the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 08 Apr 2026
Wind Tunnel Testing for Wind Energy: A State-of-the-Art Review
Abstract. Wind tunnel testing is essential for the advancement of wind energy, providing the "ground truth" necessary to validate numerical and data-driven models. However, the increasing size of modern turbines introduces significant challenges, including aerodynamic scaling, tunnel blockage, and the need to simulate complex atmospheric boundary layers. This paper reviews the state-of-the-art in wind energy experimental aerodynamics, covering inflow conditions, airfoil performance, aeroacoustics, and aeroelasticity. It further examines the integration of wind tunnel data with CFD modeling, field measurements, and data science. By synthesizing current progress and identifying inherent limitations, this review provides a critical outlook on future directions for the experimental wind engineering community.
Competing interests: At least one of the (co-)authors is a member of the editorial board of Wind Energy Science.
Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims made in the text, published maps, institutional affiliations, or any other geographical representation in this paper. While Copernicus Publications makes every effort to include appropriate place names, the final responsibility lies with the authors. Views expressed in the text are those of the authors and do not necessarily reflect the views of the publisher.- Preprint
(35500 KB) - Metadata XML
- BibTeX
- EndNote
Status: open (until 03 Jul 2026)
- CC1: 'Comment on wes-2026-51', Michaela Herr, 19 May 2026 reply
-
RC1: 'Comment on wes-2026-51', Anonymous Referee #1, 21 May 2026
reply
Wind Tunnel Testing for Wind Energy: A State-of-the-Art Review
Manolesos et al.
A review is written for experimental scaled testing in the context of wind energy. The topic is of relevance and would be of interest to scientists performing experiments. The manuscript does not read like a review. For having 145 pages, which is excessively long, the work efforts done by the community - it highlights the work mostly of the authors as observed by the images and the citations. The outlook sections do not highlight open questions and possible routes for addressing the questions - much more detail can included to guide those wanting to work on questions related to the particular topics. The organization of the manuscript by topic is well received yet the order and possible coupling between them is not present. There is no mention of requirements or scales for instrumentation. The scaling arguments are of great importance and this is also not mentioned. There is not a single, unifying voice in the manuscript. The initial sections of the manuscript lack supporting images. Concrete results are in general not included but rather discussion of global points which leaves the reader needing details for what has been found. There is a great imbalance between sections, some providing details and others do not yet the claim is that there has been work performed in each of the areas. The sample results included, which should be prevalent throughout, are not well discussed. Given these points, the manuscript is rejected.
Citation: https://doi.org/10.5194/wes-2026-51-RC1 -
RC2: 'Comment on wes-2026-51 - Aeroacoustics section', Anonymous Referee #2, 02 Jun 2026
reply
The manuscript provides a broad and valuable review of wind-tunnel testing methodologies for wind-energy applications. The aeroacoustics section covers several important elements of wind-turbine noise prediction and measurement, including turbulent-boundary-layer trailing-edge noise, Amiet-type formulations, wall-pressure-spectrum models, microphone-array measurements, PIV-based pressure reconstruction, and aeroacoustic facility design.
However, for a manuscript intended to provide a state-of-the-art review, the aeroacoustics discussion appears insufficiently connected to recent developments: only a limited number of studies from the last five years are cited, while important advances in the physical understanding of noise generation mechanisms in wind turbines are reported in this period.
For instance, I believe that the discussion of trailing-edge noise in Sections 6.2–6.3 requires a substantial update. In its present form, the section presents the wall-pressure spectrum, convection velocity, and correlation length as the primary quantities controlling trailing-edge-noise prediction, with the spanwise correlation length introduced through Corcos- or Efimtsov-type models. While this is consistent with classical semi-empirical prediction approaches, the discussion overlooks an important body of more recent work concerning the specific turbulent structures that are acoustically efficient.
In particular, the manuscript states that the correlation length provides the characteristic size of turbulent wall-pressure structures and reports empirical relations for its estimation. This treatment may give the impression that an integral correlation-length measure is sufficient to characterise the flow structures responsible for radiated trailing-edge noise. Recent studies indicate that this is not generally the case. The wall-pressure field contains energetic hydrodynamic fluctuations that may contribute strongly to conventional coherence or correlation-length measures while contributing little or nothing to the radiated acoustic field. From the standpoint of trailing-edge scattering, the relevant issue is not only the extent of hydrodynamic coherence, but whether the corresponding spanwise-wavenumber components satisfy the acoustic radiation condition.
