SHEAR STRENGTHENING OF CONTINUOUS RC BEAMS USING EXTERNALLY BONDED CFRP SHEETS

Transcription

1 SHEAR STRENGTHENING OF CONTINUOUS RC BEAMS USING EXTERNALLY BONDED CFRP SHEETS Ahmed Khalia, Gustavo Tumialan, Antonio Nanni, and Abdeldjelil Belarbi Synopsis: This paper presents the results o an experimental investigation on the response o continuous reinorced concrete (RC) beams with shear deiciencies, strengthened with externally bonded carbon iber reinorced polymer (CFRP) sheets. The experimental program consisted o nine ull-scale, two-span, continuous beams with rectangular cross section. The tested beams were grouped into three series. Three beams, one rom each series, were not strengthened and taken as reerence beams, whereas, six beams were strengthened using dierent schemes. The variables investigated in this study included the amount o steel shear reinorcement, amount o CFRP, wrapping schemes, and 90 0 /0 0 ply combination. The experimental results indicated that the contribution o externally bonded CFRP to the shear capacity o continuous RC beams is signiicant and is dependent on the tested variables. In addition, the test results were used to validate shear design algorithms. The proposed algorithms show good correlation with the test results and provided conservative estimates. 1

2 Ahmed Khalia, ACI member, is a visiting doctoral candidate in the Department o Civil Engineering, University o Missouri at Rolla, USA. He received his BSc and MSc rom Alexandria University, Egypt. His area o interest includes structural engineering, materials science, and concrete structural rehabilitation. He is a member o ASCE. Gustavo Tumialan, is a PhD candidate in the Department o Civil Engineering, University o Missouri at Rolla, USA, where he received his MSc. He pursued his undergraduate studies at the Pontiicia Universidad Catolica del Peru. His area o interest includes structural engineering, construction, concrete and masonry rehabilitation. Antonio Nanni, FACI, is the V&M Jones Proessor o Civil Engineering and Director o the University Transportation Center (UTC) at the University o Missouri-Rolla. Dr. Nanni is interested in construction materials, their structural perormance, and ield application. He is an active member in the technical committees o ACI, ASCE, ASTM and TMS. Abdeldjelil Belarbi, ACI member, is an Associate Proessor o Civil Engineering at the University o Missouri - Rolla. His area o interest is the constitutive modeling o reinorced and prestressed concrete as well as the use o advanced materials in new construction and strengthening o civil inrastructures. He is an active member o several ACI technical and educational committees. INTRODUCTION Shear collapse o RC members is catastrophic and occurs with no advance warning o distress. Existing RC beams with shear deiciencies ultimately need strengthening. Deiciencies may occur due to actors such as insuicient shear reinorcement, reduction in steel area due to corrosion, increased service load, and design/construction deects. In such situations, it has been shown that externally bonded FRP sheets increase the shear capacity signiicantly (1, 2). At present, most o the studies have speciically addressed simply supported beams. The objectives o this study were to investigate the shear behavior and mode o ailure o continuous RC beams strengthened with CFRP sheets and to validate a proposed shear design approach (3, 4). 2

3 EXPERIMENTAL PROGRAM Test Specimens and Materials The experimental program consisted o nine ull-scale, two-span, continuous RC beams with a rectangular cross section o 150 by 305 mm. The beams were grouped into three series labeled CW, CO, and CF (Fig 1). Each series had dierent longitudinal and transverse steel reinorcement ratios (a) Series CW No stirrups (b) Series CO (c) Series CF 2 D 32 2 D 16 Stirrups D = Stirrups D = D D (d) Specimen cross-section or Series CW and CO Dimensions in mm (e) Specimen cross-section or Series CF Strain gauge location Figure 1. Beam specimens detailing and dimensions Series CW consisted o two beams tested over a total span o 4,880 mm as shown in Fig. 1(a). The central support consisted o a 300-mm oset intended to 3

