Construction and Building Materials 141 (2017) 501–516

Contents lists available at ScienceDirect

Construction and Building Materials

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

Axial compressive behaviour of steel-jacket retrofitted RC columns with

recycled aggregate concrete
An He a, Jian Cai a,b, Qing-Jun Chen a,b,d,⇑, Xinpei Liu c, Hua Xue a, Chenjie Yu d
a
School of Civil Engineering and Transportation, South China University of Technology, Guangzhou 510640, PR China
b
State Key Laboratory of Subtropical Building Science, South China University of Technology, Guangzhou 510640, PR China
c
Centre for Infrastructure Engineering and Safety, The University of New South Wales, UNSW Sydney, 2052, Australia
d
Delft University of Technology, Delft, The Netherlands

h i g h l i g h t s

Axial compression tests on steel-jacket retrofitted columns with RAC were undertaken.

Variable Poisson’s ratio of material was considered in the finite element model.

Developed an axial load capacity design model for the retrofitted columns.

a r t i c l e i n f o a b s t r a c t

Article history: It is an effective solution to strengthen structural column by using steel jacket and infill concrete for rein-
Received 12 November 2016 forced concrete (RC) structure retrofitting. In this research, the recycled aggregate concrete (RAC) is
Received in revised form 1 March 2017 adopted as an infill concrete instead of using the normal concrete, in order to reduce the carbon footprint
Accepted 3 March 2017
of construction. Fifteen columns including twelve steel-jacket strengthened columns, one unstrength-
Available online 14 March 2017
ened column and two concrete filled steel tube (CFST) columns are tested to investigate the axial com-
pressive behaviour of steel-jacket retrofitted RC columns with RAC. The variables among the tested
Keywords:
specimens include the recycled coarse aggregate replacement ratio, RAC strength, steel tube thickness
Steel jacket
Retrofit
and preload of original column. Experimental results illustrate that using steel jacketing approach for ret-
Recycled aggregate concrete rofitting could significantly improve the strength, stiffness and ductility of the columns. The axial com-
Compression pressive strength of the steel-jacket retrofitted column with RAC is slightly lower than the one with
Finite element model normal infill concrete, but the influence of recycled coarse aggregate replacement ratios is negligible.
The peak strengths of steel-jacket retrofitted columns can be significantly enhanced by increasing the
thickness of steel tube. The performances of the steel-jacket retrofitted columns are similar to that of
the CFST columns. The effect of the preload of original column on the ultimate strength of steel-jacket
retrofitted column is limited. A finite element (FE) model is also developed by using ABAQUS software
to simulate the performance of steel-jacket retrofitted columns. The accuracy of the model is validated
through comparing with the experimental results. By using the developed FE model, the mechanical
behaviour of the column is further discussed in details and extended parametric studies are undertaken
to elucidate the effects of various influencing factors on the behaviour of the column and its components.
The design methods for predicting the axial compressive strength of the retrofitted column with RAC are
suggested.
Ó 2017 Published by Elsevier Ltd.

1. Introduction retrofitting effectiveness and economic efficiency, as well as ease

The steel jacketing approach is a commonly used retrofitting
solution for reinforced concrete (RC) structural column, with high
⇑ Corresponding author at: School of Civil Engineering and Transportation, South
China University of Technology, Guangzhou 510640, PR China.
E-mail address: qjchen@scut.edu.cn (Q.-J. Chen).
of construction. In this method, a steel jacket is added outside
the RC column, while the space between the tube and the original
column is filled with concrete for integration purpose. Several
experimental results on the steel jacketing approach with normal
infill concrete have been reported. Chai and Priestley [1–4] firstly
studied the influence of the circle and ellipse steel jacketing on
the lateral response of circular bridge columns, and found great

http://dx.doi.org/10.1016/j.conbuildmat.2017.03.013
0950-0618/Ó 2017 Published by Elsevier Ltd.
502 A. He et al. / Construction and Building Materials 141 (2017) 501–516
enhancement on the lateral strength, shear strength, resistance of
lap splice and ductility of the columns after being retrofitted by
steel jacket. Aboutaha [5] examined the effectiveness of the rectan-
gular solid steel jackets and partial steel jackets on improving the
ductility and strength of the columns that had inadequate shear
resistance. Xiao et al. [6,7] conducted a series of tests on the col-
umns with a square or rectangular section retrofitted by partially
stiffened steel jackets and indicated the effectiveness of the
method. Other types of steel jackets have also been proposed,
e.g. corrugated steel jacket [8] and octagonal steel jacket [9]. It
was shown that the strength, stiffness and deformation capacity
of the retrofitted column could be significantly improved due to
the confinement effects on the core concrete provided by the steel
jacket. The steel jacketing construction process is also quite effi-
cient since the steel tube could act as permanent formwork.
In the conventional approach of steel jacket retrofitting, normal
aggregate concrete is often used as padding to fill the gap between
the original column and the steel jacket. However, in recent years,
a new type of concrete named recycled aggregate concrete (RAC)
was introduced. In RAC, the natural coarse aggregates are partially
or entirely substituted by the recycled coarse aggregates. The
application of RAC is considered as alternative solution for process-
ing and recycling demolished concrete. Accordingly, this paper pre-
sents a structural column retrofitting technique by using RAC. The
proposed method is based on the steel jacketing approach, while
the RAC is adopted as an infill concrete to replace the normal con-
crete. Fig. 1 shows the typical cross-sections of the steel-jacket ret-
rofitted columns with RAC. This method not only can serve the
propose of the building or bridge retrofitting in order to avoid
reconstruction, but also take full advantage of using recycled
aggregated concrete, which therefore help the reduction of carbon
dioxide emission.
Unfortunately, RAC used to be described as a kind of green
materials but with low strength and elastic modulus, bad worka-
bility, high water infiltration and high shrinkage and creep com-
paring to normal concrete [10–13]. However, from the ongoing
research, it has been found out that confining RAC is an effective
way to enhance their mechanical properties not only because con-
finement increases the compressive strength of concrete and sup-
presses the crack development, but the core concrete is sealed
from moisture exchange with the surroundings and its lateral
deformation is significantly restricted by the steel tubes [14–16].
Chen et al. [16] present a summary review on the bond behaviour,
the static and cyclic performance of recycled aggregate concrete-
filled steel tubes members and frames. It was found the axial com-
pressive strength and deformation capacity of RAC are significantly
improved under the confinement of steel tube. Yang et al. [17,18]
conducted a series of tests on RAC filled steel tubes and RAC filled
stainless steel tubes. The test results showed that the RAC in-fill
columns had slightly lower but satisfactory ultimate strength com-
pared with the specimens filled with normal concrete. Xiao et al.
[19] undertook a series of tests on RAC that was confined by steel
tubes, which found out that both the strength and deformation
capacity of RAC are greatly increased. The results also indicated
that the peak load decreases with the increase of the recycled
aggregate replacement percentage. Wang et al. [20] investigated
the compressive behaviour of RAC filled steel tubes. It was men-
Fig. 1. Various sections of steel-jacket retrofitted column with RAC.
tioned that there was about 10% or less reduction on the compres-
sive strength of RAC filled steel tube columns than the normal
concrete filled steel tube columns due to the inferior performance
of recycled coarse aggregates, and clarified that the source of recy-
cled coarse aggregates had little effect. In summary, most of the
research conducted on RAC has mainly been concerned with
implementation of structural components. However, the applica-
tions of RAC on structural retrofitting have not received much
attention.
Although the cross-section of steel-jacket retrofitted column
with RAC is similar as that of RAC filled steel tube column, some
distinct characteristics of steel-jacket retrofitted column with
RAC need to be identified. Firstly, two types of concrete exist in
the steel tube of retrofitted column. The strength, elastic module
and shrinkage properties of the concrete in the original column
could be different from the infill RAC, which may lead to stress
redistribution on the core concrete. Secondly, there is discontinued
interface between the original column concrete and the infill RAC,
which may influence the axial compressive strength of the column.
Thirdly, during the retrofitting construction in practice, most of the
pre-existing RC columns are subjected to preloads arising from the
existing live and permanent loads from the upper floors or sub-
structure. The stress and deformation induced by the preloads
might inevitably affect the structural performance of the retro-
fitted columns [21], which was often overlooked in previous
research [1–9].
The purpose of this paper is to present a comprehensive study
on the axial compressive behaviour of steel-jacket retrofitted RC
columns with RAC. The effects of the steel tube thickness, recycled
coarse aggregate replacement ratio, RAC strength and axial
preloading level are considered. Tests on fifteen columns are
reported and the experimental results are discussed. A finite ele-
ment model verified by comparing the experimental results is also
implemented and extended for further investigation. A design
model for predicting the ultimate strength of the column is
proposed.
2. Experimental program
2.1. Test specimens
Fifteen columns including twelve steel-jacket retrofitted col-
umns (A1  A12), one unstrengthened RC column (A0), two con-
crete filled steel tube (CFST) columns (E1  E2) were fabricated
and tested. The CFST columns consisted of one normal concrete
filled steel tube column (E2) and one RAC filled steel tube column
(E1). The cross-sectional dimensions of the specimens are shown in
Fig. 2. The diameters D of steel jackets of all the specimens were
the same as 320 mm. Details of the specimens are shown in
Table 1.
The recycled coarse aggregates were produced from a concrete
recovery plant in Guangzhou with the aggregate size from
5  20 mm and the water absorption of 9.37%. The crush index of
the recycled coarse aggregates was 9.4%. The bulk density and
apparent density were 1245 kg/m3 and 2656 kg/m3 respectively.
For the infill concrete, the river sand, tap water and the ordinary
Portland cement with a 28 day compressive strength of 42.5 MPa
 A. He et al. / Construction and Building Materials 141 (2017) 501–516 503

