Article

Journal of Reinforced Plastics

and Composites

Repair and strengthening of reinforced 29(22) 3411–3424

! The Author(s) 2010
Reprints and permissions:
concrete columns with CFRPs sagepub.co.uk/journalsPermissions.nav
DOI: 10.1177/0731684410376332
jrp.sagepub.com

Okan Ozcan1, Baris Binici2, Erdem Canbay2 and

Guney Ozcebe2

Abstract
In this study, seismic retrofitting of undamaged and moderately damaged RC columns with CFRP was investigated. The
experimental program was conducted on five RC columns with inadequate tie spacing, 90-degree hooks at tie ends, and
plain (undeformed) reinforcing bars. Columns were tested under a constant axial load of 27% of axial capacity and under
reversed lateral cyclic loading. The primary objective of this study was to investigate the influence of sustained axial load
during repairing, damage, and corner-rounding radius of CFRP wraps on the seismic performance of RC columns.
Improved performance in terms of ductility, dissipated energy, and secant-stiffness degradation was observed for the
CFRP strengthened columns. Reducing the corner-rounding radius deteriorated the seismic behavior of the columns,
while the stiffness degradation response remained unaffected. The sustained axial load during repairing had only minor
effect on column behavior. After CFRP repairing, the increase in yield drift ratio and curvature was observed to be in
proportion with the corresponding consumed ductility values during damage cycles. A drift-based design equation for
FRP repaired columns was proposed that considers the damage amount and axial load level as the major parameters as
consistent with the experimental results.

Keywords
Reinforced concrete column, repair, strengthening, FRP

Introduction
In Turkey, the use of plain bars was quite common
Poor performance of reinforced concrete (RC) columns until the 1999 Kocaeli and Düzce earthquakes.
during recent earthquakes in Turkey (Kocaeli 1999, Considering inadequate bond characteristics of plain
Düzce 1999, and Bingöl 2003) demonstrated the lack bars, seismic behavior of such columns is inferior due
of deformability in building columns due to inadequate to the need of longer development length and signifi-
seismic detailing of confining steel reinforcement. cant fixed-end rotations. The literature review con-
Convenient strengthening and repairing techniques ducted reveals that there was no experimental data on
can be utilized for upgrading such columns. The tradi- CFRP repair of RC columns with plain bars to enhance
tional method of repairing or strengthening RC col- the deformability of the section. This study examines
umns by means of concrete1–5 or steel jacketing6,7 the performance of both strengthened and repaired
involves complicated and cumbersome implementation
procedures. There are several studies in the literature on
seismic upgrading of RC columns by using carbon
fiber reinforced polymer (CFRP) sheets.8–23 It was 1
Department of Civil Engineering, Akdeniz University, Antalya, Turkey.
2
shown that CFRP wrapping of columns enhances the Department of Civil Engineering, Middle East Technical University,
seismic behavior by confining the column section and Ankara, Turkey.
preventing rebar buckling. Throughout these studies,
Corresponding author:
repairing with concrete1–3 and CFRP jacketing10–13,18 Erdem Canbay, Department of Civil Engineering, Middle East Technical
was investigated only for columns that had deformed University, Ankara 06531, Turkey
rebars for longitudinal reinforcement. Email: [email protected]
3412 Journal of Reinforced Plastics and Composites 29(22)

columns with weaknesses such as inadequate transverse and S-NL-10, column corners were rounded to 30 and
reinforcement, 90-degree hooks at tie ends, and plain 10 mm, respectively. The remaining companion speci-
bars. The main test variables were selected as the pres- mens R-NL-30 and R-UL-30 were repaired under the
ence of axial load during repairing, undamaged/ absence and presence of axial load (27% of axial load
damaged column retrofitting, and rounding radius of capacity), respectively. Herein, both the CFRP repaired
column corners. In light of the experimental results pre- specimens were initially damaged to 2% drift ratio by
sented herein, it is believed that retrofit and repair applying cyclic displacement excursions. After the lat-
design of such columns can be conducted economically. eral damage cycles, the tip displacement of both speci-
mens was brought back to its original configuration.
Subsequently, specimens R-NL-30 and R-UL-30 were
Experimental program tested after CFRP repairing using a corner-rounding
radius of 30 mm. Table 1 presents the details of test
Test specimens
specimens and test parameters.
The test specimens had dimensions of 350  350 
2000 mm3 that were connected to a 1350  500 
400 mm3 foundation stub (Figure 1(a)). Longitudinal
Material properties
reinforcement ratio of 2.48% was achieved by using All the specimens were cast vertically with three batches
plain rebars with 8–22 mm diameter for all the columns. of concrete having maximum 15 mm aggregate size with
For the transverse reinforcement, plain bars of 10 mm a target compressive strength of 20 MPa. Concrete
diameter spaced at 200 mm with 90-degree hooks at tie compressive strength values at the test day of the speci-
ends were used to simulate the typical deficient con- mens are presented in Table 1. The plain longitudinal
struction practice in Turkey (Figure 1(a) and (b)). rebars had yield strength (fy) and ultimate strength (fu)
Target concrete compressive strength was selected as of 284 and 453 MPa, respectively. Yield and ultimate
20 MPa. All specimens were wrapped with one layer strengths of transverse reinforcement were 331 and
of CFRP sheet except the unwrapped reference speci- 439 MPa, respectively. Carbon fiber wrapping was
men (REF). For the strengthened specimens S-NL-30 used to strengthen and repair the test specimens.