4 represent the intersection with a column. The concrete strength was 27.5 MPa. In this series, two 32-mm bars were used as longitudinal reinorcement or both top and bottom ace o the cross section to avor shear ailure. The beams were reinorced with 10-mm stirrups throughout. The stirrup spacing in the shear span o interest was selected to orce ailure in that span. For each beam, six strain gauges were attached to three stirrups to monitor the stirrup strain during loading. Series CO consisted o three beams, and had similar longitudinal reinorcement as that o series CW (Fig. 1(b)). No steel stirrups were provided in the tested shear span. The concrete strength or this series was 20.5 MPa. Four beams were tested in series CF (Fig. 1(c)). The concrete strength or this series was 50 MPa. The beams were reinorced with two 16-mm longitudinal steel bars on both top and bottom aces with no shear reinorcement. The mechanical properties o the materials used or manuacturing the test specimens are listed in Table 1. Fabrication o the specimens including surace preparation and CFRP installation is described elsewhere (4). Material Speciications Compressive strength (MPa) Table 1 Materials properties Yield point Ultimate tensile strength (MPa) Modulus o elasticity (GPa) (MPa) Series CW Concrete Series CO Series CF D = 32 mm Steel D = 16 mm D = 10 mm CFRP* t = mm *Fiber only Strengthening Schemes and Test Set-up One beam rom each series (CW1, CO1, and CF1) was not strengthened and was considered as a reerence beam, whereas six beams were strengthened with externally bonded CFRP sheets ollowing dierent schemes. The test setup as well as the strengthening schemes are shown in Fig. 2. In series CW, beam CW2 was strengthened with two CFRP plies having perpendicular iber directions (90 0 /0 0 ). The irst ply was attached in the orm o continuous U-wrap with iber direction oriented perpendicular to the longitudinal axis o the beam (90 0 ). The second ply was bonded to the two sides o the beam with iber direction parallel to the beam axis (0 0 ). This ply may provide additional resistance to the horizontal component o the crack opening. 4

5 Load cell Load Steel distribution beam 305 Load cell (a) Beams CW1, CO1, and CF1 (reerence beams) (b) Beams CW2 and CF3 (CFRP 90 / 0 ) (c) Beam CO2 (CFRP strips) (d) Beams CO3 and CF2 (CFRP 90, U-wrap) (e) Beam CF4 (CFRP 90, totally wrapped) LVDT Strain gauge Dimensions in mm Figure 2. Test set-up and strengthening schemes Two beams were strengthened in series CO. Beam CO2 was strengthened with one-ply CFRP strips in the orm o a U-wrap with 90 0 iber orientation. The strip width was 50 mm with center-to-center spacing o 125 mm. Beam CO3 was strengthened with one-ply continuous U-wrap. 5

6 In series CF, three beams were strengthened. Beam CF2 was strengthened with one-ply continuous U-wrap. Beam CF3 was strengthened with two CFRP plies having perpendicular iber directions (90 0 /0 0 ). Beam CF4 was totally wrapped with one-ply CFRP sheets. The sheets were attached to the our sides o the beam with an overlap on the top side. Even though total wrapping may not be possible in the ield, this case is representative o the upper threshold. Specimens were tested as continuous beams under concentrated loads applied to the mid-point o each span. Two load cells were used to monitor total applied load and reaction at the span o interest. This allowed the computation o the exact shear orce in the span o interest, independently o re-distribution phenomena. The load was applied progressively in ew cycles, usually one cycle beore cracking ollowed by three cycles to ultimate. The shear orce versus delection curves shown in this study are the envelopes o these load cycles. Five linear variable dierential transormers (LVDTs) were used or each test to monitor the vertical displacement at various locations as shown in Fig. 2. O these ive LVDTs, one was placed at each support to monitor support movement. Three strain gauges were attached directly to the FRP on the sides o each strengthened beam o series CW and CO, and six in beams o series CF as shown in Fig. 2. The strain gauges were oriented in the vertical direction and located at mid-height with distances o 175, 300, and 425 mm rom the ace o the central support. Test Results and Discussion In the ollowing discussion, reerence is always made to the weak shear span (span o interest). Series CW: A diagonal crack was observed in beam CW1 close to the middle o the shear span when the load was approximately 150 kn. As the load increased, more diagonal shear cracks ormed throughout, widened and propagated up to ailure, as shown in Fig. 3, at a load o 508 kn which corresponded to a shear orce o 175kN. In beam CW2, strengthened with CFRP (90 0 /0 0 ), the cracks on the beam sides and bottom were not visible because o the wrapping. Longitudinal cracks were observed on the beam topside at total applied load o 530 kn. The cracks initiated close to the position o the applied load and extended towards the middle support. At ailure, the concrete cover on the top side was extensively damaged (Fig. 4). The ailure occurred at a total load o 623 kn with a corresponding shear orce o 214 kn, a 22% increase in shear capacity as compared to CW1. The applied shear orce versus mid-span delection curves or beams CW1 and CW2 are shown in Fig. 5. The maximum CFRP strain measured at ailure in beam CW2 was about mm/mm, which corresponded to 17% o the ultimate strain. This indicates that CFRP can be stretched urther and thus increase the shear capacity i properly used. 6