Fig. 2. Details of specimens.

Table 1
Parameters and test results of specimens.

Specimen t (mm) l (mm) g (%) Ppre (kN) 0

fc1 (MPa) 0
fc2 (MPa) Ec (MPa) Ps (kN) Pu (kN) Ps/Pu Strength ratio eb
A0 – 1000 – – 34.2 – – – 1512 – 1.00 –
A1 3.84 1000 0 0 34.2 29.8 29029 4855 4941 0.98 3.27 0.0200
A2 3.84 1000 25 0 34.2 28.3 27226 4382 4547 0.96 3.01 0.0248
A3 3.84 1000 50 0 34.2 27.0 26605 4527 4538 1.00 3.00 0.0202
A4 3.84 1000 75 0 34.2 26.9 25367 4520 4793 0.94 3.17 0.0267
A5 3.84 1000 100 0 34.2 24.7 17214 4223 4661 0.91 3.08 0.0288
A6 3.84 1000 50 0 34.2 32.9 26619 4340 4659 0.93 3.08 0.0285
A7 3.84 1300 25 125 34.2 33.7 28561 – 4682 – 3.10 0.0100
A8 3.84 1300 25 316 34.2 33.7 28561 – 4868 – 3.22 0.0080
A9 1.81 1000 0 0 34.2 30.4 29029 3213 3309 0.97 2.19 0.0142
A10 5.84 1000 0 0 34.2 30.4 29029 – 6003 – 3.97 0.0320
A11 1.81 1000 50 0 34.2 28.2 26605 3220 3386 0.95 2.24 0.0108
A12 5.84 1000 50 0 34.2 28.2 26605 – 6021 – 3.98 0.0290
E1 3.84 1000 100 0 – 23.7 – 4207 4217 1.00 2.79 0.0152
E2 3.84 1000 0 0 – 43.6 – 5400 5522 0.98 3.65 0.0102

Note: t is the thickness of steel tube; l is the length of specimen; g is the recycled coarse aggregate replacement percentage; Ppre is the preload of original column; f0c1 , f0c2 are
the average compressive cubic strengths of the original column concrete and infill concrete respectively; Ec is the elastic modulus of infill concrete; Ps is the load when the
first shear slip lines appeared on the steel jacket; Pu is the experimental maximum load; strength ratio is the ratio of the value of Pu of each retrofitted column versus the value
of Pu of unstrengthen RC column; eb is the average longitudinal strain of specimen when local buckling appeared on the steel tube.

were used in this investigation. The recycled coarse aggregates
were not pre-wetting. However, during manufacturing the RAC,
additional water were added to compensate the high water
absorption capacity of recycled coarse aggregates according to
JGJ/T 240-2011 [22]. The mixture proportions are listed in Table 2.
The measured compressive strength and elastic modulus of con-
crete used in each specimen are showed in Table 1. The measured
steel properties of the specimens are listed in Table 3.
For the steel-jacket retrofitted specimens without preloads, the
original RC columns were firstly manufactured and then retrofitted
after 28 days of concrete curing period. For the specimens with
preloads, the post-tensioned prestressing construction technique
was implemented to simulate the preload actions on the speci-
mens, as shown in Fig. 3. The original column was firstly fabricated
with a pipe along the central axis of the section. After 28 days of
concrete curing period, a steel bar attached with strain gauges
was inserted into the pipe and the post-tensioning procedures
were carried out on the steel bar to impose compressive forces
on the column. The strains of the prestressed steel bar were mon-
itored throughout the prestressing procedure until their values
became stable. The preloads on specimens A7 and A8 could be cal-
culated as 125 kN and 316 kN respectively, based on the stable
value of strain of the steel bar. Subsequently, the prestressed col-
umn was set in the middle of a steel jacket with protective steel
tube locating at each end. The protective steel tube was used to
maintain the prestressed original column upright and steady dur-
ing the infill concrete pouring, and to prevent the accidental dam-
age of the prestressed tendon anchorage at both ends of columns
when vibrating the wet concrete. Finally, the gap between the orig-
inal column and the steel jacket was filled with RAC. For the con-
crete filled steel tubular columns E1 and E2, their reinforcement
cages which are the same as the one in the original column were
located in the central section.
2.2. Test setup
The experimental test set-up and instrumentation are shown in
Fig. 4. A 15 MN capacity compression testing machine was used for
all the tests. Before actual test, a small preload was applied slowly
on the specimen and removed gradually. This operation was

Table 2
Mixing proportions of concrete.

Specimen Design grade g (%) Water (kg) Cement (kg) Sand (kg) Natural coarse aggregates (kg) Recycled coarse aggregates (kg)
A1, A9, A10, E2 C30 0 205 422 633 1079 0
A2, A7, A8 C30 25 230 422 634 809 270
A3, A11, A12 C30 50 256 422 634 540 540
A4 C30 75 281 422 634 270 810
A5, E1 C30 100 306 422 634 0 1080
A6 C35 50 254 484 615 523 523
504 A. He et al. / Construction and Building Materials 141 (2017) 501–516

Table 3
Measured steel mechanical properties of reinforcing and steel tube.

