Construction and Building Materials 158 (2018) 624–639

Contents lists available at ScienceDirect

Construction and Building Materials

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

Seismic behaviour of steel-jacket retrofitted reinforced concrete columns

with recycled aggregate concrete
An He a, Jian Cai a,b, Qing-Jun Chen a,b,⇑, Xinpei Liu c, Peizhou Huang a, Xu-Lin Tang a,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
Guangzhou Jishi Construction Group Co. Ltd, Guangzhou 510115, PR China

h i g h l i g h t s

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

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

The parametric study on the retrofitted columns was conducted.

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

Article history: This paper presents the experimental and numerical investigations on the seismic behaviour of steel-
Received 21 June 2017 jacket retrofitted reinforced concrete (SJRRC) columns with recycled aggregate concrete (RAC). An
Received in revised form 9 October 2017 unstrengthened reinforced concrete (RC) column and nine SJRRC columns tested under lateral cyclic load-
Accepted 10 October 2017
ing are reported. The experimental results manifest that by using the steel-jacket retrofitting approach
Available online 17 October 2017
with RAC, the initial stiffness, ultimate strength, deformation ductility and energy dissipation ability of
the columns are improved significantly. The peak strengths of the SJRRC columns are about 1.86–3.44
Keywords:
times of the counterpart of the original RC column. The retrofitted columns also show ductile post-
Steel jacket
Retrofit
peak load behaviour with the ductility coefficients ranging between 4.05 and 7.93. As the applied axial
Recycled aggregate concrete compressive loads increase, the failure mode of the SJRRC specimens is transited gradually from
Cyclic behaviour tension-controlled failure to compression-controlled failure. The specimens failed in compression-
Finite element model controlled mode exhibit plumper hysteresis curves, better energy dissipation ability, and higher secant
stiffness than those failed in tension-controlled mode. The specimen having 100% recycled coarse aggre-
gate replacement ratio has slightly lower lateral strength and secant stiffness than the specimens with 0%
or 50% recycled coarse aggregate replacement ratio, and shows more serious pinching effect on its hys-
teresis curve. The effects of the preload and pre-damage of original column could be unfavourable on the
cyclic performance of retrofitted column. The finite element analyses are also performed to further inves-
tigate the lateral behaviour of SJRRC columns. The proposed finite element model is validated by a com-
parison with the experimental results. By using the developed finite element model, parameter studies
are undertaken and indicate that the applied axial compressive load and the thickness of steel jacket
are the dominant factors affecting the lateral performance of SJRRC column.
Ó 2017 Published by Elsevier Ltd.

1. Introduction be damaged severely under earthquake loading due to the shortage

The reinforced concrete (RC) columns are substantial structural
components in contemporary building infrastructures. However,
most exiting RC columns built in 90’s or even earlier in China could
⇑ 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).
and incompetency of the previous RC structures design code,
which has been confirmed unfortunately during the 2008 Sichuan
earthquake [1]. In order to resolve this crucial problem, developing
advanced technologies for RC columns retrofitting are much
needed nowadays. Among various developed RC columns retrofit-
ting solutions viz. carbon fibre-reinforced polymer composite
wrapping [2], concrete jacketing and steel jacketing [3,4], the steel
jacketing method is usually the preferred one because of its high

https://doi.org/10.1016/j.conbuildmat.2017.10.053
0950-0618/Ó 2017 Published by Elsevier Ltd.
 A. He et al. / Construction and Building Materials 158 (2018) 624–639
retrofitting effectiveness and economic efficiency [5]. In this
method, a common RC column is encased by a thin steel jacket.
The gap between the steel tube and the original column is filled
with the concrete or grout for integration purpose. The steel jacket
plays a role as constraint to the core concrete. Several experimental
studies have been reported showing the effectiveness of steel jack-
eting approach on improving the seismic performance of RC col-
umns. Chai and Priestley [6–9] firstly applied the steel-jacket
retrofitting method for bridge columns in California, and found
great improvement on the lateral strength, shear strength, the
resistance of lap splice and the ductility of the columns after being
retrofitted by circle and ellipse steel jacket. The effectiveness of
steel jacket retrofitted bridge columns was also tested under
shake-table excitation [10]. Aboutaha [11] investigated the effec-
tiveness of the rectangular solid steel jackets and partial steel jack-
ets. Xiao [12,13] described the influence of partially stiffened steel
jacket and prefabricated composite jacket on enhancing the
strength and ductility of the RC column. Some other types of steel
jacket sections such as corrugated steel jacket [14] and octagonal
steel jacket [15] have also been proposed and evaluated. Recently,
Choi [16,17] introduced a new steel-jacketing method by applying
lateral pressure to attach steel jacket on the surface of original col-
umns. The retrofitted columns were tested under a combination of
axial and lateral loadings and the applicability of the proposed
method were assessed. It could be summarised that the seismic
performance of the steel-jacket retrofitted columns were enhanced
significantly compared with that of the original column.
In general, normal fine aggregate concrete or grout is adopted as
padding filling the gap between the original column and steel
jacket for integrity propose. In recent years, the recycled aggregate
concrete (RAC) in which the natural coarse aggregates are partially
or entirely substituted by the recycled coarse aggregates has
received increasing attention in academic research and industrial
application, and has been extensively studied. Accordingly, a solu-
tion of steel-jacket retrofitting with RAC was proposed [18], in
which the RAC replaces the normal fine aggregate concrete or
grout as the padding material. The typical cross-sections of the
steel-jacket retrofitted columns with RAC are shown in Fig. 1.
The application of this method not only can serve the principal
purpose for infrastructure retrofitting to avoid reconstruction,
but also takes full advantage of using RAC which can reduce carbon
dioxide emission. Experimental studies have illustrated that using
the steel-jacket retrofitting approach with RAC could improve sig-
nificantly the strength, stiffness and ductility of the RC column
under axial compression. However, the seismic performance of
steel-jacket retrofitted column with RAC has not been studied yet.
Many researches have been undertaken on the seismic beha-
viour of RAC structural members. It was illustrated that the RAC
might be unfavourable to be used for the structural members
under seismic loading compared with the use of normal aggregate
concrete. Yang [19] studied the cyclic performances of RAC-filled
steel tubular columns and found that both the load carrying capac-
ities and sectional flexural stiffness of the RAC-filled steel tubular
columns were quite comparable to those of the normal concrete
filled steel tubular columns. Wu [20,21] proposed the use of
demolished concrete with distinctly larger coarse aggregate sizes
and investigated the cyclic behaviour of the thin-walled steel tubu-
Fig. 1. Various cross-sections of steel-jacket retrofitted columns with RAC.
625
lar columns filled with demolished concrete lumps or demolished
concrete segments. They mentioned the minor reduction on the
lateral strength of the column compared with the counterpart of
the normal concrete filled steel tube column. Xiao [22] performed
experimental tests on a cast-in situ normal aggregate concrete col-
umn and a cast-in situ RAC column as well as four semi-precast
RAC columns under cyclic horizontal loading. It was concluded that
the semi-precast RAC columns had similar seismic behaviour as
the fully cast-in situ columns. Ma [23,24] processed the tests on
steel reinforced RAC column under cyclic loading. The test results
showed that the seismic performance of the columns decreased
slightly as the RCA replacement percentage increases. Shi [25,26]
firstly proposed a new construction material called geopolymeric
recycled concrete (GRC) and tested the structural properties of
GRC filled steel tube (GRCFST). It was found that with the increas-
ing of the recycled aggregate replacement ratio, the ultimate
strength of GRCFST was reduced while the peak strain and the duc-
tility of the columns were improved. Tang [27] conducted low cyc-
lic loading tests on nine seamless steel tube columns filled with
RAC and found the RAC-filled steel tube columns had appreciably
better lateral bearing capacity and ductility, but slightly lower
energy dissipation ability, than the normal concrete filled steel
tube columns.
The object of the current paper is to present a comprehensive
study on the cyclic behaviour of steel-jacket retrofitted reinforced
concrete (SJRRC) columns with RAC. The effects of the steel tube
thickness, recycled coarse aggregate replacement ratios, axial com-
pressive load, the preload and pre-damage conditions are consid-
ered in this paper. Physical tests on ten columns are reported
and the experimental results are discussed. A finite element model
verified by comparing its results against the experimental results is
also implemented and extended for further investigation.
2. Experimental program
2.1. Test specimens
Nine reinforced concrete (RC) columns with strong base ends
were fabricated. The cross section of the RC columns is 200 mm
 200 mm, as Fig. 2 shows. The SJRRC columns were fabricated
based on the RC columns. The diameters of steel jackets D and
the effective heights of all SJRRC columns were the same as 320
mm and 1320 mm respectively. The variables among the SJRRC
columns included the thickness of steel tube t, the replacement
percentage of recycled coarse aggregate g, the axial compressive
load applied on the column N0, the preload on original RC column
Npre and the damage condition of original RC column. Details of the
specimens are shown in Table 1.
Among the specimens, the unstrengthened RC column C0 was
chosen as the control sample. Specimen C8 is the SJRRC specimen
with pre-damage. After the cyclic loading testing on Specimen C0,
the damaged C0 specimen was retrofitted by steel-jacket and RAC,
and re-named as Specimen C8. Specimen C9 is the SJRRC specimen
with preload. The preload was applied on Specimen C9 by using
the post-tensioned prestressing construction technique as
shown in Fig. 3. The original RC column of Specimen C9 was firstly
626 A. He et al. / Construction and Building Materials 158 (2018) 624–639

