[1A13-16] Knowledge Management System

5:00 PM - 5:15 PM JST(8:00 AM - 8:15 AM UTC)

[1A16] TASKI:AI-based nuclear knowledge management system(4) Model Development for AI-Driven Nuclear Fuel Cycle simulator

*Koki Ono¹, Tomohiro Okamura^1,2, Takumi Abe³, Takahiro Nishihara^1,2, Masahiko Nakase^1,2, Kenji Nishihara³, Taiga Suzuki¹ (1. Institute of Science Tokyo, 2. NEUChain inc, 3. JAEA)

Keywords:

Nuclear knowledge management,Material Balance Analysis,NMB4.0,AI,Nuclear Power Utilization Scenario,Verification & Validation,TASKI

Nuclear fuel cycle simulation codes are used as decision-support tools to assess future scenarios of nuclear power generation. The Institute of Tokyo Science and the Japan Atomic Energy Agency are jointly developing a next-generation multivariable evaluation code by incorporating AI inference capabilities into NMB4, aiming to support future decision-making. However, conventional evaluations have not sufficiently incorporated actual reactor operation data or fuel histories, and improving the reproducibility and reliability of these models remains a significant challenge.In this study, we build upon previous verification efforts by Nishihara (2023) and Suzuki (2024) to develop an evaluation model that reflects the operational history of existing nuclear reactors in Japan. We present the results of this verification and discuss the potential applications of the model.