Engineering Structures 85 (2015) 162–169

Contents lists available at ScienceDirect

Engineering Structures
journal homepage: www.elsevier.com/locate/engstruct

A practical approach for the strength evaluation of RC columns

reinforced with RC jackets
Giovanni Minafò ⇑
Università degli Studi di Enna ‘‘Kore’’, Facoltà di Ingegneria e Architettura, Cittadella Universitaria, 94100 Enna, Italy

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

Article history: Reinforced concrete (RC) jacketing is nowadays one of the most common techniques adopted for seismic
Received 12 July 2014 retrofitting of existing RC columns. It is used to increase load-carrying capacity and ductility of weak
Revised 11 December 2014 existing members by means of a simple and cheap method. The structural efficiency is related to two
Accepted 12 December 2014
main effects: – the enlargement of the transverse cross section; – the confinement action provided by
Available online 29 December 2014
the external jacket to the inner core. Several theoretical and experimental studies were addressed in
the past to investigate on how it is possible to calculate the strength enhancement due to these effects
Keywords:
and to highlight the main key parameters influencing the structural behavior of jacketed columns. Most
RC columns
RC jacketing
of theoretical studies analyzed members subjected to axial compression while the case of axial force and
Retrofit bending moment was adapted only with complex formulations based on numerical approaches, which
Confinement require the use of a suitable algorithm (e.g. non-linear finite element analyses, sectional fiber models).
This paper presents a simplified approach, able to calculate the strength domains for jacketed columns
subjected to axial force and uniaxial bending moment. The model takes into account the effects of
confinement with proper stress-block parameters, the latter adapted for confined concrete, and of the
composite action of jacket and core; buckling of longitudinal bars is considered and discussed with an
appropriate stress–strain law for steel in compression. Results are compared with numerical analyses
carried-out with the fiber model approach implemented in a commercial software (SAP2000), showing
the accuracy of proposed method. Comparisons are also made with experimental results available in
the literature in order to validate the model. Finally parametric considerations are made on the basis
of adopted model, useful for design/verification purposes.
Ó 2014 Elsevier Ltd. All rights reserved.

1. Introduction strengthening technique. Particular attention has to be paid to

old-new concrete interface, which could reduce the flexural capac-
Reinforced concrete (RC) jacketing is always more frequently ity as observed in [1,2]. If concrete surface of the old member is not
adopted to retrofit existing RC columns with poor structural fea- roughened, the reduction in the effectiveness of composite column,
tures. This method consists in casting a concrete layer around in terms of flexural capacity, is almost 10%, while if interfaces are
the existing member, and reinforcing the jacket with a properly well roughened these effects are negligible [3,4]. Furthermore,
designed amount of longitudinal and transverse reinforcement the long term effects, including shrinkage, have to be carefully
(Fig. 1). The efficacy of the technique on the structural behavior taken into account, as stressed in [5].
is related to the enlargement of the transverse cross section, which From practical point of view some studies have proposed design
increases the load-carrying capacity and to the confinement pres- rules for concrete jacketing techniques [6]; specifically, these can
sure induced by the jacket in the inner column. This confining be summarized as follow: – the strength of the new materials
action allows to increase strength and ductility of the original con- utilized for the jacket must be greater than that of the column; –
crete, and to restrain buckling of longitudinal bars, especially when the thickness of the jacket should be at least 4 cm for shotcrete
stirrups in the column are largely spaced. application and 10 cm for cast-in-situ concrete; – the reinforce-
The efficiency of the RC jacketing is affected from different fac- ment should be not less than four bars for four-side jacketing
tors, which has to be taken into account when designing the and minimum bar diameter 14 mm; – the ties should be minimum
8 mm and at least 1/3 of the vertical bar diameter; – the vertical
spacing is at most 200 mm and close to the joint must not exceed
⇑ Tel.: +39 0935 536442; fax: +39 0935 536623.
100 mm. In addition, the spacing of the ties should not exceed the
E-mail address: [email protected]

http://dx.doi.org/10.1016/j.engstruct.2014.12.025
0141-0296/Ó 2014 Elsevier Ltd. All rights reserved.
 G. Minafò / Engineering Structures 85 (2015) 162–169 163

