Homework 1:

NACA 4-digit airfoils analysis using the Discrete Vortex Method

Vı́ctor Barrachina, Younes el Jarari, Gerard Gallardo, Guillem Olivella, Enric Royo
Universitat Politècnica de Catalunya, ESEIAAT, Technical University, Colom 1 Street, 08222 Terrassa, Barcelona

Delivered 29 December 2017

Aerodynamics - 220024
Professor Enrique Ortega

ABSTRACT: The results shown are calculated under the theoretical results and they show a clear parallelism to the
assumptions of the thin airfoil theory (TAT). During this theoretical results. In both curves (CL and CM,0 ) the slopes
study, the characteristics related to a given airfoil (NACA coincide with the expected by the TAT. All the calculus and
2408) are going to be analysed, in concrete, load calculations algorithms done in this project are going to be based on the
and the flap efficiency trend as a function of the flap-chord ra- Discrete Vorthex Method [1].
tio and its deflection. It has been observed that experimental
Key words: Thin airfoil theory (TAT), Discrete vortex
results resemble the computed data in almost all the cases.
method (DVM), Lift coefficient (CL ), Moment coefficient at
The characteristic curves of an airfoil have been compared to
the Leading Edge (CM,le )

INTRODUCTION 4-digits airfoils have good stall characteristics and high-speed

performance. However, the maximum lift is relatively low
An airfoil is defined as the shape of a wing that produces and the pitching moment and drag are elevated. [2] [3]
an aerodynamic force (lift and drag) when moved through a As the NACA airfoil description is not the topic of this
fluid. The NACA1 airfoils are normalized shapes for aircraft assignment, we will not go deeper. If interested, check the
wings developed by the National Advisory Committee for following bibliography for more accurate descriptions. [3]
Aeronautics. They follow a numerical code that determines
In terms of the project, this assignment is divided in
their shape and properties. In the following figure, all the
three distinguished parts: Verification assessment, Val-
airfoils properties are represented.
idation assessment and Discussion. Furthermore, the fol-
lowing study has been done using the Discrete Vortex Method
(DVM). This method is based in the the discretization of
the camber line in consecutive flat panels or lumped-vortex
elements.
Using this scheme, a computational modelling study can
be conducted in order to compute the effects that flying con-
ditions have to the loads. This scheme also let us compute
the effects of the flap by modifying the camber line, so we will
also be able to study its effects to the loads and, moreover, its
efficiency. Hence, for the given airfoil (NACA 2408) all of
this calculations are going to be computed with the support
of the software MatLab ®.
Then our calculations are verified by comparing them to
Figure 1: Profile geometry the theoretical results; common aerodynamic curves (such as
The most basic and common NACA airfoils (and also the CL vs α or CM,le vs α) will be provided. This part is focused
ones used here) are the Four-digit series. Those shapes, as mainly on the slopes of this curves, knowing that the linear
we advanced, are described by four digits: the first and the of the CL plot must be around 2π and 0 for the CM,le . This
second are referred to the camber line and represent maxi- results are deeply reviewed latter in their own sections.
mum value of camber and its position along the x-axis, both Eventually, a discussion about the results obtained and
expressed in % of chord, and the third and fourth terms rep- its matches with the theoretical onesis set.In the annexes the
resent the maximum thickness in % of chord too. NACA full MatLab®script is provide ready to compile.

