A concurrent fault diagnosis method for nuclear power plants based on wavelet neural networks

A new fault diagnosis method based on the wavelet neural network (WNN) for nuclear power plants (NPPs) was proposed to solve the problem of concurrent fault diagnosis of NPPs. The concurrent fault diagnosis modes for the condensate feed water system and nuclear power system were established by using of WNN to demonstrate the feasibility of the proposed diagnosis method. Simulation data of the typical concurrent faults of the condensate feed water system and nuclear power system were employed to demonstrate the effect of this diagnosis method. The results show that it is successful for concurrent fault diagnosis of nuclear power system by WNNs. The WNNs are more efficient compared with BP neural networks and ART-BP neural networks in concurrent fault diagnosis.