See discussions, stats, and author profiles for this publication at: [Link]

net/publication/262934814

Two-Dimensional Assessment of A Cambered and A Symmetric Spoiler at

Various Angles of Attack and Speeds: A CFD Study

Article in i-manager’s Journal on Future Engineering and Technology · April 2010

DOI: 10.26634/jfet.5.3.1139

CITATIONS READS

3 620

2 authors:

Tuncay Kamas Mohammed A Omar

33 PUBLICATIONS 141 CITATIONS

Khalifa University
237 PUBLICATIONS 7,882 CITATIONS
SEE PROFILE
SEE PROFILE

All content following this page was uploaded by Tuncay Kamas on 09 June 2014.

The user has requested enhancement of the downloaded file.

Two-Dimensional Assessment of A Cambered and A Symmetric Spoiler at Various
Angles of Attack and Speeds: A CFD Study

Tuncay Kamasa, Mohammed Omarb

aMechanical Engineering Department, Clemson University, Clemson SC 29634, USA

bClemson University International Center for Automotive Research CU-ICAR, Greenville SC

29607, USA

Keywords: Rear-wing, Computational Fluid Dynamics, downforce, RANS, k-ε turbulence

model

Abstract

A two-dimensional and three-dimensional computational study has been carried out

respectively regarding aerodynamic forces affecting a symmetric spoiler, NACA0015 and a
cambered spoiler, NACA2414. The negative lift (down force) and drag forces were predicted
through the simulation of airflows over inverted rear-wings (spoilers) in different
configurations namely; varying incidences i.e. angles of attack of the airfoils, two
dimensional cross-section of the wings, and varying speeds of initial airflow. The
downforce increases as the angle of attack increases however, at high angle of attacks the
rear-wing starts to stall which is not a desired condition affecting the vehicle stability.
Furthermore, the camberness added to the airfoil increased the downforce, which
contributes more to the stability concept of the automobile.

Corresponding Author:
Mohammed Omar, Tel: 864 283 7226, momar@[Link]
 2

Nomenclatures:
 (t ) : Expected flow property that is a function of time
 : Time averaged flow property
 '(t ) : The fluctuation of the flow property that is a function of time
 : Density of the air
U: Velocity of the airflow
P: Pressure
µ: Dynamic viscosity
µt: Turbulence viscosity

ui ' u j ' : Reynolds stress term

k: Kinetic energy per unit mass

 ij : Kronecker delta (0 if i≠j; 1 if i=j)

