Distribution system state estimation: monitoring and characterization of active distribution networks

The Data Science Discipline of the School of EECS is hosting the following guest seminar:

Distribution system state estimation: monitoring and characterization of active distribution networks

Presenter: Dr Marta Vanin (KU Leuven)
Host: Archie Chapman

NB: Martha and Fred Geth (from GridQube) will also be visiting for the day and anyone is welcome to join us for lunch and/or meetings and discussions.

Abstract: Advanced physics-based tools have been developed that can model and simulate distribution networks (DNs) in increasing detail, avoiding shortcuts like positive sequence analyses. These tools allow to build modern grid management solutions that are necessary for the energy transition. However, to exploit their full potential in practical settings (i.e., real-life DNs), accurate network data are needed as input, including consumer phase connectivity and cable impedances. 

Unfortunately, such information is usually poorly known, because of legacy issues in DN data acquisition and logging. The increasing amount of smart meters provides invaluable data that can be leveraged to estimate and identify the network properties needed for accurate digital network models. However, these measurements are often suboptimal: they are averaged and noisy, and only collected at limited network buses for limited time. This presentation illustrates a state estimation framework that can be extended to derive network asset properties, jointly to the “usual” state variables (i.e., voltages at all buses). The underlying use of state estimation implies that system identification is physics-aware, as the power flow equations are included in the estimation/identification process. This has several benefits that can compensate for the limited availability of measurements, and their noise. The presentation shows results both on synthetic data and real network+smart meter data from a Belgian system operator.

Bio: Marta Vanin received a joint MSc degree in Energy Engineering from the University of Trento and the Free University of Bolzano, in Italy. After that, she joined the electrical engineering department of KU Leuven (Belgium), where she got her PhD in 2022 under the supervision of Prof. Dirk Van Hertem and Dr. Reinhilde D’hulst from VITO (Flemish research agency). Since January 2023, she is a postdoctoral researcher at the same department. She will be in Australia for a research visit with CSIRO and GridQube until the end of May. Her main research focus is the development of optimization methods that leverage (smart meter) measurements to 1) improve digital network models, and 2) optimally manage and operate such networks.