SEISMIC RETROFIT OF RC CIRCULAR COLUMNS USING

PREFABRICATED COMPOSITE JACKETING

By Van Xiao, l Member, ASCE, and Rui Ma2

ABSTRACT: This paper describes experimental and theoretical studies on seismic retrofit of reinforced concrete
circular columns with poor lap-splice details using prefabricated composite jacketing. Three 1:2 scale model
columns have been tested. One column was tested under the condition of "as built" and two others were tested
after being retrofitted using prefabricated composite jacketing. The as-built column suffered brittle failure due
Downloaded from ascelibrary.org by INDIAN INSTITUTE OF TECHNOLOGY on 06/08/14. Copyright ASCE. For personal use only; all rights reserved.

to the deterioration of lap-spliced longitudinal reinforcement without developing its flexural capacity or any
ductility. The retrofitted columns showed significant improvement in seismic performance. The failed as-built
column was retested after being repaired. The repaired column also demonstrated significant improvement in
ductility. An analytical model, which takes into consideration the bond-slip deterioration of lap-spliced longi-
tudinal bars, has been developed for seismic assessment and retrofit design.

INTRODUCTION carbon fiber sheets were bonded to the concrete surface using
epoxy resin. Another composite wrapping system using E-
Several recent destructive earthquakes have indicated that glass fiber, which is much more economical than carbon fiber,
bridges designed and constructed based on older seismic de- has been experimentally studied by Priestley and Seible (1991)
sign provisions are vulnerable to catastrophic collapse result- and Seible and Priestly (1993). Priestley et aI.' s test results on
ing from the failure of reinforced concrete columns (Priestley 40% scale bridge piers wrapped with the glass fiber composite
1987; "Loma" 1990; Preliminary 1995). Due to the practice jacketing demonstrated significant improvement of seismic
of using elastic analyses along with much smaller earthquake performance with increased strength and ductility. Priestley
forces compared to current design standards, columns in many and Seible also developed a full design package for seismic
existing bridges typically have the following potential prob- retrofit of existing columns using different retrofit jacketing
lems: systems. Saadatmanesh et al. (1994) have proposed a wrapping
technique using glass fiber composite straps for column re-
1. Undependable flexural capacity due to poor details in trofit. Most recently, Seible et al. (1995) have experimentally
longitudinal lap splices validated a carbon fiber retrofit system that uses an automated
2. Insufficient ductility due to improper transverse confine- machine to wrap carbon bundles to form a continuous jacket.
ment Successful field construction demonstration is also reported by
3. Insufficient shear strength Seible et al. (1995).
4. Improper details and insufficient strength in the column! These composite retrofit measures can be categorized as in-
footing and column/superstructure joints situ fabricated jacketing that involves hand or automated ma-
chine placement of epoxy saturated glass or carbon fabrics on
There is an urgent need to upgrade existing older bridges the surface of existing concrete. An in-situ fabricated jacket
to current seismic design standards in regions with high seis- can match the shape of the existing column. However, due to
micity. Steel jacketing has been proved to be an effective mea- the fact of in-situ fabrication, these systems may need special
sure to retrofit bridge columns for increased strength and duc- attention to the jobsite quality control and curing of the com-
tility (Chai et al. 1991; Priestley et al. 1994; Xiao et al. 1996). posite jackets.
Although steel jacketing has been widely used in practice in A prefabricated composite jacketing system for retrofitting
California and elsewhere, the society is also looking for other reinforced concrete columns has been recently investigated at
alternatives to improve the retrofitting process for the vast the University of Southern California (USC). The retrofit sys-
number of existing, structurally deficient bridges both in the tem uses a series of prefabricated E-glass fiber reinforced com-
United States and throughout the world. One of the key goals posite cylindrical shells with slits. When a column is retrofit-
is to ease construction. ted, the shells are opened and clamped around the column in
Advanced composite materials have been recently recog- sequences with their slits staggered. Adhesive is applied to
nized and applied to bridge retrofit. The general expectations bond the shells to each other and to the column to form a
from composite retrofit systems include light weight, high continuous jacket. The slit for each layer is not butt-bonded
stiffness or strength to weight ratios, etc. Several composite and the continuity relies on the subsequent layer. For this rea-
jacketing systems have been developed and validated in lab- son, the effective layer number is considered as the total num-
oratory or field conditions. Matsuda et al. (1990) tested a sys- ber of installed layers subtracting the last layer. The prefab-
tem for bridge pier retrofit using unidirectional carbon fiber ricated jacketing system is expected to have superior
sheets wrapped longitudinally and transversely in the potential constructability in terms of the quality control and the speed
plastic hinge region or in the region of main bar cutoff. The of installation. This paper describes the experimental and an-
alytical results from a research program designed to validate
'Asst. Prof.. Dept. of Civ. Engrg., Univ. of Southern California. 3620 the effectiveness of the prefabricated composite jacketing sys-
S. Vermont Ave., Los Angeles, CA 90089-2531. tem for improving flexural ductility of bridge columns with
'Grad. Res. Asst., Dept. of Civ. Engrg.• Univ. of Southern California.
3620 S. Vermont Ave., Los Angeles, CA.
lap-spliced longitudinal reinforcement.
Note. Associate Editor: John B. Mander. Discussion open until March
I. 1998. To extend the closing date one month, a written request must EXPERIMENTAL PROGRAM
be filed with the ASCE Manager of Journals. The manuscript for this
paper was submitted for review and possible publication on May 24. Model Column Design
1996. This paper is part of the ]ourTUlI of Structural Engineering. Vol.
123. No. 10. October. 1997. ©ASCE, ISSN 0733-9445/97/0010-1357- Three model columns have been constructed and tested. The
1364/$4.00 + $.50 per page. Paper No. 13339. model columns were designed based on a 1:2 scale of typical
JOURNAL OF STRUCTURAL ENGINEERING / OCTOBER 1997/1357

