Optimization of Landing Gear Fairings

Final Report

Group7
Daniel Ablog, Michael Fuget, Seunghyun Ko, and Keisuke Tsujita
AA 322 Aerospace Laboratory II
WILLIAM E. BOEING DEPARTMENT OF AERONAUTICS & ASTRONAUTICS
University of Washington
Seattle, Washington, 98195-2400
2014-05-05

Wind tunnel and flow visualization experiments were conducted to deter-

mine the aerodynamic effects of landing gear fairings of varied length at
different angles of attack in both the University of Washington 3x3 Wind
Tunnel and water tunnel. This experiment utilized a 1:3 scale model of
the actual fairing and ratios of length 0.7d,1.0d, 1.14d, 1.5d, and 2d. Water
tunnel testing of the fairings provided a means to flow visualization around
each model. Dye was injected in front of the models to see the flow around
the fairings. For most models, the flow separated before reaching the fair-
ing and therefore, the fairing had no aerodynamic effect. However, for the
longer fairings, flow remained attached to the fairings for a longer period
of time before separating. In concordance, wind tunnel tests showed that
longer fairings generally had less drag than those of the covered fairing and
individual wheel. These results indicate that attaching longer fairings to
the wheel of a landing gear substantially reduces drag in comparison to that
of shorter fairings.

Nomenclature
A Axial Force (lb)
c Chord length (in)
C Coefficients
d Diameter of wheel (in)
D Drag(lb)
l Length of the fairings (in)
L Lift(lb)
M Moment(in-lb)
N Normal Force(lb)
q Dynamic Pressure(lb/ft2 )

1 of 17

American Institute of Aeronautics and Astronautics

Re Reynolds Number
S Surface area of test samples (in2 )
t Thickness (in)
α Angle of attack (°)
ρ Density of air (slug/ft2 )
µ absolute viscosity coefficient (slug/(ft*s))
Subscript
A Axial
D Drag
L Lift
M Moment
N Normal

I. Introduction
Small airplanes, such as Cessna 150, cover their non-retractable landing gear wheels with
the streamlined fairings which is also known as wheel pants while large airplanes have a
retractable landing gear to decrease the drag associated with the landing gear during their
flight.

Fig. 1: Landing gear fairing of Cessna 150

2 of 17

American Institute of Aeronautics and Astronautics

 Fig. 2: Fairing design that does not cover the wheel

Currently, the fairings of non-retractable landing gear cover the top half of the wheel as
shown in Fig. 1.1 Previously, the drag of several types of existing fairings were measured by
Herrnstein and Biermann.2

Fig. 3: The Smallest Drag Fairing Configuration

Should I have this part here or in the theory? From the experiment, the fairing shown in
Fig. 3 with modification 2 had the smallest drag among the six types of fairings tested. The
other configurations of fairing tested are shown in Appendix A. Herrnstein and Biermann2
concludes that the fairing which covered both sides of the wheel and had minimal frontal
area was most effective in reducing drag. Although the fairings reduce the drag, they are
the extra weight added to the airplane, which should be minimized. One way to reduce the
weight of a landing gear fairing is to utilize the new design of the fairing which does not

3 of 17

American Institute of Aeronautics and Astronautics

cover the wheel as shown in Fig. 2. The new design of the fairing does not increase the
frontal area and may reduce the weight while keeping the drag performance similar to that
of the existing fairing. The purpose of this project is to measure the aerodynamic forces
and visualize the flows around the fairings with five different lengths using the configuration
shown in Fig. 2, a fairing with the existing configuration of fairings as shown in Fig. 1, and a
wheel without a fairing. The results are compared to determine the differences in drag and
volume.

II. Theory
A. Assumptions and Simplifications
Several assumptions and simplifications were made for the test samples. First, the test
model can consist of only a wheel and fairing behind the wheel. Also, although a real wheel
is made of several parts and different materials, a circular cylinder, whose ratio of thickness
and diameter is the same as that of a real wheel, can model the wheel. Also, the test
models can be modeled by any materials as long as they are consistent for all models. These
simplifications are valid because the purpose of the project is only to investigate the effect of
the variation of length of the fairings. In addition, the assumption that the wheel modeled
by a cylinder does not rotate was made because wheels on a Cessna 150 do not rotate during
cruising conditions.

B. Physics
Title of the subsection must be changed Aerodynamic forces exerted by the flow of air are
due to the pressure distribution of the surface and shear stress on the surface.3

Fig. 4: Aerodynamic forces on the test sample

The sting mount in the 3x3 wind tunnel measures the normal and axial force shown in
Fig. 4. Drag at a certain angle of attack can be calculated from normal and axial forces as
follows.

