Machine learning approaches for elastic localization linkages in high-contrast composite materials

Problem statement Localization, as opposed to homogenization, describes the spatial distribution of the response at the microscale for an imposed loading condition (e.g., averaged strain) at macroscale. Localization is critically important in correlating various failure-related macroscale properties of the material with the specific local microstructure conformations responsible for the (local) damage initiation in the material. In this work, these two scales are to be connected through linkages extracted by data-driven processes used in machine learning systems. More specifically, we focus our effort in this study on extracting localization relationship for elastic deformation in a two-phase composite [15], [16], [18], [19]. The input into such a linkage typically includes the material microstructure (defined in a three-dimensional (3-D) microscale volume element (MVE)) and the applied macroscale loading condition (typically expressed as the…


Link to Full Article: Machine learning approaches for elastic localization linkages in high-contrast composite materials