Predicting the core thermal hydraulic parameters with a gated recurrent unit model based on the soft attention mechanism

• DOI: 10.1016/j.net.2024.01.045
• Publication Date: 2024-03-06T01:30:28Z
• Authors: Anni Zhang; Siqi Chun; Zhoukai Cheng; Pengcheng Zhao
• Keywords: Thermal hydraulics; Transient (computer programming); Mass flow rate; Mass flow

Accurately predicting the thermal hydraulic parameters of a transient reactor core under different working conditions is first step toward safety. Mass flow rate and temperature are important hydraulics, which have often been modeled as time series prediction problems. This study aims to achieve accurate continuous instantaneous conditions, well test feasibility newly constructed gated recurrent unit (GRU) model based on soft attention mechanism for parameter predictions. Herein, China Experimental Fast Reactor (CEFR) used research object, CEFR 1/2 was taken subject carry out predictive analysis conditions., while subchannel code named SUBCHANFLOW generate thermal-hydraulic parameters. The GRU predict mass core. results show that compared adaptive radial basis function neural network, network produces better results. average relative error less than 0.5 % when size 3, effect within 15 s. 5 10, in subsequent 12 not only shows higher accuracy, but also captures trends dynamic series, useful maintaining safety preventing nuclear power plant accidents. Furthermore, it can provide long-term predictions engineering applications improving