Dynamic Synchrophasor Estimation by Extended Kalman Filter

Fast synchronized measurements of the phasor, frequency, and rate of change of frequency are expected to be very important for the automated control actions in the smart grid context. In this regard, measurement latency must be kept as short as possible for an effective control implementation when networks characterized by extremely fast dynamics are concerned. Kalman filter (KF)-based estimation algorithms appear to be attractive in this context; however, the conventional implementations suffer from significant limitations in their ability to deal with different types of dynamic conditions due to approximations in the model and in the associated uncertainty. This article proposes an innovative solution, based on an extended KF algorithm using a Taylor model, which is shown to provide improved tracking ability in a vast range of dynamic conditions. A novel element in the proposed technique is the representation of model uncertainty, which takes into account the intrinsic correlation among errors that appear in the state–space description under dynamic conditions. A compatibility check between the forecast and measurement result is also introduced as an effective and metrologically sound approach to detect large unexpected changes in the tracked parameters in order to achieve a fast response of the algorithm also under those conditions. The performance of the algorithm is thoroughly investigated by means of simulation to demonstrate the significant improvement compared to other KF solutions in some conditions of practical relevance.