Você está na página 1de 22

Finite Wing in Kirstin Wind Tunnel

Jacob Brown, and Jaime Katzer, Kelsey Moser


AA 321 Aerospace Structures I (AD)
William E. Boeing Department of Aeronautics & Astronautics
University of Washington, Seattle, Washington 98195-2400
2013-02-21


In this lab, a finite wing with a NACA airfoil was tested in the Kirsten
Wind Tunnel; runs were also conducted with circular endplates
installed to simulate an infinite wing. Lift, drag, and pitching moments
were measured, and coefficients were calculated, along with the span
efficiency factor and parasitic drag. China clay runs provided a visual
representation of the flow over the wing in both 3D and simulated 2D
configurations. Reduced data roughly matched final data provided by
the Kirsten Wind Tunnel, with discrepancies caused by differences in
moment centers and in data reduction processes.




Nomenclature
a 3D lift curve slope (1/degree)
AR Aspect ratio
b Span (ft)
DRAG Drag force corrected for balance interactions (lb)
DRAGR Raw drag force (lb)
e Span efficiency factor
c Chord length (ft)
D Drag force (lb)
L Lift force (lb)
LIFT Lift force corrected for balance interactions (lb)
LIFTR Raw lift force (lb)
2
M Pitching moment (lb-in)
NDRAG Normalized fork strut drag (ft
2
)
PM Pitching moment corrected for balance interactions (lb-in)
PMR Raw pitching moment (lb-in)
PMWT Pitching moment weight tare correction factor (lb-in)
q Dynamic pressure (psf)
QA Actual dynamic pressure (psf)
Re Reynolds number
V Airflow velocity (ft/s)
Air density (slug/ft
3
)
Air viscosity (Pa-s)
a
0
2D lift curve slope (1/degree)
C
L
Coefficient of lift
C
D
Coefficient of drag
C
Dp
Coefficient of parasitic drag
C
M
Coefficient of pitching moment
S Wing reference area (ft
2
)





I. Introduction

The purpose of the experiment was to investigate the lift, drag, and pitching moment on a three
dimensional NACA 23012 airfoil straight wing. An infinite wing was also simulated by
installing circular endplates on the finite wing. Data was collected at two different dynamic
pressures starting at both a high and a low angle of attack (AOA) to examine the effects of
aerodynamic hysteresis. Flow visualization runs were conducted to provide a tangible
representation of the way the flow over the wing at a high angle of attack differed when the
endplates were installed. Coefficients of lift, drag, and pitching moment were calculated for both
the finite and infinite wing and compared. The span efficiency and parasitic drag were also
estimated.
3
II. Theory

At subsonic speeds, viscous airflow causes a boundary layer near the surface of a body moving
through air. The boundary layer can be laminar, turbulent, or separated, and each condition has
severe effects on the lift and drag forces acting on the wing. A laminar boundary layer has the
lowest amount of energy and does not easily stay attached to the surface of a wing when pitched
to high angles of attack. Turbulent flow is more energetic and stays attached to the surface of a
wing at higher angles of attack than laminar flow, but produces more drag than that of laminar
flow. For this reason, flow over a wind tunnel model wing is often tripped to induce turbulent
flow, which ensures that the state of the boundary layer is known over the entire span. When
wings are pitched to high angles of attack, the boundary layer separates and causes a severe
decrease in lift and increase in drag. The angle where separation occurs is the stall angle, which
is different for every airfoil and wing. Because of the severe change in behavior of a stalling
wing, it is important to find the stall angle empirically (often through wind tunnel testing). An
important variable that characterizes this stall angle is the Reynolds number, which is defined in
Eq. 1 [1].

(1)
The aspect ratio of a wing is defined by Eq. 2 in terms of wing span and reference area.


(2)
The span efficiency factor of a wing is defined by Eq. 3.


(3)
The drag of a finite wing is defined by Eq. 4.


(4)
The three-dimensional lift curve slope is calculated using Eq. 5.


(5)
There are many corrections that are applied to wind tunnel data that correct for balance
imperfections, center of gravity movement, strut interference, wall effects, blockage effects, test
4
section air up-flow, and lift-induced drag. Some of these corrections are not applied manually,
of which the details are not included in this report.

