RESEARCH ON ACCIDENT PREDICTION OF NUCLEAR POWER PLANTS BASED ON DEEP LEARNING

Based on the bidirectional long short-term memory network (BILSTM) and the long short-term memory network (LSTM), this paper uses the rolling update mechanism (RU) to obtain the time series training set data and test set data of the original data, and predicts the water level and pressure changes of the Pressurizer after the LOCA accident in nuclear power plants, so as to provide a certain reference for the prediction of LOCA accidents in nuclear power plants. Specifically, the following conclusions are obtained: Compared with the LSTM model, the BILSTM model has better prediction results and less prediction loss. Although the average error of the prediction results of the two models is 10(-3) orders of magnitude, the average error of BILSTM is reduced by more than 20% compared with LSTM.