Magazine of Concrete Research, 2004, 56, No.

2, March, 89–98

Jacketing of reinforced concrete columns

subjected to axial load
K. C. G. Ong,* Y. C. Kog,* C. H. Yu* and A. P. V. Sreekanth†

National University of Singapore; CPG Consultants Pte Ltd

This paper investigates the suitability of two existing confined concrete models, based on the concept of an
effectively confined concrete core, to predict the behaviour of jacketed reinforced concrete columns when subjected
to axial loads. Two types of jacketing with reinforced concrete (RC) were studied: full-length and partial jacketing.
Consideration is also given to variations in the properties of the different concretes used in the jacket and the
original column. Existing experimental data show that jacketing of RC columns using reinforced concrete enhances
the axial load-carrying capacity, strain at peak axial load and post-peak ductility. However, conventional methods
of analysis underestimate both the peak axial load and the corresponding strain. The two models were used to
predict the entire load–strain relation of jacketed RC columns subjected to axial loading. The predictions agree
well with available test results. For columns subjected to patch jacketing with only one set of stirrups, the predicted
peak loads showed good agreement with experimental results. However, the corresponding strains were over-
estimated by as much as 20%.

fcc compressive stress in core concrete (MPa)

Notation
fci unconfined compressive strength of concrete
Acr area of core (mm2 ) in core (MPa)
Aic area of inner cover (mm2 ) f c9i cylinder compressive strength of concrete in
Aij area of inner jacket (mm2 ) core (MPa)
Alcr area of longitudinal reinforcement in core fco unconfined compressive strength of concrete
(mm2 ) in jacket (MPa)
Alj area of longitudinal reinforcement in jacket f c9o cylinder compressive strength of concrete in
(mm2 ) jacket (MPa)
Amcr area of confined core area at critical section fcon compressive strength of confined concrete
(mm2 ) (MPa)
Aoj area of outer jacket (mm2 ) fcr yield strength of stirrups in core (MPa)
Ascr area of single leg of stirrup in core (mm2 ) fic compressive stress in concrete in inner cover
Asj area of single leg of stirrup in jacket (mm2 ) (MPa)
Bcr breadth of core (mm) fij compressive stress in concrete in inner jacket
Ccr distance between adjacent braced longitudinal (MPa)
bars in core (mm) fj yield strength of stirrups in jacket (MPa)
Cj distance between adjacent braced longitudinal foj compressive stress in concrete in outer jacket
bars in jacket (mm) (MPa)
Dcr width of core (mm) fscr stress in longitudinal reinforcement in core
(MPa)
fsj stress in longitudinal reinforcement in jacket
* Department of Civil Engineering, National University of Singapore, (MPa)
Block E1A, #07–03, 1 Engineering Drive 2, Singapore 117576. fycr yield strength of longitudinal reinforcement in
† CPG Consultants Pte Ltd, Civil & Structural Engineering Division, core (MPa)
#11–00, 238B, Tower B, Novena Square, Singapore 360238.
fyj yield strength of longitudinal reinforcement in
(MCR 1097) Paper received 23 October 2002; last revised 23 July jacket (MPa)
2003; accepted 10 October 2003 K total strength gain factor
89

0024-9831 # 2004 Thomas Telford Ltd

Ong et al.

Kcore strength gain factor for core strength gain and the increased post-peak ductility of the
Ke1 ratio of confined core area to total core area jacketed columns. Even though research has been car-
for inner stirrups ried out on the axial capacity, flexural strength and
Ke2 ratio of confined core area to total core area seismic resistance of jacketed RC columns, very limited
for outer stirrups work has been carried out to analytically quantify the
KS1 strength gain factor for core due to inner above-mentioned effects of jacketing. This paper inves-
stirrups tigates the suitability of two existing confined concrete
KS2 strength gain factor for core due to outer models, based on the concept of an effectively confined
stirrups concrete core, to predict the behaviour of jacketed col-
m rise of parabola inside inner core umns when subjected to axial loads. Two types of jack-
Pcr unconfined capacity of core (kN) eting were studied: full-length and partial jacketing.
si spacing of inner stirrups (mm) Consideration is also given to variations in the properties
so spacing of outer stirrups (mm) of the different concretes used in the jacket and the
xcr base length of parabola at periphery of core original column. Though the proposed method of ana-
(mm) lysing jacketed columns may not be as simple to use as

strain in general those for composite columns, the use of computer

co strain at unconfined compressive strength of spreadsheets or other programming methodology may
concrete in jacket be used to simplify the procedure to a large extent.

