Research on Intelligent Accident Warning and Simulation for Loss of Coolant Accident in Nuclear Power Plants

Both Convolutional Neural Network (CNN) and Convolutional Long-Short Term Memory (ConvLSTM) are utilized for the accident warning and simulation for Loss of Coolant Accidents (LOCA) in this work. The advantages of CNN to effectively extract features for classification are used for accident warning. ConvLSTM, as a long-term series processing model, is adopted to simulate the LOCA developing trend. Experimental verification in different ways proves the functionality and adaptability of the model. This work also lays a good foundation for the accident warning and simulation of nuclear power plant accidents based on deep learning.