A multi-scale stochastic fracture modelling calibration using Monte Carlo simulations

The validation of macroscopic fracture modelling of heterogenous materials is strongly related to the effectiveness of the homogenization principle at fine scales. Often because the complete material morphology is missing, artificial repetitive boundary conditions are deployed despite many criticisms on their inadequate local subscale assumptions. This paper applies a recently developed multiscale stochastic fracture modelling approach (MsSFrM) to simulate crack growth in FRP materials. First, image-based models are solved at micro level and
mapped onto the macro scale using a two-scales adaptive window principle; and second, the macro scale is solved by using a calibrated heterogeneous cohesive interface crack model based on [6, 7]. Two further enhancements are proposed here, both being able of representing the macro continuum in a stochastic fashion. First, the micro scale is related to various degrees of aggregations (DAG) via Voronoi tessellation statistics; and second, Karhunen-Loeve (KL) material property expansions are used for areas where images cannot be
provided.