Anda di halaman 1dari 4

FRPRCS-9 Sydney, Australia

Monday 13 Wednesday 15 July 2009

PEELING FAILURE AT THE CUT-OFF END OF CFRP STRENGTHENED RC BEAMS


Piotr RUSINOWSKI 1,2
1 2

Bjrn TLJSTEN 1

Bjrnar SAND 2

Department of Civil Engineering, Technical University of Denmark, Lyngby, Denmark Norut Northern Research Institute AS, Narvik, Norway

Keywords: debonding, FRP, end-peeling, strengthening.

INTRODUCTION

Fibre Reinforced Polymers, FRP, offer excellent corrosion resistance to environmental agents as well as the advantages of high stiffness-to-weight and strength-to-weight ratios when compared to conventional construction materials. For this reason FRP composites are becoming a material of choice in an increasing number of rehabilitation and retrofitting projects around the world. The application of bonded strengthening requires non-conventional design issues. One such design issue is the debonding problems in externally bonded FRP strengthening applications that have been a concern and a research challenge since the initial development stages of the strengthening method. In particular for externally bonded plates and sheets. One of the most common failure modes in plated members subjected to flexure is end peeling, caused by concentration of shear and normal stresses at the end of the FRP plate. This failure mode has been well known since mid seventies when research and tests of flexural strengthen concrete beams with steel plates were undertaken, [1-2]. This paper presents experimental study performed on CFRP (Carbon Fibre Reinforced Polymer) strengthened RC beams at Lule University of Technology. The aim of the study is to investigate influence of CFRP properties and distance of laminate end from the support on interfacial stresses. Furthermore, the experimental results are verified with theoretical.

EXPERIMENTAL STUDY

A total number of 6 RC beams were tested in the experimental study. All specimens were cast in one batch in order to provide possibly identical concrete properties. Five beams were strengthened by bonding CFRP laminates to their soffits and one beam remained without strengthening as a reference. All of the strengthened beams were unique, differing from each other with laminate length or laminate type. General design of tested beams is shown in Figure 1 and test matrix is presented in Table 1. Actual concrete compressive and splitting strengths for each beam were determined by testing cubes according to Swedish Standards (SS 13 72 10 and SS 13 72 13). Actual strength of reinforcing steel was determined by tensile tests. Material properties of CFRP laminates were not tested and nominal data provided by the manufacturers were used in calculations. All beams were subjected to 4-point-bending in displacement controlled loading. Mid-span deflection and support settlements were measured with LVDTs. Strain gauges were glued on each laminate in order to determine strain distribution in CFRP and estimate interfacial shear stresses, see Figure 1. More details on experimental programme can be found in [3].

Fig. 1 General design of tested beams and location of strain gauges 1

FRPRCS-9 Sydney, Australia Table 1


Beam

Monday 13 Wednesday 15 July 2009 Test matrix.

tf Ef fsy fcc fct Ec Lf a ta Ga wf ffu CFRP type [MPa] [MPa] [GPa] [mm] [mm] [mm] [GPa] [mm] [mm] [MPa] [GPa] [MPa] RB --------------527 70.0 3.8 46.0 4.9 S1 3300 150 2.1 1.4 120 2400 155 527 64.0 4.0 44.0 Sika CarboDur S 4.9 S2 3200 200 2.5 1.4 120 2400 155 527 65.0 4.1 44.0 Sika CarboDur S 4.9 S3 3000 300 3.1 1.4 120 2400 155 527 65.0 4.1 44.0 Sika CarboDur S 4.9 M1 3300 150 2.4 1.4 260 2000 210 527 70.0 4.0 46.0 Sika CarboDur M 4.9 H1 3300 150 2.1 1.4 250 1400 300 527 70.0 3.8 46.0 Sika CarboDur H Lf laminate length, a see Figure 1, ta adhesive thickness, Ga laminate shear modulus, wf laminate width, ffu laminate ultimate strength, Ef laminate modulus of elasticity, fys steel yield strength, fcc concrete compressive strength, fct concrete tensile strength, Ec concrete modulus of elasticity

ANALYTICAL STUDY

Apart from experimental studies mentioned earlier in this paper, numerous analytical models have been derived for shear and peeling stresses at laminate end. Comparison performed in [4] shows that most of known models give similar results and differ rather in complexity and range of application. One of the most recent models presented in [5] proposes very general solution considering various load cases and such factors as bond defects or end tapering. For needs of this paper model proposed in [6] was used in calculations. The solution shown in this model can be used for concentrated loads.. The model was used also further in prediction of the ultimate load assuming that the maximum principal stress at the laminate end should not exceed concrete tensile strength, as suggested in [7]. Apart from the analytical model presented in [6], the tested beams were studied with recommendations proposed by ACI 440.2R-02 design guidelines [8] and fib Bulletin 14 [9]. The ACI design guidelines do not consider the case of end-peeling separately and does not give any suggestions for required distance of strengthening end to support. The model consists in reduction of allowable strain in FRP in order to avoid debonding failure. Maximum strain in FRP, determined in Eq. 1, depends only on modulus of elasticity and thickness of strengthening material.

