NUMERICAL SIMULATIONS OF FLOW PAST A

WING WITH DESIGNED LEADING EDGE

SERRATIONS

An End Semester Project Report

Submitted in partial fulfilment of the
requirements for the Degree of
Master of Technology in Engineering Analysis and Design

by

Mayank Rawat

Under the guidance of

Prof. Arjun Sharma

Department of Applied Mechanics

Indian Institute of Technology, Delhi
Hauz Khas
New Delhi– 110 016

JUNE 2025
Acknowledgements

I thank Prof. Arjun Sharma for help with simulations and writing the report. I am also
very thankful to my committee members, Prof. Hari Vemuri, Prof. Sawan Sinha, Prof.
Anupam Dewan and Prof. Bahni Ray for their encouragement and valuable suggestions.
I also thank my family members for their support during my studies.

i
Abstract

A study on the improvement of the NACA0012 wing performance in the low Reynolds
number (Re) region at 20,000 and 43200 using a leading-edge protuberance (LEP) was
conducted. In this study, we focused on the effects of varying Re and angle of attack on
the effectivity of the LEP. Numerical simulations were performed to measure the wing
performance and to visualize the flow structures around the airfoil. At lower Re (20000),
the LEP appears to be effective at both 4° and 14° angles of attack. However, at higher
Re (43200), the LEP seems to be significantly more effective at 14° than at 4° angle of
attack. Separation control is caused by the streamwise vortices generated by LEP which
provides the momentum from the freestream flow into the separation region near the
valley section, and the mechanism is the same at all Reynolds numbers and angles of
attack in this study. The numerical analysis also showed that the protuberances located
on the leading edge of the airfoils act as vortex generators. The strength of the vortices
increase with angle of attack and flow is energized and pulled over the leading-edge
peaks, which explains the stall delay characteristics of the modified airfoil. Infinite-
span airfoil simulations provide basic insights but fail to capture the full complexity of
aerodynamic behavior of a realistic wing. Finite wing analysis is essential, as it accounts
for three-dimensional effects such as spanwise flow and wingtip vortices. In this study,
we have also performed a comparative analysis between a finite-span and an infinite-span
modified airfoil with leading-edge protuberances (LEP). Finally, we modified the airfoil
by introducing protuberances at both the leading and trailing edges. A comparative
study was then conducted between the airfoil with protuberances only at the leading
edge and the airfoil with protuberances at both the leading and trailing edges.

ii
iii
Contents

Acknowledgements i

Abstract ii

1 Introduction 1

2 Literature Review 6

3 Methodology 8
3.1 Governing equations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
3.2 Mesh for infinite-span airfoil . . . . . . . . . . . . . . . . . . . . . . . . . 10
3.3 Mesh for finite-span airfoil . . . . . . . . . . . . . . . . . . . . . . . . . . 11
3.4 Lifting-line theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
3.4.1 Solutions using lifting-line theory . . . . . . . . . . . . . . . . . . 13

4 Results and Discussion 16

4.1 Infinite-span airfoil with leading edge protuberances . . . . . . . . . . . . 16
4.2 Finite-span airfoil with leading edge protuberances . . . . . . . . . . . . 18
4.3 Infinite-span airfoil with leading and trailing edge protuberances . . . . . 21

5 Conclusions and Future Work 26

iv
List of Figures

1.1 A schematic of the flow configuration is shown. U0 is the flow speed in

the free-stream and α is the angle of attack. . . . . . . . . . . . . . . . . 3
1.2 The baseline geometry of the NACA0012 airfoil is shown without leading
edge protuberances. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.3 The modified geometry of the NACA0012 airfoil is shown (three-dimensional
and side views) with sinusoidal leading edge protuberances based on the
work by Yasuda et al. (2019). The size of computational domain in the
spanwise direction is 0.8L. . . . . . . . . . . . . . . . . . . . . . . . . . 4
1.4 The modified geometry of the NACA0012 airfoil is shown (three-dimensional
and side views) with sinusoidal leading and trailing edge protuberances.
The size of computational domain in the spanwise direction is 0.8L. . . 5
1.5 The rounded tip in case of finite-span wing is shown in the upper figure. In
the planform view (lower figure), size of the wing in the spanwise direction
is 4L. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5

