A game-theoretic analysis of wind power generator bidding strategies in electricity markets

Abstract

Third, we provide a systematic analysis of optimal WPG biddind strategies in the presence of energy storage. Electricity demand and wind resources have negative temporal correlation. Thus large energy price fluctuation over time can be expected with significant wind energy penetration. In this study, energy storage is applied for inter-temporal arbitrage to help increase WPG's revenue. In addition, the competition games between WPG and energy storage are also investigated.

Second, we investigate optimal wind power generator (WPG) bidding strategies in electricity markets by incorporating wind bidding model into the UC problems. Assuming the cost of wind power generation is zero, WPG would maximize its profit by determining the optimal amount of wind power to bid in the day-ahead market (DAM). Significant wind penetration is considered in our problems such that WPG’s actions will affect energy prices. In addition, the existence of WPG’s optimal bidding strategies in both cooperative games and non-cooperative games is shown. The proposed stochastic optimization aspects consist of the DA energy market clearing, the real-time (RT) energy market clearing, and WPG’s profit maximization. The optimization formulation also takes into account the uncertainty of wind power production. The models have been implemented using MATLAB and tested on a system with 10 conventional generators and 2 wind power generators. The proposed two-stage optimization framework is used in the solution. The effect of wind power forecast errors is also discussed.

Global renewable energy generation has been growing rapidly. The current trend on using renewable energy resources, especially wind, is mainly motivated by the increasing cost competitiveness of renewable resources and climate concerns. Most of wind energy transactions are traded with bi-lateral contracts and do not participate in electricity markets. Some Independent System Operators (ISO) such as NYISO have required wind energy to bid into real-time market (RTM), but wind energy bidding in the day-ahead market (DAM) is optional. This dissertation provides a systematic investigation of the impacts and benefits of wind energy participation in electricity markets.

In order to quantitatively study the optimal wind bidding strategies and their market impacts, the mathematical model of the unit commitment (UC) problem is developed. Power system UC is a nonconvex NP-complete problem, which is very complex to solve for a large system. Solution methods for the UC problem have been explored, with Lagrangian relaxation (LR) being one of the most popular approaches in practice. The significant reduction in numerical solution time of commercial Mixed Integer Programming (MIP) solvers makes transitioning from LR to MIP possible. The first part of this dissertation presents a MIP based two-stage optimization approach for solving the UC problem as well as the energy prices, which has served as the simulation platform for the analysis of wind bidding strategies. The proposed framework has been implemented using MATLAB and tested on two test power systems.