[19p-Z32-1~15] 23.1 Joint Session N "Informatics"

[19p-Z32-2] A neural network molecular dynamics study of the thermal properties of CeO₂

〇Nobuhiko Kato¹, Kazuki Mori¹, Masahiro Saito¹, Kenji Konashi² (1.CTC., 2.IMR Tohoku Univ.)

Keywords:Neural network MD, Cerium oxide, Thermophysical Properties

CeO2 is used as a mock test material for nuclear fuel bodies, but the understanding of thermophysical properties at high temperatures is insufficient. First-principles calculations and classical MD calculations are effective as methods for investigating the details of material properties at high temperatures, but first-principles calculations are costly and MD calculations are inadequate in the accuracy of interatomic potentials. Therefore, in this study, we evaluated the thermophysical properties of CeO2 using a neural network MD (NNMD), which is faster than first-principles calculation and more accurate than existing MD calculation.