Robust Adaptive Fading Unscented Kalman Filter for Decentralized Dynamic State Estimation in Power Systems Under Unknown Inputs and Covariance Mismatches

Dynamic state estimation (DSE) of synchronous machines is crucial to the monitoring, protection, and control of power systems. Bad data due to outliers and model uncertainties can affect significantly its accuracy. This paper proposes a robust adaptive fading (AF) unscented Kalman filter (UKF) for DSE and estimation of possible unknown inputs due to unmeasured input quantities under bad data. It utilizes the AF concept to minimize possible scale mismatches in the state and measurement noise covariance matrices of the KF to mitigate these uncertainties. A simple trace operation-based and a least squares-based approaches are proposed for estimating the fading factors, which are further tracked using a low order KF or a lower complexity recursive least squares algorithm. A robust statistics-based extension of the AF-UKF is also developed to effectively detect and suppress bad data. The stability of the proposed robust AF-UKF is studied. Its performance was compared with conventional algorithms on the Northeastern Power Coordinating Council 48-machine 140-bus and a 16-machine 68-bus Power System. Simulation results suggest that the proposed decentralized DSE algorithms yield more accurate performance than conventional methods under bad-data and noise covariance mismatches. It also yields more accurate estimation of the unknown input than conventional methods tested.