 ij : Shear stress tensor

ε: Dissipation rate
 3

Two-Dimensional Assessment of A Cambered and A Symmetric Spoiler at Various

Angles of Attack and Speeds: A CFD Study

1. Introduction:
Aerodynamics of inverted rearwings (spoilers) of race cars and road cars has been a
major concern in automobile industry due to increase of efficiency in terms of gasoline
consumption and of stability of the overall car. Therefore, design concepts of airfoils need
further attention in order to develop more efficient and more stable automobiles. In
inverted wing concept, researchers conduct computational studies regarding downforce
(lift) and drag forces which affect on either symmetric or cambered (asymmetric)rearwings,
mostly two-dimensional and three-dimensional models.
Computational investigations of both double-element and single-element inverted
wings in ground effect, for instance, have been numerically conducted by Mahon and Zhang
in different studies referred by Ref [1] and [2] respectively. The comparison among six
turbulence models was carried out and the outcomes quantitatively obtained such that k-ε
model has the best prediction of the surface pressures and the wake entities of downstream
of airflow domain concerned as the height clearance went down in Computational Analysis
of an Inverted Double-Element Airfoil in Ground Effect [1], nevertheless by the same
authors, in Computational Analysis of Pressure and Wake Characteristics of an Airfoil in
Ground Effect article regarding single-element airfoil, the k-ω SST predicts better the
surface pressure and the Realizable k-ε better the wake flow field at variety of ride heights
of single-element airfoil. Kieffer et-al [3] conducted a numerical investigation utilizing k- ε
model to provide negative lift and drag coefficients plots along with pressure coefficients
plots and velocity contour plots at different height and different angle of attack of a 2D cross
section of front wing and in different angle of attack of a 2D aerofoil of rearwing. The
research indicates the stall condition in high angle of attack (12° and 16°) complying with
the experimental validation for Formula Mazda race car. Ranzenbach et-al [4] numerically
simulated 2D cambered airfoil NACA 4412 in two sets of boundary conditions, no-slip wind
tunnel and road conditions where the ground and freestream flow moving. Ref [4]
introduces force reversal phenomena and provides a comparison to symmetrical airfoil
NACA 0015. The preliminary study [5] of symmetric NACA0015 airfoil by the same authors
as [4] carried out a comparison of experimental data to computational results in the same
set of boundary conditions as those described in [4]. The entity that shall be perceived in
this research is symmetricalness and whether there exists any contribution of asymmetric
inverted airfoil. This work shall carry out an investigation utilizing proper turbulence
models.
Furthermore, in order to computationally investigate the both symmetric and
asymmetric airfoil models, NACA 0015 symmetric and NACA 2414 asymmetric 2D airfoils
will be utilized. As computational fluid dynamics methods, the Reynolds Averaged Navier-
Stokes Equations (RANS) and the one-equation k-ε turbulence model derived from RANS
will be employed in this research. Subsonic flow will be taken into account with low
baseline Reynolds number which is based on the chord is 2.01x106 in both cases. Boundary
layers of the airfoils are meshed with structured grids and rest of the computational domain
is meshed with unstructured tridiagonal cells in order to lessen the time to compute. The
free airflow velocities are set at 30, 35, 40, 45 and 50 m/s at angles of attack of 0 and 4
degree and also set at 30 m/s at the higher levels of angles of attack of 8, 12, and 16 degree.
Second order accurate upwinding scheme and SIMPLE algorithm are utilized to descritize
the Navier-Stokes equations to approach the solution on a commercial simulation software,
Fluent 6.3.26.
 4

The significant properties of the airflow over the 2D airfoils shall be captured to
obtain qualitative results so that pressure and velocity contour plots are investigated in
regard with the airflow in the upstream and downstream domains, moreover, the pressure
coefficient values varying along the airfoils’ top and bottom edges from the leading edges
through the trailing edges of both the symmetric and the cambered airfoils. The distances of
computational domains are determined corresponding to the values used by Kieffer et-al in
reference [3]. The solver selected on Fluent was pressure based in order to gain pressure
coefficient versus position plots and other entities regarding the scheme of the solver are
chosen as implicit and steady flow.

2. COMPUTATIONAL FLUID DYNAMICS

In order to understand the physical phenomena in terms of aerodynamics forces

arisen by an airfoil over which airflow crosses, researchers have used some methodologies
to acquire and analyze data. Some computational methods are to be presented herein.

2.1 Reynolds Averaged Navier Stokes (RANS) Equations

Approaching the solution of turbulence feature precisely at every single time

requires very long computational process and too high computing power in supercomputer
system which might cost unaffordable prize. This is the main reason why researchers prefer
compute the averaged properties of turbulent flow over interval time, which is quite
satisfying in many cases. The flow properties (velocities, pressures, and stresses) are overall
symbolized by  and time averaged properties (mean velocities, mean pressures, and mean
stresses) by  and formalized as the following

