Strategic bidding for price-maker producers in predominantly hydroelectric systems

• Published in: Electric power systems research 2016-11, Vol.140, p.435-444
• Main Authors: Cruz, M.P., Finardi, E.C., de Matos, V.L., Luna, J.P.
• Format: Article
• Language: English
• Subjects: Bi-level program; Electricity markets; Equilibrium model; Hydrothermal model; Mathematical program with equilibrium constraints; Multi-stage model
• ISSN: 0378-7796; 1873-2046
• DOI: 10.1016/j.epsr.2016.05.032

•An MPEC model to determine bidding strategies in predominantly hydroelectric systems.•Price-maker firms operating hydro and/or thermal plants.•Consideration of cascaded hydro system with reservoirs managed by different firms.•Competition modeled as an equilibrium problem. This paper proposes an equilibrium model to determine price and quantity strategic bids for generation companies participating in a day-ahead electricity market, with predominantly hydroelectric generation. Each agent aims to maximize its profit by solving a bi-level optimization problem, where the upper level represents the producer revenue maximization problem, while the lower one consists of minimizing the cost of system operation, faced by the independent system operator. Price-maker companies operating both hydro and thermal plants are considered in a cascade hydro system with reservoirs managed by different owners. Consequently, a novel approach is proposed to decouple the first and second levels of the problem. We present an individual plant modeling, where the main constraints related to a hydrothermal system are considered. Through the utilization of the Karush–Kuhn–Tucker optimality conditions, the bi-level optimization model is converted to a single level nonlinear problem, known in literature as a mathematical program with equilibrium constraints (MPEC). To face the difficulties of this nonlinear, nonconvex, multi-stage problem, the MPEC complementarity conditions are replaced by the strong duality condition. Moreover, competition among several leaders is modeled as an iterative procedure, and the methodology is applied in two systems with data and configurations derived from the Brazilian hydrothermal system.