3.1 In the upper figure, side view of the computational domain is shown for
the case with airfoil of infinite span. A zoomed three-dimensional view of
the mesh around the airfoil (boxed region in the upper figure) is shown in
the lower figure. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11
3.2 A three-dimensional view of the mesh around the finite-span airfoil is shown. 12

4.1 Velocity and pressure contours are shown for the airfoil with leading-edge
protuberances at Re = 20000 and α = 4o at mid span (22 mm) . . . . . . 17

v
LIST OF FIGURES vi

4.2 Velocity and pressure contours are shown for the airfoil with leading-edge
protuberances at Re = 20000 and α = 14o at mid span (22 mm) . . . . . 18
4.3 Velocity and pressure contours are shown for the airfoil with leading-edge
protuberances at Re = 43200 and α = 4o at mid span (22 mm) . . . . . . 18
4.4 Velocity and pressure contours are shown for the airfoil with leading-edge
protuberances at Re = 43200 and α = 14o at mid span (22 mm) . . . . . 19
4.5 Vorticity contour for airfoil at 5mm, 25mm, 55mm and 100mm from the
leading edge at Re=43200 and α = 14o . . . . . . . . . . . . . . . . . . . 21
4.6 Vorticity contour for airfoil at 5mm, 25mm, 55mm and 100mm from the
leading edge at Re=43200 and α = 4o . . . . . . . . . . . . . . . . . . . . 22
4.7 Pressure variation on the suction surface of airfoil at Re = 43200, α = 4o
and α = 14o . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 23
4.8 Recirculation zone (isosurface ux=0) for a finite span airfoil having leading
edge protuberances, Re = 43200 and α = 14o . . . . . . . . . . . . . . . . 23
4.9 Velocity and Pressure contour for airfoil with leading-edge and trailing-
edge protuberances at Re = 20000, α = 4o and α = 14o . . . . . . . . . . 24
4.10 Velocity and Pressure contour for airfoil with leading-edge and trailing-
edge protuberances at Re = 43200, α = 4o and α = 14o . . . . . . . . . . 25
List of Tables

3.1 Flow parameters for simulation cases A to D . . . . . . . . . . . . . . . . 8

4.1 Lift and drag coefficients obtained for simulation cases A to D and their
comparison with experimental data. . . . . . . . . . . . . . . . . . . . . . 17
4.2 Lift and drag coefficients for a finite span airfoil with leading edge protu-
berances at varying angles of attack and Reynolds numbers. . . . . . . . 20
4.3 Lift and drag coefficients for an infinite span airfoil with leading and trail-
ing edge protuberances at varying angles of attack and Reynolds numbers. 24

vii
Chapter 1

Introduction

Boundary layer separation is a complex flow phenomenon that has significant implica-
tions towards forces and moments acting on a wing. In general for the dynamics of a
rigid two-dimensional wing, there are two forces and one moment component that need
to be considered. For a three-dimensional wing, all three force and moment components
have to considered. Small aerial vehicles typically operate in the low Reynolds number
range of 1.5×104 to 5×105 . Unmanned aerial vehicles (UAVs) typically have wing spans
. 6m and mass . 25kg. Micro aerial vehicles (MAVs) have now become possible with
dimensions as small as 15cm and mass 80g. MAVs are categorized as having span . 1m
and mass . 1kg (Mueller and DeLaurier, 2003).
The problem of boundary layer separation is very severe at low Reynolds numbers.
Laminar boundary layers develop over a substantial portion of the wing surface which
are amenable to separation in the presence of adverse pressure gradients. The value of
maximum lift coefficient and stall angle also undergo reduction at low values of Reynolds
numbers. The ratio of lift-to-drag coefficients, CL /CD , which is a key indicator of aerody-
namic efficiency for a wing, also reduces significantly. In general for a three-dimensional
finite-span wing, there is a component of drag in addition to the sectional drag (due to
pressure and shear forces) which is called the lift-induced drag. The net drag coefficient
for a finite-span wing at subsonic speed is written as (Anderson, 2010),

