Pattern Classification of Partial Discharge in Lv Winding of 1- Transformer Using Pca

Partial discharge is the commencement of failure of insulation of electrical devices used at high voltage. If such discharge is not detected then fault may increase at the working duration. Basically partial discharge is the cause of failure of insulation in electrical equipment. The insulation of electrical equipment deteriorates due to thermal, mechanical, electrical and environmental stresses. Continuous on-line monitoring of partial discharge (PD) activities involves recording large amounts of data and it is hallenging task to extract useful information from them. This paper proposes a partial discharge analysis and its pattern classification in LV winding of single phase Transformer using straight or electrical technique for analysis and principle component analysis algorithm for recognition of partial discharge pattern. Getting different patterns by changing fault location, we can classify these patterns according to their position in the LV winding. To measure the partial discharge we have different technique like direct or electrical method, acoustic measurement, balance. In this paper electrical method is used in which PD Analyzer DDX-9101 is used in which we get the 2-D pattern in sinusoidal waveform at different voltage level by changing location of fault in LV winding. When we vary voltage level with changing location of fault in LV winding we get variation in peaks level and also peaks location i.e. in +ve and –ve half cycle. By analysis of these peaks level and peaks location at different fault location we can analysie these pattern and also recognize or classify these partial discharge pattern with the help of principle component analysis (PCA) algorithm. It has been concluded that the discharge magnitude (PD level) will be very low if fautl is at near supply end, as we go onwards to the supply end discharge magnitude will increase.