Thin–Walled Structures 205 (2024) 112431

Contents lists available at ScienceDirect

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

Full length article

Local buckling and capacities of stainless steel hexagonal hollow sections

under axial compression
Yukai Zhong a , Ke Jiang b,c , Andi Su d,e, Jiyang Fu a, Airong Liu a, Ou Zhao b,*
a
Research Center for Wind Engineering and Engineering Vibration, Guangzhou University, Guangzhou, China
b
School of Civil and Environmental Engineering, Nanyang Technological University, Singapore
c
Department of Civil and Natural Resources Engineering, University of Canterbury, Christchurch, New Zealand
d
Key Lab of Structures Dynamic Behavior and Control of the Ministry of Education, Harbin Institute of Technology, Harbin, 150090, China
e
Key Lab of Smart Prevention and Mitigation of Civil Engineering Disasters of the Ministry of Industry and Information Technology, Harbin Institute of Technology,
Harbin, 150090, China

A R T I C L E I N F O A B S T R A C T

Keywords: This paper presents experimental and numerical investigations into the local buckling behaviour and capacities
Hexagonal hollow sections of stainless steel hexagonal hollow sections. A testing programme, including tensile coupon tests, initial local
Improved design approach geometric imperfection measurements and fifteen stub column tests, was firstly carried out. The key test results,
Local buckling
including failure loads, load–end shortening curves and failure modes, were reported. Subsequently, a numerical
Numerical modelling
Slenderness limits
modelling programme was conducted, where finite element models were developed and validated against the
Stainless steel test results and then used to conduct parametric studies to generate additional numerical data. The obtained test
Stub column tests and numerical data were used to evaluate the relevant local buckling design rules specified in the European code,
American specification and ASCE standard. The evaluation results revealed that the slenderness limits, as defined
in the European code and American specification, were generally accurate and safe when used for cross-section
classification of stainless steel hexagonal hollow sections, while the ASCE slenderness limit was unsafe. The
European code and American specification resulted in overall accurate and consistent cross-section compression
resistance predictions, but the predictions for non-slender cross-sections were conservative and scattered, due to
the neglect of material strain hardening. The ASCE standard led to overall scattered and relatively conservative
cross-section compression resistance predictions, but also with some unsafe predictions for those intermediate
cross-section sizes due mainly to the unsafe slenderness limit. Finally, a revised ASCE design approach was
proposed and shown to result in more accurate predictions of cross-section classification and resistances.

1. Introduction section members have been previously carried out. Aoji et al. [13]
conducted stub column tests on carbon steel polygonal hollow sections
Due to the aesthetic appearance and superior torsional and local with different numbers of sides and proposed an empirical design for­
buckling resistances, steel tubular sections, including circular and rect­ mula to predict their local buckling strengths. The local and global
angular hollow sections, have been widely used in civil and offshore interactive behaviour of polygonal hollow section columns were
engineering. In recent decades, the advancement in manufacturing experimentally investigated by Migita et al. [14], with strain energy
technology has made it possible to produce novel types of tubular sec­ analysed and a design formula proposed. The local and distortional
tions, such as oval hollow sections [1,2], flat-oval hollow sections [3–5], behaviour of polygonal hollow sections under compression, bending and
semi-oval hollow sections [6,7] and polygonal hollow sections [8–12]. torsion were theoretically and numerically investigated based on
In comparison with the currently commonly used rectangular hollow Generalised Beam Theory (GBT) [15,16], with the results revealing that
sections, polygonal hollow sections have smaller side widths (indicating the GBT approach could efficiently provide accurate closed-form
less susceptibility to local buckling) and better constructability analytical solutions. Liu et al. [17] conducted a series of tests to mea­
(enabling connections in more than four directions). sure the residual stresses and material properties of high strength steel
Studies on the structural behaviour of carbon steel hexagonal hollow hexagonal hollow sections and proposed predictive models for residual

* Corresponding author.
E-mail address: [email protected] (O. Zhao).

https://doi.org/10.1016/j.tws.2024.112431
Received 9 May 2024; Received in revised form 20 August 2024; Accepted 5 September 2024
Available online 6 September 2024
0263-8231/© 2024 Elsevier Ltd. All rights are reserved, including those for text and data mining, AI training, and similar technologies.
Y. Zhong et al. Thin-Walled Structures 205 (2024) 112431

