Aerodynamics of Large-Scale Vortex-Generator in Ground-Effect.

Joseph Katz and Frederic Morey**, SDSU, San Diego, CA

∗

Abstract

The aerodynamic performance of several vortex generators (VG) of the type used on the
lower surface of race-cars was tested in a low-speed wind tunnel. In this particular
application, the vortices emanating from the VGs create a suction force between the
vehicle and the ground, thereby improving tire adhesion and the vehicle’s
cornering/traction performance. Since the size of these devices is much larger than the
local boundary layer thickness, they are termed ‘Large-Scale’ in this study. Results of the
wind tunnel tests indicate that the aerodynamic adhesion forces increase with reduced
ground clearance while the corresponding drag increase is much smaller. The parameters
investigated in this study are the VG length, shape, and the effect of incidence angle.
Amongst the various shapes tested, the traditional rectangular VG created the largest
forces while the simple triangular design was the most efficient in terms of the
incremental lift to drag ratio.

Introduction

Vortex generators (VG), labeled here as ‘small-scale VGs’, are widely used for boundary
layer transition and flow separation control. Some of the most common applications are
on the suction side of aircraft wings [1], or in the diffuser section of wind tunnels where
(due to their presence) a reduction in drag and surface resistance is observed. Another
form of vortex generating devices are the leading-edge strakes used for lift augmentation
on high speed airplanes. These highly swept surfaces are effective at the higher angles of
attack, when a strong vortex is formed along the leading edge. The suction induced by the
swirling vortex not only creates lift on the strake but significantly augments the lift of the
main wing behind it. A survey depicting the shapes of several, frequently used leading
edge devices with focus on unsteady flow effects, is presented in Ref [2]. A quite
different utilization of vortex-flow was investigated by Rossow [3] where a trapped
vortex was used to augment the lift of airfoils. The basic idea is that if a strong vortex can
be stabilized above an airfoil, its circulation and resulting lift could be considerably
augmented. The main hurdle for a practical application of this concept remains the
stabilization of the (trapped) vortex above the airfoil. In reference [4] several fences,
placed normal to the airfoil’s upper surface, were used but a practical utilization wasn’t
demonstrated. Further efforts to stabilize the vortex above the wing were experimented
with by Riddle et al. [5]. In addition to the vertical fences they used suction (through the
end plates of a two dimensional wing) and showed that the vortex can be stabilized above
the wing. In spite of this successful experiment, the practical use of this principle remains
unclear. The above mentioned vortex stabilization problem can be resolved by high

∗
Dept. of Aerospace Engineering, Professor and Department Chair
**
Student, visiting from institut français de mécanique avancée (IFMA)
leading-edge sweep, as discussed in reference [6]; but again, no large benefits were
demonstrated.

The combination of the traditional boundary layer control VGs and the lift augmentation
oriented trapped vortex concepts led to a third application, widely used in the race-car
industry. These devices rely on creating a strong trailing vortex behind the VG, capable
of altering the local flow directions around cooling inlets, wheels, and lifting surfaces.
Another, less publicized application is aimed at increasing the vehicle’s downforce ([7]
and [8]) in order to improve high-speed cornering, braking and acceleration. The present
study is focused on the aerodynamics of this latter category, and the primary function of
these VGs may be clarified by the diagram in Fig. 1. Here, a flat plate is immersed in the
free-stream with a velocity of V . A ‘small-scale’ VG is shown at the left hand side and
∞

its height is on the order of the local boundary layer thickness.

Figure 1 Schematic description of ‘Large-‘ and ‘Small-Scale’ vortex generators.

If placed ahead of the transition point, it can reduce the likelihood of laminar separation
and will force transition from a laminar to a turbulent boundary layer. In some cases such
VGs are placed in the turbulent region in order to energize the boundary layer and delay
flow separation. On the other hand, a ‘large-scale’ VG, shown at the right hand side of
Fig. 1, is much larger and its function is to interact with the outer flow and not with the
much thinner boundary layer. Race-car designers occasionally incorporate similar VGs
into the lower surface of the car in an effort to create low pressure there. A typical
application of a single VG (per side) is shown in Fig. 2 where it is attached to the lower
side of the vehicle. In most cases, the vehicle’s lower surface is almost flat and more than
one (per side) VGs are used. Because of the relative incidence (yaw) of this plate (e.g.,
the VG) a strong tip vortex is formed, extending along the under-side of the car. The
suction force created in the vicinity of the VG and by its wake behind the VG can
significantly enhance the downforce of the vehicle (and downforce improves
performance).

