UCLA

The high temperature gradients present in additive manufacturing (AM) processes tend to produce microstructures in metallic materials that are not seen in other manufacturing processes. Large columnar grains oriented parallel to the build direction in AM materials are an important feature to quantify based on their ability to affect the mechanical properties of the output material. In this work we present a thermo-physical phase field model that combines an implicit solution to the heat equation in two dimensions while simultaneously modeling the explicit heterogeneous nucleation and growth of solidifying material in the active build layer. After construction of a simulated bulk material, finite-element analyses can be performed to determine the expected mechanical properties. Given the varied parameter space present in AM processes, an efficient method is presented to utilize machine learning methods to predict the effect of various microstructural features on the output mechanical properties of yield strength and post-yield hardening rate.