s1 ,
s2 , values of strain controlling shape of stress–

85 ,
0:3 strain curve
º ratio of effectively confined core to total core
Previous research
area at level of stirrups
1
rcr volumetric ratio of stirrups in core Aksan tested three RC columns subjected to axial
rj volumetric ratio of stirrups in jacket loads only. Two columns (hereafter referred to as the
original columns) were loaded to 70% of their ultimate
axial load capacity. One of them was jacketed after
unloading (specimen AU), whereas the other was
Introduction
jacketed while subjected to an axial load (specimen
There is a constant need to upgrade or repair existing AL). The results of these tests were compared with a
structures to suit new requirements. Currently, rein- similar but monolithic column (specimen AM) tested
forced concrete (RC) jacketing is one of the most under axial loading. It was observed that both jacketed
commonly used techniques for the strengthening of columns (AU and AL) performed as efficiently as the
columns. This type of jacketing refers to the provision monolithic column (AM), and specimen AL seemed to
of a concrete layer with reinforcement (longitudinal be unaffected by the fact that it was jacketed while
and lateral) around an existing column to enhance load- being subjected to load. Jacketing improved the
carrying capacity. This may be due to inadequate de- strength, stiffness and post-peak ductility of the col-
sign, substandard construction, functional changes in umns when compared with the performance of the
the service of the structure, changes in design require- original columns.
4
ments, additional loading or the need to increase lateral Tests reported by Mohanlal involved nine RC col-
load resistance of the structure. Jackets may also be umns subjected to patch jacketing and tested under
provided to restore the original load-carrying capacity axial loading to failure. At the central one-sixth portion
of a column after damage or deterioration with time. of the column, patch jacketing without enlargement of
Besides reinforced concrete, other types of materials column section was carried out. This was done by
are also available for jacketing. However, reinforced blocking out the central one-sixth portion of the col-
concrete enjoys advantages in the form of economy, umn specimens and exposing the longitudinal bars dur-
compatibility with the original concrete substrate, and ing casting of the column. ‘Repairing’ was carried out
the ability to enhance durability and impart fire protec- by casting the blocked-out portion using either the
tion. The main disadvantages of RC jacketing include same mix as the original column or two types of
the loss of floor space due to enlargement of the col- commercially available repair material. The study con-
umn cross-section, and difficulties that may be experi- cluded that the ultimate strength of the specimens was
enced in casting and compacting the jackets. not affected as long as the repair material used in the
In spite of the widespread use of jackets for strength- patch repair had a strength that was the same as or
1–3
ening columns, existing codes of practice do not higher than that of the original concrete. Failure always
provide adequate guidelines for the analysis or design of occurred at the section with the lowest axial capacity,
such columns. Conventional design methods analyse in either the unrepaired or the repaired zones. The
jacketed columns as a composite of old and new con- patch jacket, when applied while under load, was mon-
crete along with the longitudinal steel present therein. itored and showed that it was mobilised only to carry
Such analysis typically underestimates the actual additional loads over and above the load that was ap-
90 Magazine of Concrete Research, 2004, 56, No. 2
 Jacketing of RC columns subjected to axial load

plied while the repair was carried out. When the patch stirrups present in the original column are also strained,
jacket was applied after unloading the column, the generating considerable additional confining pressure.
jacket and the original concrete contributed their re- In this manner, RC jacketing leads to the development
spective shares in resisting the axial loads applied to of two sets of confining pressure around the concrete
the composite column. in the original column, one from the original stirrups
A study on the effect of bond on eccentrically loaded and the other from additional stirrups introduced by
jacketed reinforced columns shows that the bond be- jacketing. Thus it would be more appropriate to employ
tween the old and new concrete did not affect the models for confined concrete to predict the behaviour
capacity of the column as long as the original column of such jacketed columns, especially of the concrete in
5
and the jacket were loaded simultaneously. An analysis the original column.
6
in line with BS 8110, without considering any con- Previous studies show that circular spirals confine
finement effect of the stirrups, was used to estimate the concrete more effectively than rectilinear stirrups, and
ultimate loads of the jacketed RC columns. The pre- the mechanism of confinement afforded by circular
8
dicted loads showed good agreement with the experi- spirals is well understood. However, the present dis-
mental results, but no data were provided on the cussion is limited to jacketing of non-circular columns.
ultimate moment capacity. Research on the effective- In the case of non-circular stirrups, the confining pres-
ness of concrete jacketing as a method for seismic sure afforded varies in three dimensions, and the chal-
retrofit shows that it increases the strength, ductility, lenge lies in accurately estimating this variation.
stiffness and energy dissipation of both damaged and Researchers have used different approaches to estimate
undamaged columns when subjected to simulated seis- the strength increase arising from the provision of stir-
2,3,7
mic loading. rups. Various stress–strain relations have been pro-
8
posed, and a comparative study by Sheikh reported
Modelling of reinforced concrete jacketing that most of the analytical models are empirical in
Jacketed RC columns comprise three load-bearing nature and are effective only when used on the test
components in the axial direction: the original column, results or data from which they are derived. Models
9 10
the new concrete (jacket), and the longitudinal steel in proposed by Sheikh and Uzumeri and Mander et al.
both the original column and the jacket. A comparison used the concept of effectively confined core area to
of available experimental results using composite account for the variation in confining pressure and to
6
analysis in accordance with BS 8110 shows that the quantify the strength increase arising from confine-
behaviour of jacketed RC columns is different from ment. These models are suitable for modelling the be-
that obtained when taking into account the composite haviour of jacketed RC columns.
action of the three load-bearing components mentioned
above. For instance, analysis in line with BS 8110
6 Effectively confined core area method
underestimated the axial load capacity by up to 20% Owing to the geometric configuration of the stirrups,
1
for specimen AM in the study by Aksan. Jacketed RC the confining pressure within the jacket varies. Only
columns also experienced 40% higher axial deforma- the part of the concrete that is confined effectively is
tion at peak load than the values predicted. If the jacket capable of carrying additional load. The rest of the
behaved purely as a composite section, the increase in concrete is unaffected by the confining action of the
9
strains at peak loads reported experimentally would not jacket. Sheikh and Uzumeri proposed that, at the level
be expected, and the peak load should occur at a strain of the stirrups, the confined and unconfined sections of
close to that of a similar column, but without a jacket. the original column may be demarcated by a series of
On the other hand, the apparent gain in both the peak parabolas spanning between the longitudinal bars, with
load and corresponding strain appears logical when a rise 0.273 times their span. Parabolas may also be used
the concrete in the original column is considered as to demarcate the confined and unconfined portions of
confined concrete. This is further augmented by the the original column in the longitudinal direction. The
post-peak ductility reported in tests on jacketed RC latter may be assumed to have a rise equal to a quarter
columns, which cannot be fully accounted for by con- of their span (stirrup spacing). By thus defining the
ventional composite action. This type of behaviour is geometry of the parabolas, it was possible to calculate
similar to that of concrete confined by stirrups at a the area of the effectively confined concrete at the
high volumetric ratio. The lateral steel within the jacket critical longitudinal section that lies midway between
provides the confining force (through Poisson’s effect) two adjacent sets of stirrups, where the confining effect
countering the tendency of the concrete in the original is a minimum. From an experimental investigation,
9
column to expand laterally under axial load. This lat- Sheikh and Uzumeri obtained the strength gain of the
eral confining force, which acts all around the original confined concrete in terms of the ratio of effectively
column, changes the stressed state present in the con- confined concrete area, the volumetric ratio and the
10
crete to that of a triaxial one. As the axial and the yield strength of stirrups. Mander et al., on the other
corresponding lateral deformations experienced by typi- hand, calculated the enhancement in the strength of the
cal jacketed RC columns are very high, the original concrete by relating the ratio of confined concrete area
Magazine of Concrete Research, 2004, 56, No. 2 91
Ong et al.

