FRP Shear Reinforcement for Concrete Structures

Emile Shehata, Ryan Morphy, and Sami Rizkalla

Synopsis:

This paper summarizes an experimental program conducted at the University of Manitoba, Canada, to examine the structural performance of fiber reinforced polymer (FRP) stirrups as shear reinforcement for concrete structures. A total of ten large-scale reinforced concrete beams were tested to investigate the modes of failure and the contribution of the FRP stirrups in the beam mechanism. The ten beams included four beams reinforced with carbon FRP stirrups, four beams reinforced with glass FRP, one beam reinforced with steel stirrups and one control beam without shear reinforcement. The variables were the material type of the stirrups, the material type of the flexural reinforcement, and the stirrup spacing. Due to the unidirectional characteristics of FRP, significant reduction in the strength of the stirrup relative to the tensile strength parallel to the fibers is introduced by bending FRP bars into stirrup configuration and by the kinking action due to inclination of the diagonal shear crack with respect to the direction of the stirrups. A total of 40 specially designed panel specimens were tested to investigate the bend effect on the stirrup capacity, along with two control specimens reinforced with steel stirrups. The variables considered in the bend specimens are the material type of the stirrups, the bar diameter, the bend radius, the configuration of the stirrup anchorage, and the tail length beyond the bend portion. A total of 12 specially designed panel specimens were also tested to investigate the effect of the angle of cracks on the stirrup capacity. The two variables considered in this case are material type of the stirrups and the crack angle. Description of the experimental program, test results and design recommendations are presented.

Keywords: bend capacity, CFRP, cracks, GFRP, kink, shear, stirrups

Emile Shehata is an ACI member. He received his B.Sc. in 1989 from Ain Shams University, Cairo, Egypt, his M.Sc. in 1994 from Cairo University, Cairo, Egypt, and his Ph.D. in 1999 from the University of Manitoba, Canada. He is currently an NSERC industrial research fellow at Wardrop Engineering Inc. in Winnipeg, Manitoba. Ryan Morhy is an ACI member. He received his B.Sc. in 1996 and M.Sc. in 1999 from the University of Manitoba, Canada. He is currently a Design Engineer at Crosier Kilgour & Partners in Winnipeg, Manitoba, Canada. Sami Rizkalla is an ACI Fellow. He is currently the President of the Canadian Network of Centres of Excellence on Intelligent Sensing for Innovative Structures (ISIS Canada), and Professor in the Department of Civil Engineering, University of Manitoba, Canada. He is the chairman of ACI Committee 440, FRP Reinforcement for Concrete Structures.

INTRODUCTION

Stirrups used for shear reinforcement are normally located as an outer reinforcement with respect to the flexural reinforcement and therefore are more susceptible to severe environmental effects due to the minimum concrete cover provided. FRPs are corrosion-free materials and have recently been used as reinforcement to overcome the deterioration of concrete structures due to corrosion of steel reinforcement. The use of FRP as shear reinforcement for concrete structures has not yet been explored enough to establish a rational model to predict the shear behavior and strength of concrete members reinforced with FRP stirrups (1). This paper summarizes an experimental program conducted at the University of Manitoba, Canada, to examine the structural performance of FRP stirrups. The first phase of the experimental program evaluates the strength of a single FRP stirrup as influenced by the bend and the crack angle. The second phase of the experimental program investigates the modes of failure, shear strength and behavior of concrete beams reinforced with FRP stirrups.

EXPERIMENTAL PROGRAM

Phase I: Streneth of Sinele Stirrup

Forty-two specially designed specimens, using different types of carbon fiber reinforced polymer (CFRP), glass fiber reinforced polymer (GFRP) and steel

stirrups, were tested to study the bend effect on the strength of FRP stirrups. The configuration and dimensions of a typical specimen, are shown in Fig. 1. The debonding length of the stirrups within the blocks was achieved by using plastic tubes secured in place using duct tape. The variables considered in this

phase are the material type, the effective bar diameter, de (~4Ab / ff), the bend