1
Chapter 1. Introduction 2

CL2
CD = CD,0 + (1.1)
πeR

The lift-induced drag component (second term in equation 1.1) is proportional to

the square of lift coefficient and inversely proportional to the wing-aspect-ratio R and
the span efficiency factor e. Here the wing-aspect-ratio is defined as b2 /S where S is
the wing planform area (planform or top view) and b is the wing span. The induced
drag arises due to the formation of low pressure region within the wing-tip vortices. It
is important for low-aspect-ratio wings such as those used in UAVs and MAVs at high
values of incidence. The wing-tip vortices cause downwash and reduce the effective angle
of attack which tends to reduce the lift force. Tip vortices also contribute to generation
of lift in low-aspect-ratio wings due to the formation of low pressure region on the upper
side.
Several forms of control methods have been utilized to delay or prevent boundary
layer separation (el Hak and Bushnell, 1991). Such methods are usually classified as ac-
tive and passive control methods. Steady or pulsed wall blowing and suction are among
the active methods to control separation. Passive turbulators or vortex generators intro-
duce streamwise vorticity into an otherwise laminar boundary layer to promote transition
to turbulence. Increased momentum mixing in a turbulent boundary layer helps delay
separation and increase the stall angle. The objective of the present study is to assess the
effects of introducing passive geometric modifications at the leading and/or trailing edge
of a NACA0012 airfoil in the form of spanwise sinusoidal protuberances. Both infinite
and finite wing studies are conducted to study the effects of edge scalloping on aerody-
namic characteristics and especially on delaying stall. Three-dimensional boundary layer
separation, although a more complicated phenomenon than two-dimensional separation,
has been utilized to develop efficient methods to delay stall and increase the maximum
lift coefficient. Many aquatic animals and birds benefit from three-dimensional serrated
edges of their fins/wings with improved locomotion efficiency and reduced noise emission.
Chapter 1. Introduction 3

Figure 1.1: A schematic of the flow configuration is shown. U0 is the flow speed in the
free-stream and α is the angle of attack.

A schematic of the flow configuration considered in the present study is shown in

figure 1.1. The incoming flow in the free-stream is at speed U0 at a fixed incidence of α.
The geometries of the baseline airfoil and that with a scalloped leading edge are shown
in figures 1.2 and 1.3 respectively. The modified geometry is analytically prescribed
similar to that given in the study by Yasuda et al. (2019) with slight differences. The
trailing edge is held fixed while the chord length is sinusoidally varied by introducing
waviness at the leading edge. The maximum chord length is L = 55mm while the peak-
to-peak amplitude and wavelength of variations in chord length and thickness along the
spanwise direction are 0.05L = 2.75mm and 0.2L = 11mm. At each section along the
span, a NACA0012 airfoil geometry is maintained by calculating an appropriate scale
factor. In the case of infinite span wing, the span size is taken as 0.8L and periodic
boundary conditions are imposed at the end planes. An additional design is considered
in which serrations are introduced at both leading and trailing edges as shown in figure
1.4. In case of finite-span wing, the tip of the wing is rounded as shown in figure 1.5 and
the spanwise domain size is extended to approximately two times the span of the wing;
periodic boundary conditions are switched off at the two spanwise end-planes.
Chapter 1. Introduction 4

Figure 1.2: The baseline geometry of the NACA0012 airfoil is shown without leading
edge protuberances.

Figure 1.3: The modified geometry of the NACA0012 airfoil is shown (three-dimensional
and side views) with sinusoidal leading edge protuberances based on the work by Yasuda
et al. (2019). The size of computational domain in the spanwise direction is 0.8L.
Chapter 1. Introduction 5

Figure 1.4: The modified geometry of the NACA0012 airfoil is shown (three-dimensional
and side views) with sinusoidal leading and trailing edge protuberances. The size of
computational domain in the spanwise direction is 0.8L.

