.045
1.87
0.90
0.92
0.99
1.05
1.08
0.99
0.080
2.60
1.27
1.26
1.29
1.13
1.18
1.23
0.057
3.66
0.93
0.94
1.02
1.01
1.15
1.01
0.087
2.63
1.08
1.08
1.16
1.12
1.12
1.11
0.029
3.43
1.05
1.06
1.11
1.17
1.21
1.12
0.063
3.24
1.09
1.13
1.17
1.17
1.22
1.16
0.041
3.54
0.88
0.90
0.97
1.02
1.04
0.96
0.073
2.54
1.12
1.10
1.12
0.97
1.00
1.06
0.067
2.99
0.91
0.92
1.00
0.99
1.11
0.99
0.081
2.59
1.06
1.05
1.13
1.09
1.08
1.08
0.026
3.33
1.03
1.04
1.09
1.14
1.18
1.10
0.057
3.19
1.07
1.11
1.14
1.14
1.18
1.13
0.035
3.46
buckling stress for each element is weighted in accordance

with the ratio of the area of the element to the total area
of the section [1]. Effects of elastic local buckling on
column strength were determined based on Clause 4.7.4
of Part I-B of the AA Specication. The design rules in
the AS/NZS Standard for calculating the design strengths
of non-welded aluminum columns are generally identical
to those in the AA Specication, except that the AS/NZS
Standard reduces the yield load of the column using

a parameter kc which is not included in the AA
Specication. The EC9 Code adopts the Perry curve for
column design, and values of the imperfection factors
are listed in Table 5.6 of the Code. The effects of local
buckling on column strength are considered by replacing
the true section with an effective section. The effective
cross-section is obtained by employing a local buckling
ARTICLE IN PRESS
975
Table 5
Comparison of FEA and design strengths for welded columns of aluminum alloy 6061-T6
Specimen
T6-A-W-L500
T6-A-W-L1200
T6-A-W-L2000
T6-A-W-L2700
T6-A-W-L3500
Mean, Pm
COV, VP
Reliability index,
T6-B-W-L500
T6-B-W-L1200
T6-B-W-L2000
T6-B-W-L2700
T6-B-W-L3500
Mean, Pm
COV, VP
Reliability index,
T6-C-W-L500
T6-C-W-L1200
T6-C-W-L2000
T6-C-W-L2700
T6-C-W-L3500
Mean, Pm
COV, VP
Reliability index,
T6-D-W-L500
T6-D-W-L1200
T6-D-W-L2000
T6-D-W-L2700
T6-D-W-L3500
Mean, Pm
COV, VP
Reliability index,
T6-E-W-L500
T6-E-W-L1200
T6-E-W-L2000
T6-E-W-L2700
T6-E-W-L3500
Mean, Pm
COV, VP
Reliability index,
T6-F-W-L500
T6-F-W-L1200
T6-F-W-L2000
T6-F-W-L2700
T6-F-W-L3500
Mean, Pm
COV, VP
Reliability index,
FEA
Comparison
PFEA (kN)
PFEA/PAA
PFEA/PAS/NZS
PFEA/PEC9
PFEA/PDSM
PFEA/PDSM-W1
PFEA/PDSM-W2
40.5
40.0
39.4
35.5
29.4
126.5
123.3
110.6
82.7
66.4
47.1
45.6
46.8
39.7
37.6
146.0
146.9
140.5
109.7
101.7
56.2
55.8
54.1
51.0
42.2
179.3
175.4
159.5
142.8
122.9
1.16
1.15
1.13
1.02
0.85
1.06
0.126
2.48
1.87
1.82
1.63
1.22
0.98
1.50
0.258
2.37
0.95
0.92
0.94
0.96
1.08
0.97
0.064
2.65
1.41
1.42
1.36
1.06
0.98
1.25
0.167
2.62
1.14
1.13
1.36
1.56
1.53
1.34
0.205
2.99
1.29
1.26
1.15
1.03
0.89
1.12
0.147
2.48
1.30
1.29
1.27
1.14
0.95
1.19
0.150
2.69
2.09
2.04
1.83
1.37
1.10
1.68
0.258
2.50
1.00
0.97
1.00
0.96
1.08
1.00
0.047
2.67
1.58
1.59
1.52
1.19
1.10
1.40
0.167
2.81
1.14
1.13
1.36
1.56
1.53
1.34
0.205
2.84
1.44
1.41
1.28
1.15
0.99
1.26
0.150
2.65
1.39
1.44
1.57
1.65
1.79
1.57
0.103
3.71
1.40
1.48
1.60
1.62
1.93
1.60
0.126
3.51
1.52
1.53
1.66
1.52
1.65
1.58
0.047
4.35
1.41
1.50
1.59
1.47
1.83
1.56
0.106
3.65
1.78
1.82
1.85
1.83
1.65
1.79
0.080
4.91
1.65
1.69
1.66
1.65
1.74
1.68
0.023
4.80
0.60
0.62
0.70
0.74
0.78
0.69
0.114
1.06
0.68
0.73
0.80
0.78
0.92
0.78
0.112
1.52
0.69
0.70
0.80
0.77
0.89
0.77
0.108
1.49
0.66
0.70
0.75
0.68
0.80
0.72
0.080
1.35
0.82
0.85
0.91
0.96
0.96
0.90
0.072
2.28
0.79
0.81
0.82
0.84
0.88
0.83
0.042
2.09
0.89
0.93
1.04
1.09
1.15
1.02
0.113
2.51
1.05
1.11
1.18
1.10
1.22
1.13
0.058
3.29
0.92
0.93
1.06
1.03
1.19
1.03
0.105
2.54
1.01
1.07
1.15
1.03
1.19
1.09
0.069
3.08
1.01
1.05
1.12
1.19
1.18
1.11
0.071
3.13
1.13
1.15
1.16
1.18
1.24
1.17
0.037
3.62
0.90
0.93
1.03
1.07
1.11
1.01
0.089
2.61
1.03
1.06
1.08
0.96
1.04
1.03
0.045
3.02
0.93
0.94
1.06
1.01
1.15
1.02
0.091
2.63
1.02
1.07
1.14
1.01
1.15
1.08
0.059
3.11
1.02
1.05
1.12
1.17
1.15
1.10
0.059
3.21
1.13
1.16
1.15
1.17
1.20
1.16
0.022
3.66
coefcient rc to reduce the thickness of the element in the

