Bio-Inspired Outer-Span Hybridization of the NREL Phase VI Rotor for Enhancing Aerodynamic Performance and Flow-Field Stability

EL Qamch, Yassine; Qaissi, Khaoula; Ali Omar, Ashraf; Lahmili, Wafae; Bouchaala, Kenza

doi:10.5194/wes-2026-111

Preprints

https://doi.org/10.5194/wes-2026-111

Preprints

22 Jul 2026

| 22 Jul 2026

Status: this preprint is currently under review for the journal WES.

Bio-Inspired Outer-Span Hybridization of the NREL Phase VI Rotor for Enhancing Aerodynamic Performance and Flow-Field Stability

Yassine EL Qamch, Khaoula Qaissi, Ashraf Ali Omar, Wafae Lahmili, and Kenza Bouchaala

Abstract. The presenting work evaluates a selective bio-inspired outer-span hybridization of the NREL Phase VI wind turbine rotor. An owl-inspired airfoil (OA-T), obtained through controlled thickness scaling of a reconstructed owl-wing section, is integrated over the outer 20 % of the blade span while preserving the original chord and twist distributions. The objective is to assess whether localized geometric modification in the torque-dominant tip region can improve aerodynamic performance under increasing loading. A validated CFD framework is employed at both airfoil and rotor scales. Transition SST simulations reproduce the experimental lift behavior of the owl-inspired airfoil, while steady RANS–MRF simulations capture the torque response of the baseline Phase VI rotor. Using the same modeling setup, the hybrid configuration demonstrates a systematic increase in peak power coefficient from 0.38 to 0.42 at a tip speed ratior λ = 5.41, reaching 0.448 at higher tip-speed ratios. Flow-field analysis reveals outward redistribution of tangential loading, a shift toward a more chordwise-distributed pressure response, delayed separation progression, and a more confined wake structure at elevated wind speeds. The results show that controlled outer-span hybridization within a benchmark rotor can enhance torque retention and aerodynamic stability without full-blade redesign.

Received: 22 Jun 2026 – Discussion started: 22 Jul 2026

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.

Download & links

Yassine EL Qamch, Khaoula Qaissi, Ashraf Ali Omar, Wafae Lahmili, and Kenza Bouchaala

Status: final response (author comments only)

Subscribe to comment alert

RC1: 'Comment on wes-2026-111', Anonymous Referee #1, 31 Jul 2026

From the point of view of this reviewer, the manuscript does not comply with the quality expected for the Wind Energy Science journal. The authors are encouraged to keep working in this fascinating field. The research line fits the interests of the journal and of the community around it.
MAJOR TECHNICAL COMMENTS
Why a 3D simulation for the airfoil?
Reynolds for Table 1?
Reynolds and lambda for Figure 5? pitch?
The manuscript is about tip design, but Figure 9 clips the tip
Figure 11: effect of pitch is missing for this comparison
OTHER MAJOR COMMENTS
Fixable by re-reading the manuscript:

- Changes in font size inside the same paragraph

- Heterogeneity in size and style of figures, which are generally small

- References without year (6-9 june)

- RANS defined two times

- Check how Equation 2 is written. What is "u"?
Potential AI-abuse signs:

- Excessive use of "-" (e.g. wind-turbine)

- Numbers sometimes with ",", sometimes without
Figure 1 is very confusing
Font size in Figure 2
Figure 6 does not serve its purpose
MINOR COMMENTS
Figure 8: y axis range
One cannot see clearly the mesh from Figure 4
NREL Phase VI blade mentioned in Section 2.1.4, after results have been presented
reynold -> Reynolds
Potential abuse of subsubsections
Could be interesting to include wall BC for Figure 3
Figure 14: it the plot is essentially 2D, one could remove the 3D effect we see for the geometry. Velocity legend is hard to read
Uncommented figures in the Appendix
Line 123: (())
There are many acronyms in the abstract (not defined), that unnecessary limit its readability
ADDITIONAL REMARKS
How could this idea scale to other wind turbine models?
The proposed loading redistribution may also come with some risks, as aerodynamic performance is not the only design driver for horizontal axis wind turbines. There seems to be no mention about it in the manuscript
While k-w SST has been established as the canonical turbulence modeling approach in the field, it is still RANS. The phenomena studied are very linked to separation, and it could be premature to firmly conclude about the potential of the technology simply based on the computations. Could be a future work