radius, rb, the configuration of the stirrup anchorage (Type A or Type B), and the tail length, 1/, defined in Fig. 1. For Type A stirrups, the anchored end was debonded to simulate the performance of a standard hook. In Type B, the stirrups were debonded only at the continuous end. Detailed information about the tested bend specimens is given in Table 1. The test set-up consisted of a hydraulic jack used to apply the relative displacement between the two concrete blocks, and a load cell to measure the applied load. Characteristics of FRP and steel stirrups, as reported by the manufacturing companies and test results, are given in Table 2. The different configurations of the FRP stirrups used in the experimental program are shown in Fig. 2. Concrete used for all specimens had an average compressive strength of 50 MPa at 28 days.

Stirrup anchorage configuration

Fig. 1 Details of bend specimens Fig. 2 Stirrup Configurations

Ten specially designed specimens using different types of CFRP Leadine and GFRP were tested to study the kink effect on the strength of FRP stirrups. Two additional specimens reinforced with steel stirrups were tested as control specimens. Each specimen was reinforced with two stirrups at angle () with the central axis of the panel. The variables considered in this experimental phase were the material type, and the angle of inclination, () (2). The test set-up consisted of two hydraulic jacks connected to the same air pump to apply equal load on both sides of the specimen. The configuration and test set-up of a typical specimen are shown in Fig. 3.

Table 1. Details and test results of bend specimens material bend radius Tai11ength stirrup stress at fi/fiuv mode

type anchorage failure of rb r,/de ld* Id*/de Type ftv failure mm mm MPa

21 3 632 0.35 S-RB 42 6 639 0.35 S-R

20 3 63 9 737 0.41 S-RB 84 12 A 728 0.40 S-RB 120 18 793 0.44 S-R

Leadline --- B 715 0.40 R-B stirrups 21 3 1057 0.59 R-B

42 6 1235 0.69 R-B 63 9 A 1062 0.59 R-B

50 7 84 12 1053 0.58 R-B 120 18 962 0.53 R-B

--- B 981 0.55 R-B CFCC 15 4.2 45 9 A 916 0.51 R-B

7wire-5mm --- B 1455 0.82 R-B CFCC 15 3.4 45 9 A 983 0.53 R-B

U-5mm --- B 1187 0.64 R-B 20 3.2 45 6 A 798 0.43 R-B

22.5 3 789 0.42 R-B CFCC 45 6 1159 0.62 R-B 7wire - 30 4.8 67.5 9 A 1475 0.79 R-B 7.5mm 90 12 1846 0.98 R-B

150 20 1902 1.01 R-B --- B 1798 0.96 R-B

72 6 A 400 0.56 R-S GFRP 50 4.0 145 12 345 ff 0.48 R-B

--- B 347ft/' 0.49 R-B steel 20 3.0 40 6 A 593 0.99 Y-B

B 669 1.12 Y-B Failure modes: R-S: rupture along the straight portIOn between the concrete blocks, R-B: rupture at the bend, R-D : rupture at the end of the debonded length inside the concrete block, S:slippage of the bonded part of the stirrup, S-RB:slippage of the bonded part of the stirrup, followed by rupture at the bend, R-BD: rupture of some fibers at the bend zone and others at the end of the debonded length, Y -S: yield along the straight portion, and Y -B: yield at the bend

# average of six specimens ## average of ten specimens

Table 2. Properties ofFRP and steel bars used in the experimental program Material type CFRP CFRP CFCC GFRP steel

Leadline U - 5.0 7-wire 7-wire 7-wire bar strand Used for shear shear shear Shear flex. shear shear flex. diameter db (mm) 5xlOmm 5.0 5.0 7.5 15 12.0 6.35 15 areaAb (mm"') 38.48 15.20 10.10 30.40 113.6 113 31.67 140 effective de(mm) 7.0 4.40 3.59 6.22 12.0 12.0 6.35 13.4 strength (MPa) 1800 1842 1782 1875 1750 713 600ff 1590 ultimate strength"ff 1730 2170 1810 1910 2200 640 660fl 1860 E (GPa) 137 143 137 137 137 41 206 200

