FIP-GNN: Graph neural networks for scalable prediction of grain-level fatigue indicator parameters

Abstract

High-cycle fatigue is a critical performance metric of structural alloys for many applications. The high cost, time, and labor involved in experimental fatigue testing call for efficient and accurate computer models of fatigue life. We present FIP-GNN -- a graph neural network for polycrystals that (i) predicts fatigue indicator parameters as grain-level inelastic responses to cyclic loading quantifying the local driving force for crack initiation and (ii) generalizes these predictions to large microstructure volume elements with grain populations well beyond those used in training. These advances can make significant contributions to statistically rigorous and computationally efficient modeling of high-cycle fatigue -- a long-standing challenge in the field.

References in corpus (3)

• Geometric deep learning for computational mechanics Part I: Anisotropic Hyperelasticity
• Graph Neural Networks for an Accurate and Interpretable Prediction of the Properties of Polycrystalline Materials
• Learning from 2D: machine learning of 3D effective properties of heterogeneous materials based on 2D microstructure sections

Cited by in corpus (3)

• Machine learning of microstructure--property relationships in materials leveraging microstructure representation from foundational vision transformers
• Transfer-learned Kolosov-Muskhelishvili Informed Neural Networks for Fracture Mechanics
• A Variational Kolosov--Muskhelishvili Network for Elasticity and Fracture