7 Figure 3. Final ailure o beam CW1 Figure 4. Final ailure o beam CW Shear orce (kn) CW1 CW Mid-span delection (mm) Figure 5. Shear orce versus mid-span delection or beams o series CW 7

8 In beam CW1, only the third stirrup at distance o 300 mm rom ace o middle support yielded at ultimate. Comparisons between localized stirrup strains in beams CW1 and CW2 are shown in Fig. 6. The stirrup strains in beam CW2 were smaller than those in beam CW1 at the same load level due to the eect o CFRP Shear orce (KN) Stirrup strain (mm/m) stirrup 1 CW1 stirrup 2 CW1 stirrup 3 CW1 stirrup 1 CW2 stirrup 2 CW2 stirrup 3 CW2 Figure 6. Shear orce versus stirrup strains or beams o series CW Series CO: Series CO showed the largest increase in shear capacity compared to the other series studied in this research. In beam CO1, the reerence beam, the irst diagonal shear crack was observed at a total load o about 90 kn. As the load increased, more shear cracks appeared throughout the shear span. When the total load reached 145 kn, the corresponding shear orce peaked and remained constant thereater at about 48 kn, while the total load increased to its peak o 220 kn. The relatively large amounts o top and bottom longitudinal steel reinorcement (D=32 mm) kept the beam as one piece until total damage o concrete occurred. In beam CO2, strengthened with CFRP strips, the irst diagonal shear crack was observed at a load approximately 140 kn. Failure occurred at the total load o 265 kn due to debonding o the CFRP strips over the main shear crack. The applied shear orce at ultimate was 88 KN, an 83% increase in shear capacity over the reerence beam CO1. The maximum local vertical CFRP strain at beam ailure was mm/mm. Experimental results in terms o applied shear orce versus mid-span delection or beams o series CO are shown in Fig. 7. Beam CO3, which was strengthened with CFRP continuous U-wrap, ailed by CFRP debonding at a total load o 330 kn. The applied shear orce at ultimate was 113 kn, a 135% increase in shear capacity over the reerence beam CO1. In this beam, longitudinal cracks were observed on the top side o the beam beore ailure just as it was observed in beam CW2. 8

9 120 Shear orce (kn) Mid-span delection (mm) CO1 CO2 CO3 Figure 7. Shear orce versus mid-span delection or beams o series CO The applied shear orce versus vertical CFRP strain or beam CO3 is shown in Fig. 8. The maximum vertical strain in CFRP was about mm/mm. The strain gauges sg1, sg2, and sg3 were located at mid-height at distances o 175, 300, and 425 mm rom the ace o the middle support, respectively sg1 sg2 sg Vertical CFRPstrain (mm/m) Figure 8. Measured vertical CFRP strain or beam CO3 Series CF: The ailure mode o beam CF1 was shear compression. Failure occurred at total load o 268 kn with a corresponding applied shear orce o 93 kn. The use o continuous U-wrap oriented at 90 0 in beam CF2 caused a change in the inal ailure mode rom shear to lexural. The recorded load at ailure was 337 kn corresponding to a shear orce o 119 kn, showing an increase o 28% over the reerence beam CF1. Fig. 9 shows the experimental results o series CF in terms o shear orce versus mid-span delection. The maximum total applied load achieved by CF3 was 394 kn and the corresponding shear orce was 131 kn 9