Citation: https://doi.org/10.5194/wes-2026-111-RC1
RC2:
'Comment on wes-2026-111', Anonymous Referee #2, 06 Aug 2026
The manuscript replaces the airfoil sections over the outer 20% of the NREL Phase VI blade span with a thickness-scaled owl-derived airfoil (OA-T) while preserving chord and twist, and reports a peak power coefficient increase from 0.38 to 0.42 at λ = 5.41 (0.448 at λ = 7.58) based on steady RANS–MRF simulations. The topic is of interest and the use of a well-documented benchmark rotor is appropriate. However, the manuscript suffers from a fundamental mismatch between the low-Reynolds-number premise of the airfoil design/validation and the operating conditions of the rotor to which it is applied, from an inconsistent turbulence-modeling strategy that undermines the physical interpretation of the results, and from a modeling approach (steady MRF) that cannot support the wake and turbulence claims made. There are also unresolved inconsistencies with the authors' own TORQUE 2026 conference paper, incorrect citations for the experimental reference data, and numerous presentation deficiencies.

Prior to recommending the paper for publication there are several things to be addressed:
Major Comments:
The authors state several times in the literature that the OA-T excels in low to moderate Re numbers. How is this relevant to the high-Re number that typically characterizes Wind Turbines (WT) Rotors. This is relevant to Sections 2.1.2 and 3.1.1

Regarding the validation in section two there are two issues: a) the difference in Re between the 2D airfoil simulations and the 3D WT simulations b)Drag comparison is missing. Is the data available?

The 2D airfoil simulations employ transition model (γ-Reθ). It is not clear what the authors do for the 3D one. They mention that he k–ω SST model "was adopted for all rotor simulations due to its robustness in resolving adverse pressure gradients and separation phenomena typical of wind-turbine blade operation," “but then in the analysis they suggest the existence laminar separation bubbles (LSB) etc. (again on this below)

Are the steady state MRF simulations adequate in the deeply stalled region of NREL VI rotor?

The conference paper (Qamch et al., 2026) replaces the outermost 30% of span; the present manuscript states 20%. Note that Fig. 10 (torque–velocity) appears essentially identical to Fig. 11 of the conference paper despite different replacement extent; if the geometry changed, the torque response should differ. The validation figures (Figs. 7 and 8) are also reused from the conference paper.

How is the geometric transition at r/R = 0.80 between the S809 (t/c = 20.95%) and the OA-T (t/c = 11.5%) handled.Please describe and show the transition region.

The authors compare sectional pressure-coefficient distributions for the baseline and hybrid blades and attribute the "small-amplitude fluctuations" observed in the hybrid Cp curves to "laminar separation bubbles in the mid- to aft-chord region" (lines 330–331). This interpretation cannot be sustained with the numerical setup described in Section 2.1.3:

-All rotor simulations employ the fully turbulent k–ω SST model(?).
-The simulations are steady-state.; chordwise oscillations of the kind visible in Figs. 16(d)–(f) are, in steady fully turbulent conditions. The authors should state explicitly whether the plotted Cp remain the same throughout the simulations. The fluctuations are reported already at U = 5 m/s (line 330), where the flow is described by the authors themselves as fully attached. Also these fluctuations are not evident in the contour plots. This is also the case for the (LSB) claimed in the Cp discussion that do not appear in any of the velocity/streamline figures (Figs. 13–15), and no skin-friction or wall-shear distributions are shown.
Minor Comments:
Line 154–155: grid independence is assessed at a single angle of attack (10°). Justify why this angle is chosen and why the conclusion of grid independence automatically transfer across higher Reynolds numbers.

Fig. 5 / lines 174–178: describe how the rotor mesh is refined between the six levels (uniform scaling? local refinement zones? prism-layer growth held fixed?). Grid-convergence is not interpretable without the refinement strategy

Table 1 caption: "time dependent Lift Coefficient" — the simulations are steady; correct the caption.

