THE STRUCTURAL DESIGN OF TALL AND SPECIAL BUILDINGS

Struct. Design Tall Spec. Build. 23, 403–426 (2014)

Published online 3 October 2012 in Wiley Online Library (wileyonlinelibrary.com). DOI: 10.1002/tal.1046

Behaviour of uni-axially loaded concrete-filled-steel-tube columns

confined by external rings

M. H. Lai and J. C. M. Ho*

Department of Civil Engineering, The University of Hong Kong, Hong Kong

SUMMARY
Concrete-filled-steel-tube (CFST) columns have been widely adopted for column construction of tall buildings
due to its superior strength and ductility performance contributed by the composite action. However, this
beneficial composite action cannot be fully developed at early elastic stage as steel dilates more than concrete
and thereby causing imperfect interface bonding. Hence, it reduces the elastic strength and stiffness of the CFST
columns. To resolve the problem, external confinement in the form of steel rings is proposed in this study to
restrict the lateral dilation of concrete and steel at initial elastic stage. In this paper, CFST columns of various
dimensions cast with normal-strength or high-strength concrete and installed with external steel rings were tested
under uni-axial compression. From the results, it was evident that (a) the external steel rings could restrict the
lateral dilation of CFST columns and improve the interface bonding condition and (b) externally confined
CFST columns had uni-axial strength and stiffness larger than those of unconfined CFST columns. With the
experimental results, an analytical model taking into account the confining effects of steel tube and rings has
been developed to predict the uni-axial strength of ring-confined CFST columns. Copyright © 2012 John Wiley
& Sons, Ltd.

Received 20 April 2012; Revised 29 August 2012; Accepted 3 September 2012

KEY WORDS: analytical model; columns; concrete-filled-steel-tube; external confinement; high-strength concrete;
normal-strength concrete

1. INTRODUCTION

There are different methods for the design of tall buildings subjected to extreme loadings (e.g. earthquake
attack), which includes base-isolation methods (Khoshnoudian and Azad, 2011; Mehrparvar and
Khoshnoudian, 2011; Yamamoto et al., 2011); installation of semi-active dampers (Ribakov, 2011a,
2011b), active dampers (Ribakov and Agranovich, 2011a), passive dampers (Chung et al., 2009; Lee
et al., 2009; Ribakov and Agranovich, 2011b) and tuned mass dampers (Heo et al., 2009; Marano and
Greco, 2009; Mohebbi and Joghataie, 2012) and by reinforcement details such that a plastic hinge region
can be formed to dissipate energy through inelastic deformation without structural collapse (Pam and
Ho, 2009; Yan and Au, 2010). Amongst these measures, the last method is generally applicable to
medium-rise and tall buildings for performance-based design approach (Sabol and Nishi, 2011), while
the former methods are only cost-effectively for very tall buildings. In the last method, the locations of
potential plastic hinge regions are firstly designed on the basis of a collapse mechanism, then careful design
and detailing of reinforcement should be carried out within this regions to produce adequate strength and
ductility (Ho et al., 2012). This can usually be achieved by installing transverse steel (or stirrups) to
confine the concrete core (Ho and Pam, 2003; Havaei and Keramati, 2011; Ho, 2011; Zhou et al.,
2011). Conventionally, the confinement was provided by installing transverse steel at close spacing to
confine the core concrete. However, one of the major shortcomings of this form of confinement is that
the core concrete in both the horizontal and vertical planes between the laterally restrained transverse steel

*Correspondence to: J. C. M. Ho, Department of Civil Engineering, The University of Hong Kong, Hong Kong
E-mail: johnny.ho@hku.hk

Copyright © 2012 John Wiley & Sons, Ltd.

404 M. H. LAI AND J. C. M. HO

is not effectively confined owing to arching action (Mander et al., 1988). More importantly, the effective
confined concrete area reduces significantly as the spacing of transverse steel increases. To resolve the
problem and maximise the area of confined concrete by the same volume of steel, concrete-filled-steel-tube
(CFST) column that consists of a hollow steel tube with concrete filled inside was advocated (Uy, 1998;
Ellobody and Young, 2006; Bambach et al., 2008); the advantages of which are that it provides
a more uniform and continuous confining pressure to the concrete inside such that the confinement
effectiveness increases.
From a structural strength point of view, the steel tube acts as both the longitudinal and confining
reinforcement that enhances the strength-to-weight ratio of the columns, e.g. under uni-axial compression
(Elremaily and Azizinamini, 2002; Huang et al., 2002; Giakoumelis and Lam, 2004; Dabaon et al., 2009;
Chitawadagi et al., 2010; de Oliveira et al., 2010), uni-axial tension (Han et al., 2011), torsion (Han et al.,
2007), pure bending (Elchalakani et al., 2001; Chitawadagi and Narasimhan, 2009; Lu et al., 2009) and
combined axial compression and bending (Fam et al., 2004). The floor area saved due to the enhanced
strength is always beneficial to the developers, architects and engineers. From materials’ point of view,
CFST column is particularly suitable for high-strength concrete (HSC) columns. As HSC is more
brittle than normal-strength concrete (NSC) (Kwan, 2000), HSC columns need to be provided with more
confinement in order to avoid brittle failure under combined axial load and bending (Park, 2001; Ho et al.,
2010) during earthquake attack. In the traditional design using transverse steel as confinement, the
content of transverse steel will become too large that adversely affects the concrete placing quality
(Lam et al., 2009). This will inevitably limit the maximum concrete strength that can be adopted in column
construction while maintaining a certain minimum level of ductility. To further push up the limit of the
concrete strength that can be adopted without jeopardising the ductility, CFST column, which has a more
efficient form of concrete confinement to enhance ductility, can be used. From a practical construction
point of view, the steel tube acts as formwork such that no external formwork for concreting is required.
It saves the construction materials and shortens the construction cycle time. Considering the above
advantages, CFST columns are becoming increasingly popular.
Despite the above advantages, there is a major shortcoming of adopting CFST columns, which is the
imperfect interface bonding that occurs between concrete and steel tube during initial elastic stage. It is
because steel dilates more than concrete at the initial stage under compression owing to the difference
in materials’ Poisson’s ratios (0.2 for concrete and 0.3 for steel) (Persson, 1999; Ferretti, 2004; Lu and
Hsu, 2007; Liao et al., 2011). The imperfect interface bonding will reduce the confining pressure that
can be exerted on the core concrete by the hollow steel tube, and hence the elastic strength and stiffness
of the column. For example, the difference in uni-axial strength between bonded and unbounded CFST
columns could be up to 17% (Giakoumelis and Lam, 2004). The situation will only be improved
after some micro-cracks have been developed in the core concrete under larger axial strain. One of
the consequences of the reduced elastic strength and stiffness is that it increases the inter-storey or
overall drift ratio of tall buildings, which may not fulfil the serviceability limit state requirement. Also,
the reduced interface bonding would increase the chance of having tube buckling at elastic stage if the
wall is thin compared with its diameter (O’Shea and Bridge, 2000).
To fully utilise the hollow steel tube for concrete confinement at early elastic stage, the imperfect
interface bonding needs to be improved. Different methods have been proposed in the past that included
the addition of tab stiffeners (Petrus et al., 2010) and mechanical shear connectors (Nie et al., 2012).
Nonetheless, the installation and fabrication of these internal tab stiffeners and shear connectors was quite
difficult, especially when the column size is relatively small. Instead, external confinement provided in the
form of steel rings, which facilitates the installation onto the external face of CFST columns, is proposed in
this study. The improvement in the load-carrying capacity by external confinement can be divided into two
parts: (a) In the initial elastic stage, the external confinement restricts the lateral dilation of the concrete and
steel tube, which enhances the confining stress provided to the in-filled concrete and hence the elastic
strength and stiffness. (b) In the post-elastic stage after the concrete has started to crush, the confinement
is effective in delaying the buckling of the steel by reducing the effective depth to within the spacing of
the external confinement (since the rings provide the points of effective lateral restraints). The validity of
using the proposed external ring confinement in the strength and stiffness enhancement is investigated
by a series of uni-axial compression test on 35 CFST columns with and without external steel rings.
The effectiveness of these confined CFST columns was studied by the obtained Poisson’s ratio, axial

Copyright © 2012 John Wiley & Sons, Ltd. Struct. Design Tall Spec. Build. 23, 403–426 (2014)
DOI: 10.1002/tal
 UNI-AXIALLY LOADED CFST COLUMNS WITH EXTERNAL RINGS 405

load-carrying capacity and elastic stiffness. From the test results, it is clear that (a) the external steel rings
could improve the axial load-carrying capacity and elastic stiffness of CFST columns; (b) the axial
load-carrying capacity increases as the ring spacing decreases and (c) the shape of the ineffectively
confined concrete region can be assumed to follow the arching action. With the experimental results, an
analytical model taking into account the confining effects provided by the steel tube and rings is developed
to estimate the axial load-carrying capacity of unconfined and ring-confined CFST columns. The validity
of the analytical model is justified by comparing the predicted theoretical strength with the measured
strength in the tests, as well as those obtained by other researchers.