to the total concrete area to an effective value that stirrups, rcr is the volumetric ratio, and fcr is the yield
reflects the observed three-dimensional variation dis- strength of the stirrups in the original column.
cussed above. The strength increase was estimated The confining effect at the level of the stirrups aris-
using a multi-parameter failure surface for the concrete ing from the outer stirrups can be represented by para-
derived from triaxial tests. bolas originating from the longitudinal bars supporting
In the analysis presented herein, a typical cross- the outer stirrups. Fig. 2(a) shows two possible arrange-
section of a jacketed RC column is divided into four ments of these parabolas.
regions, taking into consideration the confinement
effect afforded. These four regions are the core, the (a) If these parabolas fall outside the inner core, then
inner cover, the inner jacket, and the outer jacket (Fig. the confining effect will be a maximum. The ratio
1). The original column encompasses the core and the of confined core area to the inner core area may be
inner cover. The core is enclosed by the centreline of taken as unity. That is:
the inner stirrups, and the inner cover comprises the º¼1 (2)
rest of the original column, outside the core. The inner
(b) If these parabolas intersect the inner core, the ratio
jacket refers to the portion of the jacket that is enclosed
º may be expressed as
by the centreline of the outer stirrups. The outer jacket
comprises the rest of the jacket outside the outer stir- X 2mxcr
Acr

rups. The core of a jacketed RC column is confined by 3
both the inner and outer sets of stirrups, whereas the º¼ (3)
Acr
inner cover and the inner jacket are confined only by pffiffiffiffiffiffiffiffiffi
the outer stirrups. The outer jacket is not confined. where xcr ¼ 1:912 Cj m is the base length of the
Enhancement in the strength of a jacketed RC column parabola at the periphery of the core, m is the rise
may thus be estimated in the following two ways. of the parabola inside the core, Acr is the area of
the core, and Cj is the centre-to-centre distance
a Original column
b Jacket
Method based on Sheikh and Uzumeri’s c Longitudinal bars
model d Stirrups
e Parabola representing confining action of outer stirrups
Applying the principles of the effectively confined
core area and the relations proposed by Sheikh and
9
Uzumeri, the strength gain factor for the core arising
d
from the inner stirrups, KS1, may be calculated as e
follows:
0 1
X e
2
B C cr C
Bcr Dcr @ A
K S1 ¼ 1 þ 1
:
140Pcr 5 5Bcr Dcr
a c a



si si : b b
3 Bcr
Dcr
(rcr f cr )0 5 (1)
2 2 (a)

where Bcr and Dcr are the dimensions of the concrete

core, Pcr is the unconfined capacity of the original
column (kN), Ccr is the distance between adjacent long-
itudinal bars in the original column, si is the spacing of
d
e