D = N sin α + A cos α (1)

4 of 17

American Institute of Aeronautics and Astronautics

 The Reynolds number, which is strongly associated with drag, can be found by.3

ρV∞ c ρV∞ (d + l)
Re = = (2)
µ µ
Since drag depends on the dynamic pressure and projected area, drag coefficient coeffi-
cient are defined as
D D
CD = = (3)
q∞ S q∞ td
As explained in Section II-A, the test sample models the wheel by a circular cylinder,
whose pressure drag is higher than the friction drag.4 The pressure drag is produced by the
pressure difference between the front and back side of the cylinder. This difference is caused
by the wake formed behind the cylinder.5

Fig. 5: Flow over a blunt body and streamlined body

The drag characteristic of the blunt body, such as a circular cylinder, and the streamlined
body are shown in Fig. 5. The blunt body has large pressure drag caused by the wake formed
behind the bod due to the flow separation. The new model shown in Fig. 2 uses the wheel
as the front portion and itself as the back portion of the streamlined body. This may lead
to less pressure drag compared to the case in which only the wheel is employed.

5 of 17

American Institute of Aeronautics and Astronautics

C. Previous work related to the design and dimensions of the fairings

Fig. 6: Pressure distribution over the cylinder with splitter plate behind

Roshko6 states that when a cylinder is placed in front of a flat splitter plate which has a
length of 5 times the diameter of the cylinder, as shown in Fig. 6, the drag coefficient of the
cylinder is reduced from 1.1 to 0.7. Fig. 6 shows the pressure distribution behind the circular
cylinder with splitter plate and that without splitter plate. The splitter plate is effective in
reducing the pressure drop behind the circular cylinder. Also, Roshko states that by using a
splitter plate with a length of 1.14 times the diameter, the vortex shedding frequency varies
depending on the location of the plate whereas the drag did not decrease.6 These previous
work indicate that the length of the splitter plate is what affects the reduction in drag. The
length of the splitter plate was used as the reference length in the project as explained in
the Section III-A.

III. Designing and Fabrication of the Models

A. Designing of the test models and Determining the test conditions
The dimensions of the wheel were determined based on the actual wheel used. The
Cessna 150 employs a wheel whose diameter is 15 in and thickness is 6 in.7 Due to the size
of the wind tunnel, the wheel’s diameter was scaled down to be 5 in, and the thickness was
decided to be 2 in in order to match the ratio of diameter and thickness of the actual wheel.
The lengths of the fairings to be tested were chosen as 0.7d, 1.0d, 1.14d, 1.5d, and 2.0d.
Even though the previous work described in Section III shows that the splitter plate, whose
length is 5d, placed behind the circular cylinder is effective in reducing the drag,6 the largest
length was decided to be 2d. This decision was made because the fairing may touch the
ground during the takeoff or landing if the length is too long. Other dimensions, which are
kept constant among the model fairings, were also decided. The width of the fairing was
decided to be the same as the thickness of the wheel because the previous experiment by
Herrnstein and Biermann2 described in Section II shows that fairing with the smallest frontal
area had the lowest drag. The spacing between the fairing and the wheel was chosen to be
0.5 in, which corresponds to 1.5 in in the real scale. This spacing was chosen because certain
clearance between the fairing and the wheel is required in case of landing on muddy ground.
The height of the fairing was decided to be 80% of the diameter of the wheel to allow the
smooth transition of the flow from the wheel to the fairing.

6 of 17

American Institute of Aeronautics and Astronautics

 It is ideal to execute the experiments under the dynamically similar condition, represented
by the Reynolds number. First, the Reynolds number of the flow which the actual landing
gear experiences was calculated. The cruising speed of the Cessna 150 at altitude of 7000 ft
is 117 mph.8 The Reynolds number for its wheel at that speed and altitude was calculated
with Eq. 2 with the cruising speed and properties of air at 7000 ft listed in Table. 1. The

