A Deep Learning Prediction Method for Growth of Micro Voids in Single-Crystal Metal

Spallation of metal under extreme dynamic loading is a multiscale problem spanning from microscopic to mesoscopic and macroscopic scales in short time. The damage mechanism is very complex, in which the void evolution is a critical process. Deep-learning methods have provided new possibilities for rapid and accurate prediction of void evolution. In order to predict the growth of micro voids in single-crystal metal, a deep neural network model based on U-Net and Transformer is constructed. The dataset is constructed by molecular dynamics simulation results of a single-crystal copper atomic model with initial double ellipsoidal voids. A data preprocessing scheme based on background meshes is proposed to perform local statistics on the simulation results. Numerical examples demonstrate that the aforementioned deep-learning method can accurately predict the global physical quantities and local details during growth of micro voids in single-crystal metal.