Application of PARSEC geometry representation to general airfoil for aerodynamic

optimization

U. Selvakumar R. Pandiyarajan S. Ganesh prasad

Department of Aeronautics Department of Aeronautics Department of Aeronautics
Noorul Islam College of Engg Noorul Islam College of Engg Noorul Islam College of Engg
Kumaracoil,India Kumaracoil,India Kumaracoil,India
e-mail: flyspan08@gmail.com e-mail: pandiya@indianspace.in e-mail: ganesh@gmail.com

Feb 25, 2009

Abstract

Generally in order to describe general airfoil we need to have atleast hun-

dred points of x and y co-ordinates. It is really difficult to optimize airfoils
with this large number of co-ordinates. Nowadays many different schemes
of parameter sets are used to describe general airfoil such as B-spline, Hicks-
Henne Bump function, PARSEC etc. The main goal of these methods is to
reduce the number of needed parameters as few as possible while control-
ling the important aerodynamic features effectively. Here the work has been
done on the PARSEC geometry representation method. The objective of
this project work is to introduce the knowledge of describing general airfoil
using twelve parameters by representing its shape as a polynomial function.
And also we have introduced the concept of genetic algorithm to optimize
the aerodynamic characteristics of a general airfoil for a specific codtions. A
fortran program has been developed to implement PARSEC, surface panel
theory and genetic algorithm. This program has been tested for a standard
airfoil NACA 2411.

0
1 PARSEC 1

1 Parsec
1.1 Introduction
Three decades ago the reporting of analytical results for airfoil shapes and their pres-
sure distribution in high speed flow were welcomed by developers of early numerical
algorithms to simulate inviscid flow. A key element of usefulness was the hodograph
formulation which allowed to see the model problem in complete symmetry between
geometry and variables of state, i. e. the velocity distribution surrounding the bound-
ary condition. Wellknown case studies in closed analytical formulation still serve as
test cases for CFD code development and the understanding of detailed aerodynamic
phenomena in the transonic regime.

The supersonic regime is characterized by its downstream influence of geometry

parameters, as hyperbolic potential theory teaches us. This results in suitably defined
wing or cross sections, contrary to subsonic design techniques derived from classical
mapping theory supported by elliptic potential equations, where all details of a bound-
ary influence the flow at any point. These well known facts and other mathematical
characteristics have strong influence on suitable and practical geometry definition,
manifest through various typical parameter choices to calibrate airfoils or body cross
sections.

Based on mathematical structures of analytical results for the above mentioned

model flow equations, a number of key parameters have been defined which not only
drive aerodynamic performance very effectively, but also are simple technical details
of the resulting airfoil. The variability of these PARametric SECtion airfoil families
to be generated is huge, as has been found by many users. Basic sets with only 11
parameters allow already for many useful airfoil shapes. With various aerodynamic re-
finements necessitated by design projects, we have added additional parameter called
lower leading edge radius for leading edge control, resulting in 12 parameters total.

A PARSEC airfoil geometry which is defined by twelve parameters is shown in the

following figure.They are highly effective in controlling the aerodynamic caracteristics
of a general airfoil. The twelve parameters are tabulated here below.
1 PARSEC 2

No Parameter Notation
01 Upper leading edge radius (Rleu )
02 Lower leading edge radius (Rlel )
03 Upper crest point (Yup )
04 Position of upper crest along chord (Xup )
05 Upper crest curavature (Y XXup )
06 Lower crest point (Ylo )
07 Position of lower crest (Xlo )
08 Lower crest curvature (Y XXlo )
09 Trailing edge direction angle (T E )
10 Trailing edge wedge angle (T E )
11 Trailing edge thickness (TT E )
12 Trailing edge offset (Tof f )
The twelve parmeters has been indicated here in the following figure. The mathe-
matical formulation for PARSEC is given by the ploynomial,

6
X
y= ai xi(1/2) (1)
i=1

Figure 1: Demonstration of PARSEC

1 PARSEC 3

It can be written for the upper and lower surfaces respectively as follows.
6
X
yu = ai xi(1/2) (2)
i=1

6
X
yl = bi xi(1/2) (3)
i=1

Here,
1. yu is the required y co-ordinate of the upper surface
2. yl is the required y co-ordinate of the lower surface
3. x is the chordwise location. The chord length is assumed to be 1m
4. ai and bi are coefficients to be solved

1.2 Evaluation of coefficients

1.2.1 Upper surface

The expanded plolynomial for the upper surface of general airfoil can be written as
follows. The diiferentiation of this equation with respect to x gives the second equa-
tion. Again differentiating this equation w.r.t x gives the third equation as given
below.

1. yup = a1 (xup )1/2 + a2 (xup )3/2 + a3 (xup )5/2 + a4 (xup )7/2 + a5 (xup )9/2 + a6 (xup )11/2

2. dyup/dx = (1/2)a1 (xup )1/2 +(3/2)a2 (xup )1/2 +(5/2)a3 (xup )3/2 +(7/2)a4 (xup )5/2 +
(9/2)a5(xup )7/2 +
(11/2)a6(xup )9/2

3. d2 yup /dx2 = (1/4)a1 (xup )3/2 +(3/4)a2 (xup )1/2 +(15/4)a3(xup )1/2 +(35/4)a4 (xup )3/2 +
(63/4)a5
(xup )5/2 + (99/4)a6(xup )7/2

The coefficient a1 = 2Rle has been obtained by assuming x as a small value.
Now we have to apply the following conditions,
1. At xup = maximum, yup = maximum.
2. At xup = maximum, dyup/dx = 0.
1 PARSEC 4

3. At xup = maximum, d2 yup /dx2 = maximum.

4. At xup = 1, yup = Tof f + TT E /2
5. At xup = 1, dyup/dx = tan(T E T E /2)
1

Now we will get the following equations for each of the above conditions respec-
tively.
1. yup a1 (xup )1/2 = a2 (xup )3/2 + a3 (xup )5/2 + a4 (xup )7/2 + a5 (xup )9/2 + a6 (xup )11/2

2. (1/2)a1 (xup )1/2 = (3/2)a2 (xup )1/2 +(5/2)a3(xup )3/2 +(7/2)a4 (xup )5/2 +(9/2)a5(xup )7/2 +

(11/2)a6(xup )9/2

3. (d2 y/dx2)up +(1/4)a1(xup )3/2 = (3/4)a2 (xup )1/2 +(15/4)a3(xup )1/2 +(35/4)a4 (xup )3/2 +

(63/4)a5(xup )5/2 + (99/4)a6 (xup )7/2

4. Tof f + TT E /2 a1 (xup )1/2 = a2 (xup)3/2 + a3 (xup )5/2 + a4 (xup )7/2 + a5 (xup )9/2 +
a6 (xup )11/2

5. tan(T E T E /2) + (1/4a1 (xup )3/2 = (3/4)a2 (xup )1/2 + (15/4)a3 (xup )1/2 +
(35/4)a4(xup )3/2 +
(63/4)a5(xup )5/2 + (99/4)a6 (xup )7/2

These equations has been reformed as a matrix as shown below. This matrix has
been solved by using the method of Gauss-Elimination. The resultant values will give
the required five coefficient values. Finally by substituting the values of coefficients
on equation (2) the required co-ordinates for the upper surface of general airfoil has
been obtained.