Table 1: Properties of Air at 7000 ft

Properties Values
9 Temperature(R°) 493.73
Pressure(lb/f t2 ) 1.6331 × 103
Density(slug/f t3 ) 1.9270 × 103

viscosity of air was found by interpolating the data in Table. 6 in Appendix B. The Reynolds
number was calculated to be 1.13× 106 with Eq. 2.
However, due to the limitation of the testing facilities, it was not possible to conduct the
experiment with the same Reynolds number. The air speed and water speed of the testing
facilities were limited as well as the maximum size of the models which can be used. The
3x3 wind tunnel can not be operated above 45 psf of dynamic pressure. Under the standard
atmosphere at sea level, the Reynolds number was calculated to be only 5.0 × 105 with
Eq. 2. The Reynolds number in water tunnel was also calculated to be 6.5×104 .
The dimensions of the test samples are shown in Fig. 7

7 of 17

American Institute of Aeronautics and Astronautics

 (a) Wheel with new model of fairing. l = 0.7d, 1.0d, 1.14d,
1.5d or 2.0d. Unit is inch

(b) Current design of the fairing. For simplicity, the fairing

and the wheel was designed as one body. Unit is inch

Fig. 7: The design of the new and current fairing test samples

Because the experiments are planned to be executed in the 3x3 wind tunnel and in the
water tunnel, a mount must be created for each tunnel. The CAD drawing for the sting
mount, which measures the aerodynamic forces and moments, was obtained to design the
extension mount for the wind tunnel. A CAD model of the extension mount for the sting
mount in the 3 x 3 wind tunnel was created as shown in Fig. 8.

8 of 17

American Institute of Aeronautics and Astronautics

 Fig. 8: Extension mount design for wind tunnel testing

In addition, the mount for the experiments in the water tunnel was designed as shown
in Fig. 9. These configurations were chosen to minimize the effect of the mount on the
experiments.

Fig. 9: Mount design for water tunnel testing

B. Fabrication of test samples

The materials purchased for fabrication are listed in Appendix B. A model wheel and five
fairings with different lengths were fabricated from wood. The test samples were cut with
the band saw in the machine shop in the Mechanical Engineering Department. The fabri-
cated test samples are shown in Fig. ??. Making a perfectly symmetrical test samples was
extremely difficult, and the thickness of the fairings became slightly smaller than what was
originally designed as they were sanded to be tapered. After sanding sealer was applied, the
models were waterproofed for the experiment in the water tunnel.tyotto The machine shop
in the Mechanical Engineering Department and Aeronautics and Astronautics Department
were used to fabricate the test samples. The 3×3 Wind Tunnel and Water Tunnel were used
to conduct the experiments.
The test samples used in the water tunnel and the wind tunnel are shown in Fig. ??.

IV. Experimental Apparatus

The setup of the experiment in the water tunnel and in the wind tunnel are included in
Fig. 11a and Fig. 11c, respectively. The mount shown in Fig. 9 was used to hold the test
samples and change angles of attack in the water tunnel.

9 of 17

American Institute of Aeronautics and Astronautics

 (a) Wheel (b) Fairing with l=0.7d

(c) Fairing with l=1.14d (d) Fairing with l=1.14d

(e) Fairing with l=1.5d (f) Fairing with l=2d

Fig. 10: Fabricated Test Samples

Table 2: Apparatus used in the water tunnel

Equipment
Water Tunnel
Dye Injection Tube
Lucks Color Blue
Lucks Color Green
Micropump
Gear Pump Controller
Water Tunnel Mount

Table 2 includes the apparatus used for the flow visualization in the water tunnel exper-
iment.

10 of 17

American Institute of Aeronautics and Astronautics

 (a) Experimental Setup for the Water (b) Dye Injection System for the Water
Tunnel Tunnel Experiment

(c) Experimental Setup for the Wind

Tunnel

Fig. 11: Experiment Setups

Table 3: Apparatus used in the 3×3 wind tunnel

Equipment Model

3x3 Wind Tunnel

Attachment Mount

Table 3 includes the apparatus used for the drag measurements in the 3×3 wind tunnel.

V. Procedure
A. Water Tunnel Test
The bare wheel and the fairings attached to the wheel were tested in the water tunnel
to visualize the flows around the samples. One of the injection port of the Micropump was
connected to the dye container with a dye injection tube. The other injection port was
connected with the dye injection tube, which was attached to the metal rod for the dye
injection in the water tunnel. The gear pump controller was connected to the Micropump to

11 of 17

American Institute of Aeronautics and Astronautics

control the amount of dye to be injected. Schematic of the dye injection system is shown in
Fig. 11b. The test samples were attached to the water tunnel mount. The following angles
of attack were tested: −10°,−7°, −2°, 0°, 10°, 15°, 20°, and 20°. While the experiment, the
dye injecting tube attached to the metal rod was moved by hand to see the flow over and
under the test samples. In order to analyze the flows over and under the test samples, green
dye and blue dye were injected from the top surface and bottom surface of the test samples,
respectively. However, there was not enough green dye for all runs thus blue dye was injected
from for both top and bottom of the flows. The frequency of the water tunnel was set to 50
Hz, which corresponds to the 50cm/sec. The conversion chart between the frequency and
the water speed is in Appendix.... Movies were taken to analyze the flows.

