Experimental investigation of the effects of floating wind turbine motion on a downstream turbine performance and loads

Fontanella, Alessandro; Cioni, Stefano; Papi, Francesco; Muggiasca, Sara; Bianchini, Alessandro; Belloli, Marco

doi:10.5194/wes-11-1821-2026

Articles | Volume 11, issue 5

https://doi.org/10.5194/wes-11-1821-2026

© Author(s) 2026. This work is distributed under
the Creative Commons Attribution 4.0 License.

https://doi.org/10.5194/wes-11-1821-2026

© Author(s) 2026. This work is distributed under
the Creative Commons Attribution 4.0 License.

Articles | Volume 11, issue 5

Research article

|

20 May 2026

Research article |

| 20 May 2026

Experimental investigation of the effects of floating wind turbine motion on a downstream turbine performance and loads

Alessandro Fontanella, Stefano Cioni, Francesco Papi, Sara Muggiasca, Alessandro Bianchini, and Marco Belloli

Download

Final revised paper (published on 20 May 2026)
Preprint (discussion started on 26 Jun 2025)

Interactive discussion

Status: closed

RC1: 'Comment on wes-2025-106', Anonymous Referee #1, 14 Jul 2025

Please see attachment.

Citation: https://doi.org/10.5194/wes-2025-106-RC1
RC2: 'Comment on wes-2025-106', Anonymous Referee #2, 11 Aug 2025

The paper presents interesting load measurements for a downwind turbine, in the wake of a dynamically moving turbine (representing offshore floating conditions). The results are important and highly valuable for the field of floating wind energy. The paper reads well, though is not the most concise. If made more to the point and concise it could help future readers. The paper presents results for many different conditions, and hypotheses are proposed for the observed behavior, but are mostly not proven or verified by a deeper analyses of the underlying physics. The reviewer considers the paper acceptable for publication given revisions are made according to the comments below and the comments made by the other reviewer. Most importantly, the paper should describe more clearly which wind measurement data is used, how it is measured, or where the rotor effective wind speed is reconstructed through thrust measurements. More details about how this is calculated is also needed in the paper: formula and used coefficients. I want to put attention to the fact that the rotor reconstructed velocity has its limitations, and if the entire analysis is based on the rotor reconstructed velocity, a second review should be performed to make sure the analysis is correct.

- L12 'a key driver' - sentence is missing for what it is a key driver? For wake recovery? unsteady loading ?
- L21 'substantial power losses, increased fatigue loads' --> The platform motions could increase or decrease wake losses and fatigue loads, as also shown in this paper. It is more unbiased and correct to word it as 'change', 'affect; or 'impact' those performance characteristics, without suggesting it will increase or decrease.

