stay-in-place fiber reinforced polymer forms for precast ...€¦ · struction. precast technology...

polymere differentThe framesin four-pointng stay-in-er surfaceprovided

ms outsideg for male-his study.

Stay-in-Place Fiber Reinforced Polymer Formsfor Precast Modular Bridge Pier System

Zhenyu Zhu1; Amir Mirmiran2; and Mohsen Shahawy3

Abstract: This paper presents an innovative modular construction of bridge pier system with stay-in-place fiber reinforced(FRP) forms filled with concrete. Two 1/6 scale precast modular frames were prepared of a prototype bridge pier system. Thretypes of connections were considered: male-female, dowel reinforced with or without tube embedment, and posttensioned.were load tested in negative and positive bending. Subsequently, the cap beams were cut from the frames and tested to failurebending. Posttensioned joints exhibited the most robust and ductile behavior and proved to be the preferred method of joiniplace forms. Even with dowel bars, the male-female joints lacked the necessary structural integrity in the pier frames. Bettpreparation for FRP units and higher quality grouting may improve the response. Embedment of the columns into the footingadditional stiffness for the connection. The study indicated that internal reinforcement is not necessary for the stay-in-place forthe connection zone. The experiments also showed the importance of maintaining appropriate tolerances and match castinfemale and embedment connections. Overall, however, feasibility of the precast modular FRP system was demonstrated in t

DOI: 10.1061/(ASCE)1090-0268(2004)8:6(560)

CE Database subject headings: Bridge construction; Bridges, piers; Fiber reinforced polymers; Concrete, precast; Splicing.

y

rsd insionth ando

con-withctionin the

vi-te iscast

iffer-endures

ecast

em-or

CtypeBeamomesup-d to

eam-are

rentand

FFTastthusedif-duc-

etterthis

units,ads,

s of-y beand

s se-

tate

ring,

2.mustoneitor.siblepaper

Introduction

In recent years, stay-in-place fiber reinforced polymer(FRP)forms have been used for columns(Mirmiran and Shahaw1995), piles (Fam et al. 2003; Mirmiran and Shahawy 2003), andgirders (Burgueno et al. 1999a, b). Of these, piles and girdehave been installed in Virginia and California and field testeFlorida. Their advantages lie in the light weight and corroresistance of FRP. Previous studies have shown the strengstiffness of concrete filled FRP tubes(CFFT) to be comparable tthose of reinforced or prestressed concrete columns(Mirmiran etal. 1999). CFFT could be used in precast or cast-in-placestruction. Precast technology provides efficient constructionless on-site labor and better quality control. Precast construhas dominated the market for short to medium span bridgesUnited States(PCI 1999).

Precast construction with modular FRP forms provides aable alternative to existing precast concrete, where concreexposed. However, the primary concern, as with any premodular system, is in the connection detailing between the dent components(Martin and Korkosz 1982). Structural joints armore vulnerable to failure and deterioration because they ehigher forces(Priestley et al. 1996). The main objective of thi

1Doctoral Candidate, Dept. of Civil Engineering, North Carolina SUniv., Raleigh, NC 27695-7533.

2Professor and Chair, Dept. of Civil and Environmental EngineeFlorida International Univ. Miami, FL 33174.

3President, SDR Engineering Consultants, Tallahassee, FL 3231Note. Discussion open until May 1, 2005. Separate discussions

be submitted for individual papers. To extend the closing date bymonth, a written request must be filed with the ASCE Managing EdThe manuscript for this paper was submitted for review and pospublication on June 5, 2003; approved on September 29, 2003. Thisis part of theJournal of Composites for Construction, Vol. 8, No. 6,
December 1, 2004. ©ASCE, ISSN 1090-0268/2004/6-560–568/$18.00.
560 / JOURNAL OF COMPOSITES FOR CONSTRUCTION © ASCE / NOVE

study was to evaluate different types of connections for prCFFT members in bridge applications.

