Impact of high levels of wind penetration on the exercises of market power in the multi-area systems

Abstract

New European energy policies have set a goal of a high share of renewable energy in electricity markets. In the presence of high levels of renewable generation, and especially wind, there is more uncertainty in the supply, which induces volatility in energy prices. This can create incentives for the generators to exercise market power by traditional means: withholding the output, bidding not the true marginal costs, or using locational market power. In addition, a new type of market power has been recently observed: exercise of market power on ramp rate.

This dissertation focuses on modeling the exercise of market power in power systems with high penetration of wind power. Chapter 2 provides the mathematical foundations for the dissertation, including the relevant concepts from game theory, optimization, and stochastic programming. Chapter 3 gives a brief description of power system modeling conventions and assumptions. Chapter 4, based on publication [J1], reviews exercise of market power in wind-integrated systems and identifies flexibility as one of the drivers for strategic behavior. Hydropower producers are main providers of flexibility in systems, where hydropower is available. Therefore, Chapter 5 focuses on the exercise of market power in hydro-dominated power systems with high share of wind power. The chapter is based on publications [J3] and [J4]. Chapter 6 reviews market power from the market design perspective. Two market design possibilities are compared and the impact of each of the design on the propensity of strategic generators to exercise market power in wind-integrated systems is discussed. The chapter is based on publication [J2]. Finally, Chapter 7 concludes the dissertation and provides the possible directions for future research.

Developed models are formulated as mathematical and equilibrium problems with equilibrium constraints (MPECs and EPECs). The models are recast as mixed-integer linear programs (MILPs) using discretization. Resulting MILPs can be solved directly by commercially-available MILP solvers, or by applying decomposition. Proposed Modified Benders Decomposition Algorithm (MBDA) significantly improves the computational efficiency.