Nowadays, most offshore wind turbines are installed on support structures that are rigidly mounted on the seabed. However, these fixed foundations are only economically feasible for water depths up to about 60 m . By contrast, floating offshore wind turbines (FOWTs) that are moored to the seabed unlock the exploitation of wind resources in deeper water. As opposed to bottom-fixed wind turbines, FOWTs are subject to motions in 6 degrees of freedom that result from the interactions between the turbine, wakes, and met-ocean conditions. Due to the complex dynamics involved in such systems, there is a necessity for a comprehensive understanding of the loads on such systems and the suitability of existing numerical models used during the design process.

Scaled model testing is an attractive way to validate numerical models and better understand the physics involved in these systems, at a lower cost than full-scale prototyping. An example of the use of experimental data by the research community is presented in , where wave basin data are utilized as a benchmark for code-to-code comparison of fully coupled aero-hydro-elastic engineering tools. Similarly, performed comparisons between wind tunnel experiments and numerical models of different fidelities to validate the aerodynamic loading on a turbine when experiencing large motions caused by a floating support structure.

In general, wave basin tests have demonstrated great utility in investigating the coupled dynamics of floating offshore wind turbines. In , the authors analysed the response of three different floaters supporting a 1:50 scale model of the NREL 5 MW turbine. As the blades of the model were only geometrically scaled, the rotor was subjected to a thrust force that was significantly lower than expected due to the Froude-scaled low-Reynolds-number wind generated at the basin. Hence, a subsequent test campaign was conducted in , where the rotor blades were redesigned, utilizing an airfoil profile specifically for Froude-scaled wind which could match the desired aerodynamic thrust forces. Other endeavours such as , , and also included servo-control capabilities to the model and mainly focused on investigating the global response of the system.

Even though testing floating offshore wind turbines in a wave basin allows us to have both the wind and the waves physically present, there are challenges involved mainly when it comes to controlling the wind field and the recirculation of the flow in the area, as shown by . Furthermore, the low Reynolds number encountered when applying Froude scaling could be detrimental to the aerodynamic response of the turbine, as discussed by . Therefore, when it comes to assessing purely the aerodynamic phenomena, it is preferred to perform wind tunnel testing.

Within the scope of the LIFES50+ project , a 1:75 model of the DTU 10 MW turbine was designed for wind tunnel testing . The model was initially mounted on top of a 2 degrees of freedom (DOF) test rig able to impose pitch and surge motions at the base of the turbine. The results of this experiment were utilized to compare the thrust and torque measurements to the results of a blade element momentum (BEM) model with dynamic wake. Discrepancies were observed between the numerical model and the gathered data, whose causes were not at first clearly identified, and hence motivated further investigation . Later, it was concluded that the outcomes of the experiment were affected by the flexibility of the tower. As a result, a new set of experiments, using a stiffer tower, was conducted. The UNsteady Aerodynamics of FLOating Wind turbines (UNAFLOW) project aimed to understand the aerodynamic response and wake of a floating offshore wind turbine undergoing large surge motions; the methodology applied to this test campaign is mainly documented in , and . Amidst the test findings, it was observed that the thrust response of the turbine presented a quasi-steady behaviour up to reduced frequencies of 0.5 Hz.

In , torque and thrust measurements for a 15 MW wind turbine model were performed for 4 DOFs. The imposed motion frequencies were defined based on the natural frequencies of each DOF according to two different floating platforms and the wave spectrum peak for a specific site. Among the findings, it was noticed that the agreement with a quasi-steady model was good for the low-frequency cases, with differences only observed for the pitch. However, at wave frequency, the rotor loads for both surge and pitch were not linearly proportional to the rotor apparent speed, indicating the presence of aerodynamic unsteadiness. In the OC6 project reported by , both pitch and surge were tested. However, for the range of frequencies considered, the aerodynamic loads presented mostly a quasi-steady response.

In addition to investigating wind turbine loads, a number of wind tunnel tests focus on floating wind turbine wakes and their development under different atmospheric conditions (e.g. ). These experimental campaigns typically rely on actuator disc models, with a focus on mid- and far-wake regions rather than turbine loads. Up to date, not many wind tunnel test campaigns with scale models of three-blade floating turbines are documented in the literature, and the number of conditions assessed is still scarce. Nevertheless, experimental data are very useful for providing low-uncertainty datasets for the validation of numerical codes. This paper builds on the previous experiments described above whilst extending the motion conditions tested. The turbine investigated here follows the same rotor scaling principles as defined in and is placed on top of a 6 DOF hexapod. Differently from the UNAFLOW project, where only surge was tested, this paper also includes pitch and yaw motions. The aim is to investigate whether the rotor is still subjected to quasi-static loads under these different prescribed motions for a range of motion amplitudes and frequencies. Additionally, prescribed motion time series extracted from 6 DOF fully coupled simulations are also applied to the hexapod, and the measured loads at the tower top are compared to the simulated values. It is shown how the present setup is capable of investigating more realistic conditions than those previously analysed in the literature.

The experimental setup is composed of a wind turbine scale model placed on top of a 6 degrees of freedom hexapod. The setup was tested in the Open Jet Facility (OJF) of the Delft University of Technology. The tunnel is a closed-loop open-jet test section facility with an octagonal nozzle with a size of 2.85 m × 2.85 m and a contraction ratio of 3:1 opening into a 13 m long, 8 m high open test section. The stream results are uniform with a turbulence intensity of 0.5 % up to 1 m from the nozzle exit, where the model is placed in the present tests, and lower than 2 % at 6 m from the nozzle. The tunnel is powered by a 500 kW fan, and the flow maximum velocity in the test section is 35 m s-1. The flow temperature is kept constant by a heat exchanger. A view of the setup in the wind tunnel test section is shown in Fig. .