section.
The strength of aluminum column with transverse welds
(welded column) depends on the location and number of
welds [1]. For SHS and RHS with transverse welds at the
ends of the columns only, the design equations given by the
AA and AS/NZS specications are identical to the design
equations of non-welded columns. However, the design

strength of welded columns is limited by the yield strength
of the welded material. The EC9 Code uses a factor rhaz to
consider the weakening effects of welding on column
strength, and rhaz is equal to 0.60 and 0.50 for the 6000
Series alloys of T5 and T6 conditions, respectively, as
shown in Table 5.2 of the Code.
ARTICLE IN PRESS
976
buckling for SHS and RHS columns, Pcrl is the Critical

elastic local column buckling load, E is the Youngs
modulus, le is the column effective length, r is the radius of
gyration of gross cross-section about the minor y-axis of
buckling.
The nominal axial strengths (PDSM) are calculated for
the two cases, as shown in Eqs. (2) and (3), respectively,
where Pne refers to the nominal axial strength for exural
buckling, and Pnl refers to the nominal axial strength for
local buckling as well as interaction of local and overall
buckling. The nominal axial strength, PDSM, is the
minimum of Pne and Pnl, as shown in Eq. (1). In calculating
the axial strengths, the critical elastic local buckling load
(Pcrl) of the cross-section was obtained from a rational
elastic nite strip buckling analysis [15].
B
Fig. 1. Denition of symbols.
5.2. Direct strength method for cold-formed steel structures

The direct strength method, which was developed for
cold-formed carbon steel members, is based on the same
underlying empirical assumption as the effective width
method: ultimate strength is a function of elastic buckling
and yielding of the material [7]. The direct strength method
has been proposed by Schafer and Pekoz [4] for laterally
braced exural members undergoing local or distortional
buckling. Subsequently, the method has been developed for
concentrically loaded pin-ended cold-formed steel columns
undergoing local, distortional, or overall buckling [5,6],
which allows for interaction of local and overall buckling
as well as distortional buckling. The SHS and RHS are
investigated in this study, and distortional buckling does
not occur on these sections. Therefore, distortional
buckling is not considered in this study. As summarized
in the NAS [8,9] for cold-formed steel structures, the
column design rules of the direct strength method that
considered the local and overall exural buckling are
shown in Eqs. (1)(3). The values of 0.15 and 0.4 are the
coefcient and exponent of the direct strength equation,
respectively, that were calibrated against test data of
concentrically loaded pin-ended cold-formed steel columns
for certain cross-sections and geometric limits:
PDSM min Pne ; Pnl ,
Pne

8
l2
>
< 0:658 c Py

>
Py
: 0:877
2
l
for lc p1:5;
(2)
for lc 41:5;
Pnl
8
>
<
>
: 1 0:15
(1)
Pne
0:4
Pcrl
Pne
Pcrl
Pne
0:4
for ll p0:776;
Pne
for ll 40:776;
(3)
p
p
where Py fy A; lc Py =Pcre ; ll Pne =Pcrl .
A is the gross cross-section area, fy is the material yield
strength which is the static 0.2% proof stress (s0.2) using
the non-welded material properties in this paper, Pcre is the
p2EA/(le/r)2, critical elastic buckling load in exural
5.3. Proposed design rules for aluminum alloy non-welded

columns
It should be noted that the direct strength method was
developed based on open sections, such as simple lipped
channel, lipped channel with web stiffeners, Zed section,
hat section and rack upright section. In this study, square
and RHSs are investigated. Therefore, the direct strength
method for cold-formed carbon steel members was
modied for aluminum alloy columns. The proposed
design equations for aluminum alloy SHS and RHS
columns without transverse welds at the ends of the
columns (non-welded columns) are as follows:
PDSMNW minPne ; Pnl ,
Pne

8
l2
>
< 0:658 c Pynw

>
Pynw
: 0:877
2
l
c
Pnl
8
>
<
>
: 1 0:15
(4)
for lc p1:5;
(5)
for lc 41:5;
Pne
0:3
Pcrl
Pne
Pcrl
Pne
0:3
for ll p0:713;
Pne
for ll 40:713;
(6)
p
fy-nw A; lc Pynw =Pcre ; and ll
where P
py-nw
Pne =Pcrl ; and fy-nw is the non-welded material yield
strength.
The design equations were veried against the numerical
results obtained from the parametric study as presented in
this paper, and the test results reported by Zhu and Young
[12]. The proposed design Eqs. (4)(6) for aluminum nonwelded columns require only small modications to the
current direct strength method for cold-formed steel
members. In Eq. (3), the value of the exponent 0.4 was
modied to 0.3, and the non-dimensional slenderness (ll)
has been adjusted to 0.713 for a smooth transition of the
elastic and inelastic buckling loads as shown in Eq. (6). As
a result, the reliability index (b0) of 2.86 was obtained for
the proposed design rules, which is closer to the target
value of 2.5 compared with the reliability index (b0) of 3.07
ARTICLE IN PRESS
977
1.5
Eqn. (8)
Welded FEA data
Welded experimental data
Eqn. (3) DSM

Eqn. (6) Modified DSM
0.8
Pu/ Py-nw
Non-Welded FEA data
1
Pu / Pne
Non-Welded experimental data

( = 0.85; = 2.86)
0.6
0.4
0.5
0.2
( = 0.85; = 3.07)
0.5
1.5
2
l
2.5
3.5
0.5
1.5
2.5
3.5
(a)
1.5
Eqn. (9)
Welded FEA data
Fig. 2. Comparison of FEA and experimental data with direct strength

curves for non-welded columns.
1
Pu / Pne
for the current direct strength method, as shown in Fig. 2.

