Journal of Constructional Steel Research 65 (2009) 16071616

Contents lists available at ScienceDirect

Journal of Constructional Steel Research

journal homepage: www.elsevier.com/locate/jcsr

Performance of concrete filled steel tube reinforced concrete columns subjected

to cyclic bending
Lin-Hai Han a, , Fei-Yu Liao a , Zhong Tao b , Zhe Hong b
a

Department of Civil Engineering, Tsinghua University, Key Laboratory of Structural Engineering and Vibration of China Education Ministry, Beijing, 100084, PR China

College of Civil Engineering, Fuzhou University, Fuzhou, Fujian Province, 350108, PR China

article

info

Article history:
Received 20 January 2009
Accepted 20 March 2009
Keywords:
Concrete filled steel tube (CFST)
Reinforced concrete (RC)
Beam column
Hysteretic behaviour
Ductility

abstract
This paper reports nine test results of concrete filled steel tube reinforced concrete (CFSTRC) columns,
which were tested under constant axial load and cyclically increasing flexural loading. The main
parameters varied in the experiments were axial load level and cross-sectional type. The influence of these
parameters on strength, ductility, stiffness and energy dissipation was investigated. It was found that, in
general, CFSTRC columns exhibit favourable energy dissipation and ductility, even when the columns
were subjected to high axial loads. This type of composite column is adoptable in practical engineering,
particularly in regions of high seismicity.
2009 Elsevier Ltd. All rights reserved.

1. Introduction
In China, an innovative type of composite column, named as
concrete filled steel tube reinforced concrete (CFSTRC) column in
this paper, has been utilised in some high-rise buildings recently.
This new composite technology has attracted more and more
research interest. Fig. 1 gives a schematic view of three typical
cross-sections.
For innovative composite columns, the following advantages
over conventional concrete filled steel tubular (CFST) columns are
expected:
(1) Higher stiffness.
(2) The beam-column joints of the CFSTRC column system can
be designed according to the well established knowledge of
conventional reinforced concrete (RC) beam-column joints.
(3) Higher fire resistance due to the protection of the inner steel
tubes provided by the concrete.
(4) Corrosion of the steel tubes can be prevented due to protection
from the outer concrete.
(5) The possibility of outward buckling of the steel tubes is
virtually prohibited due to the restraint of the outer concrete.
Compared with conventional RC columns, the following
advantages are expected:
(1) Higher ductility owing to the existence of the inner CFST.

Corresponding author. Tel.: +86 10 62797067; fax: +86 10 62781488.

E-mail addresses: lhhan@mail.tsinghua.edu.cn, lhhan@tsinghua.edu.cn
(L.-H. Han).
0143-974X/$ see front matter 2009 Elsevier Ltd. All rights reserved.
doi:10.1016/j.jcsr.2009.03.013

(2) Faster construction speed is expected. In practice, the inner

CFSTs can be built first to bear the construction load alone, and
the outer concrete can be poured later.
Some previous research has been carried out on CFSTRC
columns [1,2], which demonstrates the fact that this type of
composite column generally has good performance. However,
there is still a lack of information on the composite members under
cyclic loading. It indicates a need for further research in this area.
This paper thus tested nine CFSTRC columns under constant
axial load and cyclically increasing flexural loading. The main
objectives of this research were twofold: first, to report a series of
new tests on CFSTRC columns. And second, to investigate influence
of some parameters, including axial load level and cross-sectional
type, on the seismic behaviour of CFSTRC columns.
2. Experimental programme
2.1. General
Nine CFSTRC specimens were tested. Test parameters were
the section type and the axial load level (n). Three kinds of
cross-section were selected as shown in Fig. 1 (a), (b) and (c),
respectively, i.e., square RC column reinforced with square CFST,
square RC column reinforced with circular CFST, and circular RC
column reinforced with circular CFST. The axial load level (n) in
this paper is defined as follows, i.e.,
n=

No
Nu

(1)

1608

L.-H. Han et al. / Journal of Constructional Steel Research 65 (2009) 16071616

Fig. 1. Typical concrete filled steel tube reinforced concrete (CFSTRC) column cross-sections.