Balance imperfections are taken from the data by applying a corrective balance interactions
matrix to the data, which is calculated during a full balance calibration. A simplified version of
the matrix is found below, where raw lift, drag, and pitching moment are corrected by Eq. 6
through Eq. 8.
T TR RR R (6)
R TR RR - R (7)
TR - RR R (8)

Center of gravity movement is corrected for with weight tares, where the model is yawed and
pitched to its most extreme positions. The induced pitching and rolling moment are interpolated
for various positions and subtracted from the data so that the only pitching and rolling moments
analyzed are those due to aerodynamic phenomena. The weight tare correction is given by Eq. 9.

fter tre before tre T (9)

Drag tares are applied to all wind-on data to account for the drag of the strut below the model.
The values of normalized drag were pre-determined for various side-sweep angles and are used
for correction by Eq. 10.

R fter tre R before tre - R (10)

When a finite wing travels through air, vortices occur at the wingtips due to the change in
pressure between the top and bottom of the wing. These vortices induce a negative angle of
attack and reduce the effective angle of attack of the wing. This phenomenon causes a
component of force normal to the airfoil in the drag-direction, thus causing lift-induced drag.

The lift, drag, and moment coefficients were calculated using Eqs. 11-13, respectively.
5

(11)


(12)


(13)

III. Experimental Apparatus

This experiment was conducted at the Kirsten Wind Tunnel (KWT), which is a dual return,
subsonic wind tunnel with a 12 x 8 x 10 ft test section. The speed of the wind, measured by test
section dynamic pressure, is controlled by adjusting propeller RPMs at the test control and
acquisition module (TCAM) in the control room. Two 500 horsepower engines each run a 7-
blade propeller in the wind tunnel circuit, which are mechanically connected by a jack shaft that
forces them to rotate at the same speed (regardless of the asymmetry of blockage inside the test
section). As a result, the test section has uniform airflow with no turbulence and a standard up-
flow of 0.012 degrees, a value that is corrected for during data reduction. There is an external,
six-degree of freedom balance below the test section that measures lift, drag, and side forces as
well as pitching, yawing, and rolling moments. This balance uses a series of mechanical
connections to separate each degree of freedom into small displacements. These displacements
are measured using Eastman Pots, which utilize a prism, a light source, photodiodes, a long wire
coil, and a magnet, to counter-act the displacements and, by doing so, measure the forces and
moments acting on the model in the test section. A schematic of the overhead view of the wind
tunnel is shown in Fig. 1.

6

Fig. 1. Overhead view of the Kirsten Wind Tunnel [1]

A finite wing with a NACA 23012 airfoil was mounted in the center of the test section on the
blnce strut with the KTs -H fork strut and #2 pitch arm. The angle of attack (alpha)
encoder was calibrated using a mechanical inclinometer and zero side-sweep (psi) was calibrated
by setting the distances equal on both sides of the model between the trailing edge of the wing
tips and the back of the test section.

IV. Procedure

The span, the chord, and the maximum thickness of the wing were measured and recorded, as
well as the positions of the trip dots from the leading edge of the wing. The angle of attack was
calibrated by Kirsten Wind Tunnel employees. The wing was pitched using an encoder, which
was controlled by TCAM. The alpha calibration equated encoder values with the angle of attack
of the wing, as measured by an inclinometer placed on the wing mount during installation.

Prior to the lab, a strut drag tare was conducted by the KWT employees. The drag acting on the
strut with no model installed was recorded at various dynamic pressures and subtracted from the
7
raw drag data measured on the wing. This correction was made by KWT for the reduced data
provided, but not for the raw data.

A weight tare was conducted for each model configuration. Both the finite and infinite wings
were tested at dynamic pressures of 10 and 35 psf. At each dynamic pressure, the wing was
pitched from -
o
to
o
in steps of
o
, and from

to
o
in steps of
o
. This run was then
repeated backwards, as angle of attack varied from
o
to -
o
in the same increments.

China clay flow visualization was conducted for each configuration. China clay is a mixture of
Day-Glo
TM
powder, Kentucky ball clay, and kerosene. The kerosene dries as the airflow in the
test section increases and leaves behind the brightly colored clay mixture to provide a
visualization of the airflow across the wing. For each of the flow visualization runs, the china
clay was painted on the top surface of the wing before bringing the wind up and pitching the
wing to an angle of attack of
o
, chosen to be a larger angle than the apparent stall angle.

