Prediction of the evolution of the stress field of polycrystals undergoing elastic-plastic deformation with a hybrid neural network model

TL;DR

This paper introduces a hybrid neural network model that accurately predicts the evolution of stress fields in polycrystals during elastic-plastic deformation, reducing computational costs and enabling material characterization.

Contribution

The authors develop a convolutional neural network that predicts detailed stress fields from microstructure and loading, improving upon previous average-based models.

Findings

High-fidelity stress field predictions matching crystal plasticity data

Accurate stress distribution modeling during elastic to plastic transition

Effective application in material characterization and optimization

Abstract

Crystal plasticity theory is often employed to predict the mesoscopic states of polycrystalline metals, and is well-known to be costly to simulate. Using a neural network with convolutional layers encoding correlations in time and space, we were able to predict the evolution of the stress field given only the initial microstructure and external loading. In comparison to our recent work we were able to predict not only the spatial average of the stress response but the field itself. We show that the stress fields and their rates are in high fidelity with the crystal plasticity data and have no visible artifacts. Furthermore the distribution stress throughout the elastic to fully plastic transition match the truth provided by held out crystal plasticity data. Lastly we demonstrate the efficacy of the trained model in material characterization and optimization tasks.