The load combination of 1.2 DL (dead load)+1.6 LL (live
load) was used in calculating the reliability index as
specied in the AA Specication [1]. The resistance factor
(f) of 0.85 was used in the calculation. The reliability
analysis is detailed in Section 6 of this paper. Fig. 2 shows
the comparison of FEA and experimental results against
the current and modied direct strength curves plotted
from Eqs. (3) and (6), respectively, for the non-welded
columns. Generally, the results predicted using the current
and modied direct strength methods compared reasonably
well with the FEA and experimental results. However, the
modied direct strength method provided a reliability
index closer to the target value compared with the current
direct strength method.
0.5
0
0
Pne

8
l2
>
< 0:65 0:7 c Pynw

0:664
>
Pynw
:
2
l
c
(7)
for lc p1:56;
(8)
for lc 41:56;
1.5
2.5
3.5
Fig. 3. Comparison of FEA and experimental data with proposed design

rules (PDSM-W1) using non-welded material properties for welded columns:
(a) exural buckling and (b) interaction of local and exural buckling.
8
>
<
Pne

0:3 0:3
Pcrl
Pcrl
Pne
>
Pne
: 0:88 1 0:15 Pne
for ll p0:535;
for ll 40:535;
(9)
Two design approaches were proposed for aluminum

columns with transverse welds at both ends of the columns
(welded columns). The two design approaches were also
modied from the current direct strength method as well as
the current nominal axial strength (Pne) equations for
exural buckling. The proposed design rules were calibrated with the welded column strengths obtained from the
parametric study presented in this paper, as well as the test
results reported by Zhu and Young [12]. The rst approach
adopts the non-welded material properties in calculating the
welded column strength. The proposed design equations
are shown in Eqs. (7)(9). It should be noted that the Pcre
and Pcrl are obtained based on the non-welded material
properties:
PDSMW1 minPne ; Pnl ,
(b)
Pnl
5.4. Proposed design rules for aluminum alloy welded
columns
0.5
p
where P
fy-nw A; lc Pynw =Pcre ; and ll
py-nw
Pne =Pcrl ; and fy-nw is the non-welded material yield
strength.
Figs. 3(a) and (b) show the comparison of FEA and
experimental results against the results predicted by the
proposed design rules for exural buckling and interaction
of local and exural buckling, respectively. The proposed
design Eqs. (7)(9) for aluminum welded columns were
modied based on the proposed design Eqs. (4)(6) for
aluminum non-welded columns. In Eq. (5), the value of the
coefcients 0.658 and 0.877 were modied to 0.7 and 0.664,
respectively, as shown in Eq. (8). In addition, a coefcient
of 0.65 was multiplied to one of the equations, and the nondimensional slenderness (lc) has been adjusted to 1.56 for a
smooth transition of the elastic and inelastic buckling
loads. In Eq. (6), a coefcient of 0.88 was multiplied to one
of the equations, as shown in Eq. (9). Similarly, the nondimensional slenderness (ll) has been adjusted to 0.535 for
a smooth transition of the elastic and inelastic buckling
loads.
ARTICLE IN PRESS
978
The second approach adopts the welded material properties in calculating the welded column strength. The
proposed design equations are shown in Eqs. (10)(12),
where the Pcre and Pcrl are obtained based on the welded
material properties:
PDSMW2 minPne ; Pnl ,

8
l2
>
< 2 0:4 c Pyw for lc p1:0;

Pne
0:8
>
for lc 41:0;
Pyw
:
l2
(10)
(11)
equations. In Eq. (5), the value of the coefcients 0.658 and

0.877 were modied to 0.4 and 0.8, respectively, and a
coefcient of 2 was multiplied to one of the equations, as
shown in Eq. (11). The non-dimensional slenderness (lc)
has been adjusted to 1.0. In Eq. (6), a coefcient of 0.81
was multiplied to one of the equations, and the nondimensional slenderness (ll) has been adjusted to 0.445, as
shown in Eq. (12).
6. Reliability analysis
Pnl
8
>
<

0:81
1 0:15
>
:
Pne
0:3
Pcrl
Pne
Pcrl
Pne
0:3
for ll p0:445;
Pne
for ll 40:445;
(12)
p
p
where Py-w fy-w A; lc Pyw =Pcre ; ll Pne =Pcrl , and
fy-w is the welded material yield strength.
The results obtained from the proposed design equations
were compared with the FEA and experimental results of
welded columns, as shown in Figs. 4(a) and (b). The
proposed design Eqs. (10)(12) for aluminum welded
columns were also modied based on the proposed
design Eqs. (4)(6) for aluminum non-welded columns.
However, the welded material properties were used in these
4
Eqn. (11)
Welded FEA data
Pu/ Py-w
0
0
0.5
1.5
(a)
1.5
Eqn. (12)
Welded FEA data
Pu / Pne
0.5
The reliability of the design rules for aluminum columns

