Power system modal identification via wavelet analysis

The monitoring of dynamic performances by means of WAMS allows keeping track of oscillation modes during the system operation and consequently detecting undesirable on comings of possible undamped oscillation behaviors of electric variables. In this paper, the problem of monitoring power system oscillation modes by means of a wavelet-based spectral analysis is addressed. A monitoring architecture, aimed at achieving the online supervision of power system dynamic performances, is presented. The feasibility of the proposed approach has been proved through simulations on a test grid and in presence of system noise and measurement noise, through simulations on a real-sized model of the Italian power system. The results of the eigenvalue analysis, applied to the same test case, have been adopted as common base in order to compare the results coming from the wavelet analysis with the results coming from the prony analysis.