The Expert Diagnosis System Of Transformer Winding Deformation Based On Frequency Response Method

Winding deformation is one of the hidden dangers of the safe operation of transformer. The frequency response analysis method is widely used because of its short testing time, high detection sensitivity and stable measured spectrum. However, the frequency response method relies heavily on personnel’s experience. The current diagnostic criterion is easy to produce false diagnosis. And the specific deformation information cannot be concluded on site. Therefore, based on the frequency response analysis method, a fault diagnosis expert system of transformer winding is designed and developed. It improves the existing diagnostic criterion, and can diagnose the degree, type and location of winding deformation rapidly and accurately on site.Firstly,the paper introduces the frequency response method of the winding deformation diagnosis. With the common types of winding deformation concluded, the corresponding relationship between changes of frequency response spectrum and winding deformation types is established. With the frequency response spectrum in various cases analyzed, the characteristics of frequency response graph and the experience of winding deformation diagnosis using frequency response analysis are summarized. The influence of various field test conditions on the frequency response is identified through the study of the spectrum test process. Those conclusions constitute the knowledge base of the expert diagnose system.Secondly, the nature of the frequency response curve of winding is further analyzed; the equivalent circuit model of the winding is established based on the electrical characteristics of the transformer winding circuit; the simulation of its frequency response is carried out, and the frequency response curve of the winding model is obtained. The effects of various parameters on the frequency response spectrum at different frequency bands are acquired through the study of the frequency response spectrum at different frequency band. These works lay the foundation of the reasoning machine developing of the expert diagnose system.Finally, the transformer winding deformation diagnosis expert system is developed and implemented. According to the characteristics of winding deformation diagnosis, the knowledge base adopts the production structure, and the reasoning machine adopts a mixture of forward and reverse reasoning model. In view of the diagnosis mission required, an appropriate hardware and software platform is chosen, and the winding deformation expert system is designed. After the inspection of many actual cases, the expert system developed can diagnose the degree, type and location of winding deformation rapidly and accurately on site. With a reasonable use of national standard, the system can reduce the false diagnosis greatly.