A11 A12 A13 A14 A15

a2
B(1)

A21 A22 A23 A24 A25 a3 B(2)







A31 A32 A33 A34 A35 a4 = B(3)



A41 A42 A43 A44 A45






a5






B(4)



A A52 A53 A54 A55 a6 B(5)
51

1 This will be treated as negative since the upper surface slope is keep on reducing from the maximum point(Yup ).
1 PARSEC 5

Here,
p
a1 = (2 RLE )
A11 = (xup )3/2
A12 = (xup )5/2
A13 = (xup )7/2
A14 = (xup )9/2
A15 = (xup )11/2
p
A21 = (3/2) (xup )
A22 = (5/2)(xup )3/2
A23 = (7/2)(xup )5/2
A24 = (9/2)(xup )7/2
A25 = (11/2)(xup )9/2
A31 = (3/4)(xup )1/2
A32 = (15/4)(xup )1/2
A33 = (35/4)(xup )3/2
A34 = (63/4)(xup )5/2
A35 = (99/4)(xup )7/2
A41 = 1
A42 = 1
A43 = 1
A44 = 1
A45 = 1
A51 = 3/2
A52 = 5/2
A53 = 7/2
A54 = 9/2
A55 = 11/2
B(1) = yup a1 (xup )1/2
B(2) = (1/2)a1 (xup )1/2
B(3) = (d2 y/dx2 )up + (1/4)a1 (xup )3/2
B(4) = Tof f + TT E /2 a1 (xup )1/2
B(5) = tan(T E T E /2) + (1/4a1 (xup )3/2

1.2.2 Lower surface

The expanded polynomial for the lower surface of a general airfoil can be written as
follows. The differentiation of this equation with respect to x can be obtained from
the second equation. Again differentiating this equation with respect to x then we
get the following equations.
1 PARSEC 6

1. ylo = a1 (xlo )1/2 + a2 (xlo )3/2 + a3 (xlo )5/2 + a4 (xlo )7/2 + a5 (xlo )9/2 + a6 (xlo )11/2

2. dylo/dx = (1/2)a1 (xlo )1/2 + (3/2)a2 (xlo )1/2 + (5/2)a3 (xlo )3/2 + (7/2)a4 (xlo )5/2 +
(9/2)a5(xlo )7/2 +
(11/2)a6(xlo )9/2

3. d2 ylo/dx2 = (1/4)a1 (xlo )3/2 +(3/4)a2(xlo )1/2 +(15/4)a3(xlo )1/2 +(35/4)a4 (xlo )3/2 +

(63/4)a5(xlo )5/2 + (99/4)a6(xlo )7/2

Now we have to apply the following conditions for the above equations.
1. At xlo = maximum, ylo = maximum.
2. At xlo = maximum, dylo /dx = 0.
3. At xlo = maximum, d2 ylo /dx2 = maximum.
4. At xlo = 1, ylo = Tof f TT E /2
5. At xlo = 1, dylo /dx = tan(T E + T E /2)
2

Now we will get the following equations for each of the above conditions respec-
tively.
1. ylo a1 (xlo )1/2 = a2 (xlo )3/2 + a3 (xlo )5/2 + a4 (xlo )7/2 + a5 (xlo )9/2 + a6 (xlo )11/2

2. (1/2)a1 (xlo )1/2 = (3/2)a2 (xlo )1/2 +(5/2)a3(xlo )3/2 +(7/2)a4 (xlo )5/2 +(9/2)a5 (xlo )7/2 +
(11/2)a6(xlo )9/2

3. (d2 y/dx2)lo +(1/4)a1 (xlo )3/2 = (3/4)a2 (xlo )1/2 +(15/4)a3(xlo )1/2 +(35/4)a4 (xlo )3/2 +
(63/4)a5(xlo )5/2 + (99/4)a6(xlo )7/2

4. Tof f + TT E /2 a1 (xlo )1/2 = a2 (xlo)3/2 + a3 (xlo )5/2 + a4 (xlo )7/2 + a5 (xlo )9/2 +
a6 (xlo )11/2

5. tan(T E T E /2)+(1/4a1(xlo )3/2 = (3/4)a2 (xlo )1/2 +(15/4)a3(xlo )1/2 +(35/4)a4(xlo )3/2 +
(63/4)a5(xlo )5/2 + (99/4)a6(xlo )7/2

2 This will be treated as positive since the lower surface slope is keep on increasing from the maximum point(Ylo ).
1 PARSEC 7

These equations has been reformed as a matrix like we did for the upper surface of
the general airfoil.This matrix has been solved by using the method of Gauss-Elimination.
The resultant values will give the required five coefficient values. Finally by substi-
tuting the values of coefficients in the following equation the required co-ordinates for
the lower surface of general airfoil has been obtained.

ylo = a1 (xlo )1/2 + a2 (xlo )3/2 + a3 (xlo )5/2 + a4 (xlo )7/2 + a5 (xlo )9/2 + a6 (xlo )11/2

1.3 Numerical coding

A FORTRAN program has been developed to implement the parametric section
method in which the 12 parameters are the inputs. Execution of this program will re-
turn the x and y co-ordinates of the particular airfoil. Even we can extend this airfoil
geometry to a wing. We have the capability to produce rectangular,swept back,swept
forward and many other shapes using this PARSEC program. This program has been
combined with panel code and genetic algorithm code (these codes are available) for
finding pressure distribution,coefficient of lift for the airfoil and for the aerodynamic
optimization. They will be described later.

1.4 Demonstration of PARSEC

Here NACA 2411 and NACA 0012 airfoils have been selected for exprementing the
PARSEC method. From this available standard co-ordinate values of the airfoils we
have extracted the 12 parameters which are required by the PARSEC method. Then
those values had been given as input parameters to the PARSEC program. Then the
execution of the program has returned the co-ordinates.

We are able to get the co-ordinates with the accuracy we need. But it must have
been specified in the program. The extracted 12 parameters for both the airfoils has
been tabulated in first and second table respectively.

The standard co-ordinates of these airfoils have been given in third,eighth and
nineth tables. The co-ordinates for the upper surface of the NACA 2411 airfoil gen-
erated from the PARSEC has been given in fourth and fifth tables. The sixth and
seventh table will show the co-ordinates for the lower surface. The co-ordinates for
the upper surface of the NACA 0012 airfoil generated from the PARSEC has been
given in tenth and eleventh tables. The twelveth and thirteenth table will show the
co-ordinates for the lower surface.
1 PARSEC 8

Parameter value
Rleu 0.0216mm
Rlel 0.008mm
Xup 0.3445mm
Yup 0.07912mm
Y XXup -0.6448mm
Xlo 0.16912mm
Ylo -0.03379mm
Y XXlo 0.6748mm
Tte 0.000mm
Tof f 0.000mm
te -4.785 deg
te 15.082 deg