J. Struct. Eng. 1997.123:1357-1364.

 I Vertical Load 610

610 T 610

I
2642
L J
20 No.6 Bars(Grade 60)

11---.....11.._ Column <1>=610

38
3 layer. com olite
wrapplne

4 laytr5 co
wrappinj
5 Illyl!rS com
wrapping
site
t
1220
4 layers eomp ISte
wrappina
T
1220

j 1-1 Section

Bars(top &. bottom)

(a) (b) (e)
Downloaded from ascelibrary.org by INDIAN INSTITUTE OF TECHNOLOGY on 06/08/14. Copyright ASCE. For personal use only; all rights reserved.

FIG. 2. Retrofit and Repair Details: (a) Retrofit: C2-RT4; (b)

Retrofit: C3-RT5; (c) Repair: C4-RP4
14 No.6 straight bars(top)
432 plus 6 No.6 straight bars(bottom)
half region. The repaired column was then tested to failure.
T Retrofit and repair details are shown in Fig. 2(a)-(c). The
retrofit and repair were applied mainly to the lower portions
FIG. 1. Model Column Details
of the columns to achieve flexural enhancement. However, as
shown in Fig. 2(b), one model column was retrofitted for al-
1.2 m diameter prototype columns. As shown in Fig. 1, the most full column height with reduced jacket thickness for the
2.44 m tall and 0.61 m diameter model columns were rein- upper portion. The upper portion retrofit was mainly for con-
forced with 20 deformed #6 (nominal diameter = 6/8 in. = struction demonstration, rather than required by design. The
19.1 mm) bars, which constituted a longitudinal steel ratio of effects due to the upper portion retrofit for this column should
2% of the gross area of column section. The longitudinal steel be minimal since the moment in the column upper portion was
bars were lap-sliced for a length of 381 mm (15 in.) near the low.
bottom of the column. Round #2 (diameter = 1/4 in. = 6.4
mm) hoops spaced at 127 mm (5 in.) intervals were used as Retrofit Procedure
transverse reinforcement.
Due to the difficulties in obtaining Grade 40 (nominal Each layer of the jackets was prefabricated with unidirec-
strength = 40 ksi = 276 MPa) steel as specified in the prototype tional glass fiber sheets and 2-part epoxy. The layer thickness
bridge construction, Grade 60 (nominal strength = 60 ksi = is approximately 3.2 mm (118 in.). The fiber reinforcement is
414 MPa) steel was used. Actual yield strength of the Grade arranged in such way that 90% of the fibers are oriented in
60 steel was 462 MPa (67 ksi) based on tensile test results. the circumferential direction and 10% in the longitudinal di-
Compared to columns reinforced with Grade 40 steel with the rection. The elastic modulus and ultimate strength in the cir-
same longitudinal steel ratio, a column reinforced with Grade cumferential direction are 48,300 MPa (7,000 ksi) and 552
60 steel subjects to higher demand for shear and bond in the MPa (80 ksi), respectively, based on the tension tests on flat
lap-splice region. This implies that the test condition in this coupons of the composite jackets. Urethane-based high
program was even more severe than the condition of most strength adhesive was used to bond the prefabricated compos-
existing columns, where Grade 40 steel is common. In addi- ite shells to form a continuous jacket. Prior to applying the
tion, the test can also provide an important experimental back- first layer of the composite shells, the column surface was
ground for retrofitting older bridges where the designed Grade cleaned with a high pressure air gun. Fig. 3 shows the main
40 steel may actually have substantial overstrength. Ready retrofit procedures for one of the retrofitted test specimens.
mixed concrete was used throughout. Concrete 28-day cylin-
der strength was 44.8 MPa (6.5 ksi). Test Method
The main experimental parameter was the retrofitting
scheme. One model column was tested under the condition of The test setup was designed to subject the model columns
"as-built," while the other two columns were tested after be- to constant axial load and cyclic horizontal forces in a single
ing retrofitted with jacketing. The failed as-built column was curvature condition, as illustrated in Fig. 4. The horizontal
repaired by first replacing the loosened concrete with quick- force was applied by a :t 152 mm (:t6 in.) stroke actuator with
set cement and then installing a 4-layer jacket in the lower the capacity of 1,023 kN (230 kips) in compression and 832

(a) (b) (c)

FIG. 3. Installation of Prefabricated Composite Jackets: (a) Applying Adhesive to Column Surface; (b) Installation of Jacketa; (c)
Curing of Adhesive

1358/ JOURNAL OF STRUCTURAL ENGINEERING / OCTOBER 1997

J. Struct. Eng. 1997.123:1357-1364.

 I Axlal Load
splice region and then propagated upward. The first vertical
crack appeared in the lap-splice region at a displacement of
28 mm (1.1 in.), indicating the initiation of bond slip. After
one cycle at a peak displacement of 41 mm (1.6 in.), vertical
cracks were fully spread within the lap splice region, indicat-
ing the bond-slip failure of lap splices, as shown in Fig. 5(a).
The test was terminated when the lateral force capacity
dropped to 50% of the maximum capacity at a displacement
of 61 mm (2.4 in.).
The retrofitted columns performed in a consistently stable
fashion throughout testing. Flexural cracks were observed at
the column base where a gap of about 19 mm (0.75 in.) was
Downloaded from ascelibrary.org by INDIAN INSTITUTE OF TECHNOLOGY on 06/08/14. Copyright ASCE. For personal use only; all rights reserved.

intentionally provided to prevent the footing from overloading.