Imaximum &u (%) 1.26 1.52 1.32 1.40 1.60 1.56 2.0 4.0 guaranteed strength accordlllg to the manufacturer YIeld strength Based on tensIOn tests

I OO-mm extensometer (PI

Fig. 3 Details of kink specimens

Phase II: Beam Specimens

A total of ten reinforced concrete beams were tested. Four beams were reinforced with CFRP Leadline stirrups, four beams by GFRP C-BAR stirrups, one beam with steel stirrups and one beam without shear reinforcement was tested as a control specimen. The tested beams had a T cross-section with a total depth of 560 mm and a flange width of 600 mm, as shown in Fig. 4. Eight beams were reinforced for flexure with six 15-mm, 7 -wire steel strands with high yield strength. Two beams were reinforced for flexure using seven 15-mm 7-wire CFCC strands. All beams were designed to fail in shear while the flexural steel tendons are designed to remain in the elastic range to simulate the linear behavior of FRP. The beam without shear reinforcement was used as a control beam to account for all concrete contributions, including the dowel action of the steel strands used for flexural reinforcement which are normally weaker than conventional steel bars. Each beam consisted of a 5.0 meter simply supported span with 1.0-meter projections from each end to avoid bond-slip failure of the flexural reinforcement. Only one shear span was reinforced with FRP stirrups, while the other shear span was reinforced using two 6.35-mm­diameter closely spaced steel stirrups, as shown in Fig. 4. The variables considered were the material type of stirrups, stirrup spacing s, and the material type of flexural reinforcement. Detailed information about the tested beams specimens is given in Table 3. The beams were tested in four-point bending, with 2.0-m constant moment region. A closed-loop MTS cyclic loading testing machine was used to apply the load. Instrumentation of the beam included Linear Voltage Displacement Transducers (L VDTs) for deflection measurement, and displacement gauges (PI gauges) for crack width measurements. Electrical strain gauges were used to measure the strain in the stirrups (2).

Beam wIth ~ 7 CFCC '.' strands g;

1351

strands g; ---1f---+--+

1 135 1

Fig. 4 Details of beam specimens

Table 3. Details and test results of beam specimens Beam Stirrups &hend = Spacing Ie Shear Ultimate Max Average Mode !D" l!bendlEfv s cracking shear stirrup stirrup of

force V;esl strain at strain at failure# % Ver failure failure

MPa kN kN % % SN-O -------- 54 67.5 186.5 DT

SS-2 Steel --- d/2 54 70.0 272.5 0.95 0.44 SY

SC-2 CFRP d/2 54 75.0 277.5 1.05 0.77 SR

SC-3 Leadline 0.63 d/3 54 75.0 341.0 1.04 0.71 SR

SC-4 d/4 51 75.0 375.5 0.80 0.55 SR

SG-2 GFRP d/2 54 75.0 292.0 1.20 0.91 SR

SG-3 12mm 0.85 dl3 33 65.0 312.5 0.83 0.53 SC

SG-4 dl4 33 65.0 311.5 0.78 0.48 SC

CC-3 Leadline 0.63 d/3 50 67.5 305 0.90 0.65 SR

CG-3 GFRP 0.85 d/3 50 67.5 304.5 1.07 0.85 SR

" " " . .. " beam, that has an ID WIth S mItIal, IS remforced WIth steel for flexure, whIle "C stands for CFRP flexural reinforcement. "" d is the effective beam depth = 470 mm # DT: diagonal tension failure, SY: shear failure initiated by yielding of the steel stirrups, SR : shear failure initiated by rupture of the FRP stirrups, SC: shear compression failure /bend is the bend capacity of FRP stirrups, determined based on bend tests, jjuv is the guaranteed strength in the direction of the fibers, and Efv is the elastic modulus of the FRP stirrups

