Initial results in the detection and estimation of forced oscillations in power systems

With the phasor measurement unit (PMU) networks that have been developed to improve utilities' wide-area situational awareness, it is now possible to obtain high-fidelity observations of forced oscillations (FOs). Due to the large variation in their amplitudes, reliable methods for detecting forced oscillations and estimating their parameters are needed. In this paper, an algorithm designed for the detection and estimation of sinusoids in white Gaussian noise is modified for the colored noise environment of power systems. A novel time-localization algorithm based on the cross-correlation estimator is proposed to estimate the start and end points of a forced oscillation, and an estimator for the forced oscillation's amplitude dependent on these location estimates is developed. Using a simple model, the algorithms are shown to be capable of detecting a forced oscillation, estimating its frequency, locating it within the data, and estimating its amplitude.