This point was demonstrated directly by Sano et al. (2019), who investigated trailing-edge noise from a turbulent boundary layer over a NACA 0012 airfoil using large-eddy simulation and flow-acoustic correlations. They showed that only spanwise wavenumbers below the acoustic wavenumber contribute to radiated sound, and that isolating spanwise-coherent disturbances greatly increases the correlation between near-field hydrodynamic fluctuations and far-field acoustics. Their results point to the scattering of spanwise-coherent boundary-layer structures as the dominant mechanism of trailing-edge noise for the flow considered.
The physical nature of these structures was further investigated by Abreu et al. (2021), who identified spanwise-coherent hydrodynamic waves in turbulent flows around flat plates and airfoils using spectral proper orthogonal decomposition and resolvent analysis. These structures exhibit significant amplitude near the trailing edge and represent a more specific description of acoustically relevant flow organisation than conventional integral correlation lengths alone. Subsequent experimental and numerical studies, including recent work by Demange, Yuan, Cavalieri, Oberleithner and co-authors, have continued to examine the role of spanwise-coherent or low-spanwise-wavenumber structures in broadband trailing-edge noise generation.
Overall, I encourage the authors to revisit the aeroacoustics section as a whole and incorporate recent developments more systematically, so that it more accurately reflects the current state of research in wind-turbine noise prediction, measurement, and mitigation.
Relevant references that should be considered include:
Sano, A., Abreu, L. I., Cavalieri, A. V. G. and Wolf, W. R. (2019). Trailing-edge noise from the scattering of spanwise-coherent structures. Physical Review Fluids, 4, 094602.
Abreu, L. I., Tanarro, A., Cavalieri, A. V. G., Schlatter, P., Vinuesa, R., Hanifi, A. and Henningson, D. S. (2021). Spanwise-coherent hydrodynamic waves around flat plates and airfoils. Journal of Fluid Mechanics, 927, A1.
Demange, S., Yuan, Z., Jekosch, S., Sarradj, E., Hanifi, A., Cavalieri, A. V. G. and Oberleithner, K. (2024). Identification of structures driving trailing-edge noise. Part I: Experimental investigation. Preprint.
Yuan, Z., Demange, S., Oberleithner, K., Cavalieri, A. V. G. and Hanifi, A. (2024). Identification of structures driving trailing-edge noise. Part II: Numerical investigation. Preprint.
Citation: https://doi.org/10.5194/wes-2026-51-RC2 -
RC3: 'Comment on wes-2026-51 sections 2, 4, 5, and 7', Anonymous Referee #3, 04 Jun 2026
reply
This is an expansive piece of work that has clearly required the effort of numerous leaders in the wind energy wind tunnel testing community. I think the document can have value to the community as a reference but that it also requires some work before it can be published. I do want to say that I enjoyed reading this document and I did learn things, but broadly I am not sure it is sufficiently comprehensive, perhaps ironically, given its length. It focusses too lightly on too many topics. If length isn’t a problem, then diving deeper into some areas might be a solution.
As the manuscript is so long, different reviewers were tasked with looking at different parts of the manuscript. My review specifically focusses on sections 2, 4, 5, and 7, although I do offer some comments on other parts of the manuscript as well. Please consider that my lack of comments elsewhere does not mean I approve of those parts as is, but rather that I have not evaluated them. Below I divide my feedback into general comments, specific comments, and typographical errors.
DISCLAIMER: I will suggest some additional references throughout my review. The authors should not feel these are mandatory citations that must be included. Please only include the ones you feel are relevant after considering my comments and how they fit into your vision for this work.
GENERAL COMMENTS:
- The manuscript generally requires more homogenisation. The author transitions can be quite obvious and the depth of the subsectioning is not consistent (although at least it never goes deeper than 3 levels). Here, I mean more what qualifies as a sub-sub-section versus just a sub-section. The sections do not really refer to each other and there are no real common themes other than “wind energy” and “wind tunnels.”
- Some sections seem to be written in isolation without knowledge of each other. For example, section 2 could probably have an overall conclusion/summary that ties together the various parts. Right now, each sub-section of section 2 has its own independent outlook. Specifically, 2.2.3 and 2.3.3 appear to not know about each other even though it would be natural that they refer to each other as they are effectively adjusting the inflow using different tools.
- Some sections read more like mini-papers and others read more like cursory reviews. For example, the details given on the Pour la Cour Tunnel in section 5 and then the entireties of sections 8 and 9 read differently from the rest of the manuscript. This speaks to the lack of homogeneity already expressed.