deb =

(1 E f t f 360000 ) 60 0.90 fu for E f t f 180000 N/mm 90000 ( 60 E f t f ) 0.90 fu for E f t f >180000 N/mm

(1)

In fib recommendations debonding phenomenon is studied deeper and different models are presented depending on type of premature failure. For shear failure at laminate end a simple model is proposed where failure load does not depend on material properties of strengthening. Also simple models are used to predict anchorage failure and peeling-off at shear crack. More complex model is proposed for peeling-off at flexural crack where the ultimate load is limited by maximum allowable stress in FRP.

RESULTS

In Figures 2a and 2b load-deflection curves obtained in the experimental study are presented. In Figure 2a diagrams for beams with varying distance from laminate end to support (S1, S2, S3) are shown. Figure 2b shows curves obtained in testing beams S1, M1 and H1 where different types of laminates were applied. Furthermore, both figures include load-deflection curve from tests of the reference beam without strengthening. Since the reference beam was supposed to be repaired and used in other experimental programme, the loading was not continued until failure. All tested strengthened beams failed by debonding. The ends of the laminates of beams S1, S2 and S3 were peeled off with a thin layer of concrete which indicates that end peeling was possible mode of failure. In case of beams M1 and H1 the laminate ends were peeled off without adhesive which suggests that debonding at crack caused the failure, see Figure 2c. This can be also judged by comparing load capacities of tested beams, see Figure 2b. Specimens M1 and H1 strengthened with stiffer laminates, which should be more susceptible to end failure, reached higher loads. As it can be seen in Figure 2a the distance from laminate end to support did not have significant influence on load capacity. Beam S1 with the laminate end located closest to the support had lower load capacity than

FRPRCS-9 Sydney, Australia


200 150 Load, P [kN] 100
S1 S2 S3

Monday 13 Wednesday 15 July 2009


200
H1

Beams S1, S2, S3


M1

150 Load, P [kN] 100

S1

50 0

Ref.

50 0

Ref.

a)

10 20 30 40 Deflection [mm]

50

b)

10 20 30 40 Deflection [mm]

50

c)

Beams M1, H1

Fig. 2 a, b) Load-deflection curves, c) specimens after failure two other beams. This can be, however, a result of differences in crack pattern, improper bonding or thinner adhesive layer in beam S1, see Table 1. Table 2 presents results of the analytical studies according to ACI and fib recommendations compared with experimental results. It can be seen that fib debonding models are more conservative and, although characteristic values were used for analyses, underestimate load capacities of tested beams significantly. Considering failure modes, only beams S2 and S3 should fail at laminate end while all other beams should collapse by peeling-off at flexural crack according to fib recommendations. The ACI model underestimated load capacity of beam H1 significantly. The fact of conservative approach of this model to strengthening with laminates of high Eftf value has been commented in [10]. Load capacities of beams S1, S2 and S3 were overestimated with ACI mode. In case of beam S1, as mentioned before, the experimental result is rather unreliable. In case of beams S2 and S3 the inaccuracy of prediction is possibly due to end peeling failure not covered by ACI model. Figures 3a and 3b show interfacial shear stresses at the laminate ends at different loading stages. The experimental results, calculated from measured strains in CFRP laminates, are compared with results obtained from the model in [6]. Interfacial stresses measured in the experiments are in relatively good agreement with the theory at early stages of loading. For failure loads stresses obtained from the experiments are significantly higher than those estimated in the analytical study. The analytical model for shear and peeling stresses at laminate end assumes that both adherents are Table 2
Beam S1 S2 S3 M1 H1 Analytical model [6] 130.8 103.8 70.9 128.3 112.5

Comparison of ultimate loads (Pmax) from experiments with analytical predictions


Anchorage faiure 205.9 166.4 115.6 199.2 176.4 fib Bulletin 14 Peeling-off at Peeling-off at shear crack flexural crack 195.4 109.9 195.4 110.1 195.4 110.1 219.5 112.8 219.5 97.2 ACI 440.2R-02 Peeling-off at laminate end 115.8 108.0 97.1 119.5 119.5 Debonding 152.0 152.1 152.1 151.2 137.4 Experiments 142.8 151.0 147.7 159.1 160.2