Four types of connections may be perceived for CFFT mbers in bridge applications:(1) splicing of CFFT piles, columnsgirders;(2) connections of CFFT columns with CFFT beams;(3)connection of CFFT columns or piles to reinforced concrete(RC)beams or pier caps; and(4) connection of CFFT columns with Rfootings. The difference between these connections is in theof loads that needs to be transferred across the connection.splices are primarily used to transfer flexural loads while sshear demand is expected, especially in the proximity of theports. Column or pile splices, on the other hand, are usetransfer axial forces along with flexural and shear forces. Bcolumn connections as well as column-footing connectionsused to transfer axial, shear, and flexural forces with differelative intensities depending on the spans of the beamslengths of the columns.

Three basic concepts were developed for connecting Cmembers to each other:(1) male-female connection of precunits with surface bonding;(2) dowel-reinforced connection wiFRP or steel bars; and(3) posttensioned steel connection. Becaapplicability and effectiveness of each system may be quiteferent, the choice of connection depends on the strength ortility demands. These connections may be combined for bperformance. Other types of connections, not investigated instudy, include threaded inserts of the tubes in the precastcoupler-sleeve connections for the tubes with or without threand bolted connections with FRP or steel bolts.

Two general concepts were considered for connectionCFFT and RC members:(1) dowel reinforced; and(2) embedment of CFFT into the RC member. These two concepts macombined for better performance in terms of stiffnessstrength.

A prefabricated modular CFFT-RC pier-frame system wa
lected for a detailed study of the connections described previ-
MBER/DECEMBER 2004

entceptstions.nec-tionfter

e cut

peci-

tworelf the

ofn thecap

FFTin-

h forpierin-RC

ted.on-ssiveat 46s usebothhighrein-void

ment

bri-eetsm,600

on ofd ant in-stinglayerr capmo-

were. The

ssivecom-ly. Itmore

ue towas

were

;eottomoting

vinyle sur-werecol-

ilicon,ouredimi-

gentsidesdges

es ofwiths, allthmix

lign-weredingre firstcts, to

ously. Two pier-frame specimens were built with five differtypes of connections as combinations of three different conof male-female, dowel reinforced, and posttensioned connecFig. 1 shows the pier-frame specimens, their five types of contions, as well as their assembly process. Additional informaon the specimens are provided in the following sections. Atesting the frame specimens, the two pier cap beams werfrom the frames and tested to failure.

Concrete Filled Fiber Reinforced Polymer Tube–Reinforced Concrete Bridge Pier Frames

Specimen Preparation

Two 1/6 scale prefabricated modular CFFT-RC pier frame smens were prepared of a prototype bridge-pier system(Fig. 1).Each frame consisted of one RC footing, two CFFT columns,CFFT pier-cap beam-column units(Unit A), and one CFFT piecap interior beam(Unit B). Frame 1 did not include mild stereinforcement, except for the dowels in the connections oCFFT columns to the RC footing. Moreover, the columnsFrame 1 were not embedded into the RC footing. Frame 2, oother hand, included mild steel reinforcement in the pierCFFT Units A and dowel bars in the connections of the Ccolumns to Units A and to the RC footing. The mild steel reforcement in Units A were designed to provide shear strengtthe cap beam as well as development length for dowel bars incolumns. Additionally, the CFFT columns in Frame 2 wereserted into a 152-mm-deep premolded hollow core in thefooting. The gap between the tube and the footing was grou

All units were cast using a single batch of ready mixed ccrete with a 28-day target strength of 27.6 MPa. The comprestrength of concrete cylinders was measured as 34.5 MPadays, when the frames were tested. Grade 414 MPa steel waas reinforcement and dowel bars. The three pier cap units inframes were posttensioned using 12.7 mm ASTM Grade B-7strength steel threaded rods. The RC footings had identicalforcement for both frames. Reinforcement was designed to aany failure at the base of either frame. The location and alignof dowel bars were maintained using wood templates.

