Skip to Main Content
An online fault diagnosis method based on a hybrid artificial neural network (ANN) for nuclear power plant (NPP) is proposed in the paper. It adopts the BP ANN for a quickly group pre-diagnosis at first, then uses the RBF ANNs to verify the results of the BP ANN. Several simulation experiments are carried out using a NPP simulator while the NPP is under different operating conditions. The results show that the proposed method can not only diagnose the learned faults quickly and accurately, but also identify the unlearned faults under different operating conditions, even with noise signal in the input data. The output of the diagnosis system is a list of the possible faults with their probabilities. This makes the diagnosis result be more understandable and acceptable for the operator of NPP.