Construction and Building Materials 23 (2009) 2265–2276

Contents lists available at ScienceDirect

Construction and Building Materials

journal homepage: [Link]/locate/conbuildmat

Axially loaded RC columns strengthened by steel caging. Finite element modelling

Jose M. Adam a,*, Salvador Ivorra b, Francisco J. Pallarés c, Ester Giménez a, Pedro A. Calderón a
a
ICITECH, Departamento de Ingeniería de la Construcción y Proyectos de Ingeniería Civil, Universidad Politécnica de Valencia, Camino de Vera s/n, 46071 Valencia, Spain
b
Departamento de Ingeniería de la Construcción, Obras Pública e Infraestructura Urbana, Universidad de Alicante, Apartado de Correos 99, 03080 Alicante, Spain
c
Departamento de Física Aplicada, Universidad Politécnica de Valencia, Camino de Vera s/n, 46071 Valencia, Spain

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

Article history: Reinforced concrete (RC) columns in buildings often need strengthening either due to defects in the col-
Received 25 August 2008 umns themselves, having to support higher loads than those foreseen in the initial design of the structure,
Received in revised form 21 November 2008 or as the result of ageing or accidental damage. The use of steel caging for this purpose is now a common
Accepted 24 November 2008
practice in many countries throughout the world. Based on the results of an experimental study, this
Available online 1 January 2009
paper presents a parametric study using finite element models carried out with the aim of analysing
the behaviour of RC columns strengthened by steel caging. The results of the study are used to analyse
Keywords:
the influence that various parameters have on the behaviour of the strengthened column (size of the
RC columns
Strengthening
angles, the yield stress of the steel of the cage, the compressive strength of the concrete in the column,
Retrofitting the size of the strips, the addition of an extra strip at the ends of the cage, the friction coefficient between
Steel caging the layer of mortar and the steel of the cage). The results obtained from the parametric study allow a ser-
Numerical analysis ies of guidelines to be established.
Ó 2008 Elsevier Ltd. All rights reserved.

1. Introduction of the strengthening techniques in Spain. The data included in Figs.

1 and 2 were compiled by Adam [8] after an exhaustive review of
It is often necessary to strengthen RC columns in a building the bibliography and a survey of 73 technical specialists in foren-
either due to defective construction, or because higher loads than sics and strengthening of structures. As Figs. 1 and 2 illustrate,
those foreseen in the initial design are imposed to the structure the percentage of published articles concerning steel jacketing is
(possibly due to a change in building use) or as a result of accidents very small, especially when compared to how much this is actually
such as seismic loads. used in Spain. These data clearly indicate that there is a need for
Three principal methods are available for column strengthen- research into the behaviour of RC columns strengthened by steel
ing: concrete jacketing, steel jacketing and composite jacketing cages, since this is currently the most used steel jacketing variant
(FRP). Steel caging is a variant of the second category and is known [10].
to be an easily applied and economical strengthening technique Among the strengthening with steel cages, there are different
[1]. The method involves the use of longitudinal angle sections variants which provide a solution to the area nearest the ends of
fixed to the corners of the column, to which transverse steel strips the column:
are welded [2,3]. The space between cage and column is filled with
cement or epoxy mortar. At the present time it is a common prac- (a) Adding capitals welded to the steel cage so that they are in
tice [4] in countries such as the Czech Republic [5], Japan [6], contact with the beam, in a similar way to the specimens
Greece [7] and Spain [8]. As Wu et al. [7] have shown, this type studied by Ramírez [11], Ramírez and Bárcena [12], and
of strengthening is fully effective in increasing the strength and Ramírez et al. [13]. This ensures a direct transmission of
ductility of RC columns. CEB-FIB [9] also confirms the effectiveness loads to the strengthening [14,15].
of this strengthening technique. (b) Welding tubes to the angles of the strengthening, passing
Although the use of steel cages is widespread and highly effec- through the beam–column joint. This variant was proposed
tive, there has been little research into RC columns strengthened initially by Fernández [16] with the aim of ensuring the
by this technique. Fig. 1 compares the percentage of articles pub- transmission of loads between two sections of strengthened
lished in scientific journals relating to the most commonly used column. This variant has also been studied by Adam [8] and
strengthening techniques. Fig. 2 shows the percentage use of each Adam et al. [15].
(c) Not having any additional element at the ends of the stre-
* Corresponding author. Tel.: +34 963877562; fax: +34 963877568. ngthening. This variant coincides with one of the analyzed
E-mail address: joadmar@[Link] (J.M. Adam). by Giménez [10] and Giménez et al. [17], and is similar to

0950-0618/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/[Link].2008.11.014
2266 J.M. Adam et al. / Construction and Building Materials 23 (2009) 2265–2276

Nomenclature

Ac cross-section area of concrete Ns load supported by steel cage

As cross-section area of reinforcement PExp ultimate load obtained from experimental study
AL cross-section area of steel angles PFEM ultimate load obtained from FE models
a contact cohesion according to Coulomb’s frictional law P load applied by the hydraulic testing machine
COV coefficient of variation p contact normal pressure
Eci elastic modulus of concrete bc shear transfer coefficient (closed cracks)
Eco 2.15e104 MPa according to CEB-FIB Model Code 90 bt shear transfer coefficient (open cracks)
Es elastic modulus of steel n1 parameter which takes into account the effectiveness of
fc compressive strength of concrete the strengthening
fcmo 10 MPa according to CEB-FIB Model Code 90 n2 parameter which takes into account the effectiveness of
ft tensile strength of concrete the strengthening
fyL yield stress of steel cage l friction coefficient
fys yield stress of reinforcement steel slim limit shear stress
Nc load supported by column concrete t poisson ratio

