LITH INTERNATIONAL, ‘ONFERENCE ON COMPOSITE, SCIENCE AND TECHNOL: Experimental and Analytical Study on the Bending Behavior of Concrete-GFRP Hybrid Beams KOAIK ALAA', BEL SYLVAIN!, JURKIEWIEZ BRUNO! * Laboratoire des Matériaux Composites pour la Construction (LMC2), Claude Bernard University, 82 boulevard Niels Bohr, 69622, Villeurbanne, France e-mail: alaa.koaik(@univ-lyon|.fr Abstract Recently, the use of GFRP pultruded profiles increased in the domain of civil engineering especially in the construction of sandwiched slabs and footbridges. However under heavy loads, the risk of using these profiles increases due to their high deformability and instability as a result of their weak stiffness and orthotropic nature. A practical solution proposes the assembly of these profiles with concrete slabs to create a stiffer hybrid element to support higher loads. The connection of these two elements is established either by traditional means of steel studs (bolting in our case) or bonding technique. This paper presents experimental results of interface characterization and bending behavior of two hybrid beams, PB7 and PBB, designed and constructed using both connection techniques. Analytical methods are also developed to analyze the behavior of these structures in the elastic range and the ultimate phase. Comparisons show acceptable differences mainly due to the sensitivity of the GFRP moduli as well as the non-linearity of concrete elements, Key words Analytical model, concrete, flexural behavior, GFRP pultruded profile, hybrid structure, push-out test L. Introduction The use of fiber reinforced plastics (FRP) has been an innovation in aeronautics and marine applications since before the 1980s. Recent studies integrated these materials in the construction domain ue to their numerous interesting characteristics. However, these materials possess a low elastic modulus E as well as a low shear modulus G. In consequence, under heavy loads, high deformations may cause instability phenomena. A proposed solution is the association of these elements (glass fiber reinforced polymers profiles) with concrete elements to create a stiffer hybrid structure able to resist higher loads [1- 3], In this spirit, this paper discusses the conception, construction, and the testing of two hybrid beams PBT and PBS using the combination of the two types of connection (adhesive and bolts). The experimental program is accompanied by an analytical study of deformations, stresses, deflections and loads in both service limits and in the ultimate-failure state aExperimental and Analytical Study on the Bending Behavior of Concrete-GFRP Hybrid Beams: Koaik, A. etal 2. Experimental program 2.1 Materials The compressive and tensile strengths of concrete were 44.0 4.0 MPa and 3.4 + 1.4 MPa respectively. The GFRP clastic modulus average value obtained was 27.0 + 2.0 GPa with an ultimate tension stress in the GFRP of 350 + 60.0 MPa and a shear modulus Gy of 2.6 + 0.3 GPa, Tension tests were done on dried epoxy adhesive coupons and resulted in an average ultimate stress of 13.4 + 1.3 MPa. Class 8.8 bolts with a diameter of 10 mm were used in PB7 and in PBS class 6.8 bolts with a diameter of 12mm, 2.2 Connection (test & results) Given that push-out specimens with a mixed connection are hard to prepare, three push out series were constructed (Fig. 2-B) to evaluate the stiffness and strength of each shear connection system (table 1), The displacements were recorded and analyzed by digital image correlation method (DIC) [4] ‘Table 1 Push-out series details and results. “Maximal force / bolt (KN) ‘Or maximal stress (MPa) Stiffness (N/m) Bolling connection ( class 68 POPBA i 13mm diameter) 28.95 + 4.82 kN 10.5#2.13 Bolting connection ( class 88 . PPPBS nd 1Omm diamete) 29.65 = 3.38 kN 10.0 + 1.69 poppe Bonding connection ( epoxy 3 53 5 0.67 MPa Unquantified * bi-component adhesive) “the bonded specimens exhibited very slight displacements thai could nat be measured accurately 2.3 Beams preparation Both hybrid beams PB7 and PB8 had a total length of 5 meters spanning 4.8 m. The details of the proposed beams are represented in figures | and 2-A. Moreover, the