Partial Discharge Localization in Oil through Acoustic Emission Technique Utilizing Fuzzy Logic

Abstract

This study presents an examination of partial discharge (PD) localization in oil based on acoustic emission technique utilizing fuzzy logic. First, the PD measurement and localization systems are developed. In total, 2 coordinates of needle-plane electrodes that represent the PD location and 336 acoustic emission sensor (AES) coordinates are examined. The electrical PD is measured through an impedance matching circuit (IMC). The acoustic PD is acquired through AES and its signal is boosted by a preamplifier gain unit. The acoustic and electrical PDs are collected once the voltage reaches 30 kV. The data are first preprocessed by discrete wavelet transform (DWT) whereby it will be further analyzed by time of arrival (TOA), fuzzy logic Mamdani (FLM), and fuzzy logic Takagi–Sugeno (FLTS) methods. The arrival of time and distance between PD and AES are calculated based on TOA to find the PD location. Next, TOA and distance to PD location are used to define the input and output membership functions for FLM and FLTS to determine the PD location. It is found that FLTS has a great potential to be used for localization of the PD in oil whereby it has the lowest errors between actual PD location and AES. The performance of FLM and TOA is inferior to FLTS whereby the errors between actual PD location and AES are quite high for both PD locations.