the specimens studied by Dolce et al. [18] and Cirtek [5,19]. was 35 mm. It should be emphasised that the reinforcement used
However, in the latter, the strengthening was welded to was the minimum permitted under Spanish regulations [20] for RC
steel plates located at the ends of the column. Consequently, columns and is very close to most international codes [21,22]. The
these specimens did not reflect the true behaviour of a reinforcement of the heads was designed with the objective of
strengthened column [8,10]. avoiding interruption of the tests by early failure of this compo-
nent, as had occurred in the tests carried out by Ramírez [11],
This paper studies the behaviour of strengthened columns in Ramírez and Bárcena [12], and Ramírez et al. [13]. The columns
those cases where the ends of the strengthening are worked on were strengthened by L80.8 angles (leg size 80  80 mm and thick-
using variant (c) described above. The Institute of Concrete Science ness 8 mm) and rectangular strips measuring 270  160  8 mm3
and Technology (ICITECH) at the Technical University of Valencia is and 270  100  8 mm3. The steel grade was Fe430 [23] with a
at present researching the behaviour of RC columns strengthened yield stress of 275 MPa.
by this variant. Following the experimental study carried out by The concrete mix used in the columns was designed to simulate
Giménez [10], all the laboratory-tested specimens are modelled a column with low compressive strength in need of strengthening.
by the finite element method (FEM). After validating the FE models, Compressive strength was determined by the cylindrical specimen
a parametric study is carried out which analyses the behaviour of test carried out at the same time as the tests on the strengthened
RC columns strengthened by steel caging. columns. It should be pointed out that high strength concrete
(fc = 90 MPa) was used for the heads at both ends of the specimens
2. Summary of the experimental study to avoid failure due to stress concentration in the zones near to the
load application points. The cement mortar between cage and col-
The experimental study on axially loaded RC columns strength- umn had a cement/sand weight ratio of 1:2.
ened by a steel cage was carried out in the ICITECH laboratories of In order to measure strain and displacement in the steel cage
the Technical University of Valencia. The tested specimens were and column concrete, a minimum of 14 strain gauges and eight
considered to represent a full scale column in an actual building. LVDTs were attached to each of the specimens tested.
Total length of each specimen was 3100 mm. The columns were The tests were carried out in a steel frame and the axial load
2500 mm long with a cross-section of 300  300 mm2. The speci- was applied by a hydraulic testing machine with a maximum
mens had 300  300  600 mm3 concrete heads at both ends of capacity of 5000 kN. Load was applied in displacement control
the column, simulating the beam–column joint. mode at a constant rate of 0.5 mm/min. This load was applied until
The reinforcement of the column consisted of four 12 mm failure of the specimen. In Fig. 3 a specimen can be seen in the steel
diameter longitudinal rods with 6 mm diameter cross ties every frame ready for testing.
0.20 m. The steel yield stress was 400 MPa and the concrete cover Five types of specimens were used in the experiments, and two
specimens of each type were tested in the laboratory, giving a total
of 10 tests. The differences between the five types of specimen
Concrete jacketing were in the geometry of the cage and also in the strength of the
Steel jacketing 0.4%
1.2%
Composite jacketing-FRP
Concrete jacketing 8%
33%

Composite jacketing-FRP
98.4% Steel jacketing
59%
Fig. 1. Percentage of research papers related to each strengthening technique of RC
columns. Fig. 2. Percentage use in Spain of each strengthening technique for RC columns.
 J.M. Adam et al. / Construction and Building Materials 23 (2009) 2265–2276 2267

Table 1
Specimens studied. Comparison between test and finite element analysis.

Specimen fc (MPa) PExp (kN) PFEM (kN) PExp/PFEM

Exp-test 11.9 1352.3 1373.6 0.98
Exp-A 8.3 1954.8 1862.2 1.05
Exp-B 12.4 2324.1 2233.8 1.04
Exp-C 15.5 2599.4 2568.3 1.01
Exp-D 8.3 2451.9 2402.1 1.02
Mean – – – 1.02
COV – – – 0.026

(b) Yielding of the strips, due to the pressure caused by the tr-
ansverse deformation of the concrete under Poisson’s effect.

3. Finite element model

3.1. General approach

The specimens involved in the tests (see Section 2) were mod-

elled by the FEM using ANSYS 11.0 [24]. In order to simulate as clo-
sely as possible the actual behaviour of the specimens, the
numerical modelling took into account:

(a) The nonlinear behaviour of the steel used in the cage and in
the concrete reinforcement.
(b) The nonlinear behaviour of the concrete, using a constitutive
model that considered the behaviour of the confined
concrete.
Fig. 3. Specimen ready for testing. (c) The contact at the interface between the steel cage and the
mortar/concrete.
concrete used in the column. One of the specimens tested was not (d) The possible buckling of the steel cage and second-order
strengthened (Exp-Test) and was used as a control specimen. The effects.
other types of specimen tested can be observed in Fig. 4. Table 1
gives the compressive strength of the concrete in each specimen.
It should be emphasised that specimen Exp-D was equipped with 3.2. Description of the finite elements used, boundary conditions and
additional strips at both ends of the cage with the aim of improving loads applied
confinement in these zones. The results obtained from these tests
were analysed in depth by Giménez [10]. It is worth noting, how- All the specimens tested in the experiments had three planes of
ever, that from these tests, two possible mechanisms have been symmetry, so that for the numerical modelling only 1/8 of the
observed which could lead to failure of the strengthened column: specimen was considered (see Fig. 5), applying symmetry condi-
tions in the corresponding planes.
(a) Yielding of the angles, due to the axial loading absorbed, in The concrete and mortar were modelled using hexahedral ele-
combination with the bending produced by the transverse ments (SOLID65) with eight nodes and three degrees of freedom
deformation of the concrete in the column (Poisson’s effect). per node, allowing the treatment of nonlinear behaviour, including

a b
150 300 150 300 150 300 150 300

RC HEAD RC HEAD
160 20 300

220 100 90 160 20 300

10

10

RC COLUMN
RC COLUMN
STRIPS 270x100x8
410

STRIPS 270x160x8 STRIPS 270x160x8

160

160

ANGLES L80.8 ANGLES L80-8

410

410
160

160

Fig. 4. Specimens tested: (a) Exp-A, Exp-B and Exp-C and (b) Exp-D.
2268 J.M. Adam et al. / Construction and Building Materials 23 (2009) 2265–2276

