A strengthened primal-dual decomposition algorithm for solving bilevel SCUC problem

IEEE Transactions on Energy Markets, Policy and Regulation

Management Science and Operations

IEEE Transactions on Energy Markets, Policy and Regulation 2024 Online first

Goudarzi H;Hemsamzadeh M R;Bunn D;Fotuhi-Firuzabad M

Derek Bunn

2024

Efficient nodal pricing models and short-term unit commitment planning face continuous needs for improvement as operational requirements evolve. This paper develops a Bilevel Security-Constrained Unit Commitment (BL-SCUC) model to include both revenue-adequacy and Fast Frequency Reserve (FFR) constraints. The upper level of the BL-SCUC model represents the non-convex UC decisions as well as the revenue-adequacy constraints of the market participants (generators, loads, and battery-storage owner). The lower level is a convex economic dispatch model which produces the nodal electricity prices. To solve the proposed BL-SCUC model, it is first reformulated as a single-level Mixed-Integer Linear Program (MILP) using the standard strong-duality approach. The resulting MILP model is hard to solve using standard off-the-shelf solvers such as Cplex, partly because the Big-M parameters’ optimal tuning for linearization in the strong duality method is NP-hard. To solve this, we propose a strengthened Primal-Dual Decomposition (PDD) algorithm, which takes benefit from both Benders-like and Lagrange Dual-like algorithms. The new PDD algorithm eliminates the Big-M parameters without affecting optimal values. Accordingly, the computational burden and optimal solution sensitivity resulting from Big-M parameters are mitigated. Results from the modified IEEE 24-bus system demonstrate the effectiveness of the proposed BL-SCUC model with its PDD algorithm, whilst results from the IEEE 118-bus system show the superiority of the proposed strengthened PDD algorithm over the classic Benders algorithm.

Disjunctive programming; Primal-dual decomposition; SCUC; Revenue adequacy; Fast reserve

No