TEST RESULTS

Phase I: Strength of FRP Stirrup

Effect of Bend Radius--The strength of FRP stirrups may be as low as 35 percent of the strength parallel to the fibers, depending on the bend radius, rb, and tail length, ld *, as given in Table 1. In general, test results indicated that a decrease in the bend radius, rb, reduces the bend capacity. The strength reduction is attributed to the residual stress concentration at the bend zone. The radii of the bend, used in this study, range from 3.0 to 7.0 times the effective bar diameter, de. Fig. 5 indicates that the bend capacity varies greatly for the same type of reinforcing fiber. The JSCE proposed equation for the design stress of FRP stirrups based on the bend capacity, ./bend, is given in Fig. 5. Test results indicate that the JSCE equation can be used as a design tool for both the CFCC and GFRP stirrups; however, it overestimates the bend capacity of the Leadline stirrups even with large bend radius, rb. Based on this investigation, it is proposed to use a minimum bend radius-to-bar diameter ratio, njde, of 4.0 the CFCC and C-BAR stirrups and 7.0 for Leadline stirrups, in order to achieve a stirrup capacity at least 50 percent of the strength parallel to the fibers.

0.8

0.6

0.4

0.2

(- Leadline • eFee ... GFRP )

------------------------------1-----1r------------------------, , , ,e , , , •

• l • . ... ------------------------------.~----------

;- : - , , ! JSCE (1997): I'e.d = O.05~+O.30 : If" d. ,

O~~-+----r_--~--_+----~--~--_+--~

o 2 42 6 8 d

Fig. 5 Effect of bend radius, rb, on strength capacity of the bend,./bend

Effect of Stirrup Anchorage and Tail Length--Significant reduction in the CFCC stirrup capacity was observed in Type A anchored with a standard tail length of 6db, as compared to Type B anchored, as shown in Table 1. The strength reduction is attributed to possible slip at the bend leading to initiation of failure at a lower stress level. An increase in the tail length ld * resulted in an increase in the stirrup capacity, as given in Table1. For a taillength-to-effective diameter ratio, l/lde, higher than 12, the capacity of Type A anchored CFCC stirrups is as

high as that of Type B anchored stirrups. For Leadline stirrups, an increase in the tail length, 1/, resulted in a slight increase in the capacity. A tail length of 70 mm (10de) is sufficient to develop the bend capacity of the stirrups using rt/de of 7.0. The tail length of the GFRP stirrups tested in this study was either 6de or 12de. The bend capacity of such a minimum tail length of 6de was found to be equal to or higher than 48 percent of the guaranteed tensile strength parallel to the fibers which almost equals the average bend capacity of Type B stirrups. Therefore, it is recommended to use a tail length of six times the effective bar diameter or 70 mm, whichever is greater.

Effect of Crack Angle--All kink specimens failed either by rupture of FRP stirrups or yield of steel stirrups at the crack location. The relationship between the measured stress in the direction of the fibers of FRP stirrups at failure, !iv, and the stirrup angle, e, is shown in Fig. 6. There is no clear trend for an increase or decrease in the stress at failure with the variation of the angle e for both GFRP and CFRP stirrups. The average failure stress-to-ultimate strength parallel to the fibers ratio was found to be 0.81 with a standard deviation of 0.06. Fig. 6 shows that for kink specimens the stress in a FRP stirrup at failure could be as low as 65 percent of the guaranteed tensile strength parallel to the fibers. Meanwhile, it is observed for bend tests (Table 1) that the stress at failure could be as low as 35 percent of the guaranteed tensile strength parallel to the fibers. Therefore it was concluded that the bend effect on strength capacity of FRP stirrups is more critical than the kink effect, and consequently limits the contribution of FRP stirrups in the beams.

1.2'.--------------I;:::·=G=F=R=P=.=C=F=R:::::;,pi

fjv 1

flllv 0.8

0.6

0.4

0.2

______ ~--------------------__ ----------;;gIg~~j:L~Q~§l4~d§y.

----- --------------~----~---~--------q~~g~-----­• ------ - -------.r-----~------~---------q~~qg~~-£~~~§!q~gEY·

This zone statistically contains 90 % of the test data

~~o( "

"lv' , \

. " I'

o 0 10 20 30 40 e 50 60 70 80 90

Fig. 6 Effect of stirrup angle, e, on capacity ofFRP stirrups

Phase II: Beam Tests

All the tested beams failed in shear before yielding of the flexural steel strands or rupture of CFRP strands. No slip of the flexural reinforcement was observed during any of the beam tests. Shear failure of beams reinforced with FRP

stirrups initiated by either rupture of the FRP stirrups at the bend (shear-tension failure), as shown in Fig. 7, or crushing of the concrete in the shear span (shear­compressIOn failure). A summary of the beam test results is presented in Table 3.