The Poisson ratio used is t = 0.20, while the elastic modulus is

determined by following the recommendation of CEB-FIB Model
Code 90 [22]:

Eci ¼ Eco ðfc =fcmo Þ1=3 ð2Þ

where Eci is the elastic modulus, fc is the compressive strength, and
fcmo, Eco have values as defined in CEB-FIB Model Code 90 [22]:
fcmo = 10 MPa; Eco = 21.5 GPa.
No damage was detected in the cement mortar between cage
and column in any of the laboratory tests so, it was therefore mod-
Fig. 5. Modelling 1/8 of the specimen. elled assuming linear elastic behaviour with a Poisson ratio of
t = 0.20 and an elastic modulus of 25 GPa.
To consider the nonlinear behaviour of the steel in the concrete
cracking and crushing capabilities. The reinforcement was consid- reinforcement and steel cage, the well-known Von Mises yield cri-
ered to be smeared within the SOLID65 element [25]. terion is used with elastic perfectly-plastic behaviour. The elastic
The steel cage was modelled using quadrilateral shell elements modulus used for both types of steel is Es = 210 GPa, with
(SHELL181), with four nodes and six degrees of freedom per node. t = 0.30 Poisson ratio. The yield stress is fys = 400 MPa and fyL
This element is appropriate for modelling thin and moderately = 275 MPa, for the reinforcement and steel cage, respectively.
thick plate and shell elements, allowing the simulation of buckling.
Contact between the steel cage and mortar/concrete was simu- 3.4. Modelling of interaction between different materials
lated by the contact elements TARGE170 and CONTA174, which
allow a Coulomb friction model to be considered and produce a The contact between mortar layer and strips and between con-
gap between interface elements when tensile stresses are detected. crete and angles is modelled by contact elements. To allow for pos-
The contact algorithm used was the ‘‘Lagrange and Penalty Meth- sible relative slippage between surfaces, Coulomb’s friction model
od” available in ANSYS 11.0 [24]. For the ‘‘allowable tensile contact as expressed in Eq. (3) is considered.
pressure” and ‘‘penetration tolerance” parameters, values of 1.00
and 0.10, respectively, were used. This methodology meant the slim ¼ a þ lp ð3Þ
models have a good convergence, without adversely affecting the where slim is the limit shear stress, a is the contact adhesion, p is the
accuracy of the results obtained. contact normal pressure and l is the coefficient of friction.
The number of finite elements included in each model, and the For contact cohesion, a = 0 MPa is considered, while for the case
number of degrees of freedom, depend on the type of specimen. For of friction coefficient between steel and mortar/concrete the value
example, to model the specimen Exp-A, 7273 finite elements were l = 0.20 is adopted, following Johansson and Gylltoft [27,28] and
used, making a total of 32,241 degrees of freedom. Adam et al. [14,29].
The load applied by the hydraulic testing machine was consid-
ered as a controlled displacement on the concrete head. Each in-
4. Verification of the finite element model
crease in load meant a displacement of 0.05 mm Due to the
nonlinear nature of the models analysed, arc-length approach with
In order to verify the finite element model, a comparison be-
Full Newton–Raphson Method [24] were used to solve the system
tween the experimental results and finite element results was car-
of equations obtained.
ried out. It was first determined whether the ultimate load
obtained numerically (PFEM) matched the ultimate load obtained
3.3. Modelling of concrete, cement mortar and steel
in the experimental study (PExp). As can be observed in Table 1,
these ultimate loads have similar values. A maximum difference
During the process of loading the specimens, the concrete in the
of 5% was observed between experimental and numerical results.
columns becomes confined, since the steel cage prevents expan-
The mean value of PExp/PFEM is 1.02 and the corresponding coeffi-
sion of the concrete due to the Poisson effect. It is therefore neces-
cient of variation (COV) is 0.024, showing good agreement.
sary to consider a constitutive model that takes into account the
If the load-axial shortening curves of the experimental study
improved strength of the concrete due to confinement.
and the finite element models are compared for each of the speci-
As commented before, cracking and crushing capabilities have
mens, an excellent match can be observed. Fig. 6 shows the curves
been taken into account to properly model the concrete behaviour.
for specimens Exp-A and Exp-D.
The criterion used to separate the elastic from the inelastic behav-
iour is based on the work developed by Willam and Warnke [26]
for concrete under triaxial conditions in the tension and compres- 3000
sion regime. The two main strength parameters needed to define
the failure surface are the ultimate tensile strength (ft) and the ulti- 2500
mate compressive strength (fc). Regarding the shear transfer coef-
ficients, bt = 0.25 and bc = 0.75 are used for the open and closed 2000
Load (kN)

cracks, respectively. The reader is referred to Willam and Warnke

1500
[26] for more detailed explanations.
The tensile strength of concrete is obtained in Eq. (1) from fc, 1000 EXP (Exp-A)
which matches the one specified by CEB-FIB Model Code 90 [22] FEM (Exp-A)
500 EXP (Exp-D)
ft ¼ fctko;m ðfc =fcko Þ2=3 ð1Þ
FEM (Exp-D
where ft is the tensile strength, fc is the compressive strength and 0
0 5 10 15 20 25 30
fctko,m, fcko are the following parameters defined in CEB-FIB Model
Axial shortening (mm)
Code 90 [22]: fctko,m = 1.40 MPa; fcko = 10 MPa.
Fig. 6. Load versus axial shortening. Specimens Exp-A and Exp-D.
 J.M. Adam et al. / Construction and Building Materials 23 (2009) 2265–2276 2269

