Discrete Wavelet Transform and kNN-Based Fault Detector and Classifier for PV Integrated Microgrid

Abstract

The growing penetration of distributed energy resources (DERs) in modern power distribution networks operating as microgrid poses a great challenge for the conventional protection scheme due to significant variation in the fault current levels under the grid-connected and islanded mode of operation. In this regard, this paper has devised an efficient protection scheme based on discrete wavelet transform (DWT) and k-nearest neighbour (kNN) for fault detection/classification implemented for dual modes of microgrid operation considering the photovoltaic PV source and nonlinearity in the load. The proposed approach utilizes the three-phase voltage and current signals obtained during shunt faults in the distribution line under widely varying fault parameters. The pre-processing of signals through DWT determines the approximate coefficient. The standard deviation (SD) of the approximate coefficient so obtained is further fed as the input to the kNN-based classifier for fault detection/classification task separately for grid-connected and islanded mode. The test result analysis clearly reveals the effectiveness of the proposed approach and hence validates the performance.