Experimental verification of 2- and 3-D numerical models for bond-slip behavior of CFRP-concrete

•Verification of two variants of FRP-concrete double shear bond test setups.•Despite the good performance of 2-D FE model, the 3-D model outperforms it.•Experimental study of bond length effect on the FRP-concrete connection.•Benchmarked literature database of bond-slip models used for validation.•Simulation verifies that improved shear test setup can replace the classical one. This paper studies the bond performance between fiber-reinforced polymer (FRP) and concrete using two variations of the double shear lap test setups. Experimental tests were conducted followed by a series of finite element simulations using cohesive zone model. The experiment utilized an improved double-pull shear setup with the tests conducted on specimens of varying FRP-concrete bond lengths. The numerical simulations were carried out for the improved as well as the traditional setups using two converged meshes with both 2- and 3-D spatial idealizations to study their relative performance. The results are discussed in light of the damage mechanism, load–displacement, strain and stress distribution, bond-slip behavior, and fracture energy. The tested specimens show no effect of bond length variation on the FRP-concrete interfacial strength so long as the bond length exceeds the effective bond length – Only the case of bond length lower than the effective bond length shows otherwise. The study also succeeds in validating the 2- and 3-D numerical simulations using the experiments in both the two cases of spatial idealizations and the two types of test setups in terms of strain and stress distributions, bond-slip curve, and fracture energy. However, while the 3-D models succeed in capturing the failure load accurately, the 2-D models fail to do so. Evidence of similar performance between the improved double shear test and traditional double shear test setups reaffirms the fact that the former could be used to replace the latter.