Figure 1.5: The rounded tip in case of finite-span wing is shown in the upper figure. In
the planform view (lower figure), size of the wing in the spanwise direction is 4L.
Chapter 2

Literature Review

At higher Reynolds numbers, the introduction of leading-edge protuberances in airfoils

primarily benefits the aerodynamic performance mainly after the stall. Before reaching
the stall angle, there is a minimal difference observed between the original airfoil and the
modified version (Johari et al., 2007). Extensive research has been conducted in this field,
with various researchers proposing different types of modified airfoils to enhance their
performance. These modified airfoils feature leading-edge protuberances in triangular,
square and rectangular shapes.
Additionally, by altering the wavelength and amplitude of these protuberances, re-
searchers aim to improve the overall performance of the airfoils. The impact of altering
the amplitude of waves on leading-edge protuberances is more significant compared to
changes in wavelength. Various experimental data and simulation results have confirmed
the greater sensitivity of amplitude on the performance of airfoils (Johari et al., 2007).
For higher Reynolds numbers before reaching the stall angle, the performance of the
baseline airfoil was better. However, the major difference occurs after reaching the stall
angle.In the pre-stall angle range,baseline airfoils exhibit higherlift and lower drag val-
ues. Conversely, modified airfoils display higher lift and lower drag values in post-stall
angles. Notably, baseline airfoils experience a sudden decrease in lift after reaching the
stall angle, which is potentially hazardous. On the other hand, modified airfoils do not

6
Chapter 2. Literature Review 7

exhibit this sudden drop in lift. The influence of the Reynolds number on the effective-
ness of leading-edge protuberances has been investigated. The findings indicate that at
lower Reynolds numbers, the protuberances are effective at both higher and lower angles
of attack, whereas this effect is diminished at lower angles of attack for higher Reynolds
numbers.
Chapter 3

Methodology

Recent numerical studies on effects of leading-edge protuberances are reported based on

Large-Eddy Simulation (LES) technique (Yasuda et al., 2019). In the present study,
Reynolds-Averaged-Navier-Stokes (RANS) simulations are conducted for flow past base-
line and modified airfoil geometries at Reynolds numbers 20, 000 and 43, 200 for angles
of attack of 4o and 14o using commercial software Ansys Fluent.

Parameter CaseA CaseB CaseC CaseD

Re 2 × 104 43200 2 × 104 43200

α 4o 4o 14o 14o

Table 3.1: Flow parameters for simulation cases A to D

3.1 Governing equations

The steady RANS equations for incompressible flow in three-dimensions are given below.
The · represents a time-averaged quantity.

8
Chapter 3. Methodology 9

∂ui
=0
∂xi
∂ui 1 ∂p ∂ 2 ui ∂ 0 0
uj =− +ν − uu
∂xj ρ ∂xi ∂xj ∂xj ∂xj i j

The Shear Stress Transport k − ω turbulence model of Menter (1994) is used in this
study for the unclosed term u0i u0j . The Shear Stress Transport k − ω model is a two-
equation model for turbulence kinetic energy (k) and specific dissipation rate (ω). The
equations for k and ω are obtained as follows after combining the standard k − ω model
with a transformed k −  model using a multiplier function F1 . The combination of two
models is designed to utilize k −ω model in the near-wall region and maintain free-stream
independence of the k −  model in the outer region away from the wall.

 
D ∗ ∂ ∂k
ρk = P − β ρωk + (µ + σk µt )
Dt ∂xj ∂xj
 
D γ 2 ∂ ∂ω
ρω = P − βρω + (µ + σω µt )
Dt νt ∂xj ∂xj
ρσω2 ∂k ∂ω
+ 2(1 − F1 )
ω ∂xj ∂xj

The first, second and third terms on the right side of the above equations are the
production, dissipation and transport terms for k and ω where P = τij ∂vi /∂xj . The
fourth term on the right side of equation for ω is the cross diffusion term. The constants
from the standard k − ω model and transformed k −  model are linearly blended using
the weight function F1 defined below. The eddy viscosity is obtained using the following
relation using k and ω.
Chapter 3. Methodology 10

a1 k
νt = , a1 = 0.31
max(a1 ω; ΩF2 )
√ !
2 k 500ν
F2 = tanh(arg22 ) , arg2 = max ;
0.09ωy y 2 ω

F1 = tanh(arg14 ) ,
" √ ! #
k 500ν 4ρσω2 k
arg1 = min max ; ; 2
0.09ωy y 2 ω y CDkω

p
where Ω = 2Wij Wij is the absolute value of vorticity in terms of rotation rate
tensor Wij and the function F2 is defined such that its value is 1 for boundary layer
flows and 0 for free shear layers. CDkω is the positive part of cross diffusion term in the
equation for ω.
The equations were solved using the the SIMPLE algorithm implemented in the finite
volume solver ANSYS Fluent. A second order upwind discretization is applied for the
mean flow equations. The gradients at the cell centers are computed using the least
squares cell-based method, and to prevent numerical oscillations, a multi-dimensional
gradient limiter is applied. The turbulence quantities, namely the turbulent kinetic
energy and the specific dissipation rate are also discretized using a second order upwind
scheme.