2. EXPERIMENTAL TEST PROGRAMME

2.1. Specimens
A total of 35 CFST columns with different sectional and material properties have been fabricated and tested
under uni-axial compression. The CFST columns specimens are divided into four groups depending on the
concrete grades: (a) 12 CFST with concrete strength of 30 MPa and steel tube thickness of 5 and 8 mm;
(b) 5 CFST columns with concrete strength of 50 MPa and steel tube thickness of 10 mm; (c) 12 CFST
columns with concrete strength of 80 MPa and steel tube thickness of 5 mm and 8 mm; (d) 6 CFST
columns with concrete strength of 120 MPa and steel tube thickness of 5 mm. In each group, at least one
of the specimens were unconfined (i.e. without external rings), which was to serve as control specimen
for comparison purpose. The rest of the specimens were installed with external rings provided at different
spacing (i.e. 5t, 10t, 12.5t, 15t and 20t, where t is the thickness of the steel tube and is 5 mm, 8 mm or 10 mm
in this study). The grade of steel tube is S355 produced as per BS EN 10210-2:2006. The diameter of the
external steel tube and height of the CFST columns are listed respectively as follows: (Groups 1 and 3)
168.3 mm and 330 mm, (Group 2) 139.7 and 330 mm, and (Group 4) 114.3 and 248 mm. Figure 1 shows
the photos of typical CFST columns (Group 3 with thickness 8 mm). The external steel rings (diameter
8 mm and nominal yield strength 250 MPa) were intermittent welded to the external surface of the steel
tube at eight welding spots at each level, in which four of them were above the ring and four of them below
the ring in alternative manner. The welding spots were evenly distributed around the circumference of
the steel tube, which were separated from each other by 45
at the centre of the section. There was an
overlapping length for the steel rings of 10 times the diameter, which was to ensure that the yield strength
of the ring could be fully developed. Details of the CFST specimens with rings are shown in Figure 2.
A naming system was established to identify each of the specimens. The naming system consists of two
alphabets and four numbers. For example, CR5-5-168-30 represents a CFST column specimen (indicated
by the first letter ‘C’) confined by external steel ring (indicated by the second letter ‘R’). The spacing of the
external ring is five times the thickness of the steel tube, i.e. 5t, (indicated by the first number ‘5’). The
thickness of the steel tube is 5 mm (indicated by the second number ‘5’). The outer diameter of the steel

Figure 1. Typical tested specimens.

Copyright © 2012 John Wiley & Sons, Ltd. Struct. Design Tall Spec. Build. 23, 403–426 (2014)
DOI: 10.1002/tal
406 M. H. LAI AND J. C. M. HO

Figure 2. CFST column—confined by the external rings.

tube is about 168 mm (indicated by the third number ‘168’), and lastly, the concrete cylinder strength is
30 MPa on the testing day (indicated by the last number ‘30’). The CFST column without the provision
of steel rings is represented by CN0-5-168-30, where N0 stands for ‘no confinement’. Table 1 summarises
the material properties of the specimens.

2.2. Instrumentations
In this experiment, the SATEC Series RD Model with maximum load of 5000 kN and maximum travelling
displacement of 100 mm was adopted for the uni-axial compression test. The details of the test set-up and
instrumentation are shown in Figure 3(a–c). Three linear variable differential transducers (LVDTs)
with 100-mm stroke were installed to record the full length axial displacement of the CFST columns
by measuring the relative displacement between the top and bottom loading platens. Three numbers of
two-directional strain gauges (Tokyo Sokki Kenkyujo Co., Ltd, Tokyo, Japan) were installed at around
the mid-height of the external face of the tested specimens, which were 120
separated from each other
at the centre of the section to measure the longitudinal and transverse strains. A circumferential extensometer
with maximum measuring range of 6 mm was installed between two adjacent rows of the external rings to
measure the tube dilation near the steel rings during the initial elastic stage. The extensometer would be
removed when the lateral dilation was about to reach 6 mm. In some specimens, three to five one-directional
strain gauges (depending on the number of rings) were installed on the surface of the external rings to
monitor the strains developed.

2.3. Testing procedure

All the hollow steel tubes and CFST columns were tested under displacement control. For CFST columns,
the top surface was packed with a layer of gypsum before initial compression was applied to make sure the
gypsum is in complete contact with the loading platen. After that, specimens would be unloaded and the
test would start. The initial loading rate for specimens of 330-mm height was 0.3 mm/min and that for
248-mm height was 0.2 mm/min. The loading rate would then increase at a rate of 0.05 mm/min for every
2 mm increase in the axial displacement after the specimens had yielded. The loading application
would stop when the axial strain was larger than 0.2, when the applied load dropped to less than 70% of
the measured maximum load or when the applied load reached 90% of the machine capacity, whichever
was the earliest.

Copyright © 2012 John Wiley & Sons, Ltd. Struct. Design Tall Spec. Build. 23, 403–426 (2014)
DOI: 10.1002/tal
 UNI-AXIALLY LOADED CFST COLUMNS WITH EXTERNAL RINGS 407

Table 1. Lists of specimens.

Thickness of Spacing
Group ssy f c0 steel tube, of rings, Nexp Strength Stiffness Stiffness
no. Specimens (MPa) (MPa) t (mm) S (mm) (kN)* ratio (Gpa) ratio
1 CR5-5-168-30 365 29.1 5 25 3076 1.52 51.7 1.26
CR10-5-168-30 365 29.1 5 50 2598 1.29 50.6 1.23
CR12.5-5-168-30 365 29.1 5 62.5 2444 1.21 49.4 1.21
CR15-5-168-30 365 29.1 5 75 2372 1.17 47.2 1.15
CR20-5-168-30 365 29.1 5 100 2221 1.10 49.3 1.20
CN0-5-168-30 365 29.1 5 — 2019 1.00 40.9 1.00
CR5-8-168-30 365 38.3 8 40 3777 1.22 64.1 1.15
CR10-8-168-30 365 37.8 8 80 3483 1.13 69.7 1.25
CR12.5-8-168-30 365 38.3 8 100 3470 1.12 68.8 1.23
CR15-8-168-30 365 38.3 8 120 3434 1.11 62.4 1.12
CR20-8-168-30 365 38.3 8 160 3231 1.05 64.9 1.16
CN0-8-168-30 365 34.1 8 — 3089 1.00 55.9 1.00
2 CR5-10-139-50 365 47 10 50 3412 1.13 63.8 1.03
CR10-10-139-50 365 47 10 100 3207 1.06 — —
CR15-10-139-50 365 47 10 150 3093 1.03 70.1 1.13
CR20-10-139-50 365 47 10 200 3114 1.03 62.8 1.01
CN0-10-139-50 365 47 10 — 3014 1.00 62.1 1.00
3 CR5-5-168-80 365 85.4 5 25 3643 1.25 57.0 1.07
CR10-5-168-80 365 85.4 5 50 3205 1.10 53.6 1.01
CR12.5-5-168-80 365 85.4 5 62.5 3178 1.09 58.5 1.10
CR15-5-168-80 365 85.4 5 75 3079 1.05 51.0 0.96
CR20-5-168-80 365 85.4 5 100 3149 1.08 57.7 1.09
CN0-5-168-80 365 85.4 5 — 2926 1.00 53.1 1.00
CR5-8-168-80 365 75.2 8 40 3985 1.17 66.5 1.01
CR10-8-168-80 365 75.2 8 80 3793 1.11 81.6 1.24
CR12.5-8-168-80 365 85.4 8 100 3771 1.11 71.7 1.09
CR15-8-168-80 365 75.2 8 120 3580 1.05 77.2 1.17
CR20-8-168-80 365 75.2 8 160 3392 1.00 71.1 1.08
CN0-8-168-80 365 75.2 8 — 3408 1.00 65.8 1.00
4 CR5-5-114-120 365 114.3 5 25 2400 1.28 73.4 1.17
CR10-5-114-120 365 114.3 5 50 2167 1.16 66.4 1.06
CR12.5-5-114-120 365 114.3 5 62.5 2065 1.10 61.9 0.98
CR15-5-114-120 365 116.7 5 75 2109 1.12 69.1 1.10
CR20-5-114-120 365 114.3 5 100 1977 1.05 69.0 1.10
CN0-5-114-120 365 114.3 5 — 1876 1.00 62.9 1.00
*: Nexp is defined as either the strength at peak or the strength at 0.1 axial strain, whichever is larger.