stress distribution patterns. Stub column tests on high strength steel length; the notations of cross-section geometric parameters are illus­
hexagonal hollow sections were also performed by Liu et al. [18] to trated in Fig. 1. In terms of the labelling system of the specimen, the
study their local buckling behaviour and found that the current design specimen ID consisted of a letter ‘B’ followed by the nominal edge width
codes overestimated the load-carrying capacities for compact sections. of the cross-section as well as a letter ‘T’ followed by the nominal wall
Improved design approaches were then proposed to predict the thickness, e.g., B70-T2.
cross-section resistances of high strength steel hexagonal hollow sec­
tions under axial compression [8]. Liu et al. [19] also conducted a nu­
merical investigation on the cross-section behaviour of high strength 2.2. Material testing
steel hexagonal hollow sections under combined compression and
bending, with the conservatism of design predictions highlighted and an Tensile coupon tests were conducted to obtain the material proper­
improved design method considering the beneficial strain hardening ties and stress–strain curves of austenitic stainless steels. For each
proposed. It is worth noting that the aforementioned research was all thickness of steel sheet, two types of coupons, namely flat and corner
carried out on carbon steel hexagonal hollow section members. In coupons, were cut from the corresponding portions of the cross-section,
comparison with traditional carbon steels, stainless steels possess with their geometric dimensions being in conformity with the recom­
excellent corrosion resistance and almost 100 % recyclable nature as mendations of EN ISO 6892-1 [28]. A displacement-controlled Schenck
well as superior mechanical properties (high strength and good 250 kN hydraulic testing machine was adopted for all tensile coupon
ductility), resulting in lower maintenance costs during the whole service tests. The test setup is shown in Fig. 2, where a pair of clamps are
life and favourable structural performance [20–23]. Therefore, stainless equipped at the top and bottom of the machine to appropriately grip the
steel is an ideal sustainable construction material and has great potential coupon, an extensometer is mounted onto the central 50 mm of the
in offshore engineering. However, there has hitherto been no research coupon to record the elongations and two strain gauges are attached to
conducted on stainless steel hexagonal hollow section members, which the mid-height of the coupon to measure the tensile strains during
thus prompted the present study focusing on the local buckling behav­ testing. Two displacement rates, namely 0.05 mm/min and 0.8
iour and capacities of stainless steel hexagonal hollow sections under mm/min, were employed before and after the attainment of the nominal
axial compression. 0.2 % proof stress. Fig. 3 displays the measured stress–strain curves,
In this paper, a testing programme, including tensile coupon tests, while the key measured material properties are summarised in Table 2,
initial local geometric imperfection measurements and fifteen stub col­ including the Young’s modulus E, the 0.2 % proof stress σ0.2, the ulti­
umn tests, was firstly performed. Afterwards, a numerical modelling mate stress σu, the ultimate strain εu and the fracture strain εf as well as
programme was conducted to develop and validate finite element
models, which were then employed to perform parametric studies to
generate further numerical data. The test and numerical data were used
to evaluate the relevant design rules, as specified in EN 1993-1-4 [24],
AISC 370-21 [25] and ASCE/SEI 48-11 [26], for stainless steel hexago­
nal hollow sections.

2. Testing programme

2.1. Specimens

Fifteen stub column tests were carried out to investigate the local
buckling behaviour and capacities of stainless steel hexagonal hollow
sections under axial compression. The stub column specimens were
fabricated by welding two half sections, which were press-braked from
hot-rolled grade 304 austenitic stainless steel sheets [27] with three
nominal thicknesses of 1.5 mm, 2 mm and 4 mm. The geometric di­
mensions of all stub column specimens were carefully measured and
reported in Table 1, where B is the edge width of the cross-section, b is
the net width of the flat portion, H is the height of the cross-section, t is Fig. 1. Locations of coupons and notations of geometric parameters of hexag­
the wall thickness, ri is the inner corner radius and L is the member onal hollow section.

Table 1
Measured geometric dimensions and initial local geometric imperfection magnitudes of stainless steel hexagonal hollow section stub column specimens.
Specimen ID B (mm) b (mm) H (mm) t (mm) ri (mm) L (mm) ω0 (mm) ω0/t
B70-T1.5 69.3 65.2 119.9 1.48 2.0 363 0.07 1/21
B90-T1.5 89.0 85.0 154.2 1.43 2.0 467 0.02 1/72
B110-T1.5 109.3 105.3 189.2 1.41 2.0 571 0.03 1/47
B130-T1.5 128.8 124.8 223.0 1.42 2.0 675 0.04 1/36
B150-T1.5 149.0 145.1 258.1 1.41 2.0 779 0.08 1/18
B70-T2 69.0 63.8 119.5 1.91 2.6 363 0.08 1/24
B90-T2 89.5 84.3 155.0 1.94 2.6 467 0.15 1/13
B110-T2 108.8 103.5 188.4 1.92 2.6 571 0.10 1/19
B130-T2 128.5 123.3 222.6 1.91 2.6 675 0.17 1/11
B150-T2 149.3 144.0 258.5 1.91 2.6 779 0.20 1/10
B70-T4 68.5 58.3 118.6 3.85 5.0 363 0.07 1/55
B90-T4 88.5 78.2 153.3 3.89 5.0 467 0.09 1/43
B110-T4 108.5 98.2 187.9 3.88 5.0 571 0.11 1/35
B130-T4 128.5 118.3 222.6 3.87 5.0 675 0.13 1/30
B150-T4 148.5 138.2 257.2 3.89 5.0 779 0.08 1/49

2
Y. Zhong et al. Thin-Walled Structures 205 (2024) 112431

Fig. 2. Tensile coupon test setup.