Parameter value
Rleu 0.0155mm
Rlel 0.0155mm
Xup 0.296632mm
Yup 0.060015mm
Y XXup -0.4515mm
Xlo 0.296632mm
Ylo -0.060015mm
Y XXlo 0.45309mm
Tte 0.001260mm
Tof f 0.000mm
te 00.000 deg
te 07.36 deg
1 PARSEC 9

xstandard ystandard
1.00000 0.00000
0.99730 0.00059
0.98918 0.00235
0.97578 0.00520
0.95720 0.00906
0.93365 0.01380
0.90535 0.01927
0.87260 0.02532
0.83577 0.03175
0.79520 0.03839
0.75138 0.04506
0.70475 0.05158
0.65583 0.05777
0.60513 0.06347
0.55322 0.06852
0.50067 0.07276
0.44807 0.07606
0.39597 0.07830
0.34455 0.07912
0.29485 0.07831
0.24745 0.07588
0.20292 0.07193
0.16175 0.06659
0.12448 0.06006
0.09148 0.05253
0.06317 0.04427
0.03985 0.03553
0.02170 0.02654
0.00897 0.01752
0.00172 0.00862
0.00000 0.00000
0.00375 -0.00794
0.01290 -0.01483
0.02723 -0.02061
0.04662 -0.02531
0.07080 -0.02893
0.09948 -0.03151
0.13238 -0.03310
0.16912 -0.03379
0.20932 -0.03368
0.25255 -0.03291
0.29842 -0.03165
0.34642 -0.03005
0.39612 -0.02830
0.44740 -0.02637
0.49933 -0.02415
0.55132 -0.02174
0.60280 -0.01925
0.65320 -0.01677
0.70198 -0.01436
0.74862 -0.01208
0.79257 -0.00994
0.83337 -0.00798
0.87053 -0.00621
0.90368 -0.00463
0.93238 -0.00326
0.95633 -0.00211
0.97528 -0.00120
0.98895 -0.00054
0.99723 -0.00013
1.00000 0.00000
1 PARSEC 10

xupper yupper
0.0000000 0.0000000
1.01010101E-02 2.05418225E-02
2.02020202E-02 2.85857264E-02
3.03030312E-02 3.44710909E-02
4.04040404E-02 3.92132960E-02
5.05050495E-02 4.32145000E-02
6.06060587E-02 4.66848947E-02
7.07070678E-02 4.97512408E-02
8.08080807E-02 5.24969958E-02
9.09090936E-02 5.49804233E-02
0.10101011 5.72439507E-02
0.11111112 5.93193807E-02
0.12121213 6.12311214E-02
0.13131315 6.29982203E-02
0.14141415 6.46357015E-02
0.15151516 6.61555678E-02
0.16161616 6.75674528E-02
0.17171717 6.88791201E-02
0.18181817 7.00968653E-02
0.19191918 7.12257773E-02
0.20202018 7.22700059E-02
0.21212119 7.32329041E-02
0.22222219 7.41172060E-02
0.23232320 7.49251172E-02
0.24242420 7.56584480E-02
0.25252521 7.63186663E-02
0.26262623 7.69069642E-02
0.27272725 7.74243325E-02
0.28282827 7.78716207E-02
0.29292929 7.82495365E-02
0.30303031 7.85587281E-02
0.31313133 7.87997693E-02
0.32323235 7.89732262E-02
0.33333337 7.90796578E-02
0.34343439 7.91196227E-02
0.35353541 7.90937543E-02
0.36363643 7.90026709E-02
0.37373745 7.88470879E-02
0.38383847 7.86277875E-02
0.39393950 7.83455893E-02
0.40404052 7.80014470E-02
0.41414154 7.75963441E-02
0.42424256 7.71313831E-02
0.43434358 7.66077414E-02
0.44444460 7.60267004E-02
0.45454562 7.53896013E-02
0.46464664 7.46979117E-02
0.47474766 7.39531592E-02
0.48484868 7.31569752E-02
0.49494970 7.23110437E-02
0.50505072 7.14171454E-02
0.51515174 7.04771206E-02
0.52525276 6.94928914E-02
0.53535378 6.84664175E-02
0.54545480 6.73997253E-02
1 PARSEC 11

xupper yupper
0.55555582 6.62948787E-02
0.56565684 6.51539713E-02
0.57575786 6.39791414E-02
0.58585888 6.27725273E-02
0.59595990 6.15362972E-02
0.60606092 6.02725893E-02
0.61616194 5.89835607E-02
0.62626296 5.76713234E-02
0.63636398 5.63379712E-02
0.64646500 5.49855381E-02
0.65656602 5.36160134E-02
0.66666704 5.22313118E-02
0.67676806 5.08332588E-02
0.68686908 4.94235940E-02
0.69697011 4.80039343E-02
0.70707113 4.65757884E-02
0.71717215 4.51405160E-02
0.72727317 4.36993204E-02
0.73737419 4.22532484E-02
0.74747521 4.08031531E-02
0.75757623 3.93496975E-02
0.76767725 3.78933214E-02
0.77777827 3.64342332E-02
0.78787929 3.49724069E-02
0.79798031 3.35075408E-02
0.80808133 3.20390426E-02
0.81818235 3.05660460E-02
0.82828337 2.90873405E-02
0.83838439 2.76013985E-02
0.84848541 2.61063352E-02
0.85858643 2.45998893E-02
0.86868745 2.30794139E-02
0.87878847 2.15418506E-02
0.88888949 1.99837144E-02
0.89899051 1.84010677E-02
0.90909153 1.67895071E-02
0.91919255 1.51441414E-02
0.92929357 1.34595688E-02
0.93939459 1.17298607E-02
0.94949561 9.94853582E-03
0.95959663 8.10854416E-03
0.96969765 6.20224141E-03
0.97979867 4.22137044E-03
0.98989969 2.15703831E-03
1.0000007 -3.05303871E-07
1 PARSEC 12

xlower ylower
0.0000000 0.0000000
1.01010101E-02 -1.24577768E-02
2.02020202E-02 -1.72490627E-02
3.03030312E-02 -2.06657350E-02
4.04040404E-02 -2.33242214E-02
5.05050495E-02 -2.54690908E-02
6.06060587E-02 -2.72293054E-02
7.07070678E-02 -2.86840461E-02
8.08080807E-02 -2.98869424E-02
9.09090936E-02 -3.08769364E-02
0.10101011 -3.16838287E-02
0.11111112 -3.23313624E-02
0.12121213 -3.28390785E-02
0.13131315 -3.32234651E-02
0.14141415 -3.34987454E-02
0.15151516 -3.36773656E-02
0.16161616 -3.37703899E-02
0.17171717 -3.37877497E-02
0.18181817 -3.37384120E-02
0.19191918 -3.36305574E-02
0.20202018 -3.34716663E-02
0.21212119 -3.32686044E-02
0.22222219 -3.30276899E-02
0.23232320 -3.27547416E-02
0.24242420 -3.24551240E-02
0.25252521 -3.21337767E-02
0.26262623 -3.17952298E-02
0.27272725 -3.14436629E-02
0.28282827 -3.10828779E-02
0.29292929 -3.07163503E-02
0.30303031 -3.03472243E-02
0.31313133 -2.99783349E-02
0.32323235 -2.96122134E-02
0.33333337 -2.92511079E-02
0.34343439 -2.88969744E-02
0.35353541 -2.85515096E-02
0.36363643 -2.82161403E-02
0.37373745 -2.78920457E-02
0.38383847 -2.75801569E-02
0.39393950 -2.72811670E-02
0.40404052 -2.69955508E-02
0.41414154 -2.67235544E-02
0.42424256 -2.64652185E-02
0.43434358 -2.62203831E-02
0.44444460 -2.59886961E-02
0.45454562 -2.57696211E-02
0.46464664 -2.55624503E-02
0.47474766 -2.53663138E-02
0.48484868 -2.51801871E-02
0.49494970 -2.50029024E-02
0.50505072 -2.48331577E-02
0.51515174 -2.46695299E-02
0.52525276 -2.45104898E-02
0.53535378 -2.43544038E-02
0.54545480 -2.41995510E-02
1 PARSEC 13