Inner cover
Core

Inner jacket
a b c b a

Outer jacket (b)

Fig. 2. Two possible arrangements of parabolas representing

the effectively confined core with respect to the outer stirrups:
Fig. 1. Four regions in a typical cross-section for a jacketed (a) at the level of the stirrups (plan view); (b) between two
column adjacent levels of stirrups (elevation)
92 Magazine of Concrete Research, 2004, 56, No. 2
 Jacketing of RC columns subjected to axial load

between two adjacent longitudinal bars in the sive strength, and
0:3 is the strain corresponding to
jacket. 0:3K core f ci in the descending branch of the stress–
strain curve. Fig. 3 shows, schematically, the stress–
Considering the variation of the confined core area in strain curve for confined concrete used by Sheikh and
the longitudinal direction, the critical section may be 9
Uzumeri, in which fcon is the compressive strength of
identified to be that midway between two adjacent confined concrete. The stress–strain curve between
s2
outer stirrups. and
0:3 may not be a straight line, and
0:3 serves as a
Similarly, two different configurations of the para- transition point if needed.
bola are possible, as shown in Fig. 2(b). The strength gain factors for the inner cover and
(a) If the vertical parabola falls outside the inner core, inner jacket regions may also be calculated in a similar
the outer stirrups do not give rise to confinement manner and the respective stress–strain relations de-
effects. The effectively confined area at the critical fined accordingly. The full details of the equations for
section, Amcr , in the core may be given by the strength gain factors and the respective stress–
strain relations for the inner cover and inner jacket
Amcr ¼ ºBcr Dcr (4) regions are given in Reference 11. The stress–strain
(b) If the vertical parabola extends into the inner core, relation for concrete in the unconfined outer jacket
then the effectively confined inner core area at the may be modelled using the following equation pro-
12
critical section may be taken to be posed by Desayi and Krishnan:


  
so so 2 f co

Amcr ¼ º Bcr

2m Dcr

2m (5) f oj ¼ (9)
2 2
co 1 þ (
=
co )2

where so is the spacing of the outer stirrups. where fco is the compressive strength of the concrete in
the jacket and
co is the corresponding strain.
The strength gain factor for the core arising from the
Thus, for an imposed strain, the load carried by each
outer stirrups is thus
region can then be calculated as the product of the
Amcr : stress developed in that region and its area. At any
K S2 ¼ 1 þ (rj f j )0 5 (6)
140Pcr section, the total load carried by a jacketed column is
where rj is the volumetric ratio and fj is the yield the sum of the loads carried by the four different
strength of stirrups in the jacket. regions of the column cross-section and the longitudi-
The total strength gain factor for the core, Kcore , is nal steel in the original column and jacket. Thus load,
obtained by adding KS1 and KS2 : P, for an imposed strain of
, is given by

K core ¼ K S1 þ K S2 (7) P ¼ f cc Acr þ f ic Aic þ f ij Aij þ f oj Aoj þ f scr Alcr þ f sj Alj

The relationship of stress (fcc ) against strain (
) for (10)
the confined concrete present in the core can be ob- where fic and fij are the stresses developed in the inner
tained from Kcore in the following equations: cover and inner jacket, and fscr and fsj are the stresses
"
2 # 9
2

> developed in the longitudinal steel in the core and
For 0 ,
,
s1 : f cc ¼ K core f ci
>
>
>
> jacket respectively. Acr , Aic , Aij and Aoj are the areas of

s1
s1 >
>
>
> the core, inner cover, inner jacket and outer jacket
=
For
s1 ,
,
s2 : f cc ¼ K core f ci respectively. Alcr and Alj are the areas of the longitudi-
>
> nal steel in the original column and the jacket. In this
0:15K core f ci >
For
s2 ,
,
0:3 : f cc ¼ (

0:85 f ci
85 ) >
>
>

85

s2 >
>
>
;
For
0:3 ,
: f cc ¼ 0:3K core f ci
Stress

(8a)
where
s1 ,
s2 and
85 are values of strains that specify
the shape of the assumed stress–strain curve given by
fcon
9

s1 ¼ 80K core f co 3 10
6 >
>
" > 0·85fcon

2 # >
>

s2 248 s rscr f scr >
>
¼1þ 1
5 pffiffiffiffiffiffi =

co Ccr Bcr f co (8b) 0·3fcon
>
>
rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi >
>
Bcr >
>

85 ¼ 0:225rscr þ
s2 >
;
s εs1 εs2 εs85 Strain

where fcc is the stress developed in the core concrete Fig. 3. Schematic stress–strain curve for confined concrete
9
for any imposed strain
, fci is the unconfined compres- proposed by Sheikh and Uzumeri
Magazine of Concrete Research, 2004, 56, No. 2 93
Ong et al.

