THE LIFT AND DRAG OF A WING

15/03/2022
 Contents
Aims and Objectives ........................................................................................................................... 3
Apparatus and Instrumentation ........................................................................................................ 3
Experimental Procedure .................................................................................................................... 3
Results .................................................................................................................................................. 3
Experimental Results Table ........................................................................................................... 4
Sample Calculation ......................................................................................................................... 5
Experimental Results Table and Graphs ...................................................................................... 5
Theoretical Predictions ................................................................................................................... 6
Errors .................................................................................................................................................. 7
Discussion ............................................................................................................................................ 7
Conclusion ........................................................................................................................................... 8
References............................................................................................................................................ 9
 Aims and Objectives
The objectives of this experiment were:
- To measure the variation with incidence and Reynold’s number of the lift and drag
forces acting on a wing of moderate aspect ratio in a low-speed wind tunnel
- To compare the results, in coefficient form, with the predictions of simple theory
The aim of these objectives was to gain an understanding of the effect of speed and angle of
attack on the forces acting on a wing, and to analyse and discuss the accuracy and reliability
of theoretical predictions.

Apparatus and Instrumentation

The experiment utilised a closed-circuit low speed wind tunnel, with a contraction ratio of
5.6, to generate a flow of air over a cambered wing section placed horizontally in the tunnel.
The wing section had a span of 600mm, a chord of 100mm and a maximum thickness of
11.35mm.
The forces on the wing were measured by a three component, virtual centre mechanical
balance. This was attached to the wing in three positions, two rods connected to the underside
of the wing section and one rod connected downstream of the wing via a small connection
rod. The angle of attack could be varied using the balance by altering the height of the
downstream rod, the angle was displayed by a protractor on the balance. Load cells in the
balance measured the lift, drag and pitching moment on the wing and displayed the readings
digitally. A computer averaged the forces over one second intervals and displayed ten
readings (corresponding to ten seconds of data) for each angle of attack. A Betz manometer
measured the pressure difference between the working section and the upstream section of
the wind tunnel in mmH2O, which was used to calculate the velocity. A thermometer and
barometer measured the atmospheric temperature and pressure in °C and mmHg respectively,
which were used to calculate the density of air.

Experimental Procedure
Firstly, the atmospheric temperature and pressure were measured, and the dimensions of the
wing were noted. The wind tunnel was turned on and the velocity was adjusted such that the
Betz manometer displayed a reading of 19.3mmH2O. Once the velocity was set, the wing
was positioned at an angle of attack of -5° using the three-component balance. Once
fluctuations had subsided, the computer recorded the forces acting on the wing over 10
seconds, producing 10 values as mentioned previously. The angle was then adjusted to -3°
and the process repeated. The angle was increased in increments of 2° up until it reached 9°,
where the increment was then reduced to 1°, measurements were taken up to 17°. This whole
process was repeated for a higher velocity, corresponding to a Betz manometer reading of
32.9mmH2O.

Results
The coefficients of lift and drag were calculated using the following equations:
𝐿 𝐷
𝐶𝐿 = 1 2 𝑏𝑐
(1) 𝐶𝐷 = 1 2 𝑏𝑐
(2)
𝜌𝑢∞ 𝜌𝑢∞
2 2
(Motallebi, 2022) (Motallebi, 2022)
The following equation is based on the Bernoulli equation and gives the velocity in the
working section of the tunnel:
2∆𝑃
𝑢∞ = √ 𝐴 2
(3) (Motallebi, 2022)
𝜌[1−( 2) ]
𝐴1

The density of the air inside the tunnel was approximated to be equal to the density of the air
in the lab, meaning the temperature and pressure measured in the lab can be used to find the
density of air in the tunnel. The velocity of the air in the tunnel is calculated using the
pressure difference, P, and the contraction ratio. The density and pressure difference are
calculated as follows:
𝑃 Δ𝑃 = 𝜌𝑤𝑎𝑡𝑒𝑟 𝑔ℎ𝐵𝑒𝑡𝑧 (5)
𝜌 = 𝑅𝑇 (4) (Motallebi, 2022)
The Reynolds number must also be calculated to analyse its effect on the lift and drag forces.
Viscosity is calculated using equation 6, Sutherland’s equation, and the Reynold’s number is
calculated using equation 7, both shown below.
3
𝜇 𝑇𝑟𝑒𝑓 +𝑆 𝑇 2
𝜌𝑈∞ 𝑐
=( ) (𝑇 ) (6) (Motallebi, 2022) 𝑅𝑒 = (7)
𝜇𝑟𝑒𝑓 𝑇+𝑆 𝑟𝑒𝑓 𝜇

