A Novel Islanding Detection Method Based on Wavelet Packet Transform and Extreme Learning Machine

Abstract

In this manuscript, wavelet packet transform (WPT) with extreme learning machine (ELM) based method is proposed for detection of the islanding condition in the distribution system with the presence of multiple distributed generations (DGs). The system consists of different types of DGs like hydro turbine generator with synchronous machine and wind turbine generator with asynchronous machine. Negative sequence component based assessment and analysis of fault conditions is considered in this work. The change of energy components are calculated using WPT at different nodes and considered as feature index for a particular fault condition. Very often occurred practical islanding and non-islanding events like capacitor switching, load rejection, line to line fault, three phase fault, voltage sag and swell etc. are simulated. Based on the feature index, ELM is applied as a classifier to distinguish islanding from non-islanding events. The results are presented with comparison to other classifiers like decision tree, artificial neural network (ANN) and support vector machine (SVM). It has been found that the proposed WPT-ELM technique is highly effective to discriminate islanding events under a wide range of operating conditions from other type of disturbances in the power distribution network. The proposed scheme is fully simulated by the MATLAB/SIMULINK environment.