A probabilistic mean-field and microstructure based finite element modeling for predicting mechanical and ductile fracture behavior of the cast aluminum alloy

Abstract

A multiscale finite element simulation approach based on the probabilistic mean-field method is proposed to predict the strength and ductile fracture of the cast aluminum (Al) alloy A365-T6. The key microstructural characteristic of the cast alloy is the silicon (Si) particles embedded in the eutectic region of the Al matrix, which represents the inhomogeneous plastic deformation and large fracture strain scatter. Therefore, the theoretical homogenization using the Mori-Tanaka mean-field scheme and the determination of Si particle cracking based on the Weibull distribution function were formulated and implemented in the finite element model with the Gurson-Tver-gaard-Needleman ductile fracture law. The initial state of voids and Si particles was statistically applied to the finite element model based on experimental measurements. The underlying mechanism of fracture initiation from Si particle cracking and propagation was experimentally analyzed by scanning electron microscopy; thus, the model assumptions can be rationalized. Various tensile specimens, such as simple tension, center hole, notch tension, and in-plane shear, were utilized to experimentally investigate the effect of the stress state on the fracture initiation and validate the simulation results. The effects of macrovoids on the volume fraction and critical location were also investigated, along with those of pre-existing microvoids in the matrix and newly generated voids from Si particle cracking. In conclusion, the proposed microstructure based multiscale simulation approach predicted the strength, fracture strain, and fracture propagation path under different stress states with reasonable accuracy. Detailed analyses of the microscopic mechanism of ductile fracture in the cast Al alloy in relation to the strain, stress triaxiality, and damage evolution of each phase are also presented.