Research On Fault Diagnosis Technology Of Nuclear Power Plant Based On SDG-ES

The nuclear accident would cause huge threat to personnel and the environment because the nuclear power plant has high radioactivity during the operation. Therefore, nuclear safety has become the worldwide focus issue. The Fukushima nuclear power accident occurred in March 2011 in Japan that stagnated the development of the nuclear industry. Guarantee the safe operation of nuclear power plants has been a common goal in the world. Fault diagnosis technology is an effective means to enhance the security of the system, which is also a focus in the field of nuclear safety.It is difficult to diagnose the fault through establishing a precise mathematical model because nuclear power plant is large and complex. Signed directed graph (Signed Directed Graph, SDG) as a qualitative fault diagnosis method, can express complex causal relationships and is applicable to fault diagnosis of nuclear power plant which is complex.Based on the summary of SDG fault diagnosis methods and the deep study the characteristics of traditional SDG diagnostic methods, an approach of establishing SDG model of nuclear power plant is proposed in this paper, and reactor coolant system of PWR was taken as the research object and its SDG model has been built. The fault revealing time is taken into account to effectively solve the problem of closed-loop in SDG model which would lead to cannot diagnose. The method of combining the SDG and expert system is presented to improve the diagnostic speed and resolution of diagnosis, the method of extracting the expert system rules based on the SDG model is given. The SDG model will directly be used as an expression of the knowledge of the expert system to overcome the difficulties of traditional expert system to obtain knowledge when diagnosing the multi-fault.On the basis of the above theory, SDG-ES fault diagnosis expert system of nuclear power plant is developed with the C #3.5 language and SQL Server 2005 database technology. The test results show that the system has the advantages of friendly interface, easy operation, and rapid diagnosis.