Member Bar /6 Bar /12 Prestressed steel /32 Steel tube 1.81 mm Steel tube 3.84 mm Steel tube 5.84 mm
fy(MPa) 375 435 1080 432 400 356
fu(MPa) 593 654 – 570 550 509
Es(GPa) 182 150 170 221 184 173

Note: fy, fu, Es are the yield strength, the ultimate strength and the elastic modulus of the steel and reinforcement respectively.

Fig. 3. Execution of specimens with preloads.

Testing machine
Strain
gauges

LVDT LVDT
#1 #3

Testing machine

Fig. 4. Test set-up and instrumentation.

repeated in order to remove any slack between the specimen and
the load plate of testing machine and ensure the entire testing sys-
tem function properly. In the actual test, load control at a constant
speed of 4 kN/s was applied prior to the yielding of steel. When the
measured vertical strains of steel reached the yield strain, displace-
ment control loading at a constant rate of 0.01 mm/s was initiated.
The tests were terminated when the overlarge deformation and the
excessive local buckling of steel tube were observed on the
specimens.
Four linear voltage displacement transducers (LVDT) were
arranged at each side of the specimens to measure the longitudinal
deformations. The average longitudinal deformation was obtained
as the mean value of four LVDT readings. The average longitudinal
strain e of specimens was calculated through dividing the average
longitudinal deformation by the initial height of specimens.
3. Experimental results and discussions
3.1. Failure modes
For the unstrengthened specimen A0, cracks appeared and
extended rapidly at the peak load, following by the crash of the
protective layer of concrete. The typical failure mode is shown in
Fig. 5(a).
For the steel-jacket retrofitted columns and the CFST columns,
their failure modes were similar, as are shown in Fig. 5(b)–(f). At
the early stage of loading, the longitudinal deformation of steel
tube was small. The shear slip lines firstly appeared when the axial
compressive load approached closely to the maximum value,
which indicated that the steel tubes began to yield. Local buckling
of steel tubes was recognised after the peak load, followed by the
 A. He et al. / Construction and Building Materials 141 (2017) 501–516
Fig. 5. Experimental observations and failure modes.
wrinkles appearing at the mid-height of the steel tubes. The fea-
ture values, regarding the loads (Ps) when the first shear slip lines
appeared on the steel tube and the average longitudinal strains (eb)
when local buckling began to occur on the steel jacket were
recorded and shown in Table 1. It can be seen that the shear slip
lines were less likely to appear on the specimens with thicker steel
jackets. Take specimens A10 and A12 as examples, the steel tube
thickness is 5.84 mm and no obvious shear slip lines could be
observed on the steel tubes. With the decreasing of the steel tube
thickness, the values of eb reduce. It can also be found from speci-
mens A1–A5 that as the recycled coarse aggregate replacement
ratio increase, the eb might increase. For the preloaded specimens
A7 and A8, their shear slip lines were inconspicuous, but the spec-
imens were more susceptible to steel tube local buckling compared
with those without preloads. Moreover, the values of eb of CFST
columns were smaller than the counterpart of the steel-jacket ret-
rofitted columns given they having the same steel tube thickness.
Fig. 5(g) shows the sectional cut-view of the retrofitted column
specimens. The interface between the original column concrete
and the infill RAC can be evidently observed. The interface was
intact, which indicated good bond behaviour between the two lay-
ers of concrete.
3.2. Peak loads and load-deformation curves
Fig. 6 presents the applied load (P) versus average longitudinal
strain (e) curves for all the specimens. Table 1 lists the maximum
experimental loads Pu and the strength ratios of all the columns.
It can be seen from Fig. 6(a) that the steel-jacket retrofitted col-
umns have significant improvement on the strength, stiffness and
the deformation performance compared with the original RC col-
umn. After the peak strength was attained, the axial load on no-
retrofitted column descended dramatically while the retrofitted
columns exhibited ductile behaviour as the load remained constant
or decreased slightly. In this research, the sectional areas of the ret-
rofitted column are 2.00 times of the unstrengthen one, while the
range of strength ratio of the retrofitted column are 2.19–3.98. It
505
was demonstrated that there was remarkable enhancement on
the axial compressive strength of the column due to the confine-
ment of steel jacket. As shown in Table 1, by changing the recycled
coarse aggregate replacement ratio (g) from 0% to 100%, the
strength ratios of the retrofitted columns are 3.27, 3.01, 3.00,
3.17 and 3.08 respectively, which illustrates that the strengths of
the steel-jacket retrofitted columns with RAC are slightly lower
than the one with normal infill concrete. The differences of the ulti-
mate strength of the columns with various recycled coarse aggre-
gate replacement ratios are inconsiderable, except the one with
0% recycled coarse aggregate replacement ratio.
It can be observed from Fig. 6(b) that the increase of the infill
concrete strength could slightly increase the compressive strength
of the retrofitted columns as the strength ratio of specimens A3
and A6 are 3.00 and 3.08 respectively.
As shown in Fig. 6(c), specimen A12 exhibits relatively low duc-
tility because it failed due to the welding fracture. Except that, as
the steel jacket thickness increases, the retrofitted columns show
significant increase on the strength, stiffness and the ductility.
Fig. 6(d) presents the influence of the preload of original col-
umn. It can be seen that the initial stiffness of the preloaded and
non-preloaded specimens are almost identical. The peak load val-
ues of the preloaded and non-preloaded specimens are very close,
although the post-peak load behaviour of some specimens were
not recorded. It can be stated that the influence of preload on the
axial compressive strength of retrofitted column is negligible.
As shown in Fig. 6(e), because of having much higher compres-
sive strength of the core concrete, the specimen E2 exhibits higher
compressive strength and initial stiffness than all the steel-jacket
retrofitted columns and the specimen E1. In order to eliminate
the effect of various concrete strengths on the compressive beha-
viour of the columns, the non-dimensional load (P/Pu) versus aver-
age longitudinal strain (e) curves are adopted and shown in Fig. 7.
It can be seen that the non-dimensional skeleton curves between
the retrofitted columns and the CSFT columns are close to each
other, which indicates that the steel-jacket retrofitted columns
have similar axial compressive behaviour as the CFST columns.
506 A. He et al. / Construction and Building Materials 141 (2017) 501–516

Fig. 6. Load (P) versus average longitudinal strain (e) curves.

