부산대학교 도서관

The statistics of critical state variables and nodal fluctuations in power grids are useful for control of its dynamics and procurement of regulation resources. This paper studies estimation of the statistics of grid state variables and ambient fluctuations using time-stamped nodal measurements collected from limited PMUs in the grid. We show that in the presence of time-stamped observations that enable computation of delayed covariances, PMUs located at half of the number of grid buses are sufficient to reconstruct all state and noise statistics exactly. For lower number of available PMUs, we provide a convex optimization framework for estimating the covariance matrices that outshines standard schemes that do not utilize the availability of time-stamped measurements.