thickness of the jackets. Furthermore, the surface should be moist- for practical engineering applications, which allows taking into
ened before placing shotcrete and the existing concrete must be account the effect of confinement and effective interaction
heavily sandblasted and cleaned of all loose materials, dust and between core and jacket.
grease obtaining in this way a well-roughened surface. The current paper aims to provide a simplified formulation for
Different researches were carried out in the last twenty years to the calculation of strength domains of square columns reinforced
evaluate experimentally the efficacy of the technique on the struc- with RC jackets. The proposed approach is based on the determina-
tural behavior of RC columns. Ersoy et al. [7] tested two series of tion of some characteristic points defining the interaction domain.
jacketed columns under uniaxial compression or combined axial The corresponding values of axial force and bending moment are
load and bending moment. They studied the effectiveness of repair calculated by idealizing the constitutive laws of concrete in com-
and strengthening jackets and the differences between jackets pression with stress-blocks, the latter to be calibrated on the basis
made under load and after unloading. of the confinement pressure.
Julio et al. [2] carried out an experimental study to analyze the It has to be noted that in the proposed model, perfect bond
influence of the interface treatment on the structural behavior of between the old and the new concrete is assumed and the effect
columns strengthened by RC jacketing. After testing seven full- of jacket’s concrete shrinkage is neglected. It has been proved that
scale models of column-footing, they concluded that for undam- both parameters affect the response of the jacketed columns, so
aged columns a monolithic behavior of the composite element they should be carefully addressed when adopting the proposed
can be achieved even without increasing their surface roughness, moment design chart.
using bonding agents, or applying steel connectors before In particular, considering the effective connection between old
strengthening it by RC jacketing. and new concrete, it is well-known that the response of the com-
Takeuti et al. [8] tested twelve RC-jacketed columns under uni- posite member is complex, thus a practical design procedure
axial compression with and without preloading. The authors found should take advantage of a monolithic approach, making use of
that the entire core contributes to the axial capacity of the jacketed properly defined ‘‘monolithicity factors’’ [10]. However, if the
column, as long as adequate confinement is provided. Also, interface is well-roughened, bond between old and new concrete
preloading does not adversely affect the capacity of the jacketed can be ensured, as experimentally demonstrated in [2].
column, while it may increase its deformability. Additionally, shrinkage effects play an important role on the
From a theoretical point of view several research works were strength of jacketed columns. In RC jacketed columns concrete
addressed to this field. Among these Lampropoulos and Dritsos shrinkage is restrained by the presence of the initial column [5],
[5] analyzed the case of jacketed columns subjected to axial load so tensile stresses could develop, inducing a biaxial state of stress
and bending moment by means of non-linear finite element in the jacket. The flexural capacity reductions due to these effects
analyses. The authors studied the suitability of a proper formula- could be in the range between 23% and 46%, as discussed in [5].
tion to model the old-new concrete interface by comparing numer- The proposed approach discussed in the following should be
ical results with experimental data. More recently Campione et al. adopted in addiction with ‘‘monolithicity factors’’, and with a
[9] proposed a theoretical model to calculate proper constitutive reduction coefficient for taking into account shrinkage effects.
laws for old and new concrete and for steel, and validated their
model with experimental data available in the literature. The case 2. Constitutive laws of constituent materials
of eccentrically loaded columns was studied by considering a
numerical approach based on the discretization of the section by As discussed above, the adopted constitutive law has to take
means of the classic fiber model. into account the effect of confinement. Campione et al. [9] have
Concerning practical methods, different studies [10] focused on shown as the well-known model of Mander et al. [11] is suitable
the use of ‘‘monolithicity coefficient factors’’, which are used for to model the compressive behavior of concrete of both jacket and
the design of the strengthened elements. The application of these core. Therefore the following relationship is adopted:
factors is a ‘design approach’, proposed not only for the strength
e
 f cc  r
evaluation but also for stiffness, and deflection/rotation angle. rc ¼ ecc
ð1Þ
 r
It is clear that a combination of a simple calculation method r  1 þ eecc
with the use of ‘‘monolithicity coefficients’’ could be a useful tool
with
Ec
r¼ ð2Þ
Ec  Esec
pffiffiffiffi
where Ec ¼ 5000  f c in MPa and Esec ¼ ef cccc :
As well-known, the peak stress fcc and the peak strain ecc of con-
fined concrete have to be calculated on the basis of the effective
confinement pressure fl by means of the following relations [11]:
" sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi #
7:94  f l fl
f cc ¼ f c 2:254 1 þ  2   1:254 ð3Þ
fc fc

  
f cc
ecc ¼ eco  1 þ 5  1 ð4Þ
fc
with fc and eco the peak stress and strain of unconfined concrete.
The confining pressure is simply determinable from rigid body
equilibrium of the section in the plane of the stirrup, the latter con-
sidered to be yielded. The expressions of confinement pressure
induced from external and internal stirrups in the core assume
Fig. 1. Case study: square RC section reinforced with a RC jacket. the following form:
164 G. Minafò / Engineering Structures 85 (2015) 162–169

