Determination of concrete fracture parameters using inverse analysis: Influence of the tensile softening model

Abstract

The paper is focused on the identification of selected mechanical fracture parameters of concrete. An inverse analysis based on an artificial neural network is used for this purpose. In this approach the laboratory measurements are matched with the results gained by reproducing the same test numerically. The identification of mechanical fracture parameters is carried out from the records of three-point bending and wedge-splitting tests performed using three specimen sizes. The ATENA software is employed for the numerical simulation of the fracture tests. The material model with the exponential and bilinear tensile softening law is selected to govern the gradual evolution of localized damage. The obtained parameters are finally analyzed and discussed in terms of their dependence on the size of the initial uncracked ligament. The results show that both tensile softening models are able to capture the behavior of the specimens in the softening phase reasonably well. The tensile softening model does not affect the modulus of elasticity values but has a slight effect on the tensile strength and fracture energy. For the latter two parameters, both models detected the influence of specimen size on their values.