A New Passive Islanding Detection Method for Distributed Generations Using Time-Frequency Transform-Based EMD-HT

This paper presents a new method based on empirical mode decomposition and Hilbert trans-form for detecting islanding from non-islanding events in distributed generations. In this pa-per, the empirical mode decomposition is used to separate out intrinsic mode functions (IMF) from non-stationary power signal disturbance waveforms. Then, the Hilbert transform ap-plied on all the IMFs to extract the instantaneous amplitude and frequency components. The required index is extracted from the energy spectrum density of the signals that introduced as U1, U2, and U3 and we can detect islanding events with high accuracy, low cost and high speed in distributed generations.