The stay-in-place FRP forms for the pier cap units were facated by wrapping four layers of bidirectional carbon fiber shwith epoxy resin(CFRP) at 45° fiber orientation over Styrofoamolds. The carbon fibers had an ultimate tensile strength of 3MPa, a tensile modulus of 230 GPa, and an ultimate elongati1.4%. The laminate had a tensile strength of 600 MPa, anultimate elongation of 1.2%. The Styrofoam molds were kepside the FRP forms during shipping but removed before caconcrete. To develop negative moment continuity, a singleof a CFRP sheet was bonded to the top surface of the pieunits after posttensioning them together, in addition to thement capacity afforded by posttensioning. The columnsmade with 152-mm-diameter off-the-shelf glass GFRP tubestube wall was 2.54 mm thick with655° fiber orientation withrespect to the axis of the tube; had tensile and comprestrengths of 71 and 230 MPa, respectively; and tensile andpressive elastic moduli of 12,600 and 8,700 MPa, respectiveshould be noted that the compressive strength of the tube isthan three times as large as its tensile strength, primarily dthe fiber orientation in the tube. The above manufacturer data
verified by coupon tests(Shao 2003). The joints between Units A
JOURNAL OF COMPOSITES F

d

and B and the joints between the columns and Units Awrapped with one layer of the same CFRP sheet.

Each frame was assembled in three stages:(1) column-footing(2) column-Unit A; and(3) Unit A–Unit B. Assembling of thcolumns in Frame 1 consisted of several steps. First, the bsurface of each column and the top surface of the RC fowere wetted with EUCO WELD(Euclid, Cleveland, Ohio) bond-ing agent, which is an aqueous resin adhesive of a polyacetate base for permanently bonding damp or dry concretfaces, with a 3.45 MPa reported bond strength. The columnslowered in place to insert dowel bars in the ducts. Once theumns were placed and leveled, the sides were sealed with sand a cement grout mix with the same bonding agent was pinto the ducts from the top. The columns of Frame 2 were slarly placed into the RC footing after applying the bonding ato all surfaces of the core in the footing and the bottom andof the column. Silicon and rope caulk were used to seal the eof the column base before grouting.

Before placing the pier cap units, the top and side surfacthe columns and the inside surfaces of Units A were wettedthe bonding agent. To ensure continuity of concrete sectioncolumn joints were later drilled with a 1.6 mm drill bit at bosides along the column centerline and were filled with a groutthat included a latex polymer.

The seats in Units A were grinded and leveled for better ament of the pier cap units. The gaps between Units A and Bfilled with a grout mix made of cement, sand, and a bonagent. The threaded rods for posttensioned connections wecovered with tape and greased before being placed in the du

Fig. 1. Pier frame specimen and assembly process

avoid a potential bond with the grout in the joints.

OR CONSTRUCTION © ASCE / NOVEMBER/DECEMBER 2004 / 561

amed forpporth ten, PIr eacom-ns.wereB atmea-oten-r cap

both203-capoutyieldtwo

:t 76

fmid-sedd to

adingth anateThe

o en-

rfaceer, it

anyd atse ofbe-

ding,

s inlowernits

y behat

tions1 andega-ively,the

oticedistress

Test Setup and Instrumentation

Fig. 2 shows test setup and instrumentation for the pier-frspecimens. A steel frame with two 245 kN actuators were usethe tests. The specimens were tied to the floor with steel subeams and four 50-mm-diameter all-threaded steel bars, eacsioned to 90 kN. A total of 33 devices including inclinometersgauges, strain gauges, and potentiometers were used foframe. Each pier cap Unit A was instrumented with an inclineter along the column centerline to monitor joint rotatioDisplacement-type strain measuring devices, PI gauges,used over a 100 mm length at the joints between Units A andthe top and between the columns and Units A. Strains weresured using 30-mm-long electrical resistance foil gauges. Ptiometers were used to measure displacements of the piebeam.

Test Procedure

A hydraulic jack was used to posttension the pier cap units inframes to 142 kN, equivalent to 3.4 MPa prestress over themm-square section of Unit B, i.e., the smallest section in thebeam(see Fig. 2). This level of prestress corresponds to ab10% of the compressive strength of concrete and 40% of thestrength of the threaded rods. Load testing was carried out inseparate phases to simulate traffic loads in different lanes(1)negative bending of the cap beam with two load points aboumm from the free edges of the beam, and(2) positive bending othe cap beam with two load points about 152 mm from thespan. A roller 50 mm in diameter and 203 mm long was uunder each load point. The first phase of loading was limite

Fig. 2. Test setup and instr

about 224 kN to determine the stiffness of the frame in negative


-

h

bending up to service loads; whereas, the second phase of lowas carried out to failure. Displacement control was used wiinitial rate of 0.13 mm/min. After cracking of concrete, the rwas gradually increased to 1.3 mm/min by the time of failure.two actuators were hydraulically connected to each other tsure the same loading rate.

