Application of wavelet theory to power distribution systems for fault detection

In this paper, an investigation of the wavelet transform as a means of creating a feature extractor for artificial neural network (ANN) training is presented for application to distribution network fault location. The study includes a terrestrial-based three-phase delta-delta power distribution system. Faults were injected into the system and data was obtained from experimentation. Graphical representations of the feature extractors obtained in the time domain, the frequency domain and the wavelet domain are presented to ascertain the superiority of the wavelet transform feature extractor.