3.3. Strains distributions

Fig. 8 depicts the typical load and measured strain curve of the
steel tube, longitudinal bar and stirrup, in which el is the measured
longitudinal strain of the steel tube, eh is the measured hoop strain
of the steel tube, elb is the measured strain of the longitudinal bar,
est is the measured strain of the stirrup. It is shown that there was
an initial elastic response for all the components and the load-
strain response is closely linear. At this stage, the longitudinal
strains of the steel tube were larger than the hoop strains under
the same load. As the upper load increase to about 80% of the peak
load, the curves entered the inelastic range. The hoop strains of the
steel tube rose dramatically and even surpassed the longitudinal
strain, which demonstrated that the confinement effects from the
steel tube was activated. At the peak load, all the steels had
reached their yield strains.
Fig. 7. Non-dimensional load (P/Pu) versus average longitudinal strain (e) curves.
 A. He et al. / Construction and Building Materials 141 (2017) 501–516 507

qffiffiffiffi
0
crete is taken as 4730 f c according to ACI 318-11 [25], in which
f0c is the compressive cylinder strengths (in MPa) that is taken as
0.76 times of the compressive cubic strength. To describe the uni-
axial compressive response of the confined normal aggregate con-
crete and RAC, the stress-strain relationship recommended by
Yang et al. [26] is adopted herein. The equations of the stress-
strain (rc-ec) relationship are expressed as
(
2x  x2 ; ðx 6 1Þ
y¼ ð1Þ
x
bo ðx1Þ2 þx
; ðx > 1Þ

where

rc rc
y¼ ¼ ð2Þ
r0 a  f 0c
Fig. 8. Load (P) versus measured strain curves for steel tube and rebar of A3.
a ¼ 1  0:28g þ 0:08g2 ð3Þ

The hoop strains (eh) and the longitudinal strains (el) of the steel
ec ec
x¼ ¼   ð4Þ
jackets measured from different specimens in the test are shown in e0 b  1300 þ 12:5f 0c þ 800n0:2  106
Fig. 9. It can be observed that the hoop strains and longitudinal
strains of all specimens reached yielding strain limit almost at g
b¼1þ ð5Þ
the maximum loads, which demonstrated that the steel jacket 65:715g2  109:43g þ 48:989
could provide sufficient confinement on the core concrete. As and
shown in Fig. 9(a), (b) and (d), there are no significant differences
on the steel tube strains with the changes in the RAC strength,  ½0:25þðn0:5Þ7   0:5
0
recycled coarse aggregate replacement ratio or preload. As shown bo ¼ 2:36  105 fc  0:5 P 0:12 ð6Þ
in Fig. 9(c), with the increase of steel tube thickness, the strains
corresponding to the maximum load increase as well. In other where r0 and e0 are maximum compressive stress and the corre-
words, a thicker steel tube could provide more effective confine- sponding strain, respectively. The coefficient n representing the
ment on the core concrete. Comparison between the strains of confinement effect can be calculated by
steel-jacket retrofitted columns and CFST columns in Fig. 9(e)
As f sj
shows that their strains developments are similar, so would be n¼ 0 ð7Þ
their confining mechanisms of the steel tube on the core concrete. Ac f c

in which fsj is the tensile yield strength of the steel jacket, As and Ac
4. Finite element modelling are the cross-sectional areas of the steel tube and inner concrete
respectively.
4.1. General The uniaxial tensile response of concrete is assumed to be linear
until the tensile strength is attained. The tensile strength of con-
The software ABAQUS is used in the present study. The finite crete is taken as 0.1f0c . The tension stiffening effect of concrete is
element (FE) model of the retrofitted column is shown in Fig. 10. characterised by means of fracture energy (GF) proposed by MC
The end plates and concrete components are simulated by 3D 2010 [27] as follow:
eight-node solid reduced-integration element. Truss element
  f 0 0:7
T3D2 is used for the steel bars which are embedded in the con- 2
GF ¼ 0:0469dmax  0:5dmax þ 26 c
N=m ð8Þ
crete. The steel tube is simulated by four-node conventional shell 10
element. The end plate is assumed to be elastic rigid body. The
where dmax is the maximum size of coarse aggregate (in mm).
TIE contact is applied at the interfaces between the end plate and
The Poisson’s ratio of confined concrete is considered to be vari-
the main specimen. The surface-to-surface contact interaction with
able based on the change of the axial compressive stress. Instead of
finite sliding is applied at the interfaces between the steel tube and
using conventional Poisson’s ratio with constant value, the variable
the infill concrete, in which a hard contact property is adopted in
Poisson’s ratio is adopted herein to describe the relationship
the direction normal to the interface plane and the PENALTY option
between the transverse strain and the axial strain of the confined
is used for the tangential behaviour. The friction coefficient used in
concrete. The definition of variable Poisson’s ratio was introduced
the penalty frictional formulation is taken as 0.5 [23]. The same
firstly by Han et al. [28], as the ratio of transverse strain and axial
contact interaction is applied at the interfaces between the infill
strain of confined concrete is varied with the change of its axial
concrete and the original column concrete except that the friction
compressive stress. The equation for the variable Poisson’s ratio
coefficient is assumed to be 0.8 [24]. The bottom of the end plate is
of confined concrete is given by
restricted from moving in the x, y, z direction and the top of the 8  41
end plate is restricted from moving in the y, z direction. >
< 0:173;
rc
r0 6 0:55 þ 0:25 r041
lc ¼  3  41
>
: 0:173þ0:7036 rrc  0:4 2 r240 ; rrc > 0:55 þ 0:25 r041
4.2. Material properties 0 0

ð9Þ
Damage plasticity model is used to describe the constitutive
behaviour of the concrete. The elastic modulus of infill concrete Fig. 11 shows the curves of axial compressive stress rc verse
(Ec) is taken from Table 1, and the value for original column con- variable Poisson’s ratio lc for four types of confined concretes.
508 A. He et al. / Construction and Building Materials 141 (2017) 501–516

Fig. 9. Load (P) versus measured strain curves.

