Performance assessment of wavelet transform based islanding detection relay

This paper proposes a new technique based on timefrequency transform such as wavelet transform for islanding detection in distributed generation. Wavelets are mathematical functions with time-frequency representation for analysis of signal with transient features. The process starts with retrieving the three-phase voltage signals at the target DG location and each phase will be processed through Wavelet Transform (WT) to derive detailed co-efficient at different levels of decompositions. The next step is to compute Wavelet Singular Entropy (WSE) for each phase from the detailed coefficients. Then WSE Index (WSEI) for islanding detection is derived by adding WSE of each phase. The comparison of proposed algorithm with the widely used Rate Of Change Of Frequency (ROCOF) and Rate Of Change Of Voltage (ROCOV) relays suggests WSEI based relay working effectively in the situations where ROCOF and ROCOV fails to do so. The proposed WSEI relay is tested on standard micro-grid structures with possible variations in operating parameters and the results indicate that the WSEI relay can reliably detect islanding conditions in large micro-grids.