2020 PES GM 8/4 Panel Video: Learning to Model, Monitor, and Control Power Distribution Systems

The existing modeling, monitoring, and control capabilities of contemporary power distribution systems need to be enhanced to better integrate distributed energy resources and improve their reliability, security, and efficiency. The effectiveness of conventional model-based monitoring, control, and optimization approaches for power distribution grids are oftentimes hampered by inaccurate and incomplete feeder models, and/or obsolete knowledge of current grid conditions. Recent developments in data-driven approaches for topology identification, load and solar monitoring, grid optimization, and smart inverter control, have demonstrated great promise in addressing the aforesaid grid operation challenges. After judiciously adopting machine learning methodologies to the distribution grid domain of expertise, this panel puts forth an array of algorithmic solutions for data-driven distribution grid learning, optimization, and control, and discusses the technical challenges and opportunities involved.