Composite Structures 70 (2005) 185–199

www.elsevier.com/locate/compstruct

FRPC reinforced concrete beam-column joints under cyclic excitation

Abhijit Mukherjee *, Mangesh Joshi
Department of Civil Engineering, Indian Institute of Technology Bombay, Mumbai 400 076, India

Available online 29 September 2004

Abstract

An investigation on the performance of reinforced concrete beam-column joints under cyclic loading is reported. Joints have been
cast with adequate and deficient bond of reinforcements at the beam-column joint. FRP sheets and strips have been applied on the
joints in different configurations. The columns are subjected to an axial force while the beams are subjected to a cyclic load with
controlled displacement. The amplitude of displacement is increased monotonically using a dynamic actuator. The hysteretic curves
of the specimens have been plotted. The energy dissipation capacity of various FRP configurations has been compared. In addition,
the control specimens have been reused after testing as damaged specimens that are candidates for rehabilitation. The rehabilitation
has been carried out using FRP and their performance has been compared with that of the undamaged specimens.

2004 Elsevier Ltd. All rights reserved.

Keywords: Structural upgradation; Fiber reinforced polymer composites; Concrete; Beam-column joints; Cyclic loading; Hysteretic curves; Energy
absorption

1. Introduction joints, however, is a challenging task that poses major

practical difficulties. A variety of techniques have been
Although hundreds of thousands of successful rein- applied to joints with the most common ones being
forced cement concrete (RCC) framed structures are the construction of RCC or steel jackets [1,2]. Rein-
annually constructed worldwide, there are large num- forced concrete jackets require intensive labor and often
bers of them that deteriorate, or become unsafe due to intricate detailing of steel in the form of diagonal collars.
changes in loading, changes in use, or changes in config- Moreover, concrete jackets increase the dimensions and
uration. Occurrence of natural calamities may also lead weight of structural elements. Few attempts have been
to review of engineering notions that make reworking of made in employing plain or corrugated steel plates
existing structures inevitable. The reworking is variously [3,4]. In addition to corrosion protection, these ele-
referred as repair, rehabilitation, retrofitting, and upgra- ments require special attachment through the use of
dation with distinct meanings attached to all these either epoxy adhesives combined with bolts or special
terminologies. In this paper, we use upgradation to sug- grouting.
gest all these activities. The RCC beam-column joints FRPC based strengthening strategy could be an
may require upgradation due to deficient detailing of attractive option in order to restore the joints. In addi-
reinforcing bars, insufficient column sections or due to tion to being lighter, thinner and easier to implement
increased loading on the structure. Upgradation of these FRPC reinforced joints have the virtue of making the
joints more ductile. This property is extremely desirable
for seismic upgradation of structures. However, a direct
*
Corresponding author. Tel.: +91 22 2576 7343/578 2545; fax: +91
extension of the strategies adopted for beams and
22 576 7302/578 3480. columns is difficult as the behavior of beam-column con-
E-mail address: [email protected] (A. Mukherjee). nections is complex and still not completely understood.

0263-8223/$ - see front matter
2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.compstruct.2004.08.022
186 A. Mukherjee, M. Joshi / Composite Structures 70 (2005) 185–199

Research on this area can be classified into experiments tained experimentally are discussed along with failure
with quasistatic loading on bridge beam-column joints types.
[5,6], typical building frame joints [7,8] and frameworks
[9] report enhancement of both shear and bending capac-
ity. Cyclic load tests on joints with inadequate reinforce- 2. Experimental work
ment bond length [10], rehabilitated specimens [11],
shear strengthened T joints [12–14], have also reported 2.1. Specimen details
substantial improvements in ultimate capacities. A com-
bination of FRP in various forms bars, plates and sheets Two different types of RCC joints have been cast for
in structural enhancement have been investigated [15,16]. experimental verification. One set of joints has adequate
The suitability of adhesives of different moduli [17] has steel reinforcements with proper detailing of reinforce-
also been studied. On some occasions a combination of ments at the critical sections (Fig. 1). In the other set
FRP and steel external reinforcements has been used of specimens the beam reinforcements have deficient
[18]. FRP in seismic retrofitting of strong-beam-weak- bond lengths at the junctions with the columns (Fig.
column frameworks that are prevalent in gravity load de- 2). When the beam of the joint is transversely loaded
signed buildings have been investigated [19]. the first set is characterized by a long plastic zone (duc-
Analytical models of externally reinforced beam-col- tile) while the second set fails in reinforcement pull out
umn joints have been presented using closed form solu- and exhibits sudden failure (non-ductile).
tions [20], and finite element technique [21]. Design The concrete used is of M30 grade (30 N/mm2) and
methods have been suggested for shear [22] and bending properties of other material used are shown in Table 1.
[20]. Pushover analysis has been employed in the design
of upgradation for bridge joints [21] and buildings [23] 2.2. FRP reinforcements
in seismic conditions.
Survey of existing constructions reveals that upgrada- The specimens in Figs. 1 and 2 were strengthened by
tion of structures is necessary in three conditions: using carbon and glass FRP materials. Prior to the
application of the FRP the concrete substrate was
• The structure is inadequately designed for the present smoothed by grinding and cleaning. The cement paste
load conditions. was removed from the surface and the coarse aggregates
• The structure is inadequately detailed for the present were exposed. Acetone was used to clean the surface.
loading. This also includes those structures that are The corners of all the members were ground to create
found deficient under seismic conditions. a minimum radius of 10 mm. Epoxy putty was used to
• The structure is damaged and requires rehabilitation. fill the voids and concave areas. Figs. 3 and 4 present
schematic arrangement for two typical systems; L-over-
This paper includes joints of all the three varieties. lays and precured carbon plates respectively.
The motivation behind this program is to examine the
performance of fiber reinforced polymer composites 2.3. Type A: L-reinforcement with GFRP/CFRP sheets
(FRP) in upgradation of healthy joints with adequate
and deficient reinforcements and also in rehabilitation In this type, FRP sheets were applied in L shape to
of damaged joints. The load deflection relations ob- upgrade the joints. FRP has been applied in several