Figure 2 Typical application of large scale VGs on the lower surface of an open-wheel
race car.

Previous studies of such VGs ([7] and [8]) focused on rectangular VGs (as shown in Fig.
2) and their angle of attack and spacing effects. Note that the heights of these VGs cannot
change much because of the available ground clearance under the vehicle. The present
study, therefore, will investigate additional geometrical aspects of these VGs, such as
length and shape.
Experimental Apparatus

Figure 3 The flat plate model with the vortex generators mounted upside down at the
center of the test section. Ground clearance variations were simulated by moving the thin
ground plane down, towards the stationary model. Dimensions are in inches (meters).

The experimental setup used in the present study is similar to the one used in references
[7] and [8], and is shown schematically in Fig. 3. A closed circuit wind tunnel with a 4 ft
wide and 3 ft high (1.22 x 0.91m) test-section was used and maximum free-stream speed
could be as high as 170 Mph (76 m/sec). The lower surface of the car (sometimes called
the underwing) was represented by a flat rectangular plate, onto which the VGs were
mounted. This plate was set at zero incidence at the center of the test section and was
mounted onto the 6 component balance (located under the test section) by three struts, as
shown in the figure. Ground clearance variation was obtained by moving a larger
rectangular thin plate up or down relative to the model (see ‘Ground Plane in Fig. 3).
The advantage of this inverted setup (e.g. the ground is above the model) was that the
flat-plate model was mounted directly to the balance and only the non-metric ground
plane was moving.
Figure 4 The wind tunnel model consists of a horizontally placed flat-plate with four
VGs. Dimensions are in inches (meters) and the spacing d was kept constant at d = 2
inches (0.05 m) throughout the whole test. Note the schematic description of the vortices
emanating from the left side.

Since both the model and ground plane were made of thin flat plates, wind tunnel
blockage was small (less than 1%) and the effect of the ground plane installation (on the
blockage ratio) was negligible too. Two rectangular vortex generators per side were used
(see Fig. 4), following Rossow’s [3], [4] observation that at least two fences are required
to stabilize the vortex. The spanwise and chordwise position and the yaw angle of the
VGs was adjustable, but their vertical orientation was fixed (e.g., at 90 deg to the flat
plate). For the data presented here the separation distance between the leading edges of
the inner VGs was kept constant (at a total of 4 inches = 0.10m ) and each VG was yawed
about this point (positive β defined in the outward direction, see Fig.4). Other parameters
such as the angles β and the distance d between adjacent VGs were selected based on the
findings reported in reference [7] (e.g., β = 20, 30 deg, and d = 2 inches = 0.05 m).
Therefore, the leading edges of all VGs tested here remained at the same point, as shown
in Fig. 4. Also, the geometrical parameters of this study were dictated by the Indycar
geometry shown in Fig. 2. Thus, the flat-plate size, and the relative size, location, and
orientation of the VGs were all set by this open-wheel race-car application.

Additional dimensions of the model and test section are depicted in Fig. 3 and Fig. 4. The
free-stream speed was set at V = 120 mph (53.6 m/sec) and turbulence levels were below
∞

2%. Although previous studies (with wings [9]) reported noticeable effects when free-
stream turbulence levels were about 9%, those results do not directly relate to the present
experimental setup. Also, because of the sharp edges of the VGs it is assumed that
Reynolds number and free-stream turbulence effects are less pronounced and results of
this vortex flow experiment are scaleable. Based on the flat plate length the test Reynolds
number was 2.7 x 106. Boundary layer thickness on the flat plate and near the VGs was
about 2 – 4 % of the VG height, which is much smaller than the schematic description in
Fig. 1. It was assumed that the boundary layer is laminar ahead of the VGs, and due to
the vortex flow transitions to turbulent behind it Accuracy of the six-component balance
was about ±0.004 for CL, ±0.002 for CD. Note that in this setup positive lift is upward and
the ground clearance was measured between the upper tip of the VG and the ground
plane above it (Fig. 3).

Results

Earlier aerodynamic studies of such ‘large-scale’ VGs [7][8], focused mainly on the
effect of the spacing, d, between adjacent VGs and on their incidence, β. Based on the
‘best results’ from reference [7] and [8] the spacing d in the present study was set at 2
inches (0.05m), and the side slip of the inner VGs was set at β = 20 deg while the outer
VG was set either to β = 20 or β = 30 deg. As noted, the parameters investigated in the
present study are the VG shape and length.

