Seismic Enhancement of Circular RC Bridge Piers With Fiber Composites
Seismic Enhancement of Circular RC Bridge Piers With Fiber Composites
Seismic Enhancement of Circular RC Bridge Piers With Fiber Composites
l I I= ~, I StructuralEngineering I
Vol+6, No. 4 / December 2002
ppl 485"--493 J
Abstract
The quasi-static test was conducted in a displacement-controlled way to research the seismic behavior of reinforced con-
crete columns strengthened with glass-fiber sheets. Six concrete bridge pier specimens were constructed using the starter
bars that were lapped with the main longitudinal reinforcing steels, in a 1/2.83 dimensional scale factor. The unidirectional
glass-fiber sheets were impregnated with epoxy mortar and wrapped around the plastic hinge region of the specimens. All
specimens were tested under inelastic reversal loading while simultaneously subjected to a constant axial loadP ~. lfchAg .
Test results show that the seismic performance of retrofitted concrete columns significantly improved as a result of the con-
fining action of the glass-fiber composite straps. The straps are highly effective in confining the core concrete and preventing
the fongitndinal reinforcing steels from buckling under cyclic loading.
Keywords: tap-splice, gtass fiber sheets, seismic retrofit, reinjbrced concrete columns, quasi-static test
1. Introduction 1991, Aboutaha et at., 1999 etc.) have reported the effects
of confinement on the compressive strength and ductility of
The recent earthquakes have caused extensive damage to reinforced concrete bridge piers. They investigated the
highway bridge structures. Failure of these structures effect of lap splice lengths of the longitudinal reinforce-
exposed a number of structural deficiencies in many ment. Riyad et at. (1999) investigated the seismic repair of
bridges constructed before the new seismic design codes lap splice failures in damaged concrete coltunns. A total of
were in place. In particular, concrete columns with inade- six specimens in 18• inches cross-section were fabri-
quate lateral reinforcement contributed to the catastrophic cated and tested under axial load and cyclic lateral displace-
collapse of many bridges, and the poor detailing of the ment to investigate the effect of two confinement steel
starter bars in these columns compounded the problem of types and the repair method. Results from this cyclic lateral
seismic deficiency. The moments and lateral forces induced load test showed that the retrofitted columns reached their
by seismic loads result in large shear forces in bridge col- design strength with the enhancement of the shear strength
mnns, which are resisted mainly through lateral ties or spi- and ductility of the columns. Nader et al. (1992) indicated
rals around the longitudinal reinforcement. The lateral that compression lap splices can be designed to sustain a
reinforcement, if properly detailed and anchored, can also minimum of a dozen cycles of high intensity loads into the
prevent sudden loss of bond and buckling of the longitudi- inelastic range, where the maximum bar strain reached at
nal bars. Therefore, in columns with inadequate lateral rein- least three times the yield strain. Omar A. et at. (1998)
forcement, it is imperative to provide additional external investigated the performance of older reinforced concrete
confinement to insure ductile behavior. bridge columns under seismic loading, particularly with
Previous research has indicated that closely spaced trans- regard to the assessment of the residual strengths present in
verse reinforcement in the potential plastic hinge zone of degraded hinge regions. Eight reduced scale specimens
bridge columns increases ultimate compression strength were tested to evaluate the effects of varying shear span to
and strain of the concrete core (Priestley et at., 1996). Also, depth ratio, longitudinal steel ratio, lap splice length, and
many researchers (Saadatmanesh et at. 1997, Priestley et al. retrofitting detail on column performance. Plastic hinge
The manuscript for this paper was submitted for review on July 24, 2002.
regions with poor confinement and lap splice experienced 2. Research Significance
rapid flexural strength degradation as a result of concrete
cracking and splice slippage, but those with no splices and Many bridge failures during the recent earthquakes were
poor confinement exhibited moderate ductility with even- caused by poor pertbrmance of concrete columns, primarily
tual degradation due to longitudinal bar buckling. Also, due to inadequate lateral reinforcement and insufficient lap
these tests have shown that strengthening bridge columns length of the starter bars. This research provides an effec-
with steel jackets or with fiberglass/epoxy jackets signifi- tive and economical alternative technique for seismic up-
cantly improved the flexural and shear strengths and grading of this type of substandard concrete columns in low
increased the ductility of the column. or moderate seismic region. Furthermore, the results of this
The objective of this research is to evaluate the seismic research will provide an insight into fiber composites mate-
performance of reinforced concrete bridge piers with lap rials for retrofitting and reinforcing of concrete bridge col-
splice in the plastic hinge region, to develop the enhance- umIIS.
ment scheme of their seismic capacity by retrofitting with
glass fiber sheets, and to develop appropriate limited ductil- 3. Experimental Program
ity design concept in low or moderate seismicity region.
