Publication

Abstract : The identification of electromechanical modes in the power system has been researched extensively in the last decade. Nonlinear nonstationary techniques like variational mode decomposition (VMD) are much effective in determining mode parameters. However, these methods experiment with single-channel data and while computing for other channels needs additional computational time for the process. Many modifications have been developed for VMD, out of which multivariate VMD is the highlighted one here. It has good robustness for nonstationary signals with low SNR. Unlike multivariate empirical mode decomposition (MEMD) methods, the proposed scheme has an exact mathematical context and can better line up based on common frequency modes. To determine the mode parameters of ambient oscillations in power systems, we propose a novel technique of multivariate VMD in the above domain. The simulation results are verified with two area Kundur system and real-world phasor measurement unit (PMU) data from the Indian power grid.