Study On A Fault Diagnosis Technology Of Nuclear Power Plant Based On Expert System

To ensure nuclear safety has been the primary goal of the nuclear power plant operation.The systems and equipments of nuclear power plant are large and complex ted and which has many measuring points,in case of failure will produce a huge number of alarms which bring the great stress to operators,increasing person errors.Exploiting nuclear power plant fault diagnosis system for assessing the state of the nuclear power plant operation can quickly compress the number of alarms generated by the alarm analysis to identify the cause of alarm which leads to the abnormal system,exclude the result alarms and determine the cause of the failure has great significance to guide operators to relieve the consequence of accident.Expert system established by expert knowledge of domain experts,without having to establish a detailed mathematical model of the system and equipment,which can describe the complex causal relationship between fault symptoms and fault,support alarm analysis and fault diagnosis capabilities,with proven technology,diagnostic process clear,the traceable diagnostic results and good robustness for the development and application of nuclear power plant fault diagnosis system.In this paper,the study on fault diagnosis of nuclear power plant based on expert system has two difficult issues,the building of the faults diagnosis expert knowledge base and reasoning technology of the faults diagnosis expert knowledge base.The chapter mainly includes the following contents:(1)The construction of alarm analysis expert knowledge bases on multi-level flow model technology,on the foundation of traditional multi-level flow model,a standardizing and normalizing alarm reasoning relationship is proposed to build the alarm analysis expert knowledge bases which is combing with the multi-level flow model.(2)Faults diagnosis expert knowledge base is divided into typical faults diagnosis expert knowledge base and common fault expert knowledge base.The construction method of typical faults diagnosis expert knowledge base is combing the alarm analysis expert knowledge base with FMEA to determine the scope of failure.According to extract the typical signs of faults alarm failure from analysis expert knowledge,by using typical rules from faults to signs to build the typical faults diagnosis expert knowledge base.Rule base method is adopted in the construction of common fault expert knowledge base.(3)Taking primary coolant system of two-loop pressurized water reactor nuclear power plant as the object,establishing the multi-flow model of nuclear power plant primary coolant system,and building the alarm analysis expert knowledge bases.To build the diagnosis expert knowledge base according to FMEA and specific rules.(4)The fault diagnosis system of nuclear power plant has been designed in the principle of integration,modularization and expansibility,and the development of the fault diagnosis system of nuclear power plant designed to using real time expert system development platform and through the verification and implementation of the various modules function of fault diagnosis system.The results of this thesis show that:(1)To build alarm analysis and fault diagnosis expert knowledge base by modeling nuclear power plant using Multilevel Flow Models has many advantages such as the models are compact,easy to modify and extend.(2)The fault diagnosis system for nuclear power plant based on expert system designed and developed in this thesis can diagnose the abnormal states of nuclear power plant real-time.The research results of this paper lays the foundation for the development of the nuclear power plant operation support system,and key technologies used is mature and reliable.Test results show that fault diagnosis system has advantages of the rapid and accurate diagnosis,easy manipulation.The author believes that the design and development of the nuclear power plant fault diagnosis system can provide a reference for future study for fault diagnosis system of nuclear power plant.