See

discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/308181187

Behavior of Thin-walled Circular Hollow

Section Stub Columns Under Axial
Compression

Article in International journal of steel structures · September 2016

DOI: 10.1007/s13296-015-0159-0

CITATION READS

1 73

4 authors, including:

Lanhui Guo Yong Liu

Harbin Institute of Technology Harbin Institute of Technology
48 PUBLICATIONS 300 CITATIONS 2 PUBLICATIONS 1 CITATION

SEE PROFILE SEE PROFILE

All content following this page was uploaded by Lanhui Guo on 23 January 2017.

The user has requested enhancement of the downloaded file. All in-text references underlined in blue are added to the original document
and are linked to publications on ResearchGate, letting you access and read them immediately.
International Journal of Steel Structures 16(3): 777-787 (2016)
DOI 10.1007/s13296-015-0159-0
ISSN 1598-2351 (Print)
ISSN 2093-6311 (Online)

www.springer.com/journal/13296

Behavior of Thin-walled Circular Hollow Section

Stub Columns Under Axial Compression
Lanhui Guo1,*, Yong Liu1, Hui Jiao2, and Shilong An1
1
Ph.D., Professor, School of Civil Engineering, Harbin Institute of Technology, 73 Huanghe-road, Harbin, 150090, China
2
Ph.D., Senior Lecturer, School of Engineering and ICT, University of Tasmania, 65 Hobart, Tasmania 7001, Australia

Abstract

Thin-walled steel circular hollow section (CHS) is widely used in wind turbine towers, where the tube is subjected to axial
load combined with bending moment. Understanding the behavior of axially loaded thin-walled tubes with large diameter-to-
thickness (D/t) ratios is crucial for the design of such structures. To investigate the behavior of axially loaded thin-walled
circular hollow section steel tubes, 16 stub columns were tested with the D/t ratio ranging from 75 to 300. The experimental
results showed that the compressive strength decreased sharply with the increase of the D/t ratio. The experimental result
indicated that the tubes with smaller D/t ratios failed with outward deformation without visible local buckling. With the increase
of the D/t ratio, local buckling appeared at the peak load. The finite element method was used to model the behavior of axially
loaded stub columns. The experimental results were used to validate the accuracy of the finite element results. Based on the
finite element model, the influences of the D/t ratio, the initial geometric deformation and the initial residual stress were
analyzed. Design guidelines in Eurocode3, AISC-LRFD, AS4100 and S16-9 on thin-walled steel members were used to
compare with the experimental and calculated results. Among them, the results based on AS4100 were the closest to the
experimental results.

Keywords: thin-walled, circular hollow section, stub columns, diameter-to-thickness ratio, local buckling

1. Introduction horizontal load exerted by wind or earthquake. While the

tower tube is under the combination of the axial load and
Wind energy is gaining great attention nowadays as a the bending moment, the behavior of the tube under an
mature type of cost-effective renewable energy. The last axial load is the basis for the design of the tube under
two decades have seen significant growth in the use of combined load cases. As circular hollow section (CHS) is
wind power throughout the world (Dimopoulos and Gantes, the most common type of section used as turbine towers,
2013). With this growth, the wind turbine is starting to this paper studies the behavior of axially loaded CHS stub
play a critical role in the world power infrastructures. columns. Literatures by Batterman (1965) showed that
Various design approaches have been explored for turbine the capacity of CHS members was related to the diameter-
towers. The primary design alternatives considered for a to-thickness (D/t) ratio of the tube. Since tubes with
turbine tower are a truss or a tube structure, such as the larger D/t ratios were commonly used as wind turbine
electrical transmission towers. Comparing with truss towers, towers, it is necessary to study the behavior of CHS tubes
a tube tower is more convenient in terms of fabrication with larger D/t ratios.
and maintenance. As a result, steel tube towers are In the past, most of research was focused on steel tubes
superior to truss towers and are commonly used in wind with smaller D/t ratios (Chen and Ross, 1977; Elchalakani
turbines. et al., 2001; Jiao and Zhao, 2003; Gardner and Nethercot,
A tower tube needs to carry the vertical load from the 2004a,b; Guo et al., 2013). This may be due to the fact
rotor, the hub and the nacelle. Also, it is subjected to the that the majority of CHS tubes in the building or
automotive industry were compact or non-compact sections
Received September 28, 2015; accepted May 22, 2016; with a D/t ratio less than 200 (Liu and Day, 2008). As we
published online September 30, 2016 know, structural tubes can be manufactured by several
© KSSC and Springer 2016 methods. The manufacturing process may produce different
*Corresponding author
level of residual stresses and geometrical imperfections,
Tel: +86-451-86289100, Fax: +86-451-86289100 which will affect their structural performance. Therefore,
E-mail: [email protected] the slenderness limit given in a design code may be
778 Lanhui Guo et al. / International Journal of Steel Structures, 16(3), 777-787, 2016