Fig. 7 Beam specimen at failure

Contribution of FRP Stirrups--The shear capacity of a concrete beam without shear reinforcement, Vcr. is determined as the applied load which causes the initiation of the first shear crack. The contribution of the FRP stirrups to the shear carrying capacity of concrete beams was evaluated based on the difference between the measured shear strength, Vtes(, and the measured shear at the initiation of the first crack, Vcr. The term "Vtest- Vcr" also includes the stirrup effectiveness in enhancing the concrete contribution to the shear carrying capacity. Based on the traditional 45-deg truss model, and as observed in the beams tested in this experimental program, the effective stirrup capacity at failure,jfve, can be determined as follows:

f - (Vlesl - Vcr)s

fve- A d fv

(1)

where A/v is the area of the FRP stirrups, s is the stirrup spacing, and d is the effective depth of the beam. Fig. 8 shows the effective stress in FRP stirrups at failure for the different spacings, s, used in this study. Test results indicate that the effective capacity of FRP stirrups in beam action might be as low as 50 percent of the strength parallel to the fibers, provided that shear failure occurs due to rupture of FRP stirrups. For closely spaced stirrups, there is a higher chance for the diagonal cracks to intersect the bend zone of the stirrups, leading to a lower contribution of the FRP stirrups, as evident in Fig. 8. For beams reinforced with CFRP strands for flexure, the effective stirrup stress is less than

for the corresponding beam reinforced with steel strands. This may be attributed to the observed low values of the concrete contribution, Vel, in comparison to the concrete contribution, Ve, in equation (1).

Iv9·75

•

I • CFRP .. GFRP I

• .. /.i\ . : \

0.5 - - - - - - - - - -- -- -\~~~~ ... - - - - - - -- --- - - - - - -- - - - - - - - - --

0.25

-'-' Beams reinforced for flexure with CFRP strands

f = (v,,,,, - v'r )s foe Aft d

Beams failed due to rupture of FRP stirrups

O+----+----+----+----~--~--~----4---~

0.2 0.3 0.4 0.5 0.6 s

d Fig. 8 Effect of stirrup spacing on effective capacity ofFRP stirrups

Shear Cracking: At the early stage of loading, flexural cracks were observed in the region of pure bending at the same load level for all of the tested beams. With a further increase in load, additional flexural cracks formed in the shear spans between the applied load and the support. The shear cracking load was monitored by three techniques in addition to the visual observation of cracks. The crack width for three beams reinforced with CFRP, GFRP, and steel stirrups using stirrup spacing of dl2 are shown in Fig. 9. It can be seen that large crack widths were observed for the beam with CFRP stirrups, even though the stiffness index Ejvpjv is higher for this beam than for the one reinforced with GFRP stirrups. For the beam reinforced with GFRP stirrups, it is evident that the beam with shear reinforcement ratio of Pjv = 0.71 percent behaves similarly to the one with steel stirrups ratio of psv = 0.40 percent. This indicates that an increase in the shear reinforcement ratio Pjv of 80 percent minimizes the effect of the low modular ratio (EjlEs = 0.21) due to the good bond of GFRP stirrups. In general, it was observed that the beams reinforced with GFRP stirrups performed well despite the low elastic modulus of the GFRP material.

