A stochastic model for crack propagation in random heterogeneous media uses mesoscopic representations of elastic and fracture properties. Monte-Carlo simulations with a phase-field approach assess subscale randomness effects. Mesoscopic descriptors and random elasticity tensors are introduced for scale transitions. Fracture toughness is identified via an inverse problem. A non-Gaussian stochastic model predicts the macroscopic response, accurately matching full-scale simulations.
Hun, D., & et al. (2018). Construction of simplified models for crack propagation in random heterogeneous media. WCCM, New York. https://hdl.handle.net/2078.5/238298