| Abstract Scope |
Microstructural evolution in metals is a multiscale process, with multiscale implications. Global behavior is informed by local phenomena, but the reverse is not always true. To tackle these codependent mechanisms, researchers often use simulation and experimental techniques that yield information at the appropriate local or bulk scale, however these streams of data are often analyzed and utilized in isolation, rendering an understanding of hierarchical microstructural evolution, and eventual control, elusive. In this talk, I will review recent efforts to build multimodal machine-learning assisted algorithms to parameterize microstructural evolution and metals processing, creating a living representation, or digital twin, of coincident global and local scales. Ultimately, these frameworks will be useful in future autonomous materials processing and development of models toward rapid screening of novel, complex alloy systems. |