is evaluated using reliability analysis. Reliability analysis is
detailed in the AA Specication [1], and the ratio of dead
(DL) to live (LL) load of 0.2 was used in the analysis. In
general, a target reliability index of 2.5 for aluminum alloy
columns as a lower limit is recommended by the AA
Specication [1]. If the reliability index is greater than or
equal to 2.5 (bX2:5), then the design is considered to be
reliable. The AA and AS/NZS specications provide
different resistance factors (f) for compression members
with different failure modes. The resistance factor varies
with the slenderness parameter for exural buckling failure
mode. The resistance factor is a constant and equal to 0.85
for local buckling or interaction of local and overall
buckling failure mode. The observed failure modes of the
columns in this study included local buckling, exural
buckling, interaction of local and overall buckling, and
failure in the HAZ. Hence, the resistance factor of the
columns given by the AA and AS/NZS specications
ranged from 0.73 to 0.95. In calculating the reliability
indices of the AA and AS/NZS design rules, the resistance
factor for the columns is chosen to be equal to 0.85 for all
failure modes in this study. The EC9 Code provides a
constant resistance factor of 1/1.1 0.91 for compression
members, which is used in the reliability analysis. The
reliability of the direct strength method and proposed
design rules for aluminum columns is also evaluated, and
the resistance factor is equal to 0.85.
The load combination of 1.2 DL+1.6 LL is used in the
analysis for the AA Specication, the direct strength
method and the proposed design rules. The load combinations of 1.25 DL+1.5 LL and 1.35 DL+1.5 LL are used in
the analysis for AS/NZS and EC9 specications, respectively. The statistical parameters Mm, Fm, VM, and VF are
the mean values and coefcients of variation (COV) of
material and fabrication factors. These values are obtained
from Section 9 of Part I-B of the AA Specication [1],
where Mm 1.10, Fm 1.00, VM 0.06, and VF 0.05.
The statistical parameters Pm and VP are the mean value
and the COV of test-to-predicted load ratios, respectively.
0
0
(b)
0.5
1.5
2.5
3.5
Fig. 4. Comparison of FEA and experimental data with proposed design

rules (PDSM-W2) using welded material properties for welded columns: (a)
exural buckling and (b) interaction of local and exural buckling.
7. Comparison of numerical and experimental results with

design predictions
The nominal axial strengths (unfactored design strengths)
predicted by the AA Specication (PAA), AS/NZS Standard
ARTICLE IN PRESS
(PAS/NZS), and European Code (PEC9) for aluminum

structures, as well as the direct strength method (PDSM)
and proposed design equations (PDSM-NW, PDSM-W1, PDSMW2) are compared with the column strengths obtained from
the parametric study (PFEA) and experimental program
(PExp) [12], as shown in Tables 27. The statistical
parameters Pm and VP which are the mean value and
COV of FEA and experimental-to-predicted load ratios of
each series of specimens are shown in Tables 27.
Reliability indices (b) of the design rules for each series
of specimens are also shown in Tables 27. The FEA
results are also compared with the column design curves
obtained from the design rules, as shown in Figs. 528.
The design strengths are calculated using the material
properties for each series of specimens, as reported by Zhu
and Young [1213], and the 0.2% proof stress (s0U2) was
used as the corresponding yield stress. The design strengths
of non-welded column specimens are calculated using the
non-welded material properties. In calculating the design
strengths of welded columns, PAA and PAS/NZS are
calculated using the welded material properties, as specied
in the AA and AS/NZS specications. Whereas PEC9 are
calculated using the non-welded material properties as
required by the EC9 Code. The design strengths of the
proposed design equations PDSM-W1 are calculated using
the non-welded material properties, while PDSM-W2 are
calculated using the welded material properties, as detailed
in Section 5.4 Proposed design rules for aluminum alloy
welded columns of the paper. The column specimens were
designed as concentrically loaded compression members,
and the effective length (le) was taken as one-half of the
column length (L).
For the non-welded columns, it is shown that the column
strengths predicted by the AA and AS/NZS specications
are quite close. It should be noted that the AA and AS/
NZS predictions are unconservative for some specimens, as
shown in Tables 2 and 3. Reliability index as low as 1.98
979
and 1.77 were obtained for column Series T6-D-NW, with