2  f ysc  Astc 0 6 x < 1; rc ¼ K  f c ð1  j1  xjn Þ ð11aÞ

f l;c ¼ Due to internal stirrups ð5Þ  
ðb  cc Þ  sc K  f c  f cu
1 6 x < xu ; rc ¼ K  fc  ðx  1Þ ð11bÞ
xu  1
2  f ysj  Astj
f l;j ¼ Due to external stirrups ð6Þ
ðB  dÞ  sj
in which fcu = stress corresponding to stirrup fracture strain; K = fcc/
being Astc and Astj the area of the legs in the core and jacket stirrups. fc confinement ratio, x = normalized strain where x = ec/ecc, xu = ecu/
As well-known from the literature [11], appropriate efficiency ecc, n = Ececo/fc and n = Ececc/fcc for unconfined and confined
coefficients have to be considered in order to take into account concrete, respectively.
the effective confined concrete area in the section of transverse Eq. (11) can be applied to model concrete in compression for
reinforcement and between two successive stirrups. Consequently, both jacket and core and they are here adopted due to their suit-
different coefficients are assumed respectively for core and jacket. ability to be used for simplified sectional analyses. In particular,
Efficiency coefficients for the confining pressure exerted by the RC for concrete of the core the value of fc,core has to be introduced in
jacket are calculated as in [9]. Particularly, the in-plane coefficient Eq. (11) instead of fc, while the confinement ratio K, the peak ecc
for the pressure induced from the jacket’s stirrups to the core is and the ultimate ecu strains have to be evaluated as discussed
calculated as the ratio between the effectively confined concrete above. Analogously for the jacket, if considered as unconfined con-
area of the core and the gross area of the core’s transverse section. crete, Eq. (11) can be used by introducing the jacket’s concrete
On these basis and considering that effectively confined concrete strength fc,jacket, and by considering ec0 = 0.0036 as suggested by
can be assumed as delimited from parabolic curves, the following Collins and Mitchell [13].
expression of the efficiency coefficient reported in [9] can be Fig. 2a shows the comparison between the constitutive laws in
considered: compression for both concrete of core and jacket. In particular the
qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi exact (Eq. (1)) and the simplified (Eq. (11)) form of the stress–
2 3
kej ¼ 1   ðb þ 2cj  2dÞ  b  2cj þ 2d 6 1 ð7Þ strain law in compression is plotted. The considered unconfined
2
3b compressive strength is fc,core = 18 MPa and fc,jacket = 30 MPa for
for bd 6 1
cj
core and jacket respectively. Good accordance can be noted
2 1 d
between the two curves, except for the limited softening branch
With the same assumptions expressed in [11], the vertical effi- of unconfined concrete. Fig. 2b shows the plots of the stress–strain
ciency coefficient of confinement pressure exerted from the jacket relationships for steel of longitudinal bars in both tension and
assumes the following form: compression. It has to be noted that for an exact calculation the
 2 constitutive law of steel in tension has to consider the strain-hard-
sj
kvj ¼ 1  ð8Þ ening effect, while that in compression has to include the buckling
2  ðb þ 2d  2cj Þ
effects, especially when stirrups are largely spaced. In this case the
The equivalent confinement pressure is therefore obtained con- constitutive model proposed by Dhakal and Maekawa [14] is able
sidering the effects of internal and external stirrups separately, giv- to consider the effect of buckling, and it is not here reported for
ing the following: the sake of brevity. As recalled in [9], the key parameter to evaluate
2
!
2  f ysc  Astc 4 ðb  2  cc  2  dbc Þ
f l;core ¼  1  2 35
ðb  cc Þ  sc 6 ðb  2  cc Þ Karthik and Mander (2011)
 2 30
Jacket (unconfined)
sc 2  f ysj  Astj Mander et al. (1988)
 1 þ
2  ðb  cc Þ ðB  dÞ  sj 25
 qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi qffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
2
σ [MPa]

3 20
 1 2
 ðb þ 2cj  2dÞ  b  2cj þ 2d Core (confined)
3b
 2 15
sj
 1 ð9Þ 10
2  ðb þ 2d  2cj Þ
5
The compressive behavior of confined concrete is finally defined
by calculating the ultimate strain ecu. This can be computed as sug- 0
gested in [6] and considering both the effects of internal and exter- (a) 0 0.005 0.01 0.015 0.02 0.025
nal stirrups: ε
 
