Acta Mech 235, 7765–7787 (2024)

[Link]

O R I G I NA L PA P E R

Hariprasanth Palanivel · Rinku Mukherjee

Airfoil-shaped vortex generators for separation control and

drag reduction on wind turbine blades

Received: 28 May 2024 / Revised: 30 August 2024 / Accepted: 1 October 2024 / Published online: 21 October 2024
© The Author(s), under exclusive licence to Springer-Verlag GmbH Austria, part of Springer Nature 2024

Abstract A passive flow control device, Clark-Y airfoil-shaped vortex generator (VG) on NREL Phase VI
turbine blade, which has s809 airfoil section, is investigated. Both qualitative oil flow visualization from
wind tunnel experiments and quantitative measures of aerodynamic coefficients using steady-state CFD with
OpenFOAM are reported. Airfoil-shaped VGs are proposed and compared with traditional rectangular and
triangular VGs. The use of airfoil-shaped VGs to delay separation, improving aerodynamic efficiency, inducing
local pressure peaks and augmenting vorticity in the flow field are reported in detail. Results show that blades
equipped with airfoil-shaped VGs provide a 5% lift coefficient increase and a 27.68% drag coefficient reduction
compared to clean blades at a stall angle of α = 11◦ . Airfoil-shaped VGs also generate more vorticity
downstream compared to conventional VGs, contributing to maximum increase in peak vorticity inducing an
additional momentum to the flow to delay separation without significant drag penalty. Thus, airfoil-shaped
VGs offer a promising alternative to traditional VG designs.

List of symbols

α, AoA Angle of attack [deg]

AR Aspect ratio
b Span of blade [m]
βV G Angle of inclination of VG with incoming flow [deg]
c Chord of airfoil/blade [m]
Cl , Cd , Cm Lift, drag, momentum coefficient of 2D airfoil
CL , CD , CM Lift, drag, momentum coefficient of 3D blade
C L /C D Aerodynamic efficiency
Cp Coefficient of pressure
 Error in grid convergence study
GC I Grid convergence index [%]
HVG Height of VG
I Turbulent intensity
k Turbulent kinetic energy [m2 /s2 ]
L VG Length of VG
L cell Representative cell length [m]
μ Dynamic viscosity [kg/ms]
ω Specific dissipation rate [1/s]

H. Palanivel (B) · R. Mukherjee

Aerodynamics and Flow-Physics Laboratory, Department of Applied Mechanics and Biomedical Engineering,
Indian Institute of Technology Madras, Chennai, Tamil Nadu 600036, India
E-mail: haribabupv354@[Link]
7766 H. Palanivel, R. Mukherjee

p Pressure [N/m2 ]
φ Performance factor in grid convergence study
r Grid refinement factor
Re Reynolds number
ρ Density [kg/m3 ]
u, v, w Velocity components in stream-wise, normal, span-wise directions [m/s]
U∞ Free-stream velocity [m/s]
 Vorticity [1/s]
xVG Chord-wise location of VGs
x, y, z Cartesian coordinates in stream-wise, normal, span-wise directions [m]
z VG Distance b/w the trailing edges of single VG pair
Z VG Distance b/w two adjacent VG pairs

Acronyms and abbreviations

AEP Annual energy production

AVG Airfoil-shaped vortex generators
CFD Computational fluid dynamics
HWAT Horizontal axis wind turbines
LE Leading edge
LSB Laminar separation bubble
NREL National renewable energy laboratory
OpenFOAM Open-source field operation and manipulation
RVG Rectangular-shaped vortex generators
SST Shear stress transport
SIMPLEC Semi-implicit method for pressure linked equations-consistent
TE Trailing edge
TVG Triangular-shaped vortex generators
VG Vortex generators

1 Introduction

The declining availability of traditional fuels and the escalating demand for sustainable alternatives have
propelled wind energy to the forefront of renewable energy solutions. Wind power, comprising a mere 20%
share of global electricity production in 2010, surged to 29% by 2020 and is anticipated to surpass 80−90% by
2050 [1]. This remarkable growth stems from wind energy’s unique advantages: sustainability, accessibility,
cost-effectiveness, minimal maintenance, job creation, and environmental friendliness. As the renewable energy
sector aims for net-zero emissions by 2050 [2], the capacity of wind turbines is poised for significant expansion,
fueling research into innovative turbine blade designs to enhance efficiency and reliability.
A prominent challenge in wind turbine design lies in boundary layer separation that threatens turbine
performance by diminishing lift and escalating drag and structural damage. Various flow control methods have
emerged to tackle this issue, manipulating airflow over the blades and optimizing aerodynamic performance.
Among these methods, vortex generators (VGs) are passive devices widely integrated onto wind turbine blades
to delay boundary layer separation [3]. VGs, affixed to the blade inclined to the flow, generate stream-wise
vortices downstream, providing extra energy to the boundary layer, thus averting separation. This process
entails channeling high-energy fluid from the free stream to augment kinetic energy at the boundary layer,
thereby managing stall occurrences and enhancing the lift coefficient of the blade, consequently improving
power generation capabilities [4]. VGs height should be 10−50% of boundary layer thickness [5].
Numerous analyses have scrutinized the efficacy of VGs for mitigating stall and augmenting wind turbine
blade performance. One of the pioneering experiments was conducted by Sullivan (1984) [6] on a 2.5 MW
Boeing MOD-2 turbine rotor equipped with counter-rotating rectangular VGs, reporting a notable increase of
approximately 11% and 20% in annual energy production (AEP) for mid-span and full-span VG configurations,
respectively. Further studies corroborated these findings through experimental wind tunnel investigations [7,8].
Moon et al. [9] studied VGs’ performance on a turbine blade of 2.3 MW capacity via experimental and
CFD analysis, observing AEP increase of 1.87% in experiments while their CFD simulations predicted torque
Airfoil-Shaped Vortex Generators 7767

