Enabling Predictive Scale-Bridging Simulations through Active Learning

Northwestern will develop, in collaboration with LANL, a three-dimensional model for crack branching in layered media. Northwestern will validate, in collaboration with LANL, the model with controlled experiments at LANL that can control material, fluid and boundary conditions for crack branching. Northwestern will provide LANL data from the three-dimensional model in task 1.1.3.1 to train machine learning models that will relate crack branching to material constitutive parameters.