[1F12-15] Advanced Management Technology of Reactor Operation Management

[1F12] Development of evaluation method of LOCA break diameter using Gaussian process regression

*Akira Nakamura¹, Takayoshi Kusunoki¹ (1. INSS)

Keywords:machine learning, PWR, loss of coolant accident, numerical simulation

The method to predict LOCA break diameter automatically from the reactor coolant pressure has been developed in this research. This report describes the learning result using Gaussian process regression and the features extracted from the RCS pressure calculated by MAAP4 in order to improve prediction accuracy.