Wind Turbine Power Optimization Based on Extreme Gradient Boosting Model and Periodic Adjustment Strategy

In recent years, with the rapid development of wind energy-related technologies, wind farms have been established in more and more countries and regions. However, some wind farms still face the problems such as low power generation and low wind energy utilization. To increase the profit of wind farms and improve the efficiency of wind turbines, we propose a power optimization method based on machine learning algorithm and periodic adjustment. In this study, the optimization process consists of two steps. First, a FF-PCA-XGBOOST model based on data from the supervisory control and data acquisition (SCADA) and Pearson correlation analysis is presented for power prediction. Principal component analysis (PCA) is used to reduce the dimensionality of the power data, and we use feature fusion (FF) to remain key features. Second, we propose a novel data-driven method, which uses power prediction information obtained from the prediction model to adjust pitch angle periodically for enhancing power output. We demonstrate that the optimization based on our model increases the energy production of the wind plant.