Location

Coda C1215 Midtown (meeting link)

Title: Advances in Large-Scale Security-Constrained Economic Dispatch: Loss Modeling, Stochastic Dispatch, and Proxy Verification
Date: April 16th, 2025
Time: 1:00 - 3:00 pm ET
Location: Coda C1215 Midtown (meeting link)

Haoruo Zhao
Operations Research PhD Student
H. Milton Stewart School of Industrial and Systems Engineering
Georgia Institute of Technology

Committee:
Dr. Pascal Van Hentenryck (Advisor), H. Milton Stewart School of Industrial and Systems Engineering, Georgia Institute of Technology
Dr. Constance Crozier, H. Milton Stewart School of Industrial and Systems Engineering, Georgia Institute of Technology
Dr. Mathieu Dahan, H. Milton Stewart School of Industrial and Systems Engineering, Georgia Institute of Technology
Dr. Daniel Molzahn, School of Electrical and Computer Engineering, Georgia Institute of Technology
Dr. Hassan Hijazi, Gurobi

Abstract:
Security-Constrained Economic Dispatch (SCED) optimizes generation resources to meet demand at minimal cost while maintaining grid reliability. This thesis presents novel optimization techniques for large-scale SCED problems that improve computational tractability and solution quality for modern power systems. Chapter 1 introduces a novel linear model for line loss outer approximation (LLOA) in DC optimal power flow, a critical component of SCED that models network constraints, providing an efficient and practical approach that balances accuracy and computational tractability for large-scale power systems. Chapter 2 develops a stochastic look-ahead dispatch (SLAD) framework for real-time energy markets, demonstrating that stochastic optimization is now computationally viable for five-minute market clearing, offering greater savings compared to the traditional deterministic formulation with flexible ramping products. Chapter 3 presents an optimization-based method for bound tightening in a rolling-horizon fashion for neural network verification and Chapter 4 introduces a compact formulation for optimality verification of optimization proxies for DC optimal power flow and knapsack problems.