- The title is “Wind Tunnel Testing for Wind Energy,” yet vertical axis wind turbines (VAWTs) are almost completely absent from the review (Fig 3 is the only place they appear). While I think it is broadly fine that the authors focus on HAWTs, they are by far more used and researched, I think this should be explicit in the material at the start of the manuscript, e.g., title, abstract, introduction. I concede changing the title might be overkill.
- Complex terrain is discussed but there is no mention of buildings or the built environment. This might go hand-in-hand with my previous comment, as this is perhaps more relevant for VAWTs. If clarification is made that this work primarily deals with HAWTs, then I think perhaps incorporating buildings is of less import.
- It occurs to me that a detailed discussion of the various wind turbine models used around the world is not featured in this review. Actuator/porous disks have a section, but not the model turbines. Is this a deliberate omission? Perhaps to avoid conflict? Model turbines seem to be touched on sections 4, 8 and 9 but have no real dedicated discussion on them or the flow in their wakes. I noticed this because I also noticed no reference is made to the pressurised experiments from Princeton or MPI, e.g., Miller et al. (Phys Rev Fluids, 2019), Grundwald & Brunner (Phys Rev Fluids, 2026), amongst others. These studies certainly seems like they should be of interest here as these measurements are able to match quantities at lab scale that standard facilities cannot.
- It is perhaps understandable, but a lot of the content reflects publications from the authors, who, of course, have substantially contributed to the field. I estimate the work is 30-40% self-citations. Maybe this is fine. I am not sure. But I state it here as an observation. I wonder if this ratio is reflective of the literature in general.
SPECIFIC COMMENTS:
- (ln. 147) This line says the flow properties become homogeneous 5-10M downstream of a passive grid. I strongly disagree with this statement and I think it also partially conflates the ideas of homogeneity and evolution—although I do not think the authors did that on purpose. There is a wealth of literature saying that this is not the case, starting from Corrsin (Handbuch der Physik, 1963) to more recent studies, e.g., Ertunç et al. (JFM, 2010), Krogstad & Davidson (Phys Fluids, 2012). One can concede that the first-order statistics (e.g., mean velocity) become uniform quite quickly but higher order stats definitely have lingering inhomogeneity well beyond 10M.
- (sec. 2.2.2, paragraph starting on ln. 203) This paragraph, and the entire subsection, are quite Germany-centric. That is understandable given the author list. I would make the following observations:
- (around ln. 204) Reference is made to the “double-random asynchronous mode”, but the two largest parametric studies on what the flow does when this mode is used are not referenced, i.e., Larssen & Devenport (Exp Fluids, 2011), Hearst & Lavoie (Exp Fluids, 2015).
- (around ln. 208) Mention is made to producing mean gradients in the flow, but the originators of this approach (at least in the journal literature) are not reference, i.e., Cekli & van de Water (Exp Fluids, 2010). Note, further developments were also made contemporaneously with the referenced works, e.g., Hearst & Ganapathisubramani (Wind Energ, 2017), Cekli & van de Water (Phys Fluids, 2020), Li et al. (Renew Energ. 2020).
- Broadly, it would be nice for sec. 2.2.2 to have some figures, probably set-up-based ones, e.g., showing spires and comparing that to an active grid operating to generate shear. Such pictures are common in other parts of the manuscript. Perhaps it would be good to show some spectra as well to explicitly demonstrate the “separation of scales” mentioned in the text (ln. 237), and, if possible, explicitly compare to a spectrum measured in the field.
- (ln. 305-306) There is a comment that temperature and velocity measurements have not been performed together before with PIV. I think this needs more explicit context because broadly it is not true. In the combustion community (e.g., Fond et al. (Opt Exp, 2012)) and in liquid flows (e.g., with Rhodamine, Naumann et al. (Exp Fluids, 2026) and references therein) such measurements are fairly commonplace. If the statement is kept, it needs more qualifiers to make it true in the wind energy context. Alternatively, a nod to the communities that do already do this and a suggestion that this could be a future route would also address this.
- (sec. 2.3 and 3) There is no introduction or discussion of Richardson number, Nusselt number or even explicitly stability. These seem like relevant concepts to introduce here…. At least Richardson and stability. Richardson number is used in 2.3.2 and on ln. 911, but is never defined or explained. It would also be interesting to comment on how important generating the stability conditions are relative to just mimicking the momentum/velocity profile with spires or an active grid. This could be a linking discussion for 2.2 and 2.3.