the Ramberg–Osgood exponents n and m calculated by Eqs. (1) and (2),

where σ 0.01 is the 0.01 % proof stress [29].
ln(20)
n= (1)
ln(σ0.2 /σ 0.01 )
Fig. 3. Measured stress–strain curves.
σ 0.2
m = 1 + 3.5 (2)
σu
the magnitudes varied between t/10 to t/100. For the studied stainless
steel hexagonal hollow sections, the measured initial local geometric
2.3. Initial local geometric imperfection measurements imperfection magnitudes were shown to vary from t/10 to t/72, which
are similar to those of stainless steel square and rectangular hollow
Initial geometric imperfections are inevitably introduced in thin- sections. This can be attributed to their similar nature – tubular shape
walled metallic structural components during manufacturing, trans­ and same fabrication route (cold-forming and seam-welding).
portation and handling and can affect their structural behaviour and
resistances [30–42]. Therefore, initial local geometric imperfection 2.4. Stub column tests
measurements were carried out on the fifteen stainless steel hexagonal
hollow section stub column specimens. A similar test setup as that used The cross-section behaviour of the fifteen stainless steel hexagonal
in [30] was adopted and shown in Fig. 4, where three LVDTs, with one hollow section stub column specimens was studied by means of fixed-
located at the centre of the plate element and two located near the edge ended compression tests. A displacement-controlled Instron 2000 kN
of the corner of the plate element, are moved longitudinally along the hydraulic testing machine with fixed end platens was employed to apply
member length. Note that the measurements were only conducted on the axial compression loads on each stub column specimen at a constant rate
four plate elements without welds. For each plate element, the initial of 0.3 mm/min. To ensure a uniform stress distribution on each end
local geometric imperfections were determined as the deviations from section of the specimen, both ends of each specimen were milled flat
the data measured from the LVDT at the centre to a linear regression line before testing. Fig. 5 displays the stub column test setup, where a pair of
fitted to the data measured from the two LVDTs near the corner. Then, LVDTs are vertically attached to the top platen of the testing machine to
the initial local geometric imperfection magnitude of a stub column record the end shortenings of the stub column specimen during testing
specimen ω0 was taken as the maximum value obtained from the four and two strain gauges are adhered to the mid-height cross-section to
plate elements and reported in Table 1, together with the ω0/t ratios. For measure the longitudinal strains. Note that the data measured from the
stainless steel square and rectangular hollow sections, previous mea­ LVDTs consisted of the end shortening of the stub column specimen and
surements on initial local geometric imperfections [31–37] showed that the (approximately elastic) deformation of the two end platens of the

3
Y. Zhong et al. Thin-Walled Structures 205 (2024) 112431

Table 2
Measured material properties from tensile coupon tests.
Plate thickness (mm) Type E (MPa) σ0.2 (MPa) σu (MPa) εu (%) εf (%) R–O coefficient

n m

1.5 Flat 202826 264 749 50 64 12.7 2.2

Corner 197524 308 779 44 58 4.3 2.4
2 Flat 203531 305 810 48 62 9.6 2.3
Corner 197905 412 895 46 60 4.5 2.6
4 Flat 201719 299 756 51 63 10.7 2.4
Corner 187457 624 1056 41 50 5.1 3.1

data [43]. The corrected load–end shortening curves for the fifteen
stainless steel hexagonal hollow section stub column specimens are
shown in Fig. 6, while the key measured test results are reported in
Table 3, including the failure load Nu, the end shortening at the failure
load δu and the failure load to yield load ratio Nu/Aσ 0.2, where A is the
gross cross-section area. Fig. 7 displays the local buckling failure modes
for the fifteen stainless steel hexagonal hollow section stub column
specimens.

3. Numerical modelling

3.1. General

Numerical modelling was conducted using the nonlinear finite

element (FE) software ABAQUS [44]. FE models of stainless steel hex­
agonal hollow section stub column specimens were firstly developed and
validated against the test results. Subsequently, the validated FE models
were employed to conduct parametric studies to generate further nu­
merical data over a wide range of cross-section dimensions.

3.2. Development of FE models

The shell element S4R [44], which has been proven to accurately
Fig. 4. Test setup for initial local geometric imperfection measurements. simulate the structural behaviour of thin-walled steel members[45–51],
was adopted herein for modelling the stainless steel hexagonal hollow
section stub column specimens. Based on a prior mesh sensitivity study
examining element sizes from B/10 × B/10 to B/50 × B/50, the final
mesh size was chosen as B/20 × B/20, satisfying both computational
accuracy and efficiency. For the material modelling of austenitic stain­
less steel, the plastic material model [44] was employed and required
the (measured) engineering stress–strain curves, as shown in Fig. 3, to be
converted into the true stress–plastic strain curves. No explicit model­
ling of residual stresses was performed for the studied cold-formed
stainless steel hexagonal hollow section stub columns, given that
bending residual stresses have been inherently incorporated in the
measured stress–strain curves during tensile coupon tests and membrane
residual stresses are insignificant.
In terms of the modelling of boundary conditions in the tests, each
end section of the modelled stub column was coupled to a concentric
reference point. Subsequently, both reference points were fully
restrained except for the longitudinal translation of the top reference
point. Regarding the incorporation of the initial local geometric im­
perfections, the lowest elastic local buckling mode shape for each
stainless steel hexagonal hollow section stub column FE model, as ob­
tained from an elastic eigenvalue buckling analysis [44], was assumed to
be the local geometric imperfection distribution pattern. Then, five
imperfection magnitudes, including the measured value ω0 and four
generalised values (t/100, t/50, t/20 and t/10), were used to scale the
initial local geometric imperfection distribution patterns. Finally, static
Fig. 5. Stub column test setup.
Riks analyses [44] were performed on the developed stainless steel
hexagonal hollow section stub column FE models to obtain the numer­
testing machine [43]. Therefore, the end shortenings of each stub col­ ical failure loads, load-end shortening curves and failure modes.
umn specimen were corrected by subtracting the deformation of the end
platens of the machine from the LVDT reading, based on the strain gauge

4
Y. Zhong et al. Thin-Walled Structures 205 (2024) 112431

Table 3
Summary of stub column test results.
Specimen ID Nu (kN) δu (mm) Nu/Aσ0.2

