Fast and Accurate Estimation of Power System Frequency for Power Quality Studies

Abstract

This article presents a new approach for the estimation of power system frequency for power quality studies using bus voltage signals corrupted with noise and harmonics. The method involves application of a nonlinear filtering technique like a Kalman filter in the complex domain known as extended complex Kalman filter (ECKF). This article systematically outlines the development of the complex Kalman filter model using one phase measurement in the state space form and relevant equations to compute the Kalman gain. The nonlinear filter is then applied to several static and dynamic conditions of the power system and the frequency is estimated. Presence of harmonics and noise in bus voltage signal is taken into consideration in evaluating the performance of this new filter in terms of rejecting noise and harmonics. In the presence of decaying dc components in the signal, the new filter is found to be very effective as validated from the digital simulation tests. Further the ECKF is applied for the frequency estimation of an ac-dc system during load and line reactance changes and the estimation results reveal high accuracy and convergence. In general, it is observed that the proposed filter outperforms in terms of accuracy and speed of convergence the reported real extended Kalman filter during sudden transient changes.