manner, the load–strain curve for the entire range of involved, using the method described in this paper. The
loading can be modelled using Sheikh’s model. results may be plotted in the form of a design chart, as
shown in Fig. 4. This presents values of K for different
sizes of the original columns, jacket thicknesses, con-
Method based on Mander’s model crete covers and stirrup strengths, for a range of concrete
strength in the original column and jacket and for three
To model a jacketed column using Mander’s typical volumetric ratios of stirrups. In Fig. 4, the inner
10
model, the effectively confined area of the core, inner and outer stirrups have the same diameter and spacing.
cover and inner jacket may also be calculated as de- Similar charts for different diameters and stirrup spa-
scribed above. For the core, arising from the inner cings could be plotted. The load-carrying capacities of
stirrups, Ke1 may be defined as the ratio of the effec- jacketed RC columns without considering effects of
tively confined area to the total area of the core region. confinement are also shown in Fig. 4. Thus the use of a
Ke2 is the same arising from the outer stirrups. The chart like the one shown in Fig. 4 would enable a
confining pressure, fle , arising from the two sets of designer to calculate unconfined strengths and the va-
stirrups acting on the core can be calculated as lues of K, for use if effects of confinement are to be
2 f cr Ascr 2 f j Asj considered for jacketed RC columns. The axial load-
f le ¼ K e1 þ K e2 (11)
si Ccr so C j carrying capacity of a jacketed RC column is given by
the product of K and its unconfined strength. Note that
where Ascr and Asj are the areas of a single leg of the the contribution of longitudinal steel present in the ori-
stirrups in the original column and jacket respectively. ginal column and jacket needs to be taken into account
Details for calculating the confined strength for a given to obtain the total axial load-carrying capacity. The
confining pressure and for obtaining the stress–strain method of calculation is shown in the illustrative exam-
relation of the concrete in the core can be obtained ple given in the appendix.
from Reference 11. Thus the load corresponding to an
imposed strain,
, can be calculated as described above
together with the load–strain curve. Comparison with experimental results
The above discussion indicates that the maximum
load-carrying capacity of jacketed RC columns depends Columns with full jacketing
on various parameters, which makes the process of The model described above was used to analyse the
1
analysis more involved than that of a normal column. three columns tested to failure by Aksan. Details of
The values of K, defined as the ratio of the load- the three columns are given in Table 1. The cross-
carrying capacity of a jacketed RC column with con- section of each column was divided into the four re-
finement effects to that of one without, can be found by gions, and the equations presented earlier were used to
selecting the appropriate combination of the parameters calculate the strength gain factor for each region and

1·54 2500
Size of core ⫽ 200 mm Capacity without considering confinement: kN
Jacket thickness ⫽ 50 mm
1·49
Yield strength of stirrups ⫽ 250 MPa
1·44 2000

1·39

1·34 1500

1·29
K

1·24 1000

1·19

1·14 500

1·09

1·04 0
10 15 20 25 30
fci: MPa

fco ⫽ 30 MPa, ρcr ⫽ 0·035 fco ⫽ 40 MPa, ρcr ⫽ 0·035 fco ⫽ 30 MPa, ρcr ⫽ 0·015

fco ⫽ 40 MPa, ρcr ⫽ 0·015 fco ⫽ 30 MPa, ρcr ⫽ 0·035 fco ⫽ 40 MPa, ρcr ⫽ 0·0035

Capacity without considering confinement for fco ⫽ 40 MPa

Capacity without considering confinement for fco ⫽ 30 MPa

Fig. 4. Design curves for jacketed RC columns

94 Magazine of Concrete Research, 2004, 56, No. 2
 Jacketing of RC columns subjected to axial load
1
Table 1. Analytical and experimental results for columns tested by Aksan
Specimen Cylinder strength of concrete: MPa Ratio of predicted peak load to experimental value

Core Jacket BS 8110 Mander Sheikh

AM 28 28 0.80 0.85 0.89
AL 29 20 0.87 0.90 0.98
AU 28 14.3 0.89 0.89 0.96
Size of basic column 130 mm 3 130 mm; size of jacketed column 180 mm 3 180 mm. Main bars ( f ycr , f yj ¼ 280 MPa): original column 4 3
10 mm; jacket 4 3 10 mm. Stirrups ( f cr , f j ¼ 280 MPa): original col-

the stress–strain relationships for the concrete. The 1000 Experimental

Based on Sheikh and Uzumeri
analysis was achieved by imposing an incremental Based on Mander et al.
strain on the column, assuming strain compatibility. Based on BS 8110
The corresponding axial force contributed by each re- 750
gion and the force in the reinforcing steel were added
to obtain the total axial force. Figs 5–7 show the

Load: kN
predicted load–strain curves of specimens AM, AL and 500
AU respectively using BS 8110 and Mander’s and Shei-
kh’s models together with the experimental results. The
predicted peak loads of the three models are compared 250
with the experimental peak loads in Table 1.