400mm 400mm
#4 –6mmφ mild steel

3mmφ stirrups @ 100mm c/c

400mm

100

A
100

Section A-A
A
#4 –6mmφmild steel
400mm

3mmφ stirrups @ 100mm c/c

100

B B

100

Section B-B

Fig. 1. Specimen with ductile joint reinforcement.

 A. Mukherjee, M. Joshi / Composite Structures 70 (2005) 185–199 187

400mm 400mm
#4 –6mmφ mild steel

3mmφ stirrups @ 100mm c/c

400mm

100
A
100

Section A-A
A
400mm #4 –6mmφ mild steel
25 mm
3mmφ stirrups @ 100mm c/c

100
B B

100

Section B-B

Fig. 2. Specimen with non-ductile joint reinforcement.

Table 1
Properties of materials
Material Effective thickness (mm) Ultimate strength (MPa) Tensile modulus (GPa) Ultimate strain
Glass—G (fiber) 0.36 2250 70 0.0239
Carbon—C (fiber) 0.11 3500 230 0.0117
Carbon plate—CP (composite) 1.2 2800 165 0.017
Mild steel longitudinal reinforcement 6 mm dia 275 198 0.045
Mild steel transverse reinforcement 3 mm dia 555.13 193 0.043

layers. In step 1 FRP has been applied on the top and

bottom surfaces concrete surfaces. The fibers were along
the axes of the members (Fig. 3a). Subsequently, FRP
FRP Wrap
wraps were provided over the inner layers (Fig. 3b).
FRP Overlays The direction of fibers in wraps was perpendicular to
the axis of the members.
Each joint was used for two different lay ups. Fig. 5a
shows glass fiber sheets (80 mm wide and 250 mm long)
on either side of the joint. On one side only one layer
is provided. On the other side two layers of FRP have
been provided to evaluate the efficacy of different extents
of upgradation. Both the column and the beam are then
(a) (b)
wrapped by unidirectional glass fibers with 100 mm lap
Fig. 3. Type A strengthening system—use of composite overlays. length as shown in step 2 of Fig. 5. Same configuration
is repeated using carbon fiber sheet using 1 and 2 layers
of overlays and single wrap with 100 mm overlap. Both
adequate and deficient joints were reinforced using this
configuration.

CFRP Wrap
2.4. Type B: reinforcement with plates
S512 plate
In this type, procured carbon plate (25 mm wide and
1.2 mm thick) have been used in the beams to improve
bending stiffness. To achieve a good bond between the
plate and the concrete beam at the joint a groove
(25 mm wide and 25 mm deep) has been created inside
the concrete joint. The plates have been inserted into
(a) the joint as shown in Fig. 6. The groove was left in the
(b)
specimens at the time of casting. Surfaces of the con-
Fig. 4. Type B strengthening system—use of precured carbon plate. crete specimens were prepared as discussed earlier. The
188 A. Mukherjee, M. Joshi / Composite Structures 70 (2005) 185–199

GFRP 2L (80X0.72) GFRP 1 Wrap

250

250
250 250
GFRP 1L (80X0.36)
GFRP 2L (80X0.72)

(a) (b)

Fig. 5. Strengthening Type A with GFRP/CFRP Sheets.