Eq. (2) symbols are not clearly explained.

Terminology: sectional pressure-coefficient distributions are repeatedly called "Cp loops" and their "evolution" (lines 323–338). This is non-standard; clarify what "chordwise evolution" means.

Figure ordering and referencing: Figs. 13–14 (5 and 15 m/s) are discussed, then Fig. 15 (10 and 13 m/s) is introduced afterwards, which breaks the wind-speed progression and is confusing. Reorder figures to follow the narrative (or vice versa). Fig. 14 is referenced before Fig. 15 but placed inconsistently relative to the discussion.

Naming: "Model A" vs "Hybrid Model A" are used interchangeably (e.g., Figs. 9, 10, and text); use one designation throughout.

Legends and colorbars are unreadable in Figs. 13, 14, 17, 18, and 19; Fig. 1 lacks a legend. All must be regenerated at readable size.

Fig. 19: the 3D streamline views as presented do not clearly demonstrate the claimed wake differences; Quantitative wake cuts (e.g., velocity-deficit or vorticity contours at downstream planes, tip-vortex trajectories would clear things up

Line 17: "ratior" → "ratio"; line 11: "The presenting work" → "The present work"; general language editing is required throughout
Citation: https://doi.org/10.5194/wes-2026-111-RC2
RC3:
'Comment on wes-2026-111', Anonymous Referee #3, 09 Sep 2026
Review of "Bio-Inspired Outer-Span Hybridization of the NREL Phase VI Rotor for Enhancing Aerodynamic Performance and Flow-Field Stability" (wes-2026-111)
Recommendation: I recommend rejection of this manuscript in its present form.
General comments
The manuscript is well phrased and clearly written, with a comprehensive literature review, and I appreciate that the authors chose an incremental, controlled outer-span hybridization approach rather than jumping directly to a full-blade bio-inspired redesign — this is a sound way to isolate the effect of the airfoil substitution. However, in its present form the study has two problems severe enough that the conclusions cannot yet be trusted, and both need to be resolved before the level of this work can be considered adequate for publication.
From a numerical point of view, the manuscript repeatedly shows signs of unresolved convergence difficulties (see specific comments below): small-amplitude, spatially inconsistent oscillations in the sectional Cp curves, "laminar separation bubbles" that are invoked in the text but not evidenced in any flow-field, skin-friction, or pressure plot, and a modeling chain (steady-state, fully turbulent, MRF) that is not equipped to resolve the very unsteady, separated phenomena the authors use to explain their results. These are not cosmetic issues — they call into question the reliability of the reported flow-field interpretation and, by extension, part of the physical explanation offered for the performance gain.

From an engineering point of view, I do not see a clear justification for replacing an airfoil section that was designed and validated to operate efficiently at a Reynolds number of the order of 900,000 (i.e., representative of the outer span of a utility/benchmark-scale wind-turbine rotor) with a bio-inspired airfoil that the authors themselves, in this manuscript and in their prior work, describe as optimized for low Reynolds numbers. This mismatch between the premise under which OA-T was designed/validated and the regime in which it is actually deployed here is central to the whole study and needs to be explicitly confronted, not left implicit.

Beyond these two points, several specific methodological choices (single-angle-of-attack grid independence at the airfoil level, torque-only grid convergence at the rotor level, the handling of the abrupt thickness discontinuity at r/R = 0.80, and the use of steady RANS-MRF to support claims about stall, wake, and turbulence behavior) need clarification or additional evidence. I list these below as specific comments, followed by technical corrections.
Specific comments
Reynolds-number mismatch between the airfoil design premise and the rotor application (Sections 2.1.1, 2.1.2, 3.1.1). The owl-inspired airfoil is reconstructed from a wing cross-section at ~40% of the (biological) wingspan (lines 109–110) and is validated/discussed throughout as a low-Reynolds-number, low-Mach device. It is not clear how these inflow conditions — appropriate for a small, slow fixed wing — compare to the tip region of a large-radius, rotating wind-turbine blade, where Reynolds numbers are far higher and rotational/centrifugal effects are present. Please state explicitly the Reynolds numbers involved at each stage (owl-airfoil validation, 2D OA-T validation, 3D rotor outer span) and justify why an airfoil designed and validated for low-to-moderate Re is expected to remain effective, or even physically sensible, when it replaces a section originally selected to work at Re ≈ 9×10⁵.

