A Study of Impact of Environment Factor on Power System Equipment Partial Discharge Using Multi-Groups Analysis of Variance and Correlation Analysis

The main purpose of this paper is to propose an acoustic imager to diagnosis partial discharge for electro-mechanical equipment. A high voltage power generator applied to three types of porcelain insulators one is with 10% salt concentration another is with 20% salt concentration and the other is with salt contaminated for laboratory experiments. Taiwan power system field-measurements collect data will comparison with laboratory experiments for evaluation the prediction of diagnosing partial discharge. One-way ANOVA method which statistical analysis includes data mean and standard deviation and statics test will use to test mean difference larger or small, P value will test mean difference significance between relative humidity groups. Correlation analysis is used to test the relevance degree between humidity on the Partial discharge dB data. The research has the following results: 1.The ANOVA method had verified that when humidity is considered as a factor, there is a very significant mean difference in dB data regardless of laboratory experiments or field measurements. It also detected which between groups the mean difference occurs. 2. The mean of laboratory experiments and field measurements in the same humidity groups can be used to evaluate the predictive value of diagnosing partial discharge through the mean error formulation. 3. The correlation analysis results show that the salt concentration had a negative correlation to positive medium correlation with dB data. Field- measurements had a negative correlation to high correlation with dB data.