The ensemble average and the fluctuation of the flow properties,  '  t  , that is function of
time are superimposed so that time averaged flow property is obtained as indicated below:
 t      ' t  (1)
Reynolds Averaged Navier Stokes (RANS) equations are derived from general Navier Stokes
Equations along with the statistical assumption above. Hence;
  U iU j  P   U 
     ui ' u j ' (2)
xi x j xi  xi 
whereas μ is the dynamic viscosity and ρ is the fluid density and assumed as a constant
while fluctuations of fluid density insignificantly shows up and possesses unimportant
effects on the flow up to around Mach number 3-5 [7]. Since in this particular case of the
research, as investigating subsonic flow over a car, i.e. much smaller than Mach number 1,
the density does not importantly fluctuate, thus indeed the fluctuation is assumed as zero
and the density as a constant. The correlation between the density and both velocity
fluctuation components is named as Reynolds stress and is non-zero while two fluctuations
of the velocity components are statistically dependent on one another. Therefore Reynolds
stress term is an additional unknown and needs to be solved to obtain the resolution of the
RANS equations. Boussinesq’s assumption is suggested to relate the Reynolds stress to
constant turbulent viscosity by the following equation;
 U U j 2 U k  2
  ui ' u j '  t  i    ij    k ij (3)
 x x 3 x
 j i k  3
 5

Where k represents the kinetic energy per unit mass and the equation below defines k as
the half of the production from the correlation between two velocity fluctuation
components;
ui ' u j '
k (4)
2
The mean flow features along with the effects of turbulent fluctuations averaged out
the turbulence through Reynolds stress   ui ' u j ' . The various models have been
developed regarding turbulent flows employing RANS in the range from the one-equation
turbulent models to Reynolds-stress model. RANS models enable users to obtain sufficiently
promising resolution rapidly and do not require too much power consuming process.
Therefore RANS equation is chosen for this particular inverted rear-wing project and the
methods are further discussed in the later sections.

2.2 Two-Equation Turbulence Models

In order to include the effects from history of flow and to predict wake length, two
transport equations are derived from Navier-Stokes equation. If two transport equations
are included in the model, it is called two-equation turbulence model and many kinds of
two-equation turbulence models have been proposed over the past a few decades and one
of them shall be present in this section.

2.2.1 k- ε Two-Equation Turbulence Model

The models need to be adjusted at different levels of flow such as from wall-
bounded to free-shear flows, i.e. flow over a trailing edge of an airfoil [8]; Therefore, the
more accommodating model to various kinds of flows is, the more universal model is,
particularly in this research, for separated flows over an airfoil, transport effects on the
turbulence length scale shall be taken into account. Two equation turbulence models are
generally based upon Bousinessq approximation [8].
The second order accurate upwinding discretization scheme and SIMPLE algorithm
are chosen to approach the solution.
The turbulent kinetic equation in Cartesian coordinates for two dimensional flowfields is as
the following;
 k k k     k      k 
  u  v    (   t )      t    Pk   (5)
 t x y  x   k x  y  k  y 
where

Pk is the production of turbulence and Pk   ij ui / x j 
The dissipation rate equation in Cartesian coordinates in two dimension is;

                  2
 u  v      t       t 
  1 k c P  c2  (6)
 dt dx dy  dx     dx  dy    
dy k k

where
the constants are σk=1.0 σε=1.3 σε1=1.44 σε2=1.92
 6

k2
and μt is the turbulent viscosity and t   c in relation with ε and with cμ=0.09[9]


3. Geometric features of the airfoils and computational domains and flow properties:

The airfoils utilized in this study are NACA 2414 cambered (asymmetric) and
NACA0015 asymmetric airfoils and the point data files identifying the shape of the airfoils
are obtained from UIUC Airfoils Coordinates Database at the webpage link
[Link] . The chord of the both airfoils c
is 1000 mm and the angle of attack determined is 0,4,8,12, and 16 deg as the features of the
airfoils are indicated in Figure 1. The origin of the Cartesian coordinate system is located
right in the most front point of the leading edge of the airfoil. The structured mesh of
boundary layer region and the unstructured mesh are also demonstrated in the Figure 1.

(a) (b)
Figure 1: Geometry of the airfoils (a) NACA2414 cambered airfoil (b) NACA0015 symmetric airfoil

The Reynolds number based on chord length is determined 2.01x106 along with the
baseline freestream velocity of 30 m/s in both cambered and symmetric cases.

(-1.75, 2.59) (4, 2.59)

(-1.75, -2.59) (4, 2.59)

Figure 2: Geometry of the computational domain for NACA0015 correspond the values from Ref 13.
 7

In figure 2, the x, y coordinate of the computational domain corners are given

corresponding to the values given by Kieffer et-al.