Test Results and Observed Behavior

Posttensioning resulted in a camber of about 0.25 mm and sustrains of about 0.002–0.003 in the pier cap beams. Howevdid not cause uplift or rotation in the pier cap Units A norjoint slippage between Units A and B. No uplift was observethe base of the RC footings in either frame under either phaloading. The inclinometers monitored symmetric deformationhavior in both frames, throughout negative and positive benwith a maximum difference of 10% between the rotations.

Table 1 compares the initial stiffnesses of the two frametranslation and rotation for both loading cases. Frame 2 hadinitial stiffnesses than Frame 1, despite its reinforcement in UA and the embedment of its columns in the footing. This maattributed to a gap in the joint of the north column to Unit A twas inadvertently not fully grouted.

Fig. 3 shows the total applied load versus midspan deflecfor the two frames in negative and positive bending. Frames2 were first loaded up to 228 and 220 kN, respectively, in ntive bending, and subsequently to 303 and 289 kN, respectin positive bending. A few cracking noises were heard fromcap beams under negative bending, but no cracking was non the painted surface. On the other hand, several signs of d

tation for pier frame specimens
umen
were observed under positive bending. Fig. 3 shows two graphs

MBER/DECEMBER 2004

meframeven-

twoin the

thensetet inFRP.d into

s ofat thet theomi-etterg inn oninset

andingue tonders. Aown

wn inide ofap at

twoThe

n hy-on asty ofasimilarments

twoed for

adingin the

e

for Frame 2, where the out-of-plane instability of the fracaused a load drop in the system. Subsequent loading of theshowed loss of stiffness caused by its internal cracking. The etual cause of failure for both frames was at the column base(seeinset photograph in Fig. 3).

Fig. 4 shows load strains at the midspan section of theframes under both loading cases. The labels are definedfigure. A maximum tensile strain of about 0.005 was noted inFRP tube at the bottom of pier cap Unit B in Frame 1. The iphotograph shows buckling of the top layer of the FRP sheFrame 1 at a maximum compressive strain of 0.0035 in theFurther examination revealed that damage was not extendethe CFRP mold.

Fig. 5 shows load strains at the joints of the pier cap beamthe two frames under both loading cases. In general, strainsjoints between Units A and B were much lower than those amidspan section of the pier cap beams. This indicates the dnance of shear effects at the pier cap joints, which may be bjustified using the strut and tie model. The apparent softeninthe response of Frame 1 may be attributed to joint separatiothe south side of the pier cap beam, as evidenced by thephotograph in Fig. 5.

Fig. 6 shows load strains at the joints between the columnUnit A on the north side of the two frames under both loadcases. Joint strains under negative loading were very low dnegligible moment transfer at the sections. Large moments upositive bending resulted in nonlinear behavior of the framesimilar response for the south column in Frame 1 is not sh

Table 1. Initial Stiffness of Bridge Pier Frames

Frame number Loading case

1 Phase I: Negative bending

Phase II: Positive bending

2 Phase I: Negative bending

Phase II: Positive bendingaTranslational stiffness is based on the vertical deflection at the cenbmrad=miliradians.

Fig. 3. Load-deflection response of


due to the symmetry in response. On the other hand, as shothe Fig. 6 inset, a large separation developed on the south sthe north column in Frame 2. This may be attributed to the gthis joint.