Since the experimental apparatus of reference [7] and [8] was slightly modified (from the
structural point of view), a comparison with the earlier data is presented first. The results
for this comparison, in terms of the measured lift and drag coefficients are presented in
Fig. 5, and Fig. 6, respectively. The magnitudes of the nondimensional coefficients, are
small since the plate’s planview area was used for the data reduction (e.g. S = 480 in2 =
0.31m2). In order to visualize the magnitude of data uncertainty, error bars were added to
these two figures (only). The VGs were set at β = 20 deg and the comparison also
included the results for the flat plate model without any VGs on it (from [8]). This was
necessary because of the difference in the definition of ‘tares’ between the two
experiments.

0.18

Flat Plate, Ref. 8

0.14 F Plate + VGs, Ref. 7
CL Present Data
0.1

0.06

0.02

-0.02
0 0.05
h/c 0.1 0.15 0.2
Figure 5 Variation of the lift coefficient versus ground clearance h/c and comparison
with previously published data (rectangular VGs, at β = 20 deg). Note that uncertainty in
CL is represented by the vertical error bars!

It is expected that a flat plate at zero incidence will not produce lift. However the data
from [8] indicates that the model may have had a small negative incidence which created
negative lift that increased with reduced ground clearance (Fig. 5). This small negative
incidence was corrected in the present test and the lift (with the VGs) is therefore slightly
higher than the lift reported in [7]. (Note that the data reported in reference [7] and [8]
were taken at the same time). The increase in lift when ground clearance is reduced is
attributed to the vortex being closer to the flat plate and perhaps also being stronger due
to faster air-speeds around its edge (with reduced ground clearance). Of course, at close
to zero ground clearance, no flow passes above the VG and therefore no vortex is created,
so lift must be reduced. This extreme condition wasn’t measured in order not to damage
the model (by mistakenly running the ground plane too close to the VGs). However this
data clearly show that the lift increases, even for very small ground clearance values.

0.045

0.03375

C
D F Plate, Ref. 8
0.0225 F Plate + VGs, Ref. 7
Present data

0.01125

0
0 0.05 0.1 0.15 0.2
h/c
Figure 6 Variation of the drag coefficient versus ground clearance h/c and comparison
with previously published data (rectangular VGs, at β = 20 deg). Note that uncertainty in
CD is represented by the vertical error bars!

The drag coefficient data in Fig 6 show the difference between the methods of accounting
for ‘tares’. The earlier drag data includes the drag of the flat plate and its supporting
brackets, while the present data show the added effect of the VGs only (so in the later
data the ‘tares’ included the flat plate loads as well). Consequently, the present drag data
should be compared with the difference between the previous data minus the drag of the
flat plate. Based on this data, the drag of the flat plate, including the installation
hardware, was about CD ~ 0.035, and the increment due to the VGs was on the order of
CD ~ 0.01. This is close to the order of magnitude drag estimates for a flat plate given by
Winters [10] (e.g., the drag of the individual VG was calculated assuming an isolated flat
plate at an angle of attack β). Also, the error bars shown in Fig. 6 depict the maximum
errors in the drag force, which are much larger than the error bars in Fig. 5, mainly due to
the smaller magnitude of the drag force.

Highly swept triangular surfaces such as delta wings [11] can also function as VGs in
ground proximity. One advantage they may have is when ground clearance momentarily
reaches zero (due to suspension oscillations), then a triangular VG will still produce a
sizeable trailing vortex. The first set of experiments, therefore, focused on comparing the
performance of rectangular versus triangular VGs. At the same time the effect of VGs
length was also studied by simply reducing their length by one half. The geometry of this
set of VGs is depicted in Fig. 7.

Figure 7 Dimensions of the rectangular/triangular long/short VGs. Dimensions in inches

(meters).
A comparison between the rectangular and triangular VGs (having the same length and
height) in terms of lift and drag coefficients is presented in Fig. 8 and Fig. 9. These two
figures present the data for the long VGs (e.g., 6 inch long = 0.15m) and the data for the
shorter VGs is presented later. At the larger values of ground clearance the effect is quite
small, however with reduced h/c the rectangular design creates consistently more lift,
compared to the triangular ones. When the angle of the outer VGs was increased to β =
30 deg, the lift increased for both rectangular and triangular VGs, as seen on Fig. 8. At
the lower h/c values, the rectangular VGs show signs of a halt in the lift increase,
possibly due to blocking the flow between the VG and the ground (or vortex burst, as
suggested in reference [7]). On the other hand, the triangular design does not have this
deficiency, and its lift increases, even for very low h/c values. It was also suggested that
the vortices emanating from the VGs create a ‘local blockage’ for the flow between the
two large flat-plates. This blockage will increase the local velocity (and CL) as ground
clearance is reduced. However, data for larger VG angles (e.g. β = 30/40 deg), reported in
Ref [7], showed an earlier lift loss when ground clearance was reduced. This observation
and the fact that the lift of the triangular VGs is not half of the rectangular VGs, suggests
that the local blockage effect is not dominant.

