Learning to fail: Predicting fracture evolution in brittle material models using recurrent graph convolutional neural networks

TL;DR

This paper introduces a deep learning framework combining graph convolutional and recurrent neural networks to predict fracture evolution and failure in brittle materials efficiently, reducing computational costs significantly.

Contribution

The authors develop a novel neural network architecture with data augmentation for accurate fracture prediction, outperforming traditional simulation methods in speed and comparable in accuracy.

Findings

Predictions within 3% of simulated fracture damage and length.

Time to failure predicted within approximately 15%.

Predictions generated within seconds instead of hours.

Abstract

We propose a machine learning approach to address a key challenge in materials science: predicting how fractures propagate in brittle materials under stress, and how these materials ultimately fail. Our methods use deep learning and train on simulation data from high-fidelity models, emulating the results of these models while avoiding the overwhelming computational demands associated with running a statistically significant sample of simulations. We employ a graph convolutional network that recognizes features of the fracturing material and a recurrent neural network that models the evolution of these features, along with a novel form of data augmentation that compensates for the modest size of our training data. We simultaneously generate predictions for qualitatively distinct material properties. Results on fracture damage and length are within 3% of their simulated values, and results on time to material failure, which is notoriously difficult to predict even with high-fidelity models, are within approximately 15% of simulated values. Once trained, our neural networks generate predictions within seconds, rather than the hours needed to run a single simulation.