V. Discussion

Using the data reduction steps shown in the procedure and Eq. 6-10, the corrected lift, drag, and
moment coefficients of the wing were found. The calculations are shown in Appendix A. The
results re shown in Tble long with the sme dt processed by the Kirstin ind Tunnels
data reduction process and the calculated percent difference. The error for lift and drag
coefficient were reasonable when considering the additional corrections that the KWT applied to
correct for wall effects, blockage effects, lift-induced drag, and up flow, as well as using a
different balance interactions matrix. The large error in pitching moment coefficient is due to the
KWT data being about the model moment center at the quarter chord and the manually corrected
data being about the balance moment center, not transferred to model moment center.





8
Tble ift, drg, nd moment coefficients of the wing with q psf nd
Lift Coefficient Drag Coefficient Moment Coefficient
Corrected 1.0019 0.0616 -0.0030
Corrected by KWT 1.0526 0.0686 0.00572
Difference -4.817 % -10.20 % 152.4 %

From Eq. 2 the wing was found to have an aspect ratio of 6.0.

The lift, drag, and moment coefficients of the wing were calculated using Eqs. 11-13 and plotted
in Fig. 2-5. The calculations are shown in Appendix A. Part a) of the plots shows the lift and
drag coefficients vs. the angle of attack of the wing. The linear region of the C
L
plots, referred to
as the lift curve slope, can be clearly seen as well as the general stall area where the linear region
breaks. The drag coefficient steadily increases with an increasing angle of attack, which is
expected as more cross-flow area is exposed to the flow. Part b) of the plots shows a closer look
at the stalling region of the lift coefficient plot shown in a). The wing exhibits different
characteristics when reaching the stalling point depending on if the wing was increasing or
decreasing its AOA when approaching a given position. This is given more detail later in the
discussion. Part c) of the plots shows the moment coefficients about the quarter-chord point of
the wing as well as the lift-to-drag ratio plotted against the angle of attack of the wing. Part d) of
the plots shows the lift coefficient as a function of the drag coefficient.
9

Fig. 2. Calculated properties of wing at q = 10 psf. a) shows the lift and drag coefficient
vs. angle of attack of the airfoil, b) is a close-up of plot a) near the stalling point for
increasing and decreasing AOA, c) shows the moment coefficient about the quarter
chord point and the lift-to-drag ratio, and d) shows the lift coefficient as a function
of the drag coefficient.

From part c) of Fig. 2 it can be seen that the maximum lift-to-drag ratio of the wing was about 18
and occurred at an angle of attack of approximately 6 deg. The moment coefficient about the
quarter-chord point remained relatively constant with angle of attack, which was expected. The
coefficient of lift steadily increased with the drag coefficient as the angle of attack of the wing
increased. Upon reaching the stall point, however, the lift leveled off while the drag continued to
increase due to separation drag.

10


Fig. 3. Calculated properties of wing at q = 35 psf. a) shows the lift and drag coefficient
vs. angle of attack of the airfoil, b) is a close-up of plot a) near the stalling point for
increasing and decreasing AOA, c) shows the moment coefficient about the quarter chord
point and the lift-to-drag ratio, and d) shows the lift coefficient as a function of the drag
coefficient.

From part c) of Fig. 3 it can be seen that the maximum lift-to-drag ratio of the wing was about 19
and occurred at an angle of attack of about 6 deg. Once again, the moment coefficient about the
quarter-chord point remained relatively constant with angle of attack and the coefficient of lift
increased with the drag coefficient as angle of attack increased. Upon reaching the stall point, the
lift leveled off as before while the drag continued to increase.

11

Fig. 4. Calculated properties of wing with circular end caps at q = 10 psf. a) shows the
lift and drag coefficient vs. angle of attack of the airfoil, b) is a close-up of plot a)
near the stalling point for increasing and decreasing AOA, c) shows the moment
coefficient about the quarter-chord point and the lift-to-drag ratio, and d) shows the
lift coefficient as a function of the drag coefficient.

From part c) of Fig. 4 it can be seen that the maximum lift-to-drag ratio of the wing was about 18
and occurred at an angle of attack of about 8 deg. Once again, the moment coefficient about the
quarter-chord point remained relatively constant with angle of attack and the coefficient of lift
increased with the drag coefficient as angle of attack increased. Upon reaching the stall point, the
lift leveled off as before while the drag continued to increase.