Groove 25 mm
CFRP 1 Wrap
CFRP plate (25X1.2 mm)

CFRP plate (25X1.2 mm)

450

(a) (b)

Fig. 6. Strengthening Type B.

groove was filled by injecting epoxy resin, and the plates Table 3
were inserted in grove as shown in step 1 of Fig. 6. The Test matrix for non-ductile specimen
beams and columns were then wrapped using a single S. Specimen Details
wrap of carbon sheet. no. name
1. ND-1 –
2.5. Rehabilitated specimens 2 G1L-ND Type A with single L of GFRP at top and bottom
3 G2L-ND Type A with two L of GFRP at top and bottom
4 C1L-ND Type A with single L of CFRP at top and bottom
The control specimens without FRP were used after 5 C2L-ND Type A with two L of CFRP at top and bottom
testing to evaluate the rehabilitation of joints with 6 CP1-D Type B with CFRP plate at top and bottom
FRP. From the failed specimens loose concrete was re-
moved and the cracks and voids were filled with epoxy 2.6. Test program
and sand mortar. The smooth surfaces were restored
prior to the application of FRP. Grooves were cut up The experimental setup is as shown in Fig. 7. The col-
to 25 mm depth using a rotary concrete cutter. Type B umn was fixed at its ends on a loading frame. It was sub-
strengthening system with precured carbon plates was jected to a constant axial load of 100 KN which is 50%
used. of ultimate load carrying capacity of the column. This
The detailed test matrix for adequate and deficient is considered the service load that the column is expected
specimens is presented in Tables 2 and 3 respectively. to carry under normal loading conditions.
Cyclic load was applied using a hydraulic actuator.
The load cycle was predefined as shown in Fig. 8. The
displacement started from the neutral position and it
Table 2
oscillated harmonically about that position. It increased
Test matrix for ductile specimen
at a uniform rate 0.25 mm/cycle. Each cycle comprises of
S. Specimen Details
three full waves of same amplitude in 10 s (0.3 Hz. fre-
no. name
quency). The final tip deflection was 22.5 mm. The other
1 D-1 –
end of the beam was free. Vertical deflection of the tip of
2 G1L-D Type A with single L of GFRP at top and bottom
3 G2L-D Type A with two L of GFRP at top and bottom the beam was recorded directly by the linear variable
4 C1L-D Type A with single L of CFRP at top and bottom displacement transducer (LVDT) built in the actuator.
5 C2L-D Type A with two L of CFRP at top and bottom It was validated with another external LVDT. The data
6 CP1-D Type B with CFRP plate at top and bottom were collected using a computerized data acquisition
7 Rehab. Type B with CFRP plate at top and bottom
system.
 A. Mukherjee, M. Joshi / Composite Structures 70 (2005) 185–199 189

Fig. 7. Experimental setup.

3. Results and discussion different systems studied here the tip load–displacement
envelopes are plotted by joining the peaks of consecu-
The beam tip load–displacement characteristics for tive loops. These plots have better clarity. The rate of
all the specimens are discussed here. The displacement stiffness degradation can also be found out from these
levels of the first few cycles did not generate any nonlin- plots.
ear deformation in the structure and the loops followed Ability of the structure to survive an earthquake de-
a straight line with its slope as initial stiffness. The onset pends to a large extent on its ability to dissipate the input
of stiffness degradation was identified by simultaneous energy. Forms of energy dissipation include kinetic en-
appearance of tension cracks at the root of the cantilever ergy, viscous damping and hysteretic damping, etc. An
beam. The calculations show that at this point the steel estimate of the hysteretic damping can be found by the
started to yield and it was not capable of taking any fur- area enclosed in the load–displacement hysteresis loops.
ther load. The additional load from this point was It may be noted that a wider loop (i.e. a large difference
carried by the FRP. At this point linearity of the ascend- in ordinates in the ascending and the descending paths)
ing and the descending paths was lost and loops between would signify higher hysteretic damping. Cumulative en-
the two paths appeared. We term this phenomenon ergy dissipated was then calculated by summing up the en-
yield. The post yield behavior is signified by monotonic ergy dissipated in consecutive loops through out the test.
degradation of stiffness. To enable comparison among Failure modes for different specimens are also discussed.
190 A. Mukherjee, M. Joshi / Composite Structures 70 (2005) 185–199

30
ultimate strain in glass is also considerably higher than
20
that in carbon. The CP1 specimens do not follow that
trend. The authors believe that better bonding of CP1
Displacement (mm)

10
specimens through grooves has delayed the onset of
0
yield in them.
The load–displacement envelopes for ductile speci-
-10
men are plotted in Fig. 10. The envelopes let us compare
-20
the relative performance of the specimens. All the FRP
reinforced specimens had higher peak loads than the
-30
0 100 200 300 400 500 control specimen. The CP1 specimen had the highest
Time(Sec) peak load followed by the C2L, G2L and G1L speci-
(a) Full cycle mens. The graph shows the superiority of Type-B
arrangement over Type-A reinforcements.
2
The rate of stiffness degradation is plotted in Fig. 11.
All the FRP reinforced specimens had a total loss of
1
Displacement (mm)

stiffness at a higher displacement level than the control

specimen. This is a highly desirable phenomenon be-
0
cause the joint collapse can be deferred through FRP
reinforcement. The carbon specimens have higher initial
-1
stiffness and slower rate of degradation. The CP1 speci-
mens showed some initial gain in stiffness. The authors
-2
believe that this is due to the rearrangement in the adhe-
| | | | | | | | |