Fig. 2. Details of test specimens & locations of strain gauges (unit: mm).

Table 1
Specimen parameters.

Specimen t (mm) g (%) fcu1 (MPa) fcu2 (MPa) Ec (MPa) Npre (kN) N0 (kN) n0 Remark
C0 – – 34.2 – – – 300 0.29 Unstrengthen RC column
C1 3.82 50 34.2 29.4 17729 0 2667 0.79 SJRRC column
C2 3.82 50 34.2 29.4 17729 0 800 0.24 SJRRC column
C3 3.82 50 34.2 29.4 17729 0 1600 0.47 SJRRC column
C4 5.83 50 34.2 29.4 17729 0 2667 0.69 SJRRC column
C5 1.81 50 34.2 29.4 17729 0 2667 0.99 SJRRC column
C6 3.82 0 34.2 32.4 23136 0 1600 0.46 SJRRC column
C7 3.82 100 34.2 27.8 14296 0 1600 0.48 SJRRC column
C8 3.82 50 34.2 30.5 26160 0 800 0.23 SJRRC column with pre-damage
C9 3.82 50 34.2 29.4 17729 358 1600 0.47 SJRRC column with preload

Notes: t is the thickness of the steel jacket; g is the replacement ratio of recycled coarse aggregate; fcu1, fcu2 are the average compressive cubic strengths of the original column
concrete and infilled concrete respectively; Ec is the elastic modulus of the infilled concrete; Npre is the preload applied on the original column; N0 is the axial compressive
force; n0 is the axial compressive force ratio, which is calculated as n0 = N0/(0.76fcu1Ac1 + 0.76fcu2Ac2 + fyAs), where Ac1, Ac2 and As are the section area of the original column
concrete, infilled concrete and steel jacket respectively.

Fig. 3. Fabrication of specimen with preload.

