Modes Decomposition Based Smart Protection Solutions for Inverter Dominated Microgrids

Abstract

newlineMicrogrids are emerging as a crucial component of the modern power system newlineowing to significant developments in renewable energy-based distributed generation newlinetechnologies and worldwide growing concern about the environmental hazards newlinecaused by the conventional coal-based power generation units. However, the newlineaddition of the distributed generation units and the ability of the microgrids to newlineoperate both in the grid-connected mode and the islanded mode have thrown a new newlineset of challenges for the protection fraternity because of the bidirectional power flow, newlineuncertainty about the amount of power flow and wide variations in the fault current newlinemagnitude depending upon a varieties of factors such as mode of operation, structure newlineof the network, and type, rating, location and number of distributed generation units. newlineAdditionally, despite extensive research, detection of arcing high impedance faults in newlinethe low and medium voltage distribution lines of the microgrids remains a newlineformidable challenge. In this research work, the non-arcing solid fault and the arcing newlinehigh impedance fault issues are analyzed separately, and new protection strategies newlineusing advanced signal processing techniques are proposed for each of them. Finally, a newlinenon-pilot protection scheme using advanced artificial intelligence technology is newlineproposed, which can protect the microgrid against both solid and high impedance newlinefaults. newlineA differential protection scheme for microgrid based on the spectral energy content newlineand the phasor deviation of the current signals is proposed for the solid faults using the newlinenoise-robust downsampling empirical mode decomposition(DEMD) and the Teager newlineenergy operator. The lower computational complexity of the DEMD algorithm and newlinethe use of a simple threshold for classification, are the significant advantages of this newlinemethod. The immunity of the differential schemes to the data-synchronization error is newlinespecially addressed in this work. newlineThe second contribution is a new technique for arcing high impedance ground newlinefault detection in distribution feeders through analysis of the interharmonic content of newlinethe current signals. For effective separation of the interharmonic components, a novel newlinev newlinemethod using maximum overlap discrete wavelet packet transform (MODWPT) and newlinea newly developed special knot based empirical mode decomposition (KEMD) is newlineproposed. newlineIn the third contribution, an intelligent, non-pilot protection strategy for inverter newlinedominated microgrids using iterative filtering based empirical mode decomposition newline(IFEMD) and extreme learning machine is proposed. The instantaneous frequency newlineenvelops, and the instantaneous Teager energy envelops of the most informative newlineintrinsic mode function obtained through processing the local current signal by newlineIFEMD is used for fault detection and fault phase identification. A mathematical newlinemorphology based transient detection scheme is designed to pinpoint the fault newlineinception instants on the feature envelops so that only the most informative newlinelimited-length post-fault features can be extracted. It increases the overall newlinecomputational efficiency of the scheme and decreases the classification time of the newlineextreme learning machines. newlineThe proposed schemes are validated in different standard microgrid models with newlinewide variations in operating parameters in all possible network topologies and modes newlineof operation. The performances are analyzed in terms of dependability, security, newlinedetection-speed, and algorithm-complexity. newlinevi