0 5
| |
15
|
25
|
35
|
45
|
55
|
65
|
75
|
85
| sively bonded groove. The initial stiffness and the ulti-
Time(sec) mate displacements are summarized in Table 5.
(b) Close view of cycle The overall performance of the joints can be summa-
rized by the energy dissipation of the structure. The cumu-
Fig. 8. Predefined displacement cycle applied to beam end.
lative energy dissipation graphs for the specimens are
presented in Fig. 12. The energy dissipation of the FRP
3.1. Ductile specimens reinforced specimens follows closely that of the control
specimen. The dissipation of energy is mainly through
Beam tip load–displacement graphs for the ductile the yielding if steel. It may be remembered that the FRPs
specimens are plotted in Fig. 9. From the graphs it remain elastic until failure. Therefore, not much dissipa-
can be seen that the load at yield was considerably tion of energy is expected through the deformation of
higher in the FRP reinforced specimens than the control the FRP. The only agent of energy dissipation in FRP
specimen. Column 3 of Table 4 summarizes the percent- at this stage is debonding and delamination and the result-
age increase in the yield load. The CP1-D exhibited the ing Coulomb damping. However, the FRPs have in-
highest increase in the yield load followed by the C2L- creased the ultimate deformation of the structure to a
D, G1L-D and G2L-D specimens. It may be noted that large extent. Through higher deformation the FRP rein-
the forces at the tensile face of the beam are shared by forced structures have exhibited much higher dissipation
the steel and FRP in proportion of their relative stiff- of energy. The reasons of higher ultimate deformation
nesses. The stiffness of carbon is considerably higher in FRP reinforced specimens will be discussed in conjunc-
than that of glass. Therefore, for the same tip load, tion with the failure mode. The carbon specimens exhibi-
the tensile force in steel is lower in the carbon reinforced ted higher energy dissipation than the glass specimens.
specimen than in the glass reinforced specimens. As a re- The photographs of the failed specimens have been
sult, the steel in the carbon reinforced specimens yields presented in Fig. 13. In Fig. 13a the failed control spec-
at higher tip loads. The CP1-D specimens are anchored imen is shown. It may be noted that the beam has failed
at the joint through a groove. Therefore, they exhibit at the joint through the formation of a hinge. The hinge
higher stiffness than other sheet specimens. The speci- has formed between the two shear links of the beam.
mens with two-layer reinforcement had higher yield The concrete has spalled in such a fashion that two semi-
loads than the specimens with one layer reinforcements. circular surfaces have been created. This has resulted in
Due to FRP reinforcements the displacement at yield free rotation of the beam limb with no transfer of bend-
increased to a much lesser extent than the load (Column ing moment to the column. The reinforcements have
5, Table 4). Another interesting point is that the glass been exposed due to spalling of concrete. This has re-
reinforced specimens had much higher displacement at sulted in kinking of the reinforcements. Due to spalling
yield than the carbon reinforced specimens. This is due the cracks were too wide and therefore, they did not
to the higher stiffness of carbon than that of glass. The come in contact in compression. The FRP reinforced
 A. Mukherjee, M. Joshi / Composite Structures 70 (2005) 185–199 191

4
Control 4
G1L-D

2
2

Load (KN)
Load (KN)

0 0

-2
-2

-4
-4

-6
-15- 10 -5 0 5 10 15 -20 -10 0 10 20

(a) Displacement (mm) (b) Displacement (mm)

6 6

G2L-D C2L-D
4 4

2 2
Load (KN)

Load (KN)
0
0

-2
-2

-4
-4

-6
-6
-20 -10 0 10 20 -20 -10 0 10 20

(c) Displacement (mm) (d) Displacement (mm)

8
CP 1-D
6

2
Load (KN)

-2

-4

-6

-8

-20 -10 0 10 20

(e) Displacement (mm)

Fig. 9. Load versus deflection for different non-ductile specimens.

