About this Abstract |
Meeting |
2025 TMS Annual Meeting & Exhibition
|
Symposium
|
Algorithms Development in Materials Science and Engineering
|
Presentation Title |
Recognizing and characterizing continuous regions of materials design spaces through stochastic microstructure representations |
Author(s) |
Simon Mason, Megna Shah, Jeff Simmons, Dennis Dimiduk, Stephen Niezgoda |
On-Site Speaker (Planned) |
Simon Mason |
Abstract Scope |
Quantitative representations of microstructure can enable the accelerated and automated development and discovery of new materials through the characterization of entire processing and design spaces. While reduced-order models of microstructure have proven useful in describing material states and predicting properties, additional functionality in exploring the processing-structure-property relationship can be achieved when expanding microstructure descriptors from self-contained to globally-informative (in terms of design space).
By developing a method to construct a continuous manifold of microstructure states from quantitative descriptors, a bi-directional relationship between processing and microstructure can be learned. From there, continuous regions of the processing domain can be recognized and exploited as highly-modifiable material systems. Informed navigation of the processing domain can be used for discovering new materials within these continuous regions, as well as help define safe boundaries for potential processing conditions. |
Proceedings Inclusion? |
Planned: |
Keywords |
Computational Materials Science & Engineering, Machine Learning, Characterization |