- L59: What is not discussed in the introduction is the different time scales at which platform and wave motions can occur (Strouhal number of motions). A difference should be made between very slow motions with larger amplitude and very fast motions usually with a much smaller amplitude. The impact on the wake can be very different, and could impact the summarized conclusions in this section.
- L75: 'more realistic turbulence levels.' --> It is relevant to add for which conditions/ Strouhal numbers and amplitudes these findings were made.
- L87: The two key research questions are a bit misleading. We know the answer is yes to both questions. The real question is: how much and for which conditions. I think the real answers from this paper are about the quantification and characterization for specific test conditions.
-L131: It would be good to add to this sentence how the authors expect that the results in this paper should be interpreted to make conclusions for wind turbines with larger spacings. Likely the motion signatures in the unsteady loading will reduce for a larger spacing? Can the authors make predictions based on the tests at different locations?
- Experimental setup: Is the design of the rotor blades changed for a lower Reynolds number? What is the Reynolds number during the tests? Does the thrust and power coefficient match the full scale value?
- L154 'noticeable influence' Can the authors describe what the influence is?
- L162: A fixed rpm leads to varying tip speed ratio during pitch or surge motions. Can the authors estimate how much the tips speed ratio varied for the first wind turbine? This can possibly help understand the impacts on unsteady loading.
- L163: How much was the thrust (coefficient)? Good to document this for future reference or giving the chance to reproduce or model the results.
- L174: Can the authors describe more precisely how this was calculated? The wind turbine thrust coefficient is calibrated with tip speed, or expected to be constant in the operating range?
- L176: How is the optimal tip speed ratio determined for WT2? Tested by varying or theoretical?
In general it is not clear in the paper where real wind speed measurements are used and where the reconstructed rotor-effective wind speed is used.
- Table 3: can the authors add the reduced frequency to the table?
- Figure 5: the wake losses seem very high! Usually we expect the power of a downstream aligned turbine to be closer to 40%-50% or even a little higher, while the data shows 20% at a spacing of 5D. Is the thrust coefficient of the turbines very high? Is the incoming turbulence very low? It is important to discuss this in the paper, because it can affect the unsteady loading and its sensitivity to platform motion.
- L265: How are the velocities for this formula defined or measured? The paper is missing a description of wind speed measurements. Maybe a more clear reference to previous work would help.
- L304: how is this average wake velocity determined? Is this calculated from the measured thrust force?
- L306: I can not find the 16% on figure 5
- L315: These observations indicate a high thrust coefficient of WT1. Is the magnitude expected to be so high? any references in the literature?
- L345: 'delay of ' Assuming the disturbance travels with an average convective velocity in the wake: can the authors use this time difference to estimate the convective velocity and see how it compares to the free stream velocity?
Figure6 : How is the thrust and power coefficient determined / how is the velocity measured to determine these coefficients? This is not clear in the paper. Velocity measurements are not really discussed as part of this paper. It looks like CT for WT1 and WT2 are normalized by the same velocity? which is not conventional: usually it is defined wrt the incoming velocity of each turbine, which is lower in the case of WT2.
- L374: The higher the reduced frequency, the shorter the streamwise wavelength of the velocity perturbations caused by the motion of the turbine. Based on the incoming wind velocity and the frequency the authors could estimate the wavelength of the velocity fluctuations created by the turbine motion. If this wavelength is of the same size as the integral length scale of the incoming turbulence, or perhaps the rotor diameter, it would make sense that it dissipates or decorrelates much faster. Perhaps the highest motion frequency passes a specific treshold explaining why the turbulent structures dissipate quickly?
When the authors say weaker: do they mean smaller in amplitude, shorter in wavelength, or both?
- L390: Again not clear how the velocity is measured.
Figure 8 a and b: Are these absolute thrust values, or fluctuations of the thrust signal?
- L409: It would be good to state that this conclusion is for the specific streamwise spacing. For a larger streamwise spacing the wake spreading will be larger, possibly changing the conclusion.
- L511: In this study only the first turbine is moved. By moving it the unsteady loading also increases for WT1. How would the findings in this paper change if WT2 also moves with waves? And what if the motion of WT2 is spatially and temporally correlated to the motion of WT1 such that movements of WT2 could be in sync or anti-sync with changes in windspeed from WT1? It would be helpful for the paper if the authors could make a comment about this for context.
- L530: See previous comment. For a bit more completeness, it would be good to add to this summation as well ' spatio-temporal correlation of waves through the wind farm'.

- The discussion section is a little awkward. It doesn't seem to present a new discussion but rather an overview of previous conclusions.

Citation: https://doi.org/10.5194/wes-2025-106-RC2
AC1: 'Comment on wes-2025-106', Alessandro Fontanella, 12 Aug 2025

Dear Referees,
Thank you for your thorough reviews and constructive comments, which we will address in detail while revising the manuscript. We think your feedback helps to further develop and clarify the results presented in the paper, and we greatly appreciate it.
Both of you raised points regarding the velocity measurements performed in the wake and the estimate of the rotor effective wind speed used to regulate the rotor speed of the downstream wind turbine. We take this opportunity to briefly clarify the distinction here, while a more detailed explanation will be included in the revised manuscript.
The rotor effective wind speed estimate was used solely to set the rotor speed and tip-speed ratio (TSR) of the downstream turbine, ensuring operation close to its nominal performance curve. A dedicated Appendix will be added to explain in detail how this estimator works.
In contrast, the wind speed values discussed in the results were obtained from hot-wire probe measurements. The experimental setup and methodology for these wake measurements are described in detail in [https://doi.org/10.5194/wes-10-1369-2025]. While our original intent was for this work to complement the study presented in [https://doi.org/10.5194/wes-10-1369-2025], we agree with your suggestions that this article would be significantly strengthened by including additional results and details on the upstream turbine wake measurements, and we will incorporate these in the revision.

