Probability and Confidence Intervals
Probability and Confidence Intervals
AME 30333
University of Notre Dame
Spring 2016
Abstract
The purpose of this report is (1) to measure and analyze pressure distribution around an
airfoil, (2) to calculate lift, drag, and moment coefficients for the entire airfoil from discrete,
local pressure coefficient measurements, (3) and to compare experimental results with thin airfoil
theory. To start the lab, a NACA 0015 symmetric airfoil will be placed in the wind tunnel with
pressure taps located at known distances from the leading edge. Then pressure measurements
will be taken and recorded to the computer with a pressure transducer connected to a Scanivalve.
The local pressure coefficient will be calculated at each pressure tap on both the upper and lower
airfoil surface. Next, the normal and axial force coefficients will be tabulated for each pressure
tap location which will then be transformed to a Riemann sum to find the total lift, drag, and
quarter chord moment coefficients for the entire airfoil. By repeating these measurements at
different angles of attack, the lift curve slope can be calculated and compared to the expected
value from thin airfoil theory. Similarly, the coefficient of pressure can be used to calculate the
moment coefficients at the leading edge and the quarter chord locations.
1
2
1 Introduction
1.1 Thin Airfoil Theory
Thin Airfoil Theory is the work of Max Munk, a colleague of Ludwig Prandtl. For the interested
reader his original report was first translated into English in 1923 is NACA report 142, General
Theory of Thin Wing Sections [1].
Some preliminary kinematics are necessary for thin airfoil theory. The first component is
a vortex, defined mathematically as a point that produces purely tangential velocity inversely
proportional to the distance from the vortex:
Γ
Vθ = − . (1)
2πr
Vθ is the velocity induced at a point of interest distance r from the vortex with strength Γ, positive
when producing clockwise velocity. Now assume that this vortex is a line that extends infinitely
into and out of the page forming a vortex filament. A vortex sheet is formed by having a continuous
function of vortex filaments. It is this vortex sheet that Prandtl defined which allowed Munk to
quantify thin airfoil theory as will be seen.
The circulation, or strength, of the vortex filament per unit length is defined as γ(s). Using
Eq. 1, the infinitesimal velocity dV∞ induced due to a vortex filament is described by
γds
dVθ = − , (2)
2πr
where ds is the infinitesimal arc length along the vortex sheet. The primary assumption of thin
airfoil theory is to assume that a thin airfoil (one where the thickness is on the order of magnitude
smaller than the chord) can be replaced with a vortex sheet. This sheet corresponds to the path
of the camber line of the airfoil which is also a streamline so that the camber line is impermeable.
dz
Figure 1 shows this camber line, z(x), where the local slope can be written as arctan − dx . In
addition to the vortex sheet there is a free-stream velocity V∞ at some angle of attack α.
The velocity induced by the vortex sheet, w(s), at any point along the camber line is equal and
opposite to the free-stream velocity normal to the camber line, V∞,n . When these two velocities are
equal and opposite the camber line lies along a streamline because the flow is everywhere parallel
to streamlines by definition. Next the assumption is made that the vortex sheet instead of being
aligned along the camber line, is along the chord line. Therefore w0 (s) ≈ w(x) because the distance
between the camber line is very small, and for a symmetric airfoil the difference is zero. So for the
normal free-stream to be cancelled out by the induced velocity of the vortex sheet,
From trigonometry and Figure 2, the local normal free-stream velocity V∞,n can be written as
dz
V∞,n = V∞ sin α + arctan − . (4)
dx
Assuming small angles of attack and a thin airfoil, a small angle approximation is used (α ≈ sin α ≈
arctan α), and so (4) becomes
dz
V∞,n = V∞ sin α − . (5)
dx
Using (5) and integrating from the leading edge (LE) to the trailing edge (TE) with the dummy
variable along the chord line ξ, the total induced velocity due to the vortex sheet w(x) can then be
found as follows: Z c
γ(ξ)
w(x) = dξ. (6)
0 2π(x − ξ)
Plugging (6) and (4) into (3) with small rearrangement yields the fundamental equation of thin
airfoil theory for an uncambered airfoil:
Z c
1 γ(ξ)
dξ = V∞ α. (7)
2π 0 (x − ξ)
In this lab the NACA 0015 symmetric airfoil is being used, so the uncambered assumption
dz
that dx = 0 is accurate, and remember that in thin airfoil theory α is always measured in radians.
