[3C10-15] Dose Estimation

4:00 PM - 4:15 PM JST(7:00 AM - 7:15 AM UTC)

[3C15] Estimation of radiation source distribution in fuel debris canisters using machine learning

*Taiyo Sato¹, Yuka Kumada¹, Tsugiko Takase¹, Masaharu Matsumoto¹, Katuhiko Yamaguchi¹ (1. Fukushima University)

Keywords:

Machine learning,Radiation,Gamma ray energy spectrum,Monte Carlo simulation,Fuel debris

In this study, a three-dimensional model was devised assuming a fuel debris storage canister, with cesium-137 and cobalt-60 arranged. In this model, the state in which radioactive sources are randomly arranged was calculated using Monte Carlo simulation calculations, and this was used as learning data for machine learning. An experimental system was then constructed for this model, and the gamma-ray energy spectrum was measured. By using this as validation data for machine learning, the accuracy of the estimation of the radiation source distribution using the spectrum obtained in real space was verified. In addition, by adding a material to the model that would shield gamma rays, it became possible to make estimates in cases with higher reproducibility.