xlower ylower
0.55555582 -2.40441337E-02
0.56565684 -2.38862950E-02
0.57575786 -2.37241220E-02
0.58585888 -2.35556643E-02
0.59595990 -2.33789477E-02
0.60606092 -2.31919810E-02
0.61616194 -2.29927804E-02
0.62626296 -2.27793697E-02
0.63636398 -2.25498099E-02
0.64646500 -2.23021992E-02
0.65656602 -2.20346954E-02
0.66666704 -2.17455328E-02
0.67676806 -2.14330293E-02
0.68686908 -2.10956093E-02
0.69697011 -2.07318142E-02
0.70707113 -2.03403253E-02
0.71717215 -1.99199710E-02
0.72727317 -1.94697473E-02
0.73737419 -1.89888421E-02
0.74747521 -1.84766371E-02
0.75757623 -1.79327372E-02
0.76767725 -1.73569825E-02
0.77777827 -1.67494696E-02
0.78787929 -1.61105674E-02
0.79798031 -1.54409325E-02
0.80808133 -1.47415334E-02
0.81818235 -1.40136676E-02
0.82828337 -1.32589750E-02
0.83838439 -1.24794655E-02
0.84848541 -1.16775334E-02
0.85858643 -1.08559784E-02
0.86868745 -1.00180237E-02
0.87878847 -9.16734058E-03
0.88888949 -8.30806140E-03
0.89899051 -7.44480966E-03
0.90909153 -6.58271089E-03
0.91919255 -5.72742196E-03
0.92929357 -4.88514686E-03
0.93939459 -4.06266004E-03
0.94949561 -3.26732895E-03
0.95959663 -2.50713388E-03
0.96969765 -1.79069140E-03
0.97979867 -1.12727657E-03
0.98989969 -5.26844873E-04
1.0000007 -5.49762902E-08
1 PARSEC 14

Figure 2: NACA 2411 generated from PARSEC

1 PARSEC 15

Figure 3: NACA 2411 standard Vs NACA 2411 PARSEC

1 PARSEC 16
1 PARSEC 17

xstandard ystandard
1.000000 0.001260
0.999315 0.001356
0.997261 0.001644
0.993844 0.002120
0.989074 0.002783
0.982963 0.003626
0.975528 0.004642
0.966790 0.005826
0.956773 0.007168
0.945503 0.008658
0.933013 0.010286
0.919335 0.012042
0.904508 0.013914
0.888573 0.015891
0.871572 0.017959
0.853553 0.020107
0.834565 0.022323
0.814660 0.024593
0.793893 0.026905
0.772320 0.029246
0.750000 0.031603
0.726995 0.033962
0.703368 0.036311
0.679184 0.038634
0.654508 0.040917
0.629410 0.043147
0.603956 0.045307
0.578217 0.047383
0.552264 0.049358
0.526168 0.051216
0.500000 0.052940
0.473832 0.054515
0.447736 0.055923
0.421783 0.057148
0.396044 0.058175
0.370590 0.058989
0.345492 0.059575
0.320816 0.059921
0.296632 0.060015
0.273005 0.059848
0.250000 0.059412
0.227680 0.058702
0.206107 0.057714
0.185340 0.056447
0.165435 0.054902
0.146447 0.053083
0.128428 0.050996
0.111427 0.048648
0.095492 0.046049
0.080665 0.043211
0.066987 0.040145
0.054497 0.036867
0.043227 0.033389
0.033210 0.029726
0.024472 0.025893
0.017037 0.021904
0.010926 0.017770
0.006156 0.013503
0.002739 0.009114
0.000685 0.004611
0.000000 0.000000
1 PARSEC 18

xstandard ystandard
0.000685 -0.004611
0.002739 -0.009114
0.006156 -0.013503
0.010926 -0.017770
0.017037 -0.021904
0.024472 -0.025893
0.033210 -0.029726
0.043227 -0.033389
0.054497 -0.036867
0.066987 -0.040145
0.080665 -0.043211
0.095492 -0.046049
0.111427 -0.048648
0.128428 -0.050996
0.146447 -0.053083
0.165435 -0.054902
0.185340 -0.056447
0.206107 -0.057714
0.227680 -0.058702
0.250000 -0.059412
0.273005 -0.059848
0.296632 -0.060015
0.320816 -0.059921
0.345492 -0.059575
0.370590 -0.058989
0.396044 -0.058175
0.421783 -0.057148
0.447736 -0.055923
0.473832 -0.054515
0.500000 -0.052940
0.526168 -0.051216
0.552264 -0.049358
0.578217 -0.047383
0.603956 -0.045307
0.629410 -0.043147
0.654508 -0.040917
0.679184 -0.038634
0.703368 -0.036311
0.726995 -0.033962
0.750000 -0.031603
0.772320 -0.029246
0.793893 -0.026905
0.814660 -0.024593
0.834565 -0.022323
0.853553 -0.020107
0.871572 -0.017959
0.888573 -0.015891
0.904508 -0.013914
0.919335 -0.012042
0.933013 -0.010286
0.945503 -0.008658
0.956773 -0.007168
0.966790 -0.005826
0.975528 -0.004642
0.982963 -0.003626
0.989074 -0.002783
0.993844 -0.002120
0.997261 -0.001644
0.999315 -0.001356
1.000000 -0.001260
1 PARSEC 19

xupper yupper
0.0000000 0.0000000
1.01010101E-02 1.74102951E-02
2.02020202E-02 2.42272820E-02
3.03030312E-02 2.91992594E-02
4.04040404E-02 3.31816413E-02
5.05050495E-02 3.65125164E-02
6.06060587E-02 3.93686518E-02
7.07070678E-02 4.18570004E-02
8.08080807E-02 4.40484881E-02
9.09090936E-02 4.59931791E-02
0.10101011 4.77280952E-02
0.11111112 4.92815599E-02
0.12121213 5.06758504E-02
0.13131315 5.19288629E-02
0.14141415 5.30552343E-02
0.15151516 5.40671088E-02
0.16161616 5.49746938E-02
0.17171717 5.57866469E-02
0.18181817 5.65103777E-02
0.19191918 5.71522675E-02
0.20202018 5.77178672E-02
0.21212119 5.82120158E-02
0.22222219 5.86389601E-02
0.23232320 5.90024255E-02
0.24242420 5.93057312E-02
0.25252521 5.95517978E-02
0.26262623 5.97432368E-02
0.27272725 5.98823763E-02
0.28282827 5.99713027E-02
0.29292929 6.00118898E-02
0.30303031 6.00058213E-02
0.31313133 5.99546172E-02
0.32323235 5.98596558E-02
0.33333337 5.97222000E-02
0.34343439 5.95433787E-02
0.35353541 5.93242496E-02
0.36363643 5.90657704E-02
0.37373745 5.87688312E-02
0.38383847 5.84342591E-02
0.39393950 5.80628254E-02
0.40404052 5.76552637E-02
0.41414154 5.72122559E-02
0.42424256 5.67344651E-02
0.43434358 5.62225245E-02
0.44444460 5.56770414E-02
0.45454562 5.50986193E-02
0.46464664 5.44878431E-02
0.47474766 5.38453087E-02
0.48484868 5.31715974E-02
0.49494970 5.24672940E-02
0.50505072 5.17330021E-02
0.51515174 5.09693213E-02
0.52525276 5.01768850E-02
0.53535378 4.93563227E-02
0.54545480 4.85083014E-02
1 PARSEC 20

