Dynamic State Estimation of Power Systems Based on Extended Kalman Particle Filter

To improve the estimation performance of the extended Kalman filter in highly nonlinear power systems and to tackle the particle degeneracy problem of the particle filter, this paper proposes a dynamic state estimation of power systems based on the extended Kalman particle filter. Based on the framework of the particle filter, the proposed method utilizes the extended Kalman filter to construct the importance sampling density function in the particle filter, and therefore effectively resolves the particle degeneracy problem, and consequently achieves high-precision power system state estimation. The proposed method is tested in the CEPRI-36 node system by comparing against the original particle filter and the extended Kalman filter, the results demonstrate the advantages of the proposed method in terms of measurement filtering metrics and root mean square error.