About this Abstract |
| Meeting |
2026 TMS Annual Meeting & Exhibition
|
| Symposium
|
Frontiers in Solidification X
|
| Presentation Title |
High-Throughput Multiscale Solidification Modeling of Refractory High Entropy Alloys |
| Author(s) |
Victoria Tucker, Gregory Wagner, Shardul Kamat, Michael S. Titus |
| On-Site Speaker (Planned) |
Michael S. Titus |
| Abstract Scope |
Understanding relationships between composition, thermodynamic properties, solidification, and resulting microstructure is of great importance to increase High Entropy Alloy (HEAs) adoption into practical applications. However, predicting these characteristics and especially resulting microstructure from solidification processing present significant challenges due to the large design space with disparate thermodynamic properties and lengthscales involved. To address this challenge, we developed a multi-physics simulation framework to train a data-driven surrogate model to rapidly assess how changes in HEA composition lead to changes in solidification profiles and ultimately microstructure. The physics-based framework sequentially couples high-throughput CALPHAD calculations, computational fluid dynamics (CFD), and Cellular Automata (CA) for microstructure. The data-driven surrogate model, which can reconstruct microstructure statistics from the thermal conditions, is trained via a reduced-dimensional form of the microstructure output, and we will present results of the physics-based framework, surrogate model, and comparison to experimental microstructures to assess efficacy of the models. |
| Proceedings Inclusion? |
Planned: |
| Keywords |
Solidification, Machine Learning, High-Entropy Alloys |