SHEAR STRENGTHENING USING CFRP SHEETS FOR PRESTRESSED CONCRETE BRIDGE GIRDERS IN MANITOBA, CANADA

Transcription

1 SHEAR STRENGTHENING USING CFRP SHEETS FOR PRESTRESSED CONCRETE BRIDGE GIRDERS IN MANITOBA, CANADA R. Hutchinson, A. Abdelrahman and S. Rizkalla Corresponding Author: R. Hutchinson ISIS Canada Network of Centres of Excellence, University of Manitoba, Room 227 Engineering Building Winnipeg MB, Canada, R3T 5V6 Telephone: (204) Fax: (204) umhutch Keywords: AASHTO, CFRP, concrete, girder, prestressed, shear, strengthening

2 The use of heavier trucks demanded shear strengthening of a twenty-seven year old bridge in Winnipeg, Manitoba, Canada. Shear strengthening of the prestressed concrete girders using Carbon Fibre Reinforced Polymer (CFRP) sheets provides a low-cost solution by increasing the speed of construction and reducing traffic interruption when compared with conventional methods. Due to a lack of information on the use of CFRP sheets for shear strengthening of I-shaped concrete AASHTO girders, an experimental program has been undertaken at the University of Manitoba. Four ten meter long scale model prestressed concrete girders have been tested to failure at both ends, using six configurations for three types of CFRP. The contribution of the CFRP sheets to the enhanced shear capacity of the girder has been examined with emphasis on the effect of this particular girder shape. INTRODUCTION The city of Winnipeg, Manitoba, Canada, is considering an upgrading of the Maryland bridge in response to the demand for using increasingly heavier truck loads. The twin five-span continuous prestressed concrete structures were designed in 1969 according to the American Association of State Highway and Transportaion Officials (AASHTO) Code. Rehabilitation analysis using the current AASHTO Code indicates that while the flexural strength of the I-shaped girders is adequate, the shear strength of the girders is not sufficient to withstand the increased truck load. Carbon Fibre Reinforced Polymer

3 (CFRP) sheets provide an excellent solution for shear strengthening since they are lightweight, corrosion-free, and have a high tensile strength. When compared with conventional methods, this technique provides a low-cost solution due to significant reduction of construction time without traffic interruption. Due to a lack of information on the use of CFRP sheets for shear strengthening of I-shaped concrete AASHTO girders, an experimental program has been undertaken at the University of Manitoba, Canada, to test scale models of the Maryland bridge girders strengthened with CFRP sheets. Four prestressed concrete beams are tested to failure at each end to determine the most efficient strengthening scheme. The contribution of the CFRP sheets to the enhanced shear capacity of the girder has been examined with emphasis on the effect of this particular girder shape. This paper summarizes preliminary test results and recommendations for the use of this strengthening technique on the I-shaped AASHTO girder. EXPERIMENTAL PROGRAM F our prestressed concrete girders were tested at the University of Manitoba, Canada. The ten meter long beams are 1:3.5 scale models of the I-shaped Maryland bridge girders. All of the beams had a depth of 415 mm with a top slab of 480 mm wide and 60 mm deep as shown in Figure 1 and Figure 2. The slabs were casted a minimum of seven days after the casting of each beam. The beams were pretensioned with three 13 mm straight steel 7-wire strands and one draped strand. Two non-prestressed 13 mm steel 7-wire strands were provided to increase the flexural capacity of the beams to avoid premature failure due to flexure. The stirrup shape and spacing were the same in all of the test beams. The beams were designed to carry the same shear stress at ultimate as the Maryland bridge girders. The stirrup shape, shown in Figure 2, is identical to those used in the bridge girders with the dimensions and bar diameter scaled down accordingly II _0_m_m_--I mm 70mm 415mm -"..-jl "'-- Figure 1. Test Beam Fabrication Figure 2. Test Beam Dimensions