3. EXPERIMENTAL RESULTS AND DISCUSSIONS

3.1. Discussions about different measurement methods

There were three methods to measure the displacement/strain in this study: ‘Method 1’ represents
the measurement by the machine; ‘Method 2’ represents the measurement by LVDTs; ‘Method 3’
represents the measurement by strain gauges. It should be noted that since rapid hardening gypsum
was adopted as capping material to ensure a smooth surface between the top loading platen and the
test specimens, the reading measured by the machine and LVDTs will include the shortening of the
gypsum, which if considered in the calculation of axial strain, will overestimate the actual
value. Therefore, the readings obtained from ‘Method 1’ and ‘Method 2’ shall be adjusted in order
to eliminate the effects of the shortening of the capping material. Assuming that the thickness of
the gypsum is tg, the total height of the CFST specimen is H, the uni-axial compression applied
is Nc, and the modulus of elasticity of the specimens measured is Ei, where i is the method of
measurement, it can be deduced the following compatibility equations from Figure 4:

Eg A E3 A
Nc ¼ Δtg ¼ ΔH (1)
tg H

Copyright © 2012 John Wiley & Sons, Ltd. Struct. Design Tall Spec. Build. 23, 403–426 (2014)
DOI: 10.1002/tal
408 M. H. LAI AND J. C. M. HO

120
degree

Location of strain gauge

Location of LVDTs

LVDT1 LVDT2

Extensometer

Strain gauge

Figure 3. (a) Test set-up. (b) Plan view of test set-up. (c) Side view of test set-up.

Similarly, the following equation can be derived for E2 and E3:

     
N c tg þ H ΔH tg þ H ΔH
E2 ¼
 ¼ E3 ¼ E3 (2)
A Δtg þ ΔH Δtg þ ΔH H Δtg þ ΔH

As we can see from Equations (1) and (2), the effect of the capping material would be significant in
the early elastic stage but become insignificant at the later inelastic stage because the axial shortening
of the gypsum is very small when compared with that of the CFST specimens. In this study, the
axial strain measured by ‘Method 3’ was taken as the axial strain in early elastic stage. Therefore,

Copyright © 2012 John Wiley & Sons, Ltd. Struct. Design Tall Spec. Build. 23, 403–426 (2014)
DOI: 10.1002/tal
 UNI-AXIALLY LOADED CFST COLUMNS WITH EXTERNAL RINGS 409

Nc
Gypsum

tg
Steel
tube
Unlevelled
surface
Core
Concrete

Nc

Figure 4. Modification of capping effects.

the readings obtained from other methods were adjusted to match with that obtained from ‘Method 3’.
At the end of the elastic stage, the difference in the strain readings obtained from ‘Method 2’ and
‘Method 3’ will be subtracted from the reading obtained by Method 2 for strain adjustment in the in-
elastic stage.

3.2. Axial load against axial strain (displacement) curves

The measured axial load is plotted against the axial strain (displacement) for the tested CFST columns
in Figure 5(a–f). The axial load is obtained by the machine while the axial strain is obtained by the
adjusted LVDTs reading. From the figures, it was observed that the axial load-carrying capacity
of the CFST columns increases as the spacing of the ring decreases. This is because as the spacing
of confinement reduces, the confining pressure provided to the CFST columns increases and becomes
more uniform. It enhances the strength of the core concrete as well as the steel tube and, in turn,
improves the axial load-carrying capacity.
Figure 5(a–c) shows the axial load–axial strain (displacement) curves for the normal-strength CFST
(NSCFST) specimens. It can be seen from these graphs that initially, the axial load increases as
the axial strain increases linearly. As the axial strain increases, the Poisson’s ratio of core concrete
will increase and be larger than that of steel tube owing to the gradual development of internal
micro-cracks. At this stage, several oblique lines were observed (Figure 6) on the external face of
the steel tube, especially near the rings or the top/bottom surfaces of the specimens, which indicated
yielding of the steel tube under large deformation. At this stage, the load sharing pattern will change
(Ding et al., 2011), in which the steel tube will take less loading while the concrete will increase the
sharing of applied load. As the axial strain kept increasing to about 0.02–0.03, the steel tube
would eventually enter strain hardening stage. It is noted from previous research (Ho et al., 2005;
Bai and Au, 2009) that the ultimate strength could be 20% higher than the yield strength due to strain
hardening effect. This is why larger hoop tensile stress can be developed in the strain hardening
process and thus pushes outward the ‘Von Mises’ yield surface (Figure 7). Thus, the confining stress
provided to the in-filled concrete is improved, which makes the in-filled concrete stronger.
Figure 5(d–f) shows the axial load–axial strain (displacement) curves for the high-strength CFST
(HSCFST) specimens. After reaching the first maximum load, the axial load in the specimens dropped
because micro-cracks started to develop in the concrete core. Under this circumstance, the axial load
capacity of the concrete core decreased and thus decreased the axial load-carrying capacity of

Copyright © 2012 John Wiley & Sons, Ltd. Struct. Design Tall Spec. Build. 23, 403–426 (2014)
DOI: 10.1002/tal
410 M. H. LAI AND J. C. M. HO

Figure 5. (a) Load–strain (displacement) curves for CFST columns (group no. 1 with thickness 5 mm).
(b) Load–strain (displacement) curves for CFST columns (group no. 1 with thickness 8 mm). (c) Load–
strain (displacement) curves for CFST columns (group no. 2). (d) Load–strain (displacement) curves
for CFST columns (group no. 3 with thickness 5 mm). (e) Load–strain (displacement) curves for CFST
columns (group no. 3 with thickness 8 mm). (f) Load–strain (displacement) curves for CFST columns
(group no. 4).

Copyright © 2012 John Wiley & Sons, Ltd. Struct. Design Tall Spec. Build. 23, 403–426 (2014)
DOI: 10.1002/tal
 UNI-AXIALLY LOADED CFST COLUMNS WITH EXTERNAL RINGS 411

Figure 5. Continued.

the HSCFST columns. However, when the axial strain increased further, the lateral dilation of core
concrete increased abruptly such that larger confining pressure was provided by the external
rings. This offsets the strength degradation due to micro-cracks formation, and hence, the axial
load-carrying capacity of the HSCFST columns increased again. This phenomenon was not observed
in NSCFST columns because NSC is more ductile and hence the concrete strength degradation
occurred at a more gradual rate than the HSC.

Copyright © 2012 John Wiley & Sons, Ltd. Struct. Design Tall Spec. Build. 23, 403–426 (2014)
DOI: 10.1002/tal
412 M. H. LAI AND J. C. M. HO

Figure 6. Illustration of oblique lines.

sz

Strain hardening

Figure 7. ‘Von Mises’ yield criteria considering strain hardening effect.