suitable for certain steel tubes which are similar to those still ensure that the specimens would be stub columns
used in the calibration of the design rules but might not with little effect from the column slenderness. The details
suitable for other tubular sections. Many tubes are of the specimens are listed in Table 1. Each tube was
commonly manufactured by rolling a steel sheet into a manufactured by rolling a steel sheet into a tube and seam
tube and seam welding longitudinally. welding longitudinally. The thickness of the steel tubes
To study the behavior of axially loaded tower tubes was 2 mm. Two 10 mm thickness end plates were welded
with larger D/t ratios, a total of 16 specimens were at both ends of each steel tube.
fabricated and tested. This study aims to analyze the axial To determine the steel material properties, three tensile
behavior of CHS with a D/t ratio beyond 100. In order to coupons were cut out from each steel sheet and tested in
compare the test results with previous studies, tubes with accordance with the Chinese standard GBJ2975 (1982).
D/t ratios ranging from 75 to 100 were also tested. The The average material properties are listed in Table 2,
test results were checked against the current steel design where t is the steel plate thickness, f y, f u, Es, ν and δ are
standards, including the American ANSI/AISC 360-10 the steel yield stress, the tensile strength, the elastic modulus,
(2010), Eurocode 3 (2007), Australian AS4100 (2012) and the poisson ratio and the elongation ratio, respectively.
Canadian S16-9 (2010). For thin-walled circular hollow sections subjected to
axial load, the capacity of the tubes is sensitive to initial
2. Experimental Study geometrical imperfections (deviations from nominal perfect
cylindrical shape), especially when a specimen fails by
2.1. Design and fabrication of CHS tubes local buckling. In this study, a laser displacement meter,
Sixteen specimens with the D/t ratios ranging from 75 KEYENCE LK-G80, was used to measure the initial
to 300 were selected to cover the tube behavior from fully geometrical imperfection of each specimen. The initial
effective through to local buckling. A length-to-diameter out-of-straightness of all stub column specimens were
ratio of 3.0 was chosen to reduce the end effects and yet measured along the whole length and listed in Table 3.

Table 1. Geometric parameters of specimen 2.2. Experimental setup

The specimens were properly instrumented to measure
D t L Number of the structural response and to identify the failure modes.
Se. D/t L/D
(mm) (mm) (mm) repeats
For this purpose, a total of eight strain gauges were
DT75-1 installed at four locations on the steel tube at the mid-
DT75-2 75 150 2 450 3 3 height of each specimen. Two gauges at each location
DT75-3
were arranged at a 90o angle to measure the longitudinal
DT100-1 and transverse strains as shown in Fig. 1(b). Four linear
DT100-2 100 200 2 600 3 3 variable differential transducers (LVDTs) were used to
DT100-3 measure the axial compressive deformations. The arrangement
DT125-1 of the LVDTs used in the test is illustrated in Fig. 1(a).
DT125-2 125 250 2 750 3 3 The test was conducted on a 5,000 kN capacity compression
DT125-3 machine. The load was measured by using a load transducer.
DT150-1 The load was applied in accordance with Chinese code
DT150-2 150 300 2 900 3 3 JGJ 101-96 (1996). In the elastic range, the vertical load
DT150-3 was applied with an increment of one-tenth of the calculated
DT200-1 capacity of the specimen. Beyond the elastic range, the
200 400 2 1200 3 2
DT200-2 displacement control loading method was adopted.
DT300-1
300 600 2 1800 3 2
DT300-2 3. Experimental Phenomena
Table 2. Steel properties With the two 10 mm thick end plates that were welded
at both ends of the steel tube, both ends of the tube walls
T (mm) f y (MPa) f u (MPa) Es (MPa) ν δ
5
were restrained from rotation so that a fixed end condition
2.0 190 315 1.88×10 0.36 0.51 was provided. Most of the specimens failed at a position