12

Fig. 5. Calculated properties of wing with circular end caps at q = 35 psf. a) shows
the lift and drag coefficient vs. angle of attack of the airfoil, b) is a close-up of
plot a) near the stalling point for increasing and decreasing AOA, c) shows the
moment coefficient about the quarter chord point and the lift-to-drag ratio, and d)
shows the lift coefficient as a function of the drag coefficient.

From part c) of Fig. 5 it can be seen that the maximum lift-to-drag ratio of the wing was about 18
and occurred at an angle of attack of about 8 deg. Once again, the moment coefficient about the
quarter-chord point remained relatively constant with angle of attack and the coefficient of lift
increased with the drag coefficient as angle of attack increased. Upon reaching the stall point, the
lift leveled off as before while the drag continued to increase.
13
The span efficiency factors of the wing with and without end caps at different dynamic pressures
were calculated using Eq. 4 as shown in Appendix A. The Results are shown in Fig. 6 plotted
against the corresponding lift coefficients.


Fig. 6. Span efficiency factor of wing w/ and w/o end caps

The span efficiency factor should be between zero and one. However, due to the interference of
the wind tunnel walls and floor, higher-than-one span efficiency factors were calculated. The
span efficiency is a result of downwash and wingtip vortices; the walls of the wind tunnel
obstruct these vortices, which increases the span efficiency factor. Blockage corrections are
commonly performed to account for this effect, but they were not included in the data reduction
for this lab. The addition of the circular endcaps further obstructed the wingtip vortices, and
further increased the span efficiency, as can be seen in Fig. 6. From calculations shown in
Appendix A, the averaged efficiency factor was found to be 1.118 without the end caps, and
increased to 1.527 when the end caps were installed.
The parasitic drag values of the wing with and without the end caps installed were calculated
using Eq. 4 to be 0.01417 and 0.02023 respectively. The calculations are shown in Appendix A.
These values were validated using a squared lift coefficient vs. drag coefficient plot as seen in
0.2 0 0.2 0.4 0.6 0.8 1 1.2 1.4
0.5
0
0.5
1
1.5
2
Lift Coefficient
S
p
a
n

E
f
f
i
c
i
e
n
c
y

F
a
c
t
o
r
,

e
Span Efficiency Factor vs. Lift Coefficient


q = 10 psf
q = 35 psf
w/ end caps; q = 10 psf
w/ end caps; q = 35 psf
Student Version of MATLAB
14
Fig. 7. The parasitic drag coefficient is the value of drag coefficient when the lift coefficient is
equal to zero. As can be seen by the plot, the values calculated are very reasonable.

Fig. 7. Squared lift coefficient plotted vs. drag coefficient to validate the parasitic drag
occurring when the lift coefficient was equal to zero.

The NACA 23012 airfoil lift curve slope was obtained from NACA airfoil data and was
calculated to be 0.0744 [1]. The 3D lift curve slope of the finite wing was calculated to be
0.07411 using Eq. 5 and the span efficiency factor calculated previously. The actual lift curve
slop was found to be 0.0743, which was only 2.6% of off the predicted value. The full
calculations are shown in Appendix A.
China clay was used for flow visualization on the wings at different angles of attack, and was
used as an aid to help determine what effect the end caps had on the airflow around the wing tips.
With no end caps on the wing edges, the high-pressure flow on the bottom surface of the wing
flowed up around the wing tips towards the low-pressure air on the top surface, causing vortices.
This flow was observed as seen in Fig. 8.

0.01 0.015 0.02 0.025 0.03 0.035 0.04
0
0.05
0.1
0.15
0.2
0.25
0.3
C
D
C
L 2
Squared Lift Coefficient vs. Drag Coefficient


q = 10 psf
q = 35 psf
With end caps; q = 10 psf
With end caps; q = 35 psf
15

Fig. 8. China clay flow visualization on right wing demonstrating flow on wing tips

From Fig. 8. it can clearly be seen how the flow from under the wing flowed up around the edge
and merged with the flow on top of the wing. These vortices require energy to create and, as a
result, steps are taken to minimize them when designing a wing. One way of minimizing wing tip
vortices is to install end caps on the wing to stop air from flowing around the edge. The china
clay visualization for this technique is shown in Fig. 9.
16

Fig. 9. China clay flow visualization on wing with circular end caps demonstrating flow
on wing edges

