Prediction of insulation sensitive parameters of power transformer using detrended fluctuation analysis based method /

Abstract

Utilities prefer to perform a condition-based assessment of power transformer insulation in the least possible shutdown time. It is essential to estimate the values of various insulation condition sensitive performance parameters with higher accuracy. In the present work, a technique for the estimation of performance parameters, moisture content of paper, dissipation factor of the entire oil-paper insulation, paper conductivity, and activation energy is proposed. The evaluation of these parameters using reported techniques requires complete measurement of polarization and depolarization current profile, which requires around 20000 s of measurement time. The present method uses a forecasted polarization current profile obtained from 600 s of measured polarization current data. Detrended Fluctuation Analysis is applied to the forecasted polarization current data obtained from various in-service power transformers to obtain a suitable parameter. It is shown that this parameter maintains well-defined relationships with parameters; moisture content of paper, dissipation factor of oil-paper insulation, paper conductivity, and activation energy. The proposed technique is beneficial for utilities as it eliminates the requirement of measuring depolarization current and return voltage measurement for estimating activation energy.