An insulation condition assessment method of transformer based on set pair analysis and similarity clustering

Abstract

An insulation condition assessment method of the transformer is proposed in this paper. The set pair analysis and the expert evaluation method based on the set-valued statistics are adopted to assess insulation condition of the quantitative index and the qualitative index respectively according to the insulation condition assessment model of the transformer which is built. The subjective weight of evaluation index is determined by synthesizing the fuzzy analytic hierarchy process (FAHP) and the similarity clustering. The criteria importance through inter-criteria correlation (CRITIC) method is used to revise the subjective weight to the subjective and objective weight. The insulation condition of the transformer could be evaluated on the basis of the degradation value that is obtained with the weighted average method. The actual test data and inspection records of a distribution transformer are utilized to verify the assessment method and the numerical results testify that: we could evaluate the insulation condition of the transformer correctly and intuitively with the assessment method proposed in this paper, and considering authority of experts and objectivity of weights could give a more reasonable assessment result than ignoring them. The research in this paper could provide a feasible reference for the insulation condition assessment and technical support for the condition based maintenance of the transformer.