Table 3. Initial geometric imperfection (Unit: mm)

Specimen DT75-1 DT75-2 DT75-3 DT100-1 DT100-2 DT100-3 DT125-1 DT125-2
Imperfection 0.118 0.242 0.134 0.585 0.558 0.373 1.681 0.537
Specimen DT125-3 DT150-1 DT150-2 DT150-3 DT200-1 DT200-2 DT300-1 DT300-2
Imperfection 0.606 1.551 0.651 0.371 2.212 2.302 3.124 4.168
 Behavior of Thin-walled Circular Hollow Section Stub Columns Under Axial Compression 779

Figure 1. Experimental setup.

Figure 2. Experimental phenomena of specimens with D/t Figure 3. Experimental phenomena of specimens with D/t
of 75 and 100. of 125 and 150.

approximately D/4 away from one end of the tube, whereas reason was that the specimen was in hardening stage
some failed at the mid-height, such as specimens DT125- when it reached its peak load. The average vertical strain
2 and DT300-1.The different failure positions may be due in specimen DT75-2 was 5640 με. Beyond the peak load,
to the initial geometric deformation positions. the load-carrying capacity dropped gradually with the
For specimens with D/t of 75 and 100, the outward increase of vertical displacement. The specimen exhibited
deformation appeared when the specimens reached their good compressive behavior. The experimental phenomena
peak loads. An ‘elephant’s foot’ buckle, i.e., one prominent are shown in Fig. 2.
outwards circumferential buckle occurred at one end of For specimens with D/t of 125 and 150, the ‘elephant’s
the tube. This buckle developed into a plastic mechanism. foot’ buckle and inward wave-like multiple ripples were
The smooth outward deformation was the typical failure found on the steel tube. The reason was that the stub column
mode for compact tubes with smaller D/t ratio of 75. The failed at the elastic-plastic stage. The experimental phenomena
780 Lanhui Guo et al. / International Journal of Steel Structures, 16(3), 777-787, 2016

Figure 4. Experimental phenomena of specimens with D/t of 200 and 300.

Figure 5. Axial load-vertical displacement relationship curves.

are shown in Fig. 3. The buckle deformation was also bottom (except for specimen DT300-1) appeared at the
very smooth. The load-carrying capacity dropped dramatically peak load accompanied by a sudden drop of the load-
at the collapse. carrying capacity and the increase of vertical displacement.
Figure 4 shows the failure modes of specimens with D/t The reason was that the specimen failed at the elastic
ratios of 200 and 300. The inward local buckling near the stage of steel because of initial geometric imperfection.
 Behavior of Thin-walled Circular Hollow Section Stub Columns Under Axial Compression 781

Figure 6. Comparison of axial stress-strain relationship curves.

Figure 7. Finite element model.

From these four specimens, it can be seen that the larger

the D/t ratio, the more severe the local buckling that is
indicated by the number of buckling ripples on the steel
tubes. These failure modes are clearly different from
those of specimens with D/t less than 150. It can be seen
that local buckling failure mode appeared for specimens Figure 8. Stress-strain relationship of steel.
with D/t ratio ranging from 200 to 300.

