Upscaling the Mechanical Properties of Additively Manufactured Metals
摘要
A method is proposed for using peridynamics to obtain a larger-scale characterization of a heterogeneous microstructure from a detailed small-scale model. The method produces a set of position-dependent material properties suitable for modeling the response of a structure with a coarser discretization than the original small-scale model. The elastic-plastic properties and anisotropy are included. A machine learning technique is proposed to accelerate the method by bypassing the small-scale simulations.