AC network-constrained unit commitment via conic relaxation and convex programming

To address the AC network-constrained unit commitment problem, we propose an effective three-step solution approach. We first solve a DC (mixed-integer linear) approximation of the original problem to identify an initial solution and the set of transmission lines likely to be congested. Secondly, with the linear solution as a starting point, we solve a mixed-integer second-order conic relaxation of the original problem using an active set strategy regarding transmission constraints. Finally, using the solution of the relaxed problem as a starting point, we solve a set of increasingly accurate continuous convex optimization problems to ensure full AC feasibility. We show the effectiveness of the proposed approach using the IEEE 24-bus test system and an Illinois 200-bus system with 49 generators and 245 lines.

Automated summary

An effective three-step solution approach is proposed to address the AC network-constrained unit commitment problem, including DC approximation, second-order conic relaxation, and continuous convex optimization to ensure full AC feasibility, which is validated using test systems with multiple generators and transmission lines.