the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 29 Jan 2026
Vortex generator design for unsteady flow separation control and dynamic stall suppression on pitching thick airfoils
Abstract. This study experimentally investigates the performance of vortex generators (VGs) designed for maximising lift-to-drag ratio in steady conditions to prevent unsteady flow separation. Surface pressure measurements are conducted in the TU Delft low-speed wind tunnel on a DU-97-W-300 airfoil undergoing pitch oscillations while equipped with VGs of various vane sizes and shapes. In steady conditions, vanes with heights smaller than the local boundary layer thickness optimally balance the stall delay with maximum lift-to-drag ratio among the tested triangular vane VGs. However, these same VGs with vane heights smaller than or equal to the steady local boundary layer thickness are insufficient to suppress unsteady flow separation in all pitching cycles. VGs whose vane height exceeds the local boundary layer thickness for a larger part of the pitch cycle prevent unsteady flow separation and restrict the upstream movement of the stall vortex for a larger percentage of cycles. Flow separation is less likely at higher reduced frequencies, making the number of separated flow pitching cycles less sensitive to the VG vane size. Contrary to past literature, rectangular vanes yield a higher steady aerodynamic efficiency than triangular vanes. Rectangular vanes also suppress unsteady flow separation in all pitching cycles at all tested reduced frequencies, indicating overall stronger streamwise vortices than triangular vanes and proving to be a better VG shape for steady and unsteady stall suppression on thick airfoils.
- Preprint
(8713 KB) - Metadata XML
- BibTeX
- EndNote
Status: final response (author comments only)
- RC1: 'Comment on wes-2026-7', Anonymous Referee #1, 15 Feb 2026
-
RC2: 'Comment on wes-2026-7', Anonymous Referee #2, 19 Mar 2026
The manuscript presents an experimental investigation of vortex-generator performance for controlling steady and unsteady flow separation on a DU97-W-300 thick airfoil. The authors particularly focus on the vane height and the vane shape effects under dynamic stall conditions for pitching airfoil. The topic is relevant for modern wind-energy applications, since thick airfoils are increasingly used in large wind-turbine blades and vortex generators are widely applied as passive flow-control devices to improve aerodynamic performance under both roughness-sensitive and unsteady operating conditions. The comparison between triangular and rectangular vane geometries, together with the assessment of several reduced frequencies, provides useful experimental data in a research area where unsteady measurements remain relatively scarce. In particular, the observation that vortex-generator configurations optimized under steady conditions do not necessarily behave identically under pitching conditions is an important practical outcome for future blade-design considerations.
The experimental effort is appreciated and the manuscript is generally well structured. However, some aspects of the physical interpretation require clarification, particularly where pressure-derived separation-location results are used to infer vortex dynamics during dynamic stall.
- The first point that requires clarification is about the interpretation of the dynamic stall process, particularly the repeated description that the “stall vortex moves upstream” as inferred from the pressure-derived separation location. The separation point is identified from the onset of a pressure plateau along the suction side, which can be considered as a reasonable indicator of the upstream progression of separated flow. However, this quantity should not be directly interpreted as the motion of the stall vortex itself. In the classical dynamic stall development, trailing-edge separation indeed progresses upstream as the angle of attack increases, but the dynamic stall vortex (or leading-edge vortex) forms near the leading edge once a critical separation state is reached and subsequently advects downstream over the airfoil. Therefore, the current wording appears physically ambiguous, since the pressure-based analysis more directly reflects separation-front migration rather than vortex-core motion. A clearer distinction between these two phenomena would strengthen the physical interpretation of the experimental results presented in the manuscript.
- A related point concerns the repeated conclusion that rectangular vanes generate stronger streamwise vortices than triangular vanes, based primarily on the observed pressure distributions, stall delay, and separation-location behavior. While the experimental trends clearly indicate that the rectangular vane configuration performs better in delaying separation and reducing cycle-to-cycle variability under the tested conditions, the measurements do not directly quantify vortex strength, persistence, or topology within the boundary layer. Since the analysis relies mainly on surface-pressure information and force measurements, it would be more appropriate to formulate this conclusion as an indication of more effective boundary-layer energization rather than direct evidence of stronger streamwise vortices. A slightly more cautious interpretation would better reflect the capabilities of the employed measurement approach.