Table 1
Summary of specimen information.
Test
series

Outer section
type

Inner steel
tube type

Specimen
label

Dimension of
inner steel
tube (mm)

Axial load No
(kN)

Axial load
level n

Pue
(kN)

Kie
(kN m2 )

Kse
(kN m2 )

Ductility
ratio

Dissipated
energy Etotal
(kN m)

Square

Square
Square
Square

SS1
SS2
SS3

50 50 2.7
50 50 2.7
50 50 2.7

0
287
574

0
0.3
0.6

49.5
62.6
61.8

675
1404
1824

585
1169
1616

4.02
3.91
3.18

28.0
18.5
10.6

II

Square

Circular
Circular
Circular

SC1
SC2
SC3

60 2.0
60 2.0
60 2.0

0
282
564

0
0.3
0.6

42.6
55.8
61.8

707
1402
1778

616
1218
1611

4.15
3.89
2.36

34.3
16.2
7.8

III

Circular

Circular
Circular
Circular

CC1
CC2
CC3

60 2.0
60 2.0
60 2.0

0
219
438

0
0.3
0.6

29.2
31.0
30.2

475
799
1104

305
449
813

4.18
3.71
3.33

40.8
12.7
7.0

Notation
As
Ac
D
Ec
Es
fy
fc 0
fcu
L
M
Mu
n
No
Nu
P
Pue

Steel cross-sectional area

Concrete cross-sectional area
Sectional dimension
Concrete modulus of elasticity
Steel modulus of elasticity
Yield strength of steel
Concrete cylinder strength
Concrete cube strength
Effective buckling length of column in the plane of
bending
Moment
Moment capacity
Axial load level (= No /Nu )
Axial compressive load
Compressive capacity of the composite column
Lateral load
Ultimate lateral load capacity
Lateral displacement
Strain
Curvature
Ductility coefficient

where No is the axial load applied to the composite specimens;

Nu is the axially compressive capacity of the composite columns.
The value of Nu was determined by using the mechanics model
described by Han et al. [3]. Concrete strength measured at the time
of tests was used in the calculations.
The details of each column are presented in Table 1, where the
specimens with SS, SC and CC in the specimen labels refer
to columns with cross-sections as shown in Fig. 1 (a), (b) and (c),
respectively. For clarity, they were designated as test series I, II and
III respectively in this paper.
Owing to the loading limitation of the test machine and the
high axial load level (n), the overall sectional dimension (D) for all
specimens was designed as 150 mm. This dimension was expected

to be similar to some low rise buildings, such as some workshop

structures. However, the column dimension is smaller than that
of columns used in some engineering practice, such as high rise
buildings and long span bridges. Thus, there is still potential
research need to clarify the size effect on the column behaviour.
All specimens were longitudinally reinforced with four 10 mm
diameter deformed reinforcements and 6 mm diameter plain
transverse reinforcements spaced at 100 mm centres. The clear
cover to the stirrups was 15 mm.
2.2. Material properties
Tension tests were carried out to determine the material
properties of the steel tubes and rebars. Table 2 lists the measured
average yield strength (fy ), tensile strength (fu ) and modulus of
elasticity (Es ).
The concrete mix was designed for a compressive cube strength
(fcu ) at 28 days of approximately 40 N/mm2 . The mix proportions
were as follows:

Cement: 538 kg/m3

Water: 205 kg/m3
Sand: 598 kg/m3
Coarse aggregate: 1109 kg/m3 .

In all the concrete mixes, the fine aggregate used was silicabased sand, the coarse aggregate was carbonate stone. For each
batch of concrete mixture, three 150 mm cubes were also cast and
cured in conditions similar to the related specimens. The average
cube strengths (fcu ) of the concrete used to fabricate specimens in
series I and II at 28 days and the time of tests were 33.6 N/mm2
and 52.4 N/mm2 respectively, whilst those for the specimens in
series III were 32.2 N/mm2 and 45.0 N/mm2 respectively.
2.3. Specimens preparation
The steel tubes for all inner CFSTs were made of cold-formed
steel. The measured inner corner radius (ri ) for the square tubes
was 3 mm. The ends of the steel tube sections were cut and
machined to the required length. The insides of the tubes were wire

L.-H. Han et al. / Journal of Constructional Steel Research 65 (2009) 16071616

1609

Table 2
Material properties of steel.
Steel type

fy (N/mm2 )

fu (N/mm2 )

Es (N/mm2 )

Square tube (50 50 2.7)

Circular tube (60 2.0)
10 Reinforced bar
6 Reinforced bar

356
353
417
322

394
442
551
467

1.96 105
1.96 105
1.95 105
1.96 105

Fig. 3. Typical loading history (n = 0).

