LEGOMAT: Locally Extracted Globally Organized Microstructure Models using Markov Random Fields
Abstract
The problem of microstructure reconstruction is deeply intertwined with the theories of microstructure quantification. Currently available methods for microstructure synthesis such as geometry based (Voronoi models), physically based (Phase field models) or feature-based (Simulated annealing) methods run into various difficulties when modeling real complexities of microstructures that include non-equilibrium features, non-convex grains, twins, second phases and cell structures that arise from thermomechanical processing. These features play an important role in the properties and performance of modern aerospace alloys. Further, microstructures are stochastic and lead to location-specific variability in material properties. In this proposal, we delve into a mathematical model that is expected to provide a better alternative for microstructure synthesis: Markov random fields. In the report, theory and software for building 3D microstructural maps of engineering components through inference from 2D measurements is proposed.
Document Details
- Document Type
- Technical Report
- Publication Date
- Jun 27, 2021
- Accession Number
- AD1144431
Entities
People
- Veera Sundararaghavan
Organizations
- Board of Regents of the University of Michigan