and AEP gains of 2.8% and 0.81%, respectively, at 10 m/s. Mueller−Vahl et al. [10] found that, for effective
stall control, VGs should be placed at 15−20% chord position to obtain the optimal aerodynamics. In contrast,
Baldacchino et al. [11] emphasized that both chord-wise position and vane height play a crucial role in
delaying stall on a DU97-W-300 airfoil section. Choudhry et al. [12] explored the effect of three passive flow
control strategies through wind tunnel experiments, including VGs, on a thick NACA0021 wind turbine airfoil,
concluding that VGs significantly delay dynamic stall onset and enhance post-stall behavior. De Tavernier et al.
[13] also demonstrated VGs’ efficacy in controlling dynamic stall on the DU17DBD25 airfoil experimentally.
Computational investigations have delved into VGs’ impact on turbine aerodynamics, revealing delayed
separation due to three-dimensional rotational effects at the blade’s root [14]. Wang et al. [15] computationally
analyzed VG application on an S809 airfoil, noting increased lift coefficients and stall delay, especially with
double VGs. Gao et al. [16] studied the influence of triangular VGs’ height, length, and spacing between them
on a DU97-W-300 blunt TE airfoil computationally. The authors found that increasing the size of VGs reduced
lift and drag performance, and separation is delayed effectively when the spacing between pairs increases. Zhu
et al. [17–19] evaluated unsteady aerodynamic responses with VGs of wind turbine airfoils, observing dynamic
stall suppression and reduced lift hysteresis intensity, with chord-wise position affecting performance. Recent
studies explore VGs’ application in delaying separation on tidal turbine blades [20,21].
Despite VGs’ efficacy in stall control, a minor drag penalty persists that needs to be addressed [22]. For
this purpose, Hansen et al. [23] proposed aerodynamically shaped VGs to mitigate drag penalties, achieving
significant reductions compared to conventional shapes. Soto et al. [24] also affirmed improved aerodynamics
with airfoil-shaped VGs, highlighting ongoing efforts to refine VG design to address drag penalty.
Given the limited existing research on the potential impact of aerodynamically shaped VGs on wind
turbine blades, there is a pressing need to analyze the potential advantages such VGs could offer in enhancing
aerodynamic efficacy and optimizing power generation for wind turbines. Hence, the authors aim to identify
the probable merits and restrictions of airfoil-shaped VGs. Figure 1 shows the typical rectangular baseline
blade where the local aerodynamic characteristics are studied at the midsection of the blade.
The investigation involves oil flow visualization and the steady-state CFD simulation using the open-
source CFD solver OpenFOAM. Comparison is made between 3D rectangular clean blade and blade with
airfoil-shaped VGs in oil flow visualization. Simulations are run on a clean blade and compared with various
configurations of VGs arranged on the blade. The simulations are expected to provide useful insights on the
advantages of airfoil-shaped VGs to improve wind turbine performance.

2 Experimental setup

Experiments are carried out in the low-speed subsonic wind tunnel shown in Fig. 2a in the Department of
Applied Mechanics and Biomedical Engineering, IIT Madras. The wind tunnel has an open test section shown
in Fig. 2b with a cross section of 1 m × 1 m with length of 1.2 m. The maximum speed achievable is 20 m/s.
A mixture of oleic acid, titanium dioxide, and vacuum pump oil is sprayed over the blades and used to
visualize the flow over the surface.

Fig. 1 3D blade simulated (local aerodynamic characteristics are calculated at the root/midsection)
7768 H. Palanivel, R. Mukherjee

Fig. 2 Wind tunnel setup

Fig. 3 Airfoil profiles: a s809 airfoil for blade, b Clark-Y airfoil for VGs

Fig. 4 3D printed blade models: a Clean blade, 3D printed in 4 parts, to be glued together for full span b = 0.7m b 1/4 part of
Blade with aero-VGs

The blade models are made from polylactic acid (PLA) using the 3D printing facility at IIT Madras. The
airfoil section of the blade and the VGs are s809 and Clark Y, respectively, as shown in Fig. 3. The surface of
the blades are smoothed using spray paint and sandpapers sheets.
A full blade model has chord, c = 0.1464 m and span, b = 0.7 m, hence aspect ratio, A R = 4.78. Due to
size constraints, it was not possible to 3D print such a full-scale model. They are instead printed in four equal
parts as shown in Fig. 4, with each part now having reduced aspect ratio, A R = 1.19. Oil flow visualization
is shown for one such part, i.e., 1/4th of the full-scale blade, which has five pairs of VGs.
The VGs are of height = 7.3 mm and width = 14.6 mm, which is the least size that can be printed in
the 3D printing facility at IIT Madras without compromising the shape of the airfoil. Anything less than this
looks rectangular or triangular. They are placed at x/c = 0.2 from the leading edge along span. The distance
between two VGs in a pair is 17.5 mm and the distance between adjacent VG pairs is 35 mm. The experiments
are conducted for blades with and without VGs at Re = 1.46 × 105 , which is based on the chord of the blade.
Airfoil-Shaped Vortex Generators 7769

Fig. 5 CFD simulation cases

Fig. 6 Portion of the full wind turbine blade simulated

3 Computational details

This section describes the computational details of simulations. This includes a description of the different VG
configurations, the formulation of governing equations, the selection of numerical schemes, and turbulence
modeling. Details of mesh generation, grid convergence studies, and boundary conditions are also described.
The numerical scheme applied to the s809 airfoil and Clark-Y is validated with existing literature that
involves aerodynamic coefficients and pressure distributions. The different cases simulated in this study are
shown in Fig. 5.
The selection of the s809 airfoil for the blade used for this study is motivated by its widespread utilization
in wind turbine aerodynamics research, which has a maximum thickness at 39.5% of the blade’s chord of 21%
as depicted in Fig. 3a [25]. The VGs have a Clark-Y airfoil as a section, with a maximum thickness of 11.7%
at 28% of its chord, as shown in Fig. 3b.
A CAD design of a 2D section of the blade with an arrangement of two VGs is shown in Fig. 6. The clean
blade is compared with a blade with rectangular (RVGs), triangular (TVGs) and airfoil-shaped VGs (AVGs)
installed at 20% c.
It is important to mention here that while 3D analysis is done on 3D geometries of the VGs, only a strip of
the blade is considered with continuous boundary conditions defined for its edges. The "symmetry" boundary
conditions are used to simulate this using CFD. Hence, while the blade considered is 3D, only the strip is
analyzed using CFD, where although continuity is maintained at the edges, tip vortices are not considered. For
the VGs, on the other hand, the 3D analysis does consider tip vortices. This is done to specifically study the
3D effects of VGs without considering the 3D effects of the blade itself. Since VGs are considerably smaller
than the blade, it is essential to understand their impact on the behavior of the blade clearly.

