RESEARCH ON FAULT DIAGNOSIS OF REACTOR COOLANT ACCIDENT IN NUCLEAR POWER PLANT BASED ON RADIAL BASIS FUNCTION AND FUZZY NEURAL NETWORK

Nuclear power plants (NPPs) are widely used in the world. After three nuclear accidents, people propose higher of the safety and reliability on NPPs. Reactor coolant system (RCS) in the NPP directly affects whether the heat can be exported and radioactivity can be inclusive. It plays an important role of the NPPs safety. So, it is great significance of fault diagnosis for RCS in NPP. Although many scholar had carried out research on fault diagnosis of NPPs, different networks may lead to different results in a system. Therefore, this paper chooses a system and uses different neural networks (NN) for comparative analysis which can provide advice for follow-up research. In the paper, RCS has been analyzed and typical fault have been analyzed through PCTRAN simulator. On this basis, two kinds of NN combined with fuzzy systems: radial basis function (RBF) and back propagation (BP) are used for fault diagnosis and comparative analysis. Loss of coolant accident, single pump failure, loss of feed water are set for simulation experiment. Simulation experiment shows that BP network's hidden layer nodes is less than RBF-NN, but iteration speed of BP network is faster; accuracy of fault diagnosis based on BP-NN is higher than RBF-NN; fuzzy-NN for fault diagnosis is faster than NN.