Expert system for health index assessment of power transformers

Abstract

A fuzzy logic based technique using dissolved gas analysis to determine the health index of the oil in a power transformer is proposed in the present paper. The proposed fuzzy model integrates the DGA standards and the practical knowledge of transformer diagnostics experts in making a consistent decision on the health condition of the transformer oil. Twenty different transformer oil samples of a power utility validate the reliability of the proposed diagnostic models. Furthermore, a fuzzy logic based overall transformer health assessment model has been proposed in this paper. This model uses several diagnostic tests data of transformers including water content, break down voltage, interfacial tension. It is envisioned that the proposed fuzzy models will prove very convenient even for inexperienced engineers to determine the health index of the transformers. This shall help in initiating suitable action for proper maintenance of the transformers and enhancing their remnant life.