the corresponding load ratios PFEA/PAA and PFEA/PAS/
NZS of 0.88 and 0.91, respectively. However, the AA and
AS/NZS predictions are generally conservative or less
unconservative for short specimens. Table 6 indicates the
AA and AS/NZS specications accurately predicted the
column strengths obtained from the experimental program
with column length ranged from 300 to 1650 mm. The
design strengths calculated using the EC9 Code are
generally conservative for the non-welded columns. However, the design strengths calculated using the EC9
Code are unconservative for a few specimens with long
column lengths, such as specimens T6-C-NW-L3500 and
T6-E-NW-L3500 with the PFEA/PEC9 load ratio of
0.87 and 0.84, respectively. The reliability index for the
design rules of the EC9 Code are mostly less than the target
value of 2.5, where seven series among twelve series are less
than the target value, as shown in Tables 2 and 3. This is
due to the large resistance factor of 0.91 has been used in
the EC9 Code. The design strengths PDSM predicted by the
current direct strength method are generally accurate and
reliable for all the non-welded column series, as shown
in Tables 2, 3 and 6. The mean values of the load ratio
PFEA/PDSM for each series of FEA specimens ranged from
0.96 to 1.20, with the corresponding COV ranged from
0.010 to 0.107. In terms of the experimental program, the
mean value of the load ratio PExp/PDSM is 1.10, with the
corresponding COV of 0.063, as shown in Table 6. It is
shown that the current direct strength method could be
successfully used in the design of aluminum non-welded
columns of square and RHSs. However, for all the nonwelded column series of FEA and test specimens, the
reliability indices of the current direct strength method are
greater than the target value of 2.5, which ranged from 2.73
to 3.46. Thus, the design equations shown in Section 5.3
Proposed design rules for aluminum alloy non-welded
columns of this paper have been proposed that require
Table 6
Comparison of test strengths with design strengths for non-welded columns
Specimen
N-S1-L300
N-S1-L1000
N-S1-L1650
N-R1-L300
N-R1-L1000
N-R2-L300
N-R2-L1000
H-R1-L300
H-R1-L1000
H-R2-L300
H-R2-L1000
Mean, Pm
COV, VP
Experimental
Comparison
PExp (kN)
PExp/PAA
PExp/PAS/NZS
PExp/PEC9
PExp/PDSM
PExp/PDSM-NW
34.1
33.7
33.6
42.3
41.7
147.9
145.8
53.3
51.6
209.2
202.4
1.08
1.06
1.08
1.08
1.03
1.03
1.01
1.08
1.06
1.06
1.04
1.06
0.023
3.24
1.08
1.06
1.08
1.08
1.03
1.10
1.08
1.08
1.06
1.14
1.12
1.08
0.027
3.13
1.13
1.20
1.34
1.02
1.07
1.08
1.14
1.06
1.10
1.14
1.21
1.14
0.078
2.88
1.09
1.14
1.26
1.05
1.09
0.99
1.04
1.09
1.12
1.10
1.15
1.10
0.063
3.30
1.09
1.14
1.28
0.94
0.98
1.03
1.08
0.95
0.99
1.10
1.16
1.07
0.096
3.00
ARTICLE IN PRESS
980
Table 7
Comparison of test strengths with design strengths for welded columns
Specimen
N-S1-W-L300
N-S1-W-L1000
N-S1-W-L1650
N-S1-W-L2350
N-S1-W-L3000
Mean, Pm
COV, VP
Reliability index,
N-R1-W-L300
N-R1-W-L1000
N-R1-W-L1650
N-R1-W-L2350
N-R1-W-L3000
Mean, Pm
COV, VP
Reliability index,
N-R2-W-L300
N-R2-W-L1000
N-R2-W-L1650
N-R2-W-L2350
N-R2-W-L3000
Mean, Pm
COV, VP
Reliability index,
H-R1-W-L300
H-R1-W-L1000
H-R1-W-L1650
H-R1-W-L2350
H-R1-W-L3000
Mean, Pm
COV, VP
Reliability index,
H-R2-W-L300
H-R2-W-L1000
H-R2-W-L1650
H-R2-W-L2350
H-R2-W-L3000
Mean, Pm
COV, VP
Reliability index,
Experimental
Comparison
PExp (kN)
PExp/PAA
PExp/PAS/NZS
PExp/PEC9
PExp/PDSM
PExp/PDSM-W1
PExp/PDSM-W2
18.8
19.2
19.8
18.4
15.2
26.4
27.7
28.5
25.1
23.2
101.0
89.7
85.4
74.3
60.4
37.5
37.9
37.7
30.3
23.8
118.0
139.3
119.4
95.2
75.4
1.45
1.47
1.55
1.41
1.16
1.41
0.105
3.69
0.88
0.94
0.94
0.92
0.93
0.92
0.025
2.62
1.77
1.61
1.53
1.35
1.09
1.47
0.177
3.00
1.02
1.04
1.04
1.06
1.00
1.03
0.022
3.13
1.48
1.80
1.54
1.23
0.98
1.41
0.222
2.48
1.62
1.65
1.73
1.59
1.30
1.58
0.104
3.93
0.95
1.01
1.02
0.92
0.93
0.96
0.047
2.51
1.98
1.76
1.68
1.47
1.19
1.62
0.185
3.04
1.03
1.04
1.04
1.06
1.00
1.03
0.021
2.92
1.66
1.96
1.68
1.33
1.06
1.54
0.226
2.54
1.15
1.24
1.45
1.59
1.72
1.43
0.166
2.73
1.16
1.30
1.38
1.39
1.45
1.33
0.082
3.32
1.29
1.23
1.30
1.45
1.60
1.37
0.110
3.16
1.65
1.76
1.87
1.72
1.68
1.74
0.049
4.74
1.38
1.76
1.75
1.94
2.22
1.81
0.169
3.37
0.60
0.62
0.66
0.65
0.57
0.62
0.059
0.83
0.65
0.70
0.74
0.68
0.67
0.69
0.050
1.28
0.68
0.61
0.60
0.57
0.51
0.59
0.104
0.55
0.76
0.79
0.81
0.69
1.12
0.83
0.199
1.31
0.62
0.74
0.67
0.57
1.39
0.80
0.422
0.65
0.99
1.01
1.16
1.27
1.28
1.14
0.121
2.77
0.92
1.01
1.14
1.16
1.28
1.10
0.127
2.59
1.07
1.04
1.09
1.13
1.19
1.10
0.053
3.26
1.04
1.12
1.26
1.24
1.24
1.18
0.081
3.30
1.00
1.25
1.23
1.23
1.38
1.22
0.112
3.07
0.97
0.99
1.12
1.22
1.23
1.10
0.112
2.74
0.93
1.02
1.13
1.14
1.25
1.09
0.113
2.70
1.05
1.01
1.06
1.08
1.07
1.05
0.026
3.22
1.29
1.37
1.48
1.37
1.27
1.36
0.061
4.06
1.30
1.58
1.49
1.38
1.32
1.41
0.084
3.95
180
PAA PAS/NZS
60
Non-welded FEA
Flexural buckling
PDSM
40
20
PDSM-NW
PEC9
Column strength, Pu (kN)
80
PAA
160
Non-welded FEA
PAS/NZS
140
120
PDSM
100
80
60
40
PEC9
20
Flexural buckling
PDSM-NW
0
0
500
1500
2000
1000
Effective length, le (mm)
2500
3000
Fig. 5. Comparison of FEA and design column strengths for Series

