Graph Neural Networks for an Accurate and Interpretable Prediction of the Properties of Polycrystalline Materials

TL;DR

This paper introduces a graph neural network model that captures physical interactions among grains in polycrystalline materials, leading to accurate and interpretable property predictions, demonstrated on magnetostriction data.

Contribution

The paper develops a GNN model that incorporates grain interactions for predicting material properties, enhancing accuracy and interpretability over existing methods.

Findings

Achieved ~10% prediction error on magnetostriction of polycrystalline alloys.

Enabled feature importance quantification for individual grains.

Demonstrated robustness across diverse microstructures.

Abstract

Various machine learning models have been used to predict the properties of polycrystalline materials, but none of them directly consider the physical interactions among neighboring grains despite such microscopic interactions critically determining macroscopic material properties. Here, we develop a graph neural network (GNN) model for obtaining an embedding of polycrystalline microstructure which incorporates not only the physical features of individual grains but also their interactions. The embedding is then linked to the target property using a feed-forward neural network. Using the magnetostriction of polycrystalline Tb0.3Dy0.7Fe2 alloys as an example, we show that a single GNN model with fixed network architecture and hyperparameters allows for a low prediction error of ~10% over a group of remarkably different microstructures as well as quantifying the importance of each feature in each grain of a microstructure to its magnetostriction. Such microstructure-graph-based GNN model therefore enables an accurate and interpretable prediction of the properties of polycrystalline materials.