Interface behavior of fiber reinforced polymer composites externally glued to quasi-brittle substrates

The use of externally glued fiber-reinforced polymers (FRP) as reinforcement to overcome the tensile deficiency of quasi-brittle elements (e.g. concrete beams, shear walls, masonry arches) has gained great popularity during the last years. Experimental and theoretical studies demonstrated that, when the FRP-substrate joint is mostly stressed in shear, one of the princiapal failure mechanisms is the debonding. It occurs when the shear capacity of the system is reached and a crack develops underneath the bond plane a few millimeters inside the substrate, causing the detachment of the composite element. In the present work the interface behavior of FRP joints is studied by means of experimental and numerical studies. A new single-lap test setup is proposed allowing to stably follow, for the first time, the entire equilibrium path of this kind of reinforcement. The proposed setup is then adopted to study the bond performances of FRPs applied on both concrete and masonry. The test results highlight a dependence of the global behavior from the initial bonded length and suggest the presence of non-negligible stresses orthogonal to the bonding plane. Also, comparing the results on concrete and on masonry, it is shown how, for this latter kind of substrates, the behavior is strongly influenced by the material texture and composition. To reproduce the changes in behavior observed during the experimental campaign, a novel cohesive zone model that accounts for the presence and the coupling between normal and tangential stresses is proposed and validated. Furthermore, the problem of the fatigue failure for this joints is studied and a new thermodynamically consistent numerical model that couples damage and plasticity under pure shear conditions is formulated. The numerical simulations coming from the two proposed models are compared to experimental results coming from the performed tests as well as from the available literature. Moreover, the improvements with respect to the models to date available are highlighted. Finally, taking advantage of new experimental studies and starting from theoretical considerations, a modified practical design formula for the debonding capacity for FRP reinforcements applied on masonry substrates is proposed and calibrated over a large database of results collected form the literature.