Identification and localisation of incipient discharges in transformer insulation adopting UHF technique

Incipient discharges formed in natural ester oil due to corona activity, surface discharges and particle movement were investigated using UHF technique. When the applied voltage magnitude is increased, on the inception of discharges, the peak to peak magnitude of the voltage of the UHF signal formed also increases. After certain applied voltage magnitude increase, if it is reduced, the magnitude of UHF signal formed is much higher than the magnitude of the signal formed when the voltage is increased. This phenomenon is observed with all the three types of discharges. Cross Recurrence plot (CRP) can be effectively used to determine the Time Difference Of Arrival (TDOA) of UHF signals. Self-similarity Recurrence Quantification Analysis (RQA) parameter is found to be more suitable technique to estimate the TDOA of UHF signals. CRP method is used to estimate TDOA correctly, even with low signal to noise ratio and it is found to be superior to Cross-correlation technique in estimating the TDOA. Estimated TDOA of UHF signals are validated and incipient discharge source is localized by adopting the non-iterative technique.