4 - One of the beams was tested as a control beam while the remaining three beams were strengthened using three different types of CFRP for six different CFRP configuration schemes. The beams were tested at each end to determine the most efficient strengthening scheme. During testing of the unstrengthened first end of the control beam, premature failure occured due to the shape of the stirrups. An outward force was observed by spalling of the concrete cover of the control beam as shown in Figure 3. This force is the resultant of the tensile forces in the vertical and diagonal legs of the stirrups, causing the concrete to spall off and the stirrup to straighten. To control this outward force, the second end of the control beam was strengthened by a clamping scheme as shown in Figure 4. It should be noted that the stirrup shape is identical to those used in the bridge girders. Spalling of Concrete cover Straightening of Stirrups ReSUltant Outward Force Figure 3. Premature Stirrup Failure Figure 4. Stirrup Clamping Scheme The second test beam was strengthened using one layer of 250 mm wide vertical CFRP sheets with a 100 mm gap between each sheet to allow drainage of any moisture accumulation. The vertical CFRP sheets were applied on each side of the cross-section from the top of the beam immediately below the slab to the underside of the beam where they were overlapped for a minimum length of 100 mm. The other end of the beam was strengthened with a single layer sheet of 220 mm wide CFRP with the continuous carbon fibres in the horizontal direction on the top of the vertical sheets as shown in Figure mm Figure 5. Horizontal and Vertical Sheets Applied to Second Test Beam

5 The surface of the second beam was prepared prior to application of the CFRP sheets using a grinder, wire brush and high pressure air for cleaning the surface after grinding. Type A CFRP sheets were used to strenghten both ends of the second test beam. One end of the third test beam was strengthened by one layer of 250 mm wide CFRP sheets with the fibres oriented diagonally at 45 degrees. The sheets were applied on each side of the beam and overlapped on the underside of the beam. A 20 mm gap was provided between each diagonal sheet as shown in Figure 6. The other end of the third beam was strengthened using one layer of 250 mm wide vertical CFRP sheets similar to the second beam, but with a gap of 20 mm between sheets. Type B CFRP sheets were used for both ends of the third beam. The surface of the beam was prepared using a high pressure water-blasting technique, however, some grinding was required to round any sharp comers on the beam. Figure 6. Diagonal CFRP Sheets Applied to Third Test Beam The fourth test beam was strengthened using one layer of 250 mm wide diagonal CFRP sheets. As with the previous beams, the sheets were applied on each side and overlapped on the underside of the beam. A 100 mm gap was provided between each diagonal sheet similar to the second test beam. The other end of the beam was strengthened with a single layer sheet of 220 mm wide CFRP with the continuous carbon fibres in the horizontal direction on the top of the diagonal sheets as shown in Figure l mmyloo mm gap. [220mm Figure 7. Horizontal and Diagonal Sheets Applied to Fourth Test Beam

6 The fourth beam was strengthened using Type C CFRP sheets. Similar to the third test beam, the surface of the beam was prepared using a high pressure water blasting technique and grinding to round any sharp corners on the beam. The CFRP sheets were applied to the beams using well defined procedures recommended by each manufacturer and proprietary products supplied by each manufacturer. Following the surface preparation for each beam, an epoxy primer was applied to the surface of the beam to seal the concrete. After setting of the primer, any significant surface irregularities were filled using an epoxy putty on the second and third beams and an epoxy resin with additional filler material on the fourth beam. Additional grinding was required after setting of the putty to achieve a smooth surface on the second and third beams. For the second and third beams, the epoxy resin was applied to the surface of the beam followed by the CFRP sheets. A second layer of epoxy resin was applied as a top coat. In the case of multiple layers of sheets, the top coat of epoxy resin served as a base coat for the next layer of CFRP sheets. For the fourth beam, the CFRP sheets were impregnated with epoxy resin in a separate resin bath prior to application on the beam and subsequent layers were applied using the same technique. Material Properties The test beams were fabricated by a local precast plant in Winnipeg, Manitoba, Canada. The compressive strength of the concrete at the time of testing ranged from 44 to 55 MPa. Seven wire steel strand with a diameter of 13 mm and an ultimate tensile strength of 1860 MPa was used for both the prestressed and non-prestressed flexural reinforcement. The stirrups consisted of un deformed 5.5 mm diameter bars which were tested and found to have an average yield strain of , an average yield stress of 640 MPa and an average modulus of elasticity of 225 OPa. Three different types of CFRP sheets were used in this experimental program. The material properties of the three types of sheets are given in Table 1. Table 1. Material Properties of CFRP Sheets Property: Type A TypeB Design Thickness (mm): Fibre Areal Weight (g/m2): Tensile Strength (MPa): Tensile Strength (N/mm width): Tensile Modulus (OPa): Strain at Rupture: * All properties reported for dry fiber sheets + All properties reported for composite fiber and resin sheets -j- TypeC

