Mathematical morphology‐based islanding detection for distributed generation

The study focuses on a novel strategy for microgrid protection based on mathematical morphology. Mathematical morphology is a time‐domain signal processing tool that could be used for accurate and reliable signal component extraction. In this study, a class of non‐linear multiresolution decomposition scheme called morphological Haar wavelet (MHW) is used for detection of faults in microgrid. The proposed method uses the detail signal of MHW obtained after prediction lifting for fault detection and faulty phase selection. Accordingly, current retrieved from both ends of the feeder are processed through MHW after prediction lifting scheme to obtain the detail signals. The detail signal difference and their norms are calculated to obtain a primary protection scheme for the feeder. Also the same scheme was investigated as backup protection for very low fault resistances by taking the currents of the adjacent feeders. The performance of proposed technique was evaluated for islanded and grid connected mode of operation, for different network topologies, i.e. radial and looped system for various fault and non‐fault disturbances.