Fig. 2. Test set-up.

brushed to remove any rust and loose debris. Each tube was welded
to a circular (for circular specimen) or square (for square specimen)
steel base plate of 12 mm thickness. As soon as the inner steel tube
was prepared, the outer reinforcement cage was then fabricated
and installed. All specimens were cast in plywood moulds. In order
to ensure that the concrete in and out the steel tube have a same
strength, they were poured together in layers. The specimens were
placed upright to air-dry. During curing, a very small amount of
longitudinal shrinkage of 0.7 to 0.9 mm or so occurred at the top of
the column. A high-strength epoxy was used to fill this longitudinal
gap so that the concrete surface was flush with the steel tube at
the top.
2.4. Cyclic test apparatus
The specimens were tested under combined constant axial
load and cyclically increasing flexural load. The test length of the
specimens was 1500 mm.
The in-plane curvatures were measured at locations along the
specimen by two displacement transducers. Eight strain gauges
were used for the longitudinal bars and inner tube respectively for
each specimen to measure strains at the mid-span.
Fig. 2 gives a general view of the beam-column test setup.
The ends of the CFSTRC specimens were attached to cylindrical
bearings and were free to rotate in-plane, and thus simulating pinpin end conditions. The axial load (No ) was applied and maintained
constantly by a 1000 kN hydraulic ram. A hydraulic pump was
used to control the axial load. Precautions were made to avoid any
eccentricity in the axial load application by very careful alignment
of the test set-up.
The flexural loading was applied by imposing cyclically lateral
loading in the mid-span of the specimen. The specimen was
confined in the middle part by a very rigid stub made of high
strength steel as shown in Fig. 2. The stub was designed and was
made of two separate halves of a box with a concentric hole that
exactly fit the CFSTRC specimen. The two halves were pushed
against the CFSTRC specimen and connected together using eight
high strength bolts. The stub may provide effective confinement

along the middle 180 mm of the specimens with a circular section,

and 150 mm of the specimens with square section respectively,
which is generally similar to the effect of a rigid floor system
intersecting the column at the floor level in a building. The stub was
connected to a MTS hydraulic ram having 250 kN capacity. Both
strain gauges and string pots were used to measure the bending
curvature. The in-plane displacements were measured at locations
along the specimen test-length by displacement transducers.
The tests were performed under displacement controlled cyclic
loading, where the typical pattern of lateral loading was shown in
Fig. 3. The cyclic loading history including three cycles at the midspan displacement of 3.75 mm, 1.88 mm or 1.25 mm, respectively
for the specimen with axial load level n = 0, 0.3 or 0.6, was
determined as 1/10 of the mid-span displacement corresponding
to the peak load.
No displacements at the reaction blocks were observed until the
beam-column bowed after reaching the failure load.
3. Specimen behaviour and experimental results
All the CFSTRC specimens behaved in a relatively ductile
manner and testing proceeded in a smooth and controlled way.
Fig. 4 presents a general view of the nine specimens after
tests. From the test observations, it can be seen that all the
specimens showed a flexure-compression failure characteristic,
and no diagonal crack caused by shear stress was observed during
the whole testing process. The failure patterns were summarised
as following:
(1) Specimens with an axial load level n = 0 (SS1, SC1 and CC1).
It was found that the failure process was generally similar for
all specimens with a same axial load level of n = 0. When the
lateral displacement attained 3.75 mm, some initial cracks about
0.03 mm in width occurred at the tensile zone of a specimen near
the mid-span. At this time, the strains of the concrete, steel tube
and longitudinal steel bars grew linearly in general, and the lateral
load (P ) versus lateral displacement () curve kept in the elastic
stage.
Hereafter with increasing lateral displacement, the concrete
cracks developed successively and formed several main cracks

1610

L.-H. Han et al. / Journal of Constructional Steel Research 65 (2009) 16071616

n=0

n=0.3

n=0.6

n=0

n=0.3

n=0.6

n=0

n=0.3

n=0.6

Fig. 4. A general view of the specimens after tests.