FIG. 4. Test Setup At the final stages of loading corresponding to a reference
displacement ductility factor of 8.0, fine horizontal cracks were
observed on the surface of the prefabricated composite jacket,
leN (187 kips) in tension. An axial load of 712 leN (160 kips)
as shown in Fig. 5(b). This was expected because of the
was applied to the column by posttensioning two 51 mm (2
smaller fiber content in the longitudinal direction. The repaired
in.) diameter high-strength rods with a hydraulic jack at the
column C4-RP4 also exhibited improved performance com-
top of the column. The ratio of the applied axial load to the
column axial load capacity was about 5%, which is among the
range of typical axial load ratios of columns in multicolumn
bent bridges in California. The footing was posttensioned to a
reaction beam using eight high strength rods. Instrumentation
included all the applied forces and horizontal displacement.
Strain gauges were applied to measure the strains of reinforc-
ing steel and composite jacket.
All specimens were subjected to cyclic lateral forces and
constant axial load. The lateral loading sequence was con-
trolled by displacement increment based on the reference duc-
tility index. The initial loading cycles were applied corre-
sponding to a peak displacement increment of 2.5 mm (0.1
in.) until the column developed the calculated capacity cor-
responding to the first yield of longitudinal steel, H y , for the
as-built reference column. Then the reference yield displace-
ment, ~y, was determined using the following equation:
Hif
~y =H ~1 (1)
Y

where ~l = average of the measured displacements

corresponding to the first yield capacity, H y , in the push and
pull directions; and Hif = calculated ideal flexural capacity of
(a)
the as-built reference column based on an extreme concrete
compressive strain of 0.005 and stress-strain models recom-
mended by Mander et al. (1988). After the column developed
the reference first yield capacity, three complete loading cycles
were performed at peak displacements corresponding to the
reference ductility factors, f.LA = 1, 1.5, 2, 3, 4, 6, and 8. The
reference ductility factor f.LA was defined as the ratio of dis-
placement, ~, to the reference yield displacement, ~y. Note
that actual loading procedures for different specimens were
somewhat different from the standard loading procedure due
to the difficulties in displacement control system used in the
tests.

EXPERIMENTAL RESULTS
The main experimental observations and results are sum-
marized in this paper. Detailed experimental results can be
found elsewhere (Xiao et al. 1995).

Observations
The as-built model column, CI-A, developed an unstable
response due to premature lap-splice failure between the lon-
gitudinal reinforcement and starter bars. At a displacement of
5 mm (0.2 in.), the first flexural crack was observed at the (b)
bottom end of the column. As the displacement increased to FIG. 5. Crack Patterns: (a) As-Built Column C1-Aj (b) Retro-
13 mm (0.5 in.), flexural cracks extended into the whole lap- flted Column C2-RT4

JOURNAL OF STRUCTURAL ENGINEERING / OCTOBER 1997/1359

J. Struct. Eng. 1997.123:1357-1364.

 paring to the as-built column. The maximum circumferential tion corresponding to a displacement of 13 mm (0.5 in.). After
strain of jackets recorded for all retrofitted columns was ap- the column suffered lap-splice failure, its load carrying capac-
proximately 0.001-0.0015 at the bottom of the jacket, similar ity stabilized at about 112 kN (25 kips), which was essentially
to the effective strain of 0.001 suggested by Priestley and Sei- the restoring force due to the applied axial load as the column
ble for composite retrofit (1991). No delamination or rupture started rocking about its base.
across the circumferential fibers of the jackets was observed
in any testing of the retrofitted or repaired columns. Retrofitted Model Columns
At large displacements, deterioration of the bond in the lap Compared to as-built column CI-A, the two retrofitted col-
splices was observed for all the retrofitted and repaired col- umns, C2-RT4 and C3-RT5, exhibited significantly improved
umns as evidenced by the pullout of the starter bars during the seismic performance and demonstrated the effectiveness of the
opening of the cracks at the base of the columns. This, along retrofit, as shown in Figs. 6(b) and (c), respectively. The
with the measured strains, indicated that performance of the retrofitted column with 4-layer jacketing developed stable hys-
Downloaded from ascelibrary.org by INDIAN INSTITUTE OF TECHNOLOGY on 06/08/14. Copyright ASCE. For personal use only; all rights reserved.