4. Analysis of Experimental Results DT75-2, the maximum compressive stress was 210 MPa
with the corresponding compressive strain of 5640 με. It
Figure 5 shows the axial load-deformation relationship means that the specimen reached its hardening stage. Thus
curves. For thin-walled stub columns subjected to axial the smooth outward deformation appeared for specimens
load, the initial imperfection has significant influence on with D/t of 75, as shown in Fig. 2. When the D/t ratio is
the strength of thin-walled shells. In the test, the effect of 100, the maximum compressive stress was 171 MPa with
initial imperfection on the strength of tubes with D/t of 75 the corresponding strain of 1620 με. The maximum
seems insignificant because of the relatively small D/t compressive stress was lower than the steel yield stress,
ratios. When the D/t is over 100, the difference on load- but higher than 0.8 f y, which meant that the stub column
carrying capacities of each group becomes significant, failed at the elastic-plastic stage. For specimens DT100-
especially for specimens with D/t of 300. Hence, cares 3, DT125-3 and DT150-3, their maximum compressive
should be taken to reduce the initial geometrical imperfection stresses were near to each other. When the D/t ratio varied
in the fabrication and transportation. from 100 to 150, whose influence on the maximum
Figure 6 shows the comparison of the average compressive compressive stress was not obvious. For specimen DT200-
stress vs. vertical strain relationship curves. It can be seen 1 with D/t of 200, the maximum compressive stress was
that both the maximum compressive stress and the ductility 143.3 MPa, and the corresponding compressive strain
decrease with the increase of D/t ratio. For specimen was 900 με, which meant that the steel tube failed at the
782 Lanhui Guo et al. / International Journal of Steel Structures, 16(3), 777-787, 2016

Figure 9. Comparison between experimental and calculated results.

elastic stage. The specimen failed by local buckling as 5. FE Analysis of Thin-walled CHS
shown in Fig. 4. For specimen DT300-1 with D/t of 300, Subjected to Axial Load
the maximum compressive stress was 111.5 MPa with the
corresponding compressive strain of 640 με, which also The commercial software ABAQUS was used to develop
failed by local buckling. the finite element (FE) model to analyze the axially
From the comparison, it can be seen that the load- loaded thin-walled CHS. The shell element S4R was used
carrying capacity of the axially loaded stub columns to model the behavior of the steel tubes, which was
decreased with the increase of D/t ratio. For specimen suitable for both thin-walled and thick shell elements.
with D/t of 75, it failed by yielding with smooth outward The finite element model is shown in Fig. 7. An initial
deformation failure mode. When the D/t ratio reached deformation resembling the eigen-buckling mode was
100, the specimen failed by local buckling in the elastic- introduced into the finite element model. A value was
plastic stage. When the D/t ratio is larger than 200, the assumed as an infill panel displacement for the first mode
specimen failed by local buckling in the elastic stage of to represent an initial deformed shape of the model prior
the steel material with inward buckling ripples local buckling to the analysis. At the bottom, the deformation and rotation
failure mode. along x, y, z axes were restrained. At the top surface, the
 Behavior of Thin-walled Circular Hollow Section Stub Columns Under Axial Compression 783

Figure 10. Failure modes comparison between the calculated and experimental phenomena.

Figure 11. Influence of initial geometric imperfection on axial load-displacement relationship curves.

