Research On Distributed Fault Diagnosis Technology Of Reactor Coolant System In Nuclear Power Plant

The Fukushima nuclear power accident happened in March2011has a significant impacton the global nuclear industry, but in general the development of China’s nuclear industry isstill in an important period of strategic opportunities. Currently, how to ensure the safetyoperation of nuclear power plant has been a major problem that ought to be resolved as thenuclear safety is the worldwide focus issue.Fault Diagnosis technology has been attached great importance in the nuclear powerplant because the application of Fault Diagnosis system could reduce the deterrence of nuclearaccident and enhance the reliability of operation. The goal of this thesis is to develop aneffective Fault Diagnosis system.The thesis has been studied for the following aspects:(1) Based on the results of existing research, a study of the fundamental structures andfunctions of the equipments in the nuclear power plant primary coolant system and a analysisof the main monitoring parameters and fault mechanisms of the important equipments havebeen made to build a knowledge database for the typical fault, which is utilized as thediagnosis objection of the whole fault diagnosis system.(2) By comparing several theories, model and algoriums that is widely used and matureintelligent fault diagnosis technology, choosing a proper method which is suitable for theoperation parameters and the characters of the typical faults as the main method of the system.(3) Refer to distributed concept a distributed fault diagnosis is introduced and the faultdiagnosis structure in details is built based on the proposed distributed fault diagnosisstrategy.(4) Visual Basic6.0is employed to develop the fault diagnosis system to realize thefunctions of its modules.The distributed fault diagnosis system adopts the global diagnostic strategies before thelocal diagnostic strategies to diagnose faults, in which the global diagnostic module uses themethod of data fusion fault diagnosis for the global positioning of faults, then the overallresult produced by the global diagnostic module can trigger the local diagnostic modules todiagnose again with the method of fuzzy neural network. Simulation results showed that the fault diagnosis system had a high diagnostic accuracy and feasibility for diagnosis and wouldbe useful for the development of intelligent fault diagnosis system.