For this research, two reduced concrete columns were 3.1 Test Specimens
wrapped with high-performance glass fiber composite Circnlar solid RC piers of the Hagal Bridge in Korea
straps and then tested under reversal inelastic displacement- were adopted as a prototype of this test. The bridge had
controlled load cycles. Glass tlbers were unidirectionally been designed in accordance w4th the provisions of current
ananged and impregnated with epoxy mortar. A layer of Korea bridge design specification. Fig. 1 shows detailed
epoxy was then brushed on the strap for inteflaminar bond dimension of all test columns. As shown in Table 1, six test
while the strap was being wrapped around the column. In specimens were prepared for the quasi-static test to investi-
this research, the effect of lap splice, retrofit scheme, and gate their seismic performance. Four test columns were
transverse confinement was investigated. Lateral confine- nonseismicalty designed and made without lap splicing, or
ment could be provided by Wansvcrse reinforcing steels and with lap splicing of longitudinal steels in the lower plastic
the retrofit material like as glass fiber jacket. To provide dif- hinge region. Three of them (N-SP05-R0,1,2) are nonseis-
ferent levels of lateral confmernents, test specimens were mic test specimens with lap splicing of 50% longitudinal
designed in accordance with the pre-1992 roadway bridge reinforcement steels, and another one specimen (N-SP00-
design code, the current seismic design code, and the lim- R0) was nonseismic test specimens without lap splicing.
ited ductility design concept from Eurocode 8. Since lap Other two specimens (L,S-SP00-R0) were designed in ac-
splice of bridge piers is sometimes practically unavoidable, cordance with the limited ductile design concept of Euro-
three test specimens were made with lap splice of 50% lon- code 8, and with the 1992 seismic design code of Korea
gitudinal reinforcement steels. All test specimens were con- bridge design specification, respectively.
strutted with the aspect ratio of 4.0 using three confinement Tee diameter of the test columns was 600 ram. These
ratios, two levels of lap splice, and two retrofitting glass columns were reinforced longitudinally with 45D10 bars
fiber sheets. [diameter=10 rnm], resulting in a longitudinal reinforce-
A-A
30 54O3O 30 540 30 SECTION ~30_ 540
~ . ~tO 305,10 3O
A - A A - A B--B A - A B - B A - A
.....
---3
&
.... i I !
i
___J
el
l .... , ............
(a) Nonseismic design Co) Limited ductile design (c) Seismic design (d) Retrofit
Fig. 1. Design Details of Test Columns
ment ratio of 1.18 percent. The measured yield strength for formance in a desired direction. These sheets were wrapped
these bm~ was 450 MPa. Transverse confinement was pro- in muitiple layers around the column to form external
vided by D6 deformed steel [diameter--6 mm] with mea- hoops for improved confinement.