The frontal area of the triangular VG is one half of the rectangular VG, and therefore less
drag is expected. This is verified in Fig. 9. Again, increasing the incidence of the outer
VGs to β = 30 increases the drag for both shapes, but in the case of the rectangular VGs
the increase is the largest. As expected, the drag data follows the trends of the lift data
presented in Fig. 8. For example, the drag of the rectangular VGs reaches a maximum
before approaching minimum ground clearance and then shows a clear drop (for both β =
20/20 and β = 20/30).

0.18

CL(20/20)rec
0.135 CL(20/30)rec
CL(20/20)tri
C CL(20/30)tri
L
0.09

0.045

0
0 0.05 0.1 0.15 0.2
h/c
Figure 8 Variation of the lift coefficient versus ground clearance h/c for rectangular and
triangular VGs, and the effect of larger incidence, β (β = 20/20 means all VGs are at 20
deg, and β = 20/30 stands for outer VGs at β = 30 deg).

The lift and drag coefficient variation versus ground clearance for the shorter VGs (both
rectangular and triangular) is presented in Fig. 10, and Fig. 11. Note that the ‘short’ VGs
had one half the length of the ‘long’ or baseline VGs. Basically, the same trends remain
as seen with the long VGs, but as expected, with less lift and less drag. Although the
surface area of the VGs was reduced by 50%, the reduction in both lift and drag was far
less. This demonstrates again that the trailing vortex effect is not local and its induced
suction affects areas behind the VG. The lift coefficient data in Fig. 10 indicates that the
triangular VG generates less lift than the rectangular one, while increasing the incidence
of the outer VGs to β = 30 deg was as effective as for the longer VGs. Also note that
some of the drag coefficient data falls close to the previously mentioned uncertainty of
±0.002 for CD.

CD(20/20)rec
0.025 CD(20/30)rec
CD(20/20)tri
CD(20/30)tri
0.02

0.015
C
D
0.01

0.005

0
0 0.05 0.1 0.15 0.2
h/c
Figure 9 Variation of the drag coefficient versus ground clearance h/c for rectangular and
triangular VGs, and the effect of larger incidence, β (β = 20/20 means all VGs are at 20
deg, and β = 20/30 stands for outer VGs at β = 30 deg).
 0.18

CL(20/20)rec
0.135 CL(20/30)rec
CL(20/20)tri
C CL(20/30)tri
L
0.09

0.045

0
0 0.05 0.1 0.15 0.2
h/c
Figure 10 Variation of the lift coefficient versus ground clearance h/c for SHORT
rectangular and triangular VGs, and the effect of larger incidence, β (β = 20/20 means all
VGs are at 20 deg, and β = 20/30 stands for outer VGs at β = 30 deg).

0.025
CD(20/20)rec
0.02 CD(20/30)rec
C CD(20/20)tri
D
CD(20/30)tri
0.015

0.01

0.005

0
0 0.05 0.1 0.15 0.2
h/c
Figure 11 Variation of the drag coefficient versus ground clearance h/c for SHORT
rectangular and triangular VGs, and the effect of larger incidence, β (β = 20/20 means all
VGs are at 20 deg, and β = 20/30 stands for outer VGs at β = 30 deg).
 combined.dat
0.18

CL(20/20)rec
0.135
CL(20/20)tri
C CL(20/20)rec,short
L
CL(20/20)tri,short
0.09

0.045

0
0 0.05 0.1 0.15 0.2
h/c
Figure 12 Lift coefficient comparison for short/long and rectangular/triangular VGs, at β
= 20/20.

A direct comparison between the long and the short VGs is presented in Fig. 12.
Basically all the data presented in Figs 8 - 11 could have been crossplotted, but for
brevity only the lift with β = 20/20 deg is presented (but results for the β = 20/30 are
basically the same). Clearly, the rectangular VGs generate more lift than the triangular
ones, and the lift of the longer VG is also higher. A close examination of the lift to drag
ratio (not plotted here) shows a small advantage for the triangular design.
Figure 13 Additional VG shapes evaluated during this test. Note that only the ‘short’
versions were tested (e.g. Length = 3 inches and height is 1 inch. Or 0.08 x 0.026 m).

