Assessment of the technical condition of a transformer based on the operational parameters of its oil using the fuzzy logic method

Abstract

This article develops a mathematical model for assessing the technical condition of non-traction electrical systems of railway transport based on the operational indicators of transformer oil, as well as an algorithm for determining the integral value of technical condition. In the study, the water content, acid number, mechanical impurity content, kinematic viscosity, density, flash point, electrical strength, and dielectric losses of the transformer oil were selected as the main diagnostic parameters characterizing the technical condition of the system. A mathematical model has been developed that accounts for the relationship between these parameters, and an algorithm for determining the value based on fuzzy logic theory has been proposed for their comprehensive evaluation. For each diagnostic indicator, affiliation functions were formed, a database of fuzzy rules was developed based on expert knowledge and current regulatory documents, and an integral assessment indicator of the technical condition was calculated using the Mamdani inference algorithm and the Centroid method. The results of the study showed that the developed mathematical model and algorithm allow for a comprehensive analysis of transformer oil performance indicators, a reliable assessment of the technical condition of non-draft electrical systems, the determination of the degree of insulation system degradation, and the classification of technical condition into "very poor," "poor," "average," "good," and "excellent" categories. The proposed approach serves as a scientific and practical basis for increasing diagnostic accuracy in non-traction electrical systems of railway transport, preventing emergency situations, improving condition-based maintenance systems, and increasing the operational reliability and service life of transformers.