EXPERIMENTAL STUDY ON RC COLUMNS RETROFITTED BY FRP

AND SUBJECTED TO SEISMIC LOADING

Raphaelle SADONE, Marc QUIERTANT

Université Paris Est - IFSTTAR - SOA
58 Boulevard Lefebvre, 75015 PARIS
r.sadone@hotmail.fr*; marc.quiertant@ifsttar.fr*

Julien MERCIER
Freyssinet International
1bis, rue du Petit-Clamart, Bâtiment C, 78148 VELIZY
julien.mercier@freyssinet.com

Emmanuel FERRIER
Université Lyon 1 – LGCIE
Domaine scientifique de la Doua, 82 bd Niels Bohr, 69622 Villeurbanne Cedex
Emmanuel.Ferrier@univ-lyon1.fr

Abstract
Structures can be submitted to severe conditions, especially earthquakes. Strengthening of
reinforced concrete structures is then a matter of concern. Externally bonded Fiber-Reinforced
Polymers (FRP) are emerging as a valuable solution for structural strengthening but design
rules have to be established concerning their application for seismic strengthening. For this
purpose, an experimental campaign carried out on full-scale reinforced concrete (RC)
columns has been undertaken. Different strengthening configurations have been applied on
columns, which were then tested under combined axial and reversed cyclic lateral load. Those
tests helped us to analyse the behaviour of RC columns depending on the FRP confinement
(carbon FRP jacket) and on flexural reinforcement (carbon plates).

Keywords: concrete columns, cyclic, strengthening, flexure, FRP

1. Introduction
Many existing buildings or bridges have been designed according to old codes and may not be
resistant enough if any earthquake occurs; they do not satisfy specific seismic detailing
requirements and may collapse. For several years, new rules have been introduced in order to
avoid such deficiencies in new constructions, but some existing ones need to be retrofitted.
Considering that the major cause of constructions collapse is column failure, specific
retrofitting techniques and design rules have to be established. In this context, the use of Fibre
Reinforced Polymer (FRP) has significantly increased in construction and civil engineering
fields. Indeed, bonding of external FRP reinforcements is now recognized as an effective
technique for the strengthening of reinforced concrete (RC) structures, and it can be
particularly useful for seismic retrofitting. Most research studies conducted to date on
reinforcement of existing RC columns using externally bonded FRP have mainly been
focused on confinement efficiency (for example [1], [2], [3], [4], and [5]). However, columns
can be strengthened by associating a flexural strengthening, achieved by FRP plates bonded
longitudinally, with a confinement by wrapping. Such retrofitting scheme seems particularly
appropriate for seismic strengthening or more generally for the strengthening of structural RC

Page 1 of 9
columns that are rarely perfectly axially compressed [6]. A previous study of Hadi [7] carried
out on small cylindrical plain concrete columns showed that longitudinal CFRP straps applied
on columns could highly increase their performances in terms of strength and ductility (under
eccentric compression loading), compared to columns without any longitudinal CFRP straps.
Before being generalized, such result has to be confirmed on real-scale rectangular RC
columns.
Then this study aims to evaluate the strength and ductility enhancement resulting from the use
of longitudinally bonded FRP. For this purpose, the paper presents the results of experimental
investigation conducted on six real scale RC columns tested under a quasi-static loading path
intended to be representative of a seismic solicitation: reversed lateral cyclic loading and
constant axial load. The main parameter tested was the FRP configuration. Those tests helped
us to analyze the behaviour of RC columns depending on the FRP confinement (CFRP jacket)
and on flexural reinforcement (carbon plates) and allowed us to determine the influence of
these parameters on the ultimate strength, on ductility, on stiffness, on dissipated energy, and
then to conclude about the behaviour of the externally reinforced columns.

2. Experimental investigation
2.1 Test specimens
Specimens consisted of 0.25 x 0.37 x 2.50m3 columns connected to a 1.25 x 1.00 x 1.00m3
RC stubs. Specimen dimensions and internal reinforcement details are shown on Figure 1.
The reinforcing cage was made of two parts: one for the column and one for the stub, with the
column longitudinal rebars extending through the stub. Six 10mm steel deformed rebars were
used for longitudinal reinforcement and 6mm ties spaced at 150mm, were used for transverse
reinforcement. These design details produce a longitudinal reinforcement ratio of 0.5 %.

Figure 1. Column dimensions and reinforcing cage.

2.2 Materials
From each batch of concrete it was possible to cast two or three specimens at the same time.
The concrete strength of each specimen was determined by averaging the values obtained
using three standards 11cm by 22cm concrete cylinders. Mechanical properties of concrete,
for each column, are summarized in Table 1.