7 Test Set-up Testing was conducted at the University of Manitoba. The simply supported beams were subjected to two equivalent non-symmetric point loads. The shear span of 1940 mm was kept constant for all of the tests while the overall span was varied in order to test both ends of each beam. The monotonic static load was applied using a 5000 kn testing machine under stroke control. The strain in the stirrups and the distribution of the strain in the CFRP sheets were monitored using electrical strain gauges. Demec stations were used to measure flexural strains at the top and bottom of the beam. U-shaped PI gauges were used to form rosettes for strain measurement in three directions on the web. Deflection was measured using Linear Variable Deflection Transducers (L VDT). The strain in the prestressed and nonprestressed strands was also monitored using electrical strain gauges. TEST RESULTS AND DISCUSSION The first end of the control beam was tested to failure without strengthening and the second end was tested using a clamping scheme to control the outward force in the stirrups. The remaining beams were tested at each end using different CFRP strengthening schemes as described in Table 2 below. Table 2 also provides a comparison of the ultimate shear failure load, V U' and the initial shear cracking load, V C' detennined by the load corresponding to the initiation of the first inclined cracks. Since the presence of CFRP sheets made the observation of cracks difficult, Vc was determined from the measured strain of the stirrups and the concrete. Table 2. Comparison of Specimen Parameters and Test Results Strengthening Gap CFRP Surface f' c Vc Vu (Vu)test Scheme Size Thickness Preparation (MPa) (kn) (kn) (mm) (mm) -- & Type (V u)contro! Control (none) n.a. n.a. n.a Clamped Stirrups n.a. n.a. n.a Vertical CFRP A grinding Vertical CFRP B hydro-blast Diagonal CFRP C hydro-blast Diagonal CFRP B hydro-blast Horiz. & Vert. CFRP A grinding Horiz. & Diag. CFRP C hydro-blast

8 In all tests, flexural shear cracks were observed within the shear span and extended toward the top flange at ultimate. Figure 8 shows the control beam at failure. Just prior to failure, spalling of the concrete cover was observed due to the outward tensile force resultant causing straightening of the bent comer of the stirrups. Figure 9 illustrates the straightening behaviour of the stirrups in the control beam. Figure 8. Control Beam at Failure Figure 9. Stirrup Behaviour Due to the shape of the stirrups, only one of the stirrups in the control beam reached yield before premature failure occured due to spalling of the concrete cover and straightening of the adjacent stirrups. Figure 10 illustrates the location of the stirrups and the observed concrete spalling as well as the crack patterns just prior to failure of the control beam. 3 stirrups with strain gauges 1 stirrup yielded spalling of concrete cover Figure 10. Control Beam Immediately Prior to Failure