Concrete Filled Fiber Reinforced Polymer Tube PierCap Beams

Specimen Preparation, Instrumentation, and TestProcedure

The CFFT pier cap beams, Units A and B, were cut from thepier frames at the joints between the columns and Units A.cut surfaces were grinded smooth and were then set level idrostone. The specimens were tested in four-point flexurehinge and roller system, as shown in Fig. 7(a). Lateral supportwere provided at four locations to secure out-of-plane stabilithe specimens, as shown in Fig. 7(b). No lateral displacement wnoted throughout the load tests. The instrumentation was sto that described for the pier frames, except for the measurerelated to the columns, which were no longer applicable. Thepier cap beams were posttensioned to the same level describthe pier frames. A single 2,000 kN actuator with a 1 m strokecapacity was used for the loading of the pier cap beams. Lowas carried out in displacement control at the same rates aspier frame load tests.

Translationala stiffnessat midspan(kN/mm)

Rotational stiffnessat column centerlin

(kN/mrad)b

54 96

34 42

36 86

22 27

Unit B.

systems in negative and positive bending

ter of

frame


Fig. 4. Load strains at midspan of pier cap beams

Fig. 5. Load strains at joints of pier cap beams

Fig. 6. Load strains at pier cap beam-column joints

564 / JOURNAL OF COMPOSITES FOR CONSTRUCTION © ASCE / NOVEMBER/DECEMBER 2004

pier

mildesspeci-

l toth foralso

FRP

g ofd

er ofd theCFRPasits Aen in

weened aents

n. Onwithandside6was

pierwereing insive. Onam 2north

sec-nearoad-f Capcon-

nd

Test Results and Observed Behavior

Fig. 8 shows the load-deflection response curves for the twocap beams. The response consists of three stages:(1) initial stagebefore decompression and opening of the cracks;(2) preyielding;and(3) the postyielding plateau. Cap Beam 1, which had nosteel reinforcement, showed lower initial and preyielding stiffnwhen compared to Cap Beam 2. As discussed for the frame s

Fig. 7. Pier cap beams:(a) test setup and instrumentation; a(b) lateral supported cap beam

Fig. 8. Load-deflectio


mens, Units A in Cap Beam 2 were reinforced with mild steeprovide shear strength for cap beams and development lengdowel bars in pier columns. The strength of Cap Beam 1 waslower than that of Cap Beam 2 by about 25%.

Cap Beam 1 failed at 187 kN as a sudden rupture of the Cmold propagated to the top[Fig. 9(a) and inset in Fig. 8]. Furtherexamination of the beam primarily showed flexural crackinthe concrete core in Unit B[Fig. 9(b)]. Cap Beam 2 showevisible signs of separation between Units A and B[Figs. 9(c andd)]. A popping noise was heard at 150 kN, when the top layCFRP buckled at the midspan. A visual inspection revealedamage to be limited to the bonded CFRP sheet, and themold in Unit B was still intact. Testing of Cap Beam 2 wstopped due to excessive opening of the joint between Unand B on the north side. No damage was noted in Units A, evCap Beam 1 that had no internal reinforcement.

Fig. 10 shows the relative displacements at the joints betUnits A and B in the two pier cap beams. Cap Beam 1 showsymmetric response as evident from similar joint displacemon both sides before the rupture of the CFRP at the midspathe other hand, Cap Beam 2 showed asymmetric behaviorthe threaded rods yielding at the north joint between Units AB, rather than at the midspan. The joint opening on the northgrew to 12.7 mm(inset in Fig. 10), resembling a rotation of 0.0radians between two rigid bodies. However, no crackingfound in the CFRP mold at the midspan.

Fig. 11 shows the compressive strains on the top of thecap beams at the joints between Units A and B. The strainsmeasured using the PI gauges. Cap Beam 1 showed softenthe joint on the north side at about 165 kN, with compresstrains of up to 0.003 before eventual failure at the midspanthe other hand, compressive strains at both joints in Cap Beexceeded 0.0045 before failure of the strain gauges on theside.