xupper yupper
0.55555582 4.76335026E-02
0.56565684 4.67326418E-02
0.57575786 4.58064489E-02
0.58585888 4.48556952E-02
0.59595990 4.38811891E-02
0.60606092 4.28837575E-02
0.61616194 4.18642908E-02
0.62626296 4.08236869E-02
0.63636398 3.97629142E-02
0.64646500 3.86829711E-02
0.65656602 3.75848934E-02
0.66666704 3.64697836E-02
0.67676806 3.53387743E-02
0.68686908 3.41930576E-02
0.69697011 3.30338776E-02
0.70707113 3.18625271E-02
0.71717215 3.06803528E-02
0.72727317 2.94887628E-02
0.73737419 2.82892194E-02
0.74747521 2.70832442E-02
0.75757623 2.58724168E-02
0.76767725 2.46583782E-02
0.77777827 2.34428328E-02
0.78787929 2.22275537E-02
0.79798031 2.10143682E-02
0.80808133 1.98051780E-02
0.81818235 1.86019465E-02
0.82828337 1.74067076E-02
0.83838439 1.62215643E-02
0.84848541 1.50486883E-02
0.85858643 1.38903232E-02
0.86868745 1.27487844E-02
0.87878847 1.16264597E-02
0.88888949 1.05258115E-02
0.89899051 9.44937579E-03
0.90909153 8.39976594E-03
0.91919255 7.37967109E-03
0.92929357 6.39185589E-03
0.93939459 5.43916598E-03
0.94949561 4.52452479E-03
0.95959663 3.65093513E-03
0.96969765 2.82148039E-03
0.97979867 2.03932612E-03
0.98989969 1.30771974E-03
1.0000007 6.29990944E-04
1 PARSEC 21

xlower ylower
0.0000000 0.0000000
1.01010101E-02 -1.74094830E-02
2.02020202E-02 -2.42251866E-02
3.03030312E-02 -2.91957594E-02
4.04040404E-02 -3.31767574E-02
5.05050495E-02 -3.65063511E-02
6.06060587E-02 -3.93613577E-02
7.07070678E-02 -4.18487564E-02
8.08080807E-02 -4.40394953E-02
9.09090936E-02 -4.59836423E-02
0.10101011 -4.77182195E-02
0.11111112 -4.92715426E-02
0.12121213 -5.06658740E-02
0.13131315 -5.19191027E-02
0.14141415 -5.30458391E-02
0.15151516 -5.40582128E-02
0.16161616 -5.49664125E-02
0.17171717 -5.57790697E-02
0.18181817 -5.65035678E-02
0.19191918 -5.71462773E-02
0.20202018 -5.77127188E-02
0.21212119 -5.82077056E-02
0.22222219 -5.86354658E-02
0.23232320 -5.89997098E-02
0.24242420 -5.93037270E-02
0.25252521 -5.95504232E-02
0.26262623 -5.97423911E-02
0.27272725 -5.98819442E-02
0.28282827 -5.99711537E-02
0.29292929 -6.00118749E-02
0.30303031 -6.00057840E-02
0.31313133 -5.99543899E-02
0.32323235 -5.98590598E-02
0.33333337 -5.97210489E-02
0.34343439 -5.95414825E-02
0.35353541 -5.93214184E-02
0.36363643 -5.90617992E-02
0.37373745 -5.87635264E-02
0.38383847 -5.84274270E-02
0.39393950 -5.80542646E-02
0.40404052 -5.76447919E-02
0.41414154 -5.71996868E-02
0.42424256 -5.67196235E-02
0.43434358 -5.62052391E-02
0.44444460 -5.56571633E-02
0.45454562 -5.50759956E-02
0.46464664 -5.44623509E-02
0.47474766 -5.38168214E-02
0.48484868 -5.31400181E-02
0.49494970 -5.24325408E-02
0.50505072 -5.16950041E-02
0.51515174 -5.09280302E-02
0.52525276 -5.01322709E-02
0.53535378 -4.93083782E-02
0.54545480 -4.84570377E-02
1 PARSEC 22

xlower ylower
0.55555582 -4.75789532E-02
0.56565684 -4.66748551E-02
0.57575786 -4.57455069E-02
0.58585888 -4.47916947E-02
0.59595990 -4.38142456E-02
0.60606092 -4.28140201E-02
0.61616194 -4.17919159E-02
0.62626296 -4.07488644E-02
0.63636398 -3.96858454E-02
0.64646500 -3.86038795E-02
0.65656602 -3.75040285E-02
0.66666704 -3.63874026E-02
0.67676806 -3.52551602E-02
0.68686908 -3.41085047E-02
0.69697011 -3.29486914E-02
0.70707113 -3.17770317E-02
0.71717215 -3.05948853E-02
0.72727317 -2.94036604E-02
0.73737419 -2.82048304E-02
0.74747521 -2.69999206E-02
0.75757623 -2.57905126E-02
0.76767725 -2.45782491E-02
0.77777827 -2.33648289E-02
0.78787929 -2.21520159E-02
0.79798031 -2.09416300E-02
0.80808133 -1.97355580E-02
0.81818235 -1.85357481E-02
0.82828337 -1.73442122E-02
0.83838439 -1.61630269E-02
0.84848541 -1.49943344E-02
0.85858643 -1.38403466E-02
0.86868745 -1.27033386E-02
0.87878847 -1.15856556E-02
0.88888949 -1.04897115E-02
0.89899051 -9.41798743E-03
0.90909153 -8.37303791E-03
0.91919255 -7.35748513E-03
0.92929357 -6.37402479E-03
0.93939459 -5.42542338E-03
0.94949561 -4.51452052E-03
0.95959663 -3.64422752E-03
0.96969765 -2.81753088E-03
0.97979867 -2.03749072E-03
0.98989969 -1.30724174E-03
1.0000007 -6.29993738E-04
1 PARSEC 23

Figure 4: NACA 0012 generated from PARSEC

1 PARSEC 24

Figure 5: NACA 0012 standard Vs NACA 0012 PARSEC

1 PARSEC 25

1.5 Other generated airfoil geometries

Here some of the airfoil geometries which have been created by this PARSEC method.
They all have the following common datas except figure 6. Varying datas are shown
in each figure. In figure 1 the airfoil has both the surfaces symmetrical to each other.
Figure 2 indicates that the upper surface of the airfoil has been cusped. It shows
significant effect in the aerodynamic characteristics of the airfoil. Figure 3 is same as
figure 2, but it has the lower cusp.

