Effective Wind Speed Estimation and Prediction Based Feedforward Feedback Pitch Control for Wind Turbines

Although the wind power generation technology has gained tremendous development in recent years, how to reduce important component loads of wind turbines is still a hot topic for the researchers in control communities due to the randomness and intermittence of wind speed. In this study, an effective wind speed estimation (EWSE) and prediction based feedforward feedback pitch control scheme for wind turbines is presented. First, both EWSE and prediction models are developed through support vector regression (SVR), the training of which is conducted by utilizing data acquired through the existing supervisory control and data acquisition (SCADA) systems. Further, optimization of SVR parameters is completed by the particle swarm optimization (PSO) algorithm, and smoothness of the estimated wind speeds is guaranteed by employing a low-pass filter. Thereafter, a proportional-integral-differential (PID) based feedforward feedback control scheme is designed to obtain the control goals of generator power regulation as well as loads mitigation by combining the EWSE and prediction models. Finally, comparative simulations between the novel feedforward feedback control scheme and traditional pure feedback controller are conducted on GH bladed platform to show the effectiveness of the proposed scheme.