- (sec. 4) For the power load measurements, Gambuzza and Ganapathisubramani also have a more modern iteration of the Bastankhah and Porté-Agel approach, c.f., Gambuzza & Ganapathisubramani (JRSE, 2021; JFM, 2023), Vinnes et al. (JRSE, 2022).
- (Table 1) I understand this list is not meant to be exhaustive, but it might be worth including something like Hillestad et al. (Phys Rev Fluids, 2024 or 2026) which use the Shake-the-Box volumetric reconstruction approach. That technique is referenced later on ln. 732 but seems to be missing from the list in Table 1. STB is in increasingly common usage (in fluid mechanics in general it is more common than Tomo-PIV today).
- (ln. 655) While I agree the common practice is to sample hot-wire signals up to 50 kHz or even higher, I think in a review paper some perspective should be provided, as the wires cannot actually respond to 50 kHz in the same way they respond to 1 kHz, for example. Hutchins et al. (Exp Fluids, 2015) tested several different anemometers and concluded “…the frequency response of under- or over-damped HWA systems can only be considered approximately flat up to 5-7 kHz.”
- (sec 4.3.3 and 4.3.5) There is no discussion of Stokes number and its relevance for particle tracers. There is simply a mention that micron sized particles are used (ln. 666 and 703). In a comprehensive review focussed on wind tunnel studies that say a PhD supervisor might get a new PhD student to read at the start of their PhD, I would expect at least a sentence or two on particle sizing and Stokes number.
- (ln. 793) Consider adding Bartl et al. (Euro J Mech B, 2019) which tests an NREL S826. This also probably should be discussed in sec. 12.1 as the entire point of that study is to perform experimental measurements and simulations together. Note, the same airfoil is used in a similar way (experiments+simulations) for icing (lumped into “erosion” in the present manuscript) by Hann et al. (Aerospace, 2020).
- (sec 5.2) The title is somewhat misleading. Perhaps it should be called “Blockage corrections” as that is what it primarily deals with. It would also be nice to see an equation for a typical blockage correction approach. That would be useful if the document is meant to be used as a resource in the future (again, consider my “new PhD student” scenario).
- (ln. 924) There is a statement starting on this line whereby it seems to be implied dynamic similarity does not hold…. Maybe as compressibility effects begin to be relevant... This needs further clarification and exposition.
- (sec. 5.6) The word “erosion” is used but the lab example given is an airfoil with sandpaper added. To me “erosion” is fundamentally subtractive. Putting sandpaper on something is additive. Moreover, icing is briefly referred to on ln. 963-964; that is also additive. If these are all meant to be lumped together maybe degradation or contamination is more accurate terminology. Moreover, icing is a quite rich field in and of itself, including studies that specifically use wind turbine blades, e.g., Jasinski et al. (J Sol Energ, 1998), Bragg et al. (Prog Aero Sci, 2005), Etemaddar et al. (Wind Energ, 2014) Ansell & Bragg (AIAA J, 2014), Yirtici et al. (Wind Energ, 2019) Hann et al. (Aero, 2020), Vinnes & Hearst (Wind Energ, 2021), Kelly et al. (Wind Energ, 2022). I am not saying icing needs to be included in this review, I am just giving some examples.
- (sec. 5) It seems to me that some more internal discussion might be required as Oldenburg, which features at least one author on the author list, also has a significant amount of work on airfoils, typically for wind energy applications, but that seems to not be included, e.g., Cordes et al. (JFM, 2017), Wei et al. (JFM, 2019).
- (sec. 7.2) Does it make sense to also discuss de Jong Helvig et al. (JWEIA, 2021) and Bourhis & Buxton (Phys Rev Fluids, 2024) in this section too? These are two of the few studies in the literature that explicitly tests numerous different porous disk geometries and compare them. Many groups have conduct this sort of parametric research before their published study and as such the details are not included in their published papers.
- (ln. 1268) Castro’s (1971) results were recently re-evaluated and expanded by Cicolin et al. (JFM, 2024). Castro is also an author of that newer work. It is probably also worth inclusion.
- (sec. 7.3. 7.4, possibly also sec 11) It might be interesting to include a note on the findings of Neunaber et al. (JFM, 2025) on how periodicity in the wakes of porous discs can be induced by external forcing and that this is not the same for all porous discs. They also make comments on full scale turbine wake stability analysis.
