Introduction of Three Methods Used for the Nuclear Accident Diagnosis

In order to deal with the nuclear severe accidents, the severe accident management systems are popularly considered and developed at home and abroad recently. A severe accident management system usually includes these functional parts: accident monitor, accident diagnosis, accident simulation, accident prognosis and SAMG support. Here, the accident diagnosis part is mainly concerned, and three nuclear accident diagnosis methods are introduced here, including BP neural network method, SDG expert diagnosis technique and artificial diagnosis method, which are also applied to a severe accident management system developed by us. In this paper, firstly, the severe accident management system developed by us will be introduced briefly. Then, three accident diagnosis methods for nuclear power plant (NPP) are showed and described in detail. At last, two cases including LOCA and SGTR accidents are used for the verification of these accident diagnosis methods and some analyzing results and conclusions are given. The results show that the three diagnosis methods are very useful for the accident diagnosis of NPP, which can diagnose the accident type accurately and offer much information or support to the severe accident management system and operators. The paper offers some reference significance for the research of accident diagnosis methods and the development of severe accident management system.