Abrupt thickness discontinuity at the hybridization boundary (Section 2.1.1, r/R = 0.80). The baseline NREL Phase VI blade has t/c ≈ 20.95% at and beyond r/R = 0.80 (Table A1), while OA-T has t/c = 11.5% (line 124). This is close to a factor-of-two difference in thickness at the join. Why was OA-T not thickness-scaled to match the local baseline thickness at r/R = 0.80 instead of being inserted at a fixed 11.5%? How is the resulting geometric discontinuity handled in the blade reconstruction, and could it plausibly be the source of some of the numerical convergence difficulties observed later in the manuscript (e.g., the Cp fluctuations discussed in comment 6 below)? Please describe and, ideally, show the transition/blending region.

Grid independence at the airfoil level uses lift coefficient alone, at a single angle of attack (Section 2.1.2, lines 154–159, Table 1). Since the airfoil is specifically "designed" to handle separation, the appropriate convergence metric should include the full polar (Cl, Cd, and ideally Cm), and should be checked at higher angles of attack within and beyond stall (experimental measurements available at the cited studies), not only at 10°. Alternatively, if a single-AoA check is considered sufficient, please justify the 10° choice and support it with more detailed diagnostics (pressure and skin-friction distributions), particularly in regions of strong pressure gradient and boundary-layer development.

Mesh element size given without reference length (Section 2.1.3, line 171). The minimum surface element size on the blade is quoted as 0.01 m. Please express this relative to a meaningful length scale — the rotor radius, or better, the local chord in the OA-T region — so the resolution can be judged.

Rotor-level grid convergence uses only an integrated quantity (shaft torque) (Section 2.1.3, lines 172–179, Fig. 5). As with comment 3, an integrated metric is a weak basis for judging grid adequacy specifically in the outer-span, separation-prone region that is the focus of this study. Spanwise distributions such as Fn(r), Ft(r), or better Cp(r), Cf(r) at the operating conditions of interest would give a far more convincing (and more relevant) demonstration of mesh independence for the flow features being discussed later.

Adequacy of steady-state MRF simulations for the phenomena being claimed (Sections 2.1.3, 3.1.1, 3.3–3.5, general). The stated advantage of Hybrid Model A is improved stall handling near the blade tip, and laminar separation bubbles are repeatedly invoked to explain both the 2D validation offset (lines 211–220) and the 3D sectional Cp behavior (lines 326–338). Both stall progression and LSB dynamics are inherently unsteady phenomena. Steady-state MRF simulations are not, in general, adequate to resolve them, and this is compounded by an inconsistent turbulence-modeling strategy: the 2D airfoil validation uses the Transition SST (γ–Reθ) model, while "the k–ω SST model was adopted for all rotor simulations" (line 167) — i.e., fully turbulent, with no transition modeling. Given this, the LSB claim made in Section 3.4 for the 3D rotor results (lines 330–331) cannot be supported by the numerical setup as described:
Figure 7 (2D validation) provides no direct evidence (flowfield data, Cp, Cf, or γ distributions) that an LSB is actually resolved even at the airfoil level;

At the rotor level, the small-amplitude, repeating fluctuations seen in the sectional Cp curves (Fig. 16 d–f) are, in a steady, fully turbulent simulation, more consistent with a difficulty in the convergence of the numerical scheme than with a physical LSB. A genuine LSB would be expected to produce a single, sharp change in the slope of Cp marking transition/reattachment, not repeated small-amplitude oscillations;

These oscillations are also reported already at U∞ = 5 m/s, where the authors themselves describe the flow as fully attached, and they are not evident in the corresponding contour/streamline plots (Figs. 13–15), nor are any skin-friction or wall-shear distributions shown to support the claim. The authors should state explicitly whether the plotted Cp values are converged/steady in time, and either substantiate the LSB interpretation with the appropriate flow diagnostics or withdraw/qualify the claim.