It could be seen from Figure 5 that for all CFST columns tested in this study, the axial strength increases as
the spacing of rings decreases. More importantly, the axial strength and stiffness of confined CFST columns
were found larger than those of the unconfined CFST columns. This is because when the ring spacing
decreases, a more uniform confining pressure will be provided to the CFST columns in the axial direction,
and hence improved the strength and stiffness of the columns. To study the effectiveness of confinement due
to the provision of external rings, the maximum axial load-carrying capacity and the initial stiffness for
all CFST specimens were recorded in order to study their enhancement. It is worth noting that since the
measured axial load increased as the axial strain increased for NSCFST columns, it is somehow difficult
to determine the maximum axial load-carrying capacity. From a practical design point of view, the axial load
that occurred at very large strain may not be developed in flexural members because such a large strain
could not be reached in the tension region before failure occurs. Therefore in this study, the maximum axial
load-carrying capacity is taken as the measured axial load at the axial strain of 0.1, which is about the fracture
strain of steel. Thus, the maximum load is defined as either the strength at peak or the strength at 0.1 axial

Copyright © 2012 John Wiley & Sons, Ltd. Struct. Design Tall Spec. Build. 23, 403–426 (2014)
DOI: 10.1002/tal
 UNI-AXIALLY LOADED CFST COLUMNS WITH EXTERNAL RINGS 413

strains, whichever is larger. The obtained values of the load-carrying capacities for all specimens are listed in
Table 1. However, it should be still noted that the load-carrying capacity may not be reached in some CFST
columns which is subjected to simultaneous low axial load level and bending. On the other hand, the initial
stiffness is calculated from the initial slope of the graph and divided by the total area, Acs (Equation (4)) of the
specimens. The values of the initial stiffness for all specimens are also listed in Table 1.
The condition of the interface bonding between concrete and steel can be investigated indirectly
through the comparison of the measured composite stiffness of the unconfined CFST columns
with the theoretical value (assuming full composite action between concrete and steel). The theoretical
composite stiffness for unconfined CFST columns can be calculated by the following equation:

Ecs Acs ¼ Ec Ac þ Es As (3)

Acs ¼ Ac þ As (4)

where E and A refer to the modulus and area, and subscripts CS, C and S refer to the composite section,
concrete and steel, respectively. The theoretical values of Ecs calculated as per Equation (3) are summarised
in Table 2. Compared with the theoretical values shown in Table 1, the theoretical values are on average
11.6% smaller. It is believed that the reduced stiffness of the unconfined CFST columns was caused by
the imperfect interface bonding at initial stage.
The effects of adding steel rings on the enhancement of axial load-carrying capacity and the stiffness
improvement are studied by the strength and stiffness enhancement ratios. The enhancement ratios
were defined as the ratio of the strength (or stiffness) of the confined CFST columns to the respective
value of the unconfined CFST columns. The enhancement is shown by the ratio summarised in Table 1.
It is evident from the table that the addition of external steel rings can improve the axial load-carrying
capacity and elastic stiffness of the CFST columns. It is also observed that the strength and stiffness
enhancement ratios decrease as the concrete strength increases. The reason is because of the smaller
lateral dilation in the column as a result of larger axial stiffness.

3.3. Lateral strain against longitudinal strain curves

Because of the difference in Poisson’s ratios of steel (~0.3) and concrete (~0.2), the core concrete are not
effectively confined at initial elastic stage under compression. The elastic stiffness and axial load-carrying
capacity will be reduced. To resolve the problem, the lateral dilation of steel tube and concrete needs to be
restricted. In this study, external steel rings were installed to restrict the lateral dilation of the steel tube. The
effectiveness of providing external rings to enhance the interface bonding can studied by plotting the lateral
strains measured by strain gauges on rings (LS1) or circumferential extensometers (LS2) or strain gauges
on the exterior surface of steel tube (LS3) against axial strain of the CFST specimens during initial
elastic stage. Three different methods were adopted to determine the lateral strain because the locations
of the lateral strain measurement would affect significantly of the lateral dilation measured because of
the confining effect provided by the steel rings.

Table 2. Composite modulus of elasticity of unconfined specimens.

Composite stiffness (GPa) Percentage
Specimen difference
no. Specimens Calculated Experimental (%)
1 CN0-5-168-30 48.1 40.9 15.0
2 CN0-8-168-30 61.3 55.9 8.8
3 CN0-10-139-50 79.7 62.1 22.1
4 CN0-5-168-80 57.4 53.1 7.5
5 CN0-8-168-80 68.1 65.8 3.4
6 CN0-5-114-120 71.9 62.9 12.5
Average 11.6

Copyright © 2012 John Wiley & Sons, Ltd. Struct. Design Tall Spec. Build. 23, 403–426 (2014)
DOI: 10.1002/tal
414 M. H. LAI AND J. C. M. HO

The Poisson’s ratios of the specimens were calculated by the slope of the initial straight line portion
of the lateral strain against longitudinal strain curves and the results are listed in Table 3. The curves of
lateral strain plotted against axial strain for some of the CFST columns are shown in Figure 8(a–f).
The lateral strains LS1 are obtained by adopting the largest values of the strain gauges installed on
rings (i.e. the lateral strain of steel tube at ring level). The lateral strains LS2 were obtained by dividing
the circumferential expansion by the circumference of the steel tube (i.e. the average lateral strain of
steel tube between rings). The lateral strains LS3 were obtained by averaging the values of three strain
gauges on the steel tubes (i.e. localised strain of steel tube between rings). The axial strains were
obtained by the adjusted LVDTs reading. From the table, it can be seen that the Poisson’s ratios of
the CFST columns obtained from LS1 are smaller than the initial Poisson’s ratio of concrete (~0.2).
The Poisson ratios of the columns obtained from LS2 are similar to that of concrete for those
extensometers near the ring, whereas the Poisson ratios of the columns obtained from LS3 are much
larger than that of concrete. This is because during uni-axial compression, the steel tube would bend
slightly outward between the rings, where effective lateral restraints were provided. Hence the concrete
between the rings in the vertical plane were not completely confined during compression. The shape of
the ineffectively confined concrete region can be assumed to follow the arching action (Mander et al.,
1988) as shown in Figure 9.

Table 3. Ratios of lateral strain to longitudinal strain.

Group No. Specimens LS1 LS2 LS3
1 CR5-5-168-30 — — 0.331
CR10-5-168-30 — 0.175 0.293
CR12.5-5-168-30 — 0.198 0.325
CR15-5-168-30 — 0.196 0.280
CR20-5-168-30 — 0.197 0.390
CN0-5-168-30 — 0.294 0.320
CR5-8-168-30 — — 0.317
CR10-8-168-30 0.216 0.227 0.343
CR12.5-8-168-30 — 0.225 0.387
CR15-8-168-30 — 0.222 0.316
CR20-8-168-30 — 0.203 0.329
CN0-8-168-30 — 0.215 0.295
2 CR5-10-139-50 0.079 0.203 0.305
CR10-10-139-50 * * *
CR15-10-139-50 0.132 0.196 0.308
CR20-10-139-50 0.100 0.159 0.292
CN0-10-139-50 — 0.178 0.313
3 CR5-5-168-80 — — 0.245
CR10-5-168-80 0.221 — 0.331
CR12.5-5-168-80 — 0.261 0.288
CR15-5-168-80 0.124 0.183 0.274
CR20-5-168-80 — 0.199 0.281
CN0-5-168-80 — 0.318 0.297
CR5-8-168-80 — 0.285 0.277
CR10-8-168-80 — 0.159 0.311
CR12.5-8-168-80 — 0.189 0.269
CR15-8-168-80 — 0.293 0.325
CR20-8-168-80 — 0.277 0.291
CN0-8-168-80 — 0.328 0.308
4 CR5-5-114-120 0.107 — 0.293
CR10-5-114-120 0.221 0.224 0.321
CR12.5-5-114-120 0.120 0.193 0.304
CR15-5-114-120 0.080 0.184 0.375
CR20-5-114-120 0.103 0.195 0.326
CN0-5-114-120 — 0.255 0.270
Remarks:
—: Strain gauges/circumference extensometers were not installed.
*: Strain gauges were broken.

Copyright © 2012 John Wiley & Sons, Ltd. Struct. Design Tall Spec. Build. 23, 403–426 (2014)
DOI: 10.1002/tal
 UNI-AXIALLY LOADED CFST COLUMNS WITH EXTERNAL RINGS 415

Figure 8. (a) Lateral strain against longitudinal strain (CR5-10-139-50). (b) Lateral strain against
longitudinal strain (CR15-10-139-50). (c) Lateral strain against longitudinal strain (CR20-10-139-50).
(d) Lateral strain against longitudinal strain (CR5-5-114-120). (e) Lateral strain against longitudinal
strain (CR12.5-5-114-120). (f) Lateral strain against longitudinal strain (CR20-5-114-120).

