Power System State Fusion Estimation Based on Maximum Correntropy Criterion

By using the observations from multiple sensors, the fusion estimation of power system states, such as voltage, current, phase angle and other variables can be realized. When the observation information is affected by non-Gaussian noises, especially by some heavy tailed impulse noise, the performance of the fusion filter estimation based on Kalman filter will decline. Based on maximum correntropy criterion (MCC) and combining two fusion methods of parallel fusion and sequential fusion, this paper studies two power system state fusion estimation methods that can be used in non-Gaussian noise environment, named maximum correntropy parallel Kalman fusion filter (MC-PKFF) and maximum correntropy sequential Kalman fusion filter (MC-SKFF) respectively. The validity and the performance comparison of the proposed fusion algorithms are verified by the simulation example of three-phase voltage state estimation in power system.