| About this Abstract |
| Meeting |
2024 TMS Annual Meeting & Exhibition
|
| Symposium
|
Computational Discovery and Design of Materials
|
| Presentation Title |
Machine Learning Accelerated Thermodynamic Search for Ductile Cr-based Alloys for High-Temperature Applications Complemented by Ab-Initio Simulations |
| Author(s) |
Lassi Linnala, Mikko Tahkola, Abhishek Biswas, Matti Lindroos, Napat Vajragupta, Thomas Blackburn, Kan Ma, Alexander Knowles, Tatu Pinomaa, Anssi Laukkanen |
| On-Site Speaker (Planned) |
Lassi Linnala |
| Abstract Scope |
Chromium-based alloys are considered for high-temperature applications, such as concentrated solar power operating at ~800°C. Despite chromium's advantages of low cost, high melting point, and excellent high-temperature oxidation resistance, practical use is hindered by its high ductile-to-brittle transition temperature. Here, we propose a synergistic approach of machine learning and first-principles density functional theory (DFT) simulations for developing ductile chromium-based alloys. To explore compositional space of 11 elements: Cr, Ni, Al, Fe, Mn, Si, Ti, V, Co, Mo, and W, we first employ a machine learning-accelerated CALPHAD surrogate model. This enables a systematic search for alloy compositions satisfying multiple property constraints. While the surrogate model accelerates the search process and provides potential alloy compositions, its ability to predict mechanical properties, particularly ductility, is limited. To this end, we incorporate DFT simulations to refine the list of compositions, focusing only on those exhibiting ductile properties. These compositions can then be experimentally realized. |
| Proceedings Inclusion? |
Planned: |
| Keywords |
High-Temperature Materials, Modeling and Simulation, Machine Learning |