retrofitted or repaired columns at large displacements was also teretic response up to the first cycle corresponding to the ref-
influenced by bond deterioration of the lap-spliced longitudinal erence ductility factor of about 6.0. The retrofitted column
bars. with 5-layer wrapping developed even higher load carrying
capacity with a larger ductility factor. Although the load car-
Horizontal Force-Displacement Responses rying capacities of the retrofitted columns degraded gradually
at a ductility factor of 6.0, the columns could develop 70-
Plots of horizontal force versus displacement obtained from 80% of the ideal flexural strength, even at a displacement duc-
all four tests are shown in Fig. 6(a)-6(d). Calculated capacities tility factor as large as 8.0 or a drift ratio of about 5.5%. The
corresponding to the first yield of longitudinal steel, By, as gradual degradation of the capacity at large displacements was
ideal flexural capacity, Hij, based on the as-built reference col- considered due to the bond slip in the lap-spliced longitudinal
umn, are also shown in these diagrams by dashed lines. bars. Such bond slip mechanism can be considered acceptable
for seismic retrofit as long as the retrofitted column develops
As-Built Model Column the required load carrying capacity and ductility. A gradual
bond slip failure mechanism of the lap-spliced longitudinal
Fig. 6(a) shows that the as-built model column CI-A de- bars may be even beneficial since the total failure mode of a
veloped unstable hysteresis hoops and had rapid degradation column due to the rupture of longitudinal bars can be avoided.
in load carrying capacity due to the lap-splice failure. The
predicted ultimate flexural strength was not achieved. A max- Repaired Model Column
imum horizontal force of 231 kN (52 kips), slightly larger than The repaired column C4-RP5 developed a maximum load
the calculated first yield capacity, was noted in the pull direc- carrying capacity approximately equal to the first yield
350 - r - - - - - - - - - - - - r - - - - - - - - - - - , 350,....------------.-----------,
~
: ; 250
'AI Built' Column
Cl-A
=1 1.8 J
::: f
~q~ ~ ~ ~ ~ ~~. ~ 250
r.::l
Retrofitted Column
C2-RT4-
4--Layer 1I'rapplne

~ 150 ~ 150
o o
rz.. 50 rz.. 50

~
Z -50 ~ -50J-=:;:;~~
o
~ -150 ~-150
~
o ~
:t: -250 =-250
usc Structural Lab
_350+-~-r_....,.:..___,:_:_-r--r-.,....=U_=:S.;:.C..::S:..:tru;..=C:=:itU=::r=al:..,.L=.;a=.fb -350+-.......--,-..,...---,--,---+-----r--:::r:---r--:-r-:-......--:-i
-150 -100 -50 0 50 100 150 -150 -100 -50 0 50 1 0 150
HORIZONTAL DISPLACEMENT (rom) HORIZONTAL DISPLACEMENT (mm)
(a) (b)

350...--------......,.--------,
-g 350 ,....------------.--:--:--:---:---:---=--:--;:--,

250
Retrofitted Column
C3-R1'5
I)-Layer Wrapping
-
Z
~ 250
Cl-A
J 4
Repaired Column
4-!.ayer 1I'rappine I
~ ~
U
~
150
~ 150
0 0
rz.. 50
rz.. 50
.....l
~
~ -50 Eo< -50
Z
Z
0 0
~-150 t!-150
~ ~
0 0
:t: -250
==1= =::
:t: -250
-II.
7 USC Structural Lab _350+--,-J.'~=_6,.--_._____,J-2.,.-1
-+_...-U_SC:r:-S_tru..--c~t
...r:r:_al-L,-a-b:_:J
-35~150 -100 -50 0 50 1 0 1 0 -150 -100 -50 0 50 1 0 1 0
HORIZONTAL DISPLACEMENT (roro) HORIZONTAL DISPLACEMENT (rom)
~ ~
FIG. 6. Horizontal Force-Displacement Responses: (a) As-Built Column C1-A; (b) Retrofitted Column C2-RT4; (c) Retrofitted Column
CS-RT5; (d) Repaired Column C4-RP4

1360 I JOURNAL OF STRUCTURAL ENGINEERING I OCTOBER 1997

J. Struct. Eng. 1997.123:1357-1364.

 strength calculated on the basis of the virgin section condition, the confinement provided by the composite jacket contributed
as described in Fig. 6(d). The ductility factors based on the to the capacity increase. Another reason is that in the retro-
yield displacement defined for as-built column CI-A are fitted column, the bond deterioration is a gradual process and
shown in Fig. 6(d) as an index indicating the correlation of stress redistribution can be developed after the starter bars slip.
the displacements of the repaired column and the original as- From the deformation point of view, the plastification of the
built column. By comparing the hysteretic responses shown in hysteresis loops of the columns was mainly resulted from the
Figs. 6(a) and (d), it can be seen that the repaired column bond-slip of lap-splice and the nonlinear deformation of the
developed improved behavior than the original as-built col- confined concrete, rather than the yielding of longitudinal re-
umn. However, the hysteresis loops of the repaired column are inforcement. Thus there is a need to analyze the column per-
significantly pinched. During loading cycles corresponding to formance with taking into consideration the bond deteriora-
reference ductility factors of 6.0 and 8.0, the hydraulic relief tion.
valve in the axial loading system improperly functioned, yield-
Downloaded from ascelibrary.org by INDIAN INSTITUTE OF TECHNOLOGY on 06/08/14. Copyright ASCE. For personal use only; all rights reserved.