(a) SS3, n = 0.6.

(b) SC3, n = 0.6.

(c) CC3, n = 0.6.

Fig. 5. Distribution of deflection along length of typical specimens.

around the specimen section. When the ultimate load capacity

(Pue ) was attained, the maximum crack width reached 0.3 mm.
From the measured strains, it was found that tensile yielding had
taken place for the steel tube and longitudinal steel bars at this
time. Then the width of the cracks developed faster obviously, and
minor spalling of the concrete occurred at the compressive zone.
Upon further cycling, crushing and spalling of the concrete
occurred successively and the strength of the specimen began
to deteriorate. At the failure stage, an outward bulge of the
longitudinal bars could be observed. However, the specimen still
retained a high residual strength due to the existence of the core
CFST. A minor outward bulge for the inner steel tube, forming a
complete ring on each side of the stub, was observed after the test.

(2) Specimens with an axial load level n = 0.3 (SS2, SC2 and CC2).
For the CFSTRC specimens with the same axial load level of
n = 0.3, when the lateral displacement attained 3.755.63 mm,
some initial cracks about 0.02 mm in width occurred on each side
of the stub. Then with an increase of the lateral displacement,
the concrete cracks gradually grew forming several main cracks
around the specimen section, and reached a maximum width of
0.25 mm prior to the ultimate load capacity (Pue ).
When the ultimate load capacity (Pue ) was attained, the steel
tube and the longitudinal steel bars just reached their yield strains
and spalling of the concrete began to gradually occur.
Upon further cycling, the strength of the specimens began to
deteriorate and spalling of the concrete cover on both sides of the

L.-H. Han et al. / Journal of Constructional Steel Research 65 (2009) 16071616

1611

(1) Series I.

(2) Series II.

(3) Series III.

Fig. 6. Lateral load (P ) versus lateral displacement () hysteretic curves.

stub occurred subsequently. The failure process of the specimens

exhibited comparatively good ductility owing to the existence of
the core CFST. After removing the outer concrete, no obvious local
bucking of the inner tube was observed due to the confinement of
the outer concrete.
(3) Specimens with an axial load level n = 0.6 (SS3, SC3 and CC3).
The failure process was generally similar for all the specimens
with a same axial load level of n = 0.6. It was found that,
the occurrence of the concrete cracking for the specimens with
this higher axial load level was postponed compared with the
specimens with a lower axial load level, owing to the fact that
a higher axial load could result in higher compressive stress in
the concrete and thus postponed the occurrence of the concrete

cracking. For a test specimen with an n = 0.6, initial cracks about

0.02 mm wide just occurred on each side of the stub, when the
lateral displacement attained 6.25 mm.
Before the ultimate load capacity (Pue ) was attained, the cracks
developed slowly and reached a maximum width of 0.18 mm.
In the failure stage, the strength of the specimens deteriorated
quickly with severe crushing and spalling of the concrete cover.
The failure process showed a decrease in ductility compared with
specimens with a lower axial load level.
From the measured strains, it could be seen that yielding had
taken place for the compressive longitudinal bars and the steel
tube at the ultimate load (Pue ), whilst the tensile longitudinal
bars had not yielded at this time. The failure mode indicates that

1612

L.-H. Han et al. / Journal of Constructional Steel Research 65 (2009) 16071616

decreasing stiffness, until the ultimate moment is reached. A

careful examination of the test results revealed that, in general,
the moment versus curvature relationship enters the inelastic
stage at about 20% of the moment capacity (Mu ), so the initial
section flexural stiffness (Kie ) is defined as the secant stiffness
corresponding to the moment of 0.2Mu . And the serviceabilitylevel section flexural stiffness (Kse ) is defined as the secant stiffness
corresponding to the serviceability-level moment of 0.6Mu on the
moment versus curvature curve [6]. The points corresponding to
0.2Mu , 0.6Mu and Mu in both positive and negative directions were
labelled in Fig. 7, respectively. The initial section flexural stiffness
(Kie ) and the serviceability-level section flexural stiffness (Kse ) of
the tested specimens so determined are presented in Table 1.
The formula presented in EC4 [7] for calculating the elastic
flexural stiffness of composite columns is the following:
Fig. 7. Typical moment (M ) versus curvature () envelope relation (SC2, n = 0.3).