2:8 esuj  Asj esuc  Asc
ecu ¼ eco þ  þ ð10Þ 400 σ (MPa)
f cc sj  ðB  dÞ sc  ðb  cc Þ
300
It has to be noted that in general the in-plane efficiency calculus model (no buckling)

coefficient of the jacket is quite low [9], especially if only four bars 200
Dhakal et al. model
are placed and for the common values of concrete cover. Therefore, 100 with buckling (exact analysis)
for design/verification purposes the jacket’s concrete can be
0
considered as unconfined. -0.2 -0.15 -0.1 -0.05 0 0.05 0.1 0.15 0.2
As it could be noted, the constitutive law of confined concrete -100
ε
expressed by Eq. (1) is not suitable for a straightforward calculus model -200
calculation, due to the arising difficulty in performing its
integration. However, Karthik and Mander [12] proposed a new -300
strain-hardening model
analytical form of the constitutive law, valid for both confined
(b) (exact analysis) -400
and unconfined concrete, and able to approximate the stress–
strain relationship expressed by Eq. (1). This can be written in Fig. 2. Detailed and simplified stress strain laws. (a) Compressive constitutive laws
the following form: for concrete of core and jacket. (b) Constitutive laws for steel of longitudinal bars.
 G. Minafò / Engineering Structures 85 (2015) 162–169 165

the buckling behavior of longitudinal bars is the critical length-to- 1.00

fc,core=10 MPa
diameter ratio L/Øl, which can be calculated with a simple model of
fc,core=20 MPa
elastic beam on elastic soil. With this assumptions, Campione et al. 0.80
[9] also demonstrated that second order effects are negligible for
pitch-to-diameter s/Øl ratios less than 4.5, consequently this value 0.60 fc,core=30 MPa
is recommended as design reference for stirrups for the jacket. In

αβ
the following, elastoplastic behavior of steel is assumed for rein-
0.40
forcement of both core and jacket, in tension and in compression
for the sake of simplicity. However, it has to be stressed that a pre- sj=b/2
liminary verification of the critical length of bars in the concrete 0.20
sj=b
core [9] is necessary to confirm this assumption.
0.00
3. Stress block approach for strength calculation (a) 0 2 4 6
ε/εcc
Under the hypotheses of plane section and perfect bond
1.1
between steel and concrete, the calculation of the flexural capacity fc,core=30 MPa
of a square RC jacketed section for a generic neutral axis depth can
1.0
be written by means of the following equilibrium equations fc,core=20 MPa
(Fig. 3):
0.9
C j þ C c þ F 0j þ F 0c þ F c þ F j ¼ N ð12aÞ fc,core=10 MPa

β
0.8
C j  dj þ C c  dc þ F 0j ðxc  cj Þ þ F 0c ðxc  d  cc Þ þ F c ðb þ d  xc  cc Þ
  sj=b/2
B 0.7
þ F j ðB  xc  cc Þ ¼ M  N  xc ð12bÞ
2 sj=b

where 0.6

C j ¼ aj bj  f c;jacket xc B  ðbj xc  dÞbaj f c;jacket ð13Þ

(b) 0 2 4 6
ε/εcc
compressive force in the concrete jacket Fig. 4. Stress-block parameters for concrete of the core as a function of axial strain
(d/b = 0.33, cj/d = 0.5).
C c ¼ ac bc  f cc;core ðxc  dÞb ð14Þ

compressive force in the concrete core

distance of the resultant compressive force in the concrete core
F j ¼ rsj Asj ¼ cj  f yj  Asj F 0j ¼ r0sj A0sj ¼ c0j  f yj  A0sj ð15Þ
from the neutral axis
rsj r0
forces in steel of the jacket, being cj ¼ and c0j ¼ f yjj It has to be noted that when bjxc < d the second terms in Eqs.
f yj
(13) and (17) has to be set equal to zero, and furthermore if
F c ¼ rsc Asc ¼ cc  f yc  Asc F 0c ¼ r0c A0sc ¼ c0c  f yc  A0sc ð16Þ xc < d Eqs. (14) and (18) are also equal to zero.
forces in steel of the core, being cc ¼ rf ycsc and c0c ¼ f yc
r0
c Once that the constitutive law of concrete in compression is
  defined, the stress block parameters a and b have to be calculated
bx
B  bj  xc xc  j2 c to be used for the calculation of the flexural capacity of the jack-
dj ¼  ðbj xc  dÞ  b eted section.
B  bj  xc  b  ðbj xc  dÞ
  For the generic known value of maximum strain, the stress-
ðb x dÞ
xc  d  j 2c block parameters can be found from taking the first and second
 ð17Þ moments of area of the stress–strain law expressed by Eq. (11).
B  bj  xc  b  ðbj xc  dÞ
The following expressions result:
distance of the resultant compressive force in the concrete jacket R ec
0
rc dec
from the neutral axis ab ¼ ð19aÞ
f c ec
R ec
ac bc rc ec dec
dc ¼ xc  d  ðxc  dÞ ð18Þ b¼22 0
R ec ð19bÞ
2 ec 0 rc dec