For modelling the steel material, the flow theory using von Es. The ultimate strain of steel is taken as 0.2. The equation of
Mises’ yield criterion is used. The constitutive relationships for variable Poisson’s ratio of steel is given by [28]:
the steel tube and reinforcing bars are represented by a bilinear
rs  f p
constitutive model. The strain-hardening ratios are taken as 0.01 ls ¼0:167  þ 0:283 ð10Þ
fy  fp
 A. He et al. / Construction and Building Materials 141 (2017) 501–516
Fig. 10. Finite element model of retrofitted column.
0.5
0.4
0.3
μc
0.2 Original concrete
Infill concrete (A1)
0.1 Infill concrete (A3)
Infill concrete (A5)
0.0
0 5 10 15 20 25 30 35 40
σc (MPa)
Fig. 11. Longitudinal stress rc verse Poisson’s ratio lc.
in which rs is the longitudinal stress of steel tube, fp is the elastic
limit strength of steel tube and is taken as 0.9fy herein.
4.3. Validation
In order to validate the accuracy of the FE model, the numerical
and experimental results of the load-average longitudinal strain
curves are compared as shown in Fig. 12. It can be seen that both
results have a good agreement. The mean values and standard
deviation of the ratio of the ultimate compressive strengths pre-
dicted by using the FE model and the experimental results are
1.005 and 0.035, respectively. It is demonstrated that the devel-
oped FE model is capable of predicting the axial compressive beha-
viour of the test specimens.
In order to illustrate the importance of considering variable
Poisson’s ratio in the FE model, the load-average longitudinal strain
curves predicted by the FE model using constant Poisson’s ratio are
also included in Fig. 12. The mean values and standard deviation of
the ratio of the ultimate compressive strengths predicted by using
the FE model using constant Poisson’s ratio and the experimental
results are 0.971 and 0.029, respectively. The FE model using con-
stant Poisson’s ratio might underestimate the ultimate strength of
the columns. Furthermore, Fig. 13 shows the strains of the steel
jackets obtained from the two FE models comparing with the
experimental results. The vertical axis of Fig. 13 represents the
ratio of the hoop strains eh and longitudinal strains el of the steel
jacket. It can be seen that the FE model using constant Poisson’s
509
ratio could underestimate the value of hoop strains. The curves
predicted by the FE model with variable Poisson’s ratio have better
agreement with the experimental results. It can be concluded that
the FE model using variable Poisson’s ratio could capture the stress
mechanism between the steel jacket and core concrete with better
accuracy.
5. Mechanical behaviour
For further investigation of the behaviour of the steel-jacket ret-
rofitted columns with RAC under axial compression, a steel-jacket
retrofitted column with RAC is designed. The mechanical beha-
viour of the column is further studied by using the developed
FEM model, and is discussed in details in this section. The param-
eters of the column are as follows: D = 320 mm, t = 3.84 mm,
l = 1000 mm, Ppre = 0kN, f0c1 = f0c2 = 32.4 MPa, g = 0%.
5.1. Analysis of complete load-deformation curves
The calculated axial load (P) versus axial strain (e) response of
the retrofitted column is shown in Fig. 14. The calculated axial load
(P) versus axial strain (e) curves of the components including the
original column concrete, infill concrete, steel jacket and longitudi-
nal rebar are also shown in Fig. 14. Four characteristic points can
be recognised on the curve, viz. Point A, the demarcation point
between elastic stage and elastic-plastic stage; Point B, the retro-
fitted column reaches the ultimate strength; Point C, the curve
starts to descend stably, and Point D, the calculation is terminated.
The distributions of the longitudinal stress and confining stress of
the concrete at four characteristic points are exhibited in Figs. 15
and 16, respectively. Fig. 17 shows the contact stress distributions
between the original column concrete and the infill concrete at the
characteristic points. Based on the four characteristic points, the
axial load versus axial strain curve could be divided into four
stages.
In the stage of OA, the retrofitted column shows the elastic
behaviour. The ratio of the hoop strain and longitudinal strain of
the steel jacket maintains being 0.283, which means that the steel
jacket produces little confining stresses onto the concrete (Figs. 16
(a) and 17(a)). Each component sustains the axial load only. The
longitudinal stress of the original column concrete and infill con-
crete is uniform distributed over the cross-section, as shown in
Fig. 15(a).
In the stage of AB, the performance of the retrofitted column
enters elastic-plastic range. The confining stress from the steel
jacket onto the core concrete and the contact stress between the
original column concrete and the infill concrete begin to be gener-
ated and increase rapidly (Figs. 16(b) and 17). In this stage, the
steel jacket reaches its ultimate strength, followed by the gradual
descending of the strength due to the outward buckling. The longi-
tudinal bars of the original column initiate the strain-hardening
performance after being through yielding. At point B, the retro-
fitted column attains its ultimate strength limit, while the
strengths of the original column concrete and infill concrete are
in the ascending phase but approach closely to their ultimate
strengths. It can be seen from Fig. 15(b) that the longitudinal stress
distributions of the concrete section are inhomogeneous at the
peak load. The maximum longitudinal stresses appear on the orig-
inal column concrete.
In the stage of BC, the strength of the retrofitted column
decreases, but the strengths of the infill concrete and original col-
umn concrete are still increasing slightly and reach their peak
strengths subsequently. Afterwards, the strengths of the infill con-
crete and original column concrete decrease slowly. It is shown
from Fig. 14 that the original column concrete sustains higher com-
510 A. He et al. / Construction and Building Materials 141 (2017) 501–516

5000 5000 5000

4000 4000 4000

3000 3000 3000

P(kN)
P(kN)

P(kN)
2000 2000 2000
Experiment Experiment Experiment
1000 ABAQUS-Variable μ 1000 ABAQUS-Variable μ 1000 ABAQUS-Variable μ
ABAQUS-Constant μ ABAQUS-Constant μ ABAQUS-Constant μ
0 0 0
0.00 0.01 0.02 0.03 0.00 0.01 0.02 0.03 0.00 0.01 0.02 0.03
ε ε ε

5000 5000 5000

4000 4000 4000

3000 3000 3000

P(kN)

P(kN)

P(kN)
2000 2000 2000
Experiment Experiment Experiment
1000 ABAQUS-Variable μ 1000 ABAQUS-Variable μ 1000 ABAQUS-Variable μ
ABAQUS-Constant μ ABAQUS-Constant μ ABAQUS-Constant μ
0 0 0
0.00 0.01 0.02 0.03 0.00 0.01 0.02 0.03 0.00 0.01 0.02 0.03
ε ε ε

4000
5000 5000

4000 4000 3000

3000 3000
P(kN)

P(kN)
P(kN)

2000
2000 2000
Experiment Experiment 1000 Experiment
1000 ABAQUS-Variable μ 1000 ABAQUS-Variable μ ABAQUS-Variable μ
ABAQUS-Constant μ ABAQUS-Constant μ ABAQUS-Constant μ
0 0 0
0.00 0.01 0.02 0.03 0.00 0.01 0.02 0.03 0.00 0.01 0.02 0.03
ε ε ε

4000
6000 6000

5000 3000 5000

4000 4000
P(kN)

P(kN)
P(kN)

3000 2000 3000

2000 2000
Experiment 1000 Experiment Experiment
1000 ABAQUS-Variable μ ABAQUS-Variable μ 1000 ABAQUS-Variable μ
ABAQUS-Constant μ ABAQUS-Constant μ ABAQUS-Constant μ
0 0 0
0.00 0.01 0.02 0.03 0.00 0.01 0.02 0.03 0.00 0.01 0.02 0.03
ε ε ε

Fig. 12. Comparisons of load-displacement skeleton curves from FE analysis and experiments.

pressive forces and shows better durability when compared with
the infill concrete. For the steel jacket, the compressive strength
plateaus after the descending period. As is shown in Figs. 16(c)
and 17(c), the confining stress and the contact stress reach their
peaks at Point C.
In the stage of CD, the overall strength of the retrofitted column
keeps stable, so are all the components. The residual strength of
the retrofitted column is about 92.4% of the ultimate strength.
5.2. Comparison of CFST column and steel-jacket retrofitted column
For comparison, a FE model of the CFST column containing the
same reinforcement cage as the aforementioned example retro-
fitted column is built. The geometry and material properties of
two types of columns are same, except that there is an interactive
interface exists between the original and infill concrete in the
retrofitted column. The axial load-strain relationships of the CFST
 A. He et al. / Construction and Building Materials 141 (2017) 501–516 511

0.000 0.004 0.008 0.012 0.016

0.0 0.000 0.004 0.008 0.012 0.016 0.020
ε 0.0
ε
-0.2 TEST
-0.2 TEST
ABAQUS with variable μ
ABAQUS with variable μ
-0.4 ABAQUS with constant μ -0.4 ABAQUS with constant μ
εh/εl

εh/εl
-0.6 -0.6

-0.8 -0.8

-1.0
-1.0

0.000 0.004 0.008 0.012 0.016

0.0
ε
-0.2 TEST
ABAQUS with variable μ
-0.4 ABAQUS with constant μ
εh/εl

-0.6

-0.8

-1.0

Fig. 13. Effect of Poisson’s ratio.