9 Effect of the Stirrup Clamping Scheme The clamping scheme applied to the second end of the control beam was effective in controlling the outward force in the stirrups, no spalling of the concrete cover was observed. Since premature failure due to straightening of the stirrups was prevented, all of the measured stirrup strains exceeded the yield strain of 2.8 millistrain and the distribution of forces between the clamped stirrups was improved. Figure 11 shows the location of the stirrups and the crack patterns just prior to failure. A comparison of Figures 10 and 11 suggests that the crack patterns on the web are similar for both beams. 3 stirrups with strain gauges all 3 stirrups yielded stirrup clamping scheme Figure 11. Beam with Clamped Stirrups Immediately Prior to Failure The maximum strains In the clamped sti.rrups were all significantly higher than those observed for the same stirrups in the control beam. This behaviour is illustrated by the stirrup strain versus applied shear load curves shown in Figure 12. The increased ductility and improved distribution of forces among the clamped stirrups contributed to the observed 27 % increase in the ultimate shear capacity, V U' when compared to the un strengthened control beam lo c 100 rjl "0. -Q. 50 Q. < Stirrup Strain (millistrain) Figure 12. Stirrup Strain vs Applied Shear for Clamped & Control Beam Stirrups

10 Beams Strengthened with Vertical CFRP Sheets Two different types of CFRP sheets with similar thicknesses and material properties were '. used for the beams strengthened with vertical sheets. Two gap sizes and surface preparation techniques were also used as shown in Table 2. The ultimate shear capacity, V, of the beam with the 20 mm gap size increased by 17 % compared with the control u beam, while only a 10 % increase was observed for the beam with the 100 mm gap size. It should be noted that the reduced gap size resulted in a 32 % increase in the area of sheets per unit length of beam. The contribution of the vertical sheets with the reduced gap size could also be influenced by the use of hydro-blasting for surface preparation which, according to other researchers [I], results in better bond between the sheets and the concrete.. In both beams shown in Figure 13, failure was initiated by straightening of the stirrups and debonding of the CFRP sheets above and below the diagonal cracks. For the beam with the surface prepared by hydro-blasting, debonding of the sheets was less extensive just prior to failure when compared with the beam prepared using the grinding technique. Figure 13. Beams with Vertical CFRP Sheets and Large and Small Gap Sizes at Failure The vertical strain distributions in the CFRP sheets at various levels of applied shear up to ultimate are shown in Figure 14 and Figure 15. Figure 14 provides the vertical strain distribution for the beam with vertical sheets and a large gap size while Figure 15 shows the vertical strain distribution at the same location for the beam with closely spaced vertical sheets. In both Figure 14 and Figure 15, the increase in the vertical CFRP strain at the crack locations is evident as well as the highly localized behaviour of the CFRP sheets.

11 --E o -,-----r , E '-' E ? O=_ l CO... o "&,i\----,,,;it---==-j.--1 E o ==-----=a u = l-+'--t--+--t-t Strain in FRP Sheet (millistrain) V = 100 kn B V = 125 kn V = 150 kn -6 Vu = 151 kn --E e o _ , '-' e I---.:.-.,j_ I CO... o Ec j e o ==----- u = l c t-t-t--i Strain in FRP Sheet (millis train) V = 100 kn 8 V = 125 kn V == 150 kl -6 Vu = 161 kn Figure 14. CFRP Strain Distribution for Beam with Vertical Sheets and 100 mm Gap Size Figure 15. CFRP Strain Distribution for Beam with Vertical Sheets and 20 mm Gap Size In the beam with the 100 mm gap size, a reduction in the CFRP strain occured at the first crack location, at 125 mm from the top, at a shear load of 125 kn, as shown in Figure 14. An increase in the CFRP strain at a new crack location, at 250 mm from the top, indicates that the first crack began to close while the second crack continued to widen. The strain distribution in the beam with the 20 mm gap size also demonstrates a reduction in the CFRP strain at an initial crack, located at 250 mm from the top, at a shear load of 125 kn, as shown in Figure 15. An increase in the CFRP strain at a new crack located 120 mm from the top indicates that again the first crack began to close while the second crack continued to widen. The maximum recorded stirrup strains were similar and ranged from 2.0 to 2.6 millistrain for stirrups in both beams. The maximum strains measured in the CFRP sheets were also similar, which suggests that the force transferred to the sheets through bond was not significantly influenced by the different surface preparation techniques used. However, it can be concluded from various reports of strain distribution in CFRP sheets bonded to concrete [1,2], that the bond stresses are highly localized which may result in an observed maximum strain that is lower than the actual maximum strain occuring at other locations in the CFRP. Additional bond tests are proposed to further evaluate the effect of the different surface preparation techniques and types of CFRP sheets on the bond performance of the CFRP strengthening systems used in this experimental program.

