Mutual Information Based Time Delay Estimation of Ultra High Frequency Signals Radiated from Partial Discharge Sources

Abstract

The Partial Discharge (PD) Source localization process is needed to assess the insulation condition more efficiently. As the complete breakdown of insulators would lead to power supply interruption and would cause a degraded power quality, the PD source localization process is much helpful to get rid of detrimental breakdown occurrence as the handling personnel can be warned about a probable breakdown source at a very premature state. Amongst many techniques, Time Delay of Arrival (TDOA) technique is a simple yet effective process to proceed for the PD source localization problem. Time delay estimation is an essential primary step to proceed for the localization process of a PD source using the TDOA technique. Here, Ultra High Frequency (UHF) wave emitted from the PD source is used for the Time Delay Estimation (TDE) as well as the localization process. The UHF signal is used here for its better accuracy and stability. Here, two sensors are made using microstrip patch antenna. Different techniques to estimate the TDOA of the signals in two different sensors is studied and their advantages and limitations are analyzed. In this paper, an approach to estimate the TDOA of signals in two different sensors using the Mutual Information (MI) based algorithm has been introduced which follows statistical approach. Mutual Information is the statistical interrelation between two continuous or discrete random variables. Here the data corresponding to captured signals are taken as continuous random variables and the time delay of arrival between two signals is estimated by computing the mutual information between two random variables. The Time delay is estimated in case of simulation generated theoretical partial discharge wave, after that the algorithm has been applied to experimentally captured data with the hardware set up, done in the High Voltage laboratory.