B
5000 C D

4000
Whole section
Original column concrete
P(kN)

3000 Infill concrete

Steel jacket
2000 A Steel bars

1000

0
0.000 0.005 0.010 0.015 0.020 0.025 0.030
ε
Fig. 14. Typical axial load (P) versus axial strain (e) response.

column and retrofitted column are shown in Fig. 18. It seems that
two curves nearly overlap, which indicated that the influence of
interface on the axial load-longitudinal strains curves is marginal.
5.3. Effect of RAC
In order to illustrate the effects of RAC on the axial compressive
behaviour of retrofitted column, a group of columns with different
recycled coarse aggregate replacement percentages (g) of infill
concrete ranged from 0% to 100% are analysed. The axial load-
strain curves of the columns and their components are shown in
Fig. 19, in which the symbol ‘‘s” represents the peak load of each
component. It can be seen that as the g increases, the ultimate
strength of retrofitted column decreases and the corresponding
strain increases. This can be attributed to the degradation of the
strength and elasticity modulus of the RAC compared with the nor-
mal concrete. On the other hand, the change in the g do not have
significant influence on the behaviour of the original column con-
crete, steel jacket and steel bars.
It can also be noted from the figure that each component
reaches its peak strength inconsistently. Nevertheless, the ultimate
strengths of the original concrete and infill concrete are similar to
the strengths when the retrofitted columns reach their peak loads.
5.4. Effect of preload
The effects of preload on the axial load (P) versus strain (e)
responses of the retrofitted columns are shown in Fig. 20. It exhi-
bits that the column with higher preload reaches its ultimate
strength earlier. For the columns with preload being 0, 312 kN
and 700 kN, the strains corresponding to the peak strengths are
0.0055, 0.0053 and 0.0047, respectively. The existing initial strain
512 A. He et al. / Construction and Building Materials 141 (2017) 501–516

Fig. 15. Distributions of longitudinal stress of concrete (unit: MPa).

Fig. 16. Distributions of confining stress of concrete in Y direction (unit: MPa).

 A. He et al. / Construction and Building Materials 141 (2017) 501–516 513

Point A of original column concrete due to the preload could be the main
7
Point B factor affecting the strains at the peak load. During the loading pro-
6 Point C cess of the retrofitted columns, the concrete of the original column
Contact stress (MPa)

5 Point D with preload shows higher stress levels than the one without pre-
4 load. However, as is observed in Fig. 20, the ultimate strengths of
3 all the columns are approximately the same. It can be said that
the effect of preloads on the ultimate compressive strength of
2
the retrofitted column is negligible.
1
0 5.5. Effect of steel jacket
0 50 100 150 200 Y
Section width (mm) Fig. 21 displays the effect of steel jacket thickness on the axial
load (P) versus strain (e) response. It can be seen that the change
Fig. 17. Distributions of contact stress between original and infill concrete. in the thickness of steel jacket has significant impact on the perfor-
mance of the steel-jacket retrofitted column. With the increase of
the steel jacket thickness, the ultimate strength of the retrofitted
5000 column is increased dramatically. When the thickness of steel
jacket increases from 1.81 mm to 3.84 mm and to 5.84 mm, the
4000 ultimate strengths of the retrofitted column rise by 40.5% and
87.7%, respectively. It is because increasing the thickness of steel
3000 jacket would not only enhance the compressive strength of the
P (kN)

steel jacket itself, but also could enforce the steel jacket to provide
2000
more confining pressure onto the core concrete, which promi-
nently improves the compressive strengths of all the concrete com-
ponents. Therefore, the thickness of steel jacket is the critical factor
1000
CFST column that influences the performance of the steel-jacket retrofitted
Steel jacket retrofitted column column.
0
0.000 0.005 0.010 0.015 0.020 0.025 0.030
ε
6. Proposed design method
Fig. 18. Axial load (P) versus strain (e) responses between CFST column and
retrofitted column. Based on the observations from the experimental and numeri-
cal studies, the following assumptions are made:

5000
Pressure of original concrete(η =0%)
4000 Pressure of infill concrete (η =0%)
Pressure of total section (η =0%)
P(kN)

3000 Pressure of original concrete(η =50%)

Pressure of infill concrete (η =50%)
2000 Pressure of total section (η =50%)
Pressure of original concrete(η =100%)
1000 Pressure of infill concrete (η =100%)
Pressure of total section (η =100%)
0
0.000 0.002 0.004 0.006 0.008 0.010
ε
(a) Original column concrete and infill concrete

5000
Pressure of steel jacket (η =0%)
4000 Pressure of bars (η =0%)
Pressure of total section (η =0%)
P(kN)

3000 Pressure of steel jacket (η =50%)

Pressure of bars (η =50%)
2000 Pressure of total section (η =50%)
Pressure of steel jacket (η =100%)
1000 Pressure of bars (η =100%)
Pressure of total section (η =100%)
0
0.000 0.002 0.004 0.006 0.008 0.010
ε
(b) Steel jacket and bars
Fig. 19. Axial load (P) versus strain (e) responses with various g.
514 A. He et al. / Construction and Building Materials 141 (2017) 501–516

5000
Pressure of original concrete(Ppre=0kN)
Pressure of infill concrete (Ppre=0kN)
4000
Pressure of total section (Ppre=0kN)
3000 Pressure of original concrete(Ppre=312kN)
P(kN) Pressure of infill concrete (Ppre=312kN)
2000 Pressure of total section (Ppre=312kN)
Pressure of original concrete(Ppre=700kN)
1000 Pressure of infill concrete (Ppre=700kN)
Pressure of total section (Ppre=700kN)
0
0.000 0.002 0.004 0.006 0.008 0.010
ε
(a) Original column concrete and infill concrete

5000
Pressure of steel jacket (Ppre=0kN)
4000 Pressure of bars (Ppre=0kN)
Pressure of total section (Ppre=0kN)
3000 Pressure of steel jacket (Ppre=312kN)
P(kN)

Pressure of bars (Ppre=312kN)

2000 Pressure of total section (Ppre=312kN)
Pressure of steel jacket (Ppre=700kN)
1000 Pressure of bars (Ppre=700kN)
Pressure of total section (Ppre=700kN)
0
0.000 0.002 0.004 0.006 0.008 0.010
ε
(b) Steel jacket and bars
Fig. 20. Axial load (P) versus strain (e) response with various preloads.

(1) Plane section remains plane under deformation;
(2) The strengths of the original column concrete and infill con-
crete are approximate to their ultimate strengths when the
retrofitted column reaches its ultimate compressive
strength;
(3) Perfect bond exists between the concrete and the steel tube;
(4) The influence of infill RAC on the ultimate strengths of orig-
inal column concrete and steel jacket is ignored;
(5) The preload effect on the ultimate strength of the retrofitted
column is ignored.
The design equation for predicting the axial compressive
strength of steel-jacket retrofitted column with RAC is proposed
as:
N ¼ f cc1 Acc1 þ f cc2 Acc2 þ rsj Asj þ f sb Asb
0 0
where f0cc1 is the effective compressive strengths of original column
concrete material, Acc1 the sectional area of original column con-
crete, f0cc2 the effective compressive strengths of infill concrete
material, Acc2 the sectional area of infill concrete, rsj the effective
compressive strength of the steel jacket, Asj the sectional area of
steel jacket, fsb is the yield strength of the steel bar, Asb the total sec-
tional area of steel bars.
Following the method suggested by Tang et al. [29], f0cc1 and f0cc2
can be calculated by:
0 0
f cc1 ¼ f co1 þ m  f l1
0 0
f cc2 ¼ f co2 þ m  f l2
in which f0co1 and f0co2 are 0.76 times of the compressive cubic
strengths of original concrete material and infill concrete material,
respectively. m is an empirical coefficient that is assumed to be
4.0 [30]. fl1 and fl2 are the confining stress of original concrete mate-
rial and infill concrete material, respectively, which are given by
2tfsj
f l1 ¼ c ð14Þ
D  2t
 