 A. He et al. / Construction and Building Materials 158 (2018) 624–639
fabricated with a pipe imbedded along its central axis of section.
After 28 days of concrete curing period, the steel bars were
inserted into the pipe and the post-tensioning procedures were
carried out to apply compression on each side of the column
(Fig. 3(a)). The strains of the prestressed steel bar were monitored
throughout the prestressing procedure until their values became
stable. The preload on the original column was calculated as 358
kN, based on the stable value of strain of the steel bar. The foot
of the prestressed column was then set in the middle of the
base, where the concrete would be poured subsequently into
(Fig. 3(b)). After the solidification of the base concrete, steel
jacketing procedure was executed on the column (Fig. 3(c)).
For the infilled concrete, natural coarse aggregate, recycled
coarse aggregate, river sand, tap water and the ordinary Portland
cement with a 28 day’ compressive strength of 42.5 MPa were used
in this investigation. The recycled coarse aggregates were pro-
duced from a concrete recovery plant in Guangzhou with the
aggregate size from 5 to 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. The recycled coarse aggregates were
not pre-wetting. However, during manufacturing the RAC, addi-
tional water was added to compensate the high water absorption
capacity of recycled coarse aggregates according to JGJ/T 240–
2011 [28]. The mixture proportions of the infilled concrete are
listed in Table 2. The measured compressive strength and elastic
modulus of the concrete used in each specimen are showed in
Table 1. The measured steel properties of the specimens are listed
in Table 3.
2.2. Test set-up and measurements
The experimental set-up for the cyclic loading tests is shown in
Fig. 4. The base of the specimen was fixed on the strong floor. A
hydraulic jack (of 3 MN capacity) was used to apply a constant
axial compressive force at the top of the column. A 300 kN capacity
MTS actuator was located at the top of the specimen to provide the
reversed cyclic forces in order to simulate the seismic loading. Dis-
placement control loading procedure [29,30] was used during the
test. The cyclic loading spectrum is shown in Fig. 5. The amplitudes
of the top end horizontal displacement at each cyclic load steps
were L/500, L/300, L/150, L/100, L/75, L/50, L/30 and L/20 respec-
tively, in which L is the effective height of the columns. The loading
cycle was repeated three times at each load step. The loading pro-
cedure was attempted until the lateral load resistance of the col-
umn had deteriorated below 85% of the maximum lateral strength.
Table 2
Mixture proportion of infilled concrete.
Specimen g (%) Water (kg) Cement (kg)
C6 0 205 422
C1–C5, C8, C9 50 256 422
C7 100 306 422
Table 3
Measured material properties of rebar and steel tube.
Member Bar U6 Bar U12
fy or fyb (MPa) 439 435
fu or fub (MPa) 593 654
Es (GPa) 192 150
Notes: fy and fyb are the tensile yield strength of the steel tube and rebar, respectively; fus and fub are the tensile ultimate strength of the steel tube and rebar, respectively; Es is
the elastic modulus of steel.
627
The lateral displacement at the top of the column was measured
via a linear voltage displacement transducer (LVDT). Strain gauges
were used to measure the strains in the steel tube and rebar. The
layout of the strain gauges is shown in Figs. 2 and 4.
3. Experimental results and discussions
3.1. Test observations and failure modes
For the unstrengthened Specimen C0, no obvious phenomena
were observed on the specimen under axial compression. During
the cyclic loadings procedure, cracks firstly appeared in the vicinity
of the area located at about 200 mm above from the bottom of the
column when the lateral displacement reached L/300. The strains
of the tensile bars approached yielding at the lateral displacement
of L/150. When the lateral displacement increased to L/100, longi-
tudinal cracks appeared at the compressive side of the column bot-
tom and the RC column reached its peak strengths. Eventually, the
protective layer concrete began to crash at the lateral displacement
of L/50, when the lateral strengths of the RC column dropped to
85% of its peak strengths and the testing was terminated. Fig. 6
(a) shows the failure mode of the RC column.
For the SJRRC column specimens, all of them had similar overall
behaviour under cyclic loading except Specimen C5. As the lateral
displacement increased to L/150, the bottom of the steel jacket at
the tensile zone was pulled out from the base, as shown in Fig. 6
(b). The distance between the steel jacket bottom and the base sur-
face was extended to up to about 10 mm as the amplitude of the
lateral displacement increased. The retrofitted columns reached
their peak strengths when the lateral displacement was in the
range of L/100 and L/50. Because Specimen C5 had thinner steel
jacket and was subjected to larger axial compressive forces, it
exhibited slightly different behaviour compared with the other
SJRRC column specimens. The shear slip lines appeared gradually
on the surface of the steel jacket of Specimen C5, which subse-
quently led to the local buckling of steel jacket occurred at the lat-
eral displacement of L/75 followed by the retrofitted column
attaining its peak strengths. No obvious steel jacket pulling out
was observed on Specimen C5.
In order to identify specifically the failure modes of the SJRRC
columns, the steel jackets were cut off after the testing to exam
the damage condition of inner concrete as shown in Fig. 7. It was
found firstly that the interface between the new and old concrete
was intact in all the specimens, which indicated good bond inter-
action between the two layers of concrete. Table 4 summarises
Sand (kg) Natural coarse aggregate (kg) Recycled coarse aggregate (kg)
634 1079 0
634 540 540
634 0 1080
Steel tube t = 1.81 mm Steel tube t = 3.82 mm Steel tube t = 5.83 mm
432 400 356
570 550 509
221 184 173
628 A. He et al. / Construction and Building Materials 158 (2018) 624–639

Fig. 4. Loading configuration.

modes, viz. compression-controlled failure mode and tension-

Fig. 5. Cyclic loading spectrum.
the characteristic events and their corresponding occurrence tim-
ing during the testing for all specimens. Based on the observations
on the damage conditions of inner concrete and the yielding condi-
tions of steel rebar and jacket recorded in Table 4, two failure
controlled failure mode, were identified for the SJRRC columns.
The compression-controlled failure mode was characterised by
the crushing of infilled or original column concrete (Fig. 7
(a) and (b)). Specimens C1, C4 and C5 with the axial compression
ratio as 0.79, 0.69 and 0.99 respectively were categorised as the
specimens with compression-controlled failure mode. In these
specimens, the longitudinal bars reached compressive yielding
state relatively early at the lateral displacement of L/150. The ten-
sile yielding of longitudinal bars was initiated late at the lateral
displacement of L/30 far after the specimen reached the peak load,
or even did not take place (such as Specimen C5). The tension-
controlled failure mode was characterised by the over-yielding of
longitudinal bars of original column. At the peak load, the inner
concrete still remained integrity (Fig. 7(c)). Specimens C2, C3, C6,
C7, C8 and C9 with axial compression ratio ranged from 0.23 to
0.48 were categorised as the specimens with tension-controlled
failure mode. The longitudinal bars and steel jacket in these spec-
imens yielded in compression relatively later at the lateral dis-
placement from L/100 to L/30 compared with the specimens with

(a) Failure of (b) Pulling out of steel jacket from (c) Shear slip lines of steel
RC column base jacket
Fig. 6. Experimental observations.
 A. He et al. / Construction and Building Materials 158 (2018) 624–639 629

(a) C4 (b) C5 (c) C6

Fig. 7. Damage conditions of inner concrete.

Table 4
Characteristic events during testing.

Event Lateral displacement (mm)

L/150 L/100 L/75 L/50 L/30 Not happened
Tensile yielding of longitudinal bars – – C3, C6, C7, C9 C2, C8 C1, C4 C5
Compressive yielding of longitudinal bars C1, C4, C5 C9 – C3, C6, C7 C2, C8 –
Steel jacket yielding in loop direction – C5 C1, C3, C7 C6, C9 C2, C8 C4
Steel jacket yielding of in longitudinal direction – C1, C5 C9 C3, C6, C7 C2, C8 C4
Pulling out of steel jacket from base C1, C2, C3, C4, C6, C7, C8, C9 – – – – C5
Local buckling of steel jacket – – C5 – – C1, C2, C3, C4, C6, C7, C8, C9