(a) (b)
350 mm
48 mm

350 mm
150
15050

Figure 1. The details of test specimens: (a) full view and (b) close-up view.
Ozcan et al. 3413

Table 1. Specimen properties

CFRP
Specimen properties Longitudinal steel Axial load implementation
a
ratio (%) ratio N/N0 (%)
0
fc fy, longitudinal Reinforcement
Specimen (MPa) (MPa)
Longitudinal Transverse Ply CR (mm) Typeb

REF 20 8–22 mm 10 mm spaced 27 0  REF

S-NL-30 20 plain bars at 200 mm plain bars 1 30 S, NL
S-NL-10 22 284 2.48 1 10 S, NL
R-NL-30 19 1 30 R, NL
R-UL-30 20 1 30 R, UL
CR: Corner-rounding radius, S: Strengthened, and R: Repaired.
a
N0 ¼ 0.85 fc0 Ag + As fy: axial capacity.
b
NL, UL: CFRP application was made under No Load and Under Load, respectively.

CFRP plies with a thickness of 0.165 mm, an elasticity

Instrumentation and testing
modulus of 230,000 MPa, and an ultimate strain of
0.015 were used in retrofitting process. The repairing All columns were instrumented by linear variable dif-
mortar had a 28-day uniaxial compressive strength ferential transformers (LVDTs) to measure horizontal
of 70 MPa. deflections and dial gages for rotations at the critical
regions as shown in Figure 3(a). There were seven
LVDTs located at 350, 1000, 1750, and 2000 mm
CFRP application height measuring lateral deflections of the column. In
First, the corners of each column were rounded to a addition, four dial gages were placed at a height of
predefined radius (10 mm for specimen S-NL-10 and 350 mm. Eight strain gages were placed on longitudinal
30 mm for the remaining specimens). Afterwards, a rebars at 50 and 350 mm away from the base (Figure
thin layer of undercoat was applied on the potential 1(b)) to monitor strain values during both the damage
plastic hinge region of the strengthened columns. cycles and after column retrofitting. Load cells recorded
CFRP wrapped region was extended to 650 mm for both lateral and axial loads during the test. Columns
the repaired columns. Undercoating was applied after were guided with four rollers between the guide rails to
the removal and cleaning of spalled cover concrete assure bending only in loading plane. The test setup is
for the repaired specimens (Figure 2(a) and (b)). For shown in Figure 3(b). All the column tests were con-
the damaged specimens, the column was repaired using ducted under constant axial load and lateral displace-
epoxy-based mortar over an undercoat layer (Figure ment cycles. During the column tests, the axial load on
2(c) and (d)). To achieve a successful interlock between the columns was applied prior to lateral displacement
column and CFRP sheet and to prevent debonding of cycles. For the repaired column under sustained axial
CFRP sheets at the overlapping region, CFRP anchor- load, the axial load was maintained during repairing
ages produced in-house were placed along the test and curing of CFRPs. In the loading program, drift
region of the columns. For the carbon fiber anchor ratio (column lateral tip displacement divided by
dowels, 12 mm diameter holes with a depth of 80 mm column height) levels were incremented by 0.5% until
were drilled between two longitudinal bars at heights of 3% drift with three cycles per drift level. The following
50, 250, and 450 mm, and an extra hole at 600 mm was cycles beyond 3% drift had 1% of drift increment
used for the repaired columns. The CFRP anchor having two cycles for each drift ratio.
dowels were formed from 120  130 mm2 carbon fiber
rectangular pieces. These pieces were rolled parallel to
fiber direction and tied with a string. CFRP dowels
Test observations
130 mm long were inserted into 80 mm depth holes, Flexural failure mode (i.e., longitudinal steel yielding
and outer parts of the dowels were attached onto the before concrete crushing at the plastic hinge region)
pre-wrapped CFRP sheet (Figure 2(e)). The strength- was observed in all columns as expected, since shear
ened and repaired specimens were tested 1 week after span ratio was about 5.5 and lateral stirrups were
the CFRP application (Figure 2(f)). used for all specimens. The measured specimen
3414 Journal of Reinforced Plastics and Composites 29(22)