Next it is necessary to make a change of variables from ξ and x to θ such that ξ = 2c (1 − cos θ),
dξ = 2c sin θdθ, and x = 2c (1 − cos θ0 ). Substituting into (7), and noting that the bounds of
integration must change such that θ = 0 at the leading edge and that θ = π at the trailing edge.
1 + cos θ
γ(θ) = 2αV∞ . (8)
sin θ
Now it is necessary to find the total circulation due to the vortex sheet.
Z c
c π
Z
Γ= γ(ξ)dξ = γ(θ) sin θdθ = παcV∞ (9)
0 2 0
Equation (9) is quite useful when combined with the Kutta-Joukowski theorem: lift per unit span
(2D lift) L0 is equal to ρ∞ V∞ Γ. Hence the lift coefficient from thin airfoil theory is proportional to
angle of attack:
L0 ρ∞ V∞ παcV∞
Cl = 1 2
= 1 2
= 2πα (10)
2 ρ∞ V∞ 2 ρ∞ V∞ c
Equation (10) implies that the constant of proportionality between lift and angle of attack is the
lift slope Clα :
∂Cl ∂
Clα = = (2πα) = 2π. (11)
∂α ∂α
Another coefficient that can be obtained from thin airfoil theory is the moment coefficient
about both the leading edge and the quarter chord. Consider one vortex filament of strength
γ(ξ)dξ. Again from the Kutta-Joukowski theorem the differential lift from this vortex filament is
dL = ρ∞ V∞ dΓ. See Figure 3 for a diagram of this increment of lift.
To find the total moment per unit span about the leading edge integrate across the entire
chord: Z c Z c
0 1 2 2 πα
MLE =− ξ(dL) = −ρ∞ V∞ ξγ(ξ)dξ = − ρ∞ V∞ c . (12)
0 0 2 2
By definition of the moment coefficient per unit span (2D) and using the previous equation for
the leading edge moment, the following is found:
0
MLE πα Cl
CmLE = 1 2 2
=− =− . (13)
2 ρ∞ V∞ c
2 4
More importantly the moment coefficient at the quarter chord can be found from the leading
edge since the moment about the aerodynamic center is where all the lift acts:
Cl
Cmc/4 = CmLE + =0 (14)
4
Thin airfoil theory predicts that the moment coefficient about the quarter chord is 0, which
means that the aerodynamic center coincides with the center of pressure. No moment is required
to hold a symmetric airfoil at constant angle of attack about the quarter chord.
Finally take note that drag has been implicitly defined as zero in the derivation of thin airfoil
theory. The Kutta-Joukowski theorem relates circulation to lift for a potential flow. Using thin
airfoil theory and the Kutta-Joukowski theorem, a real airfoil can be replaced by a streamline
segment with a vorticity distribution in the incompressible flow.
Here θ is the local angle between a pressure normal vector and the line normal to the chord line. θ is
positive when measured clockwise from the normal, and negative when measured counterclockwise.
See Figure 4 for a diagram of this coordinate system.
And similarly for the axial force coefficient per unit span:
Z TE Z TE
0
A = (−pu sin θ + τu cos θ)dsu + (pl sin θ + τl cos θ)dsl (19)
LE LE
The assumption is made that the pressure terms are of greater significance than the shear stress
Seeing that ds cos θ = dx then the normal force coefficient in terms of the local pressure is
Z c Z c Z c Z c
pu x pl x pl pu x x
Cn = − d + d = − d = (Cpl − Cpu )d . (24)
0 q∞ c 0 q∞ c 0 q∞ q∞ c 0 c
By doing the same operations the axial force and leading edge moment coefficients in terms of the
local pressure coefficients follow:
Z c y
Ca = (Cpl − Cpu )d , (25)
0 c
Z c Z c
x x y y
CmLE = (Cpu − Cpl ) d + (Cpl − Cpu ) d . (26)
0 c c 0 c c
#taps
X 1 yi+1 yi
Ca = − Cpi + Cpi+1 − , (28)
2 c c
i=LE
#taps #taps
X 1 xi xi+1 xi xi+1 X 1 yi yi+1 y i yi+1
CmLE = − Cpi + Cpi+1 − − Cpi + Cpi+1 − .
