Detection of lightly damped inter-area power oscillations using extended complex Kalman Filter

Detecting unstable inter-area oscillations in electrical power systems is very essential for designing appropriate generator damping strategies. Techniques like Prony analysis and Kalman filter have been proposed to extract oscillation parameters. However, Prony analysis finds it difficult to detect time-varying parameters and the performance of Kalman Filter, which is an adaptive tracking method, is dependent on initial conditions. Therefore, in this paper, Extended Complex Kalman Filter (ECKF) is investigated to adaptively detect multiple poorly damped power system oscillations under noisy conditions. Unlike Kalman Filter, the proposed technique is nonlinear and uses Hankel Singular Value Decomposition (HSVD) to accurately estimate its initial conditions. Thus, the convergence rate can be significantly improved. Additionally, the HSVD-based ECKF is capable of approximating multiple damping factors as well as the standard frequency parameters.