Figure 2. CFRP application: (a) moderately damaged column (up to 2% drift), (b) removal of spalled cover and undercoat application,
(c) repairing with epoxy-based mortar, (d) applying the epoxy-based mortar along CFRP wrapping length, (e) anchorage insertion, and
(f) CFRP wrapped column after repairing.

responses in the tests are summarized in Table 2. The buckling occurred at 3% and 5% drift ratio for the
lateral load–tip deflection hysteretic behavior of the col- reference and retrofitted specimens, respectively. The
umns and important events during the test such as lateral strength of the column that was repaired under
column – stub interface cracking, rebar yielding, the absence of axial load was approximately 15%
CFRP debonding, CFRP rupture, and onset of rebar greater than the lateral strength observed during
buckling are presented in Figures 4–8. damage cycles.
For all specimens, evenly distributed flexural cracks
were observed along both faces at heights from 50 to
Ultimate deformations and ductility
1000 mm during the test. For the CFRP repaired col-
umns, the visible flexural cracks formed during the The ultimate values of drift ratio and curvature denote
damage state opened further after repairing and no the levels at which the lateral strength degraded to
new flexural cracks outside the repaired region were 80% of lateral load capacity (Table 2). The ultimate
observed. For the reference specimen, crushing of the curvatures were calculated using the measured data
concrete cover was observed at the column base over between 350 mm and the columns base. In order to
a height of approximately 350 mm at 2% drift ratio. obtain displacement and curvature ductility, the ulti-
In all retrofitted columns, first debonding and rupture mate displacements and curvatures were divided by the
of CFRP were observed at approximately 2% and yield values determined from strain gage measure-
5% drift ratio levels, respectively. The onset of rebar ments. The yield values of the damaged specimens
Ozcan et al. 3415

(a) (b)
Loading
head LVDT Axial Hydraulic
Dial loading actuators for
gage system lateral and
axial loading

Loading
head Load cells
2000
1750
High strength (HS)
Threaded rods
1000

Column-stub
interface f48 HS bolts
350

Universal
50
400 base

Strong floor
Units: mm
1350 500

Figure 3. The details of: (a) instrumentation and (b) test setup.

Table 2. Test results

CFRP Yield Ultimatea Ductility, mb

Lp IC DR Debond Rupture Load DR K DR K

Specimen (mm) (%) DR (%) DR (%) (kN) (%) (rad/km) (%) (rad/km)
K

REF 350 0.5 2.0 (crush) 70.9 0.8 16.0 3.3 90.6 4.1 5.7
S-NL-30 310 2.5 5.0 73.1 0.8 14.2 4.1 107.9 5.1 7.6
S-NL-10 350 2.0 5.0 67.1 0.8 12.3 3.6 82.2 4.5 6.7
R-NL-30 300 2.0 5.0 67.5 0.8 15.3 5.2 129.2 6.5 (2.5) 8.4 (2.7)
R-UL-30 280 2.0 5.0 55.4 0.8 12.9 5.0 148.9 6.3 (2.5) 11.5 (3.8)
Lp: Plastic hinging length based on visual inspection, IC: Column-footing stub interface cracking, DR: Drift ratio,
: Displacement, and K: Average
curvature.
a
Denotes 20% drop in lateral strength.
b
Values given within parentheses denote consumed ductility during damage.

(R-NL-30 and R-UL-30) were taken based on the in curvature ductility was about 3 due to the deterio-
strain measurements before CFRP repairing. In rating effect of damage cycles.
Table 2, the values that were given within parentheses The details of average curvature measurements are
point out consumed ductility levels during damage illustrated in Figure 9(a). Based on the dial gage read-
cycles. The ultimate drift ratio of the reference ings at two levels, average curvature measurements
column improved from about 3% to approximately were obtained along three regions named as 350-50,
4% and 5% for strengthening and repairing cases, 350-base, and 50-base. It can be observed that highest
respectively (Table 2). For the strengthened columns, curvatures were obtained at 50-base region (Figure
the displacement ductility increased from approxi- 9(b)–(d)). Henceforth, curvature ductility values
mately 4 to 5; however, the repaired columns could observed within 350-50 were somewhat lower com-
attain a displacement ductility of at the most 4 while pared to those measured within 350-base. The envelope
consuming a displacement ductility of 2.5 during the curves of lateral load–drift and moment–average cur-
damage cycles. Similarly, although the curvature duc- vature responses for the columns are shown in Figure
tility level improved from approximately 6 to 8 for the 10(a)–(c) for comparison. For the strengthened col-
strengthened and repaired columns, the consumption umns, the ultimate displacements and displacement
3416 Journal of Reinforced Plastics and Composites 29(22)

