Comparative analysis of oil ageing intensity in 110 kV transformers and 330 kV autotransformers

Abstract

The results of transformer oil ageing intensity analysis in 110 kV transformers and 330 kV autotransformers are presented. Using a mathematical model of variance linear regression analysis (covariance analysis model), applied to the results of periodic tests on 231 transformers of 110 kV and 49 autotransformers of 330 kV, several statistical hypotheses have been tested to estimate the intensity of drift of oil indicators during long-term operation of transformers. The following hypotheses were tested as statistical hypotheses: the hypothesis of a significant systematic drift in the values of oils during long-term operation, which allows assessing the presence of transformer oil ageing processes. The hypothesis of equality of partial angular coefficients for regression models based on test results for each of the oil in individual transformers (regression lines are parallel), which allows estimating the differences in the ageing intensity of oils in individual transformers. The hypothesis that the group averages lie on a straight line, that is, the drift of the oil in the different transformers occurs at the same rate. The hypothesis of equality of partial free terms for regression models based on test results for each of the oil indicators in individual transformers, which allows assessing the presence of differences in the values of oil indicators at the time of commissioning of transformers, that is, the actual presence of differences in the quality of the poured oil. The results of the analysis for both 110 kV transformers and 330 kV autotransformers showed not only an additive and multiplicative bias between individual series of oil parameters, but also a significant systematic component, indicating the aging of transformer oils in the analysed transformers. It was found that the intensity of drift of oil indicators in 110 kV transformers and in 330 kV autotransformers significantly differs, which should be taken into account when building models for the early recognition of the condition of transformer oils by the results of periodic tests.