Unit Commitment with Gas Network Awareness

Geunyeong Byeon, University of Michigan

This study introduces gas network awareness into an unit commitment model. A tri-level optimization problem is proposed, in which natural gas zonal prices are estimated by the dual solutions of the gas flux conservation constraint and commitment decisions are subject to the gas zonal prices for ensuring the economic viability of the committed gas-fired power plants. This tri-level problem is reformulated as a single-level Mixed-Integer Second-Order Cone program (MISOC) which can then be solved using a Benders decomposition method. We developed a Benders subproblem decomposition technique that addresses the computational challenge in solving the MISOC, which is applicable to a broader class of optimization problems: discrete-continuous bilevel optimization. The model and the solution approach are validated on a case study, the results of which show that the model mitigates risks of volatile gas prices in electricity systems operation and demonstrate the computational benefits of the solution method. 

This seminar will be streamed.  https://anlpress.cels.anl.gov/cels-seminars/