Dynamic State Estimation of Generator Using PMU Data with Unknown Inputs

In this paper, a Kalman filter and particle filter based dynamic state estimation method are proposed for nonlinear systems with unknown inputs. In the proposed method, the dynamic states of a generation system are estimated in three stages. At the first stage, the biased states are predicted using unscented transform without unknown inputs. At the second stage, the unknown inputs are estimated using a particle filter technique with phasor measurements and the predicted biased states. At the final stage, the unbiased states are estimated using an unscented Kalman filter method with the estimated unknown inputs. The proposed algorithm is implemented in Korean power system model, and is compared with dynamic state estimation performances of other estimation algorithms with unknown inputs.