Graph neural networks to predict coercivity and maximum energy product of hard magnetic microstructures

Graph neural networks (GNN) are a promising tool to predict magnetic properties of large multi-grain structures, which can speed up the search for rare-earth free permanent magnets. In this paper, we use our magnetic simulation data to train a GNN to predict coercivity of hard magnetic microstructures. We evaluate the performance of the trained GNN and quantify its uncertainty. Subsequently, we reuse the GNN architecture for predicting the maximum energy product. Out-of-distribution predictions of coercivity are also performed, following feature engineering based on the observed dependence of coercivity on system size. • Data preprocessing and feature selection improve model performance and physics insight. • Graph neural networks accurately predict coercivity and energy product in magnets. • Customized convolutional layers capture hard magnetic grain interactions effectively. • Uncertainty quantification enhances reliability and helps in quantifying the error. • Model shows strong accuracy and generalization to larger, unseen microstructures.