100
Column–stub interface cracking
Rebar yielding
Concrete crushing
Onset of rebar buckling
50
Lateral load (kN)

–50

REF
–100
–7.5 –5 –2.5 0 2.5 5 7.5
Drift ratio (%)

Figure 4. Lateral load vs. drift ratio response for specimen REF.

100
Column–stub interface cracking
Rebar yielding
CFRP debonding
CFRP rupture
50
Onset of rebar buckling
Lateral load (kN)

–50

S-NL-30
–100
–7.5 –5 –2.5 0 2.5 5 7.5
Drift ratio (%)

Figure 5. Lateral load vs. drift ratio response for specimen S-NL-30.

ductility levels were augmented due to CFRP confine-

ment while impeding premature rebar buckling. The
Rebar strains
shifting of the yield displacement after damage is The strain values in longitudinal rebars located at 50
apparent in Figure 10. The average curvatures mea- (bottom) and 350 mm (top) over the base (Figure 1(b))
sured at 350-50 mm tended to decrease due to the are plotted against the drift ratios in Figure 11(a)–(c)
damage accumulation within 50 mm height over the for the reference, strengthened and repaired specimens,
column base after the formation of the interface respectively. For all specimens, initial yielding of longi-
crack. Thus, the damage distribution was not homoge- tudinal rebars was observed at approximately 0.8%
neous along the plastic hinging length. drift and the maximum strain reached 4000 me in both
Ozcan et al. 3417

100
Column–stub interface cracking
Rebar yielding
CFRP debonding
CFRP rupture
50
Onset of rebar buckling
Lateral load (kN)

–50

S-NL-10
–100
–7.5 –5 –2.5 0 2.5 5 7.5
Drift ratio (%)

Figure 6. Lateral load vs. drift ratio response for specimen S-NL-10.

100
Column–stub interface cracking
Rebar yielding
CFRP debonding
CFRP rupture
50
Onset of rebar buckling
Lateral load (kN)

–50

Damaging
cycles
R-NL-30
–100
–7.5 –5 –2.5 0 2.5 5 7.5
Drift ratio (%)

Figure 7. Lateral load vs. drift ratio response for specimen R-NL-30.

loading directions. After unloading the specimens (R- the column base for the retrofitted and reference speci-
NL-30 and R-UL-30) for repairing and reloading mens, respectively (Table 2). The pictures of damaged
from initial zero tip-displacement level, yielding of the plastic hinge regions of specimens are shown in Figure
longitudinal bars was observed around 1.7% drift ratio. 12(a)–(e).
The strain levels monitored on reinforcing bars 350 mm
away from the column base were considerably close to
Stiffness degradation and energy dissipation
the yielding level for all specimens that can be regarded
as a strong evidence that proves the formation of plastic The secant-stiffness values were calculated by dividing
hinging zone along approximately 300–350 mm over the lateral load to the corresponding lateral deflections
3418 Journal of Reinforced Plastics and Composites 29(22)

100
Column–stub interface cracking
Rebar yielding
CFRP debonding
CFRP rupture
50
Lateral load (kN) Onset of rebar buckling

–50

Damaging
cycles
R-UL-30
–100
–7.5 –5 –2.5 0 2.5 5 7.5
Drift ratio (%)

Figure 8. Lateral load vs. drift ratio response for specimen R-UL-30.

(a) (b) 1000

Column Units: mm Kavg, 350-50 mm
350
P
Average curvature (rad/km)

Kavg, 350-base
50 Kavg, 50-base
Curvature S-H-0-00
distribution 500
Kavg, 350-50
Kavg, 350-base

Kavg, 50-base 0
0 2 4 6
Drift ratio (%)

(c) 1000 (d) 1000

Kavg, 350-50 mm Kavg, 350-50 mm
Average curvature (rad/km)

Average curvature (rad/km)

Kavg, 350-base Kavg, 350-base

Kavg, 50-base Kavg, 50-base
S-H-1-00 S-HD-1-00
500 500

0 0
0 2 4 6 0 2 4 6
Drift ratio (%) Drift ratio (%)

