Identification of transients for Islanding detection in power systems

Abstract

Abstractlyransients based methods of islanding detection are investigated and different islandingdetection methods are studied comparing their performances. A test system based on a practical power system of Sri Lanka centring the Kukulegangahydro power plant is used in PSCAD/EMTDC simulation software is used as thetool and the system was verified for an actual fault event. The verified model is used to simulate multiple events including faults, load tripping, capacitor bank switching, breaker operations and generator tripping and adding. Voltage and current waveforms are rocessed using the Discrete wavelet transform (DWT)and the third (D3) and fourth (D4) level detail outputs of DWT are identified as the best candidates for differentiating islanding transients. A tolerance band is selected to distinguish islanding phenomena from other events. Logic for filtering out non-islanding events is proposed.