From Fig. 9 it is clear that few small vortices were formed at the wing tips, and no large vortices
were formed at all. With the end caps installed, the air flowed straight over the wing edges, thus
saving energy that would normally be transferred into creating the vortices. This demonstrated
how end caps could be used to make a more efficient wing. There must be a balance, however, as
the end caps used in this lab created a considerable amount of drag that would make them
impractical for regular use.
Increasing a wings aspect ratio decreases its induced drag, thus causing the wing to be more
efficient. Once again, however, there must be a balance. The higher a wing aspect ratio is, the
longer the wing becomes. This can create structural strength and vibration problems that are
expensive and difficult to design and create. Therefore, much care must be taken in designing a
wing. The aspect ratio must be low enough that the wing can be fabricated at a reasonable price
and level of safety, but high as possible to maximize the efficiency of the wing. Winglets can
also be added to wing tips to reduce vortices but care must be taken that they are saving more
energy than the extra drag that they are causing.
17
Circular end caps were used in the lab to demonstrate the effect that wing tip vortices had on a
wing. By installing the end caps and eliminating these vortices, the wing essentially became a 2-
D wing, thus increasing the lift curve slope as seen in Fig. 10.









a) b)


Fig. 10. Lift coefficient vs. angle of attack for finite wing and simulated 2-D wing using
circular end caps at a) q = 10 psf and b) q = 35 psf

Circular end caps were used so that the blockage and flow around them remained uniform
independent of angle of attack. Although the circular end caps used in the lab would not be
practical on real aircraft, the theory behind how they eliminate vortices can be used in designing
winglets that minimize vortices and create less drag than the endcaps.
As discussed earlier, Figs. 2-5 parts b) showed the stall region of the lift coefficient vs. angle of
attack. From the plots it was evident that, for the most part, the wing had a higher coefficient of
lift near the stall region during increasing angles of attack than when its angle of attack was
decreasing after stall. This is due to the airflow separating from the wing. When the wing
approaches the stall angle from lower angles of attack, the air is initially attached to the wing and
remains attached until the stall point is reached. When the stalling point is approached from
higher angles of attack, however, the flow is already separated from the wing and continues to
stay separated until a low enough angle of attack is reached that allows the flow to become
attached once again; usually below the stall point. This hysteresis effect is a danger to airplanes
18
that are landing and taking off. If a plane is approaching or departing a runway at a high angle of
attack and the wing begins to stall, the plane will need to lower its angle of attack more than the
stall angle in order for the flow to reattach. This can be a large problem if the plane is near the
ground.

V. Conclusion

For this experiment, it was found that the behavior of the lift, drag, and pitching moment
coefficients aligned with theory for both the finite and infinite wings. Reduced data matched the
corrected data provided by the Kirsten Wind Tunnel with the exception of the pitching moment,
which differed because they were calculated about different points. The behavior of the forces
and moments also depended on which direction the angle of attack was approached from; the
flow separated and did not reattach until after the stall angle was passed when the wing began a
run at a high angle of attack. The finite wing had a smaller lift curve slope than the infinite wing,
as was expected, and the equations discussed in the theory predicted the three-dimensional slope
to within 3% of the measured value. The span efficiency for both wings was found to be larger
than one since wall corrections were not applied, and the parasitic drag was larger for a three-
dimensional wing, which was expected. When circular endcaps were installed, the wingtip
vortices visible on a three dimensional wing disappeared. This experiment could be improved by
decreasing the angle of attack step size around the stall angle, taking data from
o
to
o
in
o

increments. This would result in a smoother lift plot around the stall point of the wing, so that
the stall angle could be more accurately found. Overall, both wing configurations behaved as
expected and aligned with published wind tunnel data within reason.


References
[] Bruckner, , -Finite Wing Lecture- University of shington, Settle,
98195, February 2013





19
Appendix A

% AA 321 - Aerospace Laboratory I
% Lab #7 - Finite Wing
% 02-21-2013

clear all;clc

% Define wing constants
b=90/12; % [ft]
c=15/12; % [ft]
a=c; % [ft]
S=a*b; % [ft^2]
AR=b^2/S;


% Load wind tunnel data
DATAR=xlsread('322_Thurs_Data.xlsx','Uncorrected Data');
wnec10q=xlsread('322_Thurs_Data.xlsx','Wing Only q = 10 psf');
wnec35q=xlsread('322_Thurs_Data.xlsx','Wing Only q = 35 psf');
wec10q=xlsread('322_Thurs_Data.xlsx','Wing with Endcaps q = 10 psf');
wec35q=xlsread('322_Thurs_Data.xlsx','Wing with Endcaps q = 35 psf');
q_w_ec_35q=DATAR(1,5);
CLW=DATAR(18:37,1);
NDRAGR=DATAR(18:37,2); % [LB]