3.1 VG configurations

The study compares RVGs, TVGs and AVGS as illustrated in Fig. 7. Clark-Y airfoil shown in Fig. 3b is chosen
for AVGs as also reported by Mendez et al. [22].
7770 H. Palanivel, R. Mukherjee

Fig. 7 Schematic of simulated VGs on turbine blade

Fig. 8 A single RVG pair showing influencing parameters

Fig. 9 Computational domain

VGs are positioned from the leading edge at 20%c as reported by Mueller et al. [10], where it was suggested
to place VGs within 15−25%c for an optimal aerodynamic performance.
The height (HVG ) and length (L VG ) of each VG are fixed at 1%c and 2%c, respectively, and they are inclined
at an angle βVG = 18◦ relative to the incoming flow as shown in Fig. 8.
The distance between a single VG pair is thrice the VG height, i.e., z VG = 3 × HVG . The s809 airfoil
utilized in this study has a chord length and span of 1000 mm and 60 mm, respectively. The choice of the span
is to be twice the intra-pair distance, i.e., b = 2 × z VG of a VG pair, ensuring sufficient accommodation for
one pair. Notably, all VGs are expected to generate counter-rotating vortices.

3.2 Governing equations and numerical scheme

The continuity equation is shown in Fig. 1, and 3D steady-state incompressible Reynolds-averaged Navier–
Stokes equations are shown in Fig. 2. The k − ω SST turbulence model [26] is solved using OpenFOAM, an
Airfoil-Shaped Vortex Generators 7771

open-source C++-based CFD solver.

∂u i
=0 (1)
∂ xi
   
∂u i ∂p ∂ ∂u i ∂u j  
ρu j =− + μ + − ρ ui u j (2)
∂x j ∂ xi ∂x j ∂x j ∂ xi
The transport equations of SST k − ω model governing turbulent production and dissipation rate are explicitly
formulated in Eqs. (3) and (4), respectively.
 
∂k ∂k ∗ ∂ ∂k
+uj = Pk − β kω + (ν + σk νT ) (3)
∂t ∂x j ∂x j ∂x j
 
∂ω ∂ω ∂ ∂ω 1 ∂k ∂ω
+uj = αS 2 − βω2 + (ν + σω νT ) + 2(1 − F1 )σω2 (4)
∂t ∂x j ∂x j ∂x j ω ∂ xi ∂ xi
The turbulence viscosity is obtained from
k
νt = a 1 (5)
max (a1 ω, b1 F23 S)
The model/closure coefficients and the auxiliary equations can be found in Menter’s work [27]. The choice of
turbulence model is motivated by its proven efficacy for flows characterized by adverse pressure gradients, as
demonstrated by Menter [26]. Additionally, comparative analysis of three distinct turbulence models prevalent
in literature, namely k − , Spalart-Allmaras (S A), and k − ω SST, shows that while all models perform
admirably in predicting flows at low angles, but at high angles of attack, the k − ω SST demonstrates a closer
match with experimental data.
Finite volume method is used to discretize Eqs. (1), (1), (3) and (4), and the velocity–pressure coupling is
done using the SIMPLEC algorithm. To ensure convergence, the criterion for scaled residuals is set at 1×10−5 ,
and the aerodynamic coefficients are monitored simultaneously.

3.3 Boundary conditions

Inlet surface applied with free stream velocity boundary condition, U∞ = 15 m/s. This velocity is chosen to
achieve a chord-based Re = 1 × 106 , similar to experimental data from the literature (Hand [Link]. [28], Somers
[29]). The free-stream turbulent quantities at the inlet are determined using Fig. 6.
3 0.5
k∞
k∞ = (| U∞ | I )2 , ω∞ = 0.5 (6)
2 Cμ L
where, the turbulent intensity, I = 0.05%, the constant Cμ = 0.09 and the characteristic length, L = c.
At the outlet boundary, a zero gradient is enforced for all parameters in all spatial directions. The front
and back boundaries are modeled as symmetry planes. A no-slip condition is imposed on the blade surface
through suitable wall functions.

3.4 Mesh generation

The computational domain shown in Fig. 9a uses a C − H configuration, and the domain boundaries extend
10c and 25c upstream and downstream, respectively, and 10c on either side of the blade vertically. Using
ICEMCFD software, the domain is discretized into a structured hexahedral cell configuration using a C − H
type meshing procedure. To accurately capture the curved surface of the airfoil, an O-grid is generated as
shown in Fig. 9b. Visual representations of the mesh surrounding the blade are shown in Fig. 10.
The fully resolved meshes near VGs are done through a body-fitted approach with more grid cells, aiming
to capture the intricate finer details of the flow surrounding VGs. Visualization of these meshes around VGs
is shown in Fig. 11a, b and c.
To maintain a wall y + < 1, the first cell above the blade surface is positioned at 10−5 c, and a sample of
calculated y + is shown in Fig. 12. Ensuring this fine spatial resolution is crucial to capture the boundary layer
accurately.
7772 H. Palanivel, R. Mukherjee

Fig. 10 Meshing details

Fig. 11 Body-fitted grids around 3D vortex generators of different geometries

Fig. 12 Wall y + at mid-span (z/(b/2) = 0) on suction side of the turbine blade

3.5 Grid convergence study

The uncertainty related to grid refinement is evaluated using the Richardson extrapolation method, a well-
established technique for grid convince study by calculating errors through the Grid Convergence Index (GC I )
introduced by Roache et al. [30]. The numerical errors in grid refinement are shown in Fig. 13.
Following the methodology outlined in [31], three distinct mesh sizes (N1 , N2 and N3 ) are used to compute
the C D of the blade (φ1 , φ2 and φ3 ) at α = 8◦ . The grid refinement factor, r ≥ 1.3, with r21 = L 2 /L 1 ,
r32 = L 3 /L 2 where
 

N
L j = (1/N j ) ( Vi )
i=1

and L 1 > L 2 > L 3 . The value of an infinitely fine mesh denoted as φ0 at L cell = 0, depicted in blue on
the y-axis of Fig. 13, is obtained by extrapolation. GC I shown in Fig. 7 serves as a standard measure for
Airfoil-Shaped Vortex Generators 7773

Fig. 13 Grid convergence study at α = 8◦ . C D corresponding to infinitely fine mesh, i.e., L cell /HVG = 0 is shown in Blue

quantifying discretization errors:

1.25 ji φi − φ j
GC I ji = p ;  ji = (7)
r ji − 1 φi

where p and  denote the order of convergence and relative error, respectively.
The uncertainty analysis of grid refinement reveals that for fine mesh solutions, the GC I uncertainty stands
at 1.34%, 4.4%, 0.97%, and 3.74% for different simulation cases: clean blade, blade with RVGs, TVGs, and
AVGs. The extrapolated errors for the corresponding setups are 1.09%, 3.65%, 0.79% and 3.08%. All the
errors remain below 5%, affirming the reliability of the simulation results. Consequently, we proceed with
the final grid size used are 0.96, 1.49, 1.045, and 1.74 million for clean blade, blades with RVGs, TVGs, and
AVGs, respectively, representing an approximation to an infinitely fine mesh.