deformation along x axis and y axis was restrained, and

the rotations along x, y and z axes were fixed. This
boundary condition was applied to the reference points at
center of bottom and top surface, respectively. These
reference points were attached to the bottom and top
surfaces by Rigid Body constraint, which was used to
model the behavior of stub columns welded with the end
plates. A vertical displacement was added at the top surface,
which was controlled by displacement.
The material properties, including Young’s modulus for
steel Es of 2.0×105 MPa and Poisson’s ratio vs of 0.3,
were input in the ABAQUS model. The steel was
assumed to behave as an elastic-plastic material without
considering the hardening of steel. The idealized stress- Figure 12. Influence of initial geometric imperfection on
strain relationship curve used in the analysis is shown in f m/f y-λ relationship curves.
Fig. 8.
Figure 9 shows the comparison between the experimental
results and calculated results. In the analysis of test whereas for specimens with D/t of 200 and 300, the
specimens, the measured stress-strain relationship values inward multi-buckling ripples failure appeared.
were adopted. It can be seen that they coincide well. In the following, the influences of the initial geometric
Figure 10 shows the comparison between the calculated imperfection, the residual welding stress, the element
and experimental phenomena. For specimens with D/t of slenderness and the steel yield strength were analyzed
75, the typical failure mode is outward ring failure mode. using the FE model. The element slenderness λ is used to
For specimen D/t of 150, the inward smooth wave appeared, determine the local buckling effects. For circular hollow
784 Lanhui Guo et al. / International Journal of Steel Structures, 16(3), 777-787, 2016

Figure 13. Distribution of residual stress.

sections, the element slenderness is calculated as: the influence of the initial geometric imperfection is more
significant with the increase of the element slenderness.
D fy -⎞
λ = ---- ⎛ --------
This proves that the initial geometrical imperfections
(1)
t ⎝ 235⎠ influence the load-carrying capacity of tubes with larger
D/t ratios. This result suggests that cares must be taken to
where D and t are the outer diameter and the wall avoid significant initial geometrical imperfection during
thickness of the tube respectively; and f y is the steel yield the fabrication and transportation of steel tubes with
stress. larger D/t ratios.

5.1. Influence of initial geometric imperfection 5.2. Influence of residual stress

In this part, a circular cross section with an outside Longitudinal welding of the tube resulted in residual
diameter of 3 m was analyzed. A tube with this diameter stresses. According to Chen (1977), the measured residual
has been commonly used as wind turbine towers. The stress distribution of half section is shown in Fig. 13(a).
steel yield strength was 235 N/mm2. The D/t ratio varied In which, x is the distance to the weld position, and R is
from 50 to 500 with the corresponding tube thickness the radius of the tube. The residual stress in the distribution
changing from 60 to 6 mm. According to the ANSI/AISC represented an average stress through the wall thickness.
360-10 (2010) design guideline, the allowable geometric The tensile residual stress was assumed to be positive. It
imperfection ω can be calculated as follows: is noted that the material near the weld has effectively
yielded in tension. In any such distribution, the resultant
ω
---- = ---------- r-
1 axial force must, of course, be zero. For simplification,
(2)
t 16.5 t the distribution of residual stress is assumed as shown in
Fig. 13(b).
The initial geometric imperfections of 0.5ω, ω, 2ω, and The residual stress was analyzed in the FE model. The
3ω were selected in the FE model. Figure 11 shows the outside diameter of the tube was 3 m. The steel yield
comparison of axial load-deformation relationships under strength was 235 N/mm2. The D/t ratio varied from 100
different initial geometric imperfections. For tubes with to 500 with corresponding tube thickness varying from 30
D/t of 100, the initial geometric imperfection had no to 6 mm. Figure 14 shows the analytical results of the
significant influence on the load-deformation relationship average axial load-displacement curves. It can be seen
curves. This was due to the fact that the steel tube was an that for stub columns with D/t of 100, the influence of the
impact section. Under the axial load, the tubes reached residual stress was not significant. In specimens with
the yield strength at the peak load. higher D/t ratio, the residual stress resulted in the more
While for tubes with D/t of 300, the influence of the decrease of the load-carrying capacity. For specimens
initial geometric imperfection became significant. With the with D/t of 500, the load-carrying capacity decreased by
increase of the initial geometric imperfection, the load- 16.5% comparing with specimens without residual stress.
carrying capacity decreased obviously. Beyond the peak This may be due to the fact that the compressive residual
load, the decreasing trend of the load was more dramatic stress resulted in the early local buckling of the tube. The
with the increase of the initial geometric imperfection. residual stress should be considered for specimens with
Figure 12 shows the ratio of f m/f y versus λ curves for large D/t ratios in design and analysis.
tubes with different imperfections to illustrate the influence
of the initial geometric imperfection on the load-carrying 5.3. Influence of steel yield strength
capacity of the tubes, in which f m is the peak average The influence of the steel yield strength on the load-
stress of a tube under the axial load. It can be seen that carrying capacity was analyzed. The outside diameter of
 Behavior of Thin-walled Circular Hollow Section Stub Columns Under Axial Compression 785

