Prediction of Insulating Oil for Transformers Through Fuzzy Logic Approach

Abstract

Mineral oil is the most commonly used insulating oil in distribution and transmission transformers. It possesses good physical, chemical and insulating properties making it one of the most suitable oils for insulation purpose. However, a few of its properties like low fire point pose serious threats to the electrical power system. Moreover, its nature of nonbiodegradability makes its usage a serious concern and motivates engineers to think about other alternatives to be used as dielectric oils in transformers. Vegetable oils like palm oil, rapeseed oil etc. are being used for the insulation purposes in transformers for over last two decades. In this paper, important parameters such as breakdown voltage, relative permittivity, capacitance and dissipation factor (tan δ) have been studied for different insulator oil samples. The samples are considered as Palm Oil (P.O.), Coconut Oil (C.O.), Rapeseed Oil (R.O.), Sunflower Oil (S.O.), and mineral oil (M.O.). In addition, a fuzzy logic-based approach with total of 729 rules has been developed for the selection of insulating oil based on the given set of oil properties. The results show the validation of the fuzzy logic- based model in selecting the type of oil while comparing it with the experimental results.