Variational mode decomposition <scp>‐</scp> subspace‐K‐nearest neighbour based islanding detection in distributed generation system

Abstract

Owing to the increased penetration of renewable energy resources, the issue of unintentional islanding in distributed generation systems has been one of the most common topics of research over the last decade. In this connection, the present work is intended for a solid addition to the existing literature of unintentional islanding detection approaches (UIDA). This work provides a novel application of variational mode decomposition (VMD) and subspace-K-nearest neighbour (SSKNN) method for the recognition of un-intended islanding events. Initially, the proposed UIDA extracts the 3-phase voltage signals at the targeted distributed generator (DG) location, and uses VMD to acquire the principal modes. Afterward, four major feature indices such as relative mode energy ratio, mode instantaneous amplitude, number of zero crossings and centre frequency are formulated considering the first three modes. Lastly, the SSKNN based classifier is trained and tested for effective recognition of islanding events. The performance of the proposed method is tested under several diverse microgrid operating conditions comprising of both islanding and non-islanding events. The obtained result demonstrates the effectiveness of the proposed UIDA compared to other similar approaches preserving the requirements of IEEE 1547 islanding detection standard.