Search

This paper presents a new hardware-in-the-loop methodology for wave-basin scale-model experiments about floating offshore wind turbines and its application as a tool for the validation of control strategies. In the hardware-in-the-loop experiments, the physical Froude-scaled wind turbine model used in conventional scale-model tests is replaced by a numerical model, measurements and a multi-fan actuator. As usual, properly-scaled waves are generated in the wave basin and the floating platform is simulated by means of a scale-model. The hardware-in-the-loop methodology was used to recreate the interaction between the collective pitch controller and the platform pitch mode that, often observed in numerical studies. In addition, the blade-root load measurement available in the numerical model of the rotor was used to implement an individual pitch control strategy. Different from in conventional experiments, the hardware-in-the-loop methodology allows to recreate a realistic three-dimensional wind field that was used to demonstrate the effectiveness of the individual pitch control. The improved emulation of the rotor loads and wind field make the hardware-in-the-loop experimental methodology an effective tool for the development and validation of control strategies for floating offshore wind turbines.

Floating offshore wind turbines allow wind energy to be harvested in deep waters. However, additional dynamics and structural loads may result when the floating platform is being excited by wind and waves. In this work, the conventional wind turbine controller is complemented with a novel linear feedforward controller based on wave measurements. The objective of the feedforward controller is to attenuate rotor speed variations caused by wave forcing. To design this controller, a linear model is developed that describes the system response to incident waves. The performance of the feedback-feedforward controller is assessed by a high-fidelity numerical tool using the DTU 10MW turbine and the INNWIND.EU TripleSpar platform as references. Simulations in the presence of irregular waves and turbulent wind show that the feedforward controller effectively compensates the wave-induced rotor oscillations. The novel controller is able to reduce the rotor speed variance by 26%. As a result, the remaining rotor speed variance is only 4% higher compared to operation in still water., Comment: Initial submission to the Science of Making Torque from Wind (TORQUE) 2020 conference, 10 pages, 7 figures

Closely spaced counter-rotating vertical-axis wind turbines (VAWTs) exhibit a considerable power improvement compared to the same turbines in isolation. It thus makes sense to study their potential for wind farm production optimisation. With this objective in mind, the wake of an isolated VAWT is compared experimentally to the wake of counter-rotating VAWTs. Because of the unsteady aerodynamics, the wake of an isolated VAWT deflects towards the region behind its upwind moving blade. The direction of rotation thus directly influences the deflection of the wake. For paired configurations, it is possible to use this wake deflection as an advantage. A pair of counter-rotating VAWTs where the upwind moving blades are at the centre of the pair, exhibits a particularly narrow wake. With the benefit of a power improvement and a narrow wake, closely spaced counter-rotating VAWTs exhibit promising characteristics to improve the power density in a wind farm.

The present work deals with the issues faced by authors in the implementation of a variable-speed variable-pitch (VS-VP) controller on a wind turbine scale model for wind tunnel tests. The PoliMi 1/75 model of the DTU 10MW reference wind turbine (RWT) is presented, focusing on the mechatronic system used to control the machine. The main differences between the PoliMi wind turbine model (WTM) and the DTU 10MW RWT are highlighted out pointing out their influence on the behavior of the controlled machine. Then, the control logic adopted for the PoliMi WTM is introduced. Finally, results from wind tunnel tests on the controlled model are shown and the performance of the closed-loop system is discussed.