Constrained Iterated Unscented Kalman Filter for Dynamic State and Parameter Estimation

Abstract

This paper presents a robust dynamic state estimator for the synchronous generators with unknown parameters. The estimator uses a constrained iterated unscented Kalman filter to estimate the state variables and unknown parameters of a two-axis model of a synchronous generator. The developed estimator's performance is validated using simulations, where the estimator is subjected to arbitrary initialization and large parameter errors. The developed dynamic estimator can potentially be used not only to track the dynamic states but also to detect and identify changes in model parameters with little a priori knowledge about the parameters other than a broad range which can be specified via appropriate constraints.