B70-T1.5 121.06 0.67 0.75

B90-T1.5 122.24 0.62 0.61
B110-T1.5 125.70 0.85 0.52
B130-T1.5 130.18 1.15 0.45
B150-T1.5 134.05 1.15 0.40
B70-T2 216.12 0.91 0.90
B90-T2 237.59 0.80 0.75
B110-T2 244.95 0.83 0.64
B130-T2 245.82 1.11 0.55
B150-T2 249.46 1.85 0.48
B70-T4 692.22 8.90 1.37
B90-T4 799.83 6.90 1.23
B110-T4 893.80 3.90 1.13
B130-T4 953.20 3.49 1.03
B150-T4 1041.86 2.83 0.97

3.3. Validation of FE models

Validation of the developed stainless steel hexagonal hollow section

stub column FE models was conducted by comparing the numerical
results with their experimental counterparts. Table 4 reports the nu­
merical to test failure load ratios for all the stainless steel hexagonal
hollow section stub column specimens, revealing that (i) the measured
imperfection magnitude resulted in accurate and consistent predictions
of the test failure loads and (ii) the test failure loads were also well
predicted by using the generalised imperfection magnitude of t/50,
which were thus used in the parametric studies in Section 3.4. The
experimental and numerical load–end shortening curves for three
typical specimens B90-T1.5, B90-T2 and B90-4 are shown in Fig. 8,
demonstrating good agreement. The failure modes of the three typical
specimens are shown in Fig. 9, with the test failure modes well captured
by their numerical counterparts. In summary, the developed FE models
were capable of accurately simulating the test responses of the stainless
steel hexagonal hollow section stub column specimens and hence
regarded as validated.

3.4. Parametric studies

Parametric studies were conducted based on the validated stub col­

umn FE models, in order to generate additional numerical data over a
wide range of cross-section dimensions. The material properties of 2-
mm-thick stainless steel sheets were adopted for the parametric
studies, with flat and corner material properties assigned to the corre­
sponding flat and corner regions of the modelled hexagonal hollow
sections. In terms of the imperfection magnitude adopted, the general­
ised value of t/50 was used. The cross-section dimensions of the
modelled stainless steel hexagonal hollow section stub columns were
summarised as follows. Specifically, the edge widths of the cross-section
were chosen as 70 mm and 150 mm, covering the minimum and
maximum edge widths of the hexagonal hollow sections used in the
tests. The wall thicknesses were ranged from 1.0 mm to 15.8 mm,
resulting in the modelled stainless steel hexagonal hollow sections
covering Class 1 to Class 4 cross-sections in EN 1993-1-4 [24] and
non-slender and slender cross-sections in AISC 370-21 [25] and
ASCE/SEI 48-11 [26]. The inner corner radius of each modelled hex­
agonal hollow section stub column was taken as 1.5 times the wall
thickness, while the member length was set as three times the height of
the cross-section. Finally, a total of 100 parametric study results on
stainless steel hexagonal hollow section stub columns were generated.
Fig. 6. Measured load–end shortening curves of stainless steel hexagonal hol­
low section stub column specimens.

5
Y. Zhong et al. Thin-Walled Structures 205 (2024) 112431

Fig. 7. Test failure modes of stainless steel hexagonal hollow section stub column specimens.

is worth noting that all safety factors have been set to be unity in the
Table 4
following calculations.
Comparison between test and numerical failure loads.
Specimen ID FE Nu/Test Nu 4.2. Cross-section classification
ω0 t/10 t/20 t/50 t/100

B60-T1.5 1.04 1.05 1.09 1.15 1.18 EN 1993-1-4 [24] classifies cross-sections into four classes, namely
B75-T1.5 1.02 1.00 1.02 1.03 1.05 Class 1, 2, 3 and 4, while AISC 370-21 [25] and ASCE/SEI 48-11 [26]
B90-T1.5 1.03 1.02 1.03 1.04 1.04 categorise cross-sections into non-slender and slender cross-sections. For
B105-T1.5 1.04 1.04 1.05 1.05 1.05
Class 1–3 or non-slender cross-sections, the material 0.2 % proof stress
B120-T1.5 1.03 1.04 1.04 1.04 1.04
B60-T2 1.03 1.01 1.03 1.04 1.04
can be attained when local buckling occurs, resulting in the compression
B75-T2 1.04 1.02 1.06 1.11 1.12 resistance no less than the yield load. For Class 4 or slender
B90-T2 1.03 1.00 1.03 1.06 1.07 cross-sections, the premature local buckling occurs prior to the material
B105-T2 1.02 1.02 1.03 1.03 1.04 0.2 % proof stress, leading to the compression resistance being less than
B120-T2 1.04 1.04 1.04 1.05 1.05
the yield load. Classification of a hexagonal hollow section is made
B60-T4 0.97 0.85 0.90 0.96 0.97
B75-T4 0.95 0.87 0.91 0.96 0.99 through comparisons between the codified slenderness limits for inter­
B90-T4 0.94 0.87 0.92 0.97 0.99 nal plate elements in compression and the width-to-thickness ratio of the
B105-T4 0.95 0.89 0.93 0.98 1.01 critical constituent plate element. The three codified slenderness limits
B120-T4 0.95 0.87 0.91 0.95 0.97 √̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅ √̅̅̅̅̅̅̅̅̅̅̅̅̅
are given in Table 5, where εEC3 = 235/σ0.2 , εAISC = E/σ 0.2 and
Mean 0.99 0.94 0.97 1.01 1.02 √̅̅̅̅̅̅̅̅̅̅̅̅̅
εASCE = 1/σ 0.2 are the EC3, AISC and ASCE material parameters,
COV 0.04 0.09 0.07 0.06 0.05
respectively.
The test and numerical failure loads Nu, as normalised by yield loads
4. Assessment of existing design codes Aσ0.2, are plotted against the ratios of b/tεEC3, b/tεAISC and b/tεAISC in
Figs. 10–12, respectively, together with the corresponding codified
4.1. General slenderness limits for internal plate elements under compression. The
graphical evaluation results generally revealed that the slenderness
The current design codes, including EN 1993-1-4 (EC3) [24] and limits of EN 1993-1-4 [24] and AISC 370-21 [25] were accurate and safe
AISC 370-21 (AISC) [25] for stainless steel structures and ASCE/SEI when used for cross-section classification of stainless steel hexagonal
48-11 (ASCE) [26] for transmission pole structures, were evaluated for hollow sections under axial compression, while the slenderness limit
stainless steel hexagonal hollow sections under axial compression. Note specified in ASCE/SEI 48-11 [26] was unsafe.
that only ASCE/SEI 48-11 [26] specifies design rules typically for carbon
steel hexagonal hollow sections. For EN 1993-1-4 [24] and AISC 370-21 4.3. Cross-section compression resistances
[25], the design rules set out for stainless steel rectangular hollow sec­
tions were evaluated for their applicability to stainless steel hexagonal Both EN 1993-1-4 [24] and AISC 370-21 [25] adopt the yield load
hollow sections. Regarding the design of cross-sections under axial Aσ0.2 as the compression resistance for Class 1–3 or non-slender
compression, the same framework is employed by all three design codes cross-sections but the effective compression resistance Aeffσ0.2 for Class
– (i) cross-section classification based on slenderness limits and (ii) 4 or slender cross-sections. The effective cross-section area Aeff is taken
calculation of compression resistance of the categorised cross-sections. It as the summation of gross areas of the non-slender constituent plate

