Jtam-A4.dvi JOURNAL OF THEORETICAL AND APPLIED MECHANICS 54, 4, pp. 1137-1146, Warsaw 2016 DOI: 10.15632/jtam-pl.54.4.1137 MODE OF FAILURE FOR REINFORCED CONCRETE BEAMS WITH GFRP BARS Abdelmonem Masmoudi Mechanics, Modelling and Production Laboratory, National School of Engineering of Sfax, University of Sfax, Tunisia e-mail: abdelmonem.masmoudi@enis.rnu.tn Mongi Ben Ouezdou Université de Tunis El Manar, Ecole Nationale d’Ingénieurs de Tunis, LR03ES05 Laboratoire de Génie Civil, Tunis, Tunisia; e-mail: mongi.benouezdou@enit.rnu.tn Mohammed Haddar Mechanics, Modelling and Production Laboratory, National School of Engineering of Sfax, University of Sfax, Tunisia Crack propagation in concrete structures is a very complicated process. An experimental method to predict the stress distribution of a cracked GFRP reinforced concrete flexural beam is presented. The beam subjected to four-point bending is internally reinforced with unidirectionalGFRPbars.The aimof this investigation is to study the flexural performance of reinforced concretemembers strengthenedusingGFRPbars.For the flexural tests perfor- med on reinforced concrete beams strengthened with GFRP bars, the ultimate load of the beams strengthened with GFRP was reasonably increased. The mode of failure for beams reinforcedwithGFRP sections is slightly different comparedwith the traditional beam.The GFRP reinforced concrete beams fail either by concrete crushing at the compression zone or rupture of the GFRP reinforcement. Keywords: crack propagation, flexural failure, concrete, GFRP bars, stress distribution 1. Introduction Nowadays, the construction industries around theworld face amajor problemdue to corrosion of steel reinforcement. The cost of maintenance of any deteriorated reinforced concrete structures is very expensive. GFRP composite bars are an excellent alternative to steel bars for reinforcing concrete structures in severe environments. In the last two decades, a number of studies have been carried out to investigate the flexural responseofFRPRCbeams (Ritchie, 1991;An et al., 1991;Meier et al., 1993). Therefore, flexural failure is a commonphenomenon in the case of concretemembers.Plain concretewhen subjected to flexural loads fails due to crack propagation. In this work, the effects of several further important parameters on the formation and propagation of cracks in concrete are investigated in order to better understand the decisive phenomena of concrete failure under different loading conditions. When RC beams reach ultimate state with the first two failure modes, the load- -carrying capacity of the beams can be easily estimated by applying conventional methods derived using a plane conservation concept. Experimental studies have been performed to investigate flexural, shear, and failure modes of FRP RC beams (Swamy and Roberts, 1995; Saadatmanesh and Ehsani, 1991; Ritchie et al., 1991;Chajes et al., 1994; Sharif et al., 1994). These studies indicated failuremodes that can limit the strengthening effect of FRP retrofitted structures. Recently, four-point bending experiments were performed on FRP strengthened RC beams and compared structural behaviors of beams retrofittedbyCFRP(carbonFRP) forflexural strengtheningandbyGFRP(glassFRP) for shear 1138 A. Masmoudi et al. strengthening (Kachalakev and McCurry, 2000). The FRP reinforcement technique for bridge and continuousRCbeamapplicationswas investigated byPhamandAl-Mahaidi (2004), Ashour (2004). Both studies usedCFRP sheets and observed the failuremechanismof the reinforcement beams.The effect of initial load on the structural behavior of FRPRCbeamswas also studiedby Shin and Lee (2003),Wenwei andGuo (2006), who performed experiments onCFRPRCbeams under different sustaining loads. The former study showed that the different sustain load levels have influence on the deflection of the beam. The latter study investigated the effect of initial load on the ultimate strength of the CFRP reinforcement RC beams and a theoretical model was proposed that could predict ultimate strength of the beams under sustaining loads. Shear capacity of FRP reinforcement beamswas presented (Bencardino et al., 2007; Jayaprakash et al., 2008). The first study tested CFRP reinforcement RC beams cast without shear reinforcements and observed failure modes in shear. It was concluded that the FRP retrofitted RC beams could avoid shear failure if a carefully designed anchorage system was installed in the beams. Another study (Jayaprakash et al., 2008) showed that bi-directionally attached CFRP strips significantly increased the shear capacity of the RC beams. They also investigated the effect of orientation ofCFRPstrips onultimate shear capacity aswell as crack propagation.The cracking may significantly affect the moment redistribution in continuous members and more accurate assessment of the influence of FRP reinforcement on the behavior of RC elements (Matos et al., 2012). 