Copyright © 2012 John Wiley & Sons, Ltd. Struct. Design Tall Spec. Build. 23, 403–426 (2014)
DOI: 10.1002/tal
416 M. H. LAI AND J. C. M. HO

Figure 8. Continued.

3.4. Modes of failure

Figure 10 shows the failure modes of hollow steel tube, CFST columns with or without external steel rings.
For hollow steel tube, it is seen in Figure 10(a) that local buckling occurred at the end of the steel tube,

Copyright © 2012 John Wiley & Sons, Ltd. Struct. Design Tall Spec. Build. 23, 403–426 (2014)
DOI: 10.1002/tal
 UNI-AXIALLY LOADED CFST COLUMNS WITH EXTERNAL RINGS 417

Steel
tube

Core
Ineffective
Concrete
confinement

External
ring

Figure 9. Illustration on imperfect bonding of CFST columns with external ring confinement (side view).

A C

“End effect” Buckling

between steel
rings

Steel rings
provide
effective
lateral
restraints

B D
Buckling
between steel
rings

Fracture of steel
tube and steel
rings

Local
buckling

Figure 10. (a) Failure mode for HST (height 330 mm, thickness 5 mm). (b) Failure mode for CFST
columns (CN0-5-168-30). (c) Failure mode for NSCFST columns (CR12.5-5-168-30). (d) Failure mode
for HSCFST columns (CR10-5-114-120).

which is due to the ‘end effect’ as a result of the large axial load and frictional force at the interface between
loading platen and the column specimen. For CFST columns without steel rings, it can be observed from
Figure 10(b) that buckling of steel tube occurred close to the bottom of the specimens in addition to local
buckling due to the end effects as explained above. This was due to the expansion and crushing of the
in-filled concrete at large axial strain, which was subjected to large lateral strain and subsequently pushed

Copyright © 2012 John Wiley & Sons, Ltd. Struct. Design Tall Spec. Build. 23, 403–426 (2014)
DOI: 10.1002/tal
418 M. H. LAI AND J. C. M. HO

outward against the steel tube. Since the steel tube was also subjected to large axial load, the large lateral
strain in the in-filled concrete would cause local buckling of steel tube (Figure 10(b)). For CFST columns
installed with external steel rings, it was evident from Figure 10(c–d) that the rings provided effective
lateral restraints to the steel tube such that buckling always occurred between two adjacent rows of the
steel rings. This reduced the effective length of the columns for buckling and hence increased the axial
load-carrying capacity. For HSCFST columns, it can be observed from Figure 10(d) that fracture of steel
tube and external steel rings occurred. This phenomenon was only observed in HSCFST columns but
not in NSCFST columns. It is because HSC is more brittle than NSC under uni-axial compression. Thus,
when HSC failed, the concrete stress dropped abruptly and a large amount of axial load was transferred to
the steel tube, which increased the axial and lateral strains. Since the induced strains were too large, it
caused fracture of the steel rings and the steel tube along the internal welding location.
Apart from the above observation, it was also noted that for all CFST columns, the top and bottom
surfaces of the in-filled concrete was slightly higher than the steel surface (normally 0.5 mm to 1 mm).
This can be explained by the sudden reduction of the uni-axial loading during failure such that
it caused a reduction in the hoop stress and consequently pushed inward against the core concrete
(Cai and Jiao, 1984). Figure 11 shows the phenomenon observed in one of the test specimens.

4. PROPOSED ANALYTICAL MODEL FOR LOAD-CARRYING CAPACITY PREDICTION

4.1. Proposed model

An analytical model was presented herein to predict the axial load-carrying capacity of unconfined and
ring-confined CFST columns. The model takes into account the enhanced confined stresses of concrete
and steel tube due to the confining pressure provided by the external steel ring. The model evaluates
the total axial load-carrying capacity contributed by the steel and concrete and subsequently sums
up the individual components to give the load-carrying capacity of the CFST columns.
For the steel tube, the famous Von Mises yield criterion is adopted in this model:

s2sθ þ ssθ ssz þ s2sz ¼ s2sy (5)

where ssz is the axial stress in the steel tube, ssθ is the hoop stress in the steel tube under bi-axial state
and ssy is the uni-axial yield strength of the steel tube.
For the in-filled concrete, the following equation is adopted to evaluate the confined concrete
strength with a given confining pressure fr (Cusson and Paultre, 1994).

fcc ¼ fc ’ þ kfr (6)

where fcc is the confined concrete stress, fc0 is the unconfined concrete cylinder strength, fr is the
confining pressure and k is taken as 4.1. The confining pressure fr can be expressed in terms of the

0.5 to 1mm

Figure 11. Slippage between the core concrete and the steel tube.

Copyright © 2012 John Wiley & Sons, Ltd. Struct. Design Tall Spec. Build. 23, 403–426 (2014)
DOI: 10.1002/tal
 UNI-AXIALLY LOADED CFST COLUMNS WITH EXTERNAL RINGS 419

geometry and hoop stress of the steel tube by considering the free body diagram as shown in Figure 12.
p
fr ðD  2t ÞH ¼ 2ssθ tH þ 2fsr d2 n (7)
4

fsr ¼ Esr esr ≤fyr (8)

where D is the outer diameter of the steel tube; H is the height of the CFST columns; t is the thickness
of the steel tube; d is the diameter of rings; fsr, esr and Esr are respectively the stress, strain and elastic
modulus of steel rings; n stands for the number of rings; and fyr is the yield strength of steel rings.
Combining Equations (7) and (8), we have

2ssθ tH þ 2fsr p4 d 2 n 2ssθ tH þ 2Esr esr p4 d2 n

fr ¼ ¼ (9)
ðD  2t ÞH ðD  2t ÞH

It is reasonable to assume that the steel rings yield when the CFST columns achieve the maximum
load-carrying capacity. Therefore, fsr can be taken as fyr, and Equation (8) can be rewritten as follows:

2ssθ tH þ 2fyr p4 d2 n Ast pd2

fr ¼ ¼ ssθ þ fyr (10)
ðD  2t ÞH 2Ac 2ðD  2t ÞS

where Ast is the total area of the steel, including the area of rings, i.e.

p d 2 pD fyr
Ast ¼ As þ (11)
4 S ssy

The load-carrying capacity of the CFST columns can be obtained by summing up the individual
components contributed by the in-filled concrete and steel tube by

N ¼ ssz Ast þ fcc Ac (12)

In order to determine the load-carrying capacity, the hoop stress in the steel tube should be determined
first. Substituting the experimentally obtained load-carrying capacity of the CFST column specimens into
Equation (12), ssθ can be calculated by solving Equations (5), (6), (10) and (12). Subsequently, a regression
analysis was performed to correlate ssθ/ssy to x, which is the confinement index, taking into account the
steel rings contribution (Equation (13)). The variation of ssθ/ssy against x (x ≧ 1) is shown in Figure 13.
From the figure, it is seen that the relationship between ssθ/ssy and x is almost linear. Two boundary
conditions have to be fulfilled, which are

Figure 12. Free body diagram for confining stress (with ring confinement).

Copyright © 2012 John Wiley & Sons, Ltd. Struct. Design Tall Spec. Build. 23, 403–426 (2014)
DOI: 10.1002/tal
420 M. H. LAI AND J. C. M. HO

Figure 13. Regression analysis and proposed equations.

(1) Lower bound: ssθ = 0 and ssz = ssy. In this condition, the confinement index x is very small (taken
as x ≤ 1/75), the induced confinement effect is very weak and can be ignored. Thus, the steel tube
will serve as longitudinal reinforcement without contributing any confinement effect.
(2) Upper bound: ssθ = ssy and ssz = 0. In this case, the confinement provided is very strong, and the
specimen will become very ductile (taken as x ≥ 6.68). Thus, the steel tube will serve as confining
reinforcement without contributing any axial resistance.
From the above observations, a tri-linear curve has been proposed to relate ssθ/ssy and x for different
values of x as shown in Equation (14).