2 c c c c 2 c c c c
i=LE i=LE
(29)
Splitting the upper and lower surfaces, the discretized equations become
X 1 xi+1 xi X 1 xi+1 xi
Cn = Cpi + Cpi+1 − − Cpi + Cpi+1 − (30)
2 c c upper
2 c c
lower
" # " #
X 1 yi+1 yi X 1 yi+1 yi
Ca = − Cpi + Cpi+1 − − − Cpi + Cpi+1 − (31)
2 c c upper
2 c c
lower
" #
X 1 xi xi+1 xi xi+1 X 1 yi yi+1 y i yi+1
CmLE = − Cpi + Cpi+1 − − Cpi + Cpi+1 − −
upper
2 c c c c upper
2 c c c c
" #
X 1 xi xi+1 xi xi+1 X 1 yi yi+1 y i yi+1
− Cpi + Cpi+1 − − Cpi + Cpi+1 −
2 c c c c 2 c c c c
lower lower
(32)
2 Description of Experiment
2.1 Pressure Transducer
A pressure transducer is a device that converts a pressure into a quantity that can be measured.
One quantity is a voltage that can be measured using an analog-to-digital converter or a digital
voltmeter, and another quantity would be a height difference in a U-Tube manometer.
Two common types of transducers are strain gage and capacitance types. The strain gage
pressure transducer, which is the one used in this lab, consists of a thin circular diaphragm on the
bottom of which are bonded tiny strain gages wired as a Wheatstone bridge. When the diaphragm
experiences a pressure on its exposed upper surface that is different from the pressure in a small
cavity under the diaphragm it deflects, and the resulting bridge imbalance is a measure of the
deflection. This deflection is usually very small and will need amplification after it is converted to
a voltage.
The capacitance-based pressure transducer has a stretched membrane clamped between two
insulating discs, which also support capacitive electrodes. A difference in pressure across the
diaphragm causes it to deflect, increasing one capacitor and decreasing the other. These capacitors
are connected to an electrical, alternating-current (AC) bridge circuit, producing a high level of
voltage output.
Strain gage transducers can be made small, hence they can be internally mounted in a wind
tunnel model. Also, they have reasonably good frequency response because of the small mass of the
diaphragm and the short distance between the pressure tap and the diaphragm face. Capacitance
transducers usually are not well suited for internal mounting and such systems do not have a fast
response.
In this experiment, the pressure transducer measures the static pressure at each tap on the
airfoil (controlled by Scanivalve) referenced to the freestream static pressure. The freestream stag-
nation (total) pressure is measured after the Scanivalve has stepped through all ports on the airfoil.
This final measurement is the very last column of data in the saved .mat files saved.
2.2 Scanivalve
A Scanivalve will take multiple pressure taps and allow them to be measured using only one pressure
transducer. The tubes are all connected to one stationary disk that is on top of a moving disk. The
moving disk has a hole in it and will rotate the stainless steel tubes (coinciding with each pressure
tap) that will be measured by the pressure transducer. Pictured in Figure 5 is the 4SDS-1124
Scanivalve as installed with a stepper motor and the pressure tap tubing associated with this lab.
Figure 5: Scanivalve
2.3 Airfoil
A NACA 0015 airfoil is being used for this experiment. It has a thickness of 22.86 mm, and a chord
of 152.4 mm. This is a symmetric airfoil that is relatively thin, so thin airfoil theory will apply
reasonably well. Pressure taps have been placed on this airfoil at the x and y locations in Table 1,
with the origin at the leading edge. These values will be useful in calculations of Cn and Ca .
Upper Lower
Tap Number x (mm) y (mm) Tap Number x (mm) y (mm)
1 0.00 0.00 16 0.84 -2.44
2 0.84 2.44 17 3.33 -4.68
3 3.33 4.68 18 7.46 -6.71
4 7.46 6.71 19 13.17 -8.45
5 13.17 8.45 20 20.42 -9.85
6 20.42 9.85 21 29.11 -10.83
7 29.11 10.83 22 39.14 -11.35
8 39.14 11.35 23 50.42 -11.39
9 50.42 11.39 24 62.82 -10.96
10 62.82 10.96 25 76.20 -10.09
11 76.20 10.09 26 90.41 -8.80
12 90.41 8.80 27 105.31 -7.15
13 105.31 7.15 28 120.71 -5.16
14 120.71 5.16 29 136.47 -2.87
15 136.47 2.87
Table 1: Tables of NACA 0015 coordinates for the UPPER surface on the left and the LOWER
surface on the right.
Conversely, a downstream obstruction will cause concave up streamlines over the pressure tap.
By using the Euler-N equation one could decide how to change the geometry near the tap, and
hence the streamlines, to cause a pressure reading not indicative of the local static pressure. Also
if a pressure gradient exists, a larger pressure tap will give an average pressure over a larger area,
thus reducing its preciseness.