A velocity inlet boundary condition for the upstream boundary and an outlet vent boundary
condition for the downstream boundary are set at a uniform velocity profile of 30 m/s for
the baseline simulation and the gauge pressure is set zero and slip boundary condition is
imposed.

4. Validation of The Computational Process:

As a preliminary study the Spallart-Allmaras method had been utilized, it can be said
that the Spallart-Allmaras one-equation turbulence model can be employed to capture only
the major changes in the surface pressure features along the surfaces of the symmetric
airfoils. The cambered airfoil possesses enough complexity in terms of its feature for the
turbulence model employed not to be used to capture sharp pressure diversity such as
stagnation pressure on the leading edge and the suction peak along the suction surface
(lower one in the inverted case). Consequently it was decided to use the k-ε turbulence
model as a two-equation turbulence model since the two equation model provides more
accurate results in comparison to the one-equation model, Spallart-Allmaras model.
The validation process has also been carried out via the realization of a simulation
of the symmetric airfoil at zero angle of attack.

Figure 3: Pressure coefficient and velocity magnitude plots of NACA0015 at AOA=0

The figure 3 of pressure coefficient plots obtained from the simulation of the airflow
over NACA0015 symmetric airfoil can be used as well as the velocity magnitude plots to
verify the computational method utilized. The distribution of the value points of the
pressure coefficients on both the upper edge and lower edge should be conflicting with fully
agreement along the chord length due to the symmetry of the airfoil and the absence of an
angle of attack. Since the two curves are in very close agreement with only negligible
fluctuations in the region nearby the leading edge and with even more little ones in the
vicinity of the trailing edge, one can conclude that the turbulence method, the scheme, and
the algorithm chosen provided proper and applicable results from which more conclusions
 8

can be drawn. The similar observation can be carried out on the plot which depicts the
velocity magnitude distribution of the same symmetric airfoil at zero angle of attack.

5. Results and Discussion:

a-) b-)
Figure 4. Distribution of Pressure coefficient versus position along the lower and upper surfaces of a-)NACA
0015 airfoil and b-)NACA 2414 airfoil at AOA=4 degree

A qualitative evaluation of the comparison of two airfoils was conducted to have

better understanding of whether a contribution exists via the cambered airfoil to desired
pressure distribution. The most outstanding contribution might happen in the vicinity of the
trailing edges since significant flow phenomena such as separations, transition to
turbulence starts to form about the trailing edge and wake is also generated in the
downstream domain led by the trailing edge. Moreover, since some significant airflow
phenomenon also occurs about the leading edge such as suction peak and stagnation
pressure. Suction peak was defined as the value of maxium suction along the lower surface
of the inverted airfoil in the Ref [9] by Mahon et-al. Stagnation pressure may be defined as a
static pressure at stagnation point that is in the front of the leading edge where the local
velocity of the flow is zero thus the static pressure has its maximum magnitude. Suction
peak and subsequently stabilized surface pressure along the suction surface of the NACA
0015 symmetric airfoil was captured as -1042 by the turbulence model approximately at
x/c of 0.027 as can be seen in Figure 3. Suction peak value of Cp of the cambered airfoil was
larger than that of the symmetric airfoil and appears at further location in x direction over
which the distribution of Cp smears away. A pressure gradient on the suction surface of the
NACA 0015 is in the inclination form in the downstream of the leading edge whereas the
pressure increase is more likely linear along the suction surface of the NACA 2414 airfoil.
The distributions of the values of the pressure coefficients of the two airfoils appear
to be in close agreement along the pressure (upper) surface. They appear at the largest
values and begin sharply decreasing up to zero of Cps at the stagnation point however the
pressure declination of that of the cambered airfoil is sharper. One can realize that there
exists only a slight difference, as decreasing, between the Cp distributions of NACA 0015 and
NACA 2414. However, on the suction surfaces, a large discrepancy occurs. The large variety
on the Cp distribution along the suction surface draws attention and a contribution of the
cambered airfoil in terms of higher magnitude of downforce (inverse lift) and more stability
can be presumed through this diversity. Therefore, the pressure difference between the
upper and lower surfaces of the symmetric airfoil is significantly less in comparison to that
of the cambered airfoil. Consequently, it is evident such that the cambered airfoil causes
 9

more downforce than the symmetric one in the same angle of attack of 4 degree as might be
desired for the stability concern of the car.