Ast ssy
x¼ 0 (13)
Ac fc

8
< 0 0≤x < 1=75
ssθ
¼ 0:15x  0:002 1=75≤x < 6:68 (14)
ssy :
1 x≥6:68

In practical design of CFST column under uni-axial compression, the axial load-carrying capacity
can be determined by the following procedure:
(1) Determine the confinement index x by Equation (13) and hence the hoop stress ssθ in the steel tube
by Equation (14).
(2) Determine the axial stress ssz in steel tube under bi-axial state as per Equation (5).
(3) Determine the confining pressure fr by Equation (10) with the evaluated ssθ in (1).
(4) Determine the enhanced confined concrete stress fcc using Equation (6).
(5) Lastly, the axial load-carrying capacity of the CFST column is given by Equation (12).
For HSCFST columns, it has been found that the uni-axial load-carrying capacities were always
smaller than those of the measured values. The reason may be that the adopted confinement index x
underestimates the effect of the confining pressure provided by the steel tube and rings for a particular
concrete strength. This is because in confined HSCFST columns, the external confinement enhances
the confining stress and, hence, it improves significantly the strength and ductility of the in-filled
concrete for a given strength (Zhong, 2006). Since the derivation of x was based on unconfined CFST
columns, the expression of x should be modified to cater for the provision of external confinement.
In this study, it is proposed to convert the HSCFST columns into an equivalent section of NSCFST
columns by evaluating the equivalent steel and concrete areas using the following formulas.

Copyright © 2012 John Wiley & Sons, Ltd. Struct. Design Tall Spec. Build. 23, 403–426 (2014)
DOI: 10.1002/tal
 UNI-AXIALLY LOADED CFST COLUMNS WITH EXTERNAL RINGS 421

0
s 
sy
Ast ¼ Ast (15a)
355

0
s 
sy
As ¼ As (15b)
355

0
!
0 fc
Ac ¼ Ac (15c)
30

0
0 SD
S ¼ (15d)
D

rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
0 4
0 0
D ¼ As þ Ac (15e)
p

rffiffiffiffiffiffiffiffiffiffi
4 0
0
Dc ¼ Ac (15f)
p

where Ast0 is the equivalent area of the steel tube and steel rings (of the converted NSCFST columns),
As0 is the equivalent area of the steel tube, Ac0 is the equivalent area of concrete, S0 is the equivalent
spacing of the steel rings, D0 is the diameter of the equivalent cross-section area including in-filled
concrete and steel tube, and ssy and fy are in MPa. The corresponding axial load-carrying capacity
of the HSCFST columns can be evaluated by replacing Ast by Ast0 , As by As0 , Ac by Ac0 , S by S0 , D
by D0 and (D2t) by Dc0 in Equations (5) to (14). Subsequently, the same procedure (1)–(5) of
evaluating the axial load-carrying capacity of NSCFST columns can be adopted for the axial
load-carrying capacity of HSCFST columns.

4.2. Comparison with other researchers’ results

The validity of the proposed equations (i.e. Equations (11) and (12)) are verified by comparing with
the experimental results obtained by the authors in this study and by other previous researchers on CFST
columns (O’Shea and Bridge, 2000; Giakoumelis and Lam, 2004; Han et al., 2005; de Oliveira et al., 2010;
Petrus et al., 2010). The comparisons are shown in Tables 4 and 5, respectively. It can be seen from
the tables that the proposed equations predict very well the axial load-carrying capacity of confined CFST
columns obtained by the authors and other researchers. The differences are all lying within 10%. Therefore,
the validity of the proposed equations to predict the axial load-carrying capacity of confined CFST columns
is verified.

4.3. Limitations of the proposed model

Although the proposed equations can work well in predicting the ultimate axial strength of NSCFST
and HSCFST columns, there are some limitations:
(1) Loading condition should be quasi-static.
(2) Concrete strength should be within 20 MPa to 110 MPa (cylinder strength), and the yield strength
of steel should be from 255 MPa to 355 MPa.
(3) The range of D/t ratios should be from 14 to 100.
(4) Confinement index should be smaller than 3.5.
(5) The strength is only referring to the sectional strength without taking into account the member
behaviour such as stability, initial imperfection and residual stresses. In other words, the proposed
equations only predict well in short CFST columns.

Copyright © 2012 John Wiley & Sons, Ltd. Struct. Design Tall Spec. Build. 23, 403–426 (2014)
DOI: 10.1002/tal
422 M. H. LAI AND J. C. M. HO

Table 4. Test results and predicted results of CFST columns (authors’ results).
N exp
Specimens Nexp Confinement index x Fcal Fcal

CR5-5-168-30 3076 2.10 3073 1.00

CR10-5-168-30 2598 1.87 2547 1.02
CR12.5-5-168-30 2444 1.82 2443 1.00
CR15-5-168-30 2372 1.79 2374 1.00
CR20-5-168-30 2221 1.75 2288 0.97
CN0-5-168-30 2019 1.63 2032 0.99
CR5-8-168-80 3985 1.19 4075 0.98
CR10-8-168-80 3793 1.13 3779 1.00
CR12.5-8-168-80 3771 0.99 3861 0.98
CR15-8-168-80 3580 1.11 3681 0.97
CR20-8-168-80 3392 1.10 3632 0.93
CN0-8-168-80 3408 1.07 3485 0.98
CR5-10-139-50 3412 3.02 3652 0.93
CR10-10-139-50 3207 2.92 3424 0.94
CR15-10-139-50 3093 2.88 3348 0.92
CR20-10-139-50 3114 2.86 3310 0.94
CN0-10-139-50 3014 2.81 3197 0.94
CR5-8-168-30 3777 2.35 3794 1.00
CR10-8-168-30 3483 2.26 3461 1.01
CR12.5-8-168-30 3470 2.20 3395 1.02
CR15-8-168-30 3434 2.19 3351 1.02
CR20-8-168-30 3231 2.17 3296 0.98
CN0-8-168-30 3089 2.36 3132 0.99
CR5-5-168-80 3643 0.72 3853 0.95
CR10-5-168-80 3205 0.64 3386 0.95
CR12.5-5-168-80 3178 0.62 3293 0.97
CR15-5-168-80 3079 0.61 3231 0.95
CR20-5-168-80 3149 0.60 3154 1.00
CN0-5-168-80 2926 0.56 2924 1.00
CR5-5-114-120 2340 0.83 2458 0.95
CR10-5-114-120 2167 0.73 2138 1.01
CR12.5-5-114-120 2065 0.72 2074 1.00
CR15-5-114-120 2109 0.69 2050 1.03
CR20-5-114-120 1977 0.69 1979 1.00
CN0-5-114-120 1876 0.64 1821 1.03
Standard deviation 0.0295

5. CONCLUSIONS

External steel ring confinement was proposed in this study to improve the interface bonding between
concrete and steel in CFST columns. An experimental programme was set up, which consisted of 35 CFST
columns with or without the provision of external steel rings at various spacing (spacing of 5t, 10t, 12.5t,
15t and 20t, where t is the thickness of the steel tube). All the specimens were tested under uni-axial
compression. The effectiveness of providing external rings on the improvement of steel–concrete interface
bonding were investigated by the measured Poisson’s ratio, axial load-carrying capacity (measured at peak
or 0.1 axial strain, whichever is larger) and elastic stiffness.
From the test results, we conclude the following:
(1) The external steel rings effectively improved the axial load-carrying capacity of CFST columns.
The axial strength increased as the spacing of rings decreased.
(2) The external steel rings effectively improved the elastic stiffness of CFST columns.
(3) The external rings were effective since they would limit the lateral deformation of the core concrete
and the steel tube, especially at the locations of rings. However, at the locations furthest away from
the plane of effective confinement, the confining effect would be reduced. With the experimental
results, the shape of the ineffectively confined concrete region could be assumed to follow an
arch shape.