Figure 9. (a) Average curvature measurement illustration and average curvature–drift ratio plots along plastic hinging region for (b)
unconfined, (c) strengthened, and (d) repaired specimens.
Ozcan et al. 3419

(a) 100 200

REF REF
S-NL-30 S-NL-30
50 S-NL-10 100 S-NL-10

M (kNm)
P (kN)

0 0

–50 –100

–100 –200
–8 –4 0 4 8 –200 –100 0 100 200
Drift (%) Average curvature (rad/km)
(b) 100 200
REF REF
R-NL-30 R-NL-30
50 R-UL-30 100 R-UL-30

M (kNm)
P (kN)

0 0

–50 –100

–100 –200
–8 –4 0 4 8 –200 –100 0 100 200
Drift (%) Average curvature (rad/km)

(c) 100 200

REF REF
S-NL-30 S-NL-30
50 R-NL-10 100 R-NL-30
M (kNm)
P (kN)

0 0

–50 –100

–100 –200
–8 –4 0 4 8 –200 –100 0 100 200
Drift (%) Average curvature (rad/km)

Figure 10. Lateral load vs. drift and moment vs. average curvature envelope comparisons considering the effects of: (a) corner-
rounding radius, (b) repairing under axial load, and (c) damage (to 2% drift ratio).

for each half cycle. The average secant-stiffness values degradation in stiffness. Furthermore, the dissipated
for each drift level were calculated considering the aver- energy in each reversed displacement cycle was calcu-
age stiffness values at all half cycles. As shown in Figure lated and added up to obtain cumulative dissipated
13(a), stiffness degradation responses for the reference energy levels prior to ultimate level at which the lateral
and strengthened specimens were similar until the fail- capacity dropped 20% of peak (Figure 13(b)). The cor-
ure of the reference specimen. However, the repaired responding drift ratios were added up similarly in order
columns could sustain lower stiffness levels as com- to acquire cumulative drift ratios. As can be observed
pared to the strengthened case due to the damage in Figure 13(b), the strengthened columns could dissi-
cycles. The axial load during repairing caused slight pate approximately 1.5 times more energy than the
3420 Journal of Reinforced Plastics and Composites 29(22)

(a) 5000 (b)

REF S-NL-30

2500
Yielding

Strain (me)
0

Yielding
–2500
Drift (%)
–5000
–8 –4 0 4 8 –8 –4 0 4 8
(c) (d)
5000
R-NL-30 R-NL-30
Damage
2500
Yielding

Yielding
–2500

–5000
Bottom strain gages at 50mm height
Top strain gages at 350 mm height

Figure 11. Rebar strain–drift ratio responses for: (a) unconfined, (b) strengthened, and (c) repaired specimens with damage cycles.

reference and the repaired columns for a specific drift was consumed during the damage cycles. Strengthened
level. In addition, for further drift levels beyond the columns dissipated higher amount of energy for a spe-
capacity of strengthened columns, the dissipated cific cumulative drift level. On the other hand, for the
energy level for the repaired columns was almost repaired columns, the dissipated amount of energy
twice that of the reference. tended to reduce concerning the adverse effect of
damage. There was no significant change in secant-
stiffness degradation response of strengthened columns,
Discussion since the stiffness of the columns was not affected con-
siderably after CFRP wrapping. Because CFRP wrap-
Effect of CFRP wrapping ping could not improve the initial stiffness of damaged
Strengthening square columns with one layer of CFRP columns, repaired specimens had much lower initial
sheet improved the ultimate drift levels and conse- stiffness. This phenomenon demonstrates that CFRP
quently the displacement ductility levels by about wrapping cannot prevent stiffness degradation for the
25%, when the corners of the column were sufficiently damaged columns.
rounded. The CFRP strengthened specimens without
any damage could achieve higher levels of displacement
ductility regarding the unwrapped specimen, since
Effect of sustained axial load during repairing
CFRP confinement enforced the slippage/extension of Both of the repaired columns performed similar inelas-
the plain rebars without buckling and led to higher tic deformation responses. The presence of axial load
fixed-end rotation levels. However, a reduction in dis- (27% of axial capacity) during the repairing did not
placement and curvature ductility was observed for the significantly affect the ultimate drift ratios (5.0%).
repaired columns since a significant amount of ductility Likewise, the drift levels at which the CFRP debonding
Ozcan et al. 3421

(a) (b) (c)

Lp =350 mm Lp = 310 mm Lp = 350 mm

(d) (e)
Lp =300 mm Lp =280 mm

Figure 12. General view at failure stage of specimens: (a) REF, (b) S-NL-30, (c) S-NL-10, (d) R-NL-30, and (e) R-UL-30.