Fig. 12 shows the top and bottom strains at the midspantion of the two pier cap beams. Cap Beam 1 shows a triliresponse with a distinct postyielding behavior, similar to the ldeflection curve. On the other hand, the strain response oBeam 2 remained almost linear at midspan, indicating stresscentration and eventual failure at the joints.

ponse of pier cap beams
n res

1;
Fig. 9. Failure modes in pier cap beams:(a) fiber reinforced polymer rupture in Cap Beam 1;(b) internal crack in concrete of Cap Beam(c) side view of joint separation in Cap Beam 2; and(d) top view of joint separation in Cap Beam 2
Fig. 10. Relative joint displacement between pier cap units

566 / JOURNAL OF COMPOSITES FOR CONSTRUCTION © ASCE / NOVEMBER/DECEMBER 2004

e ofcon-caphreeowelned., andre in

e be-FFTvel ofckedrfaceay

rgelyent.

rce-nnec-

tain-s forthe

n this

n andions

rily the

Conclusions

Two 1/6 scale prefabricated modular pier frames were madprototype bridge piers. Each frame consisted of a reinforcedcrete footing, two concrete-filled FRP tubes, two CFFT pierbeam-column units, and one CFFT pier cap interior beam. Tdifferent types of connections were studied: male-female, dreinforced with or without tube embedment, and posttensioThe frames were load tested in negative and positive bendingthe cap beams were cut from the frames and tested to failufour-point bending.

Posttensioned joints exhibited the most robust and ductilhavior and proved to be the preferred method of joining Cmembers. Further research is, however, needed on the leprestress. The male-female joints, even with dowel bars, lathe necessary structural integrity in the pier frames. Better supreparation for FRP units and higher quality of grouting mimprove the response. The stiffness of the joint depends laon the stiffness of the grout and the quality of its placem

Fig. 11. Compressive stra

Fig. 12. Load strains at m

Embedment of the CFFT into the footing provides additional ben-


efit for the connection. The study indicated that internal reinfoment is not necessary for the CFFT members outside the cotion area.

The experiments further revealed the importance of maining tolerances and match casting in male-female connectionCFFT joints and embedment for CFFT-RC joints. Overall,feasibility of the precast modular system was demonstrated istudy.

Acknowledgments

This study was supported by the National Science Foundatiothe Florida Department of Transportation. Findings and opinexpressed here are those of the authors and not necessaviews of the sponsoring agencies.

References

Burgueno, R., Seible, F., and Karbhari, V. M.(1999a). “Design of an

t the joints of pier cap beams

n sections of pier cap beams

ins a

idspa

advanced composite alternate for the King’s stormwater channel


No.

, L.theep.

s

e-

s-

n-

ul

-

t-

d

clic

bridge—Design and analysis.”Structural Systems Research Rep.99-06, Univ. of California at San Diego, La Jolla, Calif.

Burgueno, R., Seible, F., Karbhari, V. M., Davol, A., and Zhao(1999b). “Manufacturing and construction support document forKing’s stormwater channel bridge.”Structural Systems Research RNo. 99-05, Univ. of California at San Diego, La Jolla, Calif.

Fam, A. Z., Pando, M., Filz, G., and Rizkalla, S. H.(2003). “Precast pilefor Route 40 bridge in Virginia using concrete filled FRP tubes.”PCIJ., 48(3), 31–40.

Martin, L. D., and Korkosz, W. J.(1982). Connections for precast prstressed concrete buildings including earthquake resistance, PCI, Chi-cago.

Mirmiran, A., and Shahawy, M.(1995). “A novel FRP-concrete compo


ite construction for the infrastructure.”Proc., 13th Structures Cogress, ASCE, New York, 1663–1666.

Mirmiran, A., and Shahawy, M.(2003). “Composite pile: A successfdrive.” Concr. Int., 25(3), 89–94.

Mirmiran, A., Shahawy, M., and Samaan, M.(1999). “Strength and ductility of hybrid FRP-concrete beam-columns.”J. Struct. Eng.,125(10), 1085–1093.

Precast/Prestressed Concrete Institute(PCI) Industry Handbook Commitee.(1999). PCI design handbook, 5th Ed., PCI, Chicago, 1–21.

Priestley, M. J. N., Seible, F., and Calvi, M.(1996). Seismic design anretrofit of bridges, Wiley, New York.

Shao, Y.(2003). “Behavior of FRP-concrete beam-columns under cyloading.” PhD thesis, North Carolina State Univ., Raleigh, N.C.

MBER/DECEMBER 2004

stay-in-place fiber reinforced polymer forms for precast ...€¦ · struction. precast technology...

Documents