compression-controlled failure mode, and the yielding of longitu-
dinal bars in tension initiated relatively earlier as the lateral dis-
placement developed in the range from L/75 to L/50.
3.2. P–D hysteresis curves
Fig. 8 shows the P–D hysteresis curves for all specimens under
cyclic loading, where P is the lateral force applied at the top of the
column, and D is the corresponding displacement. It shows the
unstrengthen Specimen C0 reaches its peak load at the lateral dis-
placement of L/100. After that, the lateral load drops quickly to 60%
of its peak load when the lateral displacement reaches L/50. For the
SJRRC column specimens, it is observed that the load-displacement
response is closely linear at the beginning. Following this, the stiff-
ness degrades and the behaviour of the specimens enters the
inelastic range. The specimens with compression-controlled failure
mode have plumper hysteresis curves than those with tension-
controlled failure mode, which exhibited their better energy dissi-
pation ability. Specimen C7 having 100% recycled coarse aggregate
replacement ratio shows more serious pinching effect on the hys-
teresis curve than the other specimens with lower recycled coarse
aggregate replacement ratio, which indicated that Specimen C7
had worse seismic behaviour than the other two. As is known
[31], the degree of pinching effect is related to the tensile crack
width of concrete, elongation strain of tensile steel bars, relative
slippage between steel bars and concrete, and the accumulation
of concrete plastic deformation. Previous studies have demon-
strated that RAC has slightly lower bond performance with steel
bar and is more fragile [32] than the normal aggregated concrete,
which could be the reasons causing Specimen C7 with the maxi-
mum percentage of recycled coarse aggregate replacement ratio
having the slightly more unfavourable seismic performance. The
pinching effect is also shown on the hysteresis curve of Specimen
C8 with pre-damage as is compared with the curve of the normal
SJRRC column Specimen C2.
3.3. Skeletal curves and ductility
Fig. 9 depicts the skeletal load-deflection curves for all speci-
mens. The characteristic loads and displacements of the skeletal
curves are listed in Table 5. Quantitatively, Table 5 also shows
the ductility coefficients at the ultimate stage lu and the average
ductility coefficients la, which are given respectively by
lu ¼ Du =Dy ð1Þ
la ¼ ðluþ þ lu Þ=2 ð2Þ
where Dy is the displacement corresponding to the yield load Py, Du
is the displacement corresponding to a 15% degradation from the
peak load [33]. lu+ and lu are the ductility coefficients in positive
and negative direction, respectively. It is worth noting that as is
shown in Table 5, the ductility coefficients of Specimen C0 in posi-
tive and negative directions are quite different with the values as
2.50 and 5.52, respectively. This difference could be caused by the
unexpected early crushing of the protective layer of concrete on
one side of the column (Fig. 6(a)), which leaded to the rapid
strength degradation on the positive loading direction of the col-
umn after peak load.
As Fig. 9 and Table 5 show, the peak strengths of the SJRRC col-
umns are about 1.86–3.44 times higher than the counterpart of the
unstrengthen RC column C0. The retrofitted columns also show
more ductile post-peak-load behaviour as their ductility coeffi-
cients la are between 4.05 and 7.93 while the ductility coefficient
la of Specimen C0 is only 4.01. These indicate that through the
steel-jacket retrofitting, the initial stiffness, bearing capacity, and
ductility of the columns are improved significantly.
As shown in Fig. 9(a), the change in the replacement ratio of
recycled coarse aggregate g might slightly affect the skeletal
load-deflection curves of retrofitted columns. Specimen C3 with
50% replacement ratio of recycled coarse aggregate reaches the
peak load faster than Specimen C7 with 100% replacement ratio
630 A. He et al. / Construction and Building Materials 158 (2018) 624–639

Fig. 8. Load-displacement hysteresis curves.

as well as Specimen C6 with 0% replacement ratio, which indicated
that Specimen C3 has larger secant stiffness before the peak load
than the others. Specimen C7 (g = 100%) has slightly lower peak
strength than Specimens C6 (g = 0%) and C3 (g = 50%). The average
ductility coefficients la of Specimen C6, C3 and C7 are 7.03, 7.17
and 7.30, respectively. This irregular effect of the replacement ratio
of recycled coarse aggregate could be caused by the effects of the
randomness of recycled coarse aggregate defects [34]. Overall, it
was demonstrated from the experimental results that the influence
of the replacement ratio of recycled coarse aggregate g on the lat-
eral behaviour of SJRRC columns is insignificant.
As Fig. 9(b) and (c) show, the increase of the axial compressive
load and steel jacket thickness could increase dramatically the ini-
tial stiffness, peak strength and ductility of the SJRRC columns.
 A. He et al. / Construction and Building Materials 158 (2018) 624–639 631

Fig. 9. Skeleton curves.

However, the strengthen efforts could be weakening once the axial
load is larger than 2667 kN or the steel jacket thickness is beyond
5.83 mm.
Fig. 9(d) shows the effect of the preload of original column. The
positive and negative peak strengths of Specimen C9 with preload
are 141.4 kN and 149.8 kN respectively, while the positive and
negative peak strengths of its counterpart one without preload
(Specimen C3) are 131.3 kN and 141.4 kN respectively, which
states that the preload may slightly increase the peak strength of
the SJRRC column. However, the preload produces little negative
effect on the post-peak behaviour of specimen. The average ductil-
ity coefficient of Specimen C3 and C9 are 7.17 and 7.06
respectively.
Fig. 9(e) shows the effect of pre-damage on the behaviour of ret-
rofitted columns. It can be seen that the peak strengths of the
SJRRC column Specimen C2 is 7% higher than that of Specimen
C8 with pre-damage, although the average ductility coefficient of
Specimen C2 (la = 6.96) is lower than Specimen C8 (la = 7.68). It
can be said that the steel-jacket retrofitting approach with RAC
would be the effective solution used for post-earthquake RC struc-
tures retrofitting.
3.4. Energy dissipation
The equivalent viscous damping coefficient factor neq provides a
measurement of the energy dissipation capacity of a structure’s
seismic performance [35], which can be calculated from:
S1
neq ¼ ð3Þ
2pðS2 þ S3 Þ
where S1 is the area of the hysteresis loop ABCDA, S2 the area of the
triangle OBF and S3 the area of the triangle ODE, as Fig. 10 shows.
632 A. He et al. / Construction and Building Materials 158 (2018) 624–639

Table 5
Ductility parameters.