(a) (b)
6 60
Cumulative dissipated energy (kNm)
Secant stiffness (kN/m)

4 40

REF
2 S-NL-30 20 REF
S-NL-10 S-NL-30
R-NL-30 S-NL-10
R-UL-30 R-NL-30
R-UL-30
0 0
0 2 4 6 8 0 10 20 30 40 50
Drift ratio (%) Cumulative drift ratio (%)

Figure 13. Specimen responses: energy dissipation and stiffness degradation.

3422 Journal of Reinforced Plastics and Composites 29(22)

and rupture took place were also similar (2% and 5%, and accurate estimated value of the expected deforma-
respectively) for both companion columns. Considering tion capacity of RC columns wrapped with FRPs in
the lateral load-carrying capacities, about 15% increase their plastic hinge regions. That study did not include
can only be acquired in the absence of axial load during the effect of damage on the performance of retrofitted
repairing due to use of high strength epoxy mortar. specimens. Based on the test results presented herein
However, when repair was conducted under sustained and considering previous studies including damage
axial load, the increase in lateral capacity cannot be prior to strengthening,13,23 a modification to this equa-
monitored. This phenomenon indicates that expected tion is presented below. The column database consists
strength increase due to use of high strength repair of six columns with either square or circular cross-
material cannot be relied on, unless the axial force is sections failing in a flexure mode. Certainly, such a
removed. small number of test specimens may not provide the
full characteristics of damage and repair afterwards.
The main objective here is to provide a safe estimate
Effect of corner-rounding radius
of drift ratios based on a limited number of tests.
The reduction in corner-rounding radius from 30 to Expansion of the test database remains to be an impor-
10 mm for the strengthened columns degraded ultimate tant issue. For the equation proposed by Ozcan et al.,22
drift levels. The ultimate drift ratio of S-NL-30 was the parameters of l (longitudinal reinforcement), n
4.1%, whereas it was 3.6% for S-NL-10, keeping in (axial load ratio), and  (confinement ratio) are calcu-
mind that the ultimate drift ratio of the reference speci- lated using Equations (1a) and (1b):
men REF was 3.3%. This response implied that reduc-
ing the corner-rounding radius deteriorated the CFRP Napplied Napplied
n¼ ¼ : ð1aÞ
confined region and diminished the displacement and N0 0:85f 0c Ag þ As fy
curvature ductility. The specimen with reduced corner
radius showed an inferior performance in energy dissi- f 0le
¼
pation characteristics. There was no significant varia- f0
tion observed in secant-stiffness degradation responses 8c  2 2

< ðbþhÞE0j "f tj 1  ðb2rÞ þðh2rÞ , rectangular columns
for both of the strengthened specimens, since reducing bh f c 3bh
¼ :
the corner-rounding radius had an insignificant effect : 2Ej "f tj , circular columns
Df 0c
on lateral strength within the radius range used in
corner rounding. ð1bÞ

Drift-based design methodology where fc 0 , Ag, As, and fy denote concrete compressive
strength, gross concrete area, total steel area, and steel
for repaired columns yield stress, respectively. D, b, h, and r represent column
In the previous study by Ozcan et al.,22 a drift-based diameter, width, height, and corner-rounding radius,
design equation was proposed to estimate the ultimate respectively. The parameters of Ej, ef, and tj are the
drift ratio of fiber reinforced polymer (FRP)-strength- elasticity modulus, rupture strain, and thickness of
ened RC columns using a database of 24 columns from FRP sheet, respectively. Employing the column data-
the literature. The equation was proven to provide safe base (Table 3), the ultimate drift ratios to which the

Table 3. Column database

Research Specimen l (%) n (%) (%) DRd (%) DRpl (%) DRur (%) DRBF BF
u,p (%) DRu,p =DRur DRD D
u,p (%) DRu,p =DRur

Iacobucci et al.13 AS7NS 2.70 33 7.94 1.78 0.370 4.64 4.89 0.95 3.44 1.35
AS8NS 2.70 56 20.84 1.03 0.623 2.83 2.86 0.99 2.38 1.19
Yau23 R1NT 2.96 54 14.78 0.48 0.185 6.35 5.65 1.12 3.11 2.04
R2NT 2.96 54 12.31 0.74 0.275 3.66 4.21 0.87 2.90 1.26
This study R-NL-30 2.48 27 9.29 2.00 0.416 5.20 5.88 0.89 4.08 1.28
R-UL-30 2.48 27 8.82 2.00 0.768 5.00 4.13 1.21 3.31 1.51
Mean 1.004 Mean 1.438
SD 0.136 SD 0.314
DRd: Utmost drift ratio of damage, DRpl: Drift ratio of plastic deformation after damage, DRur : Experimental ultimate drift ratio after repairing, DRBF
u,p , and
DRD
u,p : Ultimate drift ratio predicted by the best fit and design equations.
Ozcan et al. 3423

