Nonlinear model predictive control to reduce pitch actuation of floating offshore wind turbines

Abstract

Modern-day wind turbines use active pitch control to reduce mechanical loads on the turbines in addition to regulating generator power. These control algorithms increase blade pitch actuation, primarily to reduce the 1P (once per revolution) component of the aerodynamic load. However, it is also known that the failure of the blade pitch system is a significant source of turbine downtime. Control algorithms that increase pitch actuation will only add to this problem. Therefore, increasing blade pitch actuation to reduce mechanical loads may not be the best solution is every situation. Hence, in this paper, an individual pitch control strategy is proposed to reduce pitch actuation without deteriorating rotor speed regulation or increasing structural vibrations. The controller is developed under a Non-linear Model Predictive Control (NMPC) framework. It is assumed that a preview of the inflow wind field is available in the form of LIDAR (LIght Detection And Ranging) wind speed measurements. The results presented in this paper show that it is possible to reduce blade pitch actuation below the baseline level while maintaining rated rotor speed.