Specimen C0 C1 C2 C3 C4
Loading direction +  +  +  +  + 
Pu (kN) 67.0 46.8 152.3 132.4 95.7 103.4 131.3 141.4 160.7 137.3
Du (mm) 13.1 15.5 36.7 30.1 46.4 36.5 28.9 43.4 29.6 28.4
hu (rad) 1/101 1/85 1/36 1/44 1/28 1/36 1/46 1/30 1/45 1/46
Pmax (kN) 78.8 55 179.2 155.7 112.6 121.7 154.5 166.4 189 161.6
Dm (mm) 12.0 9.7 25.1 20.7 28.4 18.6 12.0 23.5 22.0 16.3
hm (rad) 1/110 1/136 1/53 1/64 1/46 1/71 1/110 1/56 1/60 1/81
Py (kN) 50.4 33.6 96.2 97.9 74.2 89.1 137.9 111.1 113.1 102.5
Dy (mm) 5.2 2.8 5.0 4.4 6.5 5.4 5.7 4.7 3.9 3.5
hy (rad) 1/254 1/471 1/264 1/300 1/203 1/244 1/232 1/281 1/338 1/377
lu=Du/Dy 2.50 5.52 7.34 6.88 7.20 6.72 5.12 9.23 7.67 8.19
la 4.01 7.11 6.96 7.17 7.93
Specimen C5 C6 C7 C8 C9
Loading direction + – + – + – + – + –
Pu (kN) 103.0 87.0 131.7 137.4 130.2 129.5 90.2 95.1 141.4 149.8
Du (mm) 19.8 16.0 37.1 30.3 46.0 48.3 36.6 36.6 30.9 35.3
hu (rad) 1/67 1/83 1/36 1/44 1/29 1/27 1/36 1/36 1/43 1/37
Pmax (kN) 121.1 102.3 154.9 161.7 153.2 152.3 106.1 111.9 166.3 176.3
Dm (mm) 14.3 8.9 25.6 20.9 27.2 27.1 18.4 18.2 15.4 16.3
hm (rad) 1/92 1/148 1/52 1/63 1/49 1/49 1/72 1/73 1/86 1/81
Py (kN) 78.2 69.6 82.1 94.9 95.8 90.1 69.9 87.4 108.6 97.6
Dy (mm) 4.9 4.0 4.9 4.7 6.5 6.4 4.8 4.7 5.1 4.4
hy (rad) 1/269 1/330 1/269 1/281 1/203 1/206 1/275 1/281 1/259 1/300
lu = Du/Dy 4.05 4.05 7.54 6.51 7.05 7.55 7.62 7.75 6.05 8.06
la 4.05 7.03 7.30 7.68 7.06

Notes: Py is the yield load, Dy is the corresponding displacement at Py, hy is the corresponding drift ratio at Py; Pmax is the maximum load, Dm is the corresponding
displacement at Pmax, hm is the corresponding drift ratio at Pmax; Du is the ultimate displacement, Pu is corresponding load at Du, hu is the corresponding drift ratio at Pu.

Fig. 10. Typical hysteresis loop. Fig. 11. Equivalent viscous damping coefficient curves.

Fig. 11 depicts the relationships between neq and cycle steps for all
specimens. It can be seen from Fig. 11 that the SJRRC columns tend
to induce larger amounts of dissipated energy than the unstrength-
ened RC columns. The Specimen C7 whose recycled coarse aggre-
gate replacement ratio is 100% has slightly lower dissipated
energy than the specimens with 0% or 50% recycled coarse aggre-
gate replacement ratio, although the difference is not significant.
As the increase of axial compressive forces, the equivalent viscous
damping coefficient factor of retrofitted column at the same cycle
step increases. For example, neq of Specimens C2, C3 and C1 are
0.1450, 0.1566 and 0.1570 respectively at the 5th cycle step. The
retrofitted columns with pre-damage or preload have slightly lower
values of neq.
3.5. Strength and stiffness degradation
The evolution of strength degradation is evaluated by compar-
ing the peak load between the first cycle and second cycle of each
load step, which is expressed as
k ¼ Q 2j;max =Q 1j;max ð4Þ
2
where Q j,max and Q1j,max
are the peak load for the 2st and the 1th
loading cycle at the jth load step respectively. Fig. 12 shows the
strength degradation curves for all specimens. It can be observed
that the strength degradation coefficients of Specimen C0 decreases
dramatically, but the strength degradation coefficients of most of
the retrofitted columns maintain at certain level without drastic
descending as well as are always over 0.9. Two exceptions are Spec-
imens C8 and C9, which indicate that the pre-damage or preload of
the original column could have unfavourable influence on the cyclic
behaviour of retrofitted columns.
The evolution of stiffness degradation is evaluated by comput-
ing the secant stiffness of a member with different loading cycles
at the first cycle of each load step, which can be expressed as
j þ Fij þ j  Fij
K¼ ð5Þ
j þ Xij þ j  Xij
A. He et al. / Construction and Building Materials 158 (2018) 624–639 633

Fig. 12. Strength degradation curves.

Fig. 13. Secant stiffness curves.

634 A. He et al. / Construction and Building Materials 158 (2018) 624–639

(
where +Fi (Fi) is the positive (negative) peak load at the ith load 2x  x2 ; ðx 6 1Þ
step, and +Xi (Xi) is the displacement corresponding to the positive y¼ ð6Þ
x
bo ðx1Þ2 þx
; ðx > 1Þ
(negative) peak load at the ith load step. The degradations of the
secant stiffness are plotted in Fig. 13. It can be seen that the secant where
stiffness of the column is greatly increased after being retrofitted by rc rc
the steel jacket and RAC. Fig. 13(a) shows that the increase of recy- y¼ ¼ ð7Þ
r0 a  f c
cled coarse aggregate replacement ratio has adverse effect on the
secant stiffness of SJRRC columns. On the other hand, as shown in
a ¼ 1  0:28g þ 0:08g2 ð8Þ
Fig. 13(b) and (c), the increase of the applied axial compressive
loads and steel jacket thickness could enhance the secant stiffness ec ec
of SJRRC columns. It can be observed from Fig. 13(d) and (e) that x¼ ¼ ð9Þ
e0 b  ð1300 þ 12:5f c þ 800n0:2 Þ  106
the pre-damage or preload of original column has unfavourable
influences on the secant stiffness of SJRRC columns although their g
effects are very limited. b¼1þ ð10Þ
65:715g2  109:43g þ 48:989

4. Finite element modelling and

½0:25þðn0:5Þ7 
bo ¼ ð2:36  105 Þ
0:5
4.1. General ðf c Þ  0:5 P 0:12 ð11Þ
where r0 and e0 are maximum compressive stress and the corre-
In this section, the numerical simulations are performed by sponding strain respectively, g is the recycled coarse aggregate
using commercial finite element (FE) programme ABAQUS. The replacement ratio, a and b are the impact factors of recycled coarse
developed FE model of the retrofitted column is shown in Fig. 14. aggregate replacement ratio for the ultimate strength and the cor-
The inelastic behaviour of concrete is defined by using the concrete responding strain of RAC respectively. The coefficient n representing
damaged plasticity model. The elastic modulus of infilled concrete the confinement effect is given by
(Ec) is taken from Table 1, while the value for original column con-
pffiffiffiffi As f y
crete is calculated as 4730 f c according to ACI 318–11 [36], in n¼ ð12Þ
which fc is the compressive cylinder strengths (in MPa) that is
Ac f c
taken as 0.76 times of the compressive cubic strength [37]. The in which fy is the tensile strength of steel tube, As and Ac are the
compressive equivalent stress-strain curve proposed by Yang cross-sectional areas of the steel tube and inner concrete respec-
[38] which takes the influence of recycled coarse aggregate tively. The uniaxial tensile response of concrete is assumed to be
replacement ratio into consideration was adopted to simulate the linear until the tensile strength is attained. The tensile strength of
material response of confined normal aggregate concrete and the concrete is taken as 0.1fc. Beyond the tensile strength, the ten-
RAC. The equations for the stress-strain (rc–ec) relationship are sile response of concrete is represented by mean of fracture energy
expressed as: (GF) proposed by MC 2010 [39] as:

(a) SJRRC column (b) Original concrete (c) Infilled concrete (d) Steel jacket
Fig. 14. A general view of FE model.
A. He et al. / Construction and Building Materials 158 (2018) 624–639 635

Fig. 15. Comparisons of load-displacement skeleton curves.