B. Wind Tunnel Test

The atmospheric pressure was recorded. The test samples were attached to sting mount
by the attachment mount as shown in Fig. 11c with aluminum tape to reinforce the attach-
ment. Angles of attack was changed from 0° to 20° with increment of 2°. For each run, the
dynamic pressure was gradually increased from zero to near 45 psf. LabView was used to
record dynamic pressure, temperature, axial forces, and normal forces.

VI. Discussion of Results

A. Measurement of drag

Fig. 12: Drag Coefficient measured at α = 0 for each test sample with various Reynolds
number

Fig. 12 is the plot of the drag coefficient measured at α = 0 with Reynolds number from
Re = 3.5 × 105 to 7.0 × 105 . Except in the lower Reynolds number range, all of the samples
with fairing had the lower coefficient of drag than that with the bare wheel. The current
model of the fairing and the test sample with l = 0.7 d, l = 1.0 d, and l = 1.14 d of fairing,

12 of 17

American Institute of Aeronautics and Astronautics

showed slight decrease as the Reynolds number increased. On the other hand, the sample
with l = 1.5 d, l = 2.0 d of fairing showed slight increase in drag coefficient. At zero angle of
attack, which is typical in cruise, the current model of fairing showed the best aerodynamic
performance over the range of Reynolds number tested.

Table 4: The Coefficient of drag and percent decrease compared to the bare wheel model
for low and high Re with weight ratio of the models at zero angle of attack

Low Re High Re
Weight Ratio(%)
Cd % Decrease Cd % Decrease
Bare Wheel 0.245 - 0.256 - 0
0.7 d 0.186 24.2 0.177 30.9 23.1
1.0 d 0.255 -3.83 0.226 12.0 35.8
1.14 d 0.308 -25.6 0.269 -4.76 42.4
1.5 d 0.165 32.9 0.193 24.9 56.7
2.0 d 0.194 21.0 0.196 23.6 73.4
Current Model 0.157 36.0 0.135 47.4 1

Table 4 includes the Cd and the percent decrease compared to the bare wheel for each
low and high Reynolds number with the weight ratio of the fairing compared to the current
model at zero angle of attack.
Among the new design of the fairings, at lower Reynolds number, the fairing with the
length of 1.50 d had the best aerodynamic performance as shown in Fig. ??. On the other
hand, the at higher Reynolds number, the fairing with the length of 0.7 d was the lowest
coefficient of drag. Overall, length The all the models with new fairing design had better
performance except the sample with l = 1.14 d of fairing in the high Reynolds number range.
For most of the

(a) Low Reynodls Number(Re = 4.0 × (b) High Reynodls Number(Re = 7.0 ×
105 to 4.5 × 105 ) 105 to 7.5 × 105 )

Fig. 13: Avarage drag Coefficient measured at lower and higher Reynolds number for each
test sample with various angles of attack

The drag coefficients for each fairing were plotted against varying angle of attack. Low

13 of 17

American Institute of Aeronautics and Astronautics

Reynold’s number testing data is shown in Fig. 13a and High Reynold’s number data in
Fig. 13b. The general trend shown in the plot is the increase in the coefficient of drag of all
models except the bare wheel with increasing angle of attack. In addition from these plots,
it can be seen that among the new design of the fairings, the 2d length fairing generally had
the lowest drag coefficient compared to those of the individual wheel alone over the range of
the angle of attack. However, the covered fairing still had the best aerodynamic performance
with the lowest drag coefficients. This was expected as the covered fairing has no separation
between the wheel and after body, allowing for less chance of flow separation. Regardless
of this fact, it is still good to note that the weight difference between these fairings and the
covered fairing is substantial despite having a slight difference in drag performance.
I am not sure about this. Also evident in Fig. 13 is the decrease in drag coefficient, going
from the low Reynold’s number data to the high Reynold’s number data. A comparison
between these two plots shows that, in general, the drag coefficient reduces as the Reynold’s
number is increased. This is due to turbulent flow which occurs in higher Reynold’s number
flight. Turbulent flows remain attached to the fairing surface longer than flow during low
Reynold’s number. This increases pressure acting on the rear of the model and therefore,
reduces the drag component of the landing gear.
This may not be correct. The outlier in the wind tunnel experiment was the 1.14d fairing.
Drag coefficient for this fairing came out to be much higher than the 1d fairing. Some data
points even exceeded those of the individual wheel. This may be attributed to manufacturing
error such as having substantial roughness or improper shaping of the model.