𝜇𝑟𝑒𝑓 = 1.789 × 10−5 , 𝑇𝑟𝑒𝑓 = 288.2𝐾 , 𝑆 = 110.4𝐾 (Motallebi, 2022)

The values of atmospheric pressure, P, temperature, T, density, ρ, pressure difference, P,
velocity, u∞, viscosity, μ, and Reynold’s number are available in table 1.
Table 1: Values of P, T, ρ, P, u∞, μ and Re for low and high speed tests.
P /Pa T /K ρ /kg/m3 P /Pa u∞ /m/s μ /Ns/m2 Re
-5
Low- 101963.18 293.15 1.21149 189.33 17.968 1.81277 x 10 1.201 x 105
Speed
High- 101963.18 293.15 1.21149 322.75 23.460 1.81277 x 10-5 1.568 x 105
Speed

Experimental Results Table

Table 2: Values of CL, CD, and CL2 at all angles of attack for low and high speed tests
Low Speed High Speed
Angle CL CD CL2 Angle CL CD CL2
-5 -0.325 0.041 0.106 -5 -0.282 0.038 0.080
-3 -0.132 0.032 0.017 -3 -0.108 0.029 0.012
-1 0.069 0.027 0.005 -1 0.096 0.025 0.009
1 0.290 0.030 0.084 1 0.279 0.026 0.078
3 0.454 0.033 0.206 3 0.454 0.030 0.207
5 0.590 0.038 0.348 5 0.596 0.036 0.355
7 0.733 0.047 0.537 7 0.719 0.044 0.517
9 0.811 0.059 0.657 9 0.815 0.058 0.665
10 0.857 0.068 0.734 10 0.844 0.064 0.713
11 0.892 0.078 0.796 11 0.888 0.075 0.789
12 0.919 0.091 0.845 12 0.919 0.085 0.845
13 0.797 0.162 0.635 13 0.920 0.116 0.846
14 0.718 0.187 0.515 14 0.719 0.185 0.518
15 0.715 0.211 0.511 15 0.713 0.207 0.508
16 0.720 0.236 0.518 16 0.715 0.236 0.512
17 0.710 0.249 0.504 17 0.710 0.249 0.504
 Sample Calculation
A sample calculation of CL and CD is provided below for the low-speed test at 5° angle of
attack, the mean of the measured lift and drag values is taken to give one average value:
𝐿̅ 6.77
𝐶𝐿 = = = 0.59007
1 2 1
( )( )2
2 𝜌𝑢∞ 𝑏𝑐 2 1.21149 17.97 (0.600)(0.100)
̅
𝐷 0.4385
𝐶𝐷 = = = 0.03882
1 2 1
( )( )2
2 𝜌𝑢∞ 𝑏𝑐 2 1.21149 17.97 (0.600)(0.100)
Experimental Results Table and Graphs
CL against α
1.5
Low Speed
1
High Speed
0.5 High Speed
CL

Stall Angle
0 Low Speed
-10 -5 0 5 10 15 20 Stall Angle
-0.5
α/°
Graph 1: CL against α. A trendline has been fitted to the linear region of both plots. The
colour of the trendline matches the colour of the plot for both speeds.
CD against α
0.3
0.25
0.2

CD 0.15
Low Speed
0.1
High Speed
0.05
0
-10 -5 0 5 10 15 20
α/°
Graph 2: CD against α.

CD against CL
0.3
0.25
0.2
Low Speed
CD 0.15 High Speed
0.1
0.05
0
-0.4 -0.2 0 0.2 0.4 0.6 0.8 1
CL
Graph 3: CD against CL.
 CD against CL2
0.3
0.25
0.2
Low Speed
CD 0.15 High Speed
0.1
0.05
0
-0.2 0 0.2 0.4 0.6 0.8 1
C L2
Graph 4: CD against CL2. A trendline has been fitted to the linear region of both plots. The
colour of the trendline matches the colour of the plot for both speeds.
Table 3: Experimental values of lift curve slope, a w, lift dependent drag coefficient, K, profile
drag, CD0, (CL/CD)max and angle of (CL/CD)max.
Low Speed High Speed
dCL/dα (aw) 5.6723 5.3285
2
dCD/dCL (K) 0.039500 0.042100
Intercept (CD0) 0.0254 0.0218
(CL/CD)max 15.507 16.624
Angle of (CL/CD)max 7° 5°