1500 Experimental 0
Based on Sheikh and Uzumeri 0 0·01 0·02
Based on Mander et al. Strain
Based on BS 8110
Fig. 7. Comparison of experimental and predicted load–
1000 strain curves for specimen AU
Load: kN

500 It is clear from a comparison of the experimental

results and the analytical predictions that Sheikh’s mod-
el predicts the peak load of the jacketed columns accu-
rately (Table 1). Mander’s model is more conservative,
0 and BS 8110 is the most conservative. The similarity in
0 0·01 0·02
Strain
the initial stiffness of the columns predicted using the
three models indicates that the strength enhancement
Fig. 5. Comparison of experimental and predicted load– due to confinement effects occurs largely during the
strain curves for specimen AM later stages of loading, when the stirrups are more
stressed. Note also that Mander’s model provides a
better estimate of the strain at peak load than the other
Experimental two models. Sheikh’s and Mander’s models are also
1200
Based on Sheikh and Uzumeri
Based on Mander et al.
consistent in predicting increased ductility in the post-
1000 Based on BS 8110 peak load (descending) branch of the axial load–strain
curve. BS 8110 underestimates post-peak ductility as
800 the ultimate value of the concrete strain is capped at
0.0035.
Load: kN

600 Confinement models thus confirm the experimental

observation that jacketed RC columns performed much
400 like an equivalent monolithic RC column under static
axial loading irrespective of the presence or absence of
200 loading of the original column when jacketing is car-
ried out. The three analytical models used predict very
0 similar behaviour for both specimen AU and specimen
0 0·01 0·02
Strain
AL. This could be consistent with the fact that both
were subjected to a load equivalent to 70% of the
Fig. 6. Comparison of experimental and predicted load– original column’s axial load capacity prior to jacketing.
strain curves for specimen AL Specimen AU was then unloaded, whereas the load
Magazine of Concrete Research, 2004, 56, No. 2 95
Ong et al.

applied on AL was sustained throughout the jacketing attributed to the fact that only one set of stirrups was
operation. Their behaviour could in all likelihood show present, with less significant confinement effect, as the
differences if higher levels were used. stirrup volumetric ratio is relatively small. However,
In this paper, comparisons made with three columns the strain corresponding to the peak load was overesti-
1
tested by Aksan were presented in detail, showing mated by both Mander’s and Sheikh’s models (Table 2).
good agreement with predictions for the three types of A comparison of the two possible axial load–strain
jacketed column. More comparisons were made with plots mentioned above helped in identifying the critical
other data, as detailed in Reference 11. However, these section at which failure is more likely. The location at
do not provide full information on the properties of which failure is most likely to occur (either in the
materials used and the complete load–deflection/strain monolithic section or the ‘repaired’ section), predicted
curves of the jacketed columns tested; only the ultimate as shown in Table 2, is in close agreement with the
failure loads were available and some of the material experimental results.
properties are available.

Columns with patch jacketing

4 Parametric study
Columns tested by Mohanlal, as summarised in
Table 2, were modelled in similar manner. In this case For rectangular columns, the confinement effect will
only three regions—the core, inner jacket and outer be significant only if the intermediate (not at the cor-
jacket—were involved. The core comprises that part of ners) longitudinal bars present are able to provide
the column cross-section that was first cast using the enough support to the stirrups to restrain lateral expan-
parent material, whereas the inner jacket comprises the sion. In the absence of such restraint confinement
region that lies between the centreline of the stirrup action is negligible, with no significant strength en-
and the core. The core and the inner jacket were mod- hancement effects. The authors are not aware of litera-
elled to take into account the effects of confinement by ture available on the role that longitudinal bars play
the stirrups. The outer jacket is unconfined. The unre- with regards to confinement. However, studies on
paired sections of the column were modelled as mono- jacketed masonry columns have shown that intermedi-
lithic, with the same material properties being assumed ate longitudinal bars in the jacket did not significantly
for all three regions. In the repaired sections, the core enhance the axial load-carrying capacity of masonry
11
was modelled using the properties of the parent con- columns if they were not braced by cross-links. Thus
crete, whereas the jacket regions were modelled using the parametric study reported herein is limited to
the properties of the add-on repair material used. Two square columns reinforced longitudinally with only cor-
different axial load–strain curves could be plotted, one ner bars. The total strength gain factor, K, does not
for the repaired section and the other for the unre- include the contribution of the longitudinal steel, as this
paired, monolithic section. Table 2 shows the predicted is more useful in assessing the effectiveness of confine-
values based on BS 8110 and Mander’s and Sheikh’s ment for the column section.
models for the critical sections at which failure would The effects of shrinkage of the outer core could have
probably take place. It can be seen that the predicted a strengthening effect on the inner core. However, the
peak loads based on BS 8110 and Mander’s and shrinkage properties of the repair materials used were
Sheikh’s models are relatively close. This could be largely unreported in the available literature. Also, as