- (sec 7, general) There is no real discussion or image of a wake map in general, à la Neunaber et al. (Energies, 2020), although things like the intermittency ring are alluded to. In a review of this scale, it seems such a map would be relevant (it does not need to be explicitly from that study, the present authors could come up with their own if relevant). Note, this is also somewhere where I noticed the lack of more specific wind turbine model discussion, specifically on wakes (although I do realise lab-scale models are touched on in section 4, 8 and 9).
- (ln. 2596) Active grids are mentioned as a way of making shear, but are not included, to my knowledge, in the citations provided. Some examples could be: Cekli & van de Water (Exp Fluids, 2010), Hearst & Ganapathisubramani (Wind Energ, 2017), and others already cited in section 2.2.
TYPOGRAPHICAL ERRORS:
- (ln. 128) “…changing atmospheric mostly turbulent wind conditions.” Something doesn’t read right here. It is perhaps missing commas or hyphens/dashes.
- (ln. 136) “respectively” unneeded.
- (ln. 199, 233) “aera” > “area”
- (ln. 251) “ling” > “long”
- (ln. 291) “x-wire” > “X-wire” This is stylistic, but I think it is much more commonly written with a capital X.
- (ln. 543, 793, 799, 924, 960, 964, 971) citations are in brackets that shouldn’t be in brackets.
- (ln. 578) “among others” is probably meant to refer to the listed citations in brackets, and thus should also be in the brackets.
- (ln. 650) The hot-wire “thickness” is written as 5 mm. Presumably this is wire diameter and meant to be 5 microns.
- (ln. 525, 709, 1298) “and” missing between citations.
- (ln. 807) The word “can” should be deleted.
- (ln. 865) Something seems wrong here. The sentence that ends the line does not appear to be part of the same sentence that starts the next line.
- (Fig 8) Are these figures original or are they from a published paper? If the latter, they should be cited.
- (sec. 7.3) Does the first paragraph go better in the previous section (7.2)? It is still discussing disc design.
- (sec. 7.3) The references to “Noriega et al. (2022)” which appears to be a conference paper can probably be replaced with the more recent Noriega & Mazellier (JFM, 2025).
- (ln. 1345) Period missing between the citations in the middle of the line.
Citation: https://doi.org/10.5194/wes-2026-51-RC3
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 643 | 376 | 29 | 1,048 | 23 | 28 |
- HTML: 643
- PDF: 376
- XML: 29
- Total: 1,048
- BibTeX: 23
- EndNote: 28
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
As active members of both EAWE and IEA Wind TCP Tasks 39/64/47/… we feel that important contents are missing to fulfil the requirements of a state-of-the-art review.
Particularly, we are surprised that this review does not include a description of the acoustic benchmark activities conducted in IEA Wind TCP Task 39/64 “Quiet Wind Turbine Technologies”/”RESONATE” which we would have expected to appear in either sections "Aeroacoustics" and/or "Benchmarking Wind Tunnels". IEA Wind TCP Task 47 appears more comprehensively represented in this publication.
A 2023 roadmap paper[1] published within Task 39 (and reviewed by relevant institutions and industry representatives) addresses still open challenges in aero-acoustic wind tunnel testing of low-noise wind turbine blade sections. As a consequence of the identified needs round robin tests on low-Re and high-Re number profile sections with and without noise reduction devices (different types of trailing-edge serrations) have been conducted among Task 39 partners in relevant acoustic wind tunnels to identify uncertainties and overall, to improve data quality.
An important forerunner initiative to harmonize profile trailing-edge (TE) noise datasets from different wind tunnels goes back to the year 2010 when, driven by wind energy industry, category 1 “trailing edge-noise” was founded as part of the series of AIAA/CEAS Workshops on Benchmark Problems for Airframe Noise Computations (BANC). These open data sets originating from a variety of different test facilities supplemented by the ongoing TE serration benchmarks in IEA Wind TCP Task 64 are completely omitted in the present article claiming to represent the current state of the art of (acoustic) wind tunnel testing for wind energy. One reference from the above collaborative efforts is cited in this article (Verges I Plaza et al., 2022), however, taken out of its context.
Beyond that, we have the following remarks without claiming to be exhaustive. We mainly address to aeroacoustics here although overall, there should be a better harmonization and fine-tuning also in the other sections.
Given the examples above we would recommend to further elaborate the article with major revisions.
With kind regards,
Alexandre Suryadi and Michaela Herr
[1] Bertagnolio, F., Herr, M., & Madsen, K. D. (2023). A roadmap for required technological advancements to further reduce onshore wind turbine noise impact on the environment. WIREs Energy and Environment, 12(3), e469. https://doi.org/10.1002/wene.469