Control design and optimization for the DOT500 hydraulic wind turbine

The drivetrain of most wind turbines currently being deployed commercially consists of a rotor-gearboxgenerator configuration in the nacelle. This abstract introduces the control system design and optimization for a wind turbine with a hydraulic drivetrain, based on the Delft Offshore Turbine (DOT) concept1 . This concept enables the connection of multiple turbines with only a single water pump in the nacelle to drive a centralized multi-megawatt generator at sea level: simplifying maintenance, reducing the amount of components and thus has great potential in reducing the Levelized Cost of Electricity (LCOE) of offshore wind. To evaluate the practical feasibility of the concept, field tests are performed with a full-scale retrofitted Vestas V44 600 kW wind turbine, of which the drivetrain is scaled to a 500 kW hydraulic configuration. As a substitute to generator torque control used in conventional wind turbines, an alternative torque controller is designed where fluid pressure is regulated to vary the system torque. Moreover, during the field tests, a data-driven optimization technique is employed to find the unknown fine-pitch angle for maximum rotor power extraction in the below-rated region. The system torque is controlled by varying the nozzle area with use of a spear valve, which influences the fluid pressure. In previous work1 , a hydraulic torque controller is developed, where on a laboratory test set-up it is shown that a passive torque control scheme enables the rotor to operate near maximum aerodynamic efficiency in the below-rated region. However, due to complexification of the drivetrain in the early prototype stage of the fullscale wind turbine, the passive control implementation is not yet feasible. For this reason, a grid search is performed during an indoor measurement campaign where the rotor is substituted by a 500 kW electric motor, to obtain steady-state characteristics of the hydraulic drivetrain. Fig. 1 shows the intersection between the experimentally found system torque and the Cp,max rotor torque plane, as function of rotor speed and spear valve position. The intersection is used to construct a feedforward controller for maintaining the optimal rotor power coefficient in the below-rated region. In-field evaluation results are presented in Fig. 2, and show that the torque controller, subjected to fluctuating rotor speeds, is able to maintain the optimal tip-speed ratio (TSR) of 7.43. As exact information on blade geometry and thus aerodynamic rotor characteristics is unavailable for the Vestas turbine, the fine-pitch angle for tracking Cp,max is assumed to be 0 deg during the control design phase. During field tests, the data-driven optimization algorithm Extremum Seeking Control (ESC) is employed to optimize the fine-pitch angle2 , with the objective to maximize rotor power capture. Results in Fig. 3 show that the ESC implementation converges towards an optimal fine-pitch angle of -2 deg. Future work will focus on system design and control design for a complete wind farm driven by hydraulic turbines. Moreover, in later prototype stages, the passive torque control strategy1 will be evaluated, and online datadriven optimization techniques will be applied to actual full-scale offshore hydraulic wind turbines.