Table 4 specimens, on the other hand, did not have the semicir-
Yield points of ductile specimens cular failure planes. The failure planes were approxi-
Specimen Yield load (N) Deflection at yield load mately vertical (Fig. 13b–d). The authors believe that
(mm) the difference in the failure mode is due to the presence
Value % increase Value % increase of the FRP wraps. They confined the concrete and did
Control 3.8 – 1.9 – not allow it to spall. As a result, in cyclic opening and
4.15 1.86 closing of cracks the crack surfaces could react against
G1L-D 4.61 21.32 1.71 10.00 each other in compression. Therefore, the moment
4.65 12.05 2.14 15.05
resistance of the beam was not lost. This offered higher
G2L-D 5.64 48.42 3.02 58.95
5.54 33.49 2.44 31.18 collapse loads and displacement. To summarize, the
C2L-D 6.0 57.89 1.84 3.16 main cause of superior performance of the FRP rein-
5.53 33.25 2.0 7.53 forced joints is the continuous confinement provided
CP1-D 8.215 116.18 4.975 161.84 by the FRP wraps impeded the creation of hinge
7.67 84.82 2.87 54.3
through the spalling of concrete.
192 A. Mukherjee, M. Joshi / Composite Structures 70 (2005) 185–199

10 6.5
6.0
8 CP 1
5.5
6

Energy dissipation (KN-m)

5.0

4 4.5
4.0
2 Control G2L
Load (KN)

3.5
G2L
0 3.0
G1L 2.5
-2 C2L
C2L G1L
CP 1
2.0
-4 Control
1.5
-6 1.0
0.5
-8
0.0
-25- 20 -15- 10 -5 0 5 10 15 20 25 0 2 4 6 8 10 12 14 16 18

Displacement (mm) Displacement (mm)

Fig. 10. Load versus deflection envelope plots for ductile specimens. Fig. 12. Energy dissipation versus displacement plots for ductile
specimens.

6.0
5.5 could be created through large aggregates. The mechan-
5.0 ical anchorage of the large aggregate would help increas-
4.5
C2L ing the pull out forces. However, this point needs to be
4.0
verified through tests. Another difference in these speci-
Slope (KN/mm)

3.5 CP 1
3.0 mens is that the steel reinforcements had torn com-
2.5 G2L pletely through cup and cone formation. This shows
2.0
1.5
G1L
Control
that the deformations in CP1 specimens were large en-
1.0 ough to reach the ultimate strain of steel. As a result,
0.5 the ultimate failure of the joint was not due to the yield-
0.0
0 2 4 6 8 10 12 14 16 18 20 22 ing of steel but the pull out of the CFRP plate.
Displacement (mm)

Fig. 11. Stiffness versus deflection plots for ductile specimens.

3.2. Rehabbed specimens

A beam tip load–displacement relationship for

Fig. 13d shows the failure of CP1 specimens. In this undamaged ductile specimen and rehabbed specimen is
case the CFRP plates were adhesively bonded into the plotted in Fig. 14. Also the load versus displacement
joint by inserting them inside a grove at the joint. The envelops for the same specimens are plotted in Fig. 15.
figure shows that the carbon plates have pulled out of It shows that the use of composite system did not only
the joint and there is no damage to the plate. A thin restore the original capacity of damaged specimen, but
layer of cement paste is on the surface of the pulled also upgraded the ultimate load capacity by 55%. Also
out portion. This shows that the pullout was due to displacement at ultimate load increased by 30%. Stiff-
the failure of the cement paste in the groove. The ness versus displacement curves (Fig. 16) show that
authors expect that this mode will change when larger there is 48% increase in initial stiffness. An increment
specimens would be tested; because then the groove of 57% is observed in energy dissipation capacity of

Table 5
Ultimate points in ductile specimens
Specimen Initial stiffness Ultimate deflection Energy dissipation capacity
(KN m)
Value (KN/mm) % increase Value (mm) % increase Value % increase
Control 2.32 – 12.7 – 2.37 –
12.73
G1L-D 2.80 17.14 14.56 14.65 4.68 97.47
15.86 24.59
G2L-D 4.06 75.00 16.41 29.21 4.18 76.37
17.78 39.67
C2L-D 5.59 140.94 18.14 42.83 5.96 151.48
17.86 40.30
CP1-D 3.28 41.37 15.32 20.63 5.21 119.83
17.61 38.42
 A. Mukherjee, M. Joshi / Composite Structures 70 (2005) 185–199 193

Fig. 13. Failure patterns for ductile specimens.

4 6

4
2
2
Load (KN)

Load (KN)

0 0

-2
-2
-4

-4 -6

-15 -10 -5 0 5 10 15 -10 0 10

Displacement (mm) Displacement (mm)

(a) Undamaged (b) Rehabbed

Fig. 14. Load versus deflection for undamaged and rehabbed specimens.

structure after the rehabilitation (Fig. 17). The detailed can be seen that the load at yield was considerably
results of rehabbed specimen are given in Tables 6 and 7. higher in the FRP reinforced specimens than the con-
trol specimen. Column 3 of Table 8 summarizes the per-
3.3. Non-ductile specimens centage increase in the yield load. The G2L-ND
exhibited the highest increase in the yield load followed
Beam tip load–displacement graphs for the non-duc- by the C2L-ND, CP1-ND and C1L-ND specimens.
tile specimens are plotted in Fig. 18. From the graphs it The specimens with two-layer reinforcement had higher
194 A. Mukherjee, M. Joshi / Composite Structures 70 (2005) 185–199

6 Rehabbed specimen

Undamaged specimen
4

Load (KN) 0

-2

-4

-6

-15 -10 -5 0 5 10 15
Displacement (mm)

Fig. 15. Load versus deflection envelope plots for rehabbed and undamaged specimen.