A series of experiments were performed on both a static and a moving turbine. The static tests aim at characterizing the aerodynamics of the scaled rotor and hence validating the numerical design of the model. These static tests include the reproduction of the full-scale operating points and the evaluation of the rotor performance in the whole wind speed and rotor speed operating range. In addition, tests aimed at evaluating the aerodynamic sensitivities of the rotor to wind speed and rotor speed variations were performed around the operating points considered for the moving cases.

Regarding the motion tests, sinusoidal motions on a single DOF for surge, pitch and yaw were performed. The test matrix of surge and pitch motion cases consists of different sets of frequencies and motion velocities. The frequencies are in the 0.25–5 Hz range, with a thickening in the high frequencies, and the normalized motion velocities (ΔV*) are in the 0.0125–0.125 range. The motion velocity is identified as a key parameter since the thrust variation, excluding unsteadiness, is directly proportional to it. The motion amplitude is calculated from the velocity and frequency. A similar reasoning applies to the yaw motion case selection. The complete test matrices are displayed in Tables –. Two operating points of the wind turbine are considered for these tests, listed in Table , representing the rated condition (U=4 m s-1) and an additional condition in the below-rated region (U=2.5 m s-1). The duration of each test is different and is such that the tests include 50 full motion cycles. The main scope of the single DOF tests is to study unsteady aerodynamic effects that may arise at a high frequency and assess the validity of the quasi-static theory at a low frequency. The values of reduced amplitudes and frequencies for the surge motions considered in this work are shown graphically in Fig.  and are compared with the values considered in previous experimental studies as reported in . As shown, an important contribution of this work lies in having tested higher motion frequencies than in other studies available in the literature.

Additionally, 6-DOF motion tests were performed to reproduce the realistic dynamics of a floating wind turbine under combined wind–wave loading. These are just mentioned as they are not pertinent to the findings of this study but are indeed relevant for assessing the capability of the setup and preparing for future studies.

For each test run, one or multiple offset tests with a duration of 10 s are run with the hexapod in the still homing position, with the rotor stopped and the wind tunnel off, to calibrate the load cell offset values that are subjected to drifts. Each motion test case is repeated in wind and no-wind conditions. The latter, in which only the motion is activated and both the rotor and the wind tunnel are turned off, is needed to extract the aerodynamic loads from the force measurements, purifying them of the inertial component. The procedure is explained in Sect. .

During all the tests with motions, the rotor speed is kept constant. It is important to ensure that the rotor speed variation is as little as possible in order to exclude any effect of it on rotor loads that could be summed to the effect caused by the motion. To ensure good rotor speed tracking, the speed gains in the main motor servo controller are set rather high. The excellent tracking is highlighted considering that the standard deviation in a reference motion case, e.g. a high-frequency surge motion, is around 3.5 rpm, while for comparison the standard deviation for a static test is around 3 rpm. A drawback is that this results in a highly varying motor current, and consequently also torque, found in the acquired signals. However, this does not affect the results since they focus on the motion-induced thrust force variation that is not affected by the torque variance but rather benefits from the very constant rotor speed. Torque is used in steady-state analysis, where only average values are considered, or similar.

In this study, a new experimental setup is developed to analyse the aerodynamics of a scaled wind turbine under prescribed floating motions. The setup uses a scaled version of the DTU 10 MW reference wind turbine that has been extensively studied and is suitable for floating wind applications. Since wind turbine control is outside the scope of the present work, the scaled model used here has a fixed blade pitch, which simplifies the model and avoids any additional uncertainties in the experimental results. The motions are prescribed using a commercially available hexapod that can prescribe motions in 6 degrees of freedom. The present work distinguishes itself from other studies in the literature in multiple ways. Firstly, both single and realistic 6-DOF simulated motions are investigated here, whilst to our knowledge, only single and coupled 2 DOFs have been investigated in the literature so far. Secondly, the range of imposed motion amplitudes and frequencies surpasses what is currently available in the literature.

The results obtained from the single DOF motion experiments clearly indicate the presence of unsteady effects. It was observed that the variation in thrust force increases as the reduced frequencies exceed approximately 1.2. This signifies that the thrust force experiences greater fluctuations under higher reduced frequencies. However, when the reduced frequency falls below 1.2, the thrust variation remains constant, implying that the quasi-static theory can be applied in such cases. No discernible effect of wind speed and motion velocity on thrust amplification was evident from the experimental data. It was also assessed, in preparation for future studies, that the present experimental setup is capable of reproducing the full-scale behaviour of a floating wind turbine under more realistic met-ocean conditions.

These findings contribute to a deeper understanding of the aerodynamic behaviour of floating offshore wind turbines, and the observed unsteady effects underscore the importance of considering them in the design and analysis of floating wind turbines. By expanding the range of investigated parameters and pushing the boundaries of previous studies, this experimental campaign provides valuable insights into the loading on floating wind turbine rotors and can be used to further validate numerical models.

The observations outlined in this paper provide a compelling reason to pursue additional research into the impact of floater motions on wind turbine performance. To date, experiments have primarily concentrated on investigating 1 or 2 degrees of freedom. Nevertheless, operational floating offshore wind turbines experience combined motion across all 6 degrees of freedom. Hence, additional research in this field could focus on exploring the impact that the coupling of these degrees of freedom has on the aerodynamic response of the turbine.