1 NACA: National Advisory Committee for Aeronautics

1
 0.67 9 -0.205 6
Cm
8 Theoretical Cm
0.66 Error 5
7
0.65
6 -0.21 4

0.64 5
Cl
Theoretical Cl 3
Error 4
0.63

3 -0.215 2
0.62
2
1
0.61
1

0.6 0 -0.22 0
50 100 150 200 250 50 100 150 200 250

Figure 2: CL & error vs Num. of panels Figure 3: CM,le & error vs Num. of panels

VERIFICATION ASSESSMENT The routine for obtaining the CM,le is exacly the same,
but instead of eq. (1), the following expression has been used:
In this section, we proceed to calculate the parameters in-
troduced previously in order to verify the DVM’s solution. CL π
CM,le = − + (A2 − A1 ) (3)
In this purpose, we set the flying conditions as told in the 4 4
assignment (NACA: 2408, α: 4º) where A1 and A2 are the coefficients of a Fourier’s trigono-
Then, reference exact values are calculated under the as- metrical expansion. They can be obtained from the equation
sumptions of the TAT (see Appendix). Thus we proceed to shown below:
2 π dz
Z
make a Sweep for the number of panels (in this case from 4 to
Ai = cos(iθ0 )dθ0 for i ∈ N+ (4)
250 panels), in order to obtain the CL in function of the num- π 0 dx
ber of panels. As it is supposed, the curve obtained enhance fractioning, again, the integral in the two same parts as did
its precision as we increase the number of panels. for the theoretical CL . The following results have been ob-
But the last part is yet to be determined for now. As tained from operating as described:
it can be seen in the Figure 2, the CL goes to a asymptotic
value, but we can not talk about precision if theoretical trend Theoretical Parameters
of CL is not known yet. Is at this point where the TAT starts CL 0.6667 CM,le -0.2197
having huge importance.
Once obtained this theoretical results, our MatLab
From the equations of the TAT, we can state:
®script have been executed (see Appendix), taking this the-
1 π dz
 Z 
CL = 2π α + (cos θ0 − 1)dθ0 (1) oretical values as the reference when computing the error.
π 0 dx This results are shown at the top of this page (see Figures
Where term dz/dx represents the slope of the camber line 2 and 3). In this curves we can see quite information, and
and the angle (θ0 ) and the slope are related by the following the following description is valid for both curves: Continuous
expression: traces are for computed results and slashed ones for theoret-
x0 = c/2(1 − cos θ0 ) (2) ical, the left y-axis correspond to the blue lines (CL ) and the
right y-axis gives information about the error in %.
Hence we proceed to calculate the theoretical value of CL As can be seen, the results match with a very high pre-
as it is shown in the equation (1). In order to obtain cor- cision the assumptions we made, obtaining in both cases an
rect results some particular things must be present. As we error lower than 0.1% for 250 panels. Also, as predicted, the
know for the theory lectures (see [3]), NACA Four-digit se- precision increases with the number of samples used (the er-
ries camber line’s are constructed by two parabolic curves ror decreases). The tendency in both cases is the same as
that meet at the point where the camber is maxim. Due the hyperbole, having a greater error for a very low num-
to that, the previous integral is divided in two parts, and is ber of panels (it tends to infinity when approaching 0) and
easily predictable that both slopes (dz/dx) are meant to be approaching asymptotically to a constant value (the theoreti-
different. First comes from zero to 1.369 rad and second goes cal parameter) when increasing the number of lumped-vortex
from 1.369 rad to π. This points have been obtained from the elements.
NACA (x0 ) and computing x0 in (2). This data is resumed As it can be seen in Figure 2 and 3, the error commit-
in the following table for our particular case: ted when working with 100 panels is less than 0.2 % in both
x0 0 0.4 c cases, which has been considered small enough to be accepted
θ0 0 1.369 π as the quantity of panels used for the following results.

2
 Angle of delflection vs. Flap efficiency factor
0.9

0.8

0.7

0.6

0.5

Flap-chord ratio = 0.5

Flap-chord ratio = 0.45
0.4 Flap-chord ratio = 0.4
Flap-chord ratio = 0.35
Flap-chord ratio = 0.3
Figure 5: Flap characteristics
Flap-chord ratio = 0.25
0.3 Flap-chord ratio = 0.20
Flap-chord ratio = 0.15
Flap-chord ratio = 0.10
Flap-chord ratio = 0.05 E: Flap-chord ratio = 1-xh /c; tan η: Flap deflection
0.2
2 4 6 8 10 12 14 16 18 20

Note: For non-dimensional problem with relatively low flap

deflections, we will assume that Flap-chord ratio = 1-E and
Figure 4: Flap-chord ratio vs efficiency tan η ≈ η.