6
Y. Zhong et al. Thin-Walled Structures 205 (2024) 112431

Fig. 9. Test and numerical failure modes for typical specimens.

Table 5
Codified slenderness limits.
Code Slenderness limit

EN 1993-1-4 37εEC3
AISC 370–21 1.24εAISC
ASCE/SEI 48–11 681.2εASCE

λp is the EC3 plate element slenderness and determined from Eq. (4), in
which kσ= 4 is the EC3 buckling factor for internal plate elements in
compression. The AISC effective width is calculated from Eq. (5), where
fel is the elastic local buckling stress of the plate element in compression
and determined from Eq. (6), in which k=4 is the AISC buckling factor
for internal plate elements under axial compression and the Poisson’s
ratio ν=0.3.
( )
0.772 0.188
beff ,EC3 = b − 2
≤b (3)
λp λp

b/t
λp = √̅̅̅̅̅ (4)
28.4εEC3 kσ
( √̅̅̅̅̅̅̅̅)√̅̅̅̅̅̅̅̅
fel fel
beff,AISC = 0.772b 1 − 0.1 ≤b (5)
Fig. 8. Test and numerical load–end shortening curves for typical specimens. σ0.2 σ0.2

elements and effective areas of the slender constituent plate elements. π 2 E ( t )2

fel = k (6)
The effective area for a slender plate element prone to local buckling is 12(1 − ν2 ) b
calculated as the product of the effective width and wall thickness. The
ASCE/SEI 48-11 [26] stipulates the effective stress method to
EC3 effective width for internal plate elements is given by Eq. (3), where

7
Y. Zhong et al. Thin-Walled Structures 205 (2024) 112431

Fig. 12. Evaluation of ASCE slenderness limit for internal plate element in
Fig. 10. Evaluation of EC3 slenderness limit for internal plate elements in compression.
compression.

Table 6
Comparison of test and numerical failure loads with predicted cross-section
compression resistances.
Section type Nu/Nu,EC3 Nu/Nu,AISC Nu/Nu,ASCE Nu/Nu,ASCE*

Mean COV Mean COV Mean COV Mean COV

Non-slender 1.18 0.12 1.19 0.12 1.12 0.15 1.14 0.13

slender 1.02 0.03 1.03 0.03 1.11 0.24 0.99 0.03
Total 1.06 0.10 1.06 0.09 1.11 0.22 1.04 0.11

corresponding coefficients of variation (COVs), while the graphical

comparison results are displayed in Figs. 13–15, where the test and
numerical failure loads are normalised by the predicted resistances and
plotted against the ratios of b/tεEC3, b/tεAISC and b/tεASCE, respectively.
Based on the quantitative and graphical comparison results, it can be
concluded that (i) EN 1993-1-4[24] and AISC 370-21 [25] resulted in
overall good levels of design accuracy and consistency, but the predicted
resistances for non-slender cross-sections were conservative and

Fig. 11. Evaluation of AISC slenderness limit for internal plate element in
compression.

calculate the compression resistance of a hexagonal hollow section, with

the yield load Aσ0.2 and the effective compression resistance Aσ0.2,eff
respectively taken as the cross-section compression resistances for
non-slender and slender cross-sections, where σ 0.2,eff is the effective
compressive stress and given by Eq. (7).
⎧ ( )
⎪ 0.00114 b √̅̅̅̅̅̅̅̅ 681.2 b 919.62
⎪ 1.42σ0.2 1.0 − for √̅̅̅̅̅̅̅̅ < ≤ √̅̅̅̅̅̅̅̅
⎪
⎪ σ0.2
⎨ 2.62 t σ0.2 t σ 0.2
σ0.2 ,eff =
⎪
⎪ b 919.62
⎪
⎪ fel for > √̅̅̅̅̅̅̅̅
⎩ t σ 0.2
(7)
The EC3, AISC and ASCE cross-section resistance predictions were
calculated and compared against the test and numerical data. The
quantitative comparison results are reported in Table 6, including the
mean ratios of test and numerical to predicted failure loads and the Fig. 13. Comparison of test and numerical failure loads with EC3 predicted
cross-section compression resistances.