% PART 1

% Balance interaction data
LIFTbc=0.960243*DATAR(1,6)+0.014501*DATAR(1,6)+0.000112*DATAR(1,8);
DRAGbc=-0.000560*DATAR(1,6)+0.993306*DATAR(1,7)-0.000021*DATAR(1,8);
PMbc=0.012984*DATAR(1,6)-0.170853*DATAR(1,7)+0.994981*DATAR(1,8);
CLbc=b/(q_w_ec_35q*S);

% Interpolate to find NDRAG
NDRAG=interp1(NDRAGR(13:end-1,1),CLW(13:end-1,1),CLbc);

% Determine drag and pitching moment
DRAG=DRAGbc-NDRAG*DATAR(1,5);
PMW=PMbc-DATAR(5,2);

% Determine CL, CD, and CM
CL=LIFTbc/(S*q_w_ec_35q);
CD=DRAG/(S*q_w_ec_35q);
CM=PMW/12/(S*q_w_ec_35q*c);
% PART 2

AR=AR;

% PART 3

% Determine CL, CD, CM, L/D, L, D, and M for q = 10 psf
q_nec10q=wnec10q(1:end,4);AOA_nec10q=wnec10q(1:end,6);
20
CL_nec10q=wnec10q(1:end,7);CD_nec10q=wnec10q(1:end,8);
CM_nec10q=wnec10q(1:end,9);L_nec10q=CL_nec10q.*q_nec10q*S;
D_nec10q=CD_nec10q.*q_nec10q*S;M_nec10q=CM_nec10q.*q_nec10q*S;
L_ovr_D_nec10q=L_nec10q./D_nec10q;

% Determine CL, CD, CM, L/D, L, D, and M for q = 35 psf
q_nec35q=wnec35q(1:end,4);AOA_nec35q=wnec35q(1:end,6);
CL_nec35q=wnec35q(1:end,7);CD_nec35q=wnec35q(1:end,8);
CM_nec35q=wnec35q(1:end,9);L_nec35q=CL_nec35q.*q_nec35q*S;
D_nec35q=CD_nec35q.*q_nec35q*S;M_nec35q=CM_nec35q.*q_nec35q*S;
L_ovr_D_nec35q=L_nec35q./D_nec35q;

% Determine CL, CD, CM, L/D, L, D, and M with end caps and q = 10 psf
q_ec10q=wec10q(1:end,4);AOA_ec10q=wec10q(1:end,6);
CL_ec10q=wec10q(1:end,7);CD_ec10q=wec10q(1:end,8);
CM_ec10q=wec10q(1:end,9);L_ec10q=CL_ec10q.*q_ec10q*S;
D_ec10q=CD_ec10q.*q_ec10q*S;M_ec10q=CM_ec10q.*q_ec10q*S;
L_ovr_D_ec10q=L_ec10q./D_ec10q;

% Determine CL, CD, CM, L/D, L, D, and M with end caps and q = 35 psf
q_ec35q=wec35q(1:end,4);AOA_ec35q=wec35q(1:end,6);
CL_ec35q=wec35q(1:end,7);CD_ec35q=wec35q(1:end,8);
CM_ec35q=wec35q(1:end,9);L_ec35q=CL_ec35q.*q_ec35q*S;
D_ec35q=CD_ec10q.*q_ec35q*S;M_ec35q=CM_ec35q.*q_ec35q*S;
L_ovr_D_ec35q=L_ec35q./D_ec35q;

