Independent system operators (ISOs) currently use dashboards to track conditions of the electrical grid, including energy loads and the output of renewable and non-renewable energy. None of those dashboards, however, have the capacity to compute and analyze engineering and financial risks to the power grid.
“Power grid operations will need to be driven by new risk-aware methodologies and algorithms in order to properly address new developments, including renewable energy sources and grid-scale storage, as well as financial hedges in energy markets,” says the project’s principal investigator, Daniel Bienstock, Liu Family Professor of Industrial Engineering and Operations Research and professor of applied physics and applied mathematics.
He stresses that it is also critical that power grid participants–providers of generation, balancing capabilities, and storage, along with price-motivated consumers–come to see risk as an essential attribute to be modeled, priced, and accurately accounted for in operational planning.
The researchers plan to create a dashboard that will help ISOs make sound decisions quickly, in time scales ranging from several minutes to a few days, so that they can dispatch available energy to the places most in need of it at any given moment. It will also guide ISOs and market participants towards a better understanding of renewable energy risks, and how to adequately allocate resources for reducing them. In doing so, the risk-assessing dashboard will make the U.S. power grid more efficient and reliable, and facilitate penetration of renewables.
“Our reliance on risk management techniques constitutes a fundamental shift in the practice of power system modeling,” says co-PI Agostino Capponi, an associate professor of industrial engineering and operations research. “We are challenging existing practices by bringing operations research methods based on optimization, financial risk analytics, and modern data science techniques into the hard-core engineering domain of power engineering.”
The dashboard will rely on stress-testing statistical factor models and newly designed energy asset and systemic risk metrics to continuously track system and market conditions and proactively dispatch available resources to avoid insecure operations.
“To reconcile technical risks identified using these risk metrics, we will leverage financial risk management instruments to construct hedges against losses under severe or complex energy scenarios,” notes co-PI Garud N. Iyengar, Tang Family Professor of Industrial Engineering and Operations Research. “We are aiming to inform ISOs how to use renewable, demand, and storage resources for reducing such risk and compensate these resources accordingly. Our second goal is to educate ISOs on the risk versus cost minimization tradeoff to facilitate increased penetration by renewable resources.”
The research team includes Bienstock, Capponi, and Iyengar, all at Columbia Engineering and affiliates of the Data Science Institute; Yury Dvorkin, an assistant professor of electrical and computer engineering at New York University; and Michael Cherktkov, a professor of applied mathematics at the University of Arizona.
The Columbia Engineering project is one of 10 recently announced by the DOE, which is giving $25M to PERFORM teams to develop innovative management systems that represent the relative delivery risk of each asset, like wind farms or power plants, and balance the collective risk of all assets across the grid.
“Ensuring the reliability of our nation’s critical energy infrastructure and electric grid is of the utmost importance to America’s energy security and national security,” Under Secretary of Energy Mark W. Menezes said. “Investing in new technologies and systems that minimize risk and bolster the reliability of U.S. energy will allow us to utilize all of our abundant energy resources in a more integrated and secure manner.”
The modern grid relies on conventional, bulk power plants to provide the flexibility to operate power systems reliably. In general, these assets can guarantee available capacity, except in rare events like the sudden loss of power. The existing risk management strategy protects against those rare events, and aligns well with conventional technologies. There is a need for this strategy to be reassessed due to new technologies and a shift in the power generation mix. That shift includes an increase in intermittent renewable resources, distributed energy resources, and storage technologies. As emerging technologies are deployed, management systems must be able to leverage all capabilities of these new technologies to maintain a cost-effective reliable grid. PERFORM projects address that need by developing methods to quantify and manage risk at the asset level and at the system level for the grid.