ing sudden changes in axial load, which are reflected in the ANALYTICAL APPROACH
sudden drops in the hysteresis loops beyond J.L =4.0, as shown
in Fig. 6(d). Analytical Model
In an early study, Xiao et al. (1987) developed a simple
Deterioration of Bond Stresses in Lap Splices analytical model for analyzing the lateral force and displace-
Strain gauges were mounted on the surfaces of the starter ment performance of short reinforced concrete columns con-
bars of the as-built and the retrofitted columns at their critical sidering both flexural deformation and bond slips in longitu-
sections. The stresses in the starter bars at the measured lo- dinal bars. A similar model is developed for analyzing the
cations can be calculated using the elastoplastic stress-strain performance of columns with lap splices, as shown in Fig. 8.
relationship and the measured strain data. Average bond stress, The analysis is essentially based on the moment curvature
T, along the surface of the starter bar can be calculated using analysis with taking into consideration the bond slips of the
the following equation: lap-spliced longitudinal bars.
A hinge is assumed at the bottom of the column, and the
_ 1,d'b upper portion of the column is considered as an elastic beam
T=- (2)
4Ls column. Within the hinge, the curvature is assumed uniform,
while a linearly distributed curvature is assumed for the upper
where 1, = stress at the critical section; d'b = diameter of a
portion, as shown in Fig. 8(b). Bond links are assumed for all
longitudinal bar; and Ls = length of the lap splice.
the lap splices above the hinge length, as shown in Fig. 8(c).
In Fig. 7, the average bond stresses in the extreme lap-
spliced longitudinal bars of the as-built column CI-A and ret-
rofitted specimen C2-RT4 are plotted against the applied peak Hinge Length
displacement ductility factors in the push loading direction. As Priestley and Seible (1991) have suggested the following
shown in Fig. 7, the average bond stress for the as-built col- fixed length of plastic hinges for analyzing inelastic flexural
umn deteriorated significantly after the column developed a behaviors of reinforced concrete columns with or without jack-
ductility factor of 1.8. Despite a slight increase in the rebar eting:
stress of CI-A at J.L = 1.8, the load carrying capacity degraded
compared to that at J.L = 1.0, as shown in Fig. 6(a). Such a for as-built columns: Lp=O.08h + O.022du,,{y (3a)
phenomenon probably resulted from the accumulation of the
tensile strains of the lap-spliced bars upon cycling. Bond stress for retrofitted columns: L =g + O.044du,,{y
p (3b)
in the starter bars of the retrofitted columns was much higher where L = plastic hinge length; h =height of the column; d
p lb
and the degradation was more gradual than in the as-built col- = diameter of a typical lap-spliced longitudinal bar; h = yield
umn, as shown in Fig. 7. In both cases, the bond stress cor- strength of the longitudinal reinforcement; and g = gap pro-
responding to the yielding of the steel at critical section, cal- vided at the bottom of the jacket.
culated as 5.79 MPa (0.84 ksi), could not be achieved. Despite In the proposed approach, the hinge length, L h , is assumed
this, the retrofitted column C2-RT4 was able to develop the variable corresponding to the steel stress of the extreme critical
first yield capacity, Hy , calculated based on the as-built refer- tensile bar, h. The expression takes the same format of (3a)
ence column, as shown in Fig. 6(b). One reason for this is that and (3b) but uses a variable steel stress, fr, instead of the yield
strength.h
-5.....--------------------,
a:l
y
!ii
- 4 Retrofitted Column C2-RT4

LongitUdinal Bar
4Jcy)

I
IZI
:,: Built' Column CI-A
IJ.I :l:-'M
,)<I -
l!

4
DISPLACEMENT DUCTILITY FACTOR (a) (b) (c)
FIG. 7. Degradation of Average Bond Stresses In Lap-Spllced FIG. 8. Analytical Model: (a) Columnj (b) Curvaturej (c) Bond
Rebars Links

JOURNAL OF STRUCTURAL ENGINEERING / OCTOBER 1997/1361

J. Struct. Eng. 1997.123:1357-1364.

 for as-built columns: Lh =0.08h + 0.022dJs (4a) ing results of pullout tests on deformed reinforcing bars with
different confinement conditions conducted by Giuriani et al.
for retrofitted columns: L h = g + 0.044d1lJ. (4b) (1991):
Confined Concrete
Mander et aI.'s stress-strain model (1988) is assumed a~ the
She = SbO (I + at) (10)

constitutive law of the confined concrete. An effective con- where SbO = 0.25 mm (0.01 in.); a = 75.0; and
finement stress provided by the prefabricated jacket is calcu-
lated using the following equation:
r=r.-kl!.
o r /;
(11)
2tJ
(5)
where ro = 2.0 and kr = 13.0. The maximum value for r should
Downloaded from ascelibrary.org by INDIAN INSTITUTE OF TECHNOLOGY on 06/08/14. Copyright ASCE. For personal use only; all rights reserved.

be limited to less than 1.0.