The airfoil which has the trailing edge thickness is shown in figure 4. They are
used to place some secondary control devices to the airfoil. The airfoils with trailing
edge offset have been indicated in figure 5 and 6. It has the better control over the
airfoil aerodynamic characteristics.

The airfoil with both cusped surface and trailing edge offset has been shown in
the figure 7. Like above airfoils we can generate different kinds of airfoils using the
PARSEC parameterization method. Even we have the capability to generate airfoils
with negative trailing edge offsets. It has been shown in figure 8.The followings are
the parameters used to generate the following airfoils.

Parameter value
Rleu 0.050mm
Rlel 0.002mm
Xup 0.400mm
Yup 0.090mm
Y XXup -0.200mm
Xlo 0.400mm
Ylo -0.090mm
Y XXlo 0.600mm
Tte 0.000mm
Tof f 0.000mm
te 0.000 deg
te 20.00 deg
1 PARSEC 26

Figure 6: symmetrical airfoil

Figure 7: uppercusp airfoil

1 PARSEC 27

Figure 8: lowercusp airfoil

Figure 9: airfoil with Tthickness

1 PARSEC 28

Figure 10: airfoil with Toffset

Figure 11: airfoil with Toffset

1 PARSEC 29

Figure 12: cusp airfoil with Toffset

Figure 13: cusp airfoil with negative Toffset

1 PARSEC 30

1.6 wings
Even different kind of wings can be generated from PARSEC method. Some wing geometries which are
generated by this method are shown below.

Figure 14: NACA 0012 wing with straight trailing edge and swept back leading edge
1 PARSEC 31

Figure 15: NACA 0012 wing both the edges inclined towards the centre
1 PARSEC 32

Figure 16: NACA 0012 swept forward wing

1 PARSEC 33

Figure 17: NACA 0012 rectangular wing

1 PARSEC 34

Figure 18: NACA 2411 wing with straight trailing edge and swept forward leading edge
1 PARSEC 35

Figure 19: NACA 2411 wing with swept back

1 PARSEC 36

1.7 Advantages of PARSEC

1. No baseline shape is needed.
2. Specifically designed for airfoils.
3. Wide range of airfoil shapes can be generated.
4. Typical geometric constraints on the airfoil shape (e.g thickness) can be ex-
pressed or approximated by simple bound or linear constraints.
5. The impact of individual design parameters on the aerodynamic properties of
the airfoil can be predicted more easily.

1.8 Disadvantages of PARSEC

1. Not applicable to wide class of problems as spline curves.
2. Not suitable for turbine blade designs when compared to other parameterization
methods.
2 PANEL TECHNIQUE 37

2 Panel Technique
2.1 Thin airfoil theory
Although thin airfoil theory provides invaluable insights into the generation of lift, the
Kutta-condition,the effect of camber distribution on the coefficients of lift and mo-
ment, and the location of the center of pressure and the aerodynamic center, it has
several limitations that prevent it from being used in practical applications. Among
these we can mention:

. It ignores the effects of the thickness distribution on cl and cmac .

. Pessure distributions tend to be inaccurate near stagnation points.
. Airfoils with high camber or large thickness violate the assumptions of airfoil
theory, and, therefore, the prediction accuracy degrads in these situations even
away from stagnation points.

2.2 Alternatives
We could consider the following alternatives in order to overcome the limitations of
thin airfoil theory,

. In addition to sources and vortices, we could use higher order solutions to

Laplaces equation that can enhance the accuracy of the approximation (dou-
blet, quadrupoles, octupoles, etc.). This approach falls under the denomination
of multipole expansions.
. We can use the same solutions to Laplaces equation (sources/sinks and vortices)
but place them on the surface of the body of interest, and use the exact ow
tangency boundary conditions without the approximations used in thin airfoil
theory.

This latter method can be shown to treat a wide range of reasonable problems for
the applied aerodynamicist, including multi-element airfoils. It also has the advantage
that it can be naturally extended to three-dimensional ows (unlike streamfunction or
complex variable methods).

The distribution of the sources/sinks and vortices on the surface of the body can be
either continuous or discrete. A continuous distribution leads to integral equations
similar to those we saw in thin airfoil theory which cannot be treated analytically.

If we discretize the surface of the body into a series of segments or panels, the integral
equations are transformed into an easily solvable set of simultaneous linear equations.
These methods come in many avors and are typically called as panel methods.
2 PANEL TECHNIQUE 38

2.3 Hess - smith panel method

It is based on a distribution of sources and vortices on the surface of the geometry.
In their method,

where, is the total potential function and its three components are the potentials
corresponding to the free stream, the source distribution, and the vortex distribution.
These last two distributions have potentially locally varying strengths q(s) and (s),
where s is an arc-length coordinate which spans the complete surface of the airfoil in
any way you want. The potentials created by the distribution of sources/sinks and
vortices are given by:

Figure 20: Airfoil analysis nomenclature for panel methods

Notice that in these formulae, the integration is to be carried out along the com-
plete surface of the airfoil. Using the superposition principle, any such distribution
of sources/sinks and vortices satises Laplaces equation, but we will need to nd condi-
tions for q(s) and (s) such that the ow tangency boundary condition and the Kutta
condition are satised. Notice that we have multiple options. In theory, we could:

. Use the source strength distribution to satisfy ow tangency and the vortex dis-
tribution to satisfy the Kutta condition.
. Use arbitrary combinations of both sources/sinks and vortices to satisfy both
boundary conditions simultaneously.
2 PANEL TECHNIQUE 39

Hess and Smith made the following valid simplication.

. Break up the surface into straight line segments

. Assume the source strength is constant over each line segment (panel) but has
a different value for each panel, and
. The vortex strength is constant and equal over each panel.
. The ow tangency boundary condition is satised everywhere.

The following figure illustrates the representation of a smooth surface by a series

of line segments. The numbering system starts at the lower surface trailing edge and
proceeds forward, around the leading surface and aft to the upper surface trailing
edge. N+1 points define N panels.

Figure 21: Nodes and Panels

we can discretize the first equation in the following way:

Since this equation involves integrations over each discrete panel on the surface of
the airfoil, we must somehow parameterize the variation of source and vortex strength
within each of the panels. Since the vortex strength was considered to be a constant,
we only need worry about the source strength distribution within each panel. This is
the major approximation of the panel method.

However, you can see how the importance of this approximation should decrease
as the number of panels, N (of course this will increase the cost of the computation
considerably, so there are more ecient alternatives). Hess and Smith decided to take
the simplest possible approximation, that is, to take the source strength to be con-
stant on each of the panels. q(s) = qi on panel i. where, i = 1,...,N.
2 PANEL TECHNIQUE 40

Therefore, we have N + 1 unknowns to solve for in our problem: the N panel

source strengths qi and the constant vortex strength . Consequently, we will need
N + 1 independent equations which can be obtained by formulating the ow tangency
boundary condition at each of the N panels, and by enforcing the Kutta condition
discussed previously. The solution of the problem will require the inversion of a ma-
trix of size (N + 1) (N + 1).

