Cross-wavelet transform based feature extraction for classification of noisy partial discharge signals

Partial discharge detection and classification are important for safety and reliability of power equipment. A novel cross-wavelet transform based technique is used in this work for feature extraction from partial discharge signals. Results show that cross-wavelet transform eliminates the effect of random, real-life noises and therefore the partial discharge patterns can be classified properly from the noisy waveforms. Different partial discharge patterns are recorded from the various samples prepared with known defects. Features are extracted from the raw noisy data and a rough-set based classifier is used to classify the patterns. Efficient classification of the patterns justifies the approach.