3.6 Blade model with s809 airfoil

The numerical methodology used in the present investigation is validated with existing literature using a blade
with an s809 airfoil, which makes up nearly 75% of the turbine blade. CFD analysis of s809 airfoil is conducted
and 2D aerodynamic coefficients, Cl , Cd and Cm at Re = 1 × 106 are validated with experimental results from
Delft University of Technology by Somers [29], from Ohio State University by Hand et al. [28] and CFD data
of Zhong et. al. [32], Wang [Link]. [15], as presented in Fig. 14.
7774 H. Palanivel, R. Mukherjee

Fig. 14 Current CFD results of s809 airfoil at Re = 1 × 106 : comparison with literature

Fig. 15 C p distribution on s809 airfoil at Re = 2 × 106 : comparison with literature

Note: Cm − α output is not available from Wang et. al. [15] to be plotted in Fig. 14. Cl − α curves agree
almost identically within pre-stall regions up to α = 8◦ . Beyond this, all three CFD results are over-predicted
compared to the two wind tunnel results. This is the smallest deviation from experiments achieved compared
to other RANS turbulence models for the present CFD results. All the Cd − α plots, on the other hand,
show minimal deviation from each other except for the wind tunnel results from Delft Univ., which shows a
significant spike and overshoots around α = 12◦ . The wind tunnel results for Cm − α show distinct and sharp
changes in slope, unlike the two CFD outputs.
Airfoil-Shaped Vortex Generators 7775

Fig. 16 Current CFD results of Clark-Y airfoil: comparison with literature

The present CFD results of the C p distribution at specific angles of attack at Re = 2×106 exhibit favorable
agreement with experiment [29] and capture the trends accurately without any significant deviations as shown
in Fig. 15.

3.7 Vortex generator model with Clark-Y airfoil

The 3D vortex generator defines its shape using the Clark-Y airfoil as a section. Hence, this airfoil is also
analyzed separately and the present CFD output of its Cl and Cd is validated with experiments by Demirci et
al. [33] at Re = 1 × 105 and Re = 0.5 × 105 as shown in Fig. 16.
For the Cl − α plot in Fig. 16, the current CFD results at Re = 1 × 105 captures the location of the distinct
and sharp change in slope at α ≈ 14◦ exactly although the CFD results under predict the Clmax . At lower
Re = 0.5 × 105 , the present CFD results capture the various changes in slope shown by the wind tunnel results
well and Clmax predicted is not distinctly different. However, location of Clmax is over predicted by ≈ 2◦ . The
present CFD output for Cd − α also compares reasonably well with literature as shown in Fig. 16.

4 Oil flow visualization from wind tunnel

Given that the size of the AVGs is very small compared to the size of the blade, numerical analysis using
an in-house code did not result in major differences in the aerodynamic characteristics. Hence, it became
imperative that the physical model be tested in the wind tunnel to be sure that flow is going to be affected by
the VGs, their small size notwithstanding. In other words, it is important to be sure that pressures and loads
measured for the physical model in the wind tunnel are going to be affected by the presence of the AVGs.
Therefore, oil-flow visualization is carried out as shown in Fig. 17 for a clean blade and a blade with AVGs
at Re = 1.46 × 105 for α = 0◦ , 5◦ and 10◦ to see if the presence of AVGs is of any consequence. The impact
of tip vortices is evident for both the clean and AVG-equipped blades, as the flow curls onto the upper surface
near the blade tips. This behavior, highlighted with red dotted lines, can be attributed to the three-dimensional
nature of the blade.
At α = 0◦ , a local laminar separation bubble (LSB) forms near the mid-section/root at approximately
x/c ≈ 0.35, resulting from local flow separation. This is indicated by the white region and marked with blue
dotted lines in Fig. 17a. In this region, essentially the flow is not cutting through the oil layer to move forward.
Instead, it has separated from the surface for a brief period such that the original oil layer in this region is
undisturbed and appears rather evenly distributed.
In contrast, blades equipped with AVGs as shown in Fig. 17b do not exhibit such local flow separation.
Instead, small localized white regions are observed at x/c ≈ 0.1 − 0.15, just before the VGs, which are absent
in the clean blades. Flow divides and moves past the pair of VGs, resulting in two counter-rotating vortices
generated at the tips of the VGs in each pair downstream of the VGs. This results in downwash, i.e., a downward
7776 H. Palanivel, R. Mukherjee

Fig. 17 Surface oil flow visualization of blade of A R = 1.19 at Re = 1.46 × 105

velocity component between the VGs so that the local effective angle of attack is reduced, thus diminishing
the velocity around these regions and eliminating localized LSBs.
As the angle of attack increases to α = 5◦ , the LSB on the clean blade shifts toward the leading edge,
maintaining the same size as at α = 0◦ . However, such flow structures are absent on blades with AVGs. There
is no significant local flow separation near the AVGs, even at α = 5◦ .
When the angle of attack is further increased to α = 10◦ , the LSB near the leading edge of the clean blade
reduces in size, but another distinct LSB forms near the trailing edge, indicating impending flow separation
with further increase in angle of attack. However, these structures and the patterns observed at lower angles
are not present on the blade with AVGs. Once again, no significant local flow separation is observed near the
AVGs at α = 10◦ .
Thus, it is safe to say that despite the very small size of the AVGs, they significantly influence the flow,
preventing local flow separation and the formation of LSBs by supplying additional energy to the boundary
layer, allowing it to remain attached.
Aerodynamic loads and pressures are now calculated and effect of the AVGs on the surface boundary layer
is studied using high-fidelity CFD using OpenFOAM as described in the following sections.

