Research On Condition Assessment Method Of Oil-immersed Power Transformer

Power transformer is one of the very important equipments and its quantity is very large in power system. Power system accidents because of transformer failure are very frequent. At the same time transformer condition assessment is the basis of the transformer condition based maintenance. So it is necessary to research the condition assessment method of power transformer. Dissolved gasses analysis (DGA) is one of the most valuable tools in evaluating the health of a transformer. In this paper take the dissolved gas H2,CH4,C2H6,C2H4 and C2H2 in transformer oil as the evaluation indicators, divide the transformer condition into five grades:"excellent","good","in general","pay attention to"and"fault". Set up a multi-level support vector machines classifiers and Bayesian classifier transformer condition assessment evaluation model. The method and the step have been presented detailed in the paper. At last we compared the two methods with the example data.