Robust Topology Control of AC Power Networks With Injection Uncertainties

Integration of Renewable Energy Sources (RES) prompts a need for increased grid flexibility. While Optimal Transmission Switching (OTS) harnesses this flexibility, deterministic OTS fails to provide optimal topology for injection uncertainties from loads and RESs. This paper presents a novel two-level, three-stage formulation for robust AC OTS (ACOTS), addressing injection uncertainties in the day-ahead operation of power networks. Unlike previous formulations, the proposed model has no approximations of power flow equations. The master problem is a Mixed Integer Linear Programming (MILP) formulation that determines robust optimal network topology, serving as input to slave problems. The slave problems are solved sequentially, where the first slave is a deterministic AC Optimal Power Flow (ACOPF) problem, and the second slave unveils worst-case uncertainty, thereby co-optimizing dispatch in the considered uncertainty range. The proposed formulation is solved by Generalized Benders decomposition, which is devised with an accelerated convergence technique. Consideration of budget of uncertainty constraints reduces conservatism of optimal topology. Validation through Monte-Carlo Simulation-based ACOPF, appended with load curtailment, demonstrates the setpoints' robustness on NESTA and South Carolina 500 bus systems. Comparison with previous approaches and the Column-and-Constraint Generation Generation (C&CG) algorithm underscores the proposed model's effectiveness in achieving a robust optimal network topology.