εcuj α j f cc,jacket
ε'sj ε*co σ'sj
βj xc α c f cc,core
xc σ'sc
ε'sc βc (xc-δ)

εsc σsj
εsj σsj

Strain Jacket (concrete Core (concrete

and steel) stress and steel) stress

Fig. 3. Stress block approach for the evaluation of failure condition in a RC jacketed section.
166 G. Minafò / Engineering Structures 85 (2015) 162–169

1.00 assumed in both tension and compression). The corresponding

fc,core=15 MPa neutral axis has the depth equal to:

0.90
fc,core=20 MPa ecuj
xc;1—2 ¼ c ð20Þ
ecuj þ eyj j
where ecuj is the ultimate compressive strain of the concrete of the
αβ

0.80 fc,core=30 MPa

jacket and eyj is the yield stress of longitudinal bars in the jacket.
Ultimate strain profile 2 is characterized from upper steel
0.70 yielded in compression, with stress equal to the yield strength.
The following neutral axis depth holds

0.60 ecuj
xc;2—3 ¼ c ð21Þ
0 0.1 0.2 0.3 0.4 0.5 0.6 ecuj  eyj j
ωstj
The following steel stress ratios results
0
1.10 c ¼ 1; cj ¼ 1;
j ð22Þ

(
if je0sc j > eyc ! c0c ¼ signðe0sc Þ
1.05 0
fc,core=30 MPa c ¼
c e0sc Es ;
if je0sc j < eyc ! c0c ¼ f yc
fc,core=20 MPa ( ð23Þ
1.00 if jesc j > eyc ! cc ¼ signðesc Þ
β

cc ¼ if jesc j < eyc ! cc ¼ efsc Es

fc,core=15 MPa
yc

0.95 where esc and e0 sc are the strains in the top and bottom steel of the
core, which can be evaluated with the following expressions:

0.90 d þ cc  xc;23 d þ b  cc  xc;23

0 0.1 0.2 0.3 0.4 0.5 0.6 e0sc ¼  ecuj esc ¼   ecuj ð24Þ
xc;23 xc;23
ωstj
The strain profile 3 presents bottom steel stress equal to the
Fig. 5. Stress-block parameters at failure for the core concrete as a function of the yield stress. The neutral axis depth is equal to
geometrical ratio of jacket’s transverse reinforcement.
ecuj
xc;34 ¼  ðB  cj Þ ð25Þ
ecuj þ eyj
Fig. 4 shows the variation of the stress-block parameters as a func-
tion of the normalized strain for the core concrete. The normalized The steel stress ratios are
8
jacket’s thickness is d/b = 0.33 while the cover-to-jacket thickness < if je0sj j > eyj ! c0j ¼ signðe0sj Þ
ratio is cj/d = 0.5. Curves are drawn for three different compressive c0j ¼ e0sj Es cj ¼ 1 ð26Þ
strengths of concrete and for two values of stirrup’s pitch, equal : if je0sj j < eyj ! c0j ¼ f yj
respectively to b and b/2. As it could be noted, both parameters
depend mainly from the concrete strength and low variation can (
if je0sc j > eyc ! c0c ¼ signðe0sc Þ
be observed from the amount of transverse reinforcement. More- 0
c ¼
c e0sc Es
over the first parameter (ab) reaches a constant value after the peak if je0sc j < eyc ! c0c ¼ f yc
( ð27Þ
strain, while the second parameter (b) tends to become constant if jesc j > eyc ! cc ¼ signðesc Þ
near the ultimate strain. cc ¼ if jesc j < eyc ! cc ¼ escf ycEs
Fig. 5 shows the variation of the stress-block parameters at fail-
ure as a function of the geometrical ratio of stirrups. It has to be being the steel strains equal respectively to
noted that also at failure the stress-block parameters are a function
of the unconfined concrete strength; the variation is more marked xc;34  cj d þ cc  xc;34
e0sj ¼  ecuj e0sc ¼   ecuj esc
for the first parameter and for higher concrete strengths. xc;34 xc;34
d þ b  cc  xc;34
¼  ecuj ð28Þ
xc;34
4. Definition of the m–n interaction domain In every case the strain at the top fiber of the core can be calcu-
lated as follows:
As discussed above, the evaluation of the axial force-bending
xc  d
moment domain for RC jacketed sections is a difficult task. Most eco ¼  ecuj ð29Þ
xc
of studies proposed to calculate the load-carrying capacity of a
jacketed section subjected to axial force and bending moment by The domain is finally completed with two further points (0 and
means of complex algorithms which require the use of a computer 4), which corresponds respectively to the theoretical cases of pure
software. The proposed model is based on the determination of the tensile and pure compressive axial force. The minimum axial force
M-N interaction domain with five points, the latter to be calculated is calculated as the yield tensile force carried out by longitudinal
with simple relationships on the basis of the ultimate conditions of bars, while the maximum is equal to the concrete axial capacity
the considered section, as analogously done for un-retrofitted RC of the column in correspondence of the peak strain value.
columns. Fig. 6 shows the assumed strain profiles at failure. In par- Fig. 7 shows the comparison between the M–N interaction
ticular profile 1 corresponds to failure due to tensile axial force and domain of an unreinforced member and those relative to jacketed
bending moment, being all steel bars yielded in tension in both sections with normalized jacket thicknesses d/b equal to 0.13 and
core and jacket (cj = c0 j = cc = c0 c = 1 if elastoplastic behavior is 0.33 respectively. The ratio between the concrete compressive
 G. Minafò / Engineering Structures 85 (2015) 162–169 167

ε yj ε yj
ε cuj
εccj N
4

1
3

2
2 3 4
M
1
0

Fig. 6. Proposed model for calculation of strength domains of RC jacketed sections.

R ec
5 N/ fc,core) eex rc ec dec
Unreinforced secon b¼22 R ec ð30bÞ
ec eex rc dec
4 Jacketed secon
However, further studies will be addressed on this aspect to clarify
3 δ/b=0.33 analytically and experimentally on the effect of the axial preloading
on the overall flexural capacity of the reinforced member.
2
δ/b=0.13
5. Comparisons with numerical analyses and experimental data
1
The proposed model is validated with experimental data avail-
0 able in the literature [7] and with numerical analyses carried out
0.00 0.20 0.40 0.60 0.80 1.00 1.20
with the software SAP2000 [16]. This software is chosen because
-1 M/ fc,core ) it allows one complete modeling of the analyzed case study, and
additionally it is one of the most diffused computer program
Fig. 7. Simplified strength domains for RC jacketed sections.
worldwide for structural analysis. In particular the ‘‘Section
Designer’’ package allows to analyze complex cross-sectional sec-
tions with user-defined features by means of the classic fiber
strengths of core and jacket is equal to fc,core/fc,jacket = 0.57, the geo-
method. For the examined case the jacket was divided in 100
metrical ratio of longitudinal steel in the core is equal to (Asc + A0 sc)/
square fibers while the confined region was modeled with 400
b2 = 1%, while the same for the jacket is equal to 2%. The transverse
square cells. Rebars were considered as points, and overlapping
reinforcement ratio is equal to xstc = 2% for the core and xstj = 6%
with the square concrete cells was considered by neglecting the
for the jacket. Axial force and bending moment values are normal-
single cell coincident with a bar location. Constitutive laws of con-
ized with respect to b2fc,core and b3fc,core respectively, in order to
fined, unconfined concrete and steel were preliminary calculated
stress the effect of retrofitting on the strength of the original
and introduced in the software as user-defined laws by points.
column. It can be noted as the capacity is noticeably increased,
The software takes advantage of a step-by-step numerical algo-
especially for higher values of axial force. Furthermore the enlarge-
rithm (Newton-Raphson) for the solution of the non-linear system
ment of the jacket’s thickness induces the enhancement of the
to calculate the interaction domain. The required precision is
flexural capacity, the latter increasing for great values of axial
achievable by setting the number of points defining the domain.
force. Large thicknesses are therefore required only when the
In the present analysis the number of points was assumed equal
column is subjected to great axial load values. For low levels of
to 200.
normal force, that is a common case for buildings in seismic
Fig. 8 shows the comparison between the analytical results
regions, thin jacket’s thicknesses could be more convenient,
obtained with the proposed model and those computed numeri-
although this solution would require the use of special admixtures
cally. The case study refers to a column with side b = 300 mm,
with reduced coarse aggregate size.
cover thickness cc = 20 mm, having concrete compressive strength
Particular care has to be addressed to the effect of axial preload-
ing on the existing column. As discussed by Del Rio Bueno [15] the
effective enhancement of the structural performances by concrete 8000 N [kN]
jacketing strictly depends on the existing axial load. In particular, if Numerical (SAP2000)

