Abstract. The current test process in design and certification of wind turbines (WTs) is time and cost intensive, as it depends on the wind conditions and requires the setup of the WT in the field. Efforts are made to transfer the test process to a system test bench (STB) whereby an easier installation is enabled and the load can be arbitrarily applied. However, on a STB the WT is installed without rotor and tower and the remaining drive train behaviour acts differently to the WT drive train in the field. The original behaviour must be restored by incorporating a Hardware-in.the-Loop (HiL) simulation into the operation of the STB. The HiL simulation consists of the virtual rotor and wind and the control of the applied loads. Furthermore, sensors as the wind vane and actors as the pitch drives, which are not present at the STB, are substituted by simulation models.

This contribution investigates suitable HiL control methods of the applied torque. Herein, we survey three methods of different complexity and compare them in terms of performance, actuator requirements and robustness. The simplest method emulates the divergent inertia by classical control. A more complex method based on a reference model also considers the alternated dynamic behaviour of the drive train. Model predictive control (MPC) currently constitutes the most complex HiL method, as the MPC also includes future predictions of the driving torque behaviour. Our comparison identifies that increased complexity of the control method ensures enhanced preformance. WT drive train dynamics can be reproduced up to 1, 6, and 10 Hz for IE, MRC and MPC, respectively. Yet, for higher control complexity, the requirements for the dynamic torque proliferate and the controllers robustness to model deviations decreases.