[2G14-21] AI Technology

[2G19] Integration of AI Technology and Thermal Hydraulics for the Development of a Data-Driven Methodology for Plant Safety Assessment

(4) Bubble entrainment analysis by water jet using JUPITER code and estimation of penetration length by machine learning

*Susumu Yamashita¹, Zhou Wen², Shuichiro Miwa² (1. JAEA, 2. UTokyo)

Keywords:Computational fluid dynamics, Gas-liquid two-phase flow, Water jet penetration analysis, JUPITER

The development of data-driven plant safety assessment methods by integrating artificial intelligence (AI) technology and thermal hydraulics has been attracting attention as a way to reduce the time cost, range of applicability to actual phenomena, and accuracy variability that are problems in developing physical models in the field of thermal hydraulics. To contribute to developing a data-driven analysis method that automatically interprets the results of CFD calculations using AI technology, training images for building an AI model were generated by analyzing water jet into a pool system using JUPITER. The analysis was conducted using the several parameters (nozzle height, jet velocity, etc). In this presentation, we will present the calculation results for building the AI model, a quantitative comparison of the penetration length with the experimental results, problems in simulating the water impingement phenomenon, and a prediction of the penetration length by machine learning.