sured yield strength 35 MPa. Confinement steels spaced at For Specimens N-SP05-R0,1,2, the longitudinal reinforc-
122, 50 and 30 m m in plastic hinge region for three groups ing steels of the column were extended into the footing
of specimen as shown in Fig. 1 and Table l, The target using starter bars that were lapped with the main longitudi-
compressive strength of concrete was ~,[ =29 MPa. nal reinforcing steels of the cohmm over a length of
0.007~d~, i.e., 270 mm. Columns N-SP05-R1,2 were st-
3.2 Retrofit S c h e m e rengthened using the glass fiber, SEIt 51, CAF GL-t000, of
Fiber reinforced composites have long been recognized which physical properties are shown in Table 2. The stress-
for their high-strength, good fatigue life, light weight, ease strain relationship of the composite strap was linear-elastic
of transportation and handling, and low maintenance costs. to failure. Thickness of glass fiber sheet was computed
Fiber composites, in general, are constructed of filaments using Eq. (1), which was proposed by Paulay and Priestley
such as glass, kevlar, carbon, etc., embedded in a resin (1996):
matrix. The fibers are the primary load-canying elements
0.1 ( ~ . - 0 . 0 0 4 ) D J))
within the composite. The matrix binds the fibers together t~ = f.~e., (l)
and transfers loads between them. The fibers have a strong
influence on mechanical properties of the composite, such wheretg is the thickness of glass fiber sbeet, e~, and~,,t are
as strength, elastic modulus, and deformation properties. the ultimate strain of confined concrete and glass fiber
Glass fibers are the most common of all reinforcing fibers sheet, respectively, D is the diameter of test specimen, and
for resin-matrix composites. Fiberglass fabrics are one of s andf~ are the yield stress of confined concrete and glass
the various forms of glass fibers that are commercially fiber sheets, respectively. The required thickness of glass
available and are very versatile for seismic retrofit applica- fiber sheets for this research were computed as 1.3 m m
tions. The fibers could all be oriented in one direction (~mi- (Paulay and Pfiesttey, 1996) that corresponded to .1 layer
directional), or different amounts of fibers could be oriented glass fiber sheet. Specimen N-SP00-R0 was used as the ref-
in different directions to achieve an optimmn structural per- erence specimen. Retrofitted column specimens were
strength steel rods against the concrete floor of the loading .=2 'I"A ^ i/ fi
frame. This load was applied to axial load ratio, P / f , A ~ -
0.1 on the columns. Next, the reversing lateral forces were
applied to the column by an MTS 1000 kN hydraulic actua-
tor mounted on the reaction wall. Each column was instru-
v v vv v 1/// Centinue6
t o fasilzre
hinge region of the columns. Most of the specimens failed 4.2 Load-Versus-Displacement Response
due to bond failure and buckling of the longitudinal rein- Figs 3(a) and (d) show that nonseismic test specimen
forcements except for the specimen N-SP05-R0. For most without lap splice developed more ductile hysteresis loops
of the specimens flexural cracks developed in the plastic than nonseismie test specimen with lap sphce of 50% longi-
hinge area of the cohmm during the quasi-static loading, tudinal reinforcing steels. As shown in Figs 3(e) and (f), it
and then the longitudinal reinforcing bar initially fractured was observed that the glass fiber sheets remarkably in-
at p=6~8. Table 3 shows the sequence of events that were creased the displacement ductility. In addition, Figs 3(b)
visually observed in the plastic hinge zone for each speci- and (c) showed that test specimens with more confinement
men. reinforcement steels could have greater displacement duc-
The first flexural cracking in the plastic hinge region was tility.
visually observed at the displacement ductility level of p - The measured maximum strengths and ultimate displace-
2.0~Y0 for all the columns. Concrete crack could not be ment for the retrofit and nonretrofit colmrms are summa-
visually observed in retrofired specimen because of glass rized in Table 4. The percent increase in strength and
fiber sheet, but initial cracks at cold joint between column displacement were calculated with reference to the mea-
and footing were occurred at/.t=3. sured value of the control specimen (N-SP00-R0). The ulti-
The specimen N-SP00-R0 with continuous longitudi- mate displacements of 104 m m and 98 mm, respectively,
nal reinforcement started to fail at a displacement ductil- were measured for specimen (N-SP05-RI,R2). These val-
ity level of ~t=8. At this point, the longitudinal ues were approximately 18.5 and 11.7 percent higher than
reinforcement started to buckle and the lateral load-car- that of reference specimen (N-SP00-R0). Whereas, the
rying capacity rapidly reduced until the failure o f the maximum lateral forces of all specimens are not much dif-
column. The failure sequence o f all test specimens is ferent, since the lateral confinement is effective after maxi-
summarized in Table 3. mum lateral force. Very similar trend was observed for
N-SP00-R0 ,g,
N-SP05-R0 l),
N-SP05-R1 # m
N-SP05-R2 dl, []
L-SP00-R0 9 9
,Ik
S-SP00 R0 . 9 9
@ : initial cracking, 41, " rebar buckling, 9 : cover concrete spalling l i ' rebar fracture, -k : longitudinal direction crack, 9 " crack
width expansion, dl~ : crack at cold joint between footing and columr~ # : longitudinal crack of glass fiber sheets
3e i
[N-SP00-R0
:
"
,0 ] L - S P 0 0 - R 0 . . . . ff] *~ iS.SP00_R0
~o 9 9:
: Push I
zo . 9 i
i "
~1o i -~ =88.i~mm
+A ~ g7.Tmm 131 fimm !
