Modeling of wind turbine power curve based on Gaussian process

For wind farms, the relationship between wind speed and output can be described by power curve of wind turbines, and it is an important embodiment of power performance of wind turbines. Based on the mathematical model of the power curve of wind turbine, monitoring performance of the wind turbine can be designed. Power curve model of wind turbines can be established by using Gaussian process. Within the Bayesian context, the paper aims to train the Gaussian process by using the maximum likelihood optimized approach to find the optimal hyperparameters. The model was validated by the data. Finally, based on the wind turbine power curve mathematical model, the states of the wind turbine can be monitored by using the technology of control charts.