the specimens have showed compression-dominant behaviour. No

local buckling of the inner tube was observed after removing the
outer concrete.
It was found that, in general, the measured deflection curves
are in the shape of a half-sine wave. Fig. 5 shows a typical set of
deflection curves. The sinusoids with the same maximum values
are also shown in Fig. 5 (dashed lines). Therefore, within the
limitations of the current testing, it seems that the deflection curve
of the composite members can be assumed as a half sine wave in
the analysis.
The test curves of lateral load (P ) versus the mid-span
displacement () for all specimens are shown in Fig. 6. It can be
found that, in general, the axial load level (n) has a significant
influence on the shape of the hysteretic curves, particularly in
the unloading stage. As the axial load level increases, the slope
of the curve at the post-peak stage becomes much steeper and
the pinching effect of the hysteretic loop is more evident. In
general, the features of CFSTRC columns under constant axial load
and cyclically increasing flexural loading are similar to those of
conventional CFST columns [4,5], although the CFSTRC columns do
show worse ductility and energy dissipation behaviour under high
axial load level (n).
Table 1 gives the maximum lateral loads (Pue ) obtained in all
tests, calculated as the average values of the positive and negative
maximum lateral loads obtained from the lateral load (P ) versus
lateral displacement () responses.
Specimen SC2 (with an axial load level of 0.3) is selected
to demonstrate the typical envelope relation of the measured
moment (M ) versus curvature () graphs, shown as in Fig. 7.
The moment versus curvature diagram shows that there is an
initial elastic response, then inelastic behavior with gradually

(a) Initial stiffness.

Kc = Es Is + Es Ir + 0.6 Ec Ic

(2)

where Es = 206 000 (N/mm2 ); Ec = 22 000 (fc0 /10)0.3 (N/mm2 );

Is , Ir and Ic are the second moments of area for the inner steel
tube, the longitudinal reinforcement and the concrete section
respectively.
Eq. (2) is tentatively used to predict the section flexure stiffness
(Kc ) of the CFSTRC columns in this paper. Fig. 8 (a) and (b) show
the Kc /Kie versus Kie and Kc /Kse versus Kse relations, respectively.
As can be seen from Fig. 8 (a) and (b), that, within the
limitations of current testing, the values of Kie and Kse of the CFSTRC
specimens tend to increase with the increase of the axial load
level (n) generally. This is attributed to the fact that the increased
compressive axial load may lead to an increase of the compressive
area of the CFSTRC section under bending, and thus postpones the
concrete cracking.
Compared with the test results, in general, the EC4 [7] method
gives a higher prediction of Kie and Kse of the specimens with an
axial load level of n = 0, whilst it gives a lower prediction for
the specimens with an axial load level of n = 0.3 or 0.6. The
mean values of Kc /Kie and Kc /Kse are 0.93 and 1.2 respectively, with
the standard deviations of 0.39 and 0.5 respectively. It seems that
there is further research need to develop a more rational model in
predicting the section flexure stiffness of CFSTRC.
4. Analysis of test results and discussion
4.1. Effects of axial load level (n)
Fig. 9 shows the influence of the axial load level (n) on the
lateral load (P ) versus lateral displacement () envelope curves of
the specimens. It can be found from this figure that, the axial load
level (n) not only influences the ultimate lateral load (Pue ), but also
the ductility of the specimens. Generally, a higher axial load level

(b) Serviceability-level stiffness.

Fig. 8. Comparisons between test results and EC4 code predictions.

L.-H. Han et al. / Journal of Constructional Steel Research 65 (2009) 16071616

(a) Series I.

(b) Series II.

1613

(c) Series III.

Fig. 9. Influence of axial load level on lateral load (P ) versus lateral displacement () envelope curves.

(a) Series I.

(b) Series II.

(c) Series III.

Fig. 10. Influence of axial load level on moment (M ) versus curvature () envelope curves.