D
f l2 ¼ 0:0009 þ 0:81  ð0:09g þ 0:9983Þ  f l1 ð15Þ
t
where c is an empirical coefficient of the confinement from the steel
tube and is taken as 0.6 through the regression of the experimental
results. fsj is the tensile yield stress of steel jacket.
The outward buckling of steel jacket under axial compression is
also considered in the formulation. By regression, the effective
axial compressive strength of the steel jacket rsj can be given by:
 
ð11Þ
D
rsj ¼ 8:455  104  þ 0:6148  f sj ð16Þ
t
The maximum experimental loads Pu as well as other literature
results [21] are compared with the results Pm predicted by the pro-
posed design model, as listed in Table 4. The experimental results
were also compared with the results calculated by the methods in
current design codes CECS 28-2012 [31], ANSI/AISC-360-2005 [32],
EC4 [33] and AIJ-CFT [34]. It can be summarised from Table 4 that
the proposed design method provides the most accurate prediction
of the experimental results. The mean value of Pm/Pu is 1.00 with
the coefficient of variation (COV) of 0.08. Among all the approaches
ð12Þ
recommended in current design codes, the method in EC4 (PEC4)
ð13Þ gives the most satisfactory estimation. The mean value and COV
of PEC4/Pu are 1.00 and 0.09 respectively. It can be concluded that
the design method proposed in this paper as well as the one rec-
ommended in EC4 could be used to accurately predict the ultimate
strength of the steel-jacket retrofitted columns with RAC.
 A. He et al. / Construction and Building Materials 141 (2017) 501–516 515

7000

6000
Pressure of original concrete(t=1.81mm)
5000
Pressure of infill concrete (t=1.81mm)
4000 Pressure of total section (t=1.81mm)
P(kN)

Pressure of original concrete(t=3.84mm)

3000 Pressure of infill concrete (t=3.84mm)
Pressure of total section (t=3.84mm)
2000 Pressure of original concrete(t=5.84mm)
Pressure of infill concrete (t=5.84mm)
1000 Pressure of total section (t=5.84mm)
0
0.000 0.005 0.010 0.015 0.020 0.025 0.030
ε
(a) Original column concrete and infill concrete

7000

6000
Pressure of steel jacket (t=1.81mm)
5000
Pressure of bars (t=1.81mm)
4000 Pressure of total section (t=1.81mm)
P(kN)

Pressure of steel jacket (t=3.84mm)

3000 Pressure of bars (t=3.84mm)
Pressure of total section (t=3.84mm)
2000 Pressure of steel jacket (t=5.84mm)
Pressure of bars (t=5.84mm)
1000 Pressure of total section (t=5.84mm)
0
0.000 0.005 0.010 0.015 0.020 0.025 0.030
ε
(b) Steel jacket and bars
Fig. 21. Axial load (P) versus strain (e) response with various steel jacket thickness.

Table 4
Comparisons of test results and results predicted by design codes and proposed model.