4
Table 2. Analytical and experimental results for columns tested by Mohanlal
Specimen Cube strength: Ratio of predicted strains at peak Ratio of predicted peak Failure section/
MPa load to experimental value load to experimental value predicted
failure
Parent Repair BS 8110 Mander Sheikh BS 8110 Mander Sheikh section
concrete material
AR1 37.20 37.20 1.59 1.07 1.41 0.88 0.89 0.93 U/U
AR2 37.20 36.79 1.52 1.22 1.48 0.86 0.89 0.91 U/U
AR3 41.16 38.16 1.75 1.45 1.80 1.00 1.02 1.05 U/R
AS1 42.68 58.13 1.30 0.96 1.37 1.04 1.05 1.08 U/U
AS2 44.00 61.81 1.40 1.12 1.48 0.93 0.95 0.96 U/U
AS3 39.88 54.23 1.40 1.16 1.44 0.95 0.99 1.00 U/U
AT1 42.29 43.26 0.92 0.82 1.00 0.96 1.00 1.02 R/U
AT2 42.33 38.52 1.00 0.83 1.06 1.05 1.06 1.09 U/R
AT3 38.51 40.91 1.46 1.29 1.42 1.04 1.08 1.10 U/U
Size of column 200 mm 3 200 mm 3 2000 mm; size of unrepaired section 110 mm 3 110 mm 3 200 mm. Main bars ( f ycr ¼ 596 MPa) 4 3
10 mm; stirrups ( f cr ¼ 330 MPa) 6 mm at 100 mm c/c

96 Magazine of Concrete Research, 2004, 56, No. 2

 Jacketing of RC columns subjected to axial load

this effect depends upon a large number of parameters, 1·6

Core 200 mm,
including method of surface preparation of substrate Cover 25 mm
460 MPa stirrups
concrete, method of casting repair material, amount of f ′cj ⫽ 30 MPa, f ′ci ⫽ 20 MPa, ρcr ⫽ 0·03
ρj ⫽ ρcr

restraint, and time duration, this effect was not incorpo-

1·4
rated in the modelling of jacketed columns in the pre-
sent study.

K
Typical results for a jacketed RC column with 2% f ′cj ⫽ 40 MPa, f ′ci ⫽ 20 MPa, ρcr ⫽ 0·03

longitudinal reinforcement ratio are plotted in Fig. 4. f ′cj ⫽ 30 MPa, f ′ci ⫽ 20 MPa, ρcr ⫽ 0·003
1·2
The following observations may be made.

(a) K increases with increasing volumetric ratio of f ′cj ⫽ 40 MPa, f ′ci ⫽ 20 MPa, ρcr ⫽ 0·003

stirrups as the latter leads to an increase in the 1·0

confining pressure afforded by the stirrups. This is 50 75 100
Jacket thickness: mm
evident from Fig. 4. The rate of decrease in K is
greater for an increase in the spacing of the outer Fig. 8. Variation of total strength gain factor, K, with respect
stirrups when compared with the case when the to jacket thickness
spacing of the inner stirrups is increased by the
same amount. This may be attributed to the fact
that a variation in the spacing of the inner stirrups
only affects the axial load-carrying capacity of the 1·6
Jacket thickness 50 mm, cover 25 mm
core whereas variation in the spacing of the outer f ′cj ⫽ 30 MPa, f ′ci ⫽ 20 MPa; ρj ⫽ ρcr
stirrups affects the core, inner cover and the jacket.
(b) Figure 4 shows that, for the same volumetric ratio
ρcr ⫽ 0·03
of stirrups, K decreases with increasing concrete 1·4
strength in the original column. This trend is in
line with the experimental results showing that
K

ρcr ⫽ 0·015
confinement is more effective for concrete of a
lower structural grade. For the same reason, the 1·2
effectiveness of confinement decreases with an in-
crease in the grade of structural concrete used in
the jacket. The variation of K with respect to the
concrete grade in the original column is more 1·0
200 250 300
pronounced for higher volumetric ratio of stirrups. Size of core: mm
This is illustrated in Fig. 4, where the curves corre-
sponding to a volumetric ratio of 0.03 slope more Fig. 9. Variation of total strength gain factor, K, with respect
steeply than those for a volumetric ratio of only to width of core
0.003. Studies have shown that the rate of decrease
in confinement effect is more pronounced at higher
13
magnitudes of confining pressures. Hence the ob- as stirrups with a higher strength are capable of
served variation discussed above is to be expected. providing more confining pressure.
(c) K reduces slightly when the jacket thickness is
increased. The strength gain achieved by the core
is reflected only in the slope of the post-peak Conclusions and recommendations
(descending) branch of the load–strain curve for Within the scope of the present study the following
the jacketed RC column. Typical variations in conclusions may be made.
K with respect to jacket thickness are shown in
Fig. 8. (a) Existing experimental data show that jacketing of
(d) K increases with an increase in the size of the RC columns using reinforced concrete enhances
original column when the other parameters are the axial load-carrying capacity, strain at peak ax-
kept constant. As the original column size in- ial load and post-peak ductility. However, conven-
creases, for a particular volumetric ratio and stirrup tional methods of analysis assuming composite
diameter, the stirrup spacing reduces. Conversely, action between the original column and jacket,
for a particular spacing, the stirrup diameter in- such as BS 8110, underestimate both the peak axial
creases. Both result in an increase in the confining load and the corresponding strain.
pressure, reflected in an increase in the value of K. (b) The two analytical models for jacketed RC col-
Fig. 9 illustrates this for two different volumetric umns presented (Sheikh’s model and Mander’s
ratios of stirrups. model) take into consideration the effects of con-
(e) K increases with the yield strength of the stirrups, finement and may be used to predict the entire
Magazine of Concrete Research, 2004, 56, No. 2 97
Ong et al.