Citation: https://doi.org/10.5194/wes-2025-106-AC1

Peer review completion

AR – Author's response | RR – Referee report | ED – Editor decision | EF – Editorial file upload

AR by Alessandro Fontanella on behalf of the Authors (04 Sep 2025) Author's response Author's tracked changes Manuscript

ED: Referee Nomination & Report Request started (23 Sep 2025) by Raúl Bayoán Cal

RR by Anonymous Referee #1 (02 Oct 2025)

RR by Anonymous Referee #2 (06 Oct 2025)

ED: Reconsider after major revisions (20 Oct 2025) by Raúl Bayoán Cal

AR by Alessandro Fontanella on behalf of the Authors (29 Oct 2025) Author's response Author's tracked changes Manuscript

ED: Referee Nomination & Report Request started (04 Nov 2025) by Raúl Bayoán Cal

RR by Anonymous Referee #3 (22 Jan 2026)

Suggestions for revision or reasons for rejection

The paper „Experimental investigation of the effects of floating wind turbine motion on a downstream turbine performance and loads“ is well written and of very good quality. The experiments carried out are well described and results are presented in understandable way. Nevertheless I would recommend to address a few minor aspects to improve the readability and quality even more.

Experimental set-up:

Where were the hot-wire probes positioned? Results e.g in figure 8 and 9 and others, suggest a more or less continuous scan but that is not what happened. The authors should add markings where they measured the velocities and make clear that these figures are a results of interpolation.

Does the robotic platform generate a vertical velocity profile for the turbine WT2 in the wake ?? It is not clear if the measurements in section 2.2 presented for WT2 in "free-inflow" are with or without the robotic platform in front or not. The influence of the robotic platform should be quantified.

section 2.1.1.:

- can the authors say anything about the accuracy of their calibration to exclude inertial and gravitational forces from their dynamic measurements? I know that it is a hard task, nevertheless it would be interesting to see if there could be put a number on this at least an estimation?

section 3.1.2:

- even in the fixed case the measured power is way under the expected from other experiments — again the question remains, did the authors measure the WT2 performance for „free-inflow“ with or without the robotic platform which could have an impact on the vertical profile in the wake

section 3.2:

- the loads in WT2 occur with a delay (line 466) relative to the loads on WT1. What is the advection velocity (line 473) ? Does that match the theoretical values ?

section 3.2.4 Cross wind motion

- The caption in the figure 11 is not quite clear since the meaning of the angle gamma is not clear. In the text, it is written that gamma = 0° is pure surge motion — is gamma = 90° pure sway? What describes the amplitude of 0.032m? Since it is a combination of surge and sway each motion has an amplitude. Two sinusoidal signals (surge and sway) would result in a Lissajous figure as the motion pattern of the turbine model, depending on their phases it is either a straight diagonal, a circle or an ellipses. I think it would be helpful to add a sketch of of the motion pattern of the turbine. The angle gamma does not really help to understand the motion of the turbine.

Hide

RR by Anonymous Referee #1 (11 Feb 2026)

RR by Anonymous Referee #2 (02 Mar 2026)

Suggestions for revision or reasons for rejection

The paper discusses a very detailed and unique dataset for studying the impact of platform motions of floating turbines on downstream turbines. The experiments are unique due to the detailed load measurements, and the accurate control of the motion of the first turbine platform.

The experiments allow to study idealized scenarios, as well as realistic irregular wave patterns, which delivers very interesting comparisons. However, as is the case in almost every experiment: there are some differences or approximations from full-scale offshore conditions in this experiment. In general the authors have done already a good job in discussing the limitations and context of the results, and what this means for comparison to full scale real offshore conditions (for example line 680 - 688). For example: the relatively small turbine spacing, low level of turbulence in the incoming flow, and the fixed platform of the second turbine. (I want to stress that his is not an issue, as it allows to study specific idealized scenarios which make it possible to understand better individual mechanism etc., and which are also more easily simulated.) However, one limitation should receive a bit more discussion in the paper, and that is the blockage of the ceiling. While the authors mention that they think this wont play a role in relative comparisons, I think it may affect to some degree how the wake can spread due to the different motions. In either case, a bit more context for the reader would be good to interpret the results.

In general I only have a few minor comments, and one request to discuss a more the limitations of the results due to the ceiling blockage.

Specifically:
A spacing of 0.2D between the rotor tips and the ceiling is small. This will most likely affect how the wake recovers or deflects, because it is closer or similar than the typical distance between the bottom tip and the ground, and we know that ground affects play a role in vertical wake deflection, thus the ceiling will limit the amount of wake deflection or how the wake expands to some degree.