3.5 4.0

3.5
3.0 Energy dissipation (KN-m)
3.0
2.5
Stiffness (KN/mm)

2.5
2.0
2.0
Rehabbed
1.5
Undamaged 1.5
Rehabbed
1.0
1.0
Undamaged
0.5 0.5

0.0 0.0
0 2 4 6 8 10 12 14 0 2 4 6 8 10 12 14 16 18

Displacement (mm) Displacement (mm)

Fig. 16. Stiffness versus deflection plots for undamaged and rehabbed Fig. 17. Energy dissipation versus deflection plots for undamaged and
Specimens. rehabbed specimens.

Table 6
Yield points of undamaged and rehabilitated specimens
Strengthening configuration Yield load Deflection at yield load
Value (N) % increase Value (mm) % increase
Undamaged 3.8 – 1.9 –
4.15 1.86
Rehabbed 6.31 66.05 2.46 29.47
5.97 43.86 2.42 30.11

Table 7
Ultimate points in undamaged and rehabilitated specimens
Specimen Initial stiffness Ultimate deflection Energy dissipation capacity
(KN m)
Value (KN/mm) % increase Value (mm) % increase Value % increase
Control 2.32 – 12.7 – 2.37 –
12.73
Rehabbed 3.43 47.84 12.38 2.52 3.71 56.54
14.97 17.6
 A. Mukherjee, M. Joshi / Composite Structures 70 (2005) 185–199 195

4 8

3
Control ND 6
G2L-ND
2 4

1 2

Load (KN)
Load (KN)

0 0

-1 -2

-2 -4

-3 -6

-4 -8
-10 10 -20 -10 0 10 20
0
(a) Displacement (mm) (b) Displacement (mm)

4
C1L-ND 6
C2L-ND
4
2

Load (KN)
2
Load (KN)

0 0

-2
-2
-4

-6
-4

-8
-20 -10 0 10 20 -20 -10 0 10 20

(c) Displacement (mm) (d) Displacement (mm)

6
CP 1-ND
4

2
Load (KN)

-2

-4

-6

-20 -10 0 10 20
(e) Displacement (mm)

Fig. 18. Load versus deflection for different non-ductile specimens.

yield loads than the specimens with one layer rein- pare the relative performance of the specimens. All the
forcements. FRP reinforced specimens had larger areas under the
Due to FRP reinforcements the displacement at yield envelopes than the control specimen. The G2L-ND
increased to a much lesser extent than the load (Column specimen had the largest envelope area followed by the
5, Table 8). Similar to the previous ductile specimens the C2L-ND, CP1-ND and C1L-ND specimens. The graph
glass reinforced specimens had a much higher displace- shows superiority of glass reinforcements over carbon
ment at yield than the carbon reinforced specimens. reinforcement in case of non-ductile specimen.
The CP1-ND specimens do not follow that trend. The The rate of stiffness degradation is plotted in Fig. 20.
authors believe that better bonding of CP1-ND speci- All the FRP reinforced specimens had a total loss of
mens through grooves has delayed the onset of yield in stiffness at a higher displacement level than the control
them. specimen. This is a highly desirable phenomenon be-
The load–displacement envelopes for the ductile spec- cause the joint collapse can be deferred through FRP
imens are plotted in Fig. 19. The envelopes let us com- reinforcement. The glass specimens have higher initial
196 A. Mukherjee, M. Joshi / Composite Structures 70 (2005) 185–199

Table 8
Test results for non-ductile specimens
Specimens Yield load Deflection at yield load
Value (N) % increase Value (mm) % increase
Control-ND 3.73 – 1.64 –
3.6 1.70
G2L-ND 7.6 103.75 1.829 15.24
7.43 106.38 2.195 29.11
C1L-ND 4.2 12.60 1.34 18.29
4.14 10.99 1.46 14.11
C2L-ND 6.69 79.36 1.71 4.27
6.67 85.28 2.1 23.53
CP1-ND 6.28 68.36 2.92 78.05
6.39 77.5 4.02 136.47

9 5.5
8
G2L 5.0
7
6 CP 1
4.5 G2L
5 C2L
4 4.0 C2L
3
2 3.5 C1L

Slope (KN/mm)
1
Load (KN)

3.0
0
-1 2.5
-2
-3 2.0 CP1
-4 Control
1.5
-5 C1L Control
-6 1.0
-7
-8 0.5
-9
0.0
-25 -20 -15 -10 -5 0 5 10 15 20 25
0 2 4 6 8 10 12 14 16 18 20
Dispalcement (mm)
Displacement (mm)