Effect ofFRP Longitudinal Reinforcement--The use ofCFRP strands as flexural reinforcement in two beams resulted in a reduction in the shear strength, compared to similar beams reinforced with steel strands. The relationship between the applied shear and the components of the shear resisting mechanism Vej and Vsj are presented in Fig. 10 for the beams reinforced with steel or CFRP strands for flexure and CFRP stirrups spaced at d/3 for shear. For the beam reinforced with CFRP strands, it is evident in Fig. 10 that the concrete contribution, Vej. at any load level up to failure was less than the concrete

------------------------ -

contribution, Vc, for the corresponding beams reinforced with steel strands. Similar behavior was observed for beams reinforced with GFRP stirrups (2). This behavior indicates that the use of FRP flexural reinforcement in concrete beams results in wider cracks, smaller depth of the compression zone and less dowel contribution, leading to reduction in the concrete contribution to the shear carrying mechanism, Vel-

400~--------------------------------------~ Beams reinforced with

z300 .::.:. -.... co Q)

{i;200 "0 .~ a. c. «100

steel-

yield of '" first --,

GFRP - Pv =0.71%

pjv(EJl9 = 0.15% ...........

steel strands for flexure

PIv = bA

Iv ;: shear reinforcement ratio "s

Ejv = elastic modulus of FRP stirrups Es = elastic modulus of steel = 200 OPa

0~----~----~----~~----~-----4----~ o 2 3

Crack width (mm) Fig. 9 Applied shear versus crack width - beams reinforced with stirrups

spaced at dl2

Z400.-------------------------------------~ ~ ve = shear resisting force provided by concrete // (/) Vej= shear resisting force provided by concrete in ,/// 'E beams reinforced with FRP for flexure ./ § 300 Vs = shear resisting force provided by steel stirrups ,// c. Vsj= shear resisting force provided by FRP stirrups / E / 8 /' OJ

:§200 (/)

'iii Q) .... .... CO

//

Beam reinforced // with steel strands

/'/'. VsJ Vsj /

~100 CJ)

/////_---I---i---~ ~

V Vej e CFRP strands /~ ........ -- ~:~'in reinforced with

O~---+----~--~~~~--~--~----+---~ o 100 200 300 400

Applied shear (kN)

Fig. 10 Effect of flexural reinforcement on shear resisting components

SUMMARY AND CONCLUSIONS

Fifty-two specially designed panel specimens and ten reinforced concrete beams reinforced with FRP stirrups were tested. The effects of the bend radius, the crack angle, the stirrup anchorage, the stirrup spacing and the material type of flexural reinforcement were investigated. Based on the results of the experimental program, the following conclusions can be drawn: (1) The bend effect on the strength capacity ofFRP stirrups is more critical than

the kink effect, and therefore limits the contribution of FRP stirrups in the beam action.

(2) The following limitations are proposed for detailing of FRP stirrups to achieve a capacity of at least 50 percent of the guaranteed strength parallel to the fibers: a- The bend radius, rb, should not be less than four times the effective bar

diameter or 50 mm, whichever is greater. b- The tail length, ld *, should not be less than six times the effective bar

diameter or 70 mm, whichever is greater. (3) The effective capacity ofFRP stirrups in beam action might be as low as 50

percent of the guaranteed strength parallel to the fibers, provided that the failure occurs due to rupture of FRP stirrups.

(4) Beams reinforced with CFRP strands for flexure showed less concrete contribution, Vc, than beams reinforced with steel strands. This is attributed to the wide cracks, small depth of the compression zone and poor dowel action associated with the use ofFRP as longitudinal reinforcement.

(5) Shear deformations are not affected only by the elastic modulus of the stirrup material but also by other factors such as the bond characteristics of the stirrups. The beams with GFRP stirrups showed better performance than those with CFRP stirrups.

(6) The relatively inexpensive GFRP stirrups could be a good alternative for shear reinforcement in concrete structures.

REFERENCES

l. Japanese Society of Civil Engineers, JSCE, 1997; "Recommendation for Design and Construction of Concrete Structures using Continuous Fiber Reinforcing Materials," Concrete Engineering Series 23, Edited by A. Machida, 325p.

2. Shehata, E, 1999; "FRP for Shear Reinforcement in Concrete Structures," Ph.D. dissertation submitted to the department of Civil Engineering at the University of Manitoba, 412p.