columns were subjected (DRd) were obtained. In addi- respectively. Both figures support the accuracy and
tion, the maximum plastic deformations (DRpl) during safety of the proposed equations for best fit and
damage cycles were determined. All the columns in the design. As can be observed from Table 3, the applica-
database were damaged up to 2% drift ratio (Table 3) tion of Equations (2b) and (2c) is limited to for a plastic
prior to repair. This level of damage can be thought as damage drift range of 0.18–0.77% and axial load range
a moderate level damage after a frequent earthquake. of 27–56% considering the limited repaired column
Cover spalling and longitudinal reinforcement yielding database.
are expected; however, longitudinal rebar buckling is
not observed. In order to obtain the deformation
capacity of a column which did not experience any
Summary and conclusions
damage, ultimate drift ratio (DRus) equation proposed Results from an experimental study, in which five flex-
by Ozcan et al.22 is employed (Equation (2a)). For a ure-dominated columns were tested under constant
column that experienced some damage and attained a axial load and cyclic lateral displacement excursions
plastic deformation, the ultimate best fit drift ratio that simulated seismic forces, are presented in this
(DRBF D
ur ) and design drift ratio (DRur ) can be calculated study. Four specimens that represented the strength-
using Equation (2b) and (2c), respectively. All units are ened and repaired columns were retrofitted with one
in percents in the following equations: layer of CFRP sheet. The remaining column was
termed as the reference column, since it embodied a
0:64 deficient building column that had plain longitudinal
DRus ¼ 2:47 þ 50 : ð2aÞ rebars, insufficient transverse reinforcement, and mod-
l0:35 n1:29
erate concrete compressive strength. The parameters of
CFRP wrapping, sustained axial load during repairing,


DR0:209 þ3:58n0:064
DRBF
ur ¼ DRus 10 pl : ð2bÞ and corner-rounding radius were studied. The follow-
ing conclusions were drawn within the scope of the tests
! performed:
10ðDRpl þ0:01nÞ
DRD
ur ¼ DRus 1 : ð2cÞ
40 1. Strengthening plastic hinge region of RC columns
with CFRP improved the seismic behavior in terms
of ductility, energy dissipation capacity, and secant-
The obtained results shown in Table 3 demonstrated stiffness degradation responses of the columns as
that the estimations of drift ratio capacity using the long as sufficient corner rounding is employed. For
best-fit drift equation have a mean and standard devi- repaired columns, significant amount of ductility
ation (SD) of 1.004 and 0.136, respectively. For the was consumed during the damage cycles and
design equation, these values reached up to 1.438 and secant-stiffness response was adversely affected.
0.314, considering the fact that the drift capacity of all 2. The sustained axial load level of 27% of axial capac-
test specimens is estimated on the safe side. The com- ity during repairing had an insignificant effect on the
parisons of ultimate drift predictions regarding best fit ultimate drift, ultimate curvature, and ductility of
and design equations are shown in Figure 14(a) and (b), the columns. However, due to the high compressive

(a) 8 (b) 8
Iacobucci et. al.13
Ultimate drift, experimental (%)

Ultimate drift, experimental (%)

Sheikh and Yau12

This study
6 6

4 4

Iacobucci et. al.13

2 2
Sheikh and Yau12
This study
0 0
0 2 4 6 8 0 2 4 6 8
Ultimate drift, predicted by equation (2b) (%) Ultimate drift, predicted by equation (2c) (%)

Figure 14. Comparisons of predicted ultimate drift ratios for: (a) proposed best-fit equation and (b) proposed design equation.
3424 Journal of Reinforced Plastics and Composites 29(22)

