ML training data for 'Machine Learning Based Approach to Predict Ductile Damage Model Parameters for Polycrystalline Metals'

This dataset contains training data for the machine learning model described in ‘Machine Learning Based Approach to Predict Ductile Damage Model Parameters for Polycrystalline Metals’ by the same authors. The included data were generated using FLAG simulations of a flyer plate impact scenario with the ductile damage model TEPLA for three materials: half-hard copper, annealed copper, and aluminum 6061. For each of these materials, we include the postprocessed results for the free surface velocity as a function of time as well as porosity as a function of position in the target material, and finally a derived ‘reduced’ dataset containing only characteristic features of the former; details are described in the above mentioned paper, see https://doi.org/10.1016/j.commatsci.2023.112382 or https://arxiv.org/abs/2301.07790 The code to reproduce the postprocessed data from the included raw data as well as to train the ML model is published here: https://github.com/dblaschke-LANL/LEAD/