10 with increase o 40 and 10% over CF1 and CF2, respectively. The inal ailure was controlled by lexure. In beam CF4, lexural ailure occurred at an applied load o 400 kn corresponding to a shear orce o 140 kn. By comparing to the reerence beam CF1, the capacity was increased by 50 percent. In addition, a large non-linear phase was recorded showing a notable increment in ductility. Shear orce (kn) CF1 CF2 CF3 CF Mid-span delection (mm) Figure 9. Shear orce versus mid-span delection or beam o series CF DESIGN APPROACH The design approach or computing the contribution o externally bonded CFRP reinorcement to the shear capacity o RC beams, in ACI Code ormat, was proposed in a previous research study (3). The model addressed the two possible ailure mechanisms o CFRP reinorcement (either CFRP racture or debonding). Furthermore, two limits on the contribution o CFRP shear reinorcement were proposed, the irst limit was set to control the shear crack width and loss o aggregate interlock and the second to preclude web crushing. Also, the concrete strength and CFRP wrapping schemes were incorporated as design parameters. At irst, the average bond strength and the eective bond length at CFRP debonding were based on work by Maeda et al.(5). Later, experimental and analytical results rom Miller (6) modiied the model by Maeda et al. and proposed new equations to predict the eective bond length and the ultimate load at CFRP debonding. Even though both models seemed to yield similar results in terms o ultimate load, the later has been adopted (4). Moreover, the design approach was extended to provide the shear design algorithms in Eurocode ormat in addition to ACI ormat. 10

11 In traditional shear design approaches, the shear strength o an RC section is the sum o the shear strengths o concrete and steel shear reinorcement. For beams strengthened with externally bonded FRP reinorcement, the shear capacity may be computed by the addition o a third term to account or the contribution o FRP. The proposed equation to compute CFRP contribution is similar to that or steel stirrups and consistent with ACI and Eurocode ormats. In the equation, an eective average CFRP stress, e, smaller than its nominal strength, u, is used to replace the yield stress o steel. This eective stress is computed by applying a reduction coeicient, R, to the nominal CFRP strength, which is dependent on the governing mode o ailure. Failure is governed by either racture o CFRP reinorcement (at an average eective stress level below nominal strength due to stress concentrations), or debonding o the CFRP reinorcement. In either case, an upper limit o the reduction coeicient is established (R max. =0.006/ε u where ε u is the ultimate tensile CFRP strain) (4) in order to control shear crack width and loss o aggregate interlock. This limit is such that the average eective strain in CFRP materials at ultimate con not be greater than mm/mm (without the strength reduction actor, φ, or the partial saety actor, γ ). In addition to the upper limit, two equations were proposed (one or each mode o ailure) or the computation o the reduction coeicient (and thus the contribution o FRP). The lowest o the three reduction coeicients would control design. Comparing with all available test results in the literature to date, the design model showed acceptable and conservative estimates.(4) Summary o Proposed Design Algorithms-ACI Format The proposed design algorithms or computing the shear capacity o RC beams strengthened in shear with externally bonded CFRP sheets are summarized below: The eective width o CFRP sheet, w e, may be computed irst according to the suggested bonded surace coniguration (3). w e = d L e I the sheet is in the orm o a U-wrap without end anchor (1) w e = d 2 L e I the sheet is bonded to only the sides o the beam (2) Where d is the eective depth o CFRP shear reinorcement (usually equal to the eective beam depth, d, or rectangular sections and d-t s or T sections, where t s is the slab thickness) and L e is the eective bonded length (L e =75mm). The shear capacity o the section may be ound by irst computing the reduction coeicient, R, on the ultimate strength o the CFRP. The reduction coeicient should be taken as the least o: 11

12 R = (ρ E ) (ρ E ) (3) 2 3 ( ' c ) w e 6 R = [ ( t E )] 10 ε d (4) u ε u R = (5) In the above equations, E is the elastic modulus o CFRP in GPa, ρ is the CFRP area raction (ρ = (2t /b w )(w /s )), t is the thickness o CFRP in mm, b w is the width o the beam cross section in mm, w is the width o CFRP strip (Fig. 10 shows the dimensions used to deine d, w, β, and s ), s is the spacing o CFRP strips (the maximum spacing, s max. was suggested equal to w + d/4), and c is the nominal concrete compressive strength in MPa. Note that, Equation 3 provides R or ailure mode controlled by CFRP racture and applicable or ρ E 0.7 GPa, whereas Equation 4 describes the ailure mode controlled by CFRP debonding and applicable or CFRP axial rigidity, t E, ranging rom 20 to 90 GPa. Equation 4 may be disregarded i the sheet is wrapped entirely around the beam or an eective end anchor is used. d β w s (a) w (b) s Figure 10. Dimensions used to deine the area o FRP (a) Vertical FRP strips. (b) Inclined strips The shear contribution o the CFRP, V, may then be ound rom the ollowing expressions: e = R u (6) A ( ) e sinβ + cosβ d = 2 c b wd V Vs s 3 (7) Where, A is the area o CFRP shear reinorcement (A = 2t w ), β is the angle between iber orientation and longitudinal axis o beam, and V s is the nominal shear strength provided by stirrups. Note that i continuous vertical sheets are used, w and s should be equal. The shear capacity o the beam may inally be computed rom: 12