Pressure gradient with respect to X position along the chord length of NACA0015 at
angle of attack of 4, on Figure 5, reaches the smallest value between 0.2-0.3 on the upper
edge at the speed of 50m/s. In the same interval of X position along the length, the
discrepancies also appear where the pressure value decreases as the airflow accelerates.
The pressure values at all speeds of airflow happen to meet about Cp of zero at X position of
0.85 on both lower and upper edges where the turbulence flow starts to be generated. The
reverse pressure gradient phenomenon is also to start at the same time, which forces the
velocity magnitude to dramatically decrease. This phenomenon can also be observed on the
plot which demonstrates the velocity magnitude along the upper edge of the same
symmetric airfoil. The decrease in velocity and reversed pressure result in separation and
the boundary layer around the trailing edge turns into wake generated right behind the
trailing edge. The thickness of the generated wake became larger as the airfoils were set at
the higher angle of attacks.

a-) b-)

Figure 5. Distribution of Pressure coefficient versus position along the upper surfaces of a-)NACA 0015 airfoil and
b-) NACA 2414 airfoil at AOA=4 degree
 10

Since the airfoil is symmetric, one can see the same region at X position of 0.85 on
the lower edge to be hosting the same reverse pressure gradient phenomenon however the
distribution of the pressure coefficient values are to be in different manner, gradually
increasing up to around zero value of Cp along the lower edge of the airfoil. The difference
in the magnitude between Cp along the upper edge and that along the lower edge can also
be obviously realized due to the sharp decrease of Cp value between the X position of 0.01
and 0.02 on the lower edge of NACA0015, which is above called the suction peak.

a-) b-)
Figure 6 . Distribution of Pressure coefficient versus position along the suction surfaces of a-)NACA 0015 airfoil and
b-) NACA 2414 airfoil at AOA=4 degree

The more speed the airflows over the both airfoils have, the lower surface pressure
exists along the suction edge until the region nearby the trailing edge. The pressure values
at different speeds match up about zero pressure in the vicinity of the trailing edge of the
both airfoils. The negative pressure values on the suction surface avail the airfoil to fulfill
the generation of the downforce that will be used for more stability when the car speeds up.
When one who takes a look at the Figure 6.b could easily understand that the pressure
values are even more lowering down to ≈-2500 Pa on the plot which belongs to the
cambered airfoil, NACA2414 when the car speeds up to 50 m/s so then this figure also
shows that the cambered airfoils enable an aerodynamicist designer to obtain higher
downforce mounting a cambered rearwing behind the car.
The analysis of velocity magnitude was assessed by the simulation over the both
airfoils as well. As already mentioned on the discussion of the pressure diversity, the
velocity magnitude of the flow over the airfoils also varies as seen on Figure 7. As always
referred to Bernoulli equation while explaining the relationship between pressure and
velocity, also herein, the diversity in velocity can be explained through the same
relationship concluding the change in velocity magnitude along the upper and lower edges
are strongly related to the corresponding pressure change. Figure 7.a highlights the concept
illustrating that the velocity increases on the pressure surface while the velocity on the
suction surface decreases up to the x position of 0.1 so that the largest discrepancy appears
in the upstream gradually shrinks along the leading edge of the symmetric airfoil down to
the x position of 0.2. The velocity magnitude at that point was captured as 26 m/s.
Downforce was, as postulated, generated by the difference in pressure, hence, indirectly by
the difference in velocity i.e. the lower velocity along the pressure side and the higher
velocity along the suction side there exist, the higher downforce can be generated.
Therefore, the downforce was evaluated at its highest value at the most front of the leading
 11

edge and decreases down to zero and since the difference in velocity increases along the
chord length as depicted on Figure 7.a, the downforce also increases upto the trailing edge.