B. Flow Visualization
Flow over the test samples for each run is shown in Fig.XX. The flow over the bare wheel
is shown in Fig.XX.
As indicated by Fig.XX, when the test sample has a positive angle of attack, the flow
bottom of the wheel goes in the gap between the wheel and fairing, moving up to the top
side of the test model. When the flow from the bottom side mixes with the flow on top of
the sample, the separation occurs. When the test sample has a negative angle of attack,
the flow on top side of the sample moves in the gap and causes the flow to separate on the
bottom side of the model.
In the case of model having zero and small angles of attack, the flow stays attached on
the wheel and fairing as shown in Fig.XX.
When the wheel has large angles of attack, the pressure difference between the top and
bottom sides of the models is large, causing the flow suction into the gap.

VII. Conclusion
As of May 5th, the design of test samples and mounts is completed. Also, fabrication of
the test models has begun. However, experiments have not been executed, and therefore,
data cannot be presented.

14 of 17

American Institute of Aeronautics and Astronautics

 References
1
Federal Aviation Administration, ”Aircraft Landing Gear Systems”, Aviation Maintenance Technician
Handbook-Airframe, Vol. 2, 2012, Chapter 13.
2
Herrnstein, H., and Biermann, D., ”The drag of airplane wheels, wheel fairings, and landing gears - I”
NACA Report-485,. 1935.
3
Anderson, J. D., ”Basic Aerodynamics”, Introduction to Flight, 7th ed, McGraw-Hill, New York, NW.
2012, pp 134-288.
4
Anderson, J. D., ”Introduction to the Fundamental Principles and Equations of Viscous Flow”, Fun-
damentals of Aerodynamic, 3rd ed, McGraw-Hill, New York, NW. 2001, pp 713-745.
5
Anderson, J. D., ”Fundamentals of Invisid, Incompressible Flow”, Fundamentals of Aerodynamic, 3rd
ed, McGraw-Hill, New York, NW. 2001, pp 177-275.
6
Roshko, A., ”On the Wake and Drag of Bluff Bodies”, AIAA Journal, Vol. 22, No. 2 (1955), pp. 124-132.
7
The Goodyear Tire & Rubber Company, ”Flight Special II” GOODYEAR Aviation Tires ,Ohio,
2013 [http://www.goodyearaviation.com/cfmx/web/aviattiresel/details.cfm?sortorder=35#FB ACCESSED
4/28/2014]
8
Cessna Sales and Service, ”Description and Operating Details”, Model 150 Owner’s Manual, Wichita,
Kansas, 1968.
9
Anderson, J. D., ”Standard Atmosphere, English Engineering Units”, Introduction to Flight, 7th ed,
McGraw-Hill, New York, NW, 2010, pp 770-773.
10
Cengel, Y. A., ”Conversion Factors”, Thermodynamics: Engineering Approach, 7th ed, McGraw-Hill,
New York, NW, 2012, pp 655-687.
11
The Engineering ToolBox, ”Absolute and kinematic viscosity of air at temperatures rang-
ing -40 - 1000 °C (-40 - 1500 oF) at standard atmospheric pressure - Imperial and SI Units”
[http://www.engineeringtoolbox.com/air-absolute-kinematic-viscosity-d 601.html ACCESSED 5/2/2014]

15 of 17

American Institute of Aeronautics and Astronautics

 Appendix
A. Fairings tested by Herrnstein and Biermann

(a) Sample wheel (b) Fairing with (c) Fairing with

with fairing l=0.7d l=1.14d

(d) Fairing with (e) Fairing with

l=1.5d l=2d

Fig. 14: Other types of fairings tested by Herrnstein and Biermann2

B. Item Purchased

Table 5: Items Purchased

4-4x6 12 feet wooden beams

Sanding Sealer
1-1/4 in metal rod
1-3/8 in metal rod
1-1/2 in metal rod
2-3/4 in brass pipes 5 feet long
3-3/4 in brass elbows
1/2 in dowel
5/8 in dowel
3/4 in dowel
epoxy
screws

16 of 17

American Institute of Aeronautics and Astronautics