Oscillation mode analysis for multi-mode coupling power systems with high renewables penetration using improved blind source separation

The extensive application of power electronic equipment and the increasing penetration of renewable energy generation gradually strengthen the nonlinear and modal-coupling characteristics of electromechanical oscillation of modern power systems. In this study, a data-driven method based on improved blind source separation (IBSS) combined with sparse component analysis (SCA) is proposed to extract electromechanical mode (oscillation frequency, damping ratio and mode shape) from synchrophasor measurements. First, short time Fourier transform is used to convert the modal-coupling oscillation signal to sparse domain, then, on the basis of time-frequency point clustering characteristics of source signals, the mixture matrix A is estimated by frequency energy peak point algorithm, and L1 norm is utilized to separate each mode from mixture matrix A. Finally, the Hilbert identification algorithm is applied to extract the oscillation parameters. The performance of the proposed IBSS method for the mode extraction is verified using the test signal, the simulation signal, and the measured data