8
Y. Zhong et al. Thin-Walled Structures 205 (2024) 112431

effective compressive stress less than the yield stress. The reduction
factor can be expressed by Eq. (8) [52], where a1 and a2 are the constants
depending on boundary conditions, geometric imperfections and resid­
ual stresses. The ASCE effective compressive stress for slender stainless
steel hexagonal hollow sections was revised based on Eq. (8), with the
constants a1 and a2 determined using the least square method. The final
expression is given by Eq. (9). Table 6 summarises the mean Nu/Nu,ASCE*
ratios and the corresponding COVs, with the overall mean ratio of 1.04
and the corresponding COV of 0.11, indicating a significant improve­
ment in accuracy and consistency of resistance predictions compared to
their original ASCE counterparts. The same conclusion can be drawn
from Fig. 16, where a more accurate and consistent tendency of test and
numerical data is shown.
( √̅̅̅̅̅̅̅̅ )√̅̅̅̅̅̅̅̅
fel fel
ρ = a1 − a2 (8)
σ 0.2 σ0.2

( √̅̅̅̅̅̅̅̅)√̅̅̅̅̅̅̅̅
fel fel b 515
σ0.2,eff = σ 0.2 0.842 − 0.128 for > √̅̅̅̅̅̅̅̅ (9)
σ0.2 σ0.2 t σ 0.2

Fig. 14. Comparison of test and numerical failure loads with AISC predicted 5. Conclusions
cross-section compression resistances.
The local buckling behaviour and capacities of stainless steel hex­
agonal hollow sections have been experimentally and numerically
investigated in this paper. The testing investigation was conducted on
fifteen stainless steel hexagonal hollow section stub column specimens
under axial compression. The numerical investigation included devel­
opment and validation of FE models based on the test results and
parametric studies to expand the test data pool over a wide range of
cross-section dimensions. The test and numerical data were used to
evaluate the applicability of relevant design rules, as set out in EN 1993-
1-4 [24], AISC 370-21 [25] and ASCE/SEI 48-11 [26], to stainless steel
hexagonal hollow sections. On the basis of the evaluation results, the
following conclusions can be drawn:

(i) The EC3 and AISC slenderness limits were generally accurate and
safe when used for cross-section classification of stainless steel
hexagonal hollow sections in compression, while the ASCE slen­
derness limit was found to be unsafe.
(ii) EN 1993-1-4 [24] and AISC 370-21 [25] offered overall accurate
and consistent cross-section resistance predictions for stainless
steel hexagonal hollow sections. However, conservatism was

Fig. 15. Comparison of test and numerical failure loads with ASCE predicted
cross-section compression resistances.

scattered, due to the neglect of material strain hardening of stainless

steels and (ii) ASCE/SEI 48-11 [26] provided overall scattered and
relatively conservative cross-section compression resistance predictions,
but also with some unsafe predictions for those intermediate
cross-section sizes owing mainly to the unsafe slenderness limit.

4.4. Revised ASCE design approach (ASCE*)

In this section, a revised design approach was proposed based on

ASCE/SEI 48-11 [26] to improve the design accuracy and consistency.
√̅̅̅̅̅̅̅̅
Firstly, a new slenderness limit of 515/ σ0.2 was proposed for stainless
steel hexagonal hollow sections in compression based on the test and
numerical data displayed in Fig. 12. It can be seen that the proposed
slenderness limit was accurate and safe when used for the cross-section
classification of stainless steel hexagonal hollow sections in compres­
sion. For a slender plate element, the premature local buckling leads to a Fig. 16. Comparison of test and numerical failure loads with predicted cross-
reduction in average ultimate stress of the plate element, resulting in the section compression resistances from revised ASCE design method.

