Fast and reliable detection of power islands using transient signals

Abstract

A new technique for fast detection of power islands in a distribution network, which uses transient signals generated during the islanding event is investigated. Performance comparison of several pattern recognition techniques in classifying the transient generating events as islanding or non-islanding is presented. Features for the classifiers are extracted using the Discrete Wavelet Transform of current signal transients. Using a set of extracted features from simulated current signals, (i) a decision tree classifier, (ii) a probabilistic neural network classifier, and (iii) a support vector machine classifier were trained for recognizing the transient patterns originating from the islanding events. The trained classifiers were then tested with unseen test current waveforms. The test results demonstrated that the investigated technique can potentially provide a new way for identification of islanding in distribution systems. The approach was then extended changing the feature set and sampling frequency. Proposed method is finally compared with an existing islanding detection technique.