stress–strain relation of jacketed RC columns sub- Thus K ¼ 1.21.

jected to axial loading. Given that 2% longitudinal steel of yield strength
(c) The predictions agree well with available test re- 460 MPa is provided in the original column and 1% in
sults of jacketed RC columns subjected to axial the jacket, the axial load-carrying capacity of the
loading. Sheikh’s model was able to predict peak jacketed RC column is obtained as
axial loads that agree within 10% of the experi-
Pu ¼ (1:21 3 1 742 000) þ 460 3 [(0:02 3 200 3 200)
mental results. Mander’s model was more conser-
vative in predicting the peak axial load but was þ 0:01 3 (300 3 300
200 3 200)] N
more accurate in predicting the strains correspond-
ing to the peak axial load. For columns subjected ¼ 2706 kN (without considering partial
to patch jacketing with only one set of stirrups, the safety factors for materials)
predicted peak loads using the three models pre- or 1936 kN (considering partial safety factors
sented herein showed good agreement with experi- for materials)
mental results. However, for these columns, which
were reinforced with only a single layer of stirrups,
the corresponding strains were overestimated by References
20% on average. 1. Aksan B. Jacketed Column Behaviour Under Axial Load Alone.
(d ) A parametric study conducted on square-jacketed MS thesis, Middle East Technical University, 1988.
RC columns showed that the total strength gain 2. Rodrieguez M. and Park R. Seismic load tests on reinforced
concrete columns strengthened by jacketing. ACI Structural Jour-
factor of such columns increased with increasing
nal, 1993, 91, No. 2, 150–159.
volumetric ratio of stirrups in the original column 3. Eursoy U., Tankut A. T. and Suleiman R. Behaviour of
or jacket, size of original column and yield jacketed columns. ACI Structural Journal, 1993, 90, No. 3, 288–
strength of the stirrups. The total strength gain 293.
factor decreased with increasing grade of structural 4. Mohanlal K. Performance of Repaired Axially Loaded Mem-
bers. MEng thesis, National University of Singapore, 1998.
concrete strength in the original column or jacket
5. Duangjaras C. Strengthening of reinforced concrete columns.
and with the thickness of the jacket. ACECOMS News & Views, 1999, April–June, 4, 5, 17.
6. British Standards Institution. Code of Practice for Design
and Construction for Reinforced Concrete. BSI, Milton Keynes,
Acknowledgement 1997, BS 8110.
7. Bracci J. M., Reinhorn A. M. and Mander J. B. Seismic
The financial assistance received and research facil- retrofit of reinforced concrete buildings designed for gravity
ities provided by the National University of Singapore loads: performance of structural model. ACI Structural Journal,
are gratefully acknowledged. 1995, 92, No. 6, 711–727.
8. Sheikh S. A. A comparative study of confinement models. ACI
Structural Journal, 1982, 79, No. 7, 296–306.
Appendix: Illustrative example 9. Sheikh S. A. and Uzumeri S. M. Analytical model for concrete
confinement in tied columns. Journal of Structural Engineering,
A 200 mm square column is to be provided with a ASCE, 1982, 108, No. 12, 2703–2722.
reinforced concrete jacket of thickness 50 mm. The 10. Mander J. B., Priestly J. N. and Park R. Theoretical stress–
strain relationship for confined concrete. Journal of Structural
unconfined concrete strength of the original column
Engineering, ASCE, 1988, 114, No. 8, 1804–1825.
( f ci ) is 15 MPa and that of the jacket ( fco ) is 40 MPa. 11. Sreekanth A. P. V. Behaviour of Jacketed Reinforced Concrete
The volumetric ratios of inner and outer sets of stirrups and Masonry Columns. MEng thesis, National University of
are both 0.015. The amount of longitudinal steel in the Singapore, 2000.
original column is 2% of gross area and that of the 12. Desayi P. and Krishnan S. Equation for stress–strain curve of
concrete. Journal of American Concrete Institute, 1964, 61, No.
jacket is 1% of gross jacket area. Determine the axial
3, 345–350.
load capacity of the jacketed RC column. 13. Saatcioglu M. and Razvi S. R. Strength and ductility of
From Fig. 4, the axial load capacity without consid- confined concrete. Journal of Structural Engineering, ASCE,
ering confinement effects and contribution from long- 1992, 118, No. 6, 1590–1607.
itudinal steel ¼ 1742 kN.
The total strength gain factor may be obtained from
the curve plotted in Fig. 4, corresponding to rcr ¼ Discussion contributions on this paper should reach the editor by
0.015 and fco ¼ 40 MPa. 1 September 2004

98 Magazine of Concrete Research, 2004, 56, No. 2