5 CFD analysis

In this section, the behavior of the blade with and without the VGs is investigated by studying its aerodynamic
coefficients, aerodynamic efficiency, streamline patterns, pressure distribution, TKE, vortex formation behind
VGs, and boundary layer velocity profiles. All the quantities presented in this section are turbulent mean
quantities. The stream-wise (x), normal (y) and span-wise (z) distances are normalized with chord length of
blade (c), the height of VGs (HV G ) and half of the span length of the blade (b/2), respectively.
Airfoil-Shaped Vortex Generators 7777

Fig. 18 Aerodynamic force coefficients of turbine blade with and without VGs

Table 1 CL and αstall for different blades

cases Clean blade Blade with RVGs Blade with TVGs Blade with AVGs
αstall 11◦ 12◦ 12◦ 13◦
CL at αstall 1.0563 1.3573 1.345 1.0902

Fig. 19 %age change in CD of blade with and w/o VGs

5.1 Aerodynamic characteristics

CFD analysis is conducted on a clean blade and on a blade with VGs, where one set of VGs are placed at 20%
of the blade’s chord length, i.e., x/c = 0.2 as shown earlier in Figs. 5 and 8. Three different types of VGs are
used, as shown in Fig. 7. The aerodynamic characteristics of the blades, CL − α, CD − α, CL /CD − α of all
types of blades are shown in Fig. 18.
The blade with RVGs and TVGs has higher CL − α slopes for 0 < α < 12◦ and distinctly higher CLmax
as shown in Fig. 18a. The blade with the AVGs performs similarly as the clean blade. The αstall and CLstall of
all blades are given in Fig. 1.
While the other three blades show distinct stall conditions, the blade with AVGs shows very mild stall even
at α = 20◦ . This essentially means a drastic increase in the range of operation for such a blade. For example,
a clean blade can operate at C L max ≈ 1.05 for 0 < α < 11◦ , the blades with RVGs and TVGs can operate
at C L max ≈ 1.4 for 0 < α < 12◦ while the blade with AVGs can operate at C L max ≈ 1.1 for 0 < α < 20◦ .
Consequently, the blade with AVGs also sees the least C D of all blades considered here, even for α > 10◦ ,
where there is no change in the C D − α slope as shown in Fig. 18b.
Figure 19 shows the percentage change in C D of all blades with and without VGs. Blades with RVGs and
TVGs experience more drag than clean blades.
7778 H. Palanivel, R. Mukherjee

Fig. 20 CL /CD − α

Blade with RVGs sees sharp spikes in CD for α > 8◦ after which it is continually increasing with a
maximum of 45.27% at α = 9◦ . For the blade with TVGs, several spikes in CD are observed for α > 11◦ ,
which also continually increase with a maximum of 66.65% at α = 18◦ . The blade with AVGs, on the other
hand, sees a negative percentage change in C D for the entire α-range considered here with a minimum of
−27.68% at α = 11◦ , which incidentally is also the αstall of the clean blade.
The aerodynamic efficiency, a ratio of CL and CD is shown in Fig. 20 for blades with VGs and clean blades.
The blade with AVGs performs as good as the other blades for α < 8◦ . On the other hand, α > 8◦ is more
efficient than all the different blades. It also has a smoother transition into the post-stall region, and there are
no distinct fluctuations in the efficiency for α > 8◦ . For both the blades with RVGs and TVGs and the clean
blade, there are specific changes in the efficiency slope for α > 8◦ , while the blade with AVGs has a smooth
slope.

5.2 Streamline patterns on the blades

Numerical flow visualization using CFD plots streamlines for all the blades showing on-dimensional velocity,
u/U∞ as shown in Fig. 21.
At α = 6◦ , 8◦ and 10◦ , the flow stays attached to the wall for all the blades. At α = 12◦ , a small dark zone
develops on the upper side of the clean blade near TE, indicating the beginning of flow separation. This is
absent for the blades with VGs. If the angle of attack is increased even more, then at α = 14◦ , a distinct flow
separation zone is found on the clean blade and a hint of it on the blade with TVGs. Again, no such zones are
seen on the blades with RVGs and AVGs. When the angle is increased to α = 16◦ , the flow separation zone
on clean blade and the blades with the RVGs and TVGs become more prominent and now engulf more of the
blade’s surface, moving toward the leading edge. The blade with AVGs also sees a small separation zone near
the trailing edge.
At α = 18◦ , the blade with AVGs continues to see a small flow separation zone, the smallest among all the
blades, which does not change its size with an increase in the angle. On the other hand, the separation zone
becomes more extensive for other blades, and the blade with TVGs also sees a large re-circulation zone that
engulfs ≈ 75% of the blade surface. Hence, the blade with RVGs and TVGs offers no aerodynamic advantage
compared to the clean blade. At α = 20◦ , the clean blade near the leading edge sees a large trailing edge
separation zone while the blade with the RVGs sees a secondary separation zone. The blade with TVGs now
sees counter-rotating, asymmetric re-circulation zones near the leading edge along with pockets of separation
zones. The blade with AVGs sees a slightly larger trailing edge separation zone. Overall, delayed flow separation
occurs on the blade with AVGs at ≈ α = 16◦ , and even at α = 20◦ , it continues to see a trailing edge separation
zone, the size of which has changed only minutely.
Airfoil-Shaped Vortex Generators 7779

Fig. 21 Streamline patterns on turbine blade

7780 H. Palanivel, R. Mukherjee

Fig. 22 Velocity profiles at locations downstream of VGs; α = 16◦

5.3 Effect of AVGs on boundary layer

The velocity profiles are studied to further study the effects of VGs, u/U∞ with respect to wall normal
coordinate y/c at several locations downstream of the VGs on the blade at α = 16◦ are plotted, as shown in
Fig. 22.
It is observed that the boundary layer thickens on the blade surface downstream of the location of the VGs,
as indicated by the changes in slope of ∂u/∂ y. For example, the clean blade experiences a change in velocity
from x/c = 0.4 to x/c = 0.9.
For a clean blade, it is expected that the momentum of the boundary layer gradually diminishes in the
streamwise direction along the blade chord, leading to a tendency of flow separation from the blade’s surface.
This shows up at x/c = 0.6 in Fig. 22, where the slope of ∂u/∂ y becomes negative, and there is a small region
of flow reversal. The inflection point then ascends in the normal direction as the flow develops downstream in
a streamwise direction.
The formation of longitudinal tip vortices is induced at the tips of VGs by introducing VGs on the blade,
which bring free stream air of high momentum into proximity with the boundary layer, effectively enhancing
surface momentum. As a result, the inflection points of all the blades with VGs are located at distances near
the wall further downstream, indicating that flow separation is delayed.
For a blade with AVGs, the effect of changing the blade’s angles of attack is studied on velocity profiles and
corresponding separation behavior at a location, as shown in Fig. 23 at x/c = 0.3, 0.6 and 0.9. It is observed
that only for α = 20◦ at x/c = 0.6, there is a small separation zone and corresponding slope, ∂u/∂ y changes
Airfoil-Shaped Vortex Generators 7781

Fig. 23 Change in velocity profiles for a blade with AVGs at a specific location with the angle of attack

Fig. 24 Distribution of Vorticity peaks along the blade downstream of the VGs. max is the magnitude of peak vorticity

mildly. Further downstream at x/c = 0.9, the inflection point increases in the normal direction nearly at the
trailing edge. Still, there is no significant change in the separation zone, so the slope of the curve also changes
mildly. At this location, a at α = 16◦ also begins to see a small separation zone.

