Optimal dispatch of wind farms facing market prices

At present, wind power producers (WPP) are paid following feed-in tariffs in Belgium. This system will come to an end soon due to its high cost and the producers will have to bid in the day-ahead market. As wind owners cannot forecast their production perfectly, they will face imbalance costs or revenues. Imbalance price forecasting is therefore a critical problem. In this paper, we implement a machine learning model to assess the usefulness of introducing exogenous variables in imbalance price forecasting. This method shows improved results compared to classical methods. Since the imbalance price is obtained by the marginal cost of producing the missing energy, the strategic behaviour of a WPP will influence the imbalance price. In this paper, we propose a way to represent this influence as well as a formulation of a model to obtain the optimal bidding strategy in that situation. This model has been cast as a convex quadratic program that can readily be solved using a commercial solver.