In this method the flow tangency boundary condition is imposed on the points
located at the midpoint of each of the panels. Although it suers from a slight alter-
ation of the surface geometry, it is easy to implement and yields fairly accurate results
for a reasonable number of panels. This location is also used for the imposition of
the Kutta condition (on the last panels on upper and lower surfaces of the airfoil,
assuming that their midpoints remain at equal distances from the trailing edge as the
number of panels is increased).

Finally, once you have solved the system for the unknowns of the problem, it is
easy to construct the tangential velocity at the midpoint of each panel according to
the following formula,

And knowing the tangential velocity component, we can compute the pressure co-
efficient (no approximation since Vni = 0) at the midpoint of each panel according
to the following formula,
2 PANEL TECHNIQUE 41

The pressure distribution on the surface of the NACA 2411 airfoil for 5.0 deg an-
gle of attack has been shown below. The co-efficient of lift obtained for this angle of
attack is 0.842.

Figure 22: Pressure Distribution on NACA 2411 airfoil

2 PANEL TECHNIQUE 42

The comparison between the standard pressure distribution and the pressure dis-
tribution based on the panel theory for 5.0 deg angle of attack for NACA 2411 has
been indicated here below.

Figure 23: Pressure Distribution standard Vs Panel theory

2 PANEL TECHNIQUE 43

The Pressure distribution on the surface of the NACA 0012 airfoil for 5.0 deg angle
of attack based on the panel theory has been indicated here in the following graph.
The co-efficient of lift obtained for this angle of attack is 0.622.

Figure 24: Pressure Distribution on NACA 0012 airfoil

2 PANEL TECHNIQUE 44

The comparison between the standard pressure distribution and the pressure dis-
tribution based on the panel theory for 5.0 deg angle of attack for NACA 0012 has
been indicated here below.The angle of attack and corresponding co-efficient of lift
for NACA 0012 and 2411 airfoils has been tabulated here below resectively.

Figure 25: Pressure Distribution standard Vs Panel theory

Angle of attack Cp
2.0 0.2789
5.0 0.6226
10.0 1.1793

Angle of attack Cp
2.0 0.4997
5.0 0.8421
10.0 1.3964
3 AERODYNAMIC OPTIMIZATION 45

3 Aerodynamic optimization
3.1 Genetic algorithm
Genetic algorithms are search algorithms based on the mechanics of natural selection
and natural genetics. They combine survival of the fittest among string structures
with a structured yet randomized information exchange to form a search algorithm
with some of the innovative flair of human search. In every generation, a new set of
artificial creatures (strings) is created using bits and pieces of the fittest of theold; an
occasional new part is tried for good measure. While randomized, genetic algorithms
are no simple random walk. They efficiently exploit historical information to specu-
late on new search points with expected improved performance.

. GA is a random search algorithm which keeps a pool of candidate solutions (gene

pool).
. Each solution is encoded in a binary string called a chromosome with each bit
being a gene.
. Evaluate the fitness of a solution using a selection criteria.
. Generate new chromosomes by reproduction rules, including cross-over (mating),
inversion, and mutation.
. Annihilate inferior (according to the result of evaluation using the selection cri-
teria) genes, to make room for new genes.
. Adding new genes with high fitness values into gene pool.
. Evaluate termination criteria. If not yet satisfied, continue the search process.

3.2 Incredients of GA
There are two important incredients in genetic algorithm. They are described below.
. Inheritance
. Variability
Inheritance - Off springs must retain at least some of the features that made their
parents fitter than average. Otherwise evolution is reset at every generation.
Variability - At any given time individuals with varying fitness must co-exist in
the population. Otherwise natural selection has nothing to operate upon.
3 AERODYNAMIC OPTIMIZATION 46

3.3 Operators of GA
3.3.1 Selection

According to Darwins evolution theory the best ones should survive and create new
offspring. There are many methods how to select the best chromosomes.Some of them
will be described here.
1. Roulette wheel selection
2. Rank selection
3. Stedy state selection
4. Elitism

1. Imagine a roulette wheel where are placed all chromosomes in the population,
every has its place big accordingly to its fitness function.Chromosome with big-
ger fitness will be selected more times.

2. The previous selection will have problems when the fitnesses differs very much.
For example, if the best chromosome fitness is 90 percentage of all the roulette
wheel then the other chromosomes will have very few chances to be selected.
Rank selection first ranks the population and then every chromosome receives
fitness from this ranking.

The worst will have fitness 1, second worst 2 etc. and the best will have fit-
ness N (number of chromosomes in population). After this all the chromosomes
have a chance to be selected. But this method can lead to slower convergence,
because the best chromosomes do not differ so much from other ones.

3. Stedy state selection is not particular method of selecting parents. Main idea of
this selection is that big part of chromosomes should survive to next generation.
GA then works in a following way. In every generation are selected a few (good
- with high fitness) chromosomes for creating a new offspring.

Then some (bad - with low fitness) chromosomes are removed and the new off-
spring is placed in their place. The rest of population survives to new generation.

4. Idea of elitism has been already introduced. When creating new population
by crossover and mutation, we have a big chance, that we will loose the best
3 AERODYNAMIC OPTIMIZATION 47

chromosome. Elitism is name of method, which first copies the best chromosome
(or a few best chromosomes) to new population. The rest is done in classical
way. Elitism can very rapidly increase performance of GA, because it prevents
losing the best found solution.

3.3.2 crossover

Performance of GA very depends on crossover and mutation. Type and implemen-

tation of operators depends on encoding and also on a problem.Here the cross over
probability is selected. For each bit in a bit stream, a probability test is performed.
If passed, then one of the following methods can be used: There are many ways how
to do crossover and mutation.

. Single point crossover - one crossover point is selected, binary string from begin-
ning of chromosome to the crossover point is copied from one parent, the rest is
copied from the second parent.

Figure 26: Single point crossover

. Two point crossover - two crossover point are selected, binary string from begin-
ning of chromosome to the first crossover point is copied from one parent, the
part from the first to the second crossover point is copied from the second parent
and the rest is copied from the first parent.

Figure 27: Two point crossover

3 AERODYNAMIC OPTIMIZATION 48

. Uniform crossover - bits are randomly copied from the first or from the second
parent.

Figure 28: Uniform crossover

. Arithmetic crossover - some arithmetic operation is performed to make a new

offspring.

3.3.3 Mutation

Mutation will take place with a small probability. The mutation propability is se-
lected. If it is too small then the variability will be reduced. If it is too high then the
best chromosomes will be lost unnecessarily. For each bit in a bit stream, a probabil-
ity test is performed. If passed, then one of two methods can be used:

. That bit is flipped (0 changes to 1, and vice versa).

. Randomly generate a bit. If the randomly generated bit is different from the
original bit, the original bit is flipped.

Figure 29: Mutation

There are some special genetic algorithm operators. They are as follows.
. Elitism - Good individuals are preserved during any generation
. Niche - Pool of best individuals
. Micro ga - Opeartes on selected individuals
3 AERODYNAMIC OPTIMIZATION 49

3.4 Genetic Algorithm approach

. Define limits of varaiable parameters.
. Generate a random population of designs.
. Evaulate fitness of designs.
. Mate selection.
. Crossover.
. Mutation.
. Reevaluate fitness of new population.

