AMPS: Advanced Mathematical Algorithms for Model Reduction and Stochastic Modeling for the Emerging Power Grid

Southern Methodist University, Dallas TX

Investigators

Abstract

This project will develop tools to analyze and help design robust and resilient modern power grid systems. The emergent power grid is very different from the traditional grid, which is based on technology dating back to the early 20'th century. With new technologies and regulatory policies, new challenges are arising in the design of stable power systems that can reliably deliver electric power. To tackle these challenges, mathematical and computational tools are needed to analyze these power grids with the new features. The goal of this project is to develop these mathematical and computational tools. Specifically, this project will examine model reduction, uncertainty/stochasticity, and stability, which are some of the new features that must be incorporated in the models of the modern grid. Stability will be the connecting theme in these topics since stability is required to ensure a resilient grid with reduced risks. The PI will examine (1) how synchrony in the oscillations of the grid can be used to determine coherent sets of generators and loads for accurate model reduction; (2) what are the relevant structures that must be reserved to produce accurate reduced models; (3) how to develop multilevel solvers for decentralized systems; (4) how stochasticity affects the synchrony, coherency, model reduction, and stability of the grid; and (5) how observed data can be assimilated into the stochastic models. The stochasticity will be incorporated using stochastic differential-algebraic equations. This stochasticity will be introduced using the Orstein-Uhlenbeck process that best describes the stochasticity in the components of the system.