a-) b-)
Figure 7. Distribution of velocity magnitude versus position along the lower and upper surfaces of a-) NACA
0015 airfoil and b-) NACA 2414 airfoil at AOA=4 degree

The feature of NACA0015 is more curved on the leading edge in comparison to

that of NACA2414; therefore the velocity of the airflow over NACA0015 is higher than the
velocity of the airflow over NACA2414, especially on the leading edge. However, the
discrepancy between the pressure and the suction surfaces of the cambered airfoil is larger
in comparison to that of the symmetric airfoil. As it can be concluded that the downforce
generated along the chord length of NACA2414 is higher, in most of the region, with
comparing to the downforce magnitude along the NACA0015. For instance, the inverse lift
(downforce) coefficient was assessed on the points in the vicinity of the trailing edges of
both airfoils. The data regarding downforce and drag coefficients are on Table 1. Since the
discrepancy became larger due to the camber; as a result, the cambered airfoils contribute
better to the stability due to the generation of the higher downforce.

NACA 0015 AOA=4° NACA 2414 AOA=4°

Speed
CL CD CL CD
[m/s]
30 1.586 0.403 1.67 0.598
Table 1. The data regarding drag and inverse lift coefficient of both airfoils with an angle of attack of
4° at the speed of 30 m/s

The camber also causes, as can be obviously seen on the table, higher drag
coefficient which is not the desire of a designer since it results in lower performance due to
the resistance against the airflow in road conditions. The higher drag may also be a cause of
more gasoline consumption. This increase in drag coefficient should also be taken into
account with the sustainability concerns in design concept and one might consider the
optimization with respect to drag and inverse lift coefficients.
 12

AOA=0° AOA=4°

AOA=8° AOA=12°

AOA=16°

Figure 8. The distribution of the velocity magnitude of the airflow at different angle of attacks

Figure 9. The distribution of the pressure coefficient along the airfoil NACA2414 at different angle of attacks
 13

As discussed earlier, the distribution of the velocity magnitude which can be

seen on the Figure 8 is symmetric on the both upper and lower edges of NACA0015 at angle
of attack of zero so that there exists no downforce along the chord length. At the most front
point of the leading edge in this zero AOA and symmetric airfoil case, the stagnation
pressure point where the velocity magnitude is zero can be obviously realized.
The symmetry in the distribution of the velocity magnitude disappeared when
the AOA was switched to 4 degree. The stagnation point on the leading edge slid upright as
the wake thickness in the downstream led by the trailing edge was enlarged. The
improvement of the wake thickness is evidence that the separation which is still nearby the
trailing edge occurred earlier i.e. turbulence flow was generated earlier as compared to the
airfoil in the lower angle of attack. The velocity magnitudes are in very low levels around
the stagnation point on the leading edge and in the wake generated in the downstream
domain. Beside that, the velocity values remained at the highest level, as expected, about the
most curved region of the lower edge. The angle of attacks of the higher values than 8
degree resulted in the distribution of velocity beyond the desires such as lift force instead of
downforce. The AOA up to 8 degree caused the generation of downforce since the velocity
distribution over the upper and lower edges differed from one side to another. The
separation started in the closer region to the leading edge as the airfoil was set at over 8
degree and the airflow over the lower edge kept decelerating in the bigger zone due to the
growth of the wake in terms of length and thickness. This phenomenon arisen by increase of
AOA over 8 degree may result in lift force instead of downforce as can be observed in Figure
7. The same incidence that was called stall condition by Kieffer et-al in Ref. 10 might be
observed on the demonstration of the cambered airfoil set at the corresponding angles of
attack. The same conclusion could be drawn regarding the cambered airfoil since
fundamentals of the incidences occurring in both cases should be the same eventhough the
location or the sizes of the wake and velocity distribution might be slightly different.