wherefi = effective confinement stress; V j and tj = the diameter
and the effective thickness of the jacket; Ej = elastic modulus Analytical Procedure
of the composite in the circumferential direction; and EJ = the
effective strain in the jacket. The effective strain of t j =0.001 The analytical procedure involves trial and error loops for
suggested by Priestley and Seible (1991) is used in the anal- searching the correct values of the compression zone depth
ysis. and the bond slip of each individual bar in the critical section,
corresponding to each drift displacement increment. The equi-
Bond Links librium condition between bond and tensile force is used as
the criteria in finding the bond slip of a lap-sliced bar. The
It is assumed that in the lap-slice region, the stresses in the equilibrium condition of the calculated internal axial force and
starter bars are transferred to the longitudinal column bars the applied axial load is used as the governing criterion to
through a series of bond links, which are distributed through- determine the calculation step of correct neutral axis position
out a length of L b given by the following equation: in the critical section.
(6)
Analytical Results
where Ls = length of the lap splices; d1b = diameter of a typical
lap-spliced longitudinal bar; and Is = stress of a longitudinal Analyses have been conducted to simulate the horizontal
bar. It should be noted that the length of the bond links defined force-displacement responses of the model columns tested in
by (6) is different for rebars at different location in a column this study. Figs. 9(a) and (b) show the comparisons of the
section. However, for simplicity, L b is calculated based on the analytical results and the test results of the horizontal force-
stress of the extreme tensile bar in the column section. The displacement envelopes for the as-built model column CI-A
bond stresses are actually distributed throughout the lap-splice 400.------------------..,
length, L s' with the peak bond stress occurring between L, and
zero bond stresses at its ends. In this study however, a constant ~
__ , __ Analysis (Perfect Bond)
bond stress is essentially used for simplicity. The second term
Ej300
in (6) is provided to minimize the influence due to ignoring ~
the nonlinear distribution of bond stresses within L,. In future o
ro.
studies, more detailed analysis and calibration of the bond
~200
transfer mechanism should be addressed.
The constitutive law for the bond links is expressed by bond ~
o
stress, Tb-slip Sb relationship, given by the following equation N
based on a form of Popovics' equation (1973) ~ 100
o
::t:
T!x:r(SbISbe) 'As Built' Column, Cl-A
T - --""-'-""-""""-- (7)
b - r - I + (SbIShe)' 0+--.----,----.,----.----.---,----1
o 50 1 0 1 0 2 0
HORIZONTAL DISPLACEMENT (mm)
where T;e = peak bond stress between the rebar and confined
concrete; Sbe = bond slip corresponding to T;e; and r = param- (a)
eter that defines the shape of the curve expressed by (7) for 400-,-----------------,
different transverse confinement. Peak bond stress, T;e, can be
expressed by the following equation: ~ _ _ ~a~•.!!' (Perfect Bond)

T;e = TfnJ + l.4fi (8)

---Ej 300
~
In (8), the first term, TJx" expresses the bond strength for o
ro.
steel bars in plain concrete; the second term expresses the bond ~200
strength increment as a consequence of confinement, where fi
is the effective confinement stress calculated using (5) and the ~
o
N
coefficient 1.4 is determined based on the shear friction co- ~100
efficient recommended by ACI318-89 (Building 1989). On the o
other hand, following ACI408 Committee report ("State" ::t:
Retrofitted Column, C2-RT4
1992) the first term of (8) is given by, 4 --L_a..:,.y_er_lf...,ra...;p....::p_ln-i'lr--_~--I
0..-_ _---.----.-
50 100 1 0 2 0
TfnJ = 20vJ:,ldlb :$ 5.52 MPa (9) HORIZONTAL DISPLACEMENT (mm)
Based on trial analyses the upper limit in (9) was ignored. (b)
Parameters Sbe and r in (7) are determined using the follow- FIG. 9. Analytical Results: (a) As-Built Column C1-A; (b) Ret-
ing empirical equations that were developed for this study us- rofitted Column C2-RT4

1362 I JOURNAL OF STRUCTURAL ENGiNEERING I OCTOBER 1997

J. Struct. Eng. 1997.123:1357-1364.

 and the retrofitted model column C2-RT4, respectively. Solid
lines in Fig. 9 describe the analytical results predicted using H
the proposed model, with taking into consideration the bond
slip of the lap-spliced longitudinal reinforcement. Dashed lines
in Fig. 9 show the results of the upper bound analysis in which #4 Hoops @ 305
bond slips of the lap slices are ignored.
As shown in Fig. 9, the analysis based on the proposed
model for the lap-sliced column agrees well with the test re- 26 #18 Longiludinal Bars
sults and shows reasonable conservatism for both the as-built
and the retrofitted conditions. It should be pointed out that the
analytical results, which take into consideration the bond-slip
Downloaded from ascelibrary.org by INDIAN INSTITUTE OF TECHNOLOGY on 06/08/14. Copyright ASCE. For personal use only; all rights reserved.

