Prediction of crucial nuclear power plant parameters using long short‐term memory neural networks. (13th April 2022)

• Record Type: Journal Article
• Authors: Lei, Jichong; Ren, Changan; Li, Wei; Fu, Liming; Li, Zhicai; Ni, Zining; Li, Yukun; Liu, Chengwei; Zhang, Huajian; Chen, Zhenping; Yu, Tao
• Is Part Of: International journal of energy research. Volume 46:Number 15(2022)
• Page Start: 21467
• Page End: 21479
• Publication Date: 2022-04-13
• Subjects: LSTM -- nuclear power plant -- PACTRAN -- parameter estimation; Power resources -- Periodicals; Power (Mechanics) -- Periodicals; Power resources -- Research -- Periodicals; 621.042
• DOI: 10.1002/er.7873
• Languages: English
• ISSNs: 0363-907X

Abstract

Summary: Based on the failure of critical parameter sensors at nuclear power plants (NPPs) during accidents, a prediction model for critical parameter prediction during accidents was developed utilizing a long short‐term memory (LSTM) neural network and historical‐critical parameter operation sequences. The validation results show that the critical parameters model built with the LSTM neural network accurately predicts nuclear power, pressurizer pressure, pressurizer water level, coolant flow rate, coolant average temperature, and steam generator water level under loss of coolant accident and steam generator tube rupture conditions, and can help in the event of a sensor failure of critical operating parameters. This means that NPP operators will be able to better control the unit's status and improve safety in the event of a major operating parameter sensor failure.