Fault Diagnosis Of Nuclear Power Equipment Based On HMM/SVM And Database Development

The Nuclear Power Plant (NPP) is a very complicated system, which may have serious consequences when the fault occurs. Condition monitoring and fault diagnosis play a vital role in guaranteeing the safety and the reliability of NPP. To enhance the diagnosis rate, a hybrid HMM/SVM (Hidden Markov Model and Support Vector Machine)model was introduced, which has been proved more effective and more accurate than the two single separate models in simulation experiments. Besides, in the fault diagnosis process, the required data is abundant and complex, so we need to explore database system of NPP, which could help the managers to get and handle the data quickly and effectively.The character information of the primary fault is studied, which is at the first and second loop system of the NPP. Secondly the basic theory and algorithms of the Hidden Markov Model (HMM) and Support Vector Machine (SVM) are studied, and a hybrid HMM/SVM model is established. At last the author studies the application of the technology of database in the fault diagnosis of the NPP, and explores the application of database system.The main contents of this thesis are as follows:1. The author briefly introduces the meaning of studying the faults diagnosis technologies in nuclear power plant and summaries the developing and the current situations of it. A feasibility analysis of fault diagnosis system for rotation machinery on nuclear-powered plant based on HMM/SVM model is also carried out.2. The technology of fault diagnosis of the NPP is studied, and the character information of the primary fault which is at the first and second loop system of the NPP is collected.3. The basic principles and algorithms of the Hidden Markov Models (HMM) and Support Vector Machine (SVM) are described and researched, and a hybrid HMM/SVM model is proposed.4. The ER model and demand of database are studied, and then its most structure is systematically designed.5. The simulation experiment is designed, and the fault diagnosis method based on hybrid HMM/SVM model is tested and verified.This research is based on the"Research on Condition Monitoring and Fault Diagnosis Technologies based on hidden markov model and support vector machine for Equipment of Nuclear Power System"(National High Technology Research and Development Program"863", No. 2008AA04Z407). It can promote the progress of on-line monitoring and technology of fault diagnosis, and it has important theoretical and practical value in guaranteeing the nuclear power equipment's safety .