3.2 Mesh for infinite-span airfoil

For infinite-span airfoil analysis the size of the domain along the z-axis was made equal
to the span of the airfoil, that is 44mm or 0.8L, where L = 55mm is the chord length
of airfoil. The upstream inlet boundary of the computational domain is located 500mm
or 9L from the leading edge, while the downstream outlet boundary is located 1445mm
or 26L from the trailing edge. The lateral boundaries are located 500mm or 9L away
from the airfoil surface at α = 0o . These distances change slightly as α is increased. The
domain material is set to air at standard conditions of temperature and pressure. At the
Chapter 3. Methodology 11

Figure 3.1: In the upper figure, side view of the computational domain is shown for the
case with airfoil of infinite span. A zoomed three-dimensional view of the mesh around
the airfoil (boxed region in the upper figure) is shown in the lower figure.

inlet, speeds of the incoming flow are specified as 5.42m/s and 11.71m/s for Reynolds
number values of 2 × 104 and 43200 respectively. For the incoming flow, the turbulence
intensity is 0.5% with turbulent viscosity ratio of 1. The boundary conditions included
a velocity inlet, a pressure outlet, symmetry on the side-span boundaries and no-slip
on the top and bottom boundaries, and a no-slip wall on the airfoil surface. For the
case with both leading- and trailing- edge protuberances, the mesh specifications and
boundary conditions are the same as described above.

3.3 Mesh for finite-span airfoil

RANS simulation for flow past a finite-span airfoil is also conducted. For this simulation,
the airfoil has a span of 220mm or 4L and it is placed in a semi-cylindrical computational
domain of radius five times the span length (refer to figure 3.2). The axial length of
domain is 2000mm. A sinusoidal guiding curve with an amplitude of 2.75mm and a
wavelength of 11mm is used to define the variation along the span. The far end of the
Chapter 3. Methodology 12

Figure 3.2: A three-dimensional view of the mesh around the finite-span airfoil is shown.

airfoil span is smoothed by applying the revolve command in SolidWorks. At this end,
the airfoil geometry is rotated by 180 degrees to create a smooth transition. On the
side walls symmetry boundary condition is used and velocity-inlet and pressure-outlet
conditions are specified at the front and back face of the domain.

3.4 Lifting-line theory

Prandtl’s lifting line theory provides a basic framework to calculate induced drag. It
can be used for estimating the lift and induced drag on a finite wing under the following
conditions: (a) boundary layer effects are negligible and angle of attack is low , (b) the
wing is thin and at low subsonic speeds (nearly incompressible flow), (c) there is no wing
sweep, (d) the dihedral angle between the wing and fuselage is zero (the wing emerges
perpendicular to the fuselage surface), (e) aspect ratio is large (R ≥ 4).
The lifting line theory is based on irrotational flow. The integro-differential equation
given in equation 3.1 can be solved for the unknown circulation distribution along the
wing span. The first term in the equation is the zero-lift angle and the combination of
first and second terms is the effective angle. The geometric angle is obtained by adding
the induced angle which depends on the derivative of circulation function, Γ0 (z), with
respect to the spanwise coordinate z. When solved for the circulation function Γ(z), the
lift and induced drag coefficients can be directly calculated using equation 3.2.
Chapter 3. Methodology 13

b/2
Γ0 (ζ)
Z
Γ(z) 1
α(z) = αL=0 (z) + + dζ (3.1)
πU0 c(z) 4πU0 −b/2 z−ζ

Z b/2
2
CL = Γ(ζ)dζ (3.2)
U0 S −b/2
b/2 b/2
Γ0 (ζ)
Z Z
2 1
CD,induced = Γ(ζ)αi (ζ)dζ where αi (z) = dζ
U0 S −b/2 4πU0 −b/2 z−ζ

p
For an elliptical distribution of circulation along the span, Γ(z) = Γ0 1 − 4z 2 /b2 ,
the induced angle αi (z) is a constant αi = CL /πR and the induced drag is CD,induced =
CL2 /πR. For obtaining an elliptical lift distribution at each section [= ρ0 U0 Γ(z)], the
chord must also be an elliptical function of z (when there is no aerodynamic twist and
no geometric twist). The span efficiency factor e is a measure of deviation from the
elliptical lift distribution which gives the minimum induced drag.