636 A. He et al. / Construction and Building Materials 158 (2018) 624–639

based on the change of the axial compressive stress. The equation

for the variable Poisson’s ratio of confined concrete is given by [40]
8  41
>
< 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
0 0

ð14Þ
For modelling the steel material, the flow theory using von
Mises’ yield criterion is adopted. The constitutive relationships
for the steel tube and reinforcing bars are represented by an
elastic-plastic model with linear hardening. The strain hardening
modulus is taken as 0.01Es. The ultimate strain of steel is taken
as 0.2. The equation of variable Poisson’s ratio of steel is given by
[40]

rs  f p
ls ¼ 0:167  þ 0:283 ð15Þ
fy  fp

in which rs is the longitudinal stress of steel tube, fp is the elastic

Fig. 16. Effect of N0 on lateral load (P)-displacement (D) skeleton curves.
limit strength of steel tube and is taken as 0.9fy herein.
The end plates and concrete components are simulated by using
3D eight-node solid reduced-integration element. Truss element
T3D2 of ABAQUS is used for the steel bars which are embedded
in the concrete. The steel tube is simulated by four-node conven-
tional shell element. The end plate is assumed to be elastic rigid
body. The TIE contact is applied at the interfaces between the
end plate and the main specimen. The surface-to-surface contact
interaction with finite sliding is used to simulate the contact beha-
viour between steel jacket and the infilled concrete. A contact sur-
face pair comprised of the inner surface of steel jacket and the
outer surface of infilled concrete is defined. The hard contact prop-
erty which allows the separation of the interface in tension and no
penetration of that in compression is specified in the normal direc-
tion of contact surface. The PENALTY option with the friction coef-
ficient of 0.6 is used for the tangential behaviour [41]. The same
contact interaction is applied at the interfaces between the infilled
concrete and original concrete except the friction coefficient is
assumed to be 0.8 [42]. The bottom of the end plate is fixed against
all degrees of freedom and the top of the end plate is the free end.
Fig. 17. N0–M0 interaction curve of SJRRC columns.
 0:7 4.2. Numerical results
2 f
GF ¼ ð0:0469dmax  0:5dmax þ 26Þ c N=m ð13Þ
10 The load-displacement skeleton curves predicted by FE models
are compared with the measured curves obtained from the exper-
where dmax is the maximum size of coarse aggregate (in mm). The imental tests, as shown in Fig. 15. It can be seen that the
Poisson’s ratio of the confined concrete is considered to be variable results predicted by FE models have a good agreement with the

Fig. 18. Effect of g on lateral load (P)-displacement (D) skeleton curves.