6
S3 (exp)

6 Shear stress, [MPa]


S2 (exp) H1 (exp) S1 (exp) M1 (exp) H1 (theory) M1 (theory) S1 (theory)

5 Shear stress, [MPa] 4 3 2 1 0


Experiment Theory

Shear stress, [MPa]

5 4 3 2 1 0 0 0.2 0.4 0.6 0.8 1 1.2 1.4 Pmax , Load [-]


S1 (exp) S3 (theory) S2 (theory) S1 (theory)

5 4 3 2 1 0 0

a)

b)

0.2 0.4 0.6 0.8 1 1.2 1.4 Pmax , Load [-]

c)

50 100 150 200 Dist. from plate end, x [mm]

Fig. 3 a, b) Comparison of shear stresses at laminate end, c) comparison of analytical with experimental shear stress distribution

FRPRCS-9 Sydney, Australia

Monday 13 Wednesday 15 July 2009

elastic materials. In reality, however, cracking in concrete influences actual interfacial stresses. It can be seen in Figure 3c, which shows shear stress distribution for beam S1 at the ultimate load, that maximum interfacial shear stress occur in relatively long distance from laminate end comparing to the analytical result. This phenomenon may be caused by crack in concrete instead of laminate cut-off end. Also concrete cracking is believed to be reason of divergence between results obtained in the experiments and elastic end-peeling model presented in Table 2. The model underestimates performance of tested beams, especially in case of specimens with long distance from laminate end to the support.

CONCLUSIONS

The aim of the research was to perform experimental study on RC beams with different arrangements and different types of CFRP strengthening and compare its results with analytical models. The desired failure mode was peeling-off at laminate end. The analysis according to fib recommendations shows, however, that tested beams having laminate ends located closer to the support (S1, M1, H1) are more susceptible for peeling-off at flexural crack. In fact, debonding occurs very suddenly and its type is difficult to judge without optical measurements, as shown in [11]. Using elastic end-peeling model in predicting load capacity of strengthened concrete beams seems to underestimate actual load capacity although measured maximum shear stresses were generally higher then those calculated with the analytical model. The maximum measured stresses, however, occurred in relatively long distance from laminate end and might have been caused by flexural crack.

ACKNOWLEDGEMENTS
The research programme was financed by SKANSKA, SBUF (The Development Fund of the Swedish Construction Industry) and The Norwegian Research Council through the strategic institute programme RECON at Norut Northern Research Institute AS. Former students Abderahim Aboudrar and Anders Johansson are greatly acknowledged for performing and evaluating laboratory tests.

REFERENCES
[1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] Jones R., Swamy R.N., Bloxham J. and Bjounderbalaha A., Composite Behaviour of Concrete Beams with Epoxy Bonded External Reinforcement, The International Journal of Cement Composites, Vol. 2, No. 2, 1980, pp 91-107. Johnson R.P. and Tait C.J., The Strength in Combined Bending and Tension of Concrete Beams with Externally Bonding Reinforcing Plates, Building and Environment, Vol. 16, No. 4, 1981, pp 287-299. Aboudrar A. and Johansson A., Betongbalkar frstrkta med kolfiberkomposit, Masters Thesis, Lule University of Technology, Department of Civil Engineering, Section of Structrual Engineering, ISSN: 1402 1617, 1998, p 172 (In Swedish). Smith, S.T. and Teng, J.G. Interfacial stresses in plated beams, Engineering Structures, 23, 2001, pp 857-871. Stratford, T. and Cadei, J., Elastic analysis of adhesion stresses for the design of a strengthening plate bonded to a beam, Construction and Building Materials, 20 (2006), 2006, pp 34-45. Tljsten, B., Strengthening of beams by plate bonding, Journal of Materials in Civil Engineering, Vol. 9, No. 4, 1997, pp 206-212. Tljsten, B., FRP Strengthening of existing concrete structures - Design Guidelines. Third edition. Lule, Sweden: Lule University printing office, ISBN 91-89580-03-6, 2004, p 230. ACI 440.2R-02, "Guide for the design and construction of externally bonded FRP systems for strengthening concrete structures." ACI Committee 440, 2002. fib, "Externally bonded FRP reinforcement for RC structures", Bulletin 14, 2001. Said, H. and Wu, Z., Evaluating and proposing models of predicting IC debonding, Journal of Composites for Construction, Vol. 12, Issue 3, 2008, pp 284-299. Rusinowski, P. and Tljsten, B., Interlaminar peeling in concrete beams strengthened with CFRP plates experimental study, Proceedings of the First Asia-Pacific Conference on FRP in Structures (APFIS 2007), S.T. Smith, ed. The University of Hong Kong, Hong Kong, China, 12-14 December, 2007, pp 811-816. 4

Anda mungkin juga menyukai