An intelligence-based islanding detection method using DWT and ANN

Authors: MEHRDAD HEIDARI, GHODRATOLLAH SEIFOSSADAT, MORTEZA RAZAZ

Abstract: In this paper, a new method based on artificial neural network (ANN) and discrete wavelet transform (DWT) is proposed for electrical islanding detection. Transient signals produced during an event are used in the proposed method. ANN is trained to classify the transient events as islanding and nonislanding. The required features for classifying are extracted through DWT of voltage and current transient signals. The proposed method is then simulated on a medium voltage distribution system of CIGRE with 2 kinds of DGs. Results show that this method can detect electrical islands more rapidly and accurately.

Keywords: Discrete wavelet transform, electrical islanding detection, neural network, distributed generation, distribution system and transient signals

Full Text: PDF