In order to determine whether the FE models are able to satis- (a) Specimen FEM-0. This specimen is defined as the pattern to
factorily reproduce the transmission of loads to the steel cage and be used as the standard of comparison for the influence of
also the confinement imposed on the concrete, the measurements variations in each of the analysed parameters. The geometry
recorded by a series of strain gauges placed on the steel cage (see and arrangement of the cage is identical to Exp-A, Exp-B and
Fig. 7) are compared with the measurements obtained from the FE Exp-C (see Fig. 4a). Since the aim is to simulate a column in
models. Figs. 8 and 9 (corresponding to specimens Exp-A and Exp-D) need of strengthening, the concrete compressive strength is
plot the stresses in the gauges versus the ratio P/PExp and P/PFEM considered to be 12 MPa.
(A1 and A2 quantify the transmission of loads to the cage; S1 (b) Specimens FEM-L40, FEM-L60, FEM-L100 and FEM-L120.
and S2 quantify the confinement imposed by the cage). When The only difference between these specimens and FEM-0 is
the curves that relate P/PExp (P/PFEM) with the stresses registered in the type of angles used in the cage. This provides a basis
at the reference points are compared, the experimental results for the analysis of the influence of the type of angles on
and those obtained from the FE models are seen to match closely. the behaviour of the strengthened column.
In Fig. 10, a comparison between experimental and numerical (c) Specimen FEM-fy235 and FEM-fy355. The only difference
results is made regarding behaviour patterns observed at the fail- between these specimens and FEM-0 is in the yield stress
ure of specimen Exp-B. As can be seen from this figure, the site of the steel cage elements. The yield stresses considered
of the failure in the FE models coincides with the actual site of are the standard according to Eurocode No. 3 [23].
the failure in the experimental tests. (d) Specimens FEM-fc4, FEM-fc20 and FEM-fc30. The only
difference between these specimens and FEM-0 is in the con-
5. Specimens analysed in the parametric study crete compressive strength of the column. With a compres-
sive strength of 4 MPa, the aim is to simulate a column
As has been explained in the previous section, the behaviour of with very low strength in need of strengthening. With a
a RC column strengthened by a steel cage can be predicted by compressive strength of 20 and 30 MPa, the aim is to simu-
means of FE models. It is therefore possible to carry out a paramet- late a column with acceptable strength but which, neverthe-
ric study with FE models following indications stated previously. less, needs strengthening due to a possible increase of the
This parametric study provides more detailed information on the loads considered in the initial design.
behaviour of an RC column strengthened by a steel cage, and also (e) Specimens FEM-St80, FEM-St120 and FEM-St200. These sp-
on the factors involved in this behaviour, so important conclusions ecimens are studied with the aim of determining the influ-
can be outlined. By simulating the failure of each of the specimens ence of the strip size on the behaviour of the strengthened
it will be possible to compare the ultimate load obtained from each column. The geometry of these specimens is shown in
model (PFEM) with the ultimate load proposed by different design Fig. 11a.
rules. The characteristics of the specimens analysed in the para- (f) Specimen FEM-aSt. The critical area in terms of the behav-
metric study are summarized in Table 2. Each specimen presents iour of the strengthened column is at the ends [5,8,10,14].
the following special features: For this reason, it may be advisable for the distance between

a b

S1 S1

A1 A1

A2 A2

S2 S2

Fig. 7. Position of the strain gauges: (a) specimen Exp-A and (b) specimen Exp-B.
2270 J.M. Adam et al. / Construction and Building Materials 23 (2009) 2265–2276

a 1.2 b 1.2

1 1

P/PExp (P/PFEM)
P/PExp (P/PFEM)

0.8 0.8

0.6 0.6

0.4 0.4
A1 S1
0.2 A2 0.2 S2
FEM FEM
0 0
-50 0 50 100 150 200 250 300 -50 0 50 100 150 200 250 300
Stress (MPa) Stress (MPa)

Fig. 8. Specimen Exp-A. Comparison of experimental and FEM results: (a) stress in A1 and A2 gauges versus P/PExp (P/PFEM) and (b) stress in S1 and S2 gauges versus P/PExp
(P/PFEM).

each strip in this area to be smaller. With the aim of analys- strengthened column are dealt with in this section. This table
ing the effect of this recommendation, an extra strip was shows the ultimate load obtained by the all the FE analyses
added to the ends of the column, as can be seen in Fig. 11b. performed.
(g) Specimens FEM-mu0.0 and FEM-mu0.6. With the aim of In order to determine the effectiveness of the strengthening,
improving load transmission between column and cage, a parameters n1 and n2 are defined using Eqs. (4) and (5),
layer of mortar fills the space between these two elements. respectively.
As indicated by Adam et al. [14,29], the friction coefficient
(l) between mortar and steel depends largely on the care P FEM
n1 ¼ ð4Þ
given to applying this layer of mortar. In specimens FEM- Ac  fc þ As  fys
mu0.0 and FEM-mu0.6 the effect of varying the value of l rL
n2 ¼ ð5Þ
on the behaviour of the steel-jacketed column is studied. fyL
For specimens FEM-mu0.0 and FEM-mu0.6, values of l =
0.05 and l = 0.60 were used. The FEM-mu0.0 specimen where Ac is the cross-section area of the RC column to be strength-
was used to simulate the case of a strengthening which ened, fc the compressive strength of the concrete, As the cross-sec-
had not been carried out with due care, or where the mortar tion area of the longitudinal reinforcement of the column, fys the
mix had not been right, causing a considerable degree of yield stress of the longitudinal reinforcement, rL the stress in the
shrinkage (impeding perfect contact between the mortar angles of the strengthening in the centre of the specimen (for a load
and the cage). The FEM-mu0.6 specimen simulated a case level equal to PFEM), and fyL the yield stress of the steel used in the
in which special measures had been taken to carry out the angles.
strengthening (e.g. use of special mortars which would guar- Parameter n1 is used to try to obtain an order of magnitude for
antee proper contact with the cage). the increase in strength caused by the strengthening, and also to
enable the relative importance of confinement in the concrete’s in-
creased strength to be quantified. Parameter n2 represents the
6. Influence of different parameters on the behaviour of the effectiveness of the steel cage, in terms of the load absorbed by
specimens the angles at the precise moment that the specimens fail. The val-
ues of both parameters (n1 and n2), for each of the FE models, have
6.1. General been included in Table 3.
On the FE models postulated in the parametric study, the two
Using the results given in Table 3 as reference, the influence of failure modes described at the end of Section 2 were detected.
each of the parameters considered on the behaviour of the The failure of the strengthened column occurs when the cage is