3.4.1 Solutions using lifting-line theory

To solve the lifting line equation 3.1, the variable along the span ζ is transformed using
the equation 3.3 and the circulation function is expanded in terms of sine series.

b b
ζ = − cos θ and dζ = sin θdθ (3.3)
2 2
N
X
Γ(θ) = 2bU0 An sin(nθ) and Γ0 (ζ)dζ = Γ0 (θ)dθ
n=1
N
X
= 2bU0 nAn cos(nθ)dθ
n=1

For z = − 2b cos θ0 , substituting the above transformations in the fundamental relation

3.1, and using the identity,
Chapter 3. Methodology 14

Z π
cos nθ sin nθ0
dθ = π
0 cos θ − cos θ0 sin θ0

the following system of linear equations are obtained for the coefficients An in terms
of µ(θ0 ) = πc(θ0 )/2b. In the equation 3.4, α(θ0 ) is the local angle of attack and αL=0 (θ0 )
is the local zero-lift angle of attack; the difference between the two gives the local effective
angle of attack.

N  
X nµ(θ0 )
µ(θ0 ) [α(θ0 ) − αL=0 (θ0 )] = An sin(nθ) 1 + (3.4)
n=1
sin θ

The lift coefficient for wing is calculated in terms of the wing aspect ratio R = b2 /S
as,

b/2 Z π
2b2 X
Z
2
CL = Γ(y)dy = An sin nθ sin θdθ
U0 S −b/2 S 0

= πRA1

The induced drag coefficient is calculated as

Z b/2
2
CD,i = Γ(y)αi (y)dy
U0 S −b/2

The induced angle of attack αi (z) is obtained after substituting the series expansion
for the circulation function and is given by,
Chapter 3. Methodology 15

Z b/2 0
1 Γ (ζ)
αi (z) = dζ
4πU0 −b/2 z − ζ
X sin nθ0
αi (θ0 ) = nAn
sin θ0

Thus, the induced drag coefficient is obtained in terms of the span efficiency factor e
as,

" N  2 #
X An
CD,i = πRA21 1+ n
n=2
A1
CL2
=
πeR
Chapter 4

Results and Discussion

4.1 Infinite-span airfoil with leading edge protuber-

ances
The results for the lift and drag coefficients are listed in tables 4.1 and compared with
the experimental data of Yasuda et al. (2019) for the infinite-span airfoil geometry with
leading-edge protuberances. We can observe from the simulation results of an infinite
span airfoil that at lower angle of attack, the addition of leading-edge protuberances did
not provide a significant aerodynamic advantage when the value of the Reynolds number
was 43200. However, at a higher angle of attack, the airfoil with protuberances exhibited
a higher lift and lower drag, suggesting improved stall resistance and delayed flow sep-
aration. These findings align with the expected aerodynamic benefits of protuberances,
making them effective for applications where performance at higher angles of attack is
critical. We can also observe from the results below that at a lower Reynolds number of
20,000, the protuberances were effective not only at the lower angle of attack but also at
higher angle of attack. This contrasts with the case with a higher Reynolds number of
43,200, where their effectiveness was primarily limited to higher angles of attack. The
following tables show the results of our RANS simulation and the experimental values of
Cd and Cl taken from the paper for the modified profile at Reynolds numbers of 20000

16
Chapter 4. Results and Discussion 17

Quantity CaseA CaseB CaseC CaseD

Re = 20000 Re = 43200 Re = 20000 Re = 43200
α = 4o α = 4o α = 14o α = 14o

CD (RANS) 0.036 0.028 0.17 0.17

CD [(Yasuda et al., 2019) exp.] 0.050 0.032 0.21 0.19
CD [baseline exp.] 0.055 0.036 0.224 0.21
CL (RANS) 0.37 0.37 0.59 0.66
CL [(Yasuda et al., 2019) exp.] 0.35 0.35 0.69 0.72
CL [baseline exp.] 0.18 0.423 0.55 0.561

Table 4.1: Lift and drag coefficients obtained for simulation cases A to D and their
comparison with experimental data.