Flap effects deflection is produced. This deflection also produces a Lift

and a Pitching moment. So, if previously a certain α is given
This second part of the validation assessment pretends to in-
(we considered zero for simplifying) where the lift generated
troduce the effects that flaps produce in our airfoil. As we
is zero, this angle αl0 is displaced anticlockwise.
introduced, the effect of the flap can be easily computed due
Repeating the process described, a certain function will
to the TAT. The advantages of having an airfoil with rela-
be generated that shows how the zero lift angle (αl0 ) changes
tively low thickness allows the modelling of the present air-
in function of η. Now, we just have to differentiate respect η,
craft by introducing some vortex filaments perpendicular to
so we will obtain how αl0 changes with η, the flap efficiency.
the aircraft that will generate a vortex sheet. Every vortex
Now, we have to keep in mind that in this whole process we
filament has a constant circulation, but most important they
have assumed a certain Flap-chord ratio deflected. If we re-
are placed in the camber line.
peat this process for different Flap-chord ratio, we will get
Because of that, using TAT we are giving main impor-
the flap efficiency in function of the deflection for different
tance to the camber line, because all the information given
flap-chord ratio.
by the fly conditions will be represented there. Using the
camber line as the representation of our airfoil gives a simple All this routine have been implemented in MatLab ®,
and easy way to model a flap. Considering that the flap is again for the NACA and α given, and Figure 4 have been
just a portion of the aircraft that can be deflected (upper or obtained.
downwards), we will be able to shape it by giving an extra In Figure 4, we can see many interesting things. First
camber to the aircraft in some determined points. of all, the curves show that the slopes are constant and de-
In order to decide which concrete points are these, it is pends only on the flap-chord ratio. We can also observe that
needed to define the flap geometrically and it becomes easier those slopes increases with flap-chord ratio, going to zero for
using the support of the Figure 5. There can be spotted many very small flap-chord ratios. This result could be easily ex-
parameters, but only two of them are needed in order to get pected, so when the portion deflected tends to zero is easy
all the information, and these two are shown at the bottom to see that its efficiency would go to zero (it does not exist).
of this figure (E and η). On the other hand, we can see that the efficiency increases
As explained, and it is easy to see in the Figure 5, the por- also with the flap-chord ratio. Another information that this
tion of the aircraft deflected is called Flap-chord ratio (E), and figure proves is that the main parameter when designing a
its deflection is measured by η (according to the note under flap is the flap-chord ratio, and it can be seen by observing
the figure). the values of efficiency for a constant flap-chord ratio and
Once that the flaped airfoil is characterised, it is time to for a constant deflection. Focusing in the effect of the de-
start working on its properties. To perform so, there is a pa- flection, for a certain flap-chord ratio the values of efficiency
rameter which is basic in the characterisation of the of the oscillates between 0 and 0.05 (for flap-chord ratios of 0.05 and
properties that also has a direct effect on the loads. This 0.5 respectively), however, for a given deflection the values of
parameter is called flap efficiency and it is written ∂αl0 /∂η. efficiency are bounded obtaining ranges between 4.5 or 5 (for
In order to explain the physical interpretation it is con- deflections of 1 and 20 respectively).
sidered a plain airfoil moving through a fluid at some relative In short, the range of values we can control by focusing
velocity that in this case is assumed perpendicular to the air- on the flap-chord ratio is the order of 10 times higher that the
foil2 . Now, lets consider that a portion of this airfoil starts range we would get by controlling the deflections. So that,
deflecting itself η degrees, so it produces a certain camber. we can state that the main parameter in flap designs is the
This camber induces some circulation that increases as the flap-chord ratio.
2 The total contributions to the loads will be computed applying the superposition principle, adding the contribution of the flap to

the other contributions (see [4]).

3
 2 0.9
Computed
Computed and corrected
0.8 Experimental

1.5
0.7

0.6

1 X: 0.08727
Y: 0.7724 0.5
Cm 0
Cl
0.4
0.5
X: 0
Y: 0.2256 0.3

0.2
0

0.1

-0.5 0
-0.1 -0.05 0 0.05 0.1 0.15 0.2 0.25 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5

