Second-Order Stochastic Dominance Constraints for Risk Management of a Wind Power Producer's Optimal Bidding Strategy

Abstract

Risk management is critical for wind producers to participate in electricity markets. Beside market price volatility and uncertainty, wind producers are facing an additional uncertainty in the level of wind power generation. Instead of using common risk measures, such as conditional value at risk (CVaR), this paper proposes the use of the second-order stochastic dominance constraints (SOSDCs) for risk management of wind producer's bidding strategies. As benchmark selection is the major obstacle against applying SOSDCs, a novel optimization-based benchmark selection method is proposed. Case studies are carried out for an 80 MW wind producer using the SOSDCs-based bidding model with the proposed benchmark selection method and the CVaR-based bidding model. Results demonstrate the superior flexibility of the SOSDCs in risk management. Moreover, the SOSDCs can effectively manage the negative tail of the profit distribution. Compared to the SOSDCs, the CVaR is more suitable for modeling risk rather than managing risk, as it does not use a profit target value but uses the $({1 - \boldsymbol {\alpha }})$ -quantile of the profit distribution. As the negative tail is the best representative of risk in the problem under study, the SOSDCs with the proposed benchmark selection method are more suitable than the CVaR for risk management of a wind power producer's bidding strategy.