Prediction of LOCA Break Size Based on 1D Convolutional Neural Network

Loss of Coolant accident (LOCA) is one of the typical accidents in nuclear power plant (NPP) safety analysis. Different break sizes require different countermeasures, and wrong measures may lead to core meltdown. Therefore, it is necessary to identify the size of the break accurately. In this paper, we use MAAP software to simulate LOCA with different break sizes for pressurized water reactor nuclear power plants, and generate a large amount of time series data. Based on these data, with the help of deep learning techniques, a 1D convolutional neural network (1D-CNN) is trained in this paper, and the trained model is used for break size prediction. The results of the test set illustrate that a well-trained 1D-CNN model can accurately predict the break size of LOCA, and the prediction results can assist the manipulator in making reasonable decisions when LOCA occur.