The evolution of a low-emission energy system requires improvements in the automation and control of electrical systems to effectively manage the increasing complexity of the energy mix. In this context, it is essential to enhance the observability of the grid to account for the new dynamics introduced by renewable energy generators and new electrical loads.

This aspect is particularly important for distribution networks, which are still poorly monitored and are more affected by generation and load dynamics. Effectively observing the distribution grid thus requires a significant number of measurements, including synchronized measurement systems with high update frequencies to track the most significant system dynamics.

The observability of the grid generally relies on state estimation, which identifies the magnitude and phase values of phasors characterizing the operational condition of the grid. In recent years, dynamic estimators have been developed to address the complexities of distribution networks with the increasing presence of distributed generation.

Within this context, the objective of the three-year research activity, framed by this document, is the study, development, and experimentation of dynamic estimation methods for distribution networks, with particular attention to low-voltage networks. To achieve this goal, analyses of the state of the art have been conducted, and tools have been prepared to experiment with dynamic estimation methods at the Distributed Energy Resources Test Facility (DER-TF).

An estimator that uses both measurements and knowledge of dynamics can be defined as a dynamic state estimator. Dynamic estimators are based on the prediction and correction of the state using updated measurements at each estimation instant. It has been found that most methods used for state estimation in dynamic situations are based on variants of the Kalman filter, or in a few cases, on optimization and machine learning approaches. Therefore, a Kalman filter-based estimator has been identified, which has the potential to include all types of measurements, consider frequency-associated dynamics, and be adapted to different types of networks.

The Report is available on the Italian site