conerete slabs were reinforced with ST25 bars in both directions for shrinkage. At the extremities, concrete blocks were casted on both sides of the web to rigidify the support area and prevent any local buckling or crushing of the fibers. The to-be- bonded zones of the concrete slabs were sandblasted to obtain a rough surface. The conerete slabs as well as the GFRP profiles were dusted-off using compressed air jet. The GFRP profiles were then cleaned by acetone to remove any stuck particles. The adhesive was applied on the top flange of the GFRP profiles as well as the sandblasted part of the concrete slabs. The two materials were then pressed against each other using clamps to obtain an approximate bonding thickness of 1 mm. Then bolts (6.8 12 mm in PB7 and 8.8 10 mm in PB8) were introduced and manually tightened in pre-drilled holes in reservations already positioned in the profile’s top flange. 2LITH INTERNATIONAL, ‘ONFERENCE ON COMPOSITE, SCIENCE AND TECHNOL: Cxindrcal passage (passive reinforcements omitted) forthe clamps RC gab Applied co | force Tg Conse Sens sunne tome Some tom Fig.1 Longitudinal view of PB7 and PBS Bolts 4 Passive steel Epory reinforcement adhesive joint (about Imm) +b 10mm (ay @) Fig.2 (A) Beam cross section of PB7 and PBS (B) Failure of a bolted push-out specimen 2.4 Bending tests Displacement LVDTs were used to measure deflections at mid-spans and slip at the interface. ‘The mid span section (SC) was instrumented by deformation gages on the concrete top and bottom surfaces, on a longitudinal steel bar within the concrete, and in GFRP top, middle and bottom fibers. A bracing system was applied on both extremities to prevent any lateral-torsional buckling, 3 point bending tests were conducted up to failure. The press used is a hydraulic jack with a maximum applied load of 500 IN, Fig3 shows the force-deflection curves at mid-span throughout the test. We realize that the curve is linear up to a load of 50 KN followed by a reduction of the stiffness where the load continues to increase. The interface northern side of the beam was sheared at a load slightly less than 80 KN, this was followed by shearing of the southern interface at 80.69 KN which was followed directly by the crushing of the concrete and the delamination of the bottom fibers of the GFRP. Deformation gages indicated -0.31 % and 10.7 % in top concrete and bottom GERP respectively. We notice that the deflection curve of PBS at SC confounds with that of PB7. PBS represents almost the same bi-linear behavior as PB7. Northern interface shear failure was followed by a shear aExperimental and Analytical Study on the Bending Behavior of Concrete-GFRP Hybrid Beams: Koaik, A. et al failure of the junction web-flange at a load of 81.2 KN. Deformation gages indicated -0.32 % and 9.8 % in top concrete and bottom GFRP respectively. Both beams presented the same behavior over almost 90 % of loading. The primary failure for both beams is interface shearing and the stight difference between the failure loads is insignificant (0.7 %). It can be deduced thatthe type of bolts had no effect over the behavior of the beams. ho 100 1. Northern 2 Southern —_—-inrfce fare PB?) snerfoce alr PB) : Failure of PBT by she Interface at Fan, Failure pf PBS by shearing of the profile at Fg, = 81.2 KN se detection a7] Eso 2 Fair of he z 40 Ho seeflange (PB) z E Concrete 20 ‘tuninanon of boom GPRP 0 way © 10 2% 3% 40 30 6 7% 80 9% 100 M0 120 130 Deflection (sm) Fig.3 Force-deflection diagrams for PB7 and PBS 3. Analytical approach The absence of design codes concerning the FRP materials already constitutes one of the factors, that limit their wide application. However, the elastic behavior of the FRP materials can be compared to that of steel in service limits states. A simple method based on the compatibility of strains, plane sections remaining plane, and equilibrium of the section is exploited. This method has been already used in [1] at ultimate states and gave acceptable results Having relatively slight slip values before rupture, it was decided to use a non-slip analysis to characterize the connection, In the elastic phase, the