Theoretical Predictions
Lifting Line Theory predicts the lift acting on the wing in inviscid flow. Due to the absence
of boundary layer separation, the relationship between lift coefficient and angle of attack is
linear in this model, and the predictions are usually closely matched to experimental data in
the linear region. The lift curve slope is calculated using equation 8 (shown below), where,
after taking into account viscous effects, a value of 𝑎 = 5.7 𝑟𝑎𝑑 −1 is used. Using this
equation, we obtain the theoretical lift curve slope for the wing, which, it should be noted, is
the same for both Reynold’s numbers. The value of aw is shown in table 4. Lifting Line
Theory also predicts that CD will rise linearly with CL2. The theoretical distribution of CD
against CL2 is of the form 𝐶𝐷 = 𝐶𝐷0 + 𝐾𝐶𝐿2 where CD0 is the profile drag coefficient and KCL2
is the lift dependent drag. The value of CD0 depends on the position of boundary layer
transition, which cannot accurately be determined for this experiment, however the value
should lie between around 0.01 and 0.02. Assuming the lift dependent drag is composed
entirely of induced drag, lifting line theory gives equation 9, shown below, where
calculations show that 𝛿 ≈ 0.011𝐴𝑅 for wings of rectangular planform. The theoretical value
of K is shown in table 4.
𝑎 1+𝛿
𝑎𝑤 = 𝑎 (8) (Motallebi, 2022) 𝐾 = 𝜋𝐴𝑅 (9) (Motallebi, 2022)
1+
𝜋𝐴𝑅

Table 4: Theoretical values of aw and K.

aw K
4.3765 0.56553
 Errors
The percentage error in the results is calculated as follows:
|𝑒𝑥𝑝𝑒𝑟𝑖𝑚𝑒𝑛𝑡𝑎𝑙 𝑟𝑒𝑠𝑢𝑙𝑡−𝑡ℎ𝑒𝑜𝑟𝑒𝑡𝑖𝑐𝑎𝑙 𝑟𝑒𝑠𝑢𝑙𝑡|
𝑝𝑒𝑟𝑐𝑒𝑛𝑡𝑎𝑔𝑒 𝑒𝑟𝑟𝑜𝑟 = × 100 (10)
𝑡ℎ𝑒𝑜𝑟𝑒𝑡𝑖𝑐𝑎𝑙 𝑟𝑒𝑠𝑢𝑙𝑡

Using equation 10, the percentage error is obtained for the values of aw and K at low and high
speed. The percentage errors are available in table 5.
Table 5: Percentage errors for aw and K for low and high speed tests.
Percentage Error (%)
aw K
Low Speed 29.6 30.2
High Speed 21.7 25.6