9
Y. Zhong et al. Thin-Walled Structures 205 (2024) 112431

observed for non-slender cross-sections due to the neglect of [15] R. Gonçalves, D. Camotim, Elastic buckling of uniformly compressed thin-walled
regular polygonal tubes, Thin-Walled Struct. 71 (2013) 35–45.
material strain hardening of stainless steels.
[16] R. Gonçalves, D. Camotim, Buckling behaviour of thin-walled regular polygonal
(iii) ASCE/SEI 48-11 [26] provided overall scattered and relatively tubes subjected to bending or torsion, Thin-Walled Struct. 73 (2013) 185–197.
conservative resistance predictions, but there were also some [17] J.Z. Liu, H. Fang, S. Chen, T.M. Chan, Material properties and residual stresses of
unsafe resistance predictions for those intermediate cross-section high strength steel hexagonal hollow sections, J. Constr. Steel. Res. 190 (2022)
107061.
sizes owing mainly to the unsafe slenderness limit. A revised [18] J.Z. Liu, H. Fang, T.M. Chan, Structural behaviour of high strength steel hexagonal
design approach was proposed and led to improved design ac­ hollow section stub columns under axial compression, Eng. Struct. 268 (2022)
curacy and consistency. 114653.
[19] J.Z. Liu, T.M. Chan, B. Young, Cross-section behaviour of cold-formed high
strength steel irregular hexagonal hollow section stub columns under combined
CRediT authorship contribution statement compression and bending, Adv. Struct. Eng. (2023) 13694332231157930.
[20] L. Gardner, Stability and design of stainless steel structures–Review and outlook,
Thin-Walled Struct. 141 (2019) 208–216.
Yukai Zhong: Writing – original draft, Visualization, Validation, [21] N.R. Baddoo, Stainless steel in construction: a review of research, applications,
Methodology, Investigation, Formal analysis, Conceptualization. Ke challenges and opportunities, J. Constr. Steel. Res. 64 (11) (2008) 1199–1206.
[22] M.T. Chen, Z. Gong, T. Zhang, W. Zuo, Y. Zhao, O. Zhao, et al., Mechanical
Jiang: Investigation, Formal analysis, Conceptualization. Andi Su:
behavior of austenitic stainless steels produced by wire arc additive manufacturing,
Investigation, Formal analysis. Jiyang Fu: Investigation. Airong Liu: Thin-Walled Struct. 196 (2024) 111455.
Investigation. Ou Zhao: Writing – review & editing, Supervision, Re­ [23] W. Zuo, M.T. Chen, S.W. Liu, X. Yun, O. Zhao, Y. Huang, B. Cheng, Experimental
sources, Project administration, Methodology, Investigation, Funding investigation on double-lap shear behavior of 3D printed austenitic stainless steel
bolted connections, Eng. Struct. 317 (2024) 118501.
acquisition, Formal analysis, Conceptualization. [24] prEN 1993-1-4:2022, Eurocode 3: Design of Steel Structures – Part 1.4: General
rules – Supplementary rules For Stainless Steels, European Committee for
Standardization (CEN), Brussels, 2022.
[25] AISC 370-21, Specification for structural stainless steel buildings,
Declaration of competing interest
AmericanInstitute of Steel Construction (AISC), 2021.
[26] ASCE/SEI 48-11, Design of Steel Transmission Pole Structures, American Society of
The authors declare that they have no known competing financial Civil Engineers, Reston, Virginia, 2011.
interests or personal relationships that could have appeared to influence [27] ASTM A240/A240M-18, Standard specification for chromium and chromium-
nickel stainless steel plate, sheet, and strip for pressure vessels and for general
the work reported in this paper. applications, ASTM International, 2018.
[28] ISO EN. 6892-1. Metallic materials-Tensile testing-Part 1: method of test at room
Data availability temperature. International Organization for Standardization, 2009.
[29] I. Arrayago, E. Real, L. Gardner, Description of stress–strain curves for stainless
steel alloys, Mater. Des. 87 (2015) 540–552.
Data will be made available on request. [30] B.W. S, T. P, Computational modelling of cold-formed steel: characterizing
geometric imperfections and residual stresses, J. Constr. Steel. Res. 47 (3) (1998)
193–210.
[31] M. Theofanous, L. Gardner, Testing and numerical modelling of lean duplex
Acknowledgements stainless steel hollow section columns, Eng. Struct. 31 (12) (2009) 3047–3058.
[32] S. Afshan, L. Gardner, Experimental study of cold-formed ferritic stainless steel
hollow sections, J. Struct. Eng. 139 (5) (2013) 717–728.
The authors would like to thank the Singapore Ministry of Education [33] I. Arrayago, E. Real, Experimental study on ferritic stainless steel RHS and SHS
Academic Research Fund (AcRF) Tier 1 Grant (Award Number: RG122/ cross-sectional resistance under combined loading, Structures, 4 (2015) 69–79.
21) for the financial support. [34] O. Zhao, L. Gardner, B. Young, Buckling of ferritic stainless steel members under
combined axial compression and bending, J. Constr. Steel. Res. 117 (2016) 35–48.
[35] J. Chen, Y. Huang, B. Young, Design of austenitic and duplex stainless steel SHS
References and RHS beam-columns, J. Constr. Steel. Res. 152 (2019) 143–153.
[36] B. Zheng, G. Shu, F. Xie, Q. Jiang, Design of cold-rolled stainless steel rectangular
hollow section columns, J. Constr. Steel. Res. 170 (2020) 106072.
[1] F. Nowzartash, M. Mohareb, Column curves for elliptical hollow section members,
[37] R. Kim, T. Kim, S. Im, Y. Xi, Hysteretic behavior comparison of austenitic and lean