2. Experimental study 2.1. Materials For all beams ready mixed concrete was used which had been aged for 28 days. 25MPa concrete grade was used in the manufacturing of these beams using Ordinary Portland cement and crushed aggregates with maximum size of 12mm. The initial elasticity modulus E was 34GPa and the measured tensile strength was 2.1MPa. The concrete slump was 100mm. For the internal GFRP reinforcement 4 bars HA12 (diameter of 12mm) were used. The resistance capacity in shear was provided from 6mm diameter shear reinforcement with 120mm spacing. Properties of the GFRP and steel bars used in this study and details of beam cross-section are shown in Table 1 (Schock Bauteil GmbHCombar Co., 2006; Aboutaha, 2004). The average ultimate tensile strength was 738MPa and 400MPa, respectively for GFRP and steel bars. The modulus of elasticity of the tensile reinforcement bars was 60GPa and 200GPa, respectively for GFRP and steel. The beams were casted with concrete with a cover of 20mm. For the tensile reinforcement, two 12mmdiameter bars were used, and for the construction reinforcement, two 8mmdiameter as the upper reinforcement (Fig. 1). Table 1.Properties of the GFRP and steel bars used in this study Type of bar Glass Steel Nominal diameter 8 and 12mm Tensile modulus of elasticity [GPa] 60±1.9 200±7 Ultimate tensile strength [MPa] 738±22 400±11 Density 2.2 7.85 2.2. Variables of the experiments and beams In this investigation, four point flexural testswere performedwith the experimental variables being shear strengthening or no shear strengthening in the central zone. In our experimental stu- dy, we investigated the influence of the internal reinforcement amount in the compositematerial Mode of failure for reinforced concrete beams with GFRP bars 1139 Table 2.Composition and characteristics of the concrete Cement I [kg/m3] Water [kg/m3] Sand [kg/m3] Aggregate Aggregate Compressive Slump [mm] (4/12) (12/20) Strength [kg/m3] [kg/m3] [MPa] 350 207 880 300 650 25 100 Fig. 1. GFRP bars reinforcement and the beam cross section on theRCbeam response. Two types of beams are studied,with shearGFRP internal reinforce- ment in central flexural zone in comparison andwith steel internal reinforcement. Three beams were reinforcedwithGFRPbars (designated BG) and threewith steel bars (designated BS). All beams were provided with 6mm diameter mild stirrup and were designed to fail in flexure. All reinforced concrete beamshad spamof 1650mmand rectangular cross section 200mm×150mm. The beam cross section was presented in Fig. 2. All beam tests were carried out using a cali- brated 500kN testingmachine with displacement-rate control. The data acquisition systemwas started a few seconds before load application. The displacement rate of loading was kept con- stant during the tests (0.005mm/s). The beamwasmounted on cylindrical contacts whichwere simply supported. Crack width was controlled step by step during loading. The cracks were marked by a color marker directly on each beam. Fig. 2. Beam cross section (all dimensions are in mm) The objective of this investigation is to study the mechanisms of flexural failure and stress distribution.Themechanical properties enhancement have been evaluated trough comparison of the strengthened beamsmechanical response with themechanical response of the control beam with steel internal reinforcement. The beam response has been quantified in terms of stress distribution and deflections at the mi-spam. The strain evolution as well as some components of the displacement field has also been monitored. 1140 A. Masmoudi et al. 2.3. Arrangement of strain gauges Electrical strain deformeter was installed on the flexural plate surface. To obtain an accura- te deflection reading, four Linear Variable Differential Transducer (LVDT) were also mounted and one at the mid-span and them connected to a data logger. Since the interface meso-crack propagation pattern was known, the monitoring of the local failure behavior could provide in- formation on the crack-initiation point and on the direction of crack propagation. Electrical strain deformation was installed in order to obtain deflection reading by LVDT (Fig. 3). The detectors were installed on the flexural for both steel andGFRP reinforcement and strain versus load plots were considered. The time “zero” was chosen at the beginning of the load history, or at the moment in which the load was applied. Thus, if the local failure is considered as a “moment” in the load history, the sequence of local failures in the monitored spots can reveal the interface crack initiation point and the crack propagation direction. A schematic diagram of testing arrangement is presented in Fig. 3.We used uni-axial strain gages type CEA-06-250-350 with factor gage 2.08± 