Discussion
The experiment was successful in illustrating how the lift and drag forces acting on a wing
section vary with incidence angle and Reynold’s number. Graphs 1 and 2 clearly show the
distribution of lift and drag coefficient for both speeds across all incidence angles measured,
allowing for the stall angle to be observed and the lift curve slope to be calculated. Graphs 3
and 4 show how the drag coefficient varies with the lift coefficient, allowing the profile drag
and lift dependent drag coefficient to be obtained.
The experiment was not difficult to complete, however some factors could have affected the
results. Firstly, the method of adjusting the incidence angle was not precise and could have
incurred some error in the results. Secondly, the computer took time-averaged values of the
forces for 10 seconds, which could have been insufficient time for the flow around the wing
to settle, thus affecting the data.
When looking at graph 1, the stall angle is seen to be different for the different air speeds,
with the low-speed test producing a smaller stall angle. To understand why this happened, we
must analyse what causes a wing to stall. Stall occurs when the boundary layer detaches from
the aerofoil surface, generating turbulence that breaks the pressure difference between the
upper and lower surfaces of the aerofoil profile (Sánchez, 2014). This causes lift to drop
sharply as seen at 12° for the low-speed test and at 13° for the high-speed test. The higher
Reynold’s number flow was able to maintain lift at higher angles because the flow begun to
transition to turbulent flow sooner. Turbulent boundary layers are more resistant to separation
than laminar ones, so if the flow transitions sooner, the boundary layer will be able to remain
attached to the surface for longer. This allows the wing section to continue producing more
lift as seen in graph 1. The Reynold’s number also has an impact on the drag coefficient
distribution as seen in graph 2, causing the sharp increase associated with stall to be delayed.
Aside from the differences seen in graphs 1 and 2 at 13° due to the stalls, the distributions are
almost identical. The lift curve slopes of the high and low speed flow, shown in table 3, are
very similar, with the low-speed slope only 6.45% larger than the high speed one. This
difference could be due to experimental errors that lead to a higher gradient such as an
incorrectly set incidence angle. While both distributions are very similar, they deviate
significantly from the distribution predicted by lifting line theory. This is because lifting line
theory is based on inviscid flow, which was not the case in the experiment. With inviscid
flow, the flow never separates from the wing, meaning that the lift can continue to increase as
the angle increases without encountering a stall. In reality, the viscous forces of the fluid act
on the wing section, causing a boundary layer to form which eventually separates, causing the
pressure distribution to change drastically and lift to decrease. The distributions deviate
further from the theory as the incidence angle increases. This is because the separation point
moves further forward as incidence angle increases, causing more lift to be lost. The
distributions begin to deviate from the theory at the point of first separation (on the trailing
edge). The deviation increases until the wing stalls when the separation reaches the leading
edge. We see that the values of (CL/CD)max occur at 5° and 7° for the high and low speed
flows respectively. At this angle, the lift produced is still high as separation is small, and the
drag has not yet begun to sharply increase.
Graphs 3 and 4 tell us about how the drag coefficient varies with the lift coefficient. Lifting
line theory states that CD varies linearly with CL2, and a region of linearity can be seen in
graph 4. This relationship is linear when the induced drag is totally responsible for the lift-
dependent drag. Induced drag is the component of drag caused by the vortices at the wing tips
(Shelton, 2021), it increases with lift as the vortices grow. In an inviscid flow, this will be the
only source of lift-dependent drag, however in a real flow, the lift-dependent drag is also
greatly contributed to by the viscous drag associated with boundary layer separation. The
deviation between the real plot and theoretical prediction seen in graph 4 is a result of this
viscous drag and causes the results to differ greatly. Despite the large deviation due to
viscous drag, the plots both display strongly linear behaviour between the points
corresponding to angles 1° and 7°. In this region, there is little to no separation on the wing,
so the primary contributor to the lift-dependent drag is the induced drag. The values of K for
both plots are similar, with a percentage difference of just 6.58%, however the percentage
error between the real results and the theoretical predictions are much greater, due to viscous
drag playing a role even in the linear region.
The percentage errors, available in table 5, illustrate how far the experimental results were
from the theoretical predictions. These errors are all great enough to say that the results
disagree, and the reasons for this disagreement are varied. The primary reason that the results
disagree with the theory is that the theory makes assumptions that are not true of the
experiment and have a great impact on the results. The assumption that flow is inviscid
eliminates all effects of viscous forces on the wing. This creates massive deviations between
the theoretical and experimental plots seen in graphs 1 and 4 in the regions where flow
separates, but also leads to inaccurate results for a w and K.
Based on the factors above that could have affected results, there are suggested improvements
to the experiment. Firstly, the method of adjusting the angle of attack could be altered to
utilise a knob rather than a lever. This would allow for finer control and would make the
process easier for the user. Secondly, the system could be allowed to rest for a longer time
before measurements are taken, and measurements could be taken over a larger time period.
This would reduce the effects of turbulence on the results after the wing has been rotated.

Conclusion
The experiment was a success to a high degree; however, the results did not conform to
theoretical predictions. This was mainly due to inaccurate assumptions made in the theory
which led to large percentage error in the results. The improvements to the experiment
include changing the method of adjusting the angle and allowing more time for the system to
settle.
 References
1- Motallebi, F. (2022), The Lift and Drag of a wing, DEN233 Low Speed Aerodynamics,
QMUL School of Engineering and Material Science, available at:
https://qmplus.qmul.ac.uk/pluginfile.php/2718980/mod_resource/content/6/DEN233%20Lift
%20and%20Grad%20Lab%20-%20Handout%202021_22.pdf , accessed on 15/03/2022
2- Sánchez, S.J.R. (2014), Analysis of flow separation over aerodynamic airfoils, Universidad
Carlos III de Madrid, available at: https://core.ac.uk/download/pdf/44310532.pdf , accessed
on 18/03/2022
3- Shelton, N. (2021), How Induced Drag works, Aerodynamics, BoldMethod, available at:
https://www.boldmethod.com/learn-to-fly/aerodynamics/how-induced-drag-works/ , accessed
on 20/03/2022