Figure 6: CL & error vs Num. of panels Figure 7: CM,le & error vs Num. of panels

VALIDATION ASSESSMENT Eventually we want to prove the near perfect match that
our computed results have with the expected by the TAT.
In this part of the project, the main purpose is to check if the Therefore an extra plot has been added. It shows the depen-
results obtained match with the expected from the TAT. In dence on the angle of attack of two of the most characteristic
order to do that, a table with the results obtained from Mat- parameters of the airfoil, the CL and the Cm0 (see Figure
Lab ®and the experimental ones obtained from “Abbast and 6). Asit was expected and advanced previously, the moment
Doenhoff ” will be shown bellow. The experimental results coefficient does not depend on the angle of attack and the
have been tested at a very high Reynolds in order to avoid slope of the CL vs α is 6.27 ≈ 2π(note that the units of the
any interference from compressible and viscous effects. [5] x-axis are radians. In case of using degrees, this slop must be
≈0.1 deg−1 ).
Experimental and computed data
Exp. data Comp. data
αL0 -2º -2,06º Flap effects
Cm0 -0.05º -0.0546º In order to obtain some conclusions regarding the influence of
CL,α 0.1 deg−1 0.1087 deg−1 the flap-chord ratio in the efficiency of a flap, a plot display-
ing the flap efficiency factor vs. the flap-chord ratio has been
As it can be seen, the experimental results are practically computed (Figure 7), where a flap deflection of 0-20º has been
the same as the computed ones. It can aslo be stated that the used. In the same graph, experimental data obtained from
precision of the computing results is higher than the experi- Abbast and Doenhoff is shown. The computed calculations
mental shown in the bibliography. But how have this results of the flap efficiency have also been modified to obtain the
been computed? Regarding the computed calculations, the corrected flap efficiency factor, a correction that consists on
characteristic parameters of the airfoil have been calculated doing the following calculation:
following these equations and procedures:
   
CL,α is the slope of the plotted line in the Cl vs α dia- ∂αl 0 ∂αl 0
= · f actor (5)
gram, and it has been computed as a regression line of the ∂η C ∂η T
different Cl calculated for different angles.
Clα0
Zero-lift angle of attack (αL0 :) αL0 = − C where the factor used is 0.7
L,α

L,α C
Aerodynamic centre (Xac ): Xac = − CM,LE Analysing the suitability of the corrected flap efficiency
factor, some conclusions can be drawn: For small flap-chord
Cm about Xac (Cm0): Cm0 = Cm,LE + Xac · Cl
ratios (0 - 0.15), the experimental data resembles the cor-
Where: rected computed flap efficiency factor. There is a transition
Clα0 is the Cl with α = 0,CL,α is the slope of the Cl area (0.15 - 0.3) where the experimental data can be found
graph, CM,LE is the slope of the Cm graph, CmLE is the mo- between the non-corrected and the corrected trends. Finally,
ment with respect to the LE, Xac is respect the aerodynamic for big flap-chord ratios (0.3-0.45) the experimental data is
centre and Cl is the lift coefficient. better represented with the non-corrected calculations.

4
0 0
Xmc = 10% Xmc = 10%
-1 Xmc = 20% Xmc = 20%
Xmc = 30% -0.05 Xmc = 30%
-2 Xmc = 40% Xmc = 40%
Xmc = 50% Xmc = 50%
Xmc = 60% Xmc = 60%
-3
-0.1

-4

-0.15
-5

-6
-0.2
-7

-8 -0.25
0 1 2 3 4 5 6 0 1 2 3 4 5 6

0 0
MC = 0% MC = 0%
-1 MC = 1% MC = 1%
MC = 2% -0.05 MC = 2%
-2 MC = 3% MC = 3%
MC = 4% MC = 4%
MC = 5% MC = 5%
-3
MC = 6% -0.1 MC = 6%

-4

-0.15
-5

-6
-0.2
-7

-8 -0.25
10 20 30 40 50 60 10 20 30 40 50 60

Figure 8: αl0 and Cm0 as function of f and p.

DISCUSSION of Figure 8).

Referent to the αl0 , in Figure 8a it can be seen that the
Using the same initial conditions and the same criteria (100
zero-lift angle decreases in almost linear trend with the maxi-
panels) a study of how does the maximum camber and its
mum camber. It can be interpreted as a displacement through
position affect the zero-lift angle and the Moment Coefficient
the left of the line that defines Cl vs α. This means that for
about the Aerodynamic Centre will be carried out.
a given angle of attack and f, the Cl increases with the cam-
First of all the effects of the maximum camber for its ber (always in the range admitted by the TAT). However,
given relative position will be studied. Then the maximum this would have effect in the Drag, although it has not been
camber will be fixed and how the αl0 and Cm0 changes with computed. This effect is increased when the position of the
the position of the maximum camber will be stuided. maximum camber is being moved backwards.

