Making decisions about status of furnace transformer with generalized index based on fuzzy set logic

Powerful transformers at PJSC Magnitogorsk Iron and Steel Works are equipped with continuous condition monitoring systems. Available diagnostic system uses the electric method to measure an apparent discharge and a partial discharge intensity (PDI).There is a challenge to develop a decision-making system that could detect an insulation defect and estimate its severity for further transformer operation. We propose to reduce the total of diagnostic variables to a common generalized deterministic parameter reflecting numerical and quality status of the insulation. For this purpose, we use the mathematical apparatus built upon composition of fuzzy linguistic variables and metric values of two measurable parameters of the diagnosable furnace transformer. Reliability and convenience of the proposed method are proven by corresponding calculation results based on data supplied by the industrial on-line monitoring system. These results demonstrate an adequate model behavior with regard to furnace transformers.