6. Conclusions:

The symmetric airfoil NACA0015 and NACA2414 are simulated at different

speeds and at different angles of attack through the commercial airflow simulation software
Fluent. The velocity and pressure distributions were obtained and plugged in the plots
shown comparing the differences regarding the cambered feature and the symmetric
feature of two types of the airfoils. Consequently, first of all, the downforce generated along
the chord length of NACA2414 was higher, in most of the region, with comparing to the
downforce magnitude along the NACA0015. Second of all, since the discrepancy became
larger due to the camber the cambered airfoils contribute better to the stability of the
overall car due to the generation of the higher downforce which will affect the rear axis of
the car and make the cornering and speeding under better control.
The incidence of stall condition was explained via the contour plots that depict
velocity magnitude distribution. This phenomenon arisen from the increase of the angles of
attack of 8 and over 8 degree may result in lift force instead of downforce. The same
conclusion could be drawn regarding the cambered airfoil since fundamentals of the
incidences occurring in both cases should be the same. The camber also causes higher drag
coefficient which is not the desire of a designer since it results in lower performance due to
the resistance against the airflow in road conditions. The higher drag may also result in
more gasoline consumption. This increase in drag coefficient should also be taken into
account with the sustainability concerns in design concept and one might consider the
optimization with respect to drag and inverse lift coefficients as a future work.
 14

References

[1] Mahon, S., Ph.D, & Zhang, X., Prof. (2006). Computational Analysis of An Inverted Double-
Element Airfoil in Ground Effect. Journals of Fuids Engineering, 128, 1172-1180.
[2] Mahon, S., Ph.D, & Zhang, X., Prof. (2005). Computational Analysis of Pressure and Wake
Characteristics of An Aerofoil in Ground Effect. Journals of Fuids Engineering, 127,
290-298.
[3] Kieffer, W., Moujaes, S., & Armbya, N. (2006). CFD study of Section Characteristics of
Formula Mazda Race Car Wings. Mathematical And Computer Modelling, 43, 1275-
1287.
[4] Ranzenbach, R., & Barlow, J. (1996). Cambered Airfoil in Ground Effect- An Experimental
and Computational Study. SAE Paper, 960909, 269-276.
[5] Ranzenbach, R., & Barlow, J. B. (1995). Two Dimensional Airfoil in Ground Effect, An
Experimental and Computational Study. SAE Technical Paper Series, 942509, 241-
249.
[6] Turbulence and It's Modeling. (1995). In H. K. Versteeg, & W. Malalasekera. [7] H.K.
Versteeg and W. Malalasekera “An Introduction to Computational Fluid Dynamics:
The Finite Volume Method (pp. 64-78.). New York: Longman Scientific and Technical.
[7] Turbulence Energy Equation Models. (1994). In D. C. Wilcox. Turbulence Modeling for
CFD (pp. 74-91.). California: DCW Industries, Inc.
[8] Modification of Equations of Fluid Motion. (2000). In K. A. Hoffman, & S. T. Chiang.
Computational Fluid Dynamics Volume 3 Fourth Edition (pp. 27-57.). Wichit, Kansas
USA: Engineering Education System.
[9] Vu, P., & Rubini, P. A. (Eds.). (2006). CFD study of wing tip vortices for inverted airfoils in
ground effect. Cranfield University: M.S Thesis.
[10] Munson, B. R., Young, D. F., & Okiishi, T. H. (2005). Fundamentals of Fluid Mechanics
Edition 5. New York: John Wiley and Sons, Inc.
[11] Vogt, J. W., Barber, T. J., & Leonardi, E. (2007). Flow Field Phenomena about Lift and
Downforce Generating Cambered Aerofoils in Ground Effect. [Editorial]. 16th
Australiasian Fluid Mechanics Conference , 2-7 December, 328-335.
[12] Solution of The Navier Stokes Equations. (2002). In J. H. Ferziger, & M. Peric.
Computational Methods For Fluid Dynamics Third Rev. Edition (pp. 157-204.). New
York: Springer.

View publication stats