Smart antenna-based partial discharge detection and classification system

This paper discusses the use of a one arm Archimedean Sinusoidal Spiral Antenna (ASSA) to capture high frequency signals generated by Partial Discharge (PD) in an oil filled transformer. The antenna is designed such that it blocks low frequencies which corresponds to corona and noise in the transformer and allows higher frequencies which result from harmful PD activities. The captured PD signals will be further processed and classified according to the PD type. A simulation using HFSS® is compared with the real data obtained using a Network Analyzer. Results show that the ASSA performs well in the frequency range between 1GHz and 5GHz. Data processing includes labeling, extracting, and finally classifying was performed on the collected PD signals in both online and offline modes to determine the type of the detected PD.