Global performance claims rest on the same questionable numerical approach (Section 3.2, lines 271–283). The reported gains (Cp from 0.38 to 0.42 at λ = 5.41, up to 0.448 at λ = 7.58) and their implication for rotor downsizing/efficiency are a reasonable conclusion in principle, but they are derived entirely from the steady RANS–MRF framework whose adequacy for the separated/stalled flow regime is questioned throughout this review (see comments 3, 5, and 6 above). The performance numbers cannot be assessed independently of these modeling concerns.

Local aerodynamic claims are made without a supporting pressure distribution (Section 3.2, lines 257–262). The statement that "the OA-T section's thicker and more curved profile enhances local suction and circulation while promoting a smoother pressure recovery near the tip" is asserted at this point in the manuscript with no accompanying Cp plot — the sectional pressure-coefficient comparisons are not introduced until Section 3.4, several pages later. How are these specific claims (enhanced suction, smoother pressure recovery) supported at the point they are made? Either move the relevant Cp evidence earlier or qualify the claim as anticipatory of Section 3.4.

Direct, co-located comparison plots are missing throughout. Fig. 16 presents the baseline and Hybrid Model A sectional Cp curves in separate panels (a–c vs. d–f) rather than overlaid, and Figs. 17–18 similarly show the two configurations' surface pressure contours side by side rather than directly compared. A direct overlay of baseline vs. Model A at each radial position and wind speed would be far more revealing than the current side-by-side presentation and would make the claimed differences easier to verify independently.

Technical corrections
Nomenclature (p. 1/2): keep symbol case consistent — use lower-case p for (static) pressure and upper-case P for power throughout; as listed, "P — Static pressure (Pa)" conflicts with "P — Power coefficient" / "P shaft — Shaft power (W)".

Line 189: "the outer 0% of the blade span (r/R ≥ 0.80)" — this contradicts the Abstract ("outer 20%," line 13) and the same paragraph's own later statement, "restrict the modification to the last 20% of the span" (line 193). The digit appears to have been dropped; please correct to "outer 20%" and check for any other instance where a figure may have been lost in formatting.

Lines 212–214: the sentence citing the LSB literature is awkwardly broken across a paragraph/line boundary — "...explicitly reported by (Anyoji et al. ; Qamch, Omar, et al., 2026)" is followed by a stranded ", who associated it with the formation and evolution of a laminar separation bubble." Please reflow this passage.

Figures 17–18 (pressure-contour panels): clearly label which side is the pressure side and which is the suction side within each (a)/(b) panel.

General language/consistency pass recommended throughout (e.g., "Model A" vs. "Hybrid Model A" used interchangeably).
Citation: https://doi.org/10.5194/wes-2026-111-RC3

Yassine EL Qamch, Khaoula Qaissi, Ashraf Ali Omar, Wafae Lahmili, and Kenza Bouchaala

Viewed

Total article views: 443 (including HTML, PDF, and XML)

HTML	PDF	XML	Total	BibTeX	EndNote
288	84	71	443	80	58

HTML: 288
PDF: 84
XML: 71
Total: 443
BibTeX: 80
EndNote: 58

Views and downloads (calculated since 22 Jul 2026)

Month	HTML	PDF	XML	Total
Jul 2026	95	27	10	132
Aug 2026	87	17	8	112
Sep 2026	106	40	53	199

Cumulative views and downloads (calculated since 22 Jul 2026)

Month	HTML	PDF	XML	Total
Jul 2026	95	27	10	132
Aug 2026	87	17	8	112
Sep 2026	106	40	53	199

Viewed (geographical distribution)

Total article views: 418 (including HTML, PDF, and XML) Thereof 418 with geography defined and 0 with unknown origin.

Country	#	Views	%

Latest update: 21 Sep 2026

Short summary

This study explores how wing shapes inspired by owls can improve the performance of wind turbine blades. The research was carried out to address efficiency losses and noise issues in conventional designs. Using detailed computer simulations, the proposed shapes were tested and compared with standard blades. The results show measurable improvements in energy capture and flow behavior, suggesting that nature-inspired designs can contribute to quieter and more efficient wind energy systems.


Total:	0
HTML:	0
PDF:	0
XML:	0