T5-A-NW.
500
1000
1500
2000
2500
3000

T5-B-NW.
ARTICLE IN PRESS
250
PAA
Non-welded FEA
PAS/NZS
60
Flexural buckling
40
PDSM-NW
20
PDSM
PEC9
80
0
1500
1000
2000
500
2500
150
PAA P
AS/NZS
100
PDSM-NW
PDSM
50
3000

T5-C-NW.
PEC9
0
500
1500
1000
2000
2500
3000
T5-F-NW.
240
100
200
Non-welded FEA
PAS/NZS
PAA
Non-welded FEA
Flexural buckling
200
0
0
Flexural buckling
160
120
PDSM-NW
80
PDSM
40
PEC9
PAA
80
Non-welded FEA
PAS/NZS
Flexural buckling
60
PDSM-NW
40
PDSM
20
PEC9
0
0
0
500
1000
1500
2000
2500
3000

T5-D-NW.
250
Flexural buckling
40
PAA
PAS/NZS
PDSM-NW
20
PDSM
PEC9
1000
1500
2000
PAA
Non-welded FEA
60
500
2500
3000
T6-A-NW.
80
981
Non-welded FEA
PAS/NZS
200
Flexural buckling
150
PDSM-NW
100
PDSM
50
PEC9
0
0
500
1000
1500
2000
2500
3000
500
1000
1500
2000
2500
3000

T5-E-NW.
T6-B-NW.
only small modications to the current direct strength

method and obtaining closer reliability index to the
target value. The load ratios of the FEA (PFEA) and
experimental (PExp) column strengths to the design
strengths (PDSM-NW) calculated using the proposed design
equations for non-welded columns are shown in Tables 2, 3
and 6. The mean values of load ratios PFEA/PDSM-NW and
PExp/PDSM-NW ranged from 0.93 to 1.13, with the
corresponding COV ranged from 0.012 to 0.111 for all
the non-welded column series. It should be noted that the

reliability indices of the proposed design rules for each
column series are generally closer to the target value of 2.5
compared with the current direct strength method, which
ranged from 2.61 to 3.09. The non-welded column design
curves predicted by the AA, AS/NZS and EC9 specications, as well as the current and modied direct strength
methods are shown in Figs. 516 for each non-welded
column series.
ARTICLE IN PRESS
982
300
PAA PAS/NZS
80
Flexural buckling
60
PDSM-NW
40
PDSM
20
PEC9
PEC9
Flexural buckling
200
150
PDSM
PDSM-NW
100
50
0
500
1000
1500
2000
2500
3000
T6-C-NW.
500
1000
1500
2000
2500
3000
T6-F-NW.
300
60
PAA PAS/NZS
250
Flexural buckling
200
150
PDSM
100
PDSM
Non-welded FEA
Non-welded FEA
PAS/NZS
250
PDSM-NW
50
PEC9
Welded FEA
Flexural buckling
PDSM-W1
PDSM-W2
40
20
PAA
PAS/NZS
PEC9
0
500
1000
1500
2000
2500
3000
T6-D-NW.
100
PAA
500
1000
1500
2000
2500
3000
T5-A-W.
150
PAS/NZS
Non-welded FEA
80
PEC9
Flexural buckling
60
40
PDSM-NW
20
PDSM
PAA
Non-welded FEA
100
PDSM-W1
120
Welded FEA
PDSM-W2
Flexural buckling
90
60
30
PAA
PDSM
PAS/NZS
PEC9
0
0
500
1000
1500
2000
2500
3000
500
1000
1500
2000
2500
3000
T6-E-NW.
T5-B-W.
For the welded columns, it is shown that the design

strengths calculated using the AA specication (PAA) are
generally quite conservative, as shown in Tables 4, 5 and 7.
For the column series of FEA specimens, the mean values
of the load ratio PFEA/PAA ranged from 0.93 to 1.74, with
the corresponding COV ranged from 0.044 to 0.258, and
the reliability index ranged from 2.37 to 3.43. For the
column series of test specimens, the mean values of the load

ratio PExp/PAA ranged from 0.92 to 1.47, with the
corresponding COV ranged from 0.022 to 0.222, and the
reliability index ranged from 2.48 to 3.69. The design
strengths calculated using the AS/NZS Standard are
generally more conservative than the predictions given
by the AA Specication, as shown in Tables 4, 5 and 7.
ARTICLE IN PRESS
983
200
80
PDSM
PDSM
PDSM-W1
60
Welded FEA
Flexural buckling
PDSM-W2
40
PAA
20
PAS/NZS
PEC9
0
500
1000
1500
2000
Effective length, le ( mm)
2500
100
PAA
50
3000
T5-C-W.
PAS/NZS
PDSM-W1
PEC9
PDSM-W2
PDSM-W1
Flexural buckling
PDSM-W2
80
PAA
PAS/NZS
2500
3000
T5-F-W.
Welded FEA
120
40
1000
1500
2000
80
PDSM
160
500
Flexural buckling
200
Flexural buckling
150
0
0
PEC9
60
Welded FEA
PDSM
PDSM-W1
40
PDSM-W2
20
PAA
PAS/NZS
PEC9
0
0
500
1000
1500
2000
2500
3000
T5-D-W.
500
1000
1500
2000
2500
3000
T6-A-W.
200
Welded FEA
Flexural buckling
40
PAA
20
PAS/NZS
PEC9
PDSM-W1
PDSM-W2
60
Welded FEA
Flexural buckling
Welded FEA
150
PDSM
100
PDSM-W1
PDSM-W2
50
PAA
PDSM
PAS/NZS
PEC9
0
0
500
1000
1500
2000
2500
3000
500
1000
1500
2000
2500
3000
T5-E-W.
T6-B-W.
The design strength calculated using the EC9 Code for