Fig. 19. Load versus deflection envelope plots for non-ductile Fig. 20. Stiffness versus deflection plots for non-ductile specimens.
specimens.

stiffness and slower rate of degradation. The CP1 speci- cumulative energy dissipation graphs for the specimens
mens showed highest stiffness between displacement lev- are presented in Fig. 21. The energy dissipation of the
els of 3–6 mm. The authors believe that this is due to the FRP reinforced specimens follows closely that of the
rearrangement in the adhesively bonded groove. The ini- control specimen. The dissipation of energy is mainly
tial stiffness and the ultimate displacements are summa- through the yielding if steel. In absence of adequate
rized in Table 9. development length considerable loss in energy dissipa-
The overall performance of the joints can be summa- tion capacity of control specimen is observed. FRPs
rized by the energy dissipation of the structure. The have increased the ultimate deformation of the structure

Table 9
Ultimate points in non-ductile specimens
Specimen Initial stiffness Ultimate deflection Energy dissipation capacity
(KN m)
Value (KN/mm) % increase Value (mm) % increase Value % increase
Control-ND 3.57 – 12.81 – 0.695 –
13.11
G2L-ND 5.05 41.46 15.43 20.45 2.196 215.97
16.04 22.35
C1L-ND 3.91 9.52 19.87 55.11 1.42 104.32
20.30 55.07
C2L-ND 4.73 32.49 18.11 41.37 1.47 111.51
18.72 42.79
CP1-ND 2.1 41.17 14.87 16.08 1.116 60.57
15.01 14.49
 A. Mukherjee, M. Joshi / Composite Structures 70 (2005) 185–199 197

1.6 to a large extent. Through higher deformation the

1.5
1.4 FRP reinforced structures have exhibited much higher
1.3 C1L
dissipation of energy. The reasons of higher ultimate
Energy dissipation (KN-m)

1.2
1.1 deformation in FRP reinforced specimens will be
1.0
0.9
G2 L
discussed in conjunction with the failure mode. The
0.8
0.7 CP 1
glass reinforced specimens exhibited higher energy dissi-
0.6 Control pation than the carbon-reinforced specimens. The rela-
0.5 C2L
0.4
tive increase in energy dissipation capacity is more in
0.3 case of non-ductile specimen over the ductile specimen,
0.2
0.1 due to lower energy dissipation capacity of control
0.0
0 2 4 6 8 10 12 14 16 18 20 22 24
specimen.
Displacement (mm) The photographs of the failed specimens have been
presented in Fig. 22. In Fig. 22a the failed control spec-
Fig. 21. Energy dissipation versus displacement plots for non-ductile imen is shown. It may be noted that the beam has failed
specimens.

Fig. 22. Failure patterns for non-ductile specimens.

