Computational determination of interface fracture properties for concrete by inverse analysis

The interfacial transition zone (ITZ) between cement mortar and aggregate significantly affects the cracking behavior of concrete. However, the fracture properties including the tensile strength and fracture energy of ITZ are hard to obtain directly from experiments. This paper develops an inverse numerical method for determining the fracture properties of ITZ based on a meso-scale fracture model and artificial neural network. Concrete is considered a multi-phase material, mainly consisting of aggregates, cement mortar, and ITZ. In the fracture model, cohesive elements are inserted in the mesh to achieve arbitrary cracking. The tensile strength and fracture energy of ITZ are the targeted variables to be inversely attained. A neural network is created and trained based on the simulated results, by which the optimized values of the targeted variables are obtained. Experimental results from RILEM tests are used to validate the numerical method.