a 1.2 b 1.2

1 1
P/PExp (P/PFEM)
P/PExp (P/PFEM)

0.8 0.8

0.6 0.6

0.4 0.4
S1 S1
0.2 S2 0.2 S2
FEM FEM
0 0
-50 0 50 100 150 200 250 300 -50 0 50 100 150 200 250 300
Stress (MPa) Stress (MPa)

Fig. 9. Specimen Exp-D. Comparison of experimental and numerical results: (a) stress in A1 and A2 gauges versus P/PExp (P/PFEM) and (b) stress in S1 and S2 gauges versus
P/PExp (P/PFEM).
 J.M. Adam et al. / Construction and Building Materials 23 (2009) 2265–2276 2271

Likewise, when the cross-section area is increased, the transmis-

sion of loads between column and steel cage is also improved, as
can be seen in Fig. 13a. If angle cross-section area is increased,
the surface area of concrete covered by the angles also rises. This
implies that the effectiveness of the confinement imposed by the
cage increases with angle cross-section area (see n1 values included
in Table 3). Fig. 13b shows the compressive stresses on the column
in the axial direction at failure. As can be seen, as angle cross-sec-
tion area increases the failure stress of the concrete also increases
due to the confinement effect described. As regards effectiveness of
the cage, this presents inverse behaviour, that is to say, as cross-
section area increases; effectiveness is reduced, as can be seen
from the values of parameter n2 included in Table 3.

6.3. Yield stress of the cage steel

Obviously, the higher the yield stresses of the steel in the cage,
the greater the ultimate load of the strengthened column. How-
ever, as can be seen in Fig. 14a, a higher yield stress of the steel
in the cage has little effect on the transmission of loads between
the column and the cage. This effect is also confirmed by the values
of parameter n2, which decreases as the yield stress of the steel in
the cage increases (see Table 3).
The confinement effect imposed by the cage increases when the
yield stress of the steel is increased. This can be clearly observed in
Fig. 14b, which shows the compressive stress in the axial direction
of the column at failure. As regards parameter n1, its value
increases as the yield stress of the steel in the cage increases (see
Table 3).

6.4. Concrete strength

Fig. 10. Specimen Exp-B. Failure of concrete at the end of the specimen: (a)
experimental test and (b) finite element model (cracking and crushing). Varying the compressive strength of the concrete in the column,
besides affecting the ultimate load of the strengthened column,
also affects the load distribution between the column and the steel
no longer able to confine the concrete. In Fig. 12 the failure modes cage. As can be observed in Fig. 15a, the lower the strength of the
detected can be seen. In Table 3, the failure mode for each of the FE concrete in the column, the better the load distribution obtained.
models has been included. As can be seen in Fig. 15b, the relative importance of the confine-
ment is higher for low values of concrete strength. In relation to
6.2. Cross-section area of steel angles parameters n1 and n2 which allow the effectiveness of the strength-
ening to be estimated, both increase as the strength of the concrete
As would be expected, the higher the cross-section area of the decreases (see Table 3). The variation in n1 and n2 is in line with
angles, the greater the ultimate load of the strengthened column. what was observed in Fig 15a and b.

Table 2
Specimens analysed in parametric study.

Specimen Strengthening Material properties l

3
Angles Strips (mm ) fyL (Mpa) fc (MPa)
FEM-0 L 80.8 270  160  8 275 12 0.20
FEM-L40 L 40.4 270  160  8 275 12 0.20
FEM-L60 L 60.6 270  160  8 275 12 0.20
FEM-L100 L 100.10 270  160  8 275 12 0.20
FEM-L120 L 120.12 270  160  8 275 12 0.20
FEM-fy235 L 80.8 270  160  8 235 12 0.20
FEM-fy355 L 80.8 270  160  8 355 12 0.20
FEM-fc4 L 80.8 270  160  8 275 4 0.20
FEM-fc20 L 80.8 270  160  8 275 20 0.20
FEM-fc30 L 80.8 270  160  8 275 30 0.20
FEM-St80 L 80.8 270  80  8 275 12 0.20
FEM-St120 L 80.8 270  120  8 275 12 0.20
FEM-St200 L 80. 8 270  200  8 275 12 0.20
FEM-aSt L 80.8 270  160  8 275 12 0.20
FEM-mu0.0 L 80.8 270  160  8 275 12 0.05
FEM-mu0.6 L 80.8 270  160  8 275 12 0.60
2272 J.M. Adam et al. / Construction and Building Materials 23 (2009) 2265–2276

a 150 300 150 300 b 150 300 150 300

RC HEAD RC HEAD

St 20 300

160 200 160 50 160 20 300

10

10
RC COLUMN RC COLUMN

=
STRIPS 270xStx8 STRIPS 270x160x8

St
ANGLES L80.8 ANGLES L80.8

410
=

160
St
Fig. 11. (a) Specimens FEM-St80, FEM-St120 and FEM-St200 and (b) Specimen FEM-aSt.

Table 3
Results obtained from parametric study.