5.4 Vorticity distribution and peak vorticity

As mentioned previously, the presence of VGs enhances surface momentum, the probable reason for it being
enhanced vorticity generated by the tip vortices of the VGs as also shown by the distribution of vorticity-peaks
in Fig. 24.
It is also observed that the blade with AVGs generates maximum vorticity, max , which imparts additional
momentum to the boundary layer for all angles of attack downstream of the VGs. The peak vorticity for the VGs
decays as we move downstream of the VGs in a streamwise direction. The peak occurs at about x/c ≈ 0.35 for
AVGs across all angles of attack. For instance, at α = 10◦ , the blade with AVGs shows an increase of ≈ 200%
and 76% in max at x/c = 0.35 compared to blades with RVGs and TVGs, respectively. Additionally, TVGs
generate 70% more peak vorticity than RVGs. Thus, it can be inferred that although TVGs perform better
7782 H. Palanivel, R. Mukherjee

Fig. 25 Vorticity contours on a turbine blade with AVGs at α = 10◦ : a Location of vorticity contours on blade strip, b spanwise
asymmetry in the distribution of Vorticity peaks, c close-Up of (a)

than RVGs among traditional VG shapes, AVGs exhibit superior performance in terms of inducing additional
momentum into the flow field.
With increments in angles of attack, peak vorticity reduces. It is likely that at high angles, the flow separates,
reducing momentum near the surface. A noteworthy exception to this theory occurs for the blade with TVGs
at α = 20◦ , where a sudden surge in max occurs around x/c ≈ 0.6 as shown in Fig. 24. This surge is
presumably due to the additional vorticity generated by the extensive re-circulation zone on the blade’s suction
side consisting of two counter-rotating vortices as shown in Fig. 21. The flow is curling onto the suction side
from the pressure side at very high speeds near the trailing edge, causing additional vorticity, which is expected
to increase the size of the re-circulating vortex.
A graphical representation of the effect of the VGs on enhanced momentum injected into the flow field is
shown in Fig. 25.
The VGs generate longitudinal streamwise vortices within counter-rotating arrays, injecting additional
momentum into the flow field. To study this behavior, the vorticity contours at specific downstream locations
along the chord of the 3D blade behind AVGs are shown in Fig. 25a. A close-up of the same is shown in
Fig. 25c.
Originating from the tips of the VGs, these vortices progress downstream along the blade’s chord till they
gradually disengage from the surface along the blade’s span and diffuse normally to the wall. Moreover, the
Airfoil-Shaped Vortex Generators 7783

Fig. 26 Spanwise location of vorticity peaks downstream of VGs

vortices increase in size as they travel downstream. These induced vortices play a crucial part in rejuvenating
the boundary layer, providing more resistance to separation.
It is important to report here that the distribution of the vorticity peaks exhibits asymmetry in the spanwise
direction as shown in Fig. 25b. The inset indicates the locations along the blade span at which the peak
vorticity is measured. The locations are symmetrically placed about the root of the blade. The bar chart shows
the magnitude of the peak vorticity at the locations shown in the inset for different x/c locations on a slice in
the y − z plane.
For example, the peak vorticity values downstream at x/c = 0.4 and on the slice in the y − z plane at
y/c = 0.102 are slightly different, i.e.,
 HV G z
= 51.54 at blade-span = +0.42 and
U∞ (b/2)
 HV G z
= 50.45 at blade-span = −0.42
U∞ (b/2)
Hence, the max plots shown in Fig. 24 represent only the highest of these two values at two different spanwise
locations at a specific downstream location on a slice in an y − z plane.
The locations of these peak vorticity values at different spanwise positions are shown in Fig. 26. For all
blades, there is an asymmetric distribution of the peak vorticity, albeit mostly distributed around the location
of the VGs. There is flow separation with increasing angles of attack, and hence such distribution of peak
vorticity reduces, and the vorticity due to the tip vortices dominates for α ≥ 16◦ as shown in Fig. 26. The only
disadvantage of such an asymmetric vorticity distribution is additional spanwise rolling moments. However,
the asymmetry is not prominent, and major operational hazards are not envisioned. In other words, such
asymmetrical vorticity distribution can be taken care of by airplane flight control systems or adjustments in
turbine blade alignment.
7784 H. Palanivel, R. Mukherjee

Fig. 27 Non-dimensional turbulent kinetic energy over a blade with and without AVGs

5.5 Effect of AVGs on TKE and pressure distribution

The non-dimensional turbulent kinetic energy (TKE), k/U∞ 2 , close to the blade’s surface with and without
AVGs is presented in Fig. 27.
It is noteworthy that the TKE peaks are significantly higher for the blade equipped with AVGs compared
to the clean blade.
The TKE exhibits a characteristic pattern where it starts with zero at the blade surface, i.e., y/HV G = 0,
reaches a maximum value as we keep moving in the normal direction at a certain y/HV G location and then
gradually decreases to zero as we keep moving farther away from the surface in the normal direction.
This decay of TKE in the normal direction is over a larger y/HV G distance away from the VGs as we move
downstream along the chord, suggesting that flow separation is more probable away from the leading edge.
For example, for the blade with AVGs at α = 20◦ , maximum TKE, k/U∞ 2 = 0.036 at x/c = 0.9 and

y/HVG = 15.8 while k/U∞ = 0.032 at x/c = 0.6 and y/HVG = 5.5. This suggests that the flow near
2
x/c = 0.9 sees higher TKE with a sharper slope compared to the flow near x/c = 0.6 for a larger distance
away from the surface thus increasing the chances of separation.
Both the maximum TKE and its slope with y/HVG increase with the angle of attack. For the clean blade,
TKE peaks consistently increases with both x/c and α. However, this trend is not consistent for the blade with
AVGs. Specifically, at angles of attack of α = 8◦ and 12◦ , TKE peaks decreases with an increase in x/c, while
it shows an increase at α = 16◦ and 20◦ .
At higher angles of attack, TKE is greater near the trailing edge, indicating that a thicker boundary layer
is required for its dissipation. Additionally, the higher TKE may be linked to the flow reversal in the boundary
layer near the trailing edge at α = 20◦ shown in Fig. 23. The vortices shed due to separated flow increases
TKE. This increase in TKE can also be attributed to the additional momentum imparted to the boundary layer
by the AVGs.
The presence of VGs also leads to the formation of distinct localized suction-side pressure peaks, as shown
in the C p distribution in Fig. 28, where the location of the VGs is also highlighted. Blades with RVGs and TVGs
Airfoil-Shaped Vortex Generators 7785