for the retrofitted column models, indicate a slightly more

rapid degradation after exceeding the peak horizontal load car-
rying capacity. This is probably due to the conservative nature
of the proposed simplified bond-slip law. However, if the ul- (a)
timate displacement is defined as the lateral displacement cor-
responding to the point where the horizontal load carrying ca- 2000-,------------------,
pacity drops to 85% of the peak capacity, the ductility
predicted by the analysis is close to the actual ductility ~ 9-Layer Retrofit
achieved by the test specimens, as shown in Fig. 9(b). fj1500 ---....lC-~La:2y~er. Retrofit
To demonstrate the bond slip mechanisms of the lap-spliced g§ _ 3-Layer Retrofit
bars, the calculated bond-slip distribution and strain distribu- r.... Exi.ting Column
tions of concrete and longitudinal steel in the hinge region are :i! 1000
illustrated in Figs. lO(a) and (b), respectively. The distributions Eo<
Z
shown in Fig. 10 are corresponded to a calculated column o
N
displacement of 51 mm (2.0 in.). As shown in Fig. lO(a), bond l:ia 500 COLUMN DETAILS:
slips take places in all the spliced bars in the tension zone, o Diameter=1829mm: Helght=13115mm
Long_ Bar: 25No.16, G40
::c: Trans. Bar: No.4 305mm <;40
resulting in the relaxation of tensile steel compared to the con-
r~~.;\4·fo~~';8?0"oWWPa
crete at the same position, as shown in Fig. 10(b). The con- 0+-,........,....,.-,-~~~=;:~~~-,..-.--,--.....- .........-1
100 2 0 300 4 0 5 a
sequences are the reduction of the stresses carried by the lap- HORIZONTAL DISPLACEMENT (mm)
spliced bars and a slight increase of the concrete compression
zone depth. Thus, flexural capacity of the column is reduced (b)
FIG. 11. Design Example: (a) Column; (b) Predicted Re-
sponses
Neutral Axi.
I
I
compared to the case where perfect bond exists between the
bars and concrete.
RETROFIT DESIGN RECOMMENDATIONS
AND EXAMPLES
The proposed analytical methods can be used as a tool for
seismic assessment and retrofit design of existing reinforced
concrete bridge columns with lap-spliced longitudinal rein-
forcement. The procedure involves trial designs of the layer
-4 Specimen C2-RT4
Calculated Bond-SUp. numbers of the prefabricated jackets and prediction of ultimate
at Hor. Dilpl.=51mm
-5+-.-...,--,....,..,---,--,.....,-,.....,.....,...,rl:,.:.,;:;::.;..:...;=+;:.....:=r=7=~
ductility. If the calculated ductility is less than what is required
-350 -250 -150 - 0 50 1 a 2 0 3 a for retrofit then the number of jacket layers should be in-
REBAR POSITION (mm) creased and the corresponding ductility factor recalculated un-
(a) til the predicted ductility factor exceeds the required value.
The procedure is demonstrated using a column example
0.005 ....-c-om-p-r-e.-.-:'""io-n-------,,-----------, shown in Fig. 11 (a). The analysis using the proposed model
for the existing column indicates that the column is not able
to develop its ideal flexural capacity or any ductility; thus ret-
0.000 +----~,__--+--------1 rofit is necessary. A 3-layer wrapping is first assumed for ret-
I " .... Rebar S a
-£7- - - - - - - - ..........- . . .- rofitting the example column. The ductility factor for the 3-
Z
layer retrofit is calculated as 5.0, corresponding to the 85%
~-0.005 Neutral Axi. " "
" Concrete Strain reduction of the load carrying capacity. By doubling the layers
til " to a 6-layer retrofit, a ductility factor of 6.5 is obtained and
the design requirement can be satisfied. The number of layers
-0.010 can be further increased to give further design safety. For ex-
ample, if nine layers of composite jacketing are used, a duc-
Strain Di.tribution. at tility factor as high as 8.5 can be expected. Horizontal force-
-0.0 15 +-,Te_n.....i,....0'T"n..,....,.-,-r-r-...-r...,-,+H,....or..,....,.D...,.i...;.p'-.,=,....51.....InIll-,--"f......-.-..--1 displacement responses for all the calculated cases are
-350 -250 -150 - 0 50 1 a 2 a 3 a summarized in Fig. ll(b).
REBAR POSITION (mm)
(b) CONCLUSIONS
FIG. 10. Calculated Bond Slip and Strain Distributions In Lap- A prefabricated glass fiber composite jacketing system for
Spliced Rebars: (a) Bond Slip; (b) Strains seismic retrofit of bridge columns with lap-spliced rebars has
JOURNAL OF STRUCTURAL ENGINEERING / OCTOBER 1997/1363

J. Struct. Eng. 1997.123:1357-1364.

 been experimentally and theoretically studied. Test results of Preliminary reconnaissance report on the Hyogo-Ken Nanbu earthquake
an as-built model column indicate that columns with poor lap- of January 17, 1995. (1995). Earthquake Engineering Research Insti-
tute, Berkeley, Calif.
splice details are vulnerable to premature brittle failure due to
Priestley, M. J. N. (1987). "Damage of the 1-511605 separator in the
the bond deterioration of the lap-spliced rebars. Such failure Whittier earthquake of October 1987." Earthquake Spectra, 4(2),
pattern can be delayed using the prefabricated composite jack- 389-405.
eting, as demonstrated by significantly improved hysteretic re- Priestley, M. J. N., and Seible, F. (1991). "Seismic assessment and retrofit
sponses and increased ductility of the retrofitted model spec- of bridges." Struct. Sys. Res. Pro)., Rep. No. SSRP·91/103, University
imens. of California at San Diego, Calif.
Priestley, M. 1. N., Seible, F., Xiao, Y., and Verma, R. (1994). "Steel
Bond slip in lap-spliced rebars can cause a gradual degra- jacket retrofit of squat RC bridge columns for enhanced shear strength.
dation of load carrying capacity of a retrofitted column after Part 1: Theoretical considerations and test design; and Part 2: Experi-
the required ductility is developed. Such mechanism may be mental results." ACI Struct. J., 91(4), 394-405; and 91(5), 537-551.
considered beneficial for retrofit, since a column can be ret- Saadatrnanesh, H., Ehsani, M. R., and Li, M. W. (1994). "Strength and
Downloaded from ascelibrary.org by INDIAN INSTITUTE OF TECHNOLOGY on 06/08/14. Copyright ASCE. For personal use only; all rights reserved.

