Research on Fault Diagnosis Method for Reactor Primary Circuit System Based on multi-source information fusion

Abstract

Reactor primary circuit system is a complex dynamic system, variable parameter coupling, operation safety problems are prominent. In order to reduce the risk, a multi-source information fusion diagnosis system based on signed directed graph (SDG) and particle swarm optimization BP neural network (PSO-BP) is proposed. Utilizing D-S evidence theory for neural network diagnostic information fusion, logic inference combining SDG model, to determine potential failure. Simulation test shows that the intelligent diagnosis model could estimate the faults effectively, and provides the fault alarm transmission path.