Figure 4.1: Velocity and pressure contours are shown for the airfoil with leading-edge
protuberances at Re = 20000 and α = 4o at mid span (22 mm)

and 43200.
Chapter 4. Results and Discussion 18

Figure 4.2: Velocity and pressure contours are shown for the airfoil with leading-edge
protuberances at Re = 20000 and α = 14o at mid span (22 mm)

Figure 4.3: Velocity and pressure contours are shown for the airfoil with leading-edge
protuberances at Re = 43200 and α = 4o at mid span (22 mm)

4.2 Finite-span airfoil with leading edge protuber-

ances
By comparing the simulation results of finite-span and infinite-span airfoils, it is evident
that the finite-span airfoil generates a lower lift and experiences higher drag compared
to the infinite-span configuration. This discrepancy can be attributed to the wingtip
vortices and the three-dimensional flow effects present in finite wings, which are absent
in 2D or infinite-span airfoil simulations.
Chapter 4. Results and Discussion 19

Figure 4.4: Velocity and pressure contours are shown for the airfoil with leading-edge
protuberances at Re = 43200 and α = 14o at mid span (22 mm)

Different researchers have given different theories for explaining the working mech-
anism of vortices, out all the given theories the most accepted one is the theory of
streamwise vortices. We have tried to understand this physicality by using the contours
of the vorticity magnitude. Through diagrams, we can clearly see that counterrotating
vortices are being generated because of the presence of protuberances, and these vor-
tices are moving in the streamwise direction. And once the flow passes, the airfoil is
divided into two high vorticity regions. These counterrotating vortices grow in size, and
their magnitude decreases as they move along the streamwise direction. We have taken
different planes at different locations along the stream-wise direction. We have drawn
contours of the x-vorticity in this plane, which describe the curl of the velocity as a
measure of fluid rotation in the x direction. From the contours, we can observe that the
counter-rotating vortex pair generated behind the protuberances is almost symmetric.
Further downstream, the size of the vortices increases, which represents a decrease in
intensity.
Many studies have shown that at lower angles of attack, the flow remains almost
attached to the suction surface of airfoils. Generally, in the case of a baseline airfoil, it is
seen that at lower angles of attack, a laminar separation bubble starts forming at some
distance from the leading edge, but after some distance on the airfoil, it disappears. And
Chapter 4. Results and Discussion 20

Simulation CD CL
Re = 20000, α = 4o 0.065 0.297
Re = 43200, α = 4o 0.057 0.311
Re = 20000, α = 14o 0.261 0.451
Re = 43200, α = 14o 0.258 0.471

Table 4.2: Lift and drag coefficients for a finite span airfoil with leading edge protuber-
ances at varying angles of attack and Reynolds numbers.

then again, flow separation is seen near the trailing edge of the airfoil. With the further
increase in the angle of attack, it is observed that the initial point of flow separation
starts moving towards the leading edge, and the intensity of flow separation also starts
increasing.
In case of a baseline airfoil, there is a uniform pressure distribution on the suction
surface, but due to protuberances, the pressure distribution on the suction surface be-
comes nonuniform, which results in the generation of different intensities of separation
in the spanwise direction of the airfoil.
Mostly, the flow separation occurs at the trough of the leading edge and extends
to the trailing edge of the airfoil. Many studies have also shown that sometimes the
vortices induced by protuberances negatively affect the uniform pressure distribution on
the airfoil surface, resulting in early flow separation and hence decreasing aerodynamic
performance of the wavy airfoil in the pre-stall region.
A greater flow separation is seen when the angle of attack is larger. In a low-pressure
area, the flow remains almost attached near the leading edge, but this low-pressure area
extends up to a smaller distance from the leading edge. Because of this, the separation
point is much closer to the leading edge, and flow separation occurs earlier.
Chapter 4. Results and Discussion 21

Figure 4.5: Vorticity contour for airfoil at 5mm, 25mm, 55mm and 100mm from the
leading edge at Re=43200 and α = 14o

