Correlation of partial discharge occurrence in power transformer by using self organizing map, acoustic partial discharge and DGA

Abstract

Power transformer consists of components which are under consistent thermal and electrical stresses. The major component which degrades under these stresses is the paper insulation of power transformer. The degradation is accelerated by the existence of partial discharge in the transformer. The partial discharge activity in the transformer will generate gasses such as Hydrogen (H 2) and Methane (CH 4). According to IEC 60599, hydrogen may be produced by other ways than PD itself. The current predictive maintenance methodology practiced by utility that could detect partia discharge in the transformer is the dissolved gas analysis (DGA). Hydrogen is the main gas used for decision making process. In this project, acoustic partial discharge measurement technique was used together with the monthly generation rates of hydrogen and methane in proportion to Total Dissolved Combustible Gas (TDCG) method was used to identify the PD occurrence in power transformer. In this paper, a field testing was done on 120 transformers to verify the proposed method by using classification tool such as Self Organizing Map (SOM). The monthly generation rates of hydrogen and methane in proportion to TDCG method serves as useful information in identifying the partial discharge occurrence in power transformer. The correlation between the monthly generation rates of hydrogen and methane in proportion to TDCG and PD activities has been established by using SOM. As a result, the SOM gave two regions which represents the "with PD" and "no PD" region. This result has been confirmed by doing acoustic PD measurements. © 2009 WASET.ORG.