3 Procedure
3.1 Part 1 - Thin Airfoil Theory
1. Check that the pitot tube is oriented parallel to the flow inside the tunnel.
2. Check the pressure system. The pitot tube static pressure port (parallel to flow) is hooked up
to the TOP (reference) port on the transducer. The Scanivalve is hooked up to the BOTTOM
(measurement) port on the transducer.
3. Check the Analog In system. The BNC cable from the Scanivalve is connected to the AI0
channel.
4. Check the Digital Out system. The Scanivalve’s “STEP” BNC is connected to DO1 and the
“HOME” BNC is connected to DO3.
5. Turn on the Scanivalve controller and send it to the home setting by pressing the “Home”
button.
8. Make sure Digital Out is enabled. Set “No. Steps” to 1. Leave “Direction” as “Pos.”.
12. Click the “Run” button. The program will automatically acquire the mean pressure trans-
ducer voltages for each of the 29 pressure taps and the upstream stagnation pressure. The
data collected is displayed in “Plot 2”.
13. When the acquisition is complete, click “Save .MAT” in the “Plot 2 - All Data” panel. Name
the file with the appropriate angle of attack (“aoa n05”, “aoa 00”, “aoa 05”, “aoa 07p5”,
“aoa 10”, “aoa 12p5”, “aoa 15”, “aoa 17p5”).
14. Once the data has been saved, press “Clear Data” in the “Plot 2 - All Data” panel. Click
“Yes” and “Yes” to the popup dialog boxes.
15. Repeat steps 11-14 for angles of attack of 0, 5, 7.5, 10, 12.5, 15, and 17.5 degrees using
appropriate the filenames for each.
2. Make sure that the data is cleared in “Plot 2 - All Data”. If it is not, press “Clear Data” and
click “Yes” and “Yes” to the popup dialog boxes.
5. When the acquisition is complete, click “Save .MAT” in the “Plot 2 - All Data” panel. Name
the file “aoa 05 tape”.
(a) Calculate the pressure coefficient CP at each pressure tap location. The position data
for each pressure tap is given in Table 1.
(b) Plot the negative of the pressure coefficients −Cp of both the suction and pressure sides
versus the chord-wise location in mm: −Cp versus x. Make a separate plot for each
angle of attack. (7 plots)
(a) For each angle of attack, calculate the axial force coefficients Ca , normal force coefficients
Cn , and leading edge moment coefficients Cm,LE using the split discretized equations.
(b) Calculate the lift coefficients Cl , drag coefficients Cd , and the quarter chord moment
coefficients Cm,c/4 .
(c) Plot the lift coefficient versus angle of attack in degrees: Cl versus α. In the same graph,
plot the 2π/rad theoretical slope on your plot (be careful with your units!). Provide
brief comments on the data and how it compares to theory. (1 plot)
(d) Plot the drag coefficient versus angle of attack in degrees: Cd versus α. Provide brief
comments on the data. (1 plot)
(e) Plot quarter-chord moment versus angle of attack in degrees: Cm,c/4 vs. α. Provide
brief comments on the results. (1 plot)
(a) For the data set with tape on the airfoil (“aoa 05 tape.mat”) calculate the pressure
coefficient CP at each pressure tap location.
(b) Using your Cp data from step 1(a) and 3(a), plot the negative pressure coefficients verus
chord-wise position in mm for α = 5 degrees with and without tape: −Cp versus x. How
does the tape affect the pressure upstream and downstream from it? How do you expect
the tape to affect the pressure? If you do not see any effect, try to explain why this
might be. (1 plot)
2. Cl versus α in degrees. Overlay the 2π slope line that is predicted by thin airfoil theory.
Comment on what your data shows and how well it compares with thin airfoil theory.
5. −Cp versus chordwise location showing just the upper surface for α =5o with and without
tape on the same plot. Comment on what your data shows including how the tape affects
the pressure at the nearby ports. If you do not see any effect due to the tape, try to explain
why this may be? How do you exect the tape to affect the pressure.
IMPORTANT:
• Make sure plots are printed large enough to see everything clearly.
• Make sure all plots have appropriate titles and axes labels.
• If there are multiple lines on a single plot, make sure they are labeled and dis-
tinguishable by line style, markers, and/or colors.
References
[1] Munk M.M., ”General Theory of Wing Sections,” Tech. Rep. 142, NACA, 1923.
[3] McKeon, B. and Engler, R., ”Pressure Measurement Systems,” Springer Handbook of Experi-
mental Fluid Mechanics, 2007, pp. 179-214.