Could cellular automata be used to model mechanisms (for quasi-brittle fracture) that occur at the meso-scale and then feed these mechanisms to a macro-scale finite element model? Is it possible to replace constitutive models with mechanistic models, simulating mechanisms that lead to fracture instead of formulating equations that predict failure? These are typical questions that have motivated my recent collaboration with Dr Anton Shterenlikht at the University of Bristol.
We have published a paper this month in the Royal Society that reports on our recent efforts to couple a meso-scale cellular automata code with open source software for massively parallel finite element analysis.