Figure 14. Influence of residual welding stress on axial Figure 15. Influence of steel yield strength on axial load-
load-displacement relationship curves. displacement relationship curves.

the tube was 3 m. The D/t ratio varied from 50 to 500 6. Comparison with Available Steel
with corresponding tube thickness varying from 60 to 6 Specifications
mm. The steel yield strength varied from 235 to 390
MPa. The residual stress with the distribution shown in A number of specifications are available for the design
Fig. 11 was also considered. Figure 15 shows the influence of circular tubes in compression including the American
of the steel yield strength on the axial load-axial displacement ANSI/AISC 360-10 (2010), Eurocode 3 (2007), Australian
relationship curves. It can be seen that the load-carrying AS4100 (2012) and Canadian S16-9 (2010). For design
capacity increased obviously with the increase of the steel purpose, it’s extremely important to have a routine design
yield strength. If higher strength steel was used in design, procedure which is representative, and conservative in
higher load-carrying capacity could be obtained, although predicting the ultimate strength. The experimental results
the steel yield strength had little influence on the f m /f y-λ were used to evaluate the design codes, including ANSI/
relationship curves as shown in Fig. 16. AISC 360-10 (2010), EC3 (2007), AS4100 (2012) and S16-
786 Lanhui Guo et al. / International Journal of Steel Structures, 16(3), 777-787, 2016

Figure 16. f m/f y-λ relationship curves.

9 (2010) specifications. In EC3 (2007), the fabrication following conclusions were made.
equality was considered by using different index Q. For specimens with D/t less than 100, they failed by
Some test data from other references (Pan et al., 2006; yielding with outward deformation. No local buckling
Jiao and Zhao, 2003; OShea and Bridge, 1997; Bardi was observed. With the increase of D/t ratio, local buckling
and Kyriakides, 2006) were also applied to validate the failure mode with multi-ripples became more pronounced,
design specifications. Figure 16 shows the comparison especially for specimens with D/t of 200 and 300.
results based on the above mentioned design specifications With the increase of D/t ratio, the load-carrying capacity
together with the test results. Also, the FEM results are of the tubes appeared being influenced by the initial
compared with experimental results with the initial geometric imperfection and the initial residual stress. In
geometric imperfection of 2ω. The finite element results the fabrication and transportation process, cares should be
are near to the results based on EC3 (2007) (Q=24). For taken to reduce initial geometrical imperfection. While
stubs with λ less than 200, the calculated results are near the steel yield strength had little influence on the f m/f y-λ
to the experimental results. The experimental results with relationship, the steel yield strength increased the load-
λ over 200 are lower than the calculated results, the carrying capacity of short tubes under axial load.
reason was that the large initial geometric imperfection The test data were compared with the design guidelines
resulted in the decrease of loading-capacity. As shown in of ANSI/AISC 360-10 (2010), AS4100 (2012) and EC3
Fig. 16, the different design curves were adopted by (2007) on thin-walled steel members. It was found the
different design specifications in calculating the load- results based on Canadian design code S16-09 (2010)
carrying capacities of tubes. In the comparison of the test was the most conservative with the Australian design code
data with the design guidelines of ANSI/AISC 360-10 AS4100 (2012) being the most suitable one in designing
(2010), EC3 (2007), AS4100 (2012), and S16-9 (2010) the tubes tested in this paper.
on thin-walled steel members, it was found the results
based on Canadian design code S16-09 (2010) was the Acknowledgments
most conservative. The Australian design code AS4100
(2012) seemed more suitable in designing the tubes tested The project is supported by the Harbin technological
in this paper. The American design code ANSI/AISC innovation talent project (2013RFQXJ120) and Scientific
360-10 (2010) appeared overestimating the strength of the Research Foundation for the Returned Overseas Chinese
tubes with large D/t ratios. Scholars, which are gratefully acknowledged.

