Exploiting Physical and Dynamical Structures for Real-time Inference in Electric Power Systems

Arizona State University, Scottsdale AZ

Investigators

Abstract

This NSF project aims to improve the capabilities of modern electric power systems. Recent years have seen a dramatic increase in solar and wind power throughout the grid, including inside the distribution system. This increasing penetration of distributed energy resources (DERs), which will only accelerate over time, has tremendous benefits, but also brings challenges, as DERs are substantially different from traditional large-scale generators. This project brings transformative solutions to these challenges to improve the situational awareness of the DERs throughout the system, by leveraging advanced high-fidelity sensors as well as modern data science. The intellectual merits of the project include methods to extract useful information from even a small number of high-fidelity sensors; this information can be used to make rapid control decisions to improve the stability of the overall system without relying on traditional generators. The broader impacts of the project include mentoring of graduate students and postdocs, in addition to undergraduate researchers via the Arizona State University (ASU) Summer Undergraduate Research Initiative (SURI). This project studies three main problem areas: (i) estimation of the topology of distribution networks that connect the grid edge to the bulk network; (ii) faster detection and localization of forced oscillations, which are typically caused by disturbances or devices failures and can pose significant threats to power system operations; and (iii) learning the parameters of DER dynamics, including inertia and damping, so as to maintain refined dynamic models which can be used by system operators to assess and ensure system stability. The project addresses these challenges via the unifying framework of structure: the graphical structure of grid topology, sparsity in the location of an oscillation source, and structure in dynamic models. The project exploits these structural elements to develop new algorithms to extract situational awareness from high-fidelity meters, especially Phasor Measurement Units (PMUs). Project outcomes include theoretical results as well as numerical experiments on the performance of these algorithms in realistic settings of power systems. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.