Line 405: The reviewer thinks that the small wake recovery may also be a sign of the blockage above the wind turbines. See for example the papers on vertical entrainment of mean kinetic energy in a wind farm required for wake recovery. Due to the ceiling blockage, the wake recovery has to come mostly from the sides, while the vertical contribution (usually the largest) is missing. Can this also explain the slower wake recovery?

Line 419: That WT2 operates at a higher power when moved just outside of the wake of WT1 is something that has been observed in the literature before. Several simulation studies show an acceleration around the turbine. Furthermore, the wind tunnel experiments of Bossuyt et al. 2018 (Effect of layout on .. figure 13) also notice this effect when turbines are shifted 1D apart, though in that case at a smaller streamwise spacing. The fact that the tests in this paper see this phenomena at 5D downstream may be another indication of a extra blockage effects?

Line 420: about the conclusion ‘ the strongest impact by yaw motion’. Please consider that the proximity of the ceiling will dampen or block pitch or vertical wake deflections or motions in general. Thus, my opinion is that the authors should be careful to use this dataset to make this specific conclusion. If the authors find studies in the literature confirming that yaw would indeed have the strongest impact factor, than this can be used to justify that the conclusions in the paper still hold, even considering the blockage effects.

Line 660: ‘The blockage effect … “ It is great that the reviewers put in perspective the conditions for which these conclusions are relevant. However, the blockage could also affect even the relative comparison, in my opinion. For example, the vertical blockage by the ceiling may limit how much the wake can deflect vertically, and also affect its wake recovery in deflected state. Thus, I think this sentence needs more nuance, and it would be helpful if the authors could discuss a bit more what the impact of blockage may be on wake recovery and wake deflection, and what this means for the conclusions in this paper.
For example, some thoughts:
- slower wake recovery & a more pronounced wake: this could accentuate more the individual impacts of the motions of WT1 on the loads and power of WT2. If the wake recovery would be faster, the wake would spread faster and the wake boundary would be less defined, and thus if we then move WT2 radially out of the center of WT1 (the 1R or 1D positions) it would likely see a slower decline in loads (depending on the motion of WT1)? Alternatively: if the inlet flow would contain much more turbulence, like in an atmospheric boundary layer, we can expect background turbulence to increase wake meandering, also reducing the sharp wake boundaries?
- blockage on the other hand, may reducing the effects of wake deflection in this experiment?

Other minor comments:

Line 185: In the no-wind tests: because the rotor is not spinning, the centrifugal loads are not included. Is this a problem or big impact? Would be helpful to discuss in the text.

Line 197: This one sentence contains a lot of information. If I understand correctly each hot-wire time signal is synchronized with the platform motion, and then conditionally averaged to specific positions of the platform motion? Should there be more context here about the number of samples measured per hot wire and per conditional point?

Line 366: include a relative measure of deviation (in %)

Line 386 : typo for average

Figure 10 and 9: be clear on figure y-axis label: these are not just amplitudes, but amplitudes of the oscillations, correct?

Line: 565: the apparent wind speed with opposite signs is not clear

Interesting to check: how does the amplitude of the rotor motion compare to the amount of wake deflection for sway, yaw and pitch?

Hide

ED: Reconsider after major revisions (05 Mar 2026) by Raúl Bayoán Cal

AR by Alessandro Fontanella on behalf of the Authors (19 Mar 2026) Author's response Author's tracked changes Manuscript

ED: Referee Nomination & Report Request started (20 Mar 2026) by Raúl Bayoán Cal

RR by Anonymous Referee #2 (27 Mar 2026)

RR by Anonymous Referee #1 (05 May 2026)

ED: Publish as is (07 May 2026) by Raúl Bayoán Cal

ED: Publish as is (08 May 2026) by Paul Veers (Chief editor)

AR by Alessandro Fontanella on behalf of the Authors (08 May 2026) Manuscript

Short summary

This study explores how the movement of floating wind turbines affects nearby turbines through wakes. Using wind tunnel experiments, we found that certain motions of an upstream turbine can improve the energy produced by a downstream one and change the forces it experiences. These effects depend on how the turbines are spaced and aligned. Our results show that the motion of floating turbines plays a key role in how future offshore wind farms should be designed and operated.