(n) results in a higher Pue , however the ductility of the specimens

decreases with an increase in the axial load level (n).
Fig. 10 shows the influence of the axial load level (n) on the
moment (M ) versus curvature () envelope curves. It can be
concluded that, within the limitations of current testing, both the
flexural stiffness and moment capacity of the specimens increase
with an increase of the axial load level.

load, thus this specimen exhibited tension-dominant behaviour.

For the specimen with n = 0.3, the strain of the longitudinal
bar attained yield strain both in tension and in compression under
cyclic loading. For the specimen with n = 0.6, the strain of
the longitudinal bar was mainly in compression and attained
compressive yield strain prior to the peak load. In this case, the
specimen exhibited compression-dominant behaviour.

4.2. Longitudinal strain

4.3. Ductility ratio

Typical responses of the measured lateral load (P ) versus

the measured steel strain () graphs are shown in Figs. 11(1)
and (2), respectively, for the steel tubes and the longitudinal
steel bars. The yield strains (y ) of the steel are also labelled
in Fig. 11. The presented strains in Fig. 11 are defined as the
tension corresponding to the positive values and compression to
the negative values respectively. It can be found that, generally,
the load decreases linearly in the unloading stage with comparable
stiffness in the initial loading stage. After that, the loadstrain
curves exhibit nonlinear softening.
The extreme fibres of the steel tubes for the specimens with a
relatively low axial load level (n = 0 or 0.3) were mainly in tension
under cyclic loading and finally attained tensile yielding, whilst
those for the specimens with a higher axial load level (n = 0.6)
were generally in compression and had attained yielding prior to
the peak loads, as shown in Fig. 11(1).
From Fig. 11(2), it was found that, with the increasing axial load
level (n), the compressive strains of the longitudinal bars tended
to increase whilst the tension strains decreased. For the specimen
CC1 with n = 0, the strain of the longitudinal bar was mainly
in tension and attained tensile yield strain (y ) prior to the peak

For convenience of analysis, the ductility coefficient () is

defined to quantify the ductility of CFSTRC columns subjected to
constant axial load and cyclically increasing flexural loading. It is
expressed as:

u
y

(3)

where y is the yield displacement (in this paper, y is calculated

using the method presented in Han et al. [3], as shown in Fig. 12);
and u is the displacement when the axial load falls to 85% of the
ultimate strength (Pue ), shown in Fig. 12. max is the displacement
corresponding to Pue .
Fig. 13 shows the relationship between the ductility coefficient
() and the axial load level (n), it can be found that, decreases
with an increase of the axial load level (n).
4.4. Axial shortening
Fig. 14 shows the axial shortening (d) of typical specimens
through progressive loading cycles. It can be seen that, generally,
the axial shortening increases in a smooth way during the elastic

1614

L.-H. Han et al. / Journal of Constructional Steel Research 65 (2009) 16071616

(1) Steel tube.

(2) Longitudinal steel bar.

Fig. 11. Typical lateral load (P ) versus longitudinal strain () relations.

Fig. 13. Influence of the axial load level (n) on the ductility coefficient ().
Fig. 12. Typical P envelope curve.

cycles. When the lateral displacement reached about 3y , the axial

shortening started to increase significantly and the lateral load
dropped rapidly.

in which, EI is the measured flexural stiffness of the specimens

corresponding to the different lateral column deflection () and
can be determined as Eq. (5) [8]; EIfirst is the measured flexural
stiffness under the first loading increment.

4.5. Rigidity degradation

p =

The column flexural stiffness ratio (Rk ) is used to assess the

rigidity degradation of the test columns with increasing lateral
deflections. Rk is defined as
Rk =

EI
EIfirst

(4)

Pp L3

3(tan u u)
u3

48EI


(5)

where u = 21 NoEIL , No is the axial load applied to the composite

specimen, L is the overall length of the specimen. Then EI can
be obtained by substituting Pp and p into Eq. (5) and repeating
iterations for it, where Pp and p are the measured peak load

L.-H. Han et al. / Journal of Constructional Steel Research 65 (2009) 16071616

-3

-3

-3

-6

(a) SS3, n = 0.6.

-6

(b) SC3, n = 0.6.

-6

1615

(c) CC3, n = 0.6.

Fig. 14. Axial shortening.