0.5% and resistance 350.0± 0.5% at 24◦C. Figure 4 shows the strain gages fixed to the GFRP bars and attached to the extensometer pont. Fig. 3. Schematic diagram of the test set-up Fig. 4. Gauges for longitudinal deformations 2.4. Longitudinal modulus of elasticity Uniaxial compression tests on cylindrical samples 16× 32cm were performed to determine the longitudinal modulus of elasticity of concrete used for casting the concrete test beams. The device used showed in Fig. 5 was mounted on a hydraulic press with numerical control. The maximum capacity of this press was 5000kN with loading speed compression 0.005mm/s. The longitudinal elastic modulus of the concrete measured on new specimens was identical to Mode of failure for reinforced concrete beams with GFRP bars 1141 the given average of 34122MPa. The evolution of the stress in the concrete according to the deformation is shown in Fig. 6. Fig. 5. Test of the longitudinal elasticity modulus Fig. 6. Evolution of the stress in the concrete vs deformations 3. Analytical study The term of deformation of the steel in tension εs0 in the case of uncracked section can be expressed by the following expression εs0 = Z0 Z1 ( εs1+ Nc As ) (3.1) where Z0 being the lever arm in the cracked section, Z1 – lever arm in the uncracked section, Nc – normal tensile stress of concrete. The deformation of steel in the case of cracked section ε1 can be found from a conventional calculation by the expression εs1 = d−y EcI M (3.2) where y is the position of the neutral axis in the cracked section, Ec – modulus of elasticity of the concrete, I – inertia of the cracked section. The conditions of adhesion between reinforcement bars and concrete are modeled involving a scalar variable mechanical damage D (between 0 and 1) given by the following expression εc = (1−D)εs0 n (3.3) where εc is the deformation of the concrete area, εs0 – deformation of bar reinforcement in the uncracked section. 1142 A. Masmoudi et al. From equations (3.1) and (3.3) we can express, in the case of the uncracked concrete section, the normal strains in concrete and in the bars. They are as follows εc = εs1 Z0 Z1 ( 1 1−D + C nAs ) εs0 = εs1 Z0 Z1 ( 1+ (1−D)C nAs ) (3.4) whereC is the contribution of the tensioned concrete surface andn is the equivalence coefficient of concrete steel.We compare the evolution of experimental and theoretical deformations of two reinforcement concrete beams for both BG and BS beams. 4. Presentation and discussion of the test results 4.1. Load-deflection The load-deflection behavior of all the beams tested is shown in Fig. 7. Initially, all the beams have relatively the same stiffness. However, once the beam cracked, the stiffness of the GFRPreinforced concrete beamdecreased at a faster rate comparedwith the control beam.This resulted in a larger deflection of the GFRP reinforced concrete beam. The recorded deflections near failure for all beamsBG, andBSwere, 50mm, and 32mm, respectively. It can be seen from Fig. 7 that the stiffness of the beamsBGwasmuch lower than that of the beamBS.Again, this was due to the lower elastic modulus of the GFRP sections compared with steel reinforcement. At the same load level, the deflection of beams reinforced with GFRP sections was higher by about 2 to 3 times compared with the beam BS. Thus, at the service load, deflection of the beam reinforced withGFRP sections would be higher than in the beamBS andmay not satisfy the design criteria. In addition, a larger deflection would also lead to a wider crack width of the beam. Fig. 7. Load-deflection of all tested beams 4.2. Load-reinforcement strain The tensile strain of the reinforcements was measured and recorded using electrical strain gauges. The load-reinforcement strain behavior of all the beams tested is shown in Fig. 8. It can be seen that the behavior of the load-reinforcement strain was quite similar to the load- deflection of the beams. An increase in the applied load increased the tensile strain of the reinforcement. FromFig. 8, it can be seen that the bond between concrete andGFRP and steel reinforcements was relatively good. That ensured the transfer of tensile load from concrete to the tensile reinforcements. The experimental results also indicated that the strain of the GFRP reinforcement had linear behavior up to failure. On the other hand, the steel reinforcement Mode of failure for reinforced concrete beams with GFRP bars 1143 had a yield point before failure. Thus, in the design process, the aspect of ductile behavior of the beam needs to be taken into account based on the type of tensile reinforcement used. The recorded tensile strain near failure for beams BGI and BS were about 15000 and 5000 micro strains, respectively. On the other hand, the steel reinforcement started to yield at about 3200 micro strains. Obviously, the behavior of the steel