Graph neural networks for emulating crack coalescence and propagation in brittle materials
share
High-fidelity fracture mechanics simulations of multiple microcracks interaction via physics based models quickly become computationally prohibitive as the number of microcracks increases. This work develops a Graph Neural Network (GNN) based framework to simulate fracture evolution in brittle materials due to multiple microcracks' interaction. Our framework achieves high prediction accuracy for multiple microcrack evolution (as compared to a physics based simulator) by engineering a sequence of GNN-based predictions. The first prediction stage determines Mode-I and Mode-II stress intensity factors, the second prediction layer determines which microcracks will propagate, and the final layer actually propagates crack-tip positions for the selected microcracks to the next time instant. The trained GNN framework is capable of emulating crack propagation, coalescence and corresponding stress distribution for a wide range of initial microcrack configurations without any additional modification. Lastly, the framework's time performance is independent of initial number of microcracks, thus, emulating higher-complexity problems without aggregating computational costs.
READ FULL TEXT