Validating Health Status of Transformer: Grey Relational Analysis and Fuzzy Logic

In the present paper, a new grey relational analysis-based Fuzzy Logic (FL) model has been proposed to determine the overall health index of transformers. Health of transformer is directly affected to its insulation. Hence, accurate monitoring of insulation condition in return indicates the present health state of transformer insulation. This is possible with an effective decision-making system. For the same, two decision making techniques, namely, grey relational analysis and fuzzy logic expert system are used in this paper. Further, ranking based relational analysis is used to reduce all input attribute to three input grades. The method is based on concentrations of dissolved gases, diagnostic test results of transformer oil and paper insulation. These input grades are combined with a designed expert fuzzy rule base to determine health condition of transformer insulations. Integration of grey relational analysis with fuzzy logic system resolve several issues of conventional techniques such as large number of rules, requirement of large amount of training data and complexity issues. Results obtained from the present proposed model have been compared with multi-criterion based fuzzy logic model. Data corresponding to 30 in-service transformers installed by HPSEB, Himachal Pradesh, India, has been considered. The accuracy attained of the present proposed expert system is 94%.