Specimen PFEM (kN) Effectiveness Failure mode

n1 n2
FEM-0 2185.7 1.73 0.63 Angles-first
FEM-L40 1611.2 1.28 1.33 Angles-first
FEM-L60 1842.7 1.46 0.85 Angles-first
FEM-L100 2582.5 2.05 0.51 Angles-first
FEM-L120 3065.8 2.43 0.48 Strips-first
FEM-fy235 2109.8 1.67 0.68 Angles-first
FEM-fy355 2349.4 1.86 0.56 Angles-first
FEM-fc4 1572.8 2.91 0.65 Angles-first
FEM-fc20 2944.6 1.49 0.57 Angles-first
FEM-fc30 3992.3 1.39 0.54 Angles-first
FEM-St80 1898.1 1.51 0.42 Angles-first
FEM-St120 1961.9 1.56 0.48 Angles-first
FEM-St200 2396.1 1.90 0.73 Angles-first
FEM-aSt 2678.3 2.12 1.05 Angles-centre
FEM-mu0.0 1823.8 1.45 0.19 Angles-first
FEM-mu0.6 2649.2 2.10 1.06 Angles-centre

Fig. 12. Failure modes observed in the FE models. Von Mises stresses (MPa): (a) yielding of the first of the angles (angles-first); (b) yielding of the centre angles (angles-
centre) and (c) yielding of the first of the strips (strip-first).

6.5. Strip size can be observed that greater strip size increases ultimate load. Both
the load transmission from column to cage and the confinement ef-
The size of the strips plays an important role in the behaviour of fect imposed by the cage are improved when the size of the strips is
the strengthened column. Comparing the ultimate load of speci- increased (see Fig. 16). In relation to this, the values for parameters
mens FEM-0, FEM-St80, FEM-St120 and FEM-St200 in Table 3, it n1 and n2 also increase with the size of the strips (see Table 3).
 J.M. Adam et al. / Construction and Building Materials 23 (2009) 2265–2276 2273

6.6. Including additional strips at the ends of the columns column, as well as increasing the effectiveness parameters n1 and
n2 (see Table 3).
Adding strips to the ends of the cage, as in the case of specimen As Fig. 17 illustrates, the presence of the additional strip con-
FEM-aSt, helps to increase the ultimate load of the strengthened tributes to a better transmission of the loads between the column

Nc/PFEM
a 1 0.8 0.6 0.4 0.2 0 b
1250 1250
FEM-0
FEM-L40
1000 1000 FEM-L60
FEM-L100
distance (mm)

distance (mm)
750 750 FEM-L120

12 MPa
500 FEM-0 500
FEM-L40
FEM-L60 250
250
FEM-L100
FEM-L120
0 0
0 0.2 0.4 0.6 0.8 1 0 5 10 15 20 25 30
Ns /PFEM Stress in concrete (MPa)

Fig. 13. Specimens FEM-0, FEM-L40, FEM-L60, FEM-100 and FEM-L120 (load applied coincides with PFEM): (a) load distribution between steel cage and column (Nc, load
supported by the column; Ns, load supported by the cage) and (b) compressive stress in concrete.

a 1 0.8 0.6
Nc/PFEM
0.4 0.2 0
b
1250 1250

1000 1000
distance (mm)
distance (mm)

750 750
12 MPa
FEM-fy235
500 500 FEM-fy355
FEM-0
FEM-fy235
250 250
FEM-fy355
FEM-0
0 0
0 0.2 0.4 0.6 0.8 1 0 5 10 15 20 25 30
Ns/PFEM Stress in concrete (MPa)

Fig. 14. Specimens FEM-0, FEM-fy235 and FEM-fy355 (load applied coincides with PFEM): (a) load distribution between steel cage and column (Nc, load supported by the
column; Ns, load supported by the cage) and (b) compressive stress in concrete.

a Nc/PFEM
1 0.8 0.6 0.4 0.2 0 b
1250 1250
FEM-0
FEM-fc4
1000 1000
FEM-fc20
distance (mm)

distance (mm)

FEM-fc30
750 750
4 MPa 12 MPa 20 MPa 30 MPa

500 500
FEM-0
FEM-fc4
250 250
FEM-fc20
FEM-fc30
0 0
0 0.2 0.4 0.6 0.8 1 0 10 20 30 40 50
Ns/PFEM Stress in concrete (MPa)

Fig. 15. Specimens FEM-0, FEM-fc4, FEM-fc20 and FEM-fc30 (load applied coincides with PFEM): (a) load distribution between steel cage and column (Nc, load supported by
the column; Ns, load supported by the cage) and (b) compressive stress in concrete.
2274 J.M. Adam et al. / Construction and Building Materials 23 (2009) 2265–2276

a 1 0.8 0.6
Nc/PFEM
0.4 0.2 0
b
1250 1250
FEM-0
FEM-St80
1000 1000
FEM-St120

distance (mm)
distance (mm)

FEM-St200
750 750
12 MPa
500 500
FEM-0
FEM-St80
250 250
FEM-St120
FEM-St200
0 0
0 0.2 0.4 0.6 0.8 1 0 5 10 15 20 25 30
Ns/PFEM Stress in concrete (MPa)

Fig. 16. Specimens FEM-0, FEM-St80, FEM-St120 and FEM-St 200 (load applied coincides with PFEM): (a) load distribution between steel cage and column (Nc, load supported
by the column; Ns, load supported by the cage) and (b) compressive stress in concrete.

