The U.S. Department of Energy will explore whether artificial intelligence could help electric grids handle power fluctuations, avoid failures, resist damage, and recover faster from major storms, cyberattacks, solar flares and other disruptions. A new project, called GRIP, for Grid Resilience and Intelligence Project, was awarded up to $6 million over three years on September 12 by the U.S. Department of Energy. GRIP is the first project to use artificial intelligence (AI) to help power grids deal with disturbances, says Sila Kiliccote, GRIP's principal investigator and director of the Grid Integration, Systems and Mobility lab at the SLAC National Accelerator Laboratory in Menlo Park, Calif. GRIP will develop algorithms to learn how power grids work by analyzing smart meter data, utility-scale SCADA (supervisory control and data acquisition) data, electric vehicle charging data, and even satellite and street-view imagery. "By looking at satellite and street-view imagery, we can see where vegetation is growing with respect to power lines, how long it takes to grow, and anticipate what the effects of high winds might have on that vegetation, such as pulling trees onto power lines during storms," Kiliccote says.

Growing demand is straining our existing electricity generation facilities and requires active participation of the utility and the consumers to achieve energy sustainability. One of the most effective and widely used ways to achieve this goal in the smart grid is demand response (DR), whereby consumers reduce their electricity consumption in response to a request sent from the utility whenever it anticipates a peak in demand. To successfully plan and implement demand response, the utility requires reliable estimate of reduced consumption during DR. This also helps in optimal selection of consumers and curtailment strategies during DR. While much work has been done on predicting normal consumption, reduced consumption prediction is an open problem that is under-studied. In this paper, we introduce and formalize the problem of reduced consumption prediction, and discuss the challenges associated with it. We also describe computational methods that use historical DR data as well as pre-DR conditions to make such predictions. Our experiments are conducted in the real-world setting of a university campus microgrid, and our preliminary results set the foundation for more detailed modeling.