Abstract

The electrical quantities of future power networks are expected to exhibit strong fluctuations caused by dynamic bidirectional energy flows transferred from/to a multitude of "prosumers”. Such variations have to be accurately measured in real-time either for efficient power distribution or for safety and protection purposes. This task can be accomplished by the Phasor Measurement Units (PMUs), which measure the phasor of voltage or current waveforms synchronized to the Coordinated Universal Time (UTC). Accuracy of synchrophasor measurements is one of the many open challenges that need to be addressed in order to guarantee smart grid reliability and availability. Synchrophasor measurement has gained an undisputed relevance in the research community working on power delivery issues for various reasons. Among them, state estimation (SE) of both transmission and distribution networks is one of the most important. Within this general context, this dissertation covers two complementary topics. In the first part, starting from the concept of synchrophasor and from the definition of the parameters to evaluate PMU performances, useful guidelines to design a filter-based synchrophasor estimator are provided. Afterwards, an extensive performance comparison of some state-of-the-art synchrophasor estimation algorithms is reported in most of the static and dynamic conditions described in the IEEE Standards C37.118.1-2011. Also, a novel technique able to address both static and dynamic disturbances is presented and analyzed in depth. In this respect, special attention is devoted to phasor angle estimation accuracy, which is particularly important for active distribution networks. The second part of the dissertation is focused on the role and the impact of PMUs for grid state estimation. After recalling the state estimation problem and the traditional Weighted Least Square (WLS) technique to solve it, a general uncertainty sensitivity analysis to different types of measurements is introduced and justified both theoretically and through simulations. Afterwards, the effect of a growing number of PMUs on WLS-based state estimation uncertainty is evaluated as a function of instrumental accuracy and line parameters tolerance. Finally, a Bayesian linear state estimator (BLSE) based on a linear approximation of power flow equations for distribution networks is presented. The main advantage of BLSE is that in most cases it is so accurate as the WLS technique, but it is computationally lighter, faster and more stable from the numerical point of view.