Current State of On-line Partial Discharge Diagnosis Methods in Hydro-Generator Using a Novel Noise Reduction Method

摘要

Analysis of partial discharge signal is one of the most effective and commonly usedtechniques to estimate a condition of the winding insulation of generators and requires highlyprecise analysis. To satisfy this requirement, various noise reduction methods have beenproposed and developed to eliminate external noises those were measured with partialdischarge signals. However, some external noises those could not be removed byconventional noise reduction methods have been come out and decreased the reliability of theanalysis of partial discharge signals. In this paper, therefore, we propose a novel noisereduction method to remove the external noises based on the characteristics of partialdischarge signals and particular external noises. The proposed method combines conventionalnoise reduction methods, several new noise reduction schemes and a selection code which cannumerically express the noise possessing degree of signals. The proposed method shows goodperformance in the real partial discharge signals measured from hydro-generators in Korea.