12 Beams Strengthened with Diagonal CFRP Sheets The thickness and material properties of the two types of CFRP sheets used for the diagonal configurations vary significantly as shown in Table 1 and Table 2. The gap size parameter was also varied for these two tests while the surface preparation technique was not. For both beams, similar 26 % and 29 % increases in ultimate shear capacity, V U' were achieved. Due to the shape of the girder, debonding and straightening of the CFRP sheets was observed on both beams at the lower part of the thin web. Unlike the beams with vertical sheets only, the diagonal sheets remained bonded near the top of both beams at ultimate. After the test, the CFRP sheets were removed to examine the stirrups. Due to the efficiency of the diagonal CFRP contribution and the resulting reduced stress in the stirrups at ultimate, the stirrups did not straighten significantly. The efficiency of the diagonal CFRP sheets is evident when comparing the stirrup strains for the beams in Figure 16. The stirrup strain at any level of applied shear is lower for the beams with diagonal sheets. Although the beam with horizontal and vertical sheets reached a higher ultimate shear load, the stirrup strain was greater. Figure 16 also shows that in spite of the larger gap size, the beam with thicker sheets exhibited lower stirrup strains and therefore a greater contribution from the CFRP sheets at the same level of applied shear load. 250, , Diag. Diag. Lg. gap Sm. gap Boriz. & Vert t == 0.79 t = 0.11 t= ;: 150.c 00. "C Q. Q. -< Control Beam t == thickness of FRP sheet (mm) Stirrup Strain (millistrain) Figure 16. Stirrup Strain versus Applied Shear 5 The strain measured in the diagonal CFRP sheets at a similar location on each beam can also be used to demonstrate the increased contribution of the thicker sheets. At an applied shear load of 150 kn, a maximum strain of 4.07 millistrain was measured in the Type B CFRP sheet. Based on the modulus of elasticity reported by the manufacturer,

13 this strain represents a stress of 936 MPa in the Type B sheet at this location as shown in Figure 17. For the same shear resisting force per unit width of sheet, the stress in the Type C sheet would be only 130 MPa due to the increased thickness. Based on the modulus of elasticity reported by the manufacturer, the strain in the Type C sheet would be only 1.71 millistrain = 7Jl U = '" Jl Strain in CFRP Sheet (millistrain) Figure 17. Stress & Strain for Equivalent Forces in Type B & C CFRP Sheets During the test, at an applied shear load of 150 kn, the maximum strain at a similar location in the Type C sheet was found to be 2.13 millistrain rather than 1.71 millistrain, suggesting that the thicker diagonal sheet contributes more shear resisting force at the same level of applied shear load. Using the same approach but at a lower level of applied shear, the same relationship was observed. However, the maximum strains measured at a different location on both beams indicated that the Type C sheet carried a force per unit width that was almost equivalent to the Type B sheet. As mentioned previously, the highly localized nature of the bond stresses in CFRP sheets may result in an observed maximum strain which may be lower than the actual maximum strain occuring in the CFRP sheet. Further testing is proposed to investigate the load sharing behaviour between the stirrups and the different types of diagonal CFRP sheets used in this experimental program.

