A multi-scale approach for microstructure prediction in thermo-mechanical processing of metals

The microstructure prediction methodology for thermo-mechanical processing of metals predominantly used in industrial practice exhibits strong empirical characteristics with corresponding drawbacks. This gives rise to a systematic multi-scale simulation methodology for predicting the microstructure evolution, in which the empirical microstructural relationships are replaced by numerically efficient and accurate physics-based models at the mesoscopic length scale. Currently, this approach is being implemented into a microstructure modeling system, which combines mesoscopic microstructural algorithms with continuum-based macroscopic FEM formulations through a multi-scale modeling interface. This approach is expected to overcome the shortcomings of current microstructure prediction practice employing empirical microstructural relationships, and to provide universal and accurate prediction capabilities.