Effects of the Maximum camber For the Coefficient of Moment about Aerodynamic Cen-
tre (see Figure 8b), the same explanation about the diagram
The behaviour of αl0 and CM 0 for a given f (position of max- can be applied, although it has different consequences. First
imum camber) is represented in Figures 8a and 8b (first row of all, we have to assume that a negative moment means that

5
the trend is to rotate clock-wise. Hence, a relatively big Cm0 (see Figure 8d), the same explanation about the diagram can
will make the airfoil rotate by increasing the angle of attack be applied, although with different consequences. Following
significantly. It would contribute heavily on the Drag and, for the same condition that we saw for Figure 8b, the negative
relatively big angles of attack, the boundary layer could even moments only affect at the sense of rotation. In general terms,
come off losing a huge part of lift. As in the previous case, the effect on the moment tends to increase when increasing p
the effect increases when moving the position of the maximum (moving the maximum camber backwards). As with the αl0
camber is being moved backwards. for fixed f, the effect on the Cm0 is also intensified for bigger
maximum cambers, being almost independent of p for a small
Effects of the position of Maximum Camber camber (same explanation can be applied too).

The behaviour of αl0 and CM 0 for a given p (maximum cam-

ber) is plotted in Figures 8c and 8d (second row). References
Referent to the αl0 , in Figure 8c similar results to the
[1] Ortega. Enrique, Discrete Vortex Mothod, Lecture
case of the fixed position of the maximum camber can be seen.
notes and support, UPC (ESEIAAT), 2017.
Nontheless it has some remarkable differences, specially when
moving through small maximum cambers. As expected, for [2] Wikipedia, NACA airfoil, consulted on December 05
little maximum cambers the αl0 is independent of the posi- 2017 [online].
tion of its maximum camber. Despite this particular case, the https://en.wikipedia.org/wiki/NACA_airfoil
main trend is to reduce the αl0 as the position of the max- [3] Ortega. Enrique, Airfoils, Lecture notes and support,
imum camber is being moved backwards, having the same UPC (ESEIAAT), 2017.
effects on the lift and drag as explained for the αl0 behaviour
with fixed p. Closing, as we introduced for small maximum [4] Ortega. Enrique, Thin Airfoil Theory, Lecture notes
chambers, this effect is intensified by increasing the maximum and support, UPC (ESEIAAT), 2017.
camber. [5] Abbast and Doenhoff, Theory of wing sections, 1959,
For the Coefficient of Moment about Aerodynamic Centre Appendix IV, page 474.

Signed, the group

6
Appendix: MatLab ® SCRIPT

1 function [AngleCl0, Cm0, LiftC , MomentC0, Angle] = Analisis (xw,flapangle,MC, Xmc, M)

2 %%Study of CL
3 for i = 1:20
4 alpha = −5 + (i−1);
5 LiftC (i) = DiscreteVortexMethod(alpha,xw,flapangle,MC, Xmc, M);
6 Angle (i) = alpha;
7 end
8 [r,Clalpha,ClAngle0] = regression(Angle,LiftC)
9
10 % Determination of AngleCl0
11 AngleCl0 = − ClAngle0/Clalpha;
12
13 %%Study of Cm
14 alpha = 0;
15 for i = 1:20
16 alpha = −5 + (i−1);
17 [Cl,Cm] = DiscreteVortexMethod (alpha,xw,flapangle,MC, Xmc, M);
18 MomentC(i) = Cm;
19 Angle (i) = alpha;
20
21 end
22 [r,Cmalpha,MomementC0] = regression(Angle,MomentC);
23 xac = − Cmalpha / Clalpha
24
25 %Determinaction of Cm0
26 for i = 1:20
27 alpha = −5 + (i−1);
28 [Cl,Cm] = DiscreteVortexMethod (alpha,xw,flapangle,MC, Xmc, M);
29 MomentC0(i) = MomentC(i) + xac * LiftC(i);
30 Angle (i) = alpha/180*pi;
31 end
32 Cm0 = mean (MomentC0);
33 end

1 function [Cl,Cm] = DiscreteVortexMethod (alpha,xw,flapangle,MC, Xmc, M)