0.18

CL(20/20)ogive
0.135 CL(20/20)parabolic
CL(20/20)gothic
C CL(20/20)rectangle
L
CL(20/20)triangle
0.09

0.045

0
0 0.05 0.1 0.15 0.2
h/c
Figure 14 Variation of the lift coefficient versus ground clearance h/c for several short
VGs, (β = 20/20 means all VGs are at 20 deg).
 0.025

CD(20/20)ogive
0.02
CD(20/20)parabolic
C
D CD(20/20)gothic
0.015 CD(20/20)rectangle
CD(20/20)triangle

0.01

0.005

0
0 0.05 0.1 0.15 0.2
h/c
Figure 15 Variation of the drag coefficient versus ground clearance h/c for several short
VGs, (β = 20/20 means all VGs are at 20 deg).

0.18

CL(20/30)ogive
0.135 CL(20/30)parabolic
CL(20/30)gothic
C
L CL(20/30)rectangle
0.09 CL(20/30)triangle

0.045

0
0 0.05 0.1 0.15 0.2
h/c
Figure 16 Variation of the lift coefficient versus ground clearance h/c for several short
VGs, (β = 20/30 stands for inner VGs at (β = 20 deg, and outer VGs at β = 30 deg).

A Vortex Generator may be viewed as one half of a small aspect-ratio wing at an angle of
attack. With this analogy in mind, many other shapes may be considered, and the most
common are summarized in Fig. 13. In terms of area, the gothic shape is larger than the
triangle and smaller than the rectangular design. This shape was created by simply
blending a 40 deg initial leading-edge sweep into an otherwise rectangular VG. The ogive
shape at the other hand has the same area as the triangular VG but the leading edge was
smoothed by a sinusoidal curve. The parabolic shape had the smallest area, and started at
a much larger sweep at the apex (compared to the triangular design). Thus, from the
visual aspect, the area is the main parameter. However, from the aerodynamic point of
view, vortex stability (delay of vortex burst in this case) is a more important parameter
influencing the lift coefficient of the VG.

0.025
CD(20/30)ogive
CD(20/30)parabolic
0.02
C CD(20/30)gothic
D
CD(20/30)rectangle
0.015 CD(20/30)triangle

0.01

0.005

0
0 0.05 0.1 0.15 0.2
h/c
Figure 17 Variation of the drag coefficient versus ground clearance h/c for several short
VGs, (β = 20/30 stands for inner VGs at (β = 20 deg, and outer VGs at β = 30 deg).

A comparison of the lift coefficient versus ground clearance, generated by the five short
VGs is shown in Fig 14. In this case the incidence is moderate (all VGs at β = 20/20 deg )
so vortex breakdown is not immediately behind the VG trailing edge. This way the
differences between the five designs is more representative. The data in Fig 14 clearly
shows that the lift is almost directly related to the VG area. Clearly the rectangular design
has the highest lift while the parabolic shape has the lowest. In the case of the triangular
and ogive shapes (having the same area), the ogive VG demonstrates a slight advantage,
and indeed most small aspect ratio airplane wings (e.g., Concorde) have similar planform
shapes. The drag data in Fig. 15 follows similar trends, but in this case the triangular
design has lowest drag.

When increasing the incidence of the outer VGs to β = 30 deg, as shown in Fig. 16, the
lift data follows similar trends shown earlier in Fig 14. The lift increase of the rectangular
VGs however seems to slow down at the lower ground clearances, possibly due to the
outer VGs vortex breakdown (near the VG trailing edge, as discussed in reference [7]).
This loss of lift results in a dip in the drag as shown by Fig. 17. Otherwise the drag data
repeats the conclusions drawn from Fig. 15, clearly showing the drag advantage of the
triangular design. This brief study of the VGs shape indicates that the simple geometries
of the triangle or rectangle are the most effective for this type of application.

16

14 L/D(20/20rec-long)
L/D(20/20tri-long)
12
L/D(20/20rec-short)
10 L/D(20/20tri-short)
L/D
8

2
0 0.05 0.1 0.15 0.2
h/c
Figure 18 Variation of incremental lift to drag ratio versus ground clearance h/c for the
four most efficient VGs, (β = 20/20 deg).

