| About this Abstract |
| Meeting |
2025 TMS Annual Meeting & Exhibition
|
| Symposium
|
Chemistry and Physics of Interfaces
|
| Presentation Title |
A Machine Learning Approach for Extracting Grain Boundary Mobilities From Time-Resolved Grain Maps |
| Author(s) |
Jules M. Dake, Leonard Lauber, Thomas Wilhelm, Lukas Petrich, Orkun Furat, Volker Schmidt, Carl E Krill |
| On-Site Speaker (Planned) |
Jules M. Dake |
| Abstract Scope |
When polycrystalline materials are heated to high homologous temperatures, typically about 0.5 and beyond, atoms of the individual crystallites have enough thermal energy to diffuse from one grain to another, and the interfaces between crystallites—commonly referred to as grain boundaries—begin to migrate in a manner that reduces the total grain boundary area and increases the mean grain size. This process is known as grain growth, and it has far-reaching consequences for materials properties. Accordingly, much effort has been invested in understanding the mechanisms underlying grain boundary migration. In the simple case of a pure metal, the mobility of a given grain boundary is assumed to depend on its misorientation and inclination; however, determining the exact functional dependence has proved to be elusive. Here, we present a machine learning approach for extracting grain boundary mobilities directly from time-resolved grain mappings with the help of a convolutional neural network. |
| Proceedings Inclusion? |
Planned: |
| Keywords |
Machine Learning, Characterization, Aluminum |