reinforcement was elastic-plastic while the GFRP section experienced only elastic behavior. These different strain characteristics of the reinforcement have to be considered when the GFRP section is to be used as concrete reinforcement. Fig. 8. Load-reinforcement strain of all tested beams 4.3. Failure mode The recorded experimental results show that all the beams failed in flexure by crushing of concrete at the compression zone. The total number of cracks generated for beams BG and BS were 21 and 6, respectively. Hence, the beam with a lower ultimate load due to a lower elastic modulus experienced a lower number of cracks compared with the beam that had a higher load carrying capacity. In addition, the crack spacing for the beamBSwas also larger than beamBG and BS. The measured average crack spacing for beams BG and BS were 40mm and 130mm, respectively. It was also observed that the first crack of the GFRP reinforced concrete beams BG was higher by 50% compared with the beam BS. This result was confirmed by Mias et al. (2015) who found that the use of GFRP bar reinforcement led to an increase in the average crack spacing and crack width. The first crack load for GFRP reinforced concrete beams was 5 kN while for the beam BS this value was 10kN. A schematic diagram of the cracking of all the beams tested in this study is shown in Fig. 9. Fig. 9. Diagram of the cracking of beams Crack propagations were observed during the tests. They are illustrated in Fig. 10. In the control specimen, crack initiation occurred at 2.1kN around the locations where the load was applied. As the applied load increased, cracks propagated from the beam center and loaded points. When the applied load reached 40.2kN, the reinforcing bar yielded and flexure failure was observed in the specimen. In this load case, the interface crack was initiated at the beam mi-spam and propagated towards the support. At the same time, the data collected from the strain deformeter placed on the composite material showed that the local failure took place 1144 A. Masmoudi et al. Fig. 10. Level of crack propagation simultaneously at all monitored regions. Moreover, the simultaneous local failure detected in the composite material coincided with the global failure of the beam and with the “last” local failure on the flexural steel/concrete cover interface. Figures 11 and 12 show that the curve of superior fiber strain (F-sup1 and F-sup2) decrease above the load of 25kN. This means that the concrete was damaged and plasticized. Contrary to the curve of the inferior fiber strain (F-inf1 and F-inf2), it was still linear. This result confirms the GFRP linear elastic behavior until rupture. The first cracks appeared in the beam at a load of 25kN. There were five active cracks with spacing of 62.4mm in the half span of each beam. Crack widths corresponding to the 45kN load of the beam were respectively 0.325mm, 0.315mm, 0.361, 0.324 and 0.373mm. We compared the evolution of experimental and theoretical deformations of two reinforcement concrete beams for both types BG and BS. Figures 13 highlights the transfer effort of the reinforcing bar to the concrete in tension. This finding is best seen in the case of the steel beam. An increase in the applied moment generates multiple simultaneous effects. On the one hand, the deterioration of adhesion will continue locally in the portions located between the bending cracks.When degradation is complete, the concrete will not be driven by the reinforcements and mechanically no longer participate. On the other hand, new bending cracks will form in areas closer and closer to the supports. Finally, the area of non-compliance with respect to damage to the adhesive will expand, which means that other sections located between the bending cracks will suffer degradation of adhesion. Fig. 11. Beam crackingmodes: (a) steel beam, (b) GFRP beam Fig. 12. Strain evolution Mode of failure for reinforced concrete beams with GFRP bars 1145 Fig. 13. Evolution of strain in the bar versusmoment The local mechanical behavior has been compared with experimental measurements, the results show good correlations.We can conclude that the local failuremechanismwhich controls the global failure behavior of the strengthened system takes place in the concrete cover for the beam geometry. 