| About this Abstract |
| Meeting |
2022 TMS Annual Meeting & Exhibition
|
| Symposium
|
Advanced Characterization and Modeling of Nuclear Fuels: Microstructure, Thermo-physical Properties
|
| Presentation Title |
Propose Advanced Nuclear Fuels with High Thermal Conductivity Using Machine Learning |
| Author(s) |
Meigyoku Kin, Masaya Kumagai, Yuji Ohishi, Eriko Sato, Masako Aoki, Ken Kurosaki |
| On-Site Speaker (Planned) |
Meigyoku Kin |
| Abstract Scope |
In recent years, Materials Informatics (MI), which is a fusion of materials science and information science, has been attracting attention and is becoming mainstream in the field of materials research for magnetic materials and thermoelectric materials. This research aims to propose a new nuclear fuel as an approach in the nuclear field where MI research has not been reported.
We designed a machine learning model using Starrydata, a database of experimental physical properties originally developed by our research group. The prediction accuracy was compared and verified using data without uranium compounds and data with uranium compounds. The trained machine learning models were used to comprehensively predict the thermal conductivity of uranium compounds that may exist in the world. We proposed uranium compounds with high thermal conductivity based on the prediction results, and evaluated the reliability of the model by synthesizing and measuring the proposed materials. |
| Proceedings Inclusion? |
Planned: |
| Keywords |
Machine Learning, Nuclear Materials, Other |