~ .LO
(a) -3O ]
4O 40 ..
-12~ - i ~ .so ~o 4O 2~ -12Q-1H ,,~ 6~ ~i0 ,~o ii 20 40 ~ 80 I100 i~o 2~ 4(I 60 80 LCQ lZ0 Idll
Displacernent(mm) Displ;tcement (ram) Oisptacemem(mm)
4O
: ':N-SP05-R0 ""i ~0 N - S P O S - R 1 " [ ~ : N-SP05-R2 P~h
~0 i 9 Puslt
20 ]
"1!-~..;~ ~
1o i
9~.sm
+ n ~ = at.4smm [ i PaO " + & ~ ~ t04.~mnt i m
-20 9L w -l - S-SP00-R0
4.3 Lateral Force-Displacement Envelope Curve - ~- 9L - s P 0 r I - R 0
-30
Fig. 4 shows the comparative lateral force-displacement -150 -120 -90 -60 -Zr 30 60 90 120 LZ,~
envelope curve for all test specimens in accordance with Dispk~ccmem(mm)
transverse confinement and retrofit. Fig. 4(a) shows that (a) Transverse confinelnent
30
more transverse confinement in the plastic hinge region of
RC bridge colt~nns can have greater ductility. Compared to 20
":"~.~" : , ':" ~ ~ . -'. ~ ." ~ ~ . : ;~ ...!ffk, ~ . .,r ~. : ~' "~ ~ ' 2." . 9 '" "':< 2; . ~ ~ ,. " " ,~ , ~ .."
~ " : - S ~ i i < i <: "'"" '~ '~ " " ". . . . . . . . . . '~ "= ' ~.-.....-..,, .,,.,:~ ~. ~. 9 .... ,. :-..~,,...- ~ ~ ' . . . , , ~. -
30
(a) N-SP05-R0
20
(a) N-SP05-R0
/ g
~o
lO
o ~ i 9 9
-~o
=20 -20
30 ....
-~0~ 20000 -10(300 O 10000 2~00 30800 =1008 0 1000 2000 3003 40r~3
S h a i n s (F 0 - 6 ) Strains(lO -~)
I
3O
(b) N-SP05 RI (b) N-SP05-R1
20
lO
/ i , , , , i i ~ i i
o
/
=10 -10
?0 .20
30 -30
-30000 -2"0000 =10000 0 10000 20000 30000 -t000 t 000 2000 :~300 4800
S t r a i n s ( 1 0 e) Strains(l 0 6)
(c) N-SP05-R[
2O
0 , . . . . . . . . , , , , , i , i i i
-10
20
rl~ 1
-30
-30
1101311 0 1000 2000 31300 4000
-200~0 10000 0 10000 3@000 30000
Strains(1 0 0) S t r a i n s ( 1 0 6)
Fig 6. Measured Strain of Longitudinal Steels Fig, 7. Measured Strain of Transverse Hoops
hoops, loss of confinement and rapid deterioration of the ing the load, and reduces the dilation of the core concrete.
plastic zone region. To compare the behavior of typical
hoops in the retrofit and nonretrofit specimens, the results 5. Conclusions
of strain measurements at the same location for the first
hoop above the footing of specimens N-SP05-R0, N-SP05- Following conclusions are drawn from the test results of
R1 and N-SP05-R2 are presented herein. These hoops were this research:
located approximately 60 m m above the top of the footing. 1. Reinforced concrete bridge piers, which was con-
Figs 7(a), (b) and (c) show the load versus strain in the hoop strueted with lap-spliced longitudinal reinforcement in
for above specimens. Normally, the hoops are in tension the potential plastic hinge zone designed before the
under the action of the applied lateral loads. The slight com- new seismic design provisions were in place, appear to
pression strains are due to local effects at the location of the fail at low ductility levels of/1=2.5. This is due to the
strain gages. In nonretrofit specimen N-SP05-R0, at the lat- debonding of lapped starter bars, resulting from the
eral load of approximately 147.4 KN, the swain in the hoop lack of~ansverse reinforcement and insufficient devel-
reached 16x10 -4. However, at the same lateral toad leve|, opment length of the longitudinal bars. The failure
the strain in the retrofit specimens N-SP05-R1 and N-SP05- modes are brittle because strength deterioration is very
R2 were only 3.6x10 4 and 8.4x10 -4, respectively, which rapid following debonding of longitudinal reinforce-
indicates the effectiveness of the glass fiber sheets in shar- ment.