% Create plots
wing_plots_nec10q(AOA_nec10q(1:16),CL_nec10q(1:16),CD_nec10q(1:16),...
AOA_nec10q(1:16),CL_nec10q(1:16),AOA_nec10q(17:end),...
CL_nec10q(17:end),CM_nec10q(1:16),L_ovr_D_nec10q(1:16))
wing_plots_nec35q(AOA_nec35q(1:16),CL_nec35q(1:16),CD_nec35q(1:16),...
AOA_nec35q(1:16),CL_nec35q(1:16),AOA_nec35q(17:end),...
CL_nec35q(17:end),CM_nec35q(1:16),L_ovr_D_nec35q(1:16))
wing_plots_ec10q(AOA_ec10q(1:16),CL_ec10q(1:16),CD_ec10q(1:16),...
AOA_ec10q(1:16),CL_ec10q(1:16),AOA_ec10q(17:end),CL_ec10q(17:end),...
CM_ec10q(1:16),L_ovr_D_ec10q(1:16))
wing_plots_ec35q(AOA_ec35q(1:16),CL_ec35q(1:16),CD_ec35q(1:16),...
AOA_ec35q(1:16),CL_ec35q(1:16),AOA_ec35q(17:end),CL_ec35q(17:end),...
CM_ec35q(1:16),L_ovr_D_ec35q(1:16))

% Part 4

% Determine CL and CD half way between data points
for i=1:15
CL_nec10q_2(i)=(CL_nec10q(i)+CL_nec10q(i+1))/2;
CL_nec35q_2(i)=(CL_nec35q(i)+CL_nec35q(i+1))/2;
CL_ec10q_2(i)=(CL_ec10q(i)+CL_ec10q(i+1))/2;
CL_ec35q_2(i)=(CL_ec35q(i)+CL_ec35q(i+1))/2;

CD_nec10q_2(i)=(CD_nec10q(i)+CD_nec10q(i+1))/2;
CD_nec35q_2(i)=(CD_nec35q(i)+CD_nec35q(i+1))/2;
CD_ec10q_2(i)=(CD_ec10q(i)+CD_ec10q(i+1))/2;
CD_ec35q_2(i)=(CD_ec35q(i)+CD_ec35q(i+1))/2;
end
21

% Determine span efficiency factors and plot
e_nec10q=(diff(CL_nec10q(1:16).^2)./(pi*AR.*diff(CD_nec10q(1:16))))';
e_nec35q=(diff(CL_nec35q(1:16).^2)./(pi*AR.*diff(CD_nec35q(1:16))))';
e_ec10q=(diff(CL_ec10q(1:16).^2)./(pi*AR.*diff(CD_ec10q(1:16))))';
e_ec35q=(diff(CL_ec35q(1:16).^2)./(pi*AR.*diff(CD_ec35q(1:16))))';

span_eff_plot(CL_nec10q_2,e_nec10q,CL_nec35q_2,e_nec35q,CL_ec10q_2,...
e_ec10q,CL_ec35q_2,e_ec35q)

syms CDp_nec10q CDp_nec35q CDp_ec10q CDp_ec35q e_nec10q e_nec35q...
e_ec10q e_ec35q CLx_ec10q CDx_ec10q CLx_ec35q CDx_ec35q CLx_nec10q...
CDx_nec10q CLx_nec35q CDx_nec35q
CD1=CD_nec10q(3:10);CL1=CL_nec10q(3:10);
eq_nec10q=CDp_nec10q+CLx_nec10q^2/(pi*AR*e_nec10q)-CDx_nec10q;
CD2=CD_nec35q(3:10);CL2=CL_nec35q(3:10);
eq_nec35q=CDp_nec35q+CLx_nec35q^2/(pi*AR*e_nec35q)-CDx_nec35q;
CD3=CD_ec10q(3:10);CL3=CL_ec10q(3:10);
eq_ec10q=CDp_ec10q+CLx_ec10q^2/(pi*AR*e_ec10q)-CDx_ec10q;
CD4=CD_ec35q(3:10);CL4=CL_ec35q(3:10);
eq_ec35q=CDp_ec35q+CLx_ec35q^2/(pi*AR*e_ec35q)-CDx_ec35q;

for i=1:length(CD1)
for j=1:length(CL1)
if j==i
break
end
eq11=subs(subs(eq_nec10q,CLx_nec10q,CL1(i)),CDx_nec10q,CD1(i));
eq12=subs(subs(eq_nec10q,CLx_nec10q,CL1(j)),CDx_nec10q,CD1(j));
eq21=subs(subs(eq_nec35q,CLx_nec35q,CL2(i)),CDx_nec35q,CD2(i));
eq22=subs(subs(eq_nec35q,CLx_nec35q,CL2(j)),CDx_nec35q,CD2(j));
eq31=subs(subs(eq_ec10q,CLx_ec10q,CL3(i)),CDx_ec10q,CD3(i));
eq32=subs(subs(eq_ec10q,CLx_ec10q,CL3(j)),CDx_ec10q,CD3(j));
eq41=subs(subs(eq_ec35q,CLx_ec35q,CL4(i)),CDx_ec35q,CD4(i));
eq42=subs(subs(eq_ec35q,CLx_ec35q,CL4(j)),CDx_ec35q,CD4(j));

