M-estimation based robust approach for hybrid dynamic state estimation in power systems

Abstract

The state estimation (SE) process in power systems estimates bus voltage magnitude and phase angles vital for operating the system securely and reliably. The power systems state estimation problem has been extensively solved through weighted least squares (WLS) based static approach that fails to track the system dynamics properly. Furthermore, those approaches are not inherently robust against outliers, yielding a separate bad data processing (BDP) technique. Popular Dynamic state estimation (DSE) schemes which mainly employ nonlinear Kalman filters like Extended Kalman Filter (EKF) and Unscented Kalman Filter (UKF), also suffer from providing a reasonable estimation of states in the presence of bad data. Although several generalized maximum (GM) likelihood-based DSE approaches are robust against outliers, they are mainly based on nonlinear Kalman filters, which yield an iterative process. Therefore, this article focuses on developing a robust DSE approach that gives a good estimation of states against outliers in a single iteration. This article aims to propose a robust hybrid non-iterative DSE approach that gives robust SE results in the presence of bad data. The proposed novel robust hybrid DSE (NRHDSE) approach combines the robust M-estimation with the original novel hybrid DSE (NHDSE) approach. The proposed scheme implements a suitable linear relationship between integrated hybrid measurements and complex states. The proposed method uses a linear measurement model and thus employs an optimal linear Kalman filter to correct or estimate states. The efficacy of the proposed approach has been demonstrated by applying it on IEEE 57, 118 bus test systems and one more extensive 246 Indian utility bus system namely, Northern Regional Power Grid (NRPG), and after that comparing it with the original NHDSE, and DSE methods based on traditional EKF and M-estimation based robust version of EKF (REKF). The simulation result demonstrates the superiority of the proposed approach. Obtained results clearly show the superiority of the proposed approach.