 A. He et al. / Construction and Building Materials 158 (2018) 624–639
experimental results. The mean values and standard deviation of
the ratio of the numerical results and the experimental results
are 0.967 and 0.112, respectively.
In order to illustrate the importance of considering variable
Poisson’s ratio in the FE model, the load-displacement skeleton
curves predicted by FE models using constant Poisson’s ratio are
also obtained and shown in Fig. 15 for comparison. It can be
observed that the FE model with constant Poisson’s ratio may
underestimate the ultimate strength of the columns, especially
the ones with higher axial compressive force ratio. The mean val-
ues and standard deviation of the ratio of the numerical predic-
tions using constant Poisson’s ratio and the experimental results
are 0.760 and 0.134, respectively. Given being under the same
compressive stress, the core concrete in the model considering
variable Poisson’s ratio experience larger lateral dilation than the
one with constant Poisson’s ratio, which leads to the earlier con-
finement pressure effect applied from the steel tube on the con-
crete in the model. The earlier confinement effect is superior in
increasing the concrete strength, which results in increasing the
ultimate strength of the column [43].
4.3. Parametric studies
Based on the developed FE model, parametric studies are con-
ducted to investigate the effects of variation in the axial compres-
sion applied on the column N0, replacement ratio of recycled coarse
aggregate g, preload on original RC column Npre and the thickness
of steel jacket t on the lateral load-displacement responses of
SJRRC columns. In the parametric studies, the SJRRC column having
the same geometrical and material parameters as Specimen C2 is
adopted as the reference. It should also be noted that the shear
span to depth ratio of all the columns in this study are greater than
4 [40], so the effects of shear response on the lateral behaviour and
section capacity of the column can be sensibly neglected. The lat-
eral load and shear deformation behaviour [24] of the short SJRRC
columns need to be studied in the future research.
4.3.1. Effect of N0
Fig. 16 shows the effect of the axial compression applied on the
column N0 on the lateral load-displacement skeleton curves of the
retrofitted columns. Apparently, the change in the axial load pro-
duces substantial influence on the lateral behaviour of SJRRC col-
umns. It is shown in Fig. 16 that as the axial compressive force
increase, the lateral strength of the column increases dramatically.
For example, when the N0 is changed from 800 kN to 2400 kN, the
Fig. 19. Effect of Ppre on lateral load (P)-displacement (D) skeleton curves.
637
peak lateral load of the column increases by 63.1% from 103 kN to
168 kN.
In order to illustrate comprehensively the impact of N0 on the
lateral behaviour of SJRRC columns, an additional group of columns
with different axial compressive loads widely ranged from 0 kN to
Nu (Nu is the compressive section capacity of the columns) were
analysed. The lateral load-displacement skeleton curve of each col-
umn and its corresponding maximum lateral load Pmax and lateral
displacement Dm at Pmax were recorded. The maximum moment
M0 of each column were calculated by
M0 ¼ N0 Dm þ Pmax l ð16Þ
where l is the effective length of SJRRC column. Therefore, the
N0–M0 interaction curve for the section capacity of SJRRC columns
is depicted in Fig. 17. Similar as the typical axial compression and
bending moment interaction curve for RC column, the N0–M0 inter-
action curve of Fig. 17 can be divided into two phases as the
tension-controlled failure phase and the compression-controlled
failure phase, while the junction between compression-controlled
failure and tension-controlled failure is defined as the balance fail-
ure point [44]. In the tension-controlled failure phase, an increase in
N0 leads to an increase of M0. In the compression-controlled failure
phase, an increase in N0 leads to a reverse effect causing a decrease
of M0.
4.3.2. Effect of g
The effect of recycled coarse aggregate replacement ratio g on
the lateral load-displacement skeleton curves of the retrofitted col-
umns is shown in Fig. 18. Two loading conditions, one in
compression-controlled failure phase and one in tension-
controlled failure phase, are considered. It can be seen that as
the g increases, the ultimate strength of retrofitted column slightly
decreases. Under the tension-controlled failure loading condition
(Fig. 18(a)), the ultimate strength of the SJRRC column with
g = 100% is 1.79% lower than that of the column filled with normal
aggregate concrete (g = 0%). Under the compression-controlled
failure loading condition (Fig. 18(b)), the reduction ratio slightly
increases to 3.91%.
4.3.3. Effect of Ppre
The effects of preloads on the original column Ppre on the lateral
load-displacement skeleton curves of the retrofitted columns are
shown in Fig. 19. It exhibits that the effect of preload from the orig-
inal column is negligible.
638 A. He et al. / Construction and Building Materials 158 (2018) 624–639
Fig. 20. Effect of t on lateral load (P)-displacement (D) skeleton curves.
4.3.4. Effect of t
Fig. 20 displays the effects of steel jacket thickness t on the lat-
eral load-displacement skeleton curves of the retrofitted columns.
It can be observed that the change in the thickness of steel jacket
has significant impact on the performance of SJRRC column. With
the steel jacket thickness increasing, the ultimate strengths of
SJRRC column enhance dramatically. Under the tension-
controlled failure loading condition (Fig. 20(a)), if the thickness
of steel jacket t increases from 1.81 mm to 3.82 mm and 5.83
mm, the ultimate strengths of the retrofitted column rise by
15.43% and 23.29%, respectively. Under the compression-
controlled failure loading condition (Fig. 20(b)), the influence of
increasing steel jacket thickness is even more favourable, since
by increasing t from 1.81 mm to 3.82 mm and 5.83 mm, the ulti-
mate strengths of the retrofitted column rise by 51.51% and
78.88%, respectively.
5. Conclusions
The following conclusions can be drawn based on the results of
the experimental and numerical investigations in this paper:
(1) The seismic performance of the RC columns including initial
stiffness, bearing capacity, ductility and energy dissipation
ability improved significantly after the columns being retro-
fitted by the steel jacket and RAC. In details, the peak
strengths of the SJRRC columns were 1.86–3.44 times than
that of the original column, and the retrofitted columns
showed ductile post-peak load behaviour with the ductility
coefficients ranged from 4.05 to 7.93.
(2) Two failure modes, viz. compression-controlled failure mode
and tension-controlled failure mode, were identified for the
SJRRC columns under cyclic loading. The compression-
controlled failure mode was characterised by the crushing
of infill or original column concrete, while the tensile yield-
ing of longitudinal bars was initiated late after the specimen
reached the peak load, or even did not take place. The
compression-controlled failure mode occurred in the speci-
mens with relatively high axial compression ratio ranged
from 0.69 to 0.99. On the other hand, the tension-
controlled failure mode was characterised by the over-
yielding of the longitudinal bars of original column, while
the inner concrete could remain integrity at the peak load.
The tension-controlled failure mode occurred in the speci-
mens with relatively low axial compression ratio ranged
from 0.23 to 0.48. In terms of the seismic performance, the
specimens with the compression-controlled failure mode
exhibited plumper hysteresis curves, better energy dissipa-
tion ability and higher secant stiffness than those with
tension-controlled failure mode.
(3) The specimen whose recycled coarse aggregate replacement
ratio is 100% showed more serious pinching effect on the
plumper hysteresis curve, slightly lower lateral strength
and secant stiffness than the specimens whose replacement
ratios of recycled coarse aggregate are 0% and 50%. Never-
theless, due to the effects of the randomness of recycled
coarse aggregate defects, the influence of recycled coarse
aggregate replacement ratio on the structural performance
of SJRRC columns was still questionable and need further
research, but was certainly very limited. Hence, the RACs
could be adopted in the steel jacketing approach for column
retrofitting in the seismic area.
(4) The experimental results showed that the effect of the pre-
load and pre-damage on the original column would be unfa-
