Projected unscented Kalman filter for dynamic state estimation and bad data detection in power system

Dynamic state estimation is a useful tool for bad data detection and identification. However so far no dynamic state estimator considers the state constraints like zeros injection constraints in power system. In this paper, a new dynamic state estimator based on the projected unscented Kalman filter (PUKF) is presented. It is based on the application of unscented transformation and estimate-projection combined with the Kalman filter theory. The proposed method can improve the approximation of power system nonlinearity and guarantee that the estimated state vectors obey zeros injection constraints. Four tests for different conditions including normal operating condition, gross bad data condition, sudden load change condition and topology error condition are taken to verify the performance of the proposed method.