A=solve(eq11, eq12, CDp_nec10q, e_nec10q);
B=solve(eq21, eq22, CDp_nec35q, e_nec35q);
C=solve(eq31, eq32, CDp_ec10q, e_ec10q);
D=solve(eq41, eq42, CDp_ec35q, e_ec35q);

CDp_nec10q_calc((i-1)*length(CL1)+j)=vpa(A.CDp_nec10q,4);
e_nec10q_calc((i-1)*length(CL1)+j)=vpa(A.e_nec10q,4);
CDp_nec35q_calc((i-1)*length(CL2)+j)=vpa(B.CDp_nec35q,4);
e_nec35q_calc((i-1)*length(CL2)+j)=vpa(B.e_nec35q,4);
CDp_ec10q_calc((i-1)*length(CL3)+j)=vpa(C.CDp_ec10q,4);
e_ec10q_calc((i-1)*length(CL3)+j)=vpa(C.e_ec10q,4);
CDp_ec35q_calc((i-1)*length(CL4)+j)=vpa(D.CDp_ec35q,4);
e_ec35q_calc((i-1)*length(CL4)+j)=vpa(D.e_ec35q,4);
end

end
CDp_nec10q_calc=CDp_nec10q_calc(CDp_nec10q_calc~=0);
e_nec10q_calc=e_nec10q_calc(e_nec10q_calc~=0);
22
CDp_nec35q_calc=CDp_nec35q_calc(CDp_nec35q_calc~=0);
e_nec35q_calc=e_nec35q_calc(e_nec35q_calc~=0);
CDp_ec10q_calc=CDp_ec10q_calc(CDp_ec10q_calc~=0);
e_ec10q_calc=e_ec10q_calc(e_ec10q_calc~=0);
CDp_ec35q_calc=CDp_ec35q_calc(CDp_ec35q_calc~=0);
e_ec35q_calc=e_ec35q_calc(e_ec35q_calc~=0);

% Determine average span efficiency factor
e_nec=[e_nec10q_calc e_nec35q_calc];
e_ec=[e_ec10q_calc e_ec35q_calc];
e_nec=vpa(mean(e_nec),4);
e_ec=vpa(mean(e_ec),4);

% Determine average parasitic drag
CDp_nec=[CDp_nec10q_calc,CDp_nec35q_calc];
CDp_ec=[CDp_ec10q_calc,CDp_ec35q_calc];
CDp_nec=vpa(mean(CDp_nec),4);
CDp_ec=vpa(mean(CDp_ec),4);


% Determine parasitic drag from CL^2 vs. CD plots
Paras_drag_CU(CD_nec10q(2:16),CL_nec10q(2:16).^2,CD_nec35q(2:16),...
CL_nec35q(2:16).^2,CD_ec10q(2:16),CL_ec10q(2:16).^2,CD_ec35q(2:16),...
CL_ec35q(2:16).^2)

% Part 5

% Determine lift slopes of 2-D and 3-D wings
a_nec10q=(CL_nec10q(10)-CL_nec10q(1))/(AOA_nec10q(10)-AOA_nec10q(1))
a_nec35q=(CL_nec35q(10)-CL_nec35q(1))/(AOA_nec35q(10)-AOA_nec35q(1))
a_ec10q=(CL_ec10q(10)-CL_ec10q(1))/(AOA_ec10q(10)-AOA_ec10q(1));
a_ec35q=(CL_ec35q(10)-CL_ec35q(1))/(AOA_ec35q(10)-AOA_ec35q(1));

% Determine NACA 23012 airfoil lift curve slope
a0=((0.97)-(-.22))/((12)-(-4))

% Calculate 3-D lift curve slope
a=vpa(a0/(1+a0/(pi*e_nec*AR)),4)

% Plot 2D vs 3D wing
ec_vs_nec_10q(AOA_nec10q(1:16),CL_nec10q(1:16),AOA_ec10q(1:16),CL_ec10q...
(1:16));
ec_vs_nec_35q(AOA_nec35q(1:16),CL_nec35q(1:16),AOA_ec35q(1:16),CL_ec35q...
(1:16));

Você também pode gostar