5. Conclusion In this paper, flexural tests were performed for concrete beams reinforced by GFRP bars. The main conclusions drawn from the study are as follows: • The concrete beam reinforced with GFRP sections experienced a lower load carrying ca- pacity and stiffness compared with the conventional reinforced concrete beam. This was mainly due to the lower elastic modulus of the GFRP section compared with the steel reinforcement. • The number of cracks for the beam reinforced with GFRP section was higher than in the conventional beam. In addition, the average crack spacing of theGFRPreinforced concrete beamwas also larger compared with the control beam. • The curve of superior fiber strain (F-sup1 and F-sup2) decreased above the load of 25kN. Thismeans the concretewasdamaged andplasticized.Contrary to the curve of the inferior fiber strain (F-inf1 and F-inf2) it was still linear. This result confirms the GFRP linear elastic behavior until rupture. • An increase in the appliedmoment generates the deterioration of adhesion, and new ben- ding cracks are developed. • Themodes of failure for beams reinforcedwithGFRP sections were slightly different com- paredwith the control beam.TheGFRP reinforced concrete beams fail either by concrete crushing at the compression zone or rupture of the GFRP reinforcement. Failure due to rupture of GFRP reinforcement is not recommended since it may result in catastrophic failure of the structure. Acknowledgments The authors would like to thank the manufacturer of the GFRP Combar? (Schöck, Baden-Baden, Germany) for providingGFRP bars and supporting the experimental research. The opinion and analysis presented in this paper are those by the authors. References 1. Aboutaha R., 2004, Recommended Design for the GFRP Rebar Combar, Technical Report, De- partment of Civil and Environmental Engineering, Syracuse University, USA 1146 A. Masmoudi et al. 2. An W., Saadatmanesh H., Ehsani M.R., 1991, RC beams strengthened with FRP plates. II: Analysis and parametric study,ASCE Journal of Structural Engineering, 117, 11, 3434-3455 3. Ashour A.F., El-Refaie S.A., Garrity S.W., 2004, Flexural strengthening of RC continuous beams using CFRP laminates,Cement and Concrete Composites, 26, 765-775 4. BencardinoF., SpadeaG., SwamyR.N., 2007,Theproblemof shear inRCbeams strengthened with CFRP laminates,Construction and Building Materials, 21, 1997-2006 5. Chajes M.J., Thomson Jr T.A., JanuszkaT.F., Finch Jr W.W., 1994, Flexural strengthe- ning of concrete beams using externally bonded composite materials, Construction and Building Materials, 8, 3, 191-201 6. Eurocode 2. “Calcul des structures en béton armé, partie 1.1, règles générales et règles pour les bâtiments”, AFNOR, 1992 7. Jayaprakash J., SamadA.A.A.,AbbasovichA.A., Ali A.A.A., 2008, Shear capacity of pre- cracked and non-precracked reinforced concrete shear beamswith externally bonded bi-directional CFRP strips.Construction and Building Materials, 22, 1148-1165 8. Kachalakev D., McCurry D.D., 2000, Behavior of full-scale reinforced concrete beams retro- fitted for shear and flexural with FRP laminates,Composites, Part B, 31, 445-452 9. MatosB.,Correia J.R.,CastroL.M.S.,FrançaP., 2012,Structural responseof hyperstatic concrete beams reinforced with GFRP bars: effect of increasing concrete confinement, Composite Structures, 94, 1200-1210 10. Meier U., Deuring M., Meier M., Schwegler G., 1993, Strengthening of structures with advanced composites, [In:] Alternative Materials for Reinforcement and Prestressing of Concrete, J.L. Clarke (Edit.), Blackie Academic and Professional, 153-171 11. MiasC.,Torres L.,GuadagniniM.,TuronA., 2015,Short and long-termcrackingbehaviour of GFRP reinforced concrete beams,Composites, Part B, 77, 223-231 12. PhamH.,Al-MahaidiR., 2004,Experimental investigation into flexural retrofitting of reinforced concrete bridge beams using FRP composites,Composite Structures, 66, 1/4, 617-625 13. Ritchie P.A., 1991,External reinforcement of concrete beams using fiber reinforced plastics,ACI Structural Journal, 88, 4, 490-496 14. Ritchie P.A., Thomas D.A., Lu L.W., Connelley G.M., 1991, External reinforcement of concrete beams using fiber reinforced plastics,ACI Structural Journal, 88, 4, 490-500 15. SaadatmaneshH., EhsaniM.R., 1991,RCbeams strengthenedwith FRPplates. I: Experimen- tal study,ASCE Journal of Structural Engineering, 117, 11, 3417-3433 16. Schock Bauteil GmbH Combar Co., 2006, Design Guideline for Concrete Structures Reinforced with Glass Fiber Reinforced Polymer following the Requirements of DIN 1045-1and EC2. Design Guide; Issued Germany 17. Sharif A., Al-Sulaimani G.J., Basunbul I.A., Baluch M.H., Ghaleb B.N., 1994, Streng- thening of initially loaded reinforced concrete beams using FRP, ACI Structural Journal, 91, 2, 160-168 18. Shin Y., Lee C., 2003, Flexural behavior of reinforced concrete beams strengthened with CFRP laminates at different levels of sustaining load,ACI Structural Journal, 100, 2, 231-239 19. SwamyR.N., RobertsH., 1995, Structural behavior of externally bonded steel platedRCbeams after long-term exposure, Structural Engineering, 73, 16, 255-261 20. Wenwei W., Guo L., 2006, Experimental study and analysis of RC beams strengthened with CFRP laminates under sustaining load, International Journal of Solids and Structures, 43, 1372-1387 Manuscript received September 22, 2015; accepted for print February 4, 2016