and the cage in addition to the confinement effect imposed by the parameter n2 have been included so as to underline the
strengthening. It was also noted that while the failure on most of importance of the mechanism of transmission by shear
the FE models was located on the first section of angles (see stresses.
Fig. 12a), in the case of the FEM-aSt model, the failure was located (2) Confinement imposed by steel caging. The strengthening
on the angles situated in the centre (see Fig. 12b). produces a confinement effect on the column since it pre-
vents the expansion of the concrete caused by Poisson’s
6.7. Friction coefficient effect. Thus, the compressive strength of the concrete is
increased. The confinement effect was taken into account
The friction coefficient between the mortar and the cage steel when parameter n1 was defined.
has a considerable effect on the behaviour of the strengthened col-
umn. As l increases, ultimate load of the column also rises from As the results of the parametric study show, the confinement
1823.8 kN for specimen FEM-mu0.0 to 2649.2 kN for specimen imposed by the steel cage is very important for the behaviour of
FEM-mu0.6. Furthermore, when l increases, the load transmission the strengthened column (see the values of n1 included in Table
between column and cage improves (see Fig. 18), as well as the 3). In addition to this, the parameter of effectiveness n2 is, in prac-
effectiveness parameters n1 and n2 (see Table 3). tically all cases, less than one. This is because the strengthened col-
umn does not behave as a composite element since there is no
6.8. Analysis of the results obtained compatibility in deformation between the cage and the column,
especially at the ends of the strengthened column. Thus, slipping
As Adam et al. indicate [15], for the strengthening variant pro- is produced between the strengthening and the layer of mortar,
posed, there are fundamentally two mechanisms which influence as has been observed by Adam et al. [14], Giménez [10] and Giménez
the behaviour of the strengthened column: et al. [17].
Detailed analysis of the results of the parametric study makes
(1) Transmission by shear stresses. This is the way the load is clear that:
transmitted between the column and the cage, via the layer
of mortar between them. The figures representing the distri- (a) If the size of the angles is increased, the effectiveness of the
bution of loads between the cage and the column, as well as confinement increases (parameter n1). However, the effec-

Nc/PFEM
a 1 0.8 0.6 0.4 0.2 0
b
1250 1250

FEM-aSt
1000 1000
FEM-0
distance (mm)

distance (mm)

750 750
12 MPa

500 500

PMEF-aSt
250 250
PMEF-0

0 0
0 0.2 0.4 0.6 0.8 1 0 5 10 15 20 25 30
Ns/PFEM Stress in concrete (MPa)

Fig. 17. Specimens FEM-0 and FEM-aSt (load applied coincides with PFEM): (a) load distribution between steel cage and column (Nc, load supported by the column; Ns, load
supported by the cage) and (b) compressive stress in concrete.
 J.M. Adam et al. / Construction and Building Materials 23 (2009) 2265–2276 2275

Nc/PFEM
1 0.8 0.6 0.4 0.2 0
1250

1000

distance (mm)
750

500

FEM-0
250
FEM-mu0.0
FEM-mu0.6
0
0 0.2 0.4 0.6 0.8 1
Ns/PFEM

Fig. 18. Specimens FEM-0, FEM-mu0.0 and FEM-mu0.6. (load applied coincides with PFEM). Load distribution between steel cage and column (Nc, load supported by the
column; Ns, load supported by the cage).

tiveness of the transmission of loads between the cage and between the value of l and the behaviour of the strength-
the column (parameter n2). decreases when the size of the ened column (see Section 6.7).
angles is increased This phenomenon can be explained by
the fact that in the case of the increased-size angle, the fail- 7. Conclusions and future work
ure of the specimen is conditioned by the yielding of the
strips, due to the tensile stresses they are subjected to (see An experimental program was carried out to study axially
comments in Section 2). Thus, the angles are not able to loaded RC columns strengthened by steel cages. Numerical models
absorb their full load capacity. This particularity can be seen were verified from the experimental results and a parametric study
in Fig. 12c, where the failure mechanism of the cage (yield- was carried out analysing the influence of each of the parameters
ing of the strips) is shown for the FEM-L120 model. on the behaviour of RC columns strengthened by steel cages. The
(b) A variation in the yield stress of the steel of the cage slightly parameters analysed were: the size of the angles; the yield stress
affects the behaviour of the strengthened column. When the of the steel of the cage; the compressive strength of the concrete
value of the yield stress is increased, the ultimate load on the in the column; the size of the strips; the addition of an extra strip
strengthened column increases slightly but the effectiveness at the ends of the cage; the friction coefficient between the layer of
parameter n2 decreases. Considering the increased cost inc- mortar and the steel of the cage. The results obtained from the
urred by using steels with a high yield stress, it is important parametric study allow a series of guidelines to be established
to take the above effect into account when designing (see Section 6.8).
strengthening. It would seem important for future research to consider columns
(c) When the compressive strength of the concrete increases, which are strengthened while they are subjected to loading. It would
the effectiveness of the strengthening decreases (parameters thus be possible to simulate what happens when a column forming
n1 and n2). The effect mentioned occurs because the lower part of a real structure is strengthened. This recommendation also
deformability of higher strength of concrete means that applies to the three types of strengthening most widely used nowa-
the strengthening absorbs less load. This lower deformabi- days (steel jacketing, concrete jacketing, and composite jacketing),
lity will also mean a less lateral deformation of the concrete since little research has been done on this phenomenon.
due to the Poisson effect, which will reduce the confinement
effect. In view of this, it can be stated that the strengthening Acknowledgements
will be more effective in cases where the strength of the con-
crete is low. The authors wish to express their gratitude for the financial
(d) One way to reduce the slipping which occurs between the support received from the Spanish Ministry of Science and Tech-
strengthening and the column is to increase the size of the nology under the research project MAT 2003-08075, co-financed
strips. The bigger the strips, the bigger the confinement ef- with FEDER funds.
fect imposed by the strengthening will be (see Fig. 16b). This
is due to the greater stiffness of the steel cage in the trans- References
verse direction. This improvement in confinement will also
result in a better transmission of loads between the cage [1] Oey HS, Aldrete CJ. Simple method for upgrading an existing reinforced-
concrete structure. Practice Periodical Struct Design Constr 1996;1(1):47–50.
and the column by the shear stress mechanism (see [2] Frangou M, Pilakoutas K, Dritsos S. Structural repair/strengthening of RC
Fig. 16a). columns. Constr Build Mater 1995;9(5):259–66.
(e) The critical area which conditions the behaviour of the str- [3] Dritsos S, Pilakoutas K. Composite technique for repair/strengthening of RC
members. In: Second international symposium on composite materials and
engthened column is at the ends [5,8,10,14]. A more eco- structures. China: Peking University Press; 1992.
nomical way to attain the effects mentioned in (d) would [4] Tamai S, Sato T, Okamoto M. Hysteresis model of steel jacketed RC columns for
be to put the strips closer together in the area closest to railway viaducts. In: Proceedings of the 16th congress of IABSE. Lucerne; 2000.
[5] Cirtek L. RC columns strengthened with bandage – experimental programme
the ends of the column (see Fig. 11b). Adding an extra strip
and design recommendations. Constr Build Mater 2001;15(8):341–9.
gives, at only a slight increase in strengthening costs, a [6] Fukuyama H, Sugano S. Japanese seismic rehabilitation of concrete build-
greater ultimate load and more efficient strengthening. ings after the Hyogoken-Nanbu earthquake. Cem Concr Compos 2000;22(1):
(f) Considering that l depends on the thoroughness of the con- 59–79.
[7] Wu YF, Liu T, Oehlers DJ. Fundamental principles that govern retrofitting of
struction of the steel cage [14,29], this work should be reinforced concrete columns by steel and FRP jacketing. Adv Struct Eng
undertaken with great care, since there is a strong connection 2006;9(4):507–33.
2276 J.M. Adam et al. / Construction and Building Materials 23 (2009) 2265–2276

