Requirements on Super-Short-Term Wind Speed Predictions for Model Predictive Wind Turbine Control

Modern wind turbine control has a great demand for advanced control strategies that deal with several, divergent objectives, such as load alleviation and power leveling, simultaneously. Model predictive control can handle such multi-variable control problems and enables the use of preview wind speed information to improve the control performance. In this work, we address the inclusion of available wind speed predictions in the control formulation and derive specifications for wind speed prediction trajectories. To this end, we propose a linear time variant model predictive control formulation and investigate the influence of the wind speed measurement accuracy and the impact of different prediction update times on the control performance. According to our simulations, a wind speed measurement accuracy < 5% seems to be appropriate for model predictive wind turbine control using the persistence method for wind speed prediction. Prediction updates within the prediction horizon can improve the control results but may require a retuning of the controller parametrization.