Copyright © 2012 John Wiley & Sons, Ltd. Struct. Design Tall Spec. Build. 23, 403–426 (2014)
DOI: 10.1002/tal
 UNI-AXIALLY LOADED CFST COLUMNS WITH EXTERNAL RINGS 423

Table 5. Test results and predicted results of CFST columns (other researcher’ results).
N exp
Specimens Nexp Confinement index x Fcal Fcal

(Giakoumelis and Lam, 2004)

C3 948 2.16 1009 0.94
C5 929 1.84 977 0.95
C7 1380 2.58 1323 1.04
C9 1413 1.49 1392 1.01
C10 1038 1.01 1079 0.96
C11 1067 1.02 1076 0.99
C12 998 2.05 977 1.02
C13 948 2.05 975 0.97
(Petrus et al., 2010)
D4M3C1 754 2.13 763 0.99
D4M3C2 730 1.80 770 0.95
D4M3C3 745 1.85 768 0.97
D4M4C1 822 1.32 807 1.02
D4M4C2 788 1.23 820 0.96
D4M4C3 801 1.31 808 0.99
(de Oliveira et al., 2010)
C1-30-3D-C 816 1.13 784 1.04
C2-30-3D-C 1380 2.61 1497 0.92
C2-60-3D-C 1425 1.45 1546 0.92
(O’Shea and Bridge, 2000)
S30CS50B 1662 0.54 1654 1.00
S20CS50A 1678 0.26 1480 1.13
S16CS50B 1695 0.21 1667 1.02
S12CS50A 1377 0.11 1286 1.07
S10CS50A 1350 0.09 1272 1.06
S30CS80A 2295 0.33 2265 1.01
S20CS80B 2592 0.14 2391 1.08
S16CS80A 2602 0.13 2538 1.03
S20CS10A 3360 0.10 3295 1.02
S16CS10A 3260 0.09 3300 0.99
(Han et al., 2005)
CA1-1 312 0.52 310 1.01
CA2-1 822 0.30 748 1.10
CA3-2 1670 0.20 1559 1.07
CA4-1 2783 0.15 2664 1.04
CA5-2 4102 0.12 4062 1.01
CB1-2 415 0.80 391 1.06
CB2-2 920 0.46 869 1.06
CB3-2 1743 0.30 1731 1.01
CB4-2 3011 0.22 2887 1.04
CB5-1 4442 0.18 4336 1.02
CC1-1 432 0.75 400 1.08
CC2-2 1910 0.28 1804 1.06
CC3-1 4720 0.17 4547 1.04
Standard deviation 0.0487

(4) For unconfined CFST columns, the failure mode is the overall buckling of the steel tube with local
buckling occurred near the ends.
(5) For confined CFST columns, buckling of steel tube was limited to occur within the steel rings,
which provided effective lateral restraints against buckling. For HSCFST columns, because of
the brittleness of HSC, fracture of steel tube and external rings were also observed. On the other
hand, fracture of steel tube and rings were not observed in NSCFST columns.
Lastly, an analytical model was proposed to predict the axial load-carrying capacity of uncon-
fined and ring-confined CFST columns. The model took into account the confinement effects

Copyright © 2012 John Wiley & Sons, Ltd. Struct. Design Tall Spec. Build. 23, 403–426 (2014)
DOI: 10.1002/tal
424 M. H. LAI AND J. C. M. HO

provided by the steel tube and the external rings. The proposed model was verified by comparing
with the test results obtained by the authors and other researchers. The comparison verified that the
proposed model predicted very well the axial load-carrying capacity of CFST columns, in which
most of the differences between the theoretically predicted and experimentally measured values
were lying within 10%.

LIST OF NOTATIONS

esr Tensile strain of the mild steel ring

x Confinement index
ssθ Hoop stress provided by the steel tube
ssy Initial yield stress of the steel tube
ssz Axial stress of the steel tube at yield condition (bi-axial condition)
Ac Contact concrete area
A c0 Equivalent concrete area
Acs Composite section area
As Contact steel area
A s0 Equivalent steel area
Ast Total steel area, steel rings included
Ast0 Equivalent total steel area, steel rings included
CFST Concrete-filled-steel tube
d Diameter of the mild steel, 8 mm
D Outer diameter of the steel tube
D0 Equivalent outer diameter of the steel tube
E1 Modulus measured by method 1
E2 Modulus measured by method 2
E3 Modulus measured by method 3
Ec Modulus of the concrete
Ecs Modulus of the composite section
Es Modulus of the steel tube
Esr Modulus of the mild steel ring
Fcal Calculated axial load-carrying capacity
f c0 Unconfined concrete cylinder strength
fcc Confined concrete stress
fr Confining pressure
fsr Stress of steel ring
fyr Yield stress of the mild steel ring
H Height of the steel tube
HSC High-strength concrete
HSCFST High-strength concrete-filled-steel-tube
HST Hollow steel tube
D c0 Equivalent inner diameter of steel tube
k Confinement coefficient = 4.1
n Number of external steel rings
N Axial load-carrying capacity
Nc Applied load
NSC Normal-strength concrete
NSCFST Normal-strength concrete-filled-steel-tube
RC Reinforced concrete
S Spacing of the external rings
S0 Equivalent spacing of rings for the equivalent section
t Thickness of the steel tube
tg Thickness of the gypsum

Copyright © 2012 John Wiley & Sons, Ltd. Struct. Design Tall Spec. Build. 23, 403–426 (2014)
DOI: 10.1002/tal
 UNI-AXIALLY LOADED CFST COLUMNS WITH EXTERNAL RINGS 425

ACKNOWLEDGEMENTS

The work described in this paper has been substantially supported by a grant from the Research Grants
Council of the Hong Kong Special Administrative Region, China (Project No. HKU 712310E). Technical
supports for the experimental tests provided by the laboratory staff of the Department of Civil Engineering,
The University of Hong Kong, are gratefully acknowledged.

REFERENCES
Bai ZZ, Au FTK. 2009. Effects of strain hardening of reinforcement on flexural strength and ductility of reinforced concrete
columns. The Structural Design of Tall and Special Buildings 20(7): 784–800.
Bambach MR, Jama H, Zhao XL, Grzebieta RH. 2008. Hollow and concrete filled steel hollow sections under transverse impact
loads. Engineering Structures 30(10): 2859–2870.
Cai SH, Jiao ZS. 1984. Behavior and ultimate strength of short concrete-filled steel tubular columns. Journal of Building
Structures 6: 13–28.
Chitawadagi MV, Narasimhan MC. 2009. Strength deformation behaviour of circular concrete filled steel tubes subjected to pure
bending. Journal of Constructional Steel Research 65(8–9): 1836–1845.
Chitawadagi MV, Narasimhan MC, Kulkarni SM. 2010. Axial strength of circular concrete-filled steel tube columns—DOE
approach. Journal of Constructional Steel Research 66(10): 1248–1260.
Chung HS, Moon BW, Lee SK, Park JH, Min KW. 2009. Seismic performance of friction dampers using flexure of RC shear
wall system. The Structural Design of Tall and Special Buildings 18(7): 807–822.
Cusson D, Paultre P. 1994. High strength concrete columns confined by rectangular ties. Journal of Structural Engineering
120(3): 783–804.
Dabaon M, El-Khoriby S, El-Boghdadi M, Hassanein MF. 2009. Confinement effect of stiffened and unstiffened concrete-filled
stainless steel tubular stub columns. Journal of Constructional Steel Research 65(8): 1846–1854.
Ding FX, Yu ZW, Bai Y, Gong YZ. 2011. Elasto-plastic analysis of circular concrete-filled steel tube stub columns. Journal of
Constructional Steel Research 67(10): 1567–1577.
Elchalakani M, Zhao XL, Grzebieta RH. 2001. Concrete-filled circular steel tubes subjected to pure bending. Journal of
Constructional Steel Research 57(11): 1141–1168.
Ellobody E, Young B. 2006. Design and behaviour of concrete-filled cold-formed stainless steel tube columns. Engineering
Structures 28(5): 716–728.
Elremaily A, Azizinamini A. 2002. Behavior and strength of circular concrete-filled tube columns. Journal of Constructional
Steel Research 58(12): 1567–1591.
Fam A, Qie FS, Rizkalla S. 2004. Concrete-filled steel tubes subjected to axial compression and lateral cyclic loads. Journal of
Structural Engineering 130(4): 631–640.
Ferretti E. 2004. On Poisson’s ratio and volumetric strain in concrete. International Journal of Fracture 126(3): 49–55.
Giakoumelis G, Lam D. 2004. Axial capacity of circular concrete-filled tube columns. Journal of Constructional Steel Research
60(7): 1049–1068.
Han LH, Yao GH, Zhao XL. 2005. Tests and calculations for hollow structural steel (HSS) stub columns filled with self-consolidating
concrete (SCC). Journal of Constructional Steel Research 61(9): 1241–1269.
Han LH, Yao GH, Tao Z. 2007. Performance of concrete-filled thin-walled steel tubes under pure torsion. Thin-Walled Structures
45(1): 24–36.
Han LH, He SH, Liao FY. 2011. Performance and calculations of concrete filled steel tubes (CFST) under axial tension. Journal
of Constructional Steel Research 67(11): 1699–1709.
Havaei GR, Keramati A. 2011. Experimental and numerical evaluation of the strength and ductility of regular and cross spirally
circular reinforced concrete columns for tall buildings under eccentric loading. The Structural Design of Tall and Special
Buildings 20(2): 247–256.
Heo JS, Lee SK, Park E, Lee SH, Min KW, Kim H, Jo J, Cho BH. 2009. Performance test of a tuned liquid mass damper
for reducing bidirectional responses of building structures. The Structural Design of Tall and Special Buildings 18(7):
789–805.
Ho JCM. 2011. Limited ductility design of reinforced concrete columns for tall buildings in low to moderate seismicity regions.
The Structural Design of Tall and Special Buildings 20(1): 102–120.
Ho JCM, Pam HJ. 2003. Inelastic design of low-axially loaded high-strength reinforced concrete columns. Engineering Structures
25(8): 1083–1096.
Ho JCM, Au FTK, Kwan AKH. 2005. Effects of strain hardening of steel reinforcement on flexural strength and ductility of
concrete beams. Structural Engineering and Mechanics 19(2): 185–198.
Ho JCM, Lam JYK, Kwan AKH. 2010. Effectiveness of adding confinement for ductility improvement of high-strength concrete
columns. Engineering Structures 32(3): 714–725.
Ho JCM, Lam JYK, Kwan AKH. 2012. Flexural ductility and deformability of concrete beams incorporating high performance
materials. The Structural Design of Tall and Special Buildings 21(2): 114–132.
Huang CS, Yeh YK, Liu GY, Hu HT, Tsai KC, Weng YT, Wang SH, Wu MH. 2002. Axial load behavior of stiffened concrete-filled
steel columns. Journal of Structural Engineering 128(9): 1222–1230.
Khoshnoudian F, Azad AI. 2011. Effect of two horizontal components of earthquake on nonlinear response of torsionally
coupled base isolated structures. The Structural Design of Tall and Special Buildings 20(8): 986–1018.

