This project focuses on predictive modeling of complex systems such as the power grid. The research requires capturing the physics and component interactions across vastly different temporal and spatial scales. Typically, only a subset of these interactions is captured, leading to errors in model predictions that can result in inefficient operation or system overdesign.

To overcome this problem, this project aims to develop novel mathematically rigorous, efficient, scalable, and robust numerical strategies for identifying and representing model errors in complex predictive simulations. The work is considered essential for representing uncertainty in dynamical simulations important to DOE.