Effect of concrete composition on FRP/concrete bond capacity

In recent years, external bonding of fibre reinforced plastics (FRP) to concrete members becomes more and more popular in strengthening and repairing of old concrete structures. For the strengthened member, failure may occur through FRP debonding that initiates from the bottom of a major flexural crack in the concrete member, and propagates towards the free end of the plate. This failure mechanism is often studied with the direct shear test. Various analytical models have been established based on fracture mechanics and empirical study. In existing models, the bond capacity is considered to be strongly dependent on the compressive (or tensile strength) of the concrete itself. However, theoretical analysis indicates that the shear stress distribution along the FRP/concrete interface at ultimate debonding failure is usually dominated by the frictional part, where cracking has already occurred in the concrete, and aggregate interlocking is leading to residual stress. We therefore suspect that other factors affects the interlocking effect, such as size and content of aggregates, will also play important roles in determining the bond capacity. To investigate the effect of aggregates on the bond capacity, twelve different compositions of concrete have been used to prepare specimens for the direct shear test. The test results indicate that the bond capacity has little correlation with either the concrete compressive or tensile strength. On the other hand, the bond capacity is found to be related to concrete surface tensile strength, and the aggregate content. Based on the results, an empirical expression using neural network is derived to calculate the interfacial fracture energy in the shear test. The bond capacity can then be calculated according fracture mechanics based model. Good agreement has been obtained between the simulation and experimental results.