Table 1. Mechanical properties of concrete.

PRef-1 PRef-2 PC-1 PC-2 PCL-1 PCL-2
46.23 47.43 51.5 40.3 34.7 40.3
Compressive strength [MPa]
(ξ=1.5) (ξ=0.6) (ξ=1.5) (ξ=0.6) (ξ=1.1) (ξ=0.6)
3.55 3.49 3.82 3.74 2.88 3.74
Tensile strength [MPa]
(ξ=0.1) (ξ=0.1) (ξ=0.2) (ξ=0.2) (ξ=0.3) (ξ=0.2)
(ξ : standard deviation)

Page 2 of 9
The CFRP confining jacket was made using the wet-lay up process. Saturated carbon fibre
sheets were wrapped around the column while flexural reinforcement was achieved by
bonding CFRP plates. Characteristics of FRP reinforcements are summarized in Table 2.
Before strengthening, corners of the columns were rounded to avoid premature failure of
CFRP. Confining sheets were applied one day after the laminates have been installed.

Table 2. Characteristics of FRP reinforcement (Freyssinet products).

Thickness 0.48 mm
Width Pre-cut sheets of width 300mm
CFRP sheets (TFC©)
Young’s Modulus 105 000 MPa
Ultimate strain 0.01
Thickness 1.2 mm
Width 50mm
Pultruded Plates
Young’s Modulus 160 000 MPa
Ultimate strain 0.07

To characterize internal reinforcement of RC columns, tensile tests were realised on 6mm and
10mm-diameter steel rebars. Results of the tests are presented in Table 3.

Table 3. Steel properties.

Diameter of Tensile strength Ultimate strain Yielding strain Young’s modulus

the rebar (MPa) (µdef) (µdef) (MPa)

10 mm 603.5 (ξ=4.4) 37 340 (ξ=471) 2 785 (ξ=110) 195000 (ξ=7996)

6 mm 613.7 (ξ=10.6) 20 820 (ξ=500) 3 106 (ξ=194) 185 400 (ξ=9151)

2.3 Instrumentation
Extensive instrumentation was used to collect data during tests: concrete, FRP and steel
strains at different locations, deflections along the specimen length, and axial and lateral load
were recorded. 15 strain gauges were attached to longitudinal rebars and 2 strain gauges on
ties; 18 strain gauges measured either concrete (for non reinforced columns) or FRP (for
reinforced columns) strains, and 5 LVDT measured transverse displacements of test
specimens. Displacements and loading applied by hydraulic jacks were recorded thanks to
displacement and load cells located in front of the jacks.

2.4 Strengthening configurations

For the six columns, the main studied parameter is the FRP configuration. The different
strengthening configurations are summarized in Table 4. Two specimens of each series were
tested.

Page 3 of 9
 Table 4. FRP strengthening configurations.

PRef-1 & Pref-2 PC-1 & PC-2 PCL-1 & PCL-2

1 layer of CFRP sheets + CFRP

Without strengthening 1 layer of CFRP sheets
plates

Strengthening
configuration

2.5 Test setup and loading procedure

Testing was carried out in the Structures Laboratory of IFSTTAR, located in Paris. The
specimens were tested horizontally as shown on Figure 2. Seismic load was simulated by
applying cyclic lateral displacements gradually increasing (representative of a seismic
loading), while the column was simultaneously subjected to a constant axial load (simulating
gravity load). The constant axial load of 700kN was applied through a pair of hydraulic jacks
(capacity 1100kN, 220mm) linked to prestressing tendons, and displacement controlled lateral
load was applied thanks to another hydraulic jack (capacity 2000kN, +/- 250mm).
At the beginning of each test, axial load was applied thanks to six prestressing tendons, three
on each side of the column, linked to an anchored plate at the top of the column, and to the
jacks behind the stub. After the application of the axial load, the specimen was subjected to
progressively increasing lateral displacement cycles. Two fully reversed cycles were applied
for each displacement step. Those displacement steps, referred here as “drift ratio”, were
defined as a ratio of the column height: 0.25%; 0.5%; 1%; 2%; 4%; etc. until failure. The
column was considered to have reached its ultimate condition when the specimen is unable to
sustain an applied lateral load inferior to 50% of the maximum lateral capacity observed
during the test.

Page 4 of 9
 Lateral hydraulic jack applying cyclic lateral
displacements

Prestressing
tendons

Horizontal jacks
applying constant
axial load

Figure 2. Test setup.

