Research On Faults Diagnosis Methods And System Based Hidden Markov Models (HMM) For Power Transformer

This paper researched on large scale power transformer faults diagnosis methods and system based on Hidden Markov Models (HMM). The Hidden Markov Models theories and methods are studied, then proposed the applications in faults diagnosis of large scale power transformer by HMM methods and developed the faults diagnosis system software based on HMM. The details are studied as follows:Chapter one introduced the research meaning and the general situation of faults diagnosis of large scale power transformer. The developing and current situations of the technology of power transformer faults diagnosis are summarized. The meaning of HMM methods applied to large scale power transformer faults diagnosis is discussed. At last, the research background, main contents, general structure scheme are present.Chapter two introduced the classes, structure and sorts of faults of power transformer. Then the relations of the sort of a fault to "Dissolved Gas Analysis in the Oil" (DGA) are analyzed. In the end, the usual methods faults diagnosis based on DGA are introduced, and the main shortcomings of the DGA are indicated.Chapter three introduced the basic ideas of Markov Processes, Markov Chain theories and Hidden Markov Models (HMM). Then the theories and algorithms of Hidden Markov Models are studied. The modification algorithms of HMM are proposed. In addition, the application of speech recognition uses HMM is introduced. Finally, by comparing to speech recognition, the feasibility of the HMM methods applied to power transformer faults diagnosis is discussed.Chapter four researched on large scale power transformer faults diagnosis methods based on HMM. Aim at the problem of multi observation swatch sequence, the betterments for HMM algorithms are brought forward. Then the material steps of modeling and faults diagnosis of HMM methods are propounded. Lastly, by faults diagnosis tests using practical faults data, the dependability of HMM faults diagnosis method is proved.Chapter five showed the large scale power transformer faults diagnosis system software. The overall design and the structure of system software are introduced. The developing environments, tools and mixed programming technology on this software are present. At last, the basic components and functions of the system software are illustrated.In the end, in the chapter six, all of the work in this dissertation is summed up, and the future researches on applications of HMM are prospected.