welded columns are also quite conservative. The mean
values of the load ratio PFEA/PEC9 ranged from 1.23 to
1.79 for each column series of FEA specimens, and the
mean values of the load ratio PExp/PEC9 ranged from 1.33
to 1.81 for each column series of test specimens. The
corresponding COV for the load ratios PFEA/PEC9 and
PExp/PEC9 ranged from 0.013 to 0.169, and the reliability
index ranged from 2.73 to 4.91. It is also shown that the

current direct strength method is not suitable for the design
of aluminum welded columns. The mean values of the load
ratio PFEA/PDSM ranged from 0.64 to 0.90, with the
corresponding COV ranged from 0.040 to 0.114, and the
reliability index ranged from 0.89 to 2.28 for each column
series of FEA specimens, as shown in Tables 4 and 5.
The mean values of the load ratio PExp/PDSM ranged from
ARTICLE IN PRESS
984
250
Welded FEA
PDSM
Flexural buckling
60
40
PAA
20
PAS/NZS
PDSM-W1
PDSM-W2
PEC9
0
500
1000
1500
2000
2500
3000
T6-C-W.
250
Welded FEA
PDSM
200
Flexural buckling
150
100
PAA
PAS/NZS
50
PEC9
PDSM-W1
PDSM-W2
0
0
Flexural buckling
Welded FEA
200
PDSM
PDSM-W1
150
PDSM-W2
100
PAA
50
PAS/NZS
PEC9
0
0
500
1000
1500
2000
2500
3000
T6-D-W.
80
80
PDSM
Welded FEA
Flexural buckling
60
40
PAA
20
PAS/NZS
PEC9
PDSM-W1
PDSM-W2
0
0
500
1000
1500
2000
2500
3000
T6-E-W.
0.59 to 0.83, with the corresponding COV ranged from

0.050 to 0.422, and the reliability index ranged from 0.55 to
1.31 for each column series of test specimens, as shown in
Table 7.
Two design approaches have been proposed for aluminum columns with transverse welds at both ends of the
columns (welded columns). The rst approach (PDSM-W1)
adopts the non-welded material properties in calculating the
welded column strength as detailed in Eqs. (7)(9), whereas
500
1000
1500
2000
2500
3000
T6-F-W.
the second approach (PDSM-W2) adopts the welded material

properties in the calculation, as described in Eqs. (10)(12).
The design strengths (PDSM-W1 and PDSM-W2) calculated
using the two proposed design rules for welded columns are
in good agreement with the FEA and experimental results,
as shown in Tables 4, 5 and 7. For the rst design approach
that uses the non-welded material properties, the mean
values of the load ratios PFEA/PDSM-W1 and PExp/PDSM-W1
ranged from 0.99 to 1.23, with the corresponding COV
ranged from 0.029 to 0.127. The reliability indices for this
design rules (PDSM-W1) are greater than the target value of
2.5 that ranged from 2.51 to 3.66 for each series of
specimens. Fig. 3(a) shows the comparison of the welded
column results that failed by exural buckling against the
results obtained from the proposed Eq. (8), whereas
Fig. 3(b) shows the comparison of all the welded column
results against the results obtained from the proposed
Eq. (9). The design strengths calculated using the second
design approach (PDSM-W2) that uses the welded material
properties are generally conservative, as shown in Tables 4,
5 and 7. The mean values of the load ratio PFEA/PDSM-W2
ranged from 0.96 to 1.16, with the corresponding COV
ranged from 0.022 to 0.091. The mean values of the load
ratio PExp/PDSM-W2 ranged from 1.05 to 1.41, with
the corresponding COV ranged from 0.026 to 0.113. The
reliability indices for PDSM-W2 are also greater than the
target value of 2.5 that ranged from 2.54 to 4.06 for all
welded column series. Fig. 4(a) shows the comparison of
the welded column results that failed by exural buckling
with the results obtained from the proposed Eq. (11).
Fig. 4(b) shows the comparison of all the welded column
results against the results obtained from the proposed
Eq. (12). The welded column design curves predicted by the
AA, AS/NZS and EC9 specications, as well as the current
direct strength method and the two proposed design rules
are shown in Figs. 1728 for each welded column series. It
is shown that the two proposed design rules can be used for
the design of aluminum columns of square and RHSs with
transverse welds at the ends of the columns. Generally, it is
considered that the non-welded material properties could
be obtained easily compared with the welded material
ARTICLE IN PRESS
properties. The designer may choose either one of the

