Real-time cooperative yaw control with informer-based wind predicting for wind power generation enhancement

Cooperative yaw control is an advanced strategy designed to improve the overall power generation efficiency of wind farms. Previous studies on cooperative wind farm control often optimize the yaw angle using known static incoming flows, disregarding the rapid fluctuation of wind direction over time and the slow speed of wind turbine yaw rotation. This approach results in delayed yaw action and discrepancies between optimal and actual yaw angles. To address this, a real-time cooperative control strategy is proposed, taking into account short-term wind prediction to mitigate the delay effect of yawing and maximize power production. This strategy utilizes the Informer model and Bayesian optimization (BO) model for short-term wind prediction and control optimization, and investigates the impact of prediction accuracy, yaw rotation strategy, and control time interval on the performance of cooperative yaw control. The research findings indicate that the proposed method effectively improves the overall power output of the wind farm compared to traditional greedy control strategies. It is noteworthy that the cooperative yaw control method shows greater sensitivity to wind direction prediction accuracy. Increasing the yaw rotation speed before reaching 0.6 °/s significantly enhances the optimization effect of cooperative control. The most economical control strategy is the centered yaw control with a control interval of 60 s, resulting in a power optimization improvement of up to 6.35 %.