(a) Series I.

(b) Series II.

(c) Series III.

Fig. 15. Rigidity degradation.

(a) Series I.

(b) Series II.

(c) Series III.

Fig. 16. Energy (E ) dissipation ability.

and corresponding displacement respectively under the different

lateral displacement.
Fig. 15 shows Rk versus /y relations of the tested specimens.
From the comparison, it can be found that the rigidity degradation
for members with a lower axial load level (n) was more significant
than that with a higher axial load level. This phenomenon can be
explained by the increased compression area of the cross-section
for the columns with a higher axial load level.

by the hysteretic hoops. The total energy (Etotal ) dissipated through

each test is shown in Table 1.
It was found from Fig. 16 that, generally, the dissipated energy
(E ) of a specimen with a lower axial load level (n) is higher
relatively under the same /y .
Results in Table 1 indicate that the total energy dissipated
(Etotal ) for a specimen tends to decrease with an increase of the
axial load level (n).

4.6. Dissipated energy

5. Conclusions

Energy dissipation ability is one of the most important indices

to evaluate the seismic performance of a member or structure.
Fig. 16 shows the cumulative energy (E ) versus relative lateral
displacement (/y ) curves, where E is calculated from the lateral
load (P ) versus lateral displacement () curve as the area enclosed

This study provides new data on the cyclic behaviour of nine

CFSTRC columns. The following conclusions can be drawn based
on the limited research reported in the paper.
(1) CFSTRC columns exhibit favourable energy dissipation and
ductility. Generally, the ductility and energy dissipation ability

1616

L.-H. Han et al. / Journal of Constructional Steel Research 65 (2009) 16071616

of the CFSTRC columns decreases with an increase of axial load

level (n). The rigidity degradation for specimens with a lower
axial load level (n) was more significant than that with a higher
axial load level.
(2) The features of CFSTRC columns under constant axial load and
cyclically increasing flexural loading are similar to those of
conventional CFST columns.
(3) CFSTRC columns show good seismic performance, and this
type of composite column is thus expected to be adoptable in
practical building structures.
It should be noted that, further numerical study, e.g. finite
element analysis, is needed to extend the range of test data, and
to investigate the effect of changing variables that has not been
investigated in this paper. The experimental work presented in this
paper has provided a good basis for the development of theoretical
models.
Acknowledgements
The study of this paper is financially supported by the Research
Foundation of the Ministry of Railways and Tsinghua University
(RFMOR & THU) (NO. J2008G011), the National Basic Research

Program of China (973 Program) (Grant No. 2009CB623200) and

the Program for Changjiang Scholars and Innovative Research
Team in University (IRT00736). The financial support is highly
appreciated.
References
[1] Lin YJ, Cheng WR, Xu M, Zuo J. Experimental study on limit values of
axial compression ratio of SRC columns with circular steel tube. China Civil
Engineering Journal 2001;34(6):238 [in Chinese].
[2] Wang LJ, Chi YH. Experimental research on the bearing capacity of axially loaded
composite columns with concrete core encased by steel tube. In: Proceedings
of the fourth international conference on advances in steel structures. 2005.
p. 5938.
[3] Han LH, Tao Z, Wang WD. Advanced composite and mixed structures-Testing,
theory and design approach. China Science Press 2009 [in Chinese].
[4] Han LH, Yang YF. Cyclic performance of concrete-filled steel CHS columns under
flexural loading. Journal of Constructional Steel Research 2005;61(4):42352.
[5] Han LH, Yang YF, Tao Z. Concrete-filled thin walled steel RHS beam-columns
subjected to cyclic loading. Thin-walled Structures 2003;41(9):80133.
[6] Varma AH, Ricles JM, Sause R, Lu LW. Seismic behavior and modeling of highstrength composite concrete-filled steel tube (CFT) beam-columns. Journal of
Constructional Steel Research 2002;58(5-8):72558.
[7] Eurocode 4. Design of composite steel and concrete structures, Part 1.1.
General rules and rules for Building. BS EN 1994-1-1. London: British Standards
Institution; 2004.
[8] Elremaily A, Azizinamini A. Behavior and strength of circular concrete-filled
tube columns. Journal of Constructional Steel Research 2002;58(12):156791.