3. Experimental results and discussion

3.1 General observations
For reference specimens (PRef-1 and PRef-2), a buckling of the longitudinal reinforcements
and a huge spalling of cover concrete (Figure 3) occurred during lateral displacement cycles
corresponding to a drift ratio of 4%. Cracks formed at or near the footing column interface.
Then, at the end of the second cycle at the drift ratio of 6%, an important loss of axial
resistance appeared: the column was not able to resist the loads anymore and the test was
stopped. Specimens PRef-1 and PRef-2 reached, respectively, a maximum lateral load of
51.0kN and 50.1kN, at drifts of 70mm and 58mm (during the first cycle at a 4 %drift ratio).
Retrofitted specimens produced sporadic popping sounds at the beginning of tests as the
composite jackets and laminates were stressed. For all specimens, steel longitudinal
reinforcements started buckling for a drift ratio of 2%. Then, from the lateral displacement
corresponding to the drift ratio of 6% until the end of the test, it was possible to hear the
tensile rupture of the different longitudinal rebars. The first confinement strip ruptured at
welded corner (Figure 3) during loading cycle at a 6% drift ratio for specimen PCL-1 and at a
10% drift ratio for specimen PCL-2. Specimen PC-1 reached a maximum lateral load of
49.7kN at a lateral displacement of 74.6mm (during the first cycle at a drift ratio of 4%), PC-2
reached a maximum load of 48.8kN at a drift of 41mm (during the first cycle at a drift ratio of
4%), PCL-1 reached a maximum load of 48.31kN at a drift of 40.6mm (during the first cycle
at a drift ratio of 2%) and PCL-2 reached a maximum load of 51kN at a drift of 76mm (during
the first cycle at a drift ratio of 4%).
For reference specimens PRef-1 and PRef-2, the test was terminated after successfully
completing the loading cycles corresponding to a peak drift ratio of 6%, whereas retrofitted
specimens were tested until a drift ratio of 10%. The ultimate drift displacement is defined as
the displacement recorded at failure.

Page 5 of 9
 PRef-1 PC-1 PCL-1

PRef-2 PC-2 PCL-2

Figure 3. Specimens after testing.

Many previous research have demonstrated that external confinement of RC columns by FRP
wrapping can significantly delay rebars buckling, restrain concrete cover spalling and let
compressive concrete strains reach larger values, thus resulting in important flexural ductility
enhancement (for example [8]). Results obtained during the experimental study presented in
this paper are in accordance with this previous work. At failure, large deflection was obtained
for retrofitted specimens (Figure 4). Broadly, a concentration of the damaged zone at the
junction footing-column was observed for the six specimens: cracks formed at or near the
junction and a plastic hinge appeared. As a consequence, bending deformation of the column
is essentially due to the hinge rotation. After plastic hinge formation, the column stayed
nearly straight (a phenomenon confirmed by LVDT measurement).

Figure 4. Specimen PC-2 subjected to a lateral drift of 10%.

3.2 Strength – lateral displacements hysteretic response

Figure 5 shows the applied lateral force versus displacement drift ratio envelope curves for
three specimens (one of each series). Considering that a good reproducibility of tests was

Page 6 of 9
obtained and for a better legibility of the graph, results of only one specimen by series are
plotted on Figure 5.

Figure 5. Load-lateral displacements envelope curves

On Figure 5 it can be observed that the ultimate displacement of PC-1 and PCL-1 is about
twice the ultimate displacement of reference specimen. From this result, it can be concluded
that the two strengthening configurations are efficient to enhance structural ductility.
However, no significant strength enhancement was obtained for retrofitted specimens.
Moreover, considering columns PC-1 and PCL-1 it can be noticed that the overall member
response in both specimens was similar. The addition of longitudinal CFRP plates does not
influence the behaviour of the column in terms of strength and ductility. This is a particularly
interesting result considering the installation labour cost of such strengthening systems.

3.3 Stiffness
Figure 6(a) compares the stiffness curves (instantaneous secant stiffness at a certain
displacement) of the different columns at every cycle. On this figure, a positive value of cycle
number corresponds to a loading cycle involving upward lateral displacement, and negative
values are related to cycles involving downward displacement. The initial stiffness is lower
for non retrofitted specimens (initial stiffness of PC-1 is 30 % higher than initial stiffness of
PRef-1), but there is no difference in terms of initial stiffness between the two strengthening
configurations. Moreover, from Figure 6(b), it can be observed that the degradation of the
ratio of actual stiffness over initial stiffness is quite similar whatever the column, the
retrofitted specimen even experiences a slightly faster degradation.

