A Physics-Informed Neural Network-Enhanced Material Point Method for Regression Simulation of Solid Propellant

Supplementary materials for the manuscript include a one-dimensional neural network model for predicting combustion rate and AP particle size distributions for four Miller packs. The particle size distribution is composed of the x, y coordinates and radius r of the particles.

本手稿的补充材料包括一个用于预测四种Miller包装燃烧速率及AP粒子尺寸分布的一维神经网络模型。粒子尺寸分布由粒子的x、y坐标以及半径r组成。