Enabling High-fidelity Twin Pattern Prediction in Polycrystals — A Mesoscale Grain Boundary Plasticity Model.

Abstract In this paper we propose a parameter-free grain boundary (GB) model for the meso-scale, employing it to effectively capture one-to-one twin pattern development across polycrystals (for the first time), by means of crystal plasticity finite element methods (CPFEM). Our model does not require higher-order mechanics, involve limiting length-scales, or demand special considerations for its FEM implementation. It derives all its data from that of its adjacent grains, which are modeled using single crystal plasticity. We apply our mesoscale GB model to experimentally imaged microstructures of Magnesium and Titanium. We demonstrate a clear ability to capture the prominent role of GBs in the nucleation and patterning of twins, with high spatiotemporal fidelity (approaching one-to-one), as compared to electron back-scattered diffraction (EBSD) maps.