Structure-Property Linkages for Polycrystalline Materials using Materials Knowledge Systems

Abstract

Computational tools that are capable of rapidly exploring candidate microstructures and their associated properties are required to accelerate the rate of development and deployment of novel materials. In this work, a suite of computationally efficient protocols, based on the materials knowledge system (MKS) framework, are developed to evaluate the properties and performance of polycrystalline microstructures. In the MKS approach, physics-capturing coefficients (calibrated with microstructures and their responses obtained via experiments or simulations) store the microstructure-sensitive response of the material system of interest. Once calibrated, the linkages may be employed to predict the local responses (through localization) or effective properties (through homogenization) of new microstructures at low computational expense. Specifically, protocols are developed to predict bulk properties (elastic stiffness and yield strength), local cyclic plastic strains and resistance to fatigue crack formation and early growth (in the high cycle fatigue and transition fatigue regimes). These protocols are demonstrated on a diverse set of α-titanium microstructures, which exhibit heterogeneous microstructure features, in addition to anisotropy on multiple length-scales.