Neural-based time series forecasting of loss of coolant accidents in nuclear power plants

作者：

Highlights：

• Progression of nuclear LOCA accident is predicted by expert systems.

• Time series data of the nuclear accident is generated by a simulation tool.

• Prediction accuracy of 92% or more is achieved by DNN and LSTM.

• Reduction in computational costs of order of 100,000 times is observed.

摘要

•Progression of nuclear LOCA accident is predicted by expert systems.•Time series data of the nuclear accident is generated by a simulation tool.•Prediction accuracy of 92% or more is achieved by DNN and LSTM.•Reduction in computational costs of order of 100,000 times is observed.

论文关键词：Time series,Expert systems,LOCA,DNN/LSTM,Nuclear accidents

论文评审过程：Received 8 December 2019, Revised 7 June 2020, Accepted 25 June 2020, Available online 8 July 2020, Version of Record 20 July 2020.

论文官网地址：https://doi.org/10.1016/j.eswa.2020.113699