Whither dynamic state estimation

Abstract This paper aims to present some feasible directions along which investigations on dynamic state estimation have been carried out and could be developed in the future. It is shown that the benefits which could be encountered from dynamic state estimation are linked to its predictive ability which provides the necessary information to perform preventive analysis and control. Other benefits are improvements in observability analysis, identification of bad data and detection of topology errors. In practice, however, dynamic state estimation is faced with two problems: modelling of the system dynamics and algorithmic tractability within real-time requirements. To overcome these difficulties, two dynamic estimation schemes can be considered. The one combines a short term nodal load forecasting technique to model the system dynamics with a hierarchical extended Kalman filter. In the other, more pragmatic scheme, the Kalman filtering process is replaced by a static estimation algorithm.