AMPS: Risk-Averse Optimization for Stressed Power Grids: Models, Theory, and Algorithms

George Washington University, Washington DC

Investigators

Abstract

The long-term reliability of the national power grid is threatened by the elevated incidence and severity of outage-inducing extreme events, such as wildfires. In order to hedge against the consequences of wildfire incidents in a timely and effective manner, guidance and support should be provided to power system operators to build in system resilience and to account for high unpredictability driven by the underlying environmental stressors, those inherent to electric power grids, and those resulting from human decisions. It is vital to observe the uncertainty that cannot be resolved simply with the passage of time; rather, its resolution depends on whether decisions to take preventive measures against wildfires are actually made and how they affect the resolution of uncertainty, i.e., decision-dependent uncertainty (DDU). This calls for advances in new, fast, and efficient mathematical optimization approaches that can adapt to differing levels of risk and emergency response times and capture a wide spectrum of wildfire scenarios. This project involves graduate students. To address these imperatives, the principal objective of this project is to devise a mathematical optimization framework for risk-averse stochastic programming problems that account for and respond to DDUs in which the probability distributions of random variables are distorted by decisions. The mathematical advances of this project are (i) the development of novel optimization models that feature customizable risk aversion levels and that reflect the impact of DDU on power system’s resilience against wildfires, and (ii) the design of scalable reformulation, algorithmic, and solution methods to solve these complex optimization problems. In essence, the outputs of this research will provide electric power grids with optimization and decision-making tools to effectively respond to wildfire threats while minimizing the required level of investments in capacity expansion and technology enhancements. 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.