The data presented so far indicates that the β = 20/20 deg configuration is more efficient
than the β = 20/30 deg case. But it is not yet clear which configuration has the best L/D.
Also for the sake of brevity, the L/D information wasn’t plotted for all the cases
presented here. Therefore the incremental lift to drag ratio of the four most efficient VG
configurations is summarized in Fig. 18 (basically the long versus short
triangular/rectangular shapes). The results for all four VG shapes are very close but the
triangular VGs reach a slightly higher L/D. Note that the L/D curves increase almost like
the previously presented CL versus h/c data, demonstrating the effectiveness of this
device for race-car applications. Also, as noted in the discussion on Fig. 6, the drag of the
large flat plate was not included in the present data. Therefore, the L/D numbers in Fig 18
are much higher than in reference. [7]. Furthermore, since the flat plate’s drag (which
was included in the tares) was almost three times larger than the drag of the VGs alone,
the L/D data of Fig. 18 must be viewed in terms of ‘general trend’ only.

Concluding remarks

Results of this study demonstrate the potential of such VGs to generate sizeable levels of
downforce in race-car applications. For most shapes of the VGs tested, the lift (or
downforce for a race car) increases with reduced ground clearance. Near the zero ground
clearance level, the triangular VGs were found to be less likely to lose lift, an important
feature for race-car applications. The data presented here consists of the integral lift and
drag. However, many aspects of the flow, such as the vortex rollup or break down behind
the VGs and the mechanism of skin friction on the flat plate remain unclear. Therefore,
further studies must focus on additional flow visualizations, measurement of pressure
distributions, and numerical simulations, to better explain the data presented here.

References

1. Bragg M. B., “Experimental Study of Airfoil Performance with Vortex Generators,” J.

Aircraft, Vol 24, No. 5, May 1987, pp. 305 – 309.

2. Katz, J., “Wing/Vortex Interactions and Wing Rock Review ,” Progress in Aerospace
Sciences, Vol. 35, No. 7, 1999, pp. 727-750

3. Rossow, V. J., "Lift Enhancement by an Externally Trapped Vortex ," J. of Aircraft,

Vol. 15, No. 9, Sept. 1978, pp. 618-625.

4. Rossow, V. J., "Aerodynamics of Airfoils with Vortex Trapped by Two Spanwise

Fences," J. of Aircraft, Vol. 31, No. 1, Jan-Feb.. 1994, pp. 146-153.

5. Riddle, T. W., Wadcock A. J., Tso J., and Cummins R. M.,"An Experimental Analysis
of Vortex Trapping Techniques," Journal of Fluids Engineering , Vol. 121, No. 3, 1999,
pp. 555-559.

6. Buchholz, M. D., and Tso, J., "Lift Augmentation on delta Wing with Leading-Edge
fences and Gurney Flap," J. of Aircraft, Vol. 37, No. 6, Nov. Dec., 2000, pp. 1050-1057.

7. Garcia D., and Katz, J. "Trapped Vortex in Ground Effect," AIAA 2002-3307,
presented at the 32nd AIAA Fluid Dynamics Conference, St. Louis, Missouri, June 26
2002. Also AIAA J. Vol. 41, No. 4, April 2003, pp. 674-678.

8. Katz, J and Garcia D., “Aerodynamic Effects of Indy Car Components,” SAE 2002-01-
3311, SAE Journal of Passenger Cars: Mechanical Systems, 2002, pp. 2322-2330.

9. Hoffmann, J.A., Effects of Freestream Turbulence on the Performance Characteristics

of an Airfoil, AIAA Journal, Vol.29 No.9, 1991, pp. 1353-1354.

10 Winter, H., "Flow Phenomena on Plates and Airfoils

of Short Span," NACA TM 798, 1937

11. Katz J. and Levin D., "Measurement of Ground Effect for Delta Wings", J. of
Aircraft, Vol. 21, No. 6, 1984, pp. 442-443.
Nomenclature

a = VG height (1 inch)
c = plate chord (30 inches)
CD = drag coefficient = D/(0.5 ρ V 2 S)
∞

CL = lift coefficient = L/(0.5 ρ V 2 S)

∞

d = lateral spacing between VGs (2 inches)

D = drag
h = ground clearance
l = VG length (6 inches)
L = lift
L/D = Lift/Drag ratio
S = Ref. Area (30 x 16 = 480 in2)
VG = vortex generator
V = free stream speed
∞

w = plate width (16 inches)

β = VG yaw angle