Uncertainty Quantification for Cascading Failures in Power Systems
Abstract: Cascading failures in power grids are chain reactions that result in large-scale blackouts with significant economic and societal costs. Understanding the mechanisms that drive them and quantifying their probability of occurrence is therefore critical for ensuring grid reliability and developing effective mitigation strategies. This talk presents novel approaches for this challenge. First, I will introduce a statistical model that characterizes the initiation of cascades as a weather-driven process. Next, I will describe a reduced-order method that propagates uncertainty through the system’s dynamics by combining probability evolution equations with learned closures. Finally, I will discuss a software framework for performing sensitivity analysis and uncertainty quantification via adjoint methods.
Bio: Adrian Maldonado is an Assistant Scientist. He received his BS in Industrial Engineering from the Polytechnic University of Valencia and his Ph.D. in Electrical Engineering from the Illinois Institute of Technology.
See all upcoming talks at https://www.anl.gov/mcs/lans-seminars.