Condition assessment of in-situ generator transformers by frequency domain analysis using time domain data

The present paper shows that time domain data measured from different in-situ generator transformers can be successfully transformed to frequency domain. It is also shown that the transformed frequency domain data can be used to determine insulation condition. For the present work, Polarization-Depolarization Current is measured using a low-cost instrument developed at High Tension Laboratory of Jadavpur University. The transformation of data from time to frequency domain is tested using data recorded from several generator transformers belonging to NTPC Ltd. The paper reports the results obtained from some of these transformers. The value of paper moisture predicted using the measured time domain data and the transformed frequency domain data are also observed to be in good agreement with each other. Presence of a small relaxation peak in the plot of ε"_r(ω) at low frequencies, below the frequency of minimum ε"_r(ω), is an indication of high paper moisture content in oil-paper insulation. Such peaks are clearly visible in the transformed frequency domain data of the transformers which have high moisture content.