14 - Horizontal CFRP Sheets Combined with Vertical or Diagonal Sheets Both beams with a single layer of horizontal CFRP sheets applied on top of the vertical or diagonal sheets, achieved similar 34 % and 36 % increases in ultimate shear capacity, V u, when compared with the control beam. It should be noted that the level of applied load causing shear failure in both beams is within 5 % of the load predicted to cause flexural failure in the beam. Similar to the beams with vertical CFRP sheets only, the beam with both horizontal and vertical sheets, demonstrated spalling on the lower part of the thin web due to the outward force in the stirrups at this location. However, due to the presence of both horizontal and vertical sheets, spalling was observed at a higher load level than in the control beam or the beam with vertical sheets only. For the beam with both horizontal and diagonal sheets, debonding and straightening of the diagonal sheets was observed at the lower part of the thin web similar to the beams with diagonal sheets only, however, the observed debonding was less extensive. In both beams at ultimate, rupture and some debonding of the CFRP sheets was observed very close to the top of the beam as illustrated in Figure 18. Figure 18 shows the beam with both horizontal and vertical sheets at failure. There did not appear to be any significant bond problem between the CFRP and the concrete in either beam. Figure 18. Beam with Horizontal & Vertical Sheets at Failure

15 CONCLUSIONS Four ten meter long prestressed concrete beams were tested to determine the most efficient configuration and type of CFRP sheets that can be used to strengthen existing I-shaped AASHTO girders for shear. The beams are scale models of the Maryland bridge girders which require shear capacity upgrading in order to carry increasing truck loads. Preliminary results have been presented and suggest that shear strengthening using CFRP sheets is an efficient solution for the Maryland bridge. The following conclusions and recommendations are based on preliminary analysis of test results available at the time of submission of this paper: 1. Due to the shape of the stirrups used in the original bridge girders, an outward force is created under increasing tensile force in the stirrups causing spalling off of the concrete cover followed by straightening of the stirrups and premature failure. CFRP sheets are effective in reducing the tensile force in the stirrups under the same applied shear load. 2. The clamping scheme was effective in controlling the outward force in the stirrups. When compared with the control beam, higher stirrup forces were distributed to more stirrups within the shear span, allowing the stirrups to yield and contributing to a 27 % increase in the ultimate shear capacity. 3. The reduced gap size for beams with vertical CFRP sheets provided more area of CFRP per unit length of beam and increased the improvement in ultimate shear capacity from 10 % to 17 %. Use of the hydro-blasting surface preparation technique may also have contributed to the improvement in the shear capacity and will be evaluated further in proposed bond tests. 4. Diagonal CFRP sheets are the most efficient configuration in reducing the tensile force in the stirrups at the same level of applied shear load. The ultimate shear capacity was increased by 26 % to 29 % with the application of diagonal sheets. 5. Due to the shape of the girder and the efficiency of the diagonal configuration, debonding and straightening of the CFRP sheets was observed on the lower part of the web in beams with diagonal sheets. The diagonal sheets remained bonded near the top of the beam. 6. Thicker diagonal CFRP sheets appear to contribute more shear resisting force than thinner sheets at the same level of applied shear load. The effect of the different types of CFRP sheets used on the load sharing between the sheets and the stirrups will be evaluated in further testing. 7. The application of a single layer of horizontal CFRP sheets on top of the vertical or diagonal sheets was shown to increase the ultimate shear capacity by 34 % to 36 % which approaches the flexural capacity of the beam.

16 ACKNOWLEDGEMENTS Funds for this program were provided by the ISIS Canada Network of Centres of Excellence on Intelligent Sensing for Innovative Structures, the City of Winnipeg, Dillon Consulting Engineers, Vector Construction Ltd., Concrete Restoration Services Ltd., the Hexcel Fyfe Co., the Mitsubishi Chemical Corp., and the Tonen Corp. Special assistance was provided by Mr. Moray McVey and Mr. David Donald of the ISIS Canada Network. REFERENCES 1. Yoshizawa, H., Myojo, T., Okoshi, M., and Kliger, H., "Effect of Sheet Bonding Condition on Concrete Members Having Externally Bonded Carbon Fiber Sheet," presented at the 4th Materials Engineering Conference, ASCE Annual Convention, November 1996, Washington D.C. 2. Chajes, M.J., Finch, W.W., Januszka, T.F., and Thomson, T.A., "Bond and Force Transfer of Composite Material Plates Bonded to Concrete," ACI Structural Journal, 93(2),