2 %−−−−− Declaration of variables −−−−
3 %MC corresponds to Max camber in %of chord
4 %Xmc corresponds to Position of Max Camber in %of chord
5 %ecorresponds to Max. Thickness in %of chord
6 %M corresponds to n? of panels
7 %xw corresponds to flap−chord ratio
8 %flapangle corresponds to flap angle
9
10 %−−−−−−−PROBLEM ADIMENSIONALIZATION−−−−−−−
11 c=1; %Chord
12 Uinf=1; %Non−disturbed relative velocity
13 N=M+1; %Numer of x
14
15 %−−−−−−−− Computation of camberline −−−−−−
16 COORD=ones(N,2); %Matrix with the coordinates of points of the camber line
17 for k = 1:N
18 xi = 1/2*(1−cos((k−1)/(N−1)*pi));
19 if xi < Xmc/10
20 zi = (MC/100) / (Xmc/10)ˆ2 * (2 * (Xmc/10) * xi − xiˆ2);
21 else
22 zi = (MC/100) / (1−(Xmc/10))ˆ2 * (1 − 2 * (Xmc/10) +...
23 2 * (Xmc/10) * xi − xiˆ2);
24 end
25 COORD(k,1)=xi;
26 COORD(k,2)=zi;
27 end
28
29 %Determination of zw
30 zw = 0;
31 for k = 1:(N−1)
32 if ((COORD(k,1)<xw) && (COORD(k+1,1)≥xw))

7
 33 zw = COORD (k,2)+(COORD (k+1,2)−COORD (k,2))/...
34 (COORD (k+1,1)−COORD (k,1))*(xw−COORD (k,1));
35 end
36 end
37 for k = 1:N
38 if (COORD(k,1)>xw)
39 %Determination of r
40 flap (k,1) = sqrt((COORD (k,1) − xw)ˆ2 + ((COORD (k,2)) − zw)ˆ2);
41 %Determination of initial angle
42 flap (k,2) = asin((COORD (k,1) − xw)/flap (k,1));
43 end
44 end
45 for k = 1:N
46 if (COORD(k,1)>xw)
47 %New COORD after applying flap deflection
48 COORD (k,1) = xw + flap (k,1)*sin (flap (k,2)−flapangle/180*pi);
49 COORD (k,2) = zw − flap (k,1)*cos (flap (k,2)−flapangle/180*pi);
50 end
51 end
52
53 %−−−−− Computation of normal & tangent vector and Cp & CV coordinates −−−−−
54 NORMALS = ones(M,2); %Normal vectors in components (nx,ny) of each panel
55 CPS = ones (M,2); %Coordinates of Control Point of each panel
56 VS = ones(M,2); %Coordinates of the vortex point of each panel
57
58 for k = 1:M
59 dxi = COORD (k+1, 1) − COORD (k, 1);
60 dzi = COORD (k+1, 2) − COORD (k, 2);
61 ci = sqrt (dxiˆ2 + dziˆ2);
62
63 NORMALS (k,1) = −dzi / ci;
64 NORMALS (k,2) = dxi / ci;
65
66 CPS (k,1) = COORD (k,1) + 3/4 * dxi;
67 CPS (k,2) = COORD (k,2) + 3/4 * dzi;
68
69 VS (k,1) = COORD (k,1) + 1/4 * dxi;
70 VS (k,2) = COORD (k,2) + 1/4 * dzi;
71
72 end
73
74 A = ones(M,M); %Influence coeficients Aij
75 RHS= ones (M,1); %Right hand side matrix
76 GAMMA= ones (M,1); %Circularion distribution
77
78 for k1=1:M %loop over the control points
79
80 xi = CPS(k1,1);
81 zi = CPS(k1,2);
82 nxi = NORMALS(k1,1);
83 nzi = NORMALS(k1,2);
84
85 for k2=1:M %loop over the panels
86
87 x0i=VS(k2,1);
88 z0i=VS(k2,2);
89
90 r2=(xi−x0i)ˆ2+(zi−z0i)ˆ2;
91 u=1/(2*pi)*(zi−z0i)/r2;
92 w=−1/(2*pi)*(xi−x0i)/r2;
93 A(k1,k2)=u*nxi+w*nzi;
94 end
95
96 RHS(k1,1)=−Uinf*(cos(alpha/ 180 * pi)*nxi+sin(alpha/ 180 * pi)*nzi);
97 end
98 %−−−− Loads Calculations −−−−
99 CIRCULATIONS = inv(A) * RHS;
100 Cl = 2 * ones (1,M) * CIRCULATIONS;
101 Cm = − 2 * VS (:,1).' * CIRCULATIONS * cos (alpha / 180 * pi);
102 end