[P] P130~P143 Energy and Related Materials

[P130] Machine-Learning Potentials for Hydrogen–Vacancy Complexes in Tungsten

○Yuuki Noguchi¹, Daiji Kato^2,1, Koichi Sato³ (1. Kyushu Univ., 2. NIFS, 3. Kagoshima Univ.)

Keywords:

Fusion reactor material,Irradiation defects,Tungsten,Machine-learning

This study investigates hydrogen–defect interactions in tungsten, a plasma-facing material for fusion reactor divertors. We constructed an on-the-fly machine-learning potential trained on DFT data and demonstrated that it accurately reproduces DFT results.