Modal identification by curve-fiting of power spectra

This paper proposes a parametric identification method, which estimates a modal transfer function by curve-fitting power spectra of ambient data in a wide-area monitoring system (WAMS). The power spectra are calculated through autocorrelation of a difference sequence between two sets of frequency data. The modal parameters are calculated from the estimated transfer function. The proposed method is compared to the modified extended Yule Walker (MEYW) method in simulations on signal-to-noise ratio (SNR).