4.3 Infinite-span airfoil with leading and trailing edge

protuberances
Having discussed the aerodynamic effects of introducing protuberances at the leading
edge of the airfoil, we now turn our attention to an airfoil configuration that incorporates
protuberances on both the leading and trailing edges. In this section, we will present
and analyze the corresponding simulation results. These findings will then be compared
with those of the airfoil featuring protuberances only at the leading edge, to evaluate
which configuration offers superior aerodynamic performance.
To enhance the aerodynamic performance of the airfoil, protuberances were intro-
duced at both the leading and trailing edges. However, the simulation results revealed
Chapter 4. Results and Discussion 22

Figure 4.6: Vorticity contour for airfoil at 5mm, 25mm, 55mm and 100mm from the
leading edge at Re=43200 and α = 4o

that this configuration did not yield a significant improvement over the case with pro-
tuberances only at the leading edge. On the contrary, adding protuberances to both
edges led to a noticeable increase in drag and a reduction in lift. This is likely due to in-
creased flow disturbances and enhanced separation effects near the trailing edge, which
can negatively impact pressure recovery and overall lift generation. The trailing-edge
protuberances may interfere with the naturally smooth flow deceleration in that region,
thereby reducing aerodynamic efficiency. Based on these observations, it is more benefi-
cial to limit the placement of protuberances to the leading edge, where they effectively
delay flow separation and improve stall performance without compromising lift-to-drag
ratio.
Chapter 4. Results and Discussion 23

Figure 4.7: Pressure variation on the suction surface of airfoil at Re = 43200, α = 4o

and α = 14o

Figure 4.8: Recirculation zone (isosurface ux=0) for a finite span airfoil having leading
edge protuberances, Re = 43200 and α = 14o
Chapter 4. Results and Discussion 24

Simulation CD CL
Re = 20000, α = 4o 0.059 0.338
Re = 43200, α = 4o 0.039 0.354
Re = 20000, α = 14o 0.214 0.579
Re = 43200, α = 14o 0.215 0.595

Table 4.3: Lift and drag coefficients for an infinite span airfoil with leading and trailing
edge protuberances at varying angles of attack and Reynolds numbers.

Figure 4.9: Velocity and Pressure contour for airfoil with leading-edge and trailing-edge
protuberances at Re = 20000, α = 4o and α = 14o
Chapter 4. Results and Discussion 25

Figure 4.10: Velocity and Pressure contour for airfoil with leading-edge and trailing-edge
protuberances at Re = 43200, α = 4o and α = 14o
Chapter 5

Conclusions and Future Work

This study demonstrates that leading-edge protuberances (LEPs) are effective passive
flow control devices for the NACA0012 airfoil, especially at low Reynolds numbers. At
lower Re, LEPs enhance performance at both low and high angles of attack, while at
higher Re, their effectiveness becomes more pronounced at higher angles only. The
separation control is attributed to streamwise vortices generated by the LEPs, which re-
energize the boundary layer. RANS simulations further confirm that finite-span wings
experience higher drag and lower lift due to 3D effects like tip vortices, underscoring
the need for finite-span analysis in LEP studies. Additionally, adding protuberances to
both the leading and trailing edges degraded aerodynamic performance by increasing
drag and reducing lift. Therefore, incorporating protuberances only at the leading edge
is recommended for optimal aerodynamic efficiency.

26
References

Anderson, J. D. (2010). Fundamentals of Aerodynamics. McGraw-Hill.

el Hak, M. G. and Bushnell, D. M. (1991). Separation control: Review. Journal of Fluids

Engineering, 13:5–30.

Johari, H., Henoch, C., Custodio, D., and Levshin, A. (2007). Effects of leading-edge
protuberances on airfoil performance. AIAA Journal, 45:2634–2642.

Menter, F. R. (1994). Two-equation eddy-viscosity turbulence models for engineering

applications. AIAA Journal, 32:1598–1605.

Mueller, T. J. and DeLaurier, J. D. (2003). Aerodynamics of small vehicles. Annual

Review of Fluid Mechanics, 35(1):89–111.

Yasuda, T., Fukui, K., Matsuo, K., Minagawa, H., and Kurimoto, R. (2019). Effect
of the reynolds number on the performance of a naca0012 wing with leading edge
protuberance at low reynolds numbers. Flow, Turbulence and Combustion, 102:435–
455.

27