strength of the repair mortar regarding the existing 9. Seible F, Priestley MJN, Hegemier GA and Innamorato
concrete, 15% gain in lateral strength was observed D. Seismic retrofitting of RC columns with continuous
in the absence of axial load. For the column repaired carbon fiber jackets. J Compos Constr 1997; 1(2): 52–62.
under sustained axial load, the section capacity was 10. Saadatmanesh H, Ehsani MR and Jin L. Repair of earth-
quake – damaged RC columns with FRP wraps. ACI
not affected since the axial load was not perceived by
Struct J 1997; 94(2): 206–215.
the repaired section of the column.
11. Xiao Y and Ma R. Seismic retrofit of RC circular col-
3. For a predefined drift level, the energy dissipation umns using prefabricated composite jacketing. J Struct
capacities of the repaired columns were lower than Eng 1997; 123(10): 1356–1364.
the strengthened columns due to the pre-formed 12. Sheikh SA and Yau G. Seismic behavior of concrete col-
cracks. However, the repaired columns could dissi- umns confined with steel and fiber-reinforced polymers.
pate the same level of energy at further drift levels. ACI Struct J 2002; 99(1): 72–80.
4. Reduction of corner-rounding radius had an adverse 13. Iacobucci RD, Sheikh SA and Bayrak O. Retrofit of square
concrete columns with carbon-fiber-reinforced-polymer
influence on seismic behavior in terms of ultimate
for seismic resistance. ACI Struct J 2003; 100(6): 785–793.
drift, curvature, ductility, and energy dissipation
14. Sause R, Harries KA, Walkup SL, Pessiki S and Ricles
responses for the columns. No significant difference JM. Flexural behavior of concrete columns retrofitted
was observed for secant-stiffness degradation with carbon fiber reinforced polymer jackets. ACI
responses. Struct J 2004; 101(5): 708–716.
5. A simple equation to estimate the drift capacity of 15. Bousias SN, Triantafillou TC, Fardis MN, Spathis L and
damaged RC columns repaired with FRPs was pro- O’Regan BA. Fiber-reinforced polymer retrofitting of
posed based on a limited number of experiments. rectangular reinforced concrete columns with or without
The equation is a modification of a recently pro- corrosion. ACI Struct J 2004; 101(4): 512–520.
posed equation incorporating the effect of damage 16. Harajli MH and Rteil AA. Effect of confinement using
fiber-reinforced polymer or fiber-reinforced concrete on
via the use of plastic deformation.
seismic performance of gravity load-designed columns.
ACI Struct J 2004; 101(1): 47–56.
17. Haroun MA and Elsanadedy HM. Behavior of cyclically
loaded squat reinforced concrete bridge columns
References
upgraded with advanced composite-material jackets. J
1. Bett BJ, Klingner RE and Jirsa JO. Lateral load response Bridge Eng ASCE 2005; 10(6): 741–748.
of strengthened and repaired reinforced concrete columns. 18. Elsanadedy HM and Haroun MA. Seismic design guide-
ACI Struct J 1988; 85(5): 499–508. lines for squat composite-jacketed circular and rectangu-
2. Ersoy U, Tankut AT and Suleiman R. Behavior of lar reinforced concrete bridge columns. ACI Struct J
jacketed columns. ACI Struct J 1993; 90(3): 288–293. 2005; 102(4): 505–514.
3. Julio ES, Branco F and Silva VD. Structural rehabilitation 19. Yalcin C, Kaya O and Sinangil M. Seismic retrofitting of
of columns with reinforced concrete jacketing. Prog Struct R/C columns having plain rebars using CFRP sheets for
Mater Eng 2003; 5: 29–37. improved strength and ductility. Constr Build Mater
4. Vandoros KG and Dritsos SE. Axial preloading effects 2006; 8(17): 295–307.
when reinforced concrete columns are strengthened by 20. Ghosh KK and Sheikh SA. Seismic upgrade with carbon-
concrete jackets. Prog Struct Mater Eng 2006; 8: 79–92. fiber reinforced polymer of columns containing lap-
5. Vandoros KG and Dritsos SE. Concrete jacket construc- spliced reinforcing bars. ACI Struct J 2007; 104(2):
tion detail effectiveness when strengthening RC columns. 227–236.
Constr Build Mater 2008; 22: 264–276. 21. Ozcan O, Binici B and Ozcebe G. Improving seismic per-
6. Chai YH, Priestley MJN and Seible F. Seismic retrofit of formance of deficient reinforced concrete columns using
circular bridge columns for enhanced flexural perfor- carbon fiber reinforced polymers. Eng Struct 2008; 30(6):
mance. ACI Struct J 1991; 88(5): 572–584. 1632–1646.
7. Aboutaha RS and Machado R. Seismic resistance of steel 22. Ozcan O, Binici B and Ozcebe G. Seismic strengthening
confined reinforced concrete (SCRC) columns. Struct Des of rectangular reinforced concrete columns using fiber
Tall Build 1998; 7: 251–260. reinforced polymers. Eng Struct 2010; 32(4): 964–973.
8. Saadatmanesh H, Ehsani MR and Li MW. Strength and 23. Yau G. Repair and strengthening of columns with fibre
ductility of concrete columns externally reinforced with reinforced composites, PhD Dissertation, University of
fiber composite straps. ACI Struct J 1994; 91(4): 434–447. Toronto, 1998.