Source Specimen labels Test result CECS28-2012 ANSI/AISC EC4 AIJ-CFT Proposed model
Pu (kN) PCECS (kN) PCECS/Pu PANSI (kN) PANSI/Pu PEC4 (kN) PEC4/Pu PAIJ (kN) PAIJ/Pu Pm (kN) Pm/Pu
This paper A1 4941 5446 1.10 3601 0.73 4651 0.94 3994 0.81 4687 0.95
A2 4623 5528 1.20 3606 0.78 4679 1.01 4000 0.87 4660 1.03
A3 4537 5369 1.18 3540 0.78 4594 1.01 3962 0.87 4638 1.02
A4 4793 5279 1.10 3463 0.72 4479 0.93 3837 0.80 4650 0.97
A5 4738 5064 1.07 3449 0.73 4485 0.95 3971 0.84 4603 0.99
A6 4660 5513 1.18 3662 0.79 4705 1.01 4024 0.86 4802 1.03
A7 4682 5546 1.19 3689 0.79 4736 1.01 4050 0.87 4810 1.03
A8 4868 5472 1.12 3652 0.75 4681 0.96 4005 0.82 4810 0.99
A9 3310 3905 1.18 2826 0.85 3490 1.05 2986 0.90 3497 1.06
A10 6003 6534 1.09 4138 0.69 5496 0.92 4731 0.79 5538 0.92
A11 3387 3872 1.14 2780 0.82 3445 1.02 2949 0.87 3446 1.02
A12 6021 6600 1.10 4141 0.69 5529 0.92 4758 0.79 5518 0.92
[21] BZ1 2202 2945 1.34 2500 1.14 2951 1.34 2332 1.06 2852 1.30
BZ2 2990 3553 1.19 2804 0.94 3409 1.14 2780 0.93 3332 1.11
BZ3 3820 4007 1.05 3032 0.79 3761 0.99 3055 0.80 3674 0.96
BZ4 4180 4477 1.07 3266 0.78 4114 0.98 3328 0.80 4048 0.97
BZ5 4460 4684 1.05 3370 0.76 4277 0.96 3460 0.78 4199 0.94
CZ1 2780 2978 1.07 2324 0.84 2847 1.02 2257 0.81 2782 1.00
CZ2 3678 4057 1.10 2862 0.78 3672 1.00 3005 0.82 3637 0.99
CZ3 3030 2978 0.98 2324 0.77 2847 0.94 2257 0.75 2782 0.92
DZ1 4290 4521 1.05 3235 0.75 4121 0.96 3334 0.78 4044 0.94
DZ2 4230 4521 1.07 3235 0.77 4121 0.97 3334 0.79 4044 0.96
Mean value 1.12 0.79 1.00 0.84 1.00
COV 0.08 0.09 0.09 0.07 0.08
516 A. He et al. / Construction and Building Materials 141 (2017) 501–516
7. Conclusions
Based on the experimental and numerical studies on the axial
compressive behaviour of the steel-jacket retrofitted RC columns
with RAC, the main findings can be summarized in the following.
The experimental results manifested that the strength, stiffness
and ductility of the column specimens would be greatly enhanced
through the steel jacketing retrofitting approach with RAC,
although the strengths of the steel-jacket retrofitted columns with
RAC could be slightly lower than the one with normal infill con-
crete. It was also shown that the influence of recycled coarse aggre-
gate replacement ratios as well as the preload of original column
on the compressive strength might be negligible. The peak
strengths of the steel-jacket retrofitted columns were significantly
increased with the increase of steel tube thickness. The perfor-
mances of the steel-jacket retrofitted columns were similar to that
of the CFST columns.
The finite element analyses using the commercial finite element
program ABAQUS were performed to investigate the performance
of the retrofitted columns. The proposed FE model considering
the variable Poisson’s ratio of core concrete and steel jacket during
loading was validated through comparing its results with the test
results. By using the developed FE model, the mechanical beha-
viours of the steel-jacket retrofitted column with RAC were further
studied. It was observed that the longitudinal stress of the concrete
was inhomogeneous distributed over the cross-section of the col-
umn. The original column concrete sustained higher compressive
forces and showed better durability compared with the infill con-
crete. Although each component of the column reached its peak
strength inconsistently, the strengths of the original column con-
crete and the infill concrete was approximate to their correspond-
ing ultimate strengths when the retrofitted column reached its
peak load. Through comparing the axial compressive behaviours
between CFST column and steel jacket retrofitted column, it is indi-
cated that the influence of the interface between the original and
infill concrete was inconsiderable. Extended parameter studies
were also undertaken and the results showed that the thickness
of steel jacket is the dominant factors affecting the performance
of the steel-jacket retrofitted column.
A design model for predicting the axial compressive strength of
the retrofitted column with RAC was proposed. The results calcu-
lated by the proposed design model showed good agreement with
the experimental results. The design methods proposed in this
paper as well as the one suggested in EC4 were recommended
for predicting the ultimate strength of the steel-jacket retrofitted
columns with RAC.
Acknowledgement
This study was funded by the National Natural Science Founda-
tion of China (51578246), Science and Technology Program of
Guangzhou, China (201604010053), the State Key Lab of Subtrop-
ical Building Science, South China University of Technology
(2015ZC18), Natural Science Foundation of Guangdong Province,
China (2016A030310418), the Fundamental Research Funds for
the Central Universities (2014ZZ0026).
References
[1] Y.H. Chai, M.J.N. Priestley, F. Seible, Seismic retrofit of circular bridge columns
for enhanced flexural performance, ACI Struct. J. 88 (5) (1991) 572–584.
[2] Y.H. Chai, An analysis of the seismic characteristics of steel-jacketed circular
bridge columns, Earthquake Eng. Struct. Dynam. 25 (2) (1996) 149–161.
[3] M.J.N. Priestley, F. Seible, Y. Xiao, R. Verma, Steel jacket retrofitting of
reinforced concrete bridge columns for enhanced shear strength – Part 1:
theoretical considerations and test design, ACI Struct. J. 91 (4) (1994) 394–405.
[4] M.J.N. Priestley, F. Seible, Y. Xiao, R. Verma, Steel jacket retrofitting of
reinforced concrete bridge columns for enhanced shear strength – Part 2:
Test results and comparison with theory, ACI Struct. J. 91 (5) (1994) 537–551.
[5] R.S. Aboutaha, M.D. Engelhardt, J.O. Jirsa, M.E. Kreger, Rehabilitation of shear
critical concrete columns by use of rectangular steel jackets, ACI Struct. J. 96
(1) (1999) 68–78.
[6] Y. Xiao, R. Ma, Seismic retrofit of RC circular columns using prefabricated
composite jacketing, J. Struct. Eng. 123 (10) (1997) 1357–1364.
[7] Y. Xiao, H. Wu, Retrofit of reinforced concrete columns using partially stiffened
steel jackets, J. Struct. Eng. 129 (6) (2003) 725–732.
[8] A. Ghobarah, A. Biddah, M. Mahgoub, Rehabilitation of reinforced concrete
columns using corrugated steel jacketing, J. Earthquake Eng. 1 (4) (1997) 651–
673.
[9] M. Lin, P. Chen, K. Tsai, Y. Yu, J. Liu, Seismic steel jacketing of rectangular RC
bridge columns for the mitigation of lap-splice failures, Earthquake Eng. Struct.
Dynam. 39 (2010) 1687–1710.
[10] A.D. Buck, Recycled aggregate concrete as a source of aggregate, ACI Struct. J.
74 (5) (1977) 212–219.
[11] T.C. Hansen, E. Boegh, Elasticity and drying shrinkage of recycled-aggregate,
ACI J. Proc. 82 (5) (1985) 648–652.
[12] Y.H. Loo, C.T. Tam, R.S. Ravindrarajah, Recycled concrete as fine and coarse
aggregate in concrete, Mag. Concr. Res. 39 (141) (1987) 214–220.
[13] A. Ajdukiewicz, A. Kliszczewicz, Influences of recycled aggregates on
mechanical properties of HS/HPC, Cement Concr. Compos. 24 (2) (2002)
269–279.
[14] M. Saatcioglu, S.R. Razvi, Strength and ductility of confined concrete, J. Struct.
Eng. 118 (6) (1992) 1590–1607.
[15] Y. Xiao, W. He, K. Choi, Confined concrete-filled tubular columns, J. Struct. Eng.
131 (3) (2005) 488–497.
[16] Z. Chen, J. Xu, Y. Chen, E.M. Lui, Recycling and reuse of construction and
demolition waste in concrete-filled steel tubes: a review, Constr. Build. Mater.
126 (2016) 641–660.
[17] Y.F. Yang, L.H. Han, Experimental behaviour of recycled aggregate concrete
filled steel tubular columns, J. Constr. Steel Res. 62 (12) (2006) 1310–1324.
[18] Y.F. Yang, G.L. Ma, Experimental behaviour of recycled aggregate concrete
filled stainless steel tube stub columns and beams, Thin-Walled Struct. 66
(2013) 62–75.
[19] J. Xiao, Y. Huang, J. Yang, C. Zhang, Mechanical properties of confined recycled
aggregate concrete under axial compression, Constr. Build. Mater. 26 (1)
(2012) 591–603.
[20] Y. Wang, J. Chen, Y. Geng, Testing and analysis of axially loaded normal-
strength recycled aggregate concrete filled steel tubular stub columns, Eng.
Struct. 86 (2015) 192–212.
[21] A. He, J. Cai, Q. Chen, X. Liu, J. Xu, Behaviour of steel-jacket retrofitted RC
columns with preload effects, Thin-Walled Struct. 109 (2016) 25–39.
[22] JGJ/T 240-2011, Technical Specification for Application of Recycled Concrete,
China Architecture & Building Press, Beijing, 2011.
[23] F.X. Ding, Z. Li, S. Cheng, Z.W. Yu, Composite action of octagonal concrete-filled
steel tubular stub columns under axial loading, Thin-Walled Struct. 107 (2016)
453–461.
[24] A.H. Mattock, Shear transfer in concrete having reinforcement at an angle to
the shear plane, ACI Spec. Publ. 42 (1974) 17–42.
[25] ACI318-11, Building Code Requirements for Structural Concrete and
Commentary, American Concrete Institute, Detroit (USA), 2011.
[26] Y.F. Yang, Z.C. Zhang, F. Fu, Experimental and numerical study on square
RACFST members under lateral impact loading, J. Constr. Steel Res. 111 (2015)
43–56.
[27] CEB-FIP MODEL CODE 2010 Volume 1. Switzerland: International Federation
for Structural Concrete (fib), 2012.
[28] L.H. Han, Concrete Filled Steel Tubular Structure, Science Press, Beijing, 2007.
[29] J. Tang, S. Hino, I. Kuroda, T. Ohta, Modeling of stress-strain relationships for
steel and concrete in concrete filled circular steel tubular columns, Steel
Constr. Eng. 3 (11) (1996) 35–46.
[30] F. Aslani, B. Uy, J. Hura, P. Carinoa, Behaviour and design of hollow and
concrete-filled spiral welded steel tube columns subjected to axial
compression, J. Constr. Steel Res. 128 (2017) 261–288.
[31] CECS28:2012, Technical Specification for Concrete-filled Steel Tubular
Structures, China Planning Press, Beijing, 2012.
[32] ANSI/AISC 360-05, Specification for Structural Steel Buildings, American
Institute of Steel Construction, Chicago, 2005.
[33] Eurocode 4 (EC4). Design of steel and concrete structures. Part1 1: General
rules and rules for building. Brussels: European Committee for
Standardization, 2004.
[34] Architectural Institute of Japan (AIJ), Recommendations for Design and
Construction of Concrete Filled Steel Tubular Structures, Architectural
Institute of Japan, Tokyo, 1997.