J. Constr. Steel. Res. 67 (10) (2011) 1525–1536.
duplex stainless steel square hollow section members under cyclic axial loading,
[2] C. Fang, F. Zhou, W. Wu, Performance of elliptical hollow sections under combined
Eng. Struct. 237 (2021) 112227.
compression and cyclic bending, J. Struct. Eng. 144 (8) (2018) 04018102.
[38] W. Zuo, M.T. Chen, B. Young, Structural behaviour of cold-formed steel elliptical
[3] Q. Shen, J. Wang, J.R. Liew, B. Gao, Q. Xiao, Experimental study and strength
hollow section stub columns after exposure to ISO-834 fire curve, Thin-Walled
evaluation of axially loaded welded tubular joints with round-ended oval hollow
Struct. 197 (2024) 111309.
sections, Thin-Walled Struct. 154 (2020) 106846.
[39] A. Su, H. Yang, Y. Wang, Y. Wang, Experimental and numerical investigations of
[4] C. Fang, F. Wang, C. Wang, Y. Zheng, Cyclic behavior of oval hollow section (OHS)
S960 hot-rolled ultra-high strength steel seamless circular hollow section
beam-columns, Thin-Walled Struct. 161 (2021) 107430.
beam–columns, Thin-Walled Struct. 200 (2024) 111909.
[5] M.S. Abdelaty, S.S. Safar, Experimental and numerical investigation on strength of
[40] J. Zhang, A. Su, H. Yang, J. Zhao, Y. Wang, Experimental and numerical
eccentrically loaded steel oval hollow sections, Design Eng. (2021) 15123–15141.
investigations of S890 and S960 ultra-high strength steel circular hollow section
[6] M.T. Chen, B. Young, Experimental and numerical investigation on cold-formed
columns, Eng. Struct. 311 (2024) 118143.
steel semi-oval hollow section compression members, J. Constr. Steel. Res. 151
[41] Y. Cui, J. Zhang, C. Ma, M. Niu, K. Jiang, S. Li, A. Su, Testing, numerical modelling
(2018) 174–184.
and design of G550 high strength cold-formed steel built-up section columns, Thin-
[7] T.C. Michael, A.R. Veerappan, S. Shanmugam, Suitability of assumed cross sections
Walled Struct. 196 (2024) 111529.
to include ovality in the limit analysis of pipe bends under in-plane closing moment
[42] Y. Wang, A. Su, H. Yang, J. Zhao, Y. Wang, S890 hot-rolled ultra-high strength steel
and internal pressure, Thin-Walled Struct. 122 (2018) 545–553.
(UHSS) seamless circular hollow section beam–columns: testing, modelling and
[8] J.Z. Liu, H. Fang, J. Guo, S. Li, T.M. Chan, Numerical investigation and design for
design, Thin-Walled Struct. 197 (2024) 111583.
the local buckling behaviour of high strength steel hexagonal hollow section stub
[43] Centre for Advanced Structural Engineering, Compression Tests of Stainless Steel
columns under axial compression, Thin-Walled Struct. 186 (2023) 110717.
Tubular Columns. Investigation Report S770, University of Sydney, 1990.
[9] W. Xu, L.H. Han, W. Li, Performance of hexagonal CFST members under axial
[44] ABAQUS/Standard user’s manual volumes I-III and ABAQUS CAE Manual. Version
compression and bending, J. Constr. Steel. Res. 123 (2016) 162–175.
6.14, Karlsson Hibbitt, Inc Sorensen, PawtucketUSA, 2014.
[10] M.F. Hassanein, V.I. Patel, M. Bock, Behaviour and design of hexagonal concrete-
[45] C. Buchanan, E. Real, L. Gardner, Testing, simulation and design of cold-formed
filled steel tubular short columns under axial compression, Eng. Struct. 153 (2017)
stainless steel CHS columns, Thin-Walled Struct. 130 (2018) 297–312.
732–748.
[46] A. Mohammed, S. Afshan, Numerical modelling and fire design of stainless steel
[11] F.X. Ding, Z. Li, S. Cheng, Z.W. Yu, Composite action of hexagonal concrete-filled
hollow section columns, Thin-Walled Struct. 144 (2019) 106243.
steel tubular stub columns under axial loading, Thin-Walled Struct. 107 (2016)
[47] M.L. Patton, K.D. Singh, Numerical modeling of lean duplex stainless steel hollow
502–513.
columns of square, L-, T-, and+-shaped cross sections under pure axial
[12] M. Ahmed, Q.Q. Liang, Numerical analysis of concentrically loaded hexagonal
compression, Thin-Walled Struct. 53 (2012) 1–8.
concrete-filled steel tubular short columns incorporating concrete confinement,
[48] A. Su, Y. Wang, K.J. Rasmussen, L. Gardner, Structural performance and design of
Adv. Struct. Eng. 24 (15) (2021) 3472–3487.
S960 ultra-high strength steel non-slender welded I-sections subjected to combined
[13] T. Aoki, Y. Migita, Y. Fukumoto, Local buckling strength of closed polygon folded
loading, Eng. Struct. 293 (2023) 116593.
section columns, J. Constr. Steel. Res. 20 (4) (1991) 259–270.
[14] Y. Migita, T. Aoki, Y. Fukumoto, Local and interaction buckling of polygonal
section steel columns, J. Struct. Eng. 118 (10) (1992) 2659–2676.

10
Y. Zhong et al. Thin-Walled Structures 205 (2024) 112431

[49] A. Su, Y. Wang, Y. Wang, K.J. Rasmussen, L. Gardner, Behaviour and design of [51] W. Zuo, M.T. Chen, B. Young, Post-fire behavior of cold-formed steel semi-oval
S960 ultra-high strength steel non-slender welded I-section beam–columns, Eng. hollow stub columns, J. Struct. Eng. 150 (10) (2024) 04024141.
Struct. 304 (2024) 117602. [52] N.S. Trahair, M.A. Bradford, D. Nethercot, L. Gardner, The Behaviour and Design of
[50] J. Zhang, A. Su, K. Jiang, Post-fire behavior and residual resistances of S890 ultra- Steel Structures to EC3, CRC Press, 2017.
high strength steel circular hollow sections under combined compression and
bending, Eng. Struct. 316 (2024) 118519.

11