the axial load value corresponds to the peak compressive strain of Analycal (proposed model)
6000
unconfined concrete or greater, a very small increase of the axial
capacity and a limited enhancement of ductility could be obtained. δ/b=0.33
Therefore the proposed method has to be applied only when the 4000
axial load on the inner column is lower than that corresponding δ/b=0.17
to the peak strain of unconfined concrete. Proposed stress block 2000 δ/b=0
parameters could take into account of the existing axial shortening
eex of the column, simply by changing the integration limits in 0
Eq. (19) as it follows: 0 100 200 300 400 500 600
R ec M [KNm]
rc dec -2000
ab ¼ eex ð30aÞ
f c ec
Fig. 8. Comparison between numerical and analytical results.
168 G. Minafò / Engineering Structures 85 (2015) 162–169

2000 N [kN] 8 mm diameter at pitch 100 mm and yield stress 450 MPa. These
Numerical (SAP2000) features are chosen in order to simulate a poor existing RC columns
Analycal (proposed model) designed only for gravity loads and reinforced with the RC jacket-
1500
ing technique.
Three analysis case are plotted, referring to the unreinforced
1000 Experimental section (d/b = 0), and the two limit cases of jacket’s thickness
(Specimen MM - Ersoy et al.)
suggested [6], equal respectively to d/b = 0.17 and d/b = 0.33. Good
500 accordance can be noted between the numerical and the analytical
solution, highlighting also that proposed method is slightly
conservative with respect to the numerical model due to the
0
0 20 40 60 80 100 120 stress-block approximation. The comparison stresses again that
M [KNm] the increase of the strength enhancement with the jacket’s
-500 thickness is negligible for low levels of axial force.
A further comparison is shown in Fig. 9, which shows the inter-
2500
N [kN] action domains obtained theoretically (numerical and analytical
Numerical (SAP2000)
solutions) together with the experimental results determined in
2000 Analycal (proposed model) [7]. The column has b = 160 mm, cover thickness cc = 5 mm and it
is reinforced with four longitudinal bars having diameter 12 mm.
1500 The RC jacket has thickness equal to d = 35 mm, cover thickness
cj = 5 mm and four longitudinal bars with diameter 12 mm.
1000 Experimental - Specimen SR The yield strength of longitudinal reinforcement is equal to
(Ersoy et al.1999)
fyc = 300 MPa and fyj = 280 MPa for core and jacket respectively.
500 Stirrups in the core have 4 mm diameter and pitch equal to
100 mm, while stirrups in the jacket have 8 mm diameter and
0 pitch 100 mm. The concrete compressive strength varies for each
0 20 40 60 80 100 120
case analyzed and further details can be found in [7].
-500 M [KNm]
Also in this case the analytical solution achieved with the
proposed model fits the numerical solutions with good accuracy
2500 N [kN] and with differences less than 5%. In all examined cases the exper-
Numerical (SAP2000)
imental result is close to the theoretical predictions with the
2000 Analycal (proposed model)
exception of specimen MM, where a difference can be observed.
This deviation could be addressed to different aspects concerning
1500 the test specimen, such as the effective strength of materials
(especially steel), since the two theoretical approaches (numerical
1000 Experimental - Specimen MR
and analytical) lead to similar results. However it has to be noted
(Ersoy et al.1999)
that the result is quite conservative with respect to safety in all
500
examined cases. From this preliminary verification the model
could be considered as an useful tool for design purposes of RC
0
0 20 40 60 80 100 120
jacketed columns. Further experimental investigation should be
M [KNm] addressed to verify in deep the suitability of the model.
-500

2500 N [kN] 6. Conclusions

Numerical (SAP2000)