equilibrium of the section leads to the determination of the curvature and the deformations, The deflection is obtained directly from the double integration of the calculated curvature. However, the participation of the shear displacements in the case of pultruded GFRP profiles is significant. In consequence, the shear deflection term is added: oFE 8 KFL aGAg ay In this equation, Fis the applied force at mid-span, L is the span length, K is the geometric shear PaLITH INTERNATIONAL, ‘ONFERENCE ON COMPOSITE, SCIENCE AND TECHNOL: factor (equals 2 in I section), 4, the area of the web and 6; is a parameter depending on the geometric and ‘mechanical properties of the section. ‘The beams sections were designed to have the neutral axis in the concrete slab and the failure by concrete crushing. This failure is the most ductile and safe among the other possible failure criteria, At ultimate theoretical state, for an ultimate deformation of concrete in compression of eq = 0.35 %, the deformations at all section layers are obtained and finally the ultimate moment is calculated. The position of the neutral axis in such a case is obtained by the equation: O.BbafaZ2 + ea pAyly ~ eakpAp(ha + 0.5hp) = 0 ® The ultimate moment can be expressed by: My = Nae + Npadyn + Npatya + Npadys 6) ‘The different parameters are shown in Fig. 4. Moreover, the ultimate stress at the interface is calculated Fig. 4 Stress-strain diagrams in the ultimate phase ‘The numerical values of PB7 and PBS (tables 2 and 3) are similar due to the hypothesis of absence of slippage. It is noticed that the mid span deflection theoretical values as well as the deformations of the top fibers are coherent with the experimental ones despite very slight variations attributed mainly to the sensitivity of the GFRP moduli as well as the non-linearity of concrete elements. It is noticed that the ultimate stress at the interface is less than the one obtained from the push-out series (POPBS). This can be contributed to the peeling stresses at the extremities as explained by [1]. This failure caused a semi 2Experimental and Analytical Study on the Bending Behavior of Concrete-GFRP Hybrid Beams: Koaik, A. etal separation between the two materials which increased the deformations in their extreme fibers. In consequence, we obtain deformation values which are close (o the ultimate theoretical ones however the failure load is about half the theoretical one. ‘Table 2 Comparison between the analytical and experimental results at mid-span section under 40 KN. PBT PBS ‘Mid span deflection* (mim) 38.944 BEIT | 38.944 (37.209) Deformation in top concrete® (Tos) | 0.673 (0.818) | -0.673 (0818) Deformation in bottom GFRP *(oo) | __ 3-4 (3.326) 343.16) ‘Table 3 Comparison between the analytical and experimental results at mid-span section at failure PBT PRE Failure Toad * (KN) 169.2(80.7) 169.2(81.2) Deformation in top concrete™ 0.0035(0.0031) 0,0035(0.0032) Deformation in bottom GFRP* 0.0162(0.0107) 0.0162(0.0098) Shear stress at the interface (MPa) 42 42 Failure mode Conerete crushing (Interface- | Concrete erushing (Interface- Delamination of fibers) Shearing of the web) Experimental values are within parentheses 3. Conclusion ‘The tests allowed us to conclude that the variation of the type of bolts had no effect on the overall behavior of the beam because the adhesive controls the failure of the interface. Besides, the proposed analytical method offers an acceptable analysis in the elastic domain, Moreover, the prediction of the shear failure load should be improved. Reference 1, Correia JR, Branco PA, Ferreira JG : Flexural behaviour of GFRP-conerete hybrid beams with interconnection slip. Composite Structures; 77:6678 (2007). 2. Correia JR, Branco FA, Ferreira J. GFRP-concrete hybrid cross-sections for floors of buildings. Engineering Structures; 31:1331-43(2009), 3. El-Hacha R, Chen D. Behaviour of hybrid FRP-UHPC beams subjected to static flexural loading. Composites Part B: Engineering;43:582-93(2012). 4, Koaik A, Bel S, Jurkiewier B. MechComp. Shear connections between GFRP pultruded profiles and concrete: a comparison between bolting and bonding, Munich: 2016. Ey