2. The use of continuous reinforcement through the plas- Caltrans (1998). Bridge design specifieation manua~ California
tic hinge region improves moderately the lateral-dis- Department of Transportation, Division of Structures, Sacra-
placement hysteresis loops. The structural degradation mento, California, USA.
Eurocode 8 Part 2 (1996), Design provisionsfbr earthquake resis-
is likely to be delayed until ductility levels of/.t=4 are tant structures-bridg~ European Committee for Standardiza-
reached. The colunm failure is caused by the buckling tion.
o f longitudinal reinforcement within the hinge region Harold Saadatmanesh, Mohanmmad R. Ehsani, and Limin Jin
due to the lack o f lateral confinement. (1997). "Repair of earthquake-damaged RC columns with
3, Limited ductile and seismic design specimens show FRP wraps." A Journal of the American Concrete Instttut~
USA.
sufficient capacity in both strength and ductility in low
M.J. Nigel Pristley, Yuk Hun Chai, and Frieder Seible (1991).
seismicity region like Korean peninsula. "Seismic retrofit of circular bridge columns for enhanced flex-
4. The specimens retrofitted with the glass fiber sheets in ural performance."A Journal of the American Concrete Insti-
the potential plastic hinge region showed a significant tute, USA.
improvement in displacement ductility. The retrofit Mander, J.B., Priestley M./.N,, and Park R. (1988). "Theoretical
specimens developed very stable load displacement stress-strain model for confined conerete."JournaI of Struc-
tural Engineering, ASCE, Vol. 114, No. 8, Aug., pp. 1804-
hysteresis loops up to a displacement ductility level of 1826. ,-~,
~ - 4 , without evidence of significant stmcttwai deterio- Nader Panahshahi, Richard N. White, and Peter Gergely (1992).
ration associated with the bond failure of lapped started "Reinforced Concrete Compression Lap Splices under Inelas-
bars or longitudinal reinforcement buckling. tic Cyclic Loading", ACI Structural Journa~ Vol. 89, No. 2,
5. The retrofit schemes provided additional confinement Mar,-Apr., pp. t64-175.
NZS 3101 Part l&2 (1995). Design of concrete structures New
to existing core concrete, and were highly effective in
Zealand Standard.
preventing the columns from bond failure or longitudi- Omar A. Jaradat, David 1. McLean, and M, Lee Marsh (1998).
nal bar buckling. "Performance of existing bridge columns under cyclic load-
ing,part 1: experimental results and observed behavior,"ACI
Acknowledgments Structural JournaL Vol. 95, No. 6, Nov.~Dec., pp, 685-704.
Park, R. (1998). "Ductility evaluation from laboratory and analyti-
cal testing." Proc. of the Ninth Wortd Conference on Earth-
This Reseanzh was Supported by the Chung-Ang Univer- quake Engineering, Tokyo-Kyoto, Japan, Vol. , Batkema,
sity Research Grants in 2002. This support is gratefully Rotterdam, pp, 605-616.
acknowledged. Priestly, M,J.N., SeiNe, E, and Calvi, G.M, (1996).Seismic design
and retrofit ofbridg~ JOHN WILEY & SONS, Inc.
Riyad S. Aboutaha, Michael D. Engelhardt, James O, Jirsa, and
References Michael E. Kreger (1999), "Experimental investigation of
seismic repair of lap splice failures in damaged concrete col-
ATC-32 (1996). Improved seismic design criteria ./'or california urmls." A CI Str~wtural Journal, Vol. 96, No. 6, Nov,-Dec., pp.
bridges :provisional recommendation4 Applied Technology 988-996.
Council, USA. T, Paulay, M, J.N. Priestley (1991), Seismic design of reinforced
ATC-40 (1996). Seismic" evaluation and retrofit of concrete build- concrete and masonry buzTding~ JOHN WILEY & SONS,
ing, ATC Applied Technology Council, California Seismic Inc.
Safety Commission, Report SSC 96-01.