Classification of defects in ceramic insulators using partial discharge signatures extracted from radio frequency (RF) signals

The presence of insulation defects results in the initiation of partial discharge (PD) activities. As insulation failure and hence a failure of power equipment can occur due to the cumulative adverse effects of partial discharges, it is important to detect PD activities at early stages. The current techniques used in PD off-line analyses are not suitable for detecting defected insulators in the field. In this work, several cases of insulator defects are investigated in an effort to develop an on-line RF based PD monitoring technique using ceramic disc insulators with different types of defects. Like, an intentionally cracked ceramic insulator disc, a disc with a hole through the cap which results in internal discharges, and a completely broken insulator disc forming the first three classes. An external corona noise using a point to plane setup comprised the forth class. The defected discs are incorporated into strings of 2, 3 and 4 insulators. To analyze the discharges from different types of defects, the Artificial Neural Network (ANN) and Support Vector Machine (SVM) algorithms are applied to the extracted features, and the recognition rates for each class are reported.