Prediction-based Wind Turbine Operation for Active Participation in the Day-Ahead and Reserve Markets

Electricity markets around the world are opening up to a greater contribution from wind power producers (WPPs). In this regard, WPPs are incentivised to participate in both energy and reserve market floors while being responsible for real-time deviations from their submitted bids. Therefore, despite uncertainties in wind speed and system frequency, effective control systems should be developed to enable WPPs to respond reliably concerning their committed day-ahead bids, as flexible conventional power plants do. However, designing a control system for WPP to regulate their capacity margin and output power as per the offered reserve bid is challenging, as a fast response with respect to the offered balancing service is required. This paper proposes an effective control system that allows WPP to regulate their set-points so as to provide the committed reserve power while considering the real-time wind variations. A machine-learning algorithm based on the Adaptive Neuro-Fuzzy Inference System (ANFIS) is used to predict the wind speed of the following instances, to be used as input to the control system. Several wind profiles are generated to simulate a practical case study, including real and predicted cases with varying levels of turbulence. Finally, the effectiveness of proposed control strategies on the WPP's profit is evaluated.