13 φv n = 0.85(V c + V s ) V (8) Where φ is the strength reduction actor (suggested equal to 0.7 or CFRP contribution), V n is the nominal shear strength, V c is the nominal shear strength provided by concrete. Shear Capacity o a CFRP Strengthened Section-Eurocode Format The proposed design equation (Eq. 7) or computing the contribution o CFRP reinorcement may be rewritten in Eurocode ormat as Equation (9). ( γ ) ( 0.9d )( 1+ cotβ) A sinβ = (9) [ V ( V V )] e V d Rd2 Rd1 + s Where V d is the design shear contribution o CFRP to the shear capacity, γ is the partial saety actor or CFRP materials (suggested equal to 1.3), V Rd2 is the maximum design shear orce that can be carried without web ailure, V Rd1 is the design shear capacity o concrete, and V wd is the design contribution o steel shear reinorcement. wd Comparison between Test Result and Calculated Values The computed design contributions o CFRP in ACI code ormat, including the φ actor, to the shear strength o the tested beams CW2, CO2, and CO3 were 24.6, 16.5, and 41 kn, respectively. Compared to the experimental contributions, 39, 40, and 65 kn, the design algorithms give acceptable and conservative results. CONCLUSIONS In this study, the shear behavior and modes o ailure o two-span continuous RC beams strengthened with CFRP sheets were investigated. The test results indicated that the externally bonded reinorcement can be used to enhance the shear capacity o the beams in positive and negative moment regions. For the beams tested in the experimental program, increases in shear strength ranged rom 22 to 135%. Test results also indicated that the CFRP contribution is enhanced to a large degree or beams without stirrups than or beams with adequate steel shear reinorcement. The test results were used to validate design algorithms or computing the shear contribution o externally bonded CFRP sheets. The calculated values gave conservative results. 13

14 ACKNOWLEDGEMENTS This work was conducted with partial support rom the University Transportation Center on Advanced Materials and Non-Destructive Testing (NDT) Technologies based at the University o Missouri - Rolla. The Egyptian Cultural and Educational Bureau provided support to the irst author. REFERENCES (1) Arduini, M., Nanni, A., Di Tommaso, A., and Focacci, F., Shear Response o Continuous RC Beams Strengthened with Carbon FRP Sheets, Non- Metallic (FRP) Reinorcement or Concrete Structures, Proceeding o the Third Symposium, Vol. 1, Japan, Oct. 1997, pp (2) Triantaillou, T.C., Shear Strengthening o Reinorced Concrete Beams Using Epoxy-Bonded FRP Composites, ACI Structural Journal, Vol. 95 No. 2, March-April 1998, pp (3) Khalia, A., Gold, W., Nanni, A., and Abdel -Aziz M. I., Contribution o Externally Bonded FRP to the Shear Capacity o RC Flexural Members, Journal o Composites or Construction - ASCE, Vol. 2, No. 4, Nov. 1998, pp (4) Khalia, A., Shear Perormance o Reinorced Concrete Beams Strengthened with Composites, Ph.D Thesis, Structural Engineering Department, Alexandria University, Egypt, (5) Maeda, T., Asano, Y., Sato, Y., Ueda, T., and Kakuta, Y., A Study on Bond Mechanism o Carbon Fiber Sheet, Non-Metallic (FRP) Reinorcement or Concrete Structures, Proceeding o the Third Symposium, Vol. 1, Japan, Oct. 1997, pp (6) Miller, B., Bond between Carbon Fiber Reinorced Polymer Sheets and Concrete, MSc Thesis, Department o Civil Engineering, University o Missouri, Rolla, MO,

15 Keywords: Bond; Carbon iber; Continuous beam; Externally bonded reinorcement; Fiber reinorced polymer (FRP); Flexural strength; Reinorced Concrete; Shear strength; Strengthening 15