(a) (b)
Figure 6. Stiffness curves (a) and stiffness degradation curves (b) for three specimens.

Page 7 of 9
3.4 Dissipated energy
The cumulative energy dissipation versus the cumulative lateral displacement for PRef-1, PC-
1 and PCL-1 is presented on Figure 7. The wobbles in the curves are due to elastic energy
release at the unloading paths of each cycle. From Figure 7 it can be seen again that there is
no noticeable difference between the two strengthening configurations in terms of energy
absorption capacity. For a same displacement path, the same energy dissipation occurred in
both specimens. Furthermore, if overall behaviours of FRP-retrofitted specimens and
reference specimens are compared, it can be concluded that retrofitting permits to dissipate
more energy because retrofitted columns can experience bigger deflections. This is mainly
due to the plastic hinge zone confinement. However, if we focus on the first part of the graph
(until failure of PRef-1), there is no significant difference between the three curves. None of
the specimens dissipate faster the energy than the others.

Figure 7. Energy dissipation of cyclically loaded columns.

4. Conclusion and perspectives

An experimental study was undertaken to investigate the behaviour of realistically sized RC
columns strengthened with externally bonded FRP and subjected to a seismic loading. Two
strengthening configurations were tested: columns confined with CFRP and columns
combining CFRP confinement with flexural reinforcement achieved by bonded longitudinal
CFRP plates. The specimens were tested under constant axial load simulating gravity load and
lateral reversed cyclic load simulating earthquake force. First, usual trends were confirmed:
- better performances, in terms of ductility, are observed for confined columns
compared to reference specimens,
- the damaged zone is concentrated at the junction footing-column, with a plastic hinge
located in this zone.
But, within the limits of this study, the following original conclusion has also been drawn:
- longitudinal reinforcement, applied in addition to confinement, does not noticeably
change the behaviour of the columns in terms of ductility, strength, or dissipated
energy (compared to simply confined column) and are then considered ineffective.
It then appears that an optimized strengthening scheme for RC columns have to combine
plastic hinge zone confinement, zones without external reinforcement for energy dissipation
and anchorage system for longitudinal CFRP straps. Those anchors have been successfully
tested on concrete blocks ([9]) but have to be validated on realistically sized structures.

Page 8 of 9
5. References
[1] Iacobucci R.D., Sheikh S.A., and Bayrak O., “Retrofit of Square Concrete Columns
with Carbon Fiber-Reinforced Polymer for seismic resistance”, ACI Structural Journal,
100-S81, November/December 2003, pp. 785-794.
[2] Harries K.A., Ricles J.R., Pessiki S., Sause R., “Seismic retrofit of lap splices in non
ductile square columns using carbon fiber-reinforced jackets”, ACI Structural Journal,
103-S89, November/December 2006, pp. 874-884.
[3] Mohamed H. H., Dagher F., “Seismic Strengthening of Bond-Critical Regions in
Rectangular Reinforced Concrete Columns Using Fiber-Reinforced Polymer Wraps”,
ACI Structural Journal, Vol 105(1), 2008, p. 68-77.
[4] Promis G., Ferrier E., Hamelin P., “Effect of external FRP retrofitting on reinforced
concrete short columns for seismic strengthening”, Composite structures, Vol. 88(3),
2009, pp. 367-379.
[5] Gu D.S., Wu G., Wu Z.S., Wu Y.F., “Confinement Effectiveness of FRP in Retrofitting
Circular Concrete Columns under Simulated Seismic Load”, Journal of composites for
construction, Vol. 14 (5), 2010, pp. 531-540.
[6] Quiertant M., Clement J.-L., “Behavior of RC columns strengthened with different
CFRP systems under eccentric loading”, Construction and Building Materials, Vol
25(2), 2010, pp. 452-460.
[7] Hadi MNS., “Behaviour of FRP strengthened concrete columns under eccentric
compression loading”, Composites Structures, Vol 77, 2007, pp. 92-96.
[8] Hadi MNS., “Behaviour of FRP wrapped normal strength concrete columns under
eccentric loading”, Composite Structures, Vol 72(4), 2006, pp.503–11.
[9] Sadone R., Quiertant M., Chataigner S., Mercier J., “Behavior of an innovative end-
anchored externally bonded CFRP strengthening system under low cycle fatigue”, 5th
International Conference on FRP Composites in Civil Engineering CICE 2010, Beijing,
27-29 Sept. 2010.

Page 9 of 9