rofitted with lesser jacket thickness and the total failure mode ductility of concrete columns externally reinforced with fiber composite
of the column due to the rupture of longitudinal bars can be straps." ACI Struct. J., 434.
Seible, F., and Priestley, M. J. N. (1993). "Retrofit of rectangular flexural
avoided. columns with composite fiber jackets." Proc., 2nd Annu. Seismic Res.
An analytical approach that takes into consideration the de- Workshop.
formation due to the bond slip in lap-spliced longitudinal re- Seible, F., Hegemier, G. A., and Innamorato, D. (1995). "Developments
inforcement has been developed. Comparison of analytical and in bridge column jacketing using advance composites." Proc., Nat.
test results indicates that the analysis provides a rational ex- Seismic Con! on Bridges and Hwy., Federal Highway Administration
planation and prediction to the behavior of columns with lap- and California Department of Transportation.
"State-of-the-art report on bond under cyclic loads." (1992). Report ACI
spliced longitudinal reinforcement. The proposed method can 408.2R-92, American Concrete institute, Detroit, Mich.
be used as a useful tool for seismic assessment and retrofit Xiao, Y., Priestley, M. J. N., and Seible, F. (1996). "Seismic assessment
design of columns with lap-spliced rebars as demonstrated by and retrofit of bridge column footings." ACI Struct. J., 93(1), 79-94.
a design example. Xiao, Y., Sakino, K., and Tomii, M. (1987). "Ultimate moment and me-
chanical behavior of reinforced concrete short circular columns con-
fined in steel tube." Trans. Japan Concrete Inst., 9, Tokyo, Japan,
ACKNOWLEDGMENTS 389-396.
The research described in this paper was funded by C. C. Myers In- Xiao, Y., Martin, G., Yin, Z., and Ma, R. (1995). Bridge column retrofit
vestment, Inc. The prefabricated composite jackets used in the project using snap-tite composite jacketing for improved seismic performance.
were manufactured by a joint venture of C. C. Myers, Inc. and NCF Structural Engineering Research Program, University of Southern Cal-
Industries, Inc. The writers wish to thank G. R. Martin, professor of USC ifornia, Los Angeles, Calif.
for his cooperation. Valuable contributions from graduate research assis-
tant Philip Z. Yin of USC in the experimental work are also gratefully
APPENDIX II. NOTATION
appreciated. The comments and conclusions described in this paper are The following symbols are used in this paper:
solely those of the writers and do not necessarily reflect the views of the
funding agency. dlb diameter of longitudinal bar;
Ii transverse confinement stress;
APPENDIX I. REFERENCES Is= steel stress of extreme critical tensile bar;
t= steel yield strength;
Building Code Requirements for Reinforced Concrete and Commentary f: concrete strength;
(ACI318-89/ACI418R-89). (1989). American Concrete Institute, De- g = gap provided at bottom of jacket;
troit, Mich. Hi[ = ideal flexural capacity for as-built reference column;
Chai, Y. H., Priestley, M. J. N., and Seible, F. (1991). "Seismic retrofit
Hy = first yield capacity for as-built reference column;
of circular bridge columns for enhancing flexural performance." ACI
Struct. J., 88(5), 572-584. Lb bond link length;
Giuriani, E., Plizzari, G., and Schumm, C. (1991). "Role of stirrups and Lh variable length of hinge;
residual tensile strength of cracked concrete on bond." J. Struct. Lp length of plastic hinge;
Engrg., ASCE, 117(1), 1-18. L, lap-splice length;
"Loma Prieta earthquake reconnaissance report." (1990). Earthquake r = shape parameter for bond stress-strain curve;
Spectra, Earthquake Engineering Research Institute, Berkeley, Calif. Sb = bond slip;
Mander, 1. B., Priestley, M. J. N., and Park, R. (1988). "Theoretical Sbe bond slip corresponding to T;e;
stress-strain model for confined concrete." J. Struct. Engrg., 114(8), .:1 1 average of measured displacements corresponding to Hy ;
1827-1849. .:1y reference yield displacement;
Matsuda, T., Sato, H., Fujiwara, H., and Higashira, N. (1990). "Effect of
carbon fiber reinforcement as a strengthening measure for reinforced J.La = reference displacement ductility factor;
concrete bridge piers." Proc., 1st U.S.-Japan Workshop on Seismic Tb bond stress;
Retrofit of Bridges. T~ = bond strength for steel bars in plain concrete;
Popovics, S. (1973). "A numerical approach to the complete stress-strain 7;c peak bond stress between rebar and confined concrete; and
curves for concrete." Cement and Concrete Res., 3(5), 583-599. 'f average bond stress along surface of starter bar.

1364 / JOURNAL OF STRUCTURAL ENGINEERING / OCTOBER 1997

J. Struct. Eng. 1997.123:1357-1364.