vourable but limited to the cyclic performance.
(5) The finite element analyses considering the variable of
material Poisson’s ratio were also performed to investigate
the structural performance of the retrofitted columns under
combined axial compression and bending moment. The pro-
posed FE model was validated through comparing its results
with the test results. The mean values and standard devia-
tion of the ratio of the numerical and experimental results
are 0.967 and 0.112, respectively.
(6) By using the developed FE model, parameter studies were
undertaken. It can be seen that the influence of the change
in the axial load on the lateral behaviour of SJRRC columns
is substantial. The interaction curve of combined axial load
and bending moment for the SJRRC columns was depicted.
The interaction curve can be divided into two phases as
the tension-controlled failure phase and the compression-
controlled failure phase. In the tension-controlled failure
phase, an increase of axial load leads to an increase of the
lateral strength of column. In the compression-controlled
failure phase, an increase of axial load leads to a reverse
effect causing a decrease of the lateral strength of column.
(7) It was also confirmed from the parametric studies that,
although the effects of the recycled coarse aggregate
replacement ratio and preloads on the lateral response of
SJRRC column could be negative, their influence was mar-
ginal with the maximum reduction ratio of lateral strength
being only 3.91%.
(8) With the increase of the thickness of steel jacket, the ulti-
mate strengths of SJRRC column enhance dramatically. The
effect of steel jacket thickness on the ultimate strength of
 A. He et al. / Construction and Building Materials 158 (2018) 624–639
SJRRC column is more significant for the specimens with the
compression-controlled failure mode rather than the one
with the tension-controlled failure mode.
Acknowledgements
This study was funded by the National Natural Science Founda-
tion of China (51578246), the Fundamental Research Funds for the
Central Universities (2014ZZ0026), the State Key Lab of Subtropical
Building Science, South China University of Technology
(2015ZC18), and Natural Science Foundation of Guangdong
Province, China (2016A030310418, 2017A030313263).
References
[1] Civil and Structural Groups of Tsinghua University, Xinan Jiaotong University
and Beijing Jiaotong University. Analysis on seismic damage of buildings in the
Wenchuan earthquake, Journal of Building Structures, 2008, 29, 4, pp. 1–9.
[2] D. Lavorato, C. Nuti, Pseudo-dynamic tests on reinforced concrete bridges
repaired and retrofitted after seismic damage, Eng. Struct. 94 (2015) 96–112.
[3] 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.
[4] G.T. Truong, J.C. Kim, K.K. Choi, Seismic performance of reinforced concrete
columns retrofitted by various methods, Eng. Struct. 134 (2017) 217–235.
[5] H. Sezen, E.A. Miller, Experimental evaluation of axial behavior of strengthened
circular reinforced-concrete columns, J. Bridge Eng. 16 (2) (2011) 238–247.
[6] 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.
[7] Y.H. Chai, An analysis of the seismic characteristics of steel-jacketed circular
bridge columns, Earthquake Eng. Struct. Dyn. 25 (2) (1996) 149–161.
[8] 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.
[9] 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.
[10] P.N. Laplace, D.H. Sanders, M.S. Saiidi, B.M. Douglas, S. El-Azazy, Retrofitted
concrete bridge columns under shaketable excitation, ACI Struct. J. 102 (4)
(2005) 622–628.
[11] 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.
[12] Y. Xiao, R. Ma, Seismic retrofit of RC circular columns using prefabricated
composite jacketing, J. Struct. Eng. 123 (10) (1997) 1357–1364.
[13] Y. Xiao, H. Wu, Retrofit of reinforced concrete columns using partially stiffened
steel jackets, J. Struct. Eng. 129 (6) (2003) 725–732.
[14] A. Ghobarah, A. Biddah, M. Mahgoub, Rehabilitation of reinforced concrete
columns using corrugated steel jacketing, J. Earthquake Eng. 1 (4) (1997) 651–
673.
[15] 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. Dyn. 39 (2010) 1687–1710.
[16] E. Choi, Y.S. Chung, K. Park, B.S. Cho, Behavior of reinforced concrete columns
confined by new steel-jacketing method, ACI Struct. J. 107 (6) (2010) 654–662.
[17] E. Choi, Y.S. Chung, K. Park, J.S. Jeon, Effect of steel wrapping jackets on the
bond strength of concrete and the lateral performance of circular RC columns,
Eng. Struct. 48 (2013) 43–54.
[18] A. He, J. Cai, Q.J. Chen, X. Liu, H. Xue, C. Yu, Axial compressive behaviour of
steel-jacket retrofitted RC columns with recycled aggregate concrete, Constr.
Build. Mater. 141 (2017) 501–516.
639
[19] Y.F. Yang, L.H. Han, L.T. Zhu, Experimental performance of recycled aggregate
concrete-filled circular steel tubular columns subjected to cyclic flexural
loadings, Adv. Struct. Eng. 12 (2) (2009) 183–194.
[20] B. Wu, X.Y. Zhao, J.S. Zhang, Cyclic behavior of thin-walled square steel tubular
columns filled with demolished concrete lumps and fresh concrete, J. Constr.
Steel Res. 77 (2012) 69–81.
[21] B. Wu, X.Y. Zhao, J.S. Zhang, Y. Yang, Cyclic testing of thin-walled circular steel
tubular columns filled with demolished concrete blocks and fresh concrete,
Thin-Walled Struct. 66 (2013) 50–61.
[22] 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.
[23] H. Ma, J. Xue, X. Zhang, D. Luo, Seismic performance of steel-reinforced
recycled concrete columns under low cyclic loads, Constr. Build. Mater. 48
(2013) 229–237.
[24] H. Ma, J. Xue, Y. Liu, X. Zhang, Cyclic loading tests and shear strength of steel
reinforced recycled concrete short columns, Eng. Struct. 92 (2015) 55–68.
[25] X. Shi, Q. Wang, X. Zhao, F.G. Collins, Structural behaviour of geopolymeric
recycled concrete filled steel tubular columns under axial loading, Constr.
Build. Mater. 81 (2015) 187–197.
[26] X. Shi, Q. Wang, X. Zhao, F.G. Collins, Strength and ductility of recycled
aggregate concrete filled composite tubular stub columns, Incorporating
Sustainable Practice in Mechanics of Structures and Materials, London, UK,
2010, pp. 83–89.
[27] Y.C. Tang, L.J. Li, W.X. Feng, F. Liu, B. Liao, Seismic performance of recycled
aggregate concrete – filled steel tube columns, J. Constr. Steel Res. 133 (2017)
112–124.
[28] JGJ/T 240-2011, Technical Specification for Application of Recycled Concrete,
Architecture & Building Press, Beijing: China, 2011.
[29] Y.T. Wang, J. Cai, Y.L. Long, Hysteretic behavior of square CFT columns with
binding bars, J. Constr. Steel Res. 131 (2017) 162–175.
[30] J.G. Nie, H.S. Hu, J.S. Fan, et al., Experimental study on seismic behavior of high-
strength concrete filled double-steel-plate composite walls, J. Constr. Steel Res.
88 (9) (2013) 206–219.
[31] Z. Guo, X. Shi, Reinforced Concrete Theory and Analyse, Tsinghua University
Press, Beijing, 2011.
[32] 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.
[33] L.H. Han, Z. Tao, W.D. Wang, Advanced Composite and Mixed Structures -
Testing, Theory and Design Approach, Science Press, Beijing, 2009.
[34] J.J. Xu, Z.P. Chen, Y. Xiao, C. Demartino, J.H. Wang, Recycled aggregate concrete
in FRP-confined columns: a review of experimental results, Compos. Struct.
174 (2017) 277–291.
[35] X.L. Tang, J. Cai, Q.J. Chen, X. Liu, A. He, Seismic behaviour of through-beam
connection between square CFST columns and RC beams, J. Constr. Steel Res.
122 (2016) 151–166.
[36] ACI318-11, Building code Requirements for Structural Concrete and
Commentary, American Concrete Institute, Detroit (USA), 2011.
[37] GB50010-2010, Code for Design of Concrete Structures, China Building
Industry Press, Beijing, 2010.
[38] 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.
[39] CEB-FIP MODEL CODE, Volumn 1, International Federation for Structural
Concrete (fib), Switzerland, 2010, p. 2012.
[40] L.H. Han, Concrete Filled Steel Tubular Structure, Science Press, Beijing, 2007.
[41] Z. Tao, Z.B. Wang, Q. Yu, Finite element modelling of concrete-filled steel stub
columns under axial compression, J. Constr. Steel Res. 89 (2013) 121–131.
[42] A.H. Mattock, Shear transfer in concrete having reinforcement at an angle to
the shear plane, ACI Spec. Publ. 42 (1974) 17–42.
[43] F.E. Richart, A. Brandtzaeg, R.L. Brow, A Study of the Failure of Concrete Under
Combined Compressive Stress Bulletin No. 185, Univ. of Illinois Engineering
Experiment Station, Urbana, IL, 1928.
[44] Y.F. An, L.H. Han, Behaviour of concrete – encased CFST columns under
combined compression and bending, J. Constr. Steel Res. 101 (2014) 314–330.