- A further point concerns the interpretation based on the ratio between vane height and local boundary-layer thickness, which is used throughout the manuscript to explain the relative effectiveness of the tested VG configurations. The comparison is made using boundary-layer thickness values obtained from steady RFOIL calculations on the baseline airfoil without VGs. While this provides a useful reference, the central conclusions of the paper concern pitching conditions involving strongly unsteady separation and reattachment, where the instantaneous boundary-layer characteristics may differ significantly from the steady baseline estimate. It would therefore be useful to discuss more explicitly the limitations of using a steady reference boundary-layer thickness when interpreting VG effectiveness under dynamic stall conditions, particularly since the h/δ argument is central to several of the manuscript’s conclusions.
- One of the most interesting observations in the manuscript is the appearance of two distinct reattachment paths during the downstroke phase for some triangular VG configurations, particularly visible in the normal-force hysteresis loops and corresponding separation-location maps. Since this appears to represent genuine cycle-to-cycle variability, a quantitative indication of the fraction of cycles following each path at different reduced frequencies would further strengthen this result. In the corresponding discussion, the physical explanation is again formulated in terms of stall-vortex movement, whereas the presented pressure-based evidence more directly indicates differences in separation-front evolution between cycles. A clearer distinction here would help interpret the origin of the observed dual reattachment behavior more robustly.
- A brief discussion of finite-span effects would also be useful, considering that the tested model has an aspect ratio of approximately 1.92 and the vortex-generator arrays do not extend fully to the tunnel walls. Since both dynamic stall development and vortex-generator-induced boundary-layer modification are inherently three-dimensional phenomena, some comment on the possible influence of sidewall proximity and spanwise non-uniformity would help define the limits of the present interpretation, particularly where the results are discussed in quasi-two-dimensional terms.
- The manuscript would benefit from a clearer definition of the criterion used when stating that one VG configuration performs “better” than another, since different sections refer alternately to stall delay, aerodynamic efficiency, cycle-to-cycle variability, and suppression of separated cycles.
- Since all VG arrays are tested only at x/c = 0.30, a brief discussion on how the conclusions may depend on chordwise placement would improve practical interpretation, especially for modern blade applications where more downstream locations are also used.
- The discussion on reduced-frequency effects could be expanded slightly, particularly regarding whether the reduced sensitivity to VG size at higher k is attributed solely to phase lag or also to modified vortex formation timing.
Citation: https://doi.org/10.5194/wes-2026-7-RC2
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 249 | 100 | 17 | 366 | 18 | 36 |
- HTML: 249
- PDF: 100
- XML: 17
- Total: 366
- BibTeX: 18
- EndNote: 36
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
1) Review
I appreciate the effort invested in this work and its focus on aerodynamic topics that are highly relevant to modern ultra-large wind turbine blades. In contemporary designs, thicker airfoils are increasingly used in the more outboard sections, and the application of vortex generators (VGs) around the 30% thick airfoil is becoming more common. With that context in mind, I offer the following suggestions to help improve clarity and strengthen the manuscript:
# Clarification of unsteady flow separation in the abstract: I recommend providing a sharper and more explicit explanation of “unsteady flow separation” in the abstract. A clearer definition or brief description of the phenomenon and its relevance to the study would help readers immediately grasp the motivation and significance of the work.
# Comparison and evaluation of VG shapes: The manuscript suggests that the rectangular vane shape may be a preferable option. However, as indicated in Figure 6, the rectangular or larger VG array appears to induce more abrupt separation and/or unexpected hysteresis characteristics. It would be helpful to clarify the criteria used to determine which VG shape performs better. Are the conclusions primarily based on polar curves, stall delay, hysteresis behavior, or overall aerodynamic efficiency? A more explicit definition of “better” performance would strengthen the argument.
# Chordwise location of VGs: The study considers only a 30% chordwise VG location. However, for modern large blades, mid-chord placements (approximately 40–60% chord) are also commonly considered and may be more representative of practical applications. Since VG effectiveness is highly sensitive to chordwise position, the conclusions may depend strongly on this parameter. It would be valuable to discuss how different chordwise placements might influence the results and whether the current conclusions are specific to the 30% location.
# Conciseness of the conclusion section: A more concise and focused conclusion section may improve the overall impact of the paper by clearly summarizing the key findings and their practical implications.
I hope these comments are helpful in further strengthening the manuscript.
2) Some minor errors to be corrected
# Consistent angle of attack range between Figure 6 and Figure 8
# Line 44 => vane heights
# Line 65 => were one of the first to show
# Line 86 => energising
# Line 88 => is proportional to the vane height
# Line 93 => create an additional variable
# Line 150 => etc., have a significantly smaller impact
# Line 157 => high angles of attack
# Line 165 => shows that dynamic stall…
# Line 166 => exceeds
# Line 169 => its ultimate upstroke location
# Consistency: 10◦±10◦ or 10◦± 10
# Consistency: vane type VGs or vane-type VGs