Publications Details
Mesoscopic simulations of recrystallization
The application of computer simulation to grain growth and recrystallization was strongly stimulated in the early 80s by the realization that Monte Carlo models could be applied to problems of grain structure evolution. By extension of the Ising model for domain modeling of magnetic domains to the Potts model (with generalized spin numbers) it was then possible to represent discretely grains (domains) by regions of similarly oriented sets of material (lattice) points. In parallel with this fascinating development, there also occured notable work on analytical models, especially by Abbruzzese and Bunge, which has been particularly useful for understanding the variation of texture (crystallographic preferred orientation) during grain growth processes. Geometric models of recrystallization, worked on most recently and productively by Nes et al., have been useful in connection with grain size prediction as a result of recrystallization. Also, mesh-based models have been developed to a high degree by Kawasaki, Fradkov and others, and, rather recently, by Humphreys to model not just grain growth but also the nucleation process in recrystallization. These models have the strength that they deal with the essential features of grains, i.e. the nodes, but have some limitations when second phases must be considered. These various approaches to modeling of recrystallization processes will be reviewed, with a special emphasis on practical approaches to implementing the Potts model. This model has been remarkably successful in modeling such diverse phenomena as dynamic recrystallization, secondary recrystallization (abnormal grain growth), particle-inhibited recrystallization, and grain structure evolution in soldering and welding. In summary, the application of mesoscopic simulation to the phenomenon of recrystallization has yielded much new insight into some longstanding deficiencies in our understanding.