Abstract
Accurate modelling of wind speed-power curve of wind turbines plays an important role in wind power prediction, state detection and performance evaluation. While model selection is one of the keys to improve the accuracy of wind power curve (WPC) modelling. To improve the accuracy of models and accurately characterize the overall output behavior of wind turbines, a method of WPC modelling based on the mixed Richards model is proposed in this paper. By using the measured data of two wind fields, the method proposed in this paper is compared and verified with the sixth to ninth order polynomials and the four-parameter and five-parameter logistic function models based on the genetic least square method through five indicators include the root mean squared error, the coefficient of determination R2, the mean absolute percentage error, the improved Akaike information criterion and the Bayesian information criterion. Finally, based on the measured data of a wind field in Jiangsu Province, the two-fold mixed Richards model is used to predict the wind turbine power. The results show that the two-fold mixed Richards model is the optimal option with the highest fitting accuracy, effectively avoids the model’s over-fitting, and can accurately predict wind turbine output power.