Wind direction prediction for yaw control of wind turbines

Depending on historical signals from wind direction sensors, conventional yaw control methods provide general performance and may be optimized by taking advantage of wind direction prediction. This paper presents two wind direction prediction methods based on time series models. The first method adopts a univariate ARIMA (auto-regressive integrated moving average) model, while the second one uses a hybrid model that integrates the ARIMA model into a Kalman Filter (KF). Since the predicted results are used to optimize yaw control of wind turbines, six prediction models are developed using three types of mean wind directions. Finally, industrial data is used to develop, validate and test the proposed models. From obtained results, it is shown that the hybrid models outperform other ones in terms of three performance indexes and different types of wind direction time series.