Fig. 28 Pressure distribution on turbine blade at mid-span of the blade

also lead to an overall decline in C p , whereas a blade with AVG experiences an increase in C p . Particularly at
a higher angle, α = 16◦ , identical C p values are consistently observed.
At α = 20◦ , on the other hand, the C p curves for the blades with RVG and TVG deviate significantly from
the trend observed for the clean blade. The blade with AVG, however, continues a similar trend with a local
C p spike near the VG. The reason for this is due to the separation zones formed throughout the suction side of
the blades with RVGs and TVGs at α = 20◦ as shown in Fig. 21, which result in the pressure plateaus shown
in Fig. 28. Hence, conventional VGs, unlike the AVGs, offer no specific advantage at higher angles of attack.

6 Conclusion

Vortex generators (VGs) effectively postpone flow separation, enhancing the wind turbine blade’s aerodynamic
performance. A strip of a 3D rectangular blade with a pair of three different shapes of VGs, whose height is
about 1%c, is studied. Since the VGs are considerably smaller than the blade, the strip of the blade is considered
without accounting for tip vortices by using continuous boundary conditions at both edges. For the VGs, 3D
geometries are considered, and their 3D aerodynamic behavior is studied in detail without the effect of blade
tip vortices. Key observations from the study are summarized here.
7786 H. Palanivel, R. Mukherjee

In comparison with a clean blade without VGs that stalls at αstall ≈ 11◦ , blades with RVGs and TVGs
operate at 27% higher CLmax and minimally delayed stall at αstall ≈ 12◦ . On the other hand, a blade with
AVGs operates around 5% higher CLmax compared to the clean blade but sees minimal stall even at α ≈ 20◦ .
Therefore, for a blade with AVGs, there is a significant increase in the range of operation. It also sees a negative
percentage change in C D throughout the operating α-range with a minimum value of −27.68% at α = 11◦ .
In terms of efficiency measured with CL /CD , a blade with AVGs performs the best among all the blades
for α > 8◦ with a maximum value of over 40% around 7 − 8◦ and a minimum of ≈ 10% at α = 20◦ . Unlike
other blades, a blade with AVGs transitions smoothly into the post-stall region without significant fluctuations
in its C L or C D .
Delayed separation on a blade with AVGs is due to a mild change in its velocity profile near the surface
along with a small flow-separated region near the blade’s trailing edge. The locations of inflection points are
notably at much smaller y/c vertical distances on y − z planes at several downstream streamwise locations. A
blade with AVGs also sheds additional vorticity into the flow downstream of the AVGs that inject momentum
into the flow field, which in turn helps delay separation. The vorticity shed by the blade with AVGs is maximum
among all the blades considered here. However, there is an asymmetry in the strength of the vorticity shed
along the blade span, which decreases as we move downstream and is almost nil near the trailing edge. This
asymmetry is not significant and is not expected to cause concern.
Therefore, the authors feel it safe to suggest that employing airfoil-shaped VGs on a wind turbine blade’s
suction side will inject additional momentum into the boundary layer, thereby delaying flow separation without
any drastic drag penalty and improving overall aerodynamic performance. Also, the 3D effects of the AVGs
studied here without the impact of blade tip vortices are significant, although they are considerably small in
size. This gives the authors confidence to suggest using a larger size of the AVGs for easier machining and
practical purposes.

Acknowledgements We acknowledge the Ministry of Education, Government of India, for funding the research of the first
author on this paper, doctoral candidate at the AMBE Department, IIT Madras, Mr. Hariprasanth Palanivel through the Prime
Minister’s Research Fellowship (PMRF) Scheme (Project No: SB22230936AMPMRF008306). The authors also acknowledge
the use of the computing resources at HPCE, IIT Madras, Chennai.

Funding No funding was received for conducting this study.

Declarations

Conflict of interest The authors have no relevant financial or non-financial interests to disclose.

References

1. Ram, M., Bogdanov, D., Aghahosseini, A., Gulagi, A., Oyewo, A.S., Odai Mensah, T.N., Child, M., Caldera, U., Sadovskaia,
K., Barbosa, L.D.S.N.S., Fasihi, M., Khalili, S., Traber, T., Breyer, C.: Global energy transition to 100% renewables by 2050:
not fiction, but much needed impetus for developing economies to leapfrog into a sustainable future. Energy 246, 123419
(2022). [Link]
2. Dykes, K.L., Veers, P.S., Lantz, E.J., Holttinen, H., Carlson, O., Tuohy, A., Sempreviva, A.M., Clifton, A., Rodrigo, J.S.,
Berry, D.S., et al.: Iea wind tcp: Results of iea wind tcp workshop on a grand vision for wind energy technology. Technical
report, National Renewable Energy Lab., Golden, CO.(US) (2019). [Link]
3. Zhao, Z., Jiang, R., Feng, J., Liu, H., Wang, T., Shen, W., Chen, M., Wang, D., Liu, Y.: Researches on vortex generators
applied to wind turbines: a review. Ocean Eng. 253, 111266 (2022). [Link]
4. Xue, S., Johnson, B., Chao, D., Sareen, A., Westergaard, C.: Advanced aerodynamic modeling of vortex generators for wind
turbine applications. In: European Wind Energy Conference (EWEC), Warsaw Poland (2010). [Link]
net/publication/289260092
5. Lin, J.C.: Review of research on low-profile vortex generators to control boundary layer separation. Progress Aerosp. Sci.
38(4–5), 389–420 (2002). [Link]
6. Sullivan, T.L.: Effect of vortex generators on the power conversion performance and structural dynamic loads of the mod-2
wind turbine. Technical report, NASA Lewis Research Center, Cleveland, OH (United States) (1984). [Link]
citations/19840021278
7. Corrigan, R., Savino, J.: Vortex generators as a means for increasing rotor performance. Technical report, NASA Lewis
Research Center, Cleveland, OH (United States) (1985). [Link]
8. Miller, G.: Comparative performance tests on the mod-2, 2.5-mw wind turbine with and without vortex generators. In:
DASCON Engineering, Collected Papers on Wind Turbine Technology (1995). [Link]
Airfoil-Shaped Vortex Generators 7787