7. Summary and Conclusions Nomenclature

D: outside diameter of tubes
This paper presented the results of an experimental t: thickness of steel tubes
study on thin-walled steel circular hollow section (CHS) r: outer radius of steel tubes
subjected to axial compression. A total of sixteen specimens ω: initial geometric imperfection
were fabricated and tested, with D/t ratio ranging from 75 Es : young’s modulus of elasticity
to 300. Also, the finite element software ABAQUS was f m : peak compressive stress
used to analyzed the axial loading behavior of thin-walled f y : steel yield stress
tubes. Based on the experimental and analytical study, the f u : ultimate tension strength
 Behavior of Thin-walled Circular Hollow Section Stub Columns Under Axial Compression 787

L: length of specimen Eurocode 3 (2007). Design of steel structures. part 1-6:

t: thickness of steel tube strength and stability of shell structures, pp. 40-75.
λ: generalized section slenderness Gardner, L. and Nethercot, D. A. (2004a). “Experiments on
ν: poisson ratio of steel stainless steel hollow sections-Part 1: Material and cross-
δ: elongation ratio of steel sectional behavior.” Journal of Constructional Steel
Research, 60, pp. 1291-1318.
Gardner, L. and Nethercot, D. A. (2004b). “Experiments on
References stainless steel hollow sections-Part 2: Member behaviour
of columns and beams.” Journal of Constructional Steel
ANSI/AISC 360-10 (2010). Specification for structural steel Research, 60, pp. 1319-1332.
buildings. American Institute of Steel Construction. Guo, L. H., Yang, S. J., and Jiao, H. (2013). “Behavior of
AS4100 (2012). SAA steel structures code. Standards thin-walled circular hollow section tubes subjected to
Association of Australia. bending.” Thin-Walled Structures, 43(12), pp. 281-829.
Bardi, F. C., Kyriakides, S. (2006). Plastic buckling of Jiao, H., and Zhao X. L. (2003). “Imperfection, residual
circular tubes under axial compression-part I: Experiments. stress and yield slenderness limit of very high strength
International Journal of Mechanical Sciences, 48(8), pp. (VHS) circular steel tubes.” Journal of Constructional
830-841. Steel Research, 59, pp. 233-249.
Batterman, S. C. (1965). “Plastic buckling of axially Liu, Y. C. and Day, M. L. (2008). “Bending collapse of thin-
compressed cylindrical shells.” AIAA Journal, 3, pp. 316- walled circular tubes and computational application.”
325. Thin-Walled Structures, 46, pp. 442-450.
Chen, W. F. and Ross, D. A. (1977). “Tests of fabricated O’Shea M. D., Bridge R. Q. (1997). Local Buckling of thin-
tubular columns.” Journal of Structural Engineering, walled circular steel sections with or without internal
103(3), pp. 619-634. restraint. Journal of Constructional Steel Research, 41(2-
Dimopoulos, C. A. and Gantes, C. J. (2013). “Comparison of 3), pp. 137-157.
stiffening types of the cutout in tubular wind turbine Pan, H., Liang S., Pei, S. and et al. (2006). Stability analysis
towers.” Journal of Constructional Steel Research, 83(4), of large-sized thin-walled tubes members subjected to
pp. 62-74. axial force. Steel Construction, 4(21), pp. 5-8. (in Chinese)
Elchalakani, M., Zhao, X. L., and Grzebieta, R. H. (2001). S16-09 (2010). Design of steel structures. Canadian Standards
“Plastic slenderness limits for cold-formed circular hollow Association, pp. 35-40.
sections.” Australian Journal of Structural Engineering,
Institute of Engineers, 3(3), pp. 127-141.

View publication stats