[8] Adam JM. Structural behaviour of RC columns strengthened by steel jacketing. [18] Dolce M, Masi A, Cappa T, Nigro D, Ferrini M. Experimental evaluation of
Research report, Code 1940/66. Technical University of Valencia; 2007. effectiveness of local strengthening on columns of R/C existing structures. In:
[9] CEB-FIB. Seismic assessment and retrofit of reinforced concrete buildings. Proceedings of fib-symposium concrete structures in earthquake regions,
Bulletin no. 24, Task Group 7.1; 2003. Athens, Greece; 2003.
[10] Giménez E. Experimental and numerical study of reinforced concrete columns [19] Cirtek L. Mathematical model of RC banded column behaviour. Constr Build
strengthened with steel angles and strips subjected to axial loads. PhD thesis. Mater 2001;15(8):351–9.
Technical University of Valencia; 2007 [in Spanish]. Available in <http:// [20] de Fomento Ministerio. Instrucción de hormigón estructural. Madrid: EHE;
[Link]>. 1998 [in Spanish].
[11] Ramírez JL. Ten concrete column repair methods. Constr Build Mater [21] ENV 1992-1-1 (Eurocode No. 2). Design of concrete structures. Part 1: General
1996;10(3):195–202. rules and rules for buildings; 1991.
[12] Ramírez JL, Bárcena JM. Eficacia resistente de pilares de hormigón armado de [22] CEB-FIB Model Code 90. Laussane; 1991.
baja calidad reforzados por dos procedimientos diferentes. Informes de la [23] ENV 1993-1-1 (Eurocode No. 3). Design of steel structures. Part 1: General
Construcción 1975;272:89–98 [in Spanish]. rules and rules for buildings; 1993.
[13] Ramírez JL, Bárcena JM, Feijóo JM. Comparación resistente de cuatro métodos [24] ANSYS 11.0. Theory reference. ANSYS Inc.; 2006.
de refuerzo de pilares de hormigón armado. Informes de la construcción [25] ANSYS 11.0. Element reference. ANSYS Inc.; 2006.
1977;290:57–68 [in Spanish]. [26] Willam KJ, Warnke ED. Constitutive model for the triaxial behaviour of
[14] Adam JM, Ivorra S, Giménez E, Moragues JJ, Miguel P, Miragall C, Calderón PA. concrete. In: Proceedings international association of bridge and structural
Behaviour of axially loaded RC columns strengthened by steel angles and engineering. Bergamo, Italy: ISMES; 1974.
strips. Steel Compos Struct 2007;7(5):405–19. [27] Johansson M, Gylltoft K. Structural behaviour of slender circular steel–
[15] Adam JM, Giménez E, Calderón PA, Pallarés FJ, Ivorra S. Experimental study of concrete composite columns under various means of load application. Steel
beam-columns joints in axially loaded RC columns strengthened by steel Compos Struct 2001;4:393–410.
angles and strips. Steel Compos Struct 2008;8(4):329–42. [28] Johansson M, Gylltoft K. Mechanical behaviour or circular steel-concrete
[16] Fernández M. Patología y terapéutica del hormigón armado. Colegio de composite stub columns. J Struct Eng 2002;128(8):1073–81.
Ingenieros de Caminos, Canales y Puertos, Madrid; 1994 [in Spanish]. [29] Adam JM, Calderón PA, Giménez E, Hidalgo C, Ivorra S. A study of the behaviour
[17] Giménez E, Adam JM, Ivorra S, Moragues JJ, Calderón PA. Full-scale testing of of the cement mortar interface in reinforced concrete columns strengthened
axially loaded RC columns strengthened by steel angles and strips. Adv Struct by means of steel angles and strips. In: Structure Faults Repair –
Eng; accepted for publication. 2006. Edinburgh; 2006.