9. Moon, H., Jeong, J., Park, S., Ha, K., Jeong, J.-H.: Numerical and experimental validation of vortex generator effect on
power performance improvement in mw-class wind turbine blade. Renew. Energy 212, 443–454 (2023). [Link]
1016/[Link].2023.04.104
10. Mueller-Vahl, H., Pechlivanoglou, G., Nayeri, C., Paschereit, C.: Vortex generators for wind turbine blades: A combined
wind tunnel and wind turbine parametric study. In: Turbo Expo: Power for Land, Sea, and Air, vol. 44724, pp. 899–914
(2012). American Society of Mechanical Engineers. [Link]
11. Baldacchino, D., Ferreira, C., Tavernier, D.D., Timmer, W., Van Bussel, G.: Experimental parameter study for passive vortex
generators on a 30% thick airfoil. Wind Energy 21(9), 745–765 (2018). [Link]
12. Choudhry, A., Arjomandi, M., Kelso, R.: Methods to control dynamic stall for wind turbine applications. Renew. Energy 86,
26–37 (2016). [Link]
13. De Tavernier, D., Ferreira, C., Viré, A., LeBlanc, B., Bernardy, S.: Controlling dynamic stall using vortex generators on a
wind turbine airfoil. Renew. Energy 172, 1194–1211 (2021). [Link]
14. Jiang, R., Zhao, Z., Liu, H., Wang, T., Chen, M., Feng, J., Wang, D.: Numerical study on the influence of vortex generators
on wind turbine aerodynamic performance considering rotational effect. Renew. Energy 186, 730–741 (2022). [Link]
org/10.1016/[Link].2022.01.026
15. Wang, H., Zhang, B., Qiu, Q., Xu, X.: Flow control on the NREL s809 wind turbine airfoil using vortex generators. Energy
118, 1210–1221 (2017). [Link]
16. Gao, L., Zhang, H., Liu, Y., Han, S.: Effects of vortex generators on a blunt trailing-edge airfoil for wind turbines. Renew.
Energy 76, 303–311 (2015). [Link]
17. Zhu, C., Wang, T., Wu, J.: Numerical investigation of passive vortex generators on a wind turbine airfoil undergoing pitch
oscillations. Energies 12(4), 654 (2019). [Link]
18. Zhu, C., Chen, J., Wu, J., Wang, T.: Dynamic stall control of the wind turbine airfoil via single-row and double-row passive
vortex generators. Energy 189, 116272 (2019). [Link]
19. Zhu, C., Wang, T., Chen, J., Zhong, W.: Effect of single-row and double-row passive vortex generators on the deep dynamic
stall of a wind turbine airfoil. Energies 13(10), 2535 (2020). [Link]
20. Manolesos, M., Chng, L., Kaufmann, N., Ouro, P., Ntouras, D., Papadakis, G.: Using vortex generators for flow separation
control on tidal turbine profiles and blades. Renew. Energy 205, 1025–1039 (2023). [Link]
02.009
21. Wang, P., Wang, L., Huang, B., Wu, R., Wang, Y.: The effects of vortex generators on the characteristics of the tip hydrofoil
and the horizontal axis tidal turbine blade. Renew. Energy (2024). [Link]
22. Méndez, B., Gutiérrez, R.: Non-conventional vortex generators calculated with cfd. In: Journal of Physics: Conference Series,
vol. 1037, p. 022029 (2018). IOP Publishing. [Link]
23. Hansen, M.O.L., Velte, C.M., Øye, S., Hansen, R., Sørensen, N.N., Madsen, J., Mikkelsen, R.: Aerodynamically shaped
vortex generators. Wind Energy 19(3), 563–567 (2016). [Link]
24. Soto, R., Bartholomay, S., Manolesos, M., Nayeri, C., Paschereit, C.O.: Airfoil shaped vortex generators applied on a research
wind turbine. In: AIAA Scitech 2021 Forum, p. 1413 (2021). [Link]
25. Tangler, J.L., Somers, D.M.: Nrel airfoil families for hawts. Technical report, National Renewable Energy Lab., Golden,
CO.(US) (1995). [Link]
26. Menter, F.R.: Improved two-equation k-omega turbulence models for aerodynamic flows. Technical report, National Aero-
nautics and Space Administration (1992). [Link]
27. Menter, F.R.: Two-equation eddy-viscosity turbulence models for engineering applications. AIAA J. 32(8), 1598–1605
(1994). [Link]
28. Hand, M.M., Simms, D., Fingersh, L., Jager, D., Cotrell, J., Schreck, S., Larwood, S.: Unsteady aerodynamics experiment
phase vi: Wind tunnel test configurations and available data campaigns. Technical report, National Renewable Energy Lab.,
Golden, CO.(US) (2001). [Link]
29. Somers, D.M.: Design and experimental results for the s809 airfoil. Technical report, National Renewable Energy Lab.,
Golden, CO.(US) (1997). [Link]
30. Roache, P.J.: Perspective: a method for uniform reporting of grid refinement studies. J. Fluids Eng. 116(3), 405–413 (1994).
[Link]
31. Celik, I.B., Ghia, U., Roache, P.J., Freitas, C.J.: Procedure for estimation and reporting of uncertainty due to discretization
in cfd applications. J. Fluids Eng. (2008). [Link]
32. Zhong, J., Li, J., Guo, P.: Effects of leading-edge rod on dynamic stall performance of a wind turbine airfoil. Proc. Inst.
Mech. Eng. Part A J. Power Energy 231(8), 753–769 (2017). [Link]
33. Demirci, V., Seyhan, M., Sarioglu, M.: Investigation of aerodynamic performance of clark-y airfoil with more realistic
tubercle model and internal slots. Phys. Fluids (2023). [Link]

Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional
affiliations.

Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement
with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely
governed by the terms of such publishing agreement and applicable law.