2000 Analycal (proposed model)

In this paper, a simplified analytical method is presented able to
calculate the strength domains for RC jacketed columns. The model
1500 is based on the determination of some characteristic points of the
interaction domains, and it is based on the stress-block approach.
1000
Experimental - Specimen RM From derived results and from comparisons with numerical analy-
(Ersoy et al.1999) ses carried-out with a commercial software (SAP2000), the follow-
500
ing conclusions can be drawn:
0
0 20 40 60 80 100 120  the stress-block approach is applicable to RC jacketed sections if
M [KNm] parameters are well-calibrated;
-500
 for common confinement levels, stress block parameters are
Fig. 9. Comparison between analytical, numerical solutions and experimental mainly a function of concrete compressive strength and their
results of Ersoy et al. [7]. value tends to be constant after that peak strain is reached;
 results derived with the proposed method are in good accor-
dance with those obtained numerically. Also comparisons with
equal to fc,core = 18 MPa, reinforced with four longitudinal bars hav- experimental data have shown good agreement, even if with a
ing diameter 12 mm and yield strength equal to fyc = 280 MPa. The limited number of comparisons. Further experimental investi-
RC jacket is supposed to have concrete compressive strength gations need to be addressed to verify the applicability of this
fc,jacket = 25 MPa, cover thickness cj = 25 mm and four longitudinal approach;
bars with diameter 12 mm placed at the corners with yield stress  the proposed model allows one to calculate strength domains in
fyj = 450 MPa. Stirrups of the core have 8 mm diameter, pitch equal easy manner (calculation by hand) for RC jacketed sections. A
to 200 mm and yield stress 280 MPa; stirrups of the jacket have combination of this method with the use of ‘‘monolithicity
 G. Minafò / Engineering Structures 85 (2015) 162–169 169

coefficient’’ and safety factors considering shrinkage effects, [8] Takeuti AR, Bento de Hanai JB, Mirmiran A. Preloaded RC columns
strengthened with high-strength concrete jackets under uniaxial
could represent an useful tool for practical engineering
compression. Mater Struct 2008;41:1251–62.
applications. [9] Campione G, Fossetti M, Giacchino C, Minafò G. RC columns reinforced with RC
jackets. Mater Struct 2013. http://dx.doi.org/10.1617/s11527-013-0146-x.
[10] Thermou GE, Papanikolau VK, Kappos AJ. Cyclic response of R/C jacketed
columns including modelling of the interface behaviour. In: Proceedings of the
References 15th WCEE, World Congress on Earthquake Engineering, Lisbon; 2012.
[11] Mander JB, Priestley MJN, Park R. Theoretical stress–strain model for confined
[1] Vandoros KG, Dritsos SE. Concrete jacket construction detail effectiveness concrete. J Struct Eng ASCE 1988;114(8):1804–26.
when strengthening RC columns. Constr Build Mater 2008;22:264–76. [12] Karthik MM, Mander JB. Stress-block parameters for unconfined and confined
[2] Julio ES, Branco F, Silva VD. Concrete-to-concrete bond strength. Influence of concrete based on a unified stress–strain model. ASCE J Struct Eng
the roughness of the substrate surface. Constr Build Mater 2004;18(2004): 2011;137(2):270–3.
675–81. [13] Collins MF, Mitchell D. Prestressed concrete structures. Englewood Cliffs,
[3] Altun F. An experimental study of the jacketed reinforced-concrete beams NJ: Prentice-Hall; 1994.
under bending. Constr Build Mater 2004;18:611–8. [14] Dhakal RP, Maekawa K. Modelling of post-yield buckling of reinforcement.
[4] Branco F, Julio ES, Silva VD. Reinforced concrete jacketing—interface influence ASCE J Struct Eng 2002;128(9):1139–47.
on cyclic loading response. ACI Struct J 2008;105(4):471–7. [15] Del Río Bueno A. Aportaciones al Refuerzo de Estructuras de Hormigón
[5] Lampropoulos AP, Dritsos SE. Modeling of RC columns strengthened with RC Armado de Edificación. Ph.D. dissertation, Universidad Politécnica de Madrid.
jackets. Earthquake Eng Struct Dyn 2011;40:1689–705. Madrid, September 1987 [only available in Spanish].
[6] Penelis GG, Kappos AJ. Earthquake resistant concrete structures. E&Spoon; [16] Computers and Structures, SAP2000: Integrated Software for Structural
1997. p. 572. Analysis and Design CSI, ver. 14, Berkeley, California, USA; 2010.
[7] Ersoy R, Tugrul TA, Suleiman U. Behavior of jacketed columns. ACI Struct J
1993;90:3.