Research on typical fault diagnosis for nuclear power plant based on BP-SDG

It is a significant issue for nuclear power development to keep the nuclear power plant operating safely; therefore, a variety of methods has been proposed to fault diagnosis, in order to assist the work of the operators. Back Propagation (BP) neural network has a strong nonlinear mapping capacity and it can process, identify and class the complex information rapidly, so it can be used to identify and judge the state of the system equipment. The complex relationship between the parameters and the fault propagation paths can be showed clearly by signed directed graph (SDG), and it has advantages of establishing model conveniently, inference flexible and so on. SDG is used to verify the diagnosis results and achieve the fault propagation path, which is advantageous for the operators to take effective measures to ensure the safe operation of nuclear power plant. BP neural network was used to diagnose the operating state recognition and the SDG method was used to verify the diagnosis results in this paper. Finally, the proposed method is to be verified by simulator. (authors)