Abstract

This paper will describe modeling microstructure evolution using a combination of our Gradient-Weighted Moving Finite Elements code, Grain3D and our 3-D unstructured grid generation and optimization code, LaGriT. Grain boundaries evolve by mean curvature motion, and Grain3D allows for the incorporation of grain boundary orientation dependence modeled as anisotropic mobility and energy. We also describe the process of generating an initial computational grid from images obtained from electron backscatter diffraction. We present the grid optimization operations developed to respond to changes in the physical topology such as the collapse of grains and to maintain uniform computational grid quality. For 3-D columnar microstructures, validation of the method is demonstrated by comparison with experiments. For large systems of fully 3-D microstructures, simulations compare favorably to the parabolic law of normal grain growth.

Author Keywords: microstructure evolution, motion by mean curvature, gradient-weighted moving finite elements, unstructured tetrahedral meshes, deforming grids, changing grid topology

Corresponding author.