About this Abstract |
| Meeting |
2027 TMS Annual Meeting & Exhibition
|
| Symposium
|
Modeling, AI Applications, and Method Development in Nuclear Materials
|
| Presentation Title |
Development of Machine-Learned Interatomic Potentials for Tritium Permeation Barrier Materials as Fusion Structural Materials |
| Author(s) |
Anne Davis, Mary Alice Cusentino |
| On-Site Speaker (Planned) |
Anne Davis |
| Abstract Scope |
Modelling materials in fusion reactors is a challenge due to the extreme conditions present, including high neutron irradiation flux and high temperature. Molecular dynamics (MD) simulations of primary knock-on atom (PKA) events are a useful method for modelling irradiation damage in materials, allowing defect densities and types to be quantified. To predict microstructures that arise from PKA events with MD simulations, accurate interatomic potentials that reproduce defect formation energies must be used. Machine-learned interatomic potentials (MLIAPs) contain the accuracy of density functional theory methods without sacrificing the speed and size capabilities of classical potentials. In this work, we present the development of atomic cluster expansion MLIAPs for two tritium permeation barrier materials, Al2O3 and SiC. We run PKA simulations in these materials to evaluate radiation damage and quantify the microstructures that form. SNL is managed and operated by NTESS under DOE NNSA contract DE-NA0003525. |
| Proceedings Inclusion? |
Planned: |
| Keywords |
Nuclear Materials, Computational Materials Science & Engineering, Machine Learning |