198 A. Mukherjee, M. Joshi / Composite Structures 70 (2005) 185–199

at the joint by pull out of steel reinforcement. Lack of financial support received from Board of Nuclear re-
development length prevented the formation of plastic search, India is thankfully acknowledged.
hinge. The FRP reinforced specimens, on the other
hand, did not have the reinforcement pull out. The fail-
ure planes were approximately vertical (Fig. 22b–d). The References
authors believe that the difference in the failure mode is
due to the presence of the FRP overlays and wraps. [1] Alcocer S, Jisra JO. Strengthening of reinforced concrete frame
They confined the concrete and did not allow it to spall. connection rehabilitated by jacketing. ACI Struct J 1993;90(3):
249–61.
FRP Overlays helped to hold the beam in its position. [2] Tsonos AG. Lateral load response of strengthened rein-
As a result, the cracked concrete surfaces closed and forced concrete beam to column joint. ACI Struct J 1999;96(1):
they could react against each other at subsequent cycles. 46–56.
Therefore, the moment resistance of the beam was not [3] Beres A, EI-Borgi S, White RN, Gergely P. Experimental results
of repaired and retrofitted beam-column joint tests in lightly
lost. This offered higher collapse loads and displace-
reinforced concrete frame building. Technical report NCEER-92-
ment. To summarize, the main cause of superior per- 0025, State University New York at Buffalo, 1992.
formance of the FRP reinforced joints is the [4] Ghobarah A, Aziz TS, Biddah A. Rehabilitation reinforced
continuous confinement provided by the FRP wraps concrete frame connections using corrugated steel jacketing. ACI
and overlays impeded the pull out of reinforcement Struct J 1997;4(3):283–94.
through the inadequately detailed specimens. [5] Pantelides CP, Gergely I, Reaveley LD, Nuismer RJ. Rehabili-
tation of cap beam-column joints with carbon fiber jackets. Proc
Fig. 22d shows the failure of CP1-ND specimens. The 3rd Int Symp on non-metallic (FRP) reinforcement for concrete
failure pattern is similar to its counterpart ductile spec- structures, Sapparo Japan, vol. 1. Tokyo: Japan Concrete Insti-
imen discussed previously. The figure shows that both tute; 1997. p. 587–95.
the CFRP plates and steel reinforcement have pulled [6] Pantelides CP, Gergely J, Reaveley LD, Volnyy VA. Retrofit of
out of the joint and there is no damage to the plate. A PC bridge pier with CFRP advance composites. J Struct Eng
ASCE 1999;125(10):1094–9.
thin layer of cement paste is on the surface of the pulled [7] Li J, Bakoss SL, Samali B, Ye L. Reinforcement of concrete
out portion. This shows that the pullout was due to the beam-column connections with hybrid FRP sheet. Compos Struct
failure of the cement paste in the groove. As a result, the 2000;47:805–12.
ultimate failure of the joint was due to the pull out of [8] Granata JP, Parvin A. An experimental study on Kevlar
both the steel bars and the CFRP plate. strengthening of beam column connection. Compos Struct 2001;
53:163–71.
[9] Castellani A, Negro P, Colombo A, Grandi A, Ghisalberti G,
Castellani M. Carbon fiber reinforced polymers (CFRP) for
strengthening and repairing under seismic actions. Special Pub-
4. Conclusions lication No. I.99.41, European Laboratory for Structural Assess-
ment, Joint Research Center, Ispra, Italy, 1999.
Both glass and carbon composite materials can be [10] Geng ZJ, Chajes MJ, Chou TW, Pan DYC. The retrofitting of
reinforced concrete column-to-beam connections. Compos Sci
efficiently used for seismic retrofitting as well as rehabil- Technol 1998;58:1297–305.
itation of reinforced concrete joints. Joints exhibit en- [11] Mosallam A. Seismic repair and upgrade of structural capacity of
hanced strength regardless of reinforcement detailing reinforced concrete connections: Another opportunity for poly-
and damage state. Considerable increase in yield load mer composites. In: Proc Int Composites Expo. Cincinnati: 1999.
can be achieved by use of these materials. Yield load p. 1–8.
[12] Gergely I, Pantelides CP, Nuismer RJ, Reaveley LD. Bridge pier
and initial stiffness depends on numbers of overlays pro- retrofit using fiber reinforced plastic composite. J Compos
vided in the joint area. Specimens strengthened using Construct, ASCE 1998;2(4):165–74.
CFRPs show stiffer behavior than GFRP strengthened [13] Gergely J, Pantelides CP, Reaveley LD. Shear strengthening of
specimens. Energy dissipation capacity can be increased RCT-joints using CFRP composites. J Compos Construct, ASCE
with the use of small amount of composites. Test on 2000;4(2):56–64.
[14] Ghobarah A, Said A. Shear strengthening of beam-column Joints.
rehabbed specimen suggests that FRP not only restores Eng Struct 2001;24:881–8.
its original strength but there is considerable enhance- [15] Prota A, Nani A, Manfredi G, Cosenza E. Seismic upgrade of
ment in its yield load, initial stiffness and energy dissipa- beam-column joints with FRP reinforcement . Industria Italiana
tion capacity. del Cemento, November 2000. p. 1–17.
[16] Antonopoulos CP, Triantafillou TC. Experimental investigation
of FRP-strengthened RC beam-column joints. J Compos Con-
struct 2003;7(1):39–49.
[17] Mosallam SA. Strength and ductility of reinforced concrete
Acknowledgment moment frame connections strengthened with quasi-isotropic
laminates. Composites: Part B 2000;31:481–97.
The authors would like to acknowledge the donation [18] El-Amoury T, Ghobarah A. Seismic rehabilitation of beam-
of fiber and adhesive materials from Sika AG. The column joint using GFRP sheets. Eng Struct 2002;24:1397–407.
 A. Mukherjee, M. Joshi / Composite Structures 70 (2005) 185–199 199

[19] DÕAyala D, Penford A, Valentini S. Use of FRP fabric for [22] Gergely J, Pantelides CP, Reaveley LD. Shear strengthening of
strengthening of Reinforced concrete beam-column joints. In: 10th RCT-joints using CFRP composites. J Compos Construct ASCE
Int. conference on structural faults and repair. London: July 2003. 2000;4(2):56–64.
[20] Antonopoulos CP, Triantafillou TC. Analysis of FRP-strength- [23] Mukerjee A, Kalyani AR. Upgradation design of reinforced
ened RC beam-column joints. J Compos Construct 2002;6(1): concrete frames with fiber reinforced composites. J Compos
41–50. Construct ASCE, submitted for publication.
[21] Granata JP, Parvin A. Investigation on the effects of fiber
composites at concrete joints. Composites: Part B 2000;31:
499–509.