Copyright © 2012 John Wiley & Sons, Ltd. Struct. Design Tall Spec. Build. 23, 403–426 (2014)
DOI: 10.1002/tal
426 M. H. LAI AND J. C. M. HO

Kwan AKH. 2000. Use of condensed silica fume for making high-strength, self-consolidating concrete. Canadian Journal of
Civil Engineering 27(4): 620–627.
Lam JYK, Ho JCM, Kwan AKH. 2009. Maximum axial load level and minimum confinement for limited ductility design of
concrete columns. Computers and Concrete 6(5): 357–376.
Lee SK, Lee SH, Min KW, Moon BW, Youn KJ, Hwang JS. 2009. Performance evaluation of an MR damper in building
structures considering soil–structure interaction effects. The Structural Design of Tall and Special Buildings 18(1): 105–115.
Liao FY, Han LH, He SH. 2011. Behavior of CFST short column and beam with initial concrete imperfection: experiments.
Journal of Constructional Steel Research 67(12): 1922–1935.
Lu X, Hsu CTT. 2007. Tangent Poisson’s ratio of high-strength concrete in triaxial compression. Magazine of Concrete
Research, 59(1), 69–77.
Lu H, Han LH, Zhao XL. 2009. Analytical behavior of circular concrete-filled thin-walled steel tubes subjected to bending.
Thin-Walled Structures 47(3): 346–358.
Mander JB, Priestley MJN, Park R. 1988. Theoretical stress strain model for confined concrete. Journal of Structural Engineering
114(8): 1804–1826.
Marano GC, Greco R. 2009, Robust optimum design of tuned mass dampers for high rise buildings under moderate earthquakes.
The Structural Design of Tall and Special Buildings 18(8): 823–838.
Mehrparvar B, Khoshnoudian F. 2011. Efficiency of active systems in controlling base isolated buildings subjected to near fault
earthquakes. The Structural Design of Tall and Special Buildings 20(8): 1019–1034.
Mohebbi M, Joghataie A. 2012. Designing optimal tuned mass dampers for nonlinear frames by distributed genetic algorithms.
The Structural Design of Tall and Special Buildings 21(1): 57–76.
Nie JG, Huang Y, Yi WJ, Fan JS. 2012. Seismic behavior of CFRSTC composite frames considering slab effects. Journal of
Constructional Steel Research 68(1): 165–175.
de Oliveira WLA, De Nardin S, EI Debs ALHC, El Debs MK. 2010. Evaluation of passive confinement in CFT columns.
Journal of Constructional Steel Research 66(4): 487–495.
O’Shea MD, Bridge RQ. 2000. Design of circular thin-walled concrete filled steel tubes. Journal of Structural Engineering
126(11): 1295–1303.
Pam HJ, Ho JCM. 2009. Length of critical region for confinement steel in limited ductility high-strength reinforced concrete
columns. Engineering Structures 31(12): 2896–2908.
Park R. 2001. Improving the resistance of structures to earthquakes. Bulletin of the New Zealand National Society for Earthquake
Engineering 34(1): 1–39.
Persson B. 1999. Poisson’s ratio of high-performance concrete. Cement and Concrete Research 29(10): 1647–1653.
Petrus C, Hamid HA, Ibrahim A, Parke G. 2010. Experimental behaviour of concrete filled thin walled steel tubes with tab
stiffeners. Journal of Constructional Steel Research 66(7): 915–922.
Ribakov Y. 2011a. Base isolated structures with selective controlled semi active friction dampers. The Structural Design of Tall
and Special Buildings 20(7): 757–766.
Ribakov Y. 2011b. Using viscous and variable friction dampers for improving structural seismic response. The Structural Design
of Tall and Special Buildings 20(5): 579–593.
Ribakov Y, Agranovich G. 2011a. Control of structural seismic response by a limited set of active dampers. The Structural
Design of Tall and Special Buildings 20(5): 594–611.
Ribakov Y, Agranovich G. 2011b. A method for design of seismic resistant structures with viscoelastic dampers. The Structural
Design of Tall and Special Buildings 20(5): 566–578.
Sabol TA, Nishi DE. 2011. Application of performance based design to an eccentrically braced frame structure. The Structural
Design of Tall and Special Buildings 20(S1): S76–S84.
Uy B. 1998. Ductility, strength and stability of concrete-filled fabricated steel box columns for tall buildings. The Structural
Design of Tall and Special Buildings 7(2): 113–133.
Yamamoto K, Fujita K, Takewaki I. 2011. Instantaneous earthquake input energy and sensitivity in base isolated building. The
Structural Design of Tall and Special Buildings 20(6): 631–648.
Yan ZH, Au FTK. 2010. Nonlinear dynamic analysis of frames with plastic hinges at arbitrary locations. The Structural Design
of Tall and Special Buildings 19(7): 778–801.
Zhong ST. 2006. Unified Theory of CFST: Research and Application (first edn). Tsinghua University Press: Beijing, China. [In Chinese]
Zhou Y, Lu XL, Lu WS, Qian J. 2011. Study on the seismic performance of a multi tower connected structure. The Structural
Design of Tall and Special Buildings 20(3): 387–401.

AUTHORS’ BIOGRAPHIES
Mianheng H. Lai is PhD student at the Department of Civil Engineering, The University of Hong
Kong. Research interests are on flexural strength, ductility and deformability of concrete-filled-steel-
tube columns with and without confinements.

Johnny C. M. Ho (MHKIE, MIEAust, MIStructE) is an assistant professor at the Department of Civil

Engineering, The University of Hong Kong. His research interests are on ductility and deformability of
high-strength concrete beams and columns, plastic hinge analysis of reinforced concrete members and
behaviour of concrete-filled-steel-tube columns with external confinement.

Copyright © 2012 John Wiley & Sons, Ltd. Struct. Design Tall Spec. Build. 23, 403–426 (2014)
DOI: 10.1002/tal