proposed design rules for the welded columns based on the
available material properties.
8. Conclusions
This paper presents a parametric study of aluminum
alloy square and RHS columns using FEA. A non-linear
FEM was used in the parametric study that has been
veried against experimental results. The research program contained aluminum columns of 6063-T5 and 6061T6 heat-treated aluminum alloys with and without
transverse welds at both ends of the columns. The
parametric study included 24 column series with different
cross-section geometry and type of aluminum alloy. Each
series contained 5 specimens with the column lengths
ranged from 500 to 3500 mm. The column strengths
obtained from experimental and numerical investigations
were compared with the design strengths calculated using
the current American, Australian/New Zealand and
European specications for aluminum structures. The
column strengths were also compared with the design
strengths calculated using the direct strength method that
was developed for cold-formed steel members. Based on
the available data obtained from experimental and
numerical investigations, design rules modied from the
direct strength method were proposed for aluminum
columns with and without transverse welds at the ends
of the columns. Reliability analysis was performed to
evaluate the reliability of the design rules. It is shown that
the design strengths calculated using the proposed design
rules are in good agreement with the test and numerical
results. It is also shown that the proposed design rules are
reliable.
985
References
[1] AA. Aluminum design manual. Washington, DC: The Aluminum
Association; 2005.
[2] AS/NZS, Aluminum structures Part 1: Limit state design, Australian/
New Zealand Standard AS/NZS 1664.1:1997. Sydney, Australia:
Standards Australia, 1997.
[3] EC9. Eurocode 9: Design of aluminum structuresPart 1-1: General
rulesgeneral rules and rules for buildings, DD ENV 1999-1-1:2000.
Final Draft October 2000, European Committee for Standardization,
2000.
[4] Schafer BW, Pekoz T. Direct strength prediction of cold-formed steel
members using numerical elastic buckling solutions. In: Proceedings
of the 14th international specialty conference on cold-formed steel
structures, University of Missouri-Rolla, Rolla, Mo 1998. p. 6976.
[5] Schafer BW. Distortional buckling of cold-formed steel columns.
August Final Report to the American Iron and Steel Institute,
Washington, DC, 2000.
[6] Schafer BW. Local, distortional, and Euler buckling of thin-walled
columns. J Struct Eng 2002;128(3):28999.
[7] Schafer BW. Progress on the direct strength method. In: Proceeding
of 16th international specialty conference on cold-formed steel
structures, Orlando, FL, 2002. p. 64762.
[8] North American Specication for the design of cold-formed steel
structural members. Washington, DC: American Iron and Steel
Institute, 2001.
[9] Supplement to the North American Specication for the design of
cold-formed steel structural member. Washington, DC: American
Iron and Steel Institute; 2004.
[10] Mazzolani FM. Aluminum alloy structures. 2nd ed. London: E & FN
Spon; 1995.
[11] Sharp ML. Behaviour and design of aluminum structures. New York:
McGraw-Hill; 1993.
[12] Zhu JH, Young B. Tests and design of aluminum alloy compression
members. J Struct Eng 2006;132(7):1096107.
[13] Zhu JH, Young B. Aluminum alloy tubular columnsPart I: Finite
element modeling and test verication. Thin-Walled Struct 2006, in
press, doi:10.1016/j.tws.2006.08.011.
[14] ABAQUS Analysis Users Manual, Version 6.5. ABAQUS, Inc., 2004.
[15] Papangelis JP, Hancock GJ. Computer analysis of thin-walled
structural members. Comput Struct 1995;56(1):15776.

Aluminum Alloy Tubular Columns-Part II: Parametric Study and Design Using Direct Strength Method

Transféré par

Informations du document

Titre original

Copyright

Formats disponibles

Partager ce document

Partager ou intégrer le document

Options de partage

Avez-vous trouvé ce document utile ?

Ce contenu est-il inapproprié ?

Droits d'auteur :

Formats disponibles

Aluminum Alloy Tubular Columns-Part II: Parametric Study and Design Using Direct Strength Method

Transféré par

Droits d'auteur :

Formats disponibles

ARTICLE IN PRESS

Thin-Walled Structures 44 (2006) 969985

Aluminum alloy tubular columnsPart II: Parametric study and design

E-mail address: young@hku.hk (B. Young).

design rules for compression members. The design rules in

J.-H. Zhu, B. Young / Thin-Walled Structures 44 (2006) 969985

gross cross-section area

PDSM-W2 welded column design strength calculated

ling ( Py =Pcre ; Pynw =Pcre or Pyw =Pcre )

local and exural buckling ( Pne =Pcrl )

with the design strengths calculated using the American

specimens were determined by longitudinal tensile coupon

Note: 1 in. 25.4 mm, NW non-welded column series, W welded

For example, the label T5-A-NW-L500 denes the

The rst letter indicates the type of material of the

The material properties of the specimens of 6063-T5

Note: 1 kip 4.45 kN.

are identical to the material properties of Series H-R2 in

10 10 mm2 (length width) for the non-welded columns.

Note: 1 kip 4.45 kN.

Note: 1 kip 4.45 kN.

buckling stress for each element is weighted in accordance

Standard reduces the yield load of the column using

Note: 1 kip 4.45 kN.

coefcient rc to reduce the thickness of the element in the

equations of non-welded columns. However, the design

buckling for SHS and RHS columns, Pcrl is the Critical

5.2. Direct strength method for cold-formed steel structures

5.3. Proposed design rules for aluminum alloy non-welded

Eqn. (3) DSM

Non-Welded FEA data

Non-Welded experimental data

Fig. 2. Comparison of FEA and experimental data with direct strength

for the current direct strength method, as shown in Fig. 2.

Fig. 3. Comparison of FEA and experimental data with proposed design

Two design approaches were proposed for aluminum

equations. In Eq. (5), the value of the coefcients 0.658 and

The reliability of the design rules for aluminum columns

Fig. 4. Comparison of FEA and experimental data with proposed design

7. Comparison of numerical and experimental results with

(PAS/NZS), and European Code (PEC9) for aluminum

and 1.77 were obtained for column Series T6-D-NW, with

Note: 1 kip 4.45 kN.

Column strength, Pu (kN)

Column strength, Pu (kN)

Fig. 5. Comparison of FEA and design column strengths for Series

Fig. 6. Comparison of FEA and design column strengths for Series

Column strength, Pu (kN)

Column strength, Pu (kN)

Fig. 7. Comparison of FEA and design column strengths for Series

Column strength, Pu (kN)

Fig. 8. Comparison of FEA and design column strengths for Series

Column strength, Pu (kN)

Fig. 9. Comparison of FEA and design column strengths for Series

only small modications to the current direct strength

the non-welded column series. It should be noted that the

Column strength, Pu (kN)

Column strength, Pu (kN)

Column strength, Pu (kN)

Column strength, Pu (kN)

For the welded columns, it is shown that the design

column series of test specimens, the mean values of the load

Column strength, Pu (kN)