Mathematical Morphology aided Random Forest Classifier based High Voltage Porcelain Insulator Contamination level Classification

In the present work, the influence of different surface contaminations on the performance of a porcelain insulator is investigated. The measurement of the Leakage current on different classes of the contaminated porcelain insulators is performed with the help of an experimental setup. Necessary features from the recorded leakage current data are thereafter identified and used for analysis purposes. The entire data analysis has been implemented at the software platform. In the reported work the authors used Mathematical Morphological function and different statistical operations for the feature extraction purpose. By using the Random Forest classifier, authors have achieved the optimal feature classification. The accuracy level of the Random Forest classifier for different classes of contamination is also reported in the paper.