Synchronous Generator Parameters Estimation

What is it about?

This paper compares performance of extended Kalman filter (EKF), unscented Kalman filter (UKF), and least-squares (LS) algorithms in parameters estimation of synchronous generator (SG) using phasor measurement unit (PMU) data, where it is illustrated that the LS algorithm can be a better choice for SG parameters estimation. Moreover, it is shown that the LS algorithm is incapable of estimating the SG parameters accurately when the initial values of an SG model state variables are inaccurate. Hence, a modified LS (MLS) method, estimating the initial values of the state variables in addition to the parameters, is proposed. Using the proposed MLS method, parameters of SGs can be determined with acceptable accuracy even if the valid initial values of the state variables are inaccessible.

Featured Image

Why is it important?

The reasons this work is important are: - To compare performance of well-known methods such as EKF, UKF, and LS in SG parameters estimation using PMU data under various scenarios; - To resolve the problem of dependence of the LS algorithm on the state variables initial values in SG parameters estimation by proposing an MLS algorithm; - To estimate whole electromagnetic parameters and rotor inertia constant of an SG full-order model, in loaded condition and without a need for rotor angle measurement that is hard to be obtained; and - To study the impact of mechanical torque signal unavailability on accuracy of the proposed algorithm.