3.5 Representation of chromosomes

. Binary string to represent real values of an independent variable x
. Minimum value of x - Xmin
. Maximum value of x - X max
. Accuracy Required - d decimal places
. Domain length - 10d (Xmax Xmin )
. Find n such that - 2(n1) < 10d (Xmax Xmin ) < 2n

3.6 Converting binary strings into real numbers

The mapping from binary string of 22 bits m =< m21 , ..., m1 , m0 | into a real number
x is completed in two steps.
. Convert the binary string m=< m21 , ..., m 1 , m0 | from base 2 to base 10
P21 i
(m =< m21 , ..., m1 , m0 |)2 = i=0 (mi .2 ) = x : and
10
 
22
. Map x to real value (linear mapping): x = xmin + x .(xmax xmin )/(2 1) ,
where xmaxis the upper bound and xminis the lower bound

An example for binary encoding of real values has been given in the following fig.
3 AERODYNAMIC OPTIMIZATION 50

Figure 30: Binary encoding of real values

3.7 Advantages of GA
. It works with a population of individuals (solutions).
. It uses probabilistic (non-deterministic) operators.
. Not affected by local discontinuties.
. Fitness function typically evaluates the individual (solution) as a whole, rather
than its parts.
. Global search: solutions from several parts of the search space are evaluated and
crossed-over at each generation.
. Robust search method, less likely to get trapped into local maximum peaks in
the search space.

3.8 Disadvantages of GA
. Large number of function evaluations.
. Difficult in finding exact global maxima or minima.
. Ideal population size is not easy to choose.
. Being stocahstic in nature, covergence is rather slow.

3.9 Two variable optimization problem - Himmelbleau function

The Himmelbleau function f (x) = (x21 + x2 11)2 + (x22 + x1 7)2 has been selected
for the experimental optimization problem. In this function when f (x1 , x2 ) = 0 min-
imum occurs at x1 = 3.0, x2 = 2.0. Those values have been obtained by using the
genetic algorithm as follows.
3 AERODYNAMIC OPTIMIZATION 51

3.9.1 Inputs

The inputs given to the program is as follows.

. npopsize = 6
. ngeneration = 500
. nparams = 2
. crossover probability = 0.7
. mutation probability = 0.2
. lower bounds of variables = 0.0, 0.0
. upper bounds of variables = 5.0, 5.0
. npossible values = 8192, 8192
. nich flag = 0, 0
. option for elitism = 0 and micro ga = 1
. creep mutation = 0 and its probabilty = 0
. option for tournament selection = 0
. number of childern for a given pair of parents = 1
. detailed output / number of generations for restart file writing = 500
. dummy integer for random number generation = -5000

3.9.2 Outputs

The outputs of the program are,

. Individuals(in real numbers), their fitness values and their binary strings at every
generation.
. Average and best fitness value at every generation.
. Elistic reproduction.
. Intitiation of micro ga
. summary of best individuals at every generation.
3 AERODYNAMIC OPTIMIZATION 52

3.9.3 Subroutines

The subroutines used in the program has been listed below.

. code = Codes floating point value to binary string.

. crosovr = Performs crossover (single-point or uniform).
. decode = Decodes binary string to floating point value.
. evalout = Evaluates the fitness of each individual and outputs
. generational information to the ga.out file.
. func = The function which is being evaluated.
. gamicro = Implements the micro-GA technique.
. input = Inputs information from the ga.inp file.
. initial = Program initialization and inputs information from the ga.restart file.
. mutate = Performs mutation (jump and/or creep).
. newgen = Writes child array back into parent array for new generation;
also checks to see if best individual was replicated (elitism).
. niche = Performs niching (sharing) on population.
. possibl = Checks to see if decoded binary string falls within specified range of
parmin and parmax.
. ran3 = The random number generator.
. restart = Writes the ga.restart file.
. select = A subroutine of selectn.
. selectn = Performs selection; tournament selection is the only option in this
version of the code.
. shuffle = Shuffles the population randomly for selection.
3 AERODYNAMIC OPTIMIZATION 53

3.9.4 Results
The function evaluation subroutine has been shown below. Here we have denoted the
function value as negative in order to determine the minimum. This is because ga
assumes that the objective function is always positive.

Figure 31: Function evaluation subroutine

3 AERODYNAMIC OPTIMIZATION 54

Figure 32: First Generation

3 AERODYNAMIC OPTIMIZATION 55

The best individual occurs at 474th generation. It has been indicated here below.
We can see that the optimized individual occurs suddenly during the evolution. Even
after the optimized individual has been arised, the variability exists.

Figure 33: List of best individuals

3 AERODYNAMIC OPTIMIZATION 56

Figure 34: Optimized offspring

Figure 35: Fitness Vs Generation

3 AERODYNAMIC OPTIMIZATION 57

Figure 36: 3 - Dimensional plot of Himmelblauc function minimization

3 AERODYNAMIC OPTIMIZATION 58

3.9.5 Important observations

. Cross over probability should be reasonable. It should not be too high or too
low.
. Mutation probability should be very low.
. Increse of population size above some value will not increase the effectiveness of
ga in finding the solution. Rather it will increase the execution time unneces-
sarily.
. Increase of generation will not always increase the fitness of best individual.
. Suitability of mate selection is based on the particular problem.
. Suitability of encoding is based on the particular problem.
3 AERODYNAMIC OPTIMIZATION 59

3.10 Airfoil optimization problem

Here NACA 2411 airfoil has been chosen for aerodynamic optimization. It has been
done using genetic algorithm. The objective of this optimization is to get the twelve
parameters of the airfoil such that they will improve the aerodynamic characteristics
of the airfoil for a specfic environment which will be described later.

3.10.1 Approach to optimization

. Define the nature of the problem clearly.

. Define the structural constraints of the problem.
. Select the objective function.
. Use an optimization method.
. Analyse the results.
. Implement the result or reevluate the problem.

The structural constraints of this problem are,

. The angle of attack is 5.0 deg.

. The flow is incompressible and subsonic.
. The trailing edge thickness of the airfoil is zero.
. The trailing edge offset of the airfoil is zero.

The objective function of this problem is,

. Maximizing co-efficient of lift.

3.10.2 Inputs

The inputs given to the program to solve this problem are,

. npopsize = 5
. ngeneration = 1500
3 AERODYNAMIC OPTIMIZATION 60

. nparams = 12
. crossover probability = 0.7
. mutation probability = 0.2
. lower bounds of variables = 0.0200, 0.0060, 0.3000, 0.0770, -0.6300, 0.1600, -
0.03200, 0.66, -4.7, 15.0, 0,0
. upper bounds of variables = 0.0230, 0.0100, 0.3800, 0.0800, -0.6500, 0.1800,
-0.03400, 0.68, -4.8, 15.1, 0,0
. npossible values = 32, 1024, 1024, 32, 256, 256, 256, 256, 256, 128, 128, 0, 0
. nich flag = 0 for all parameters
. option for elitism = 0 and micro ga = 1
. creep mutation = 0 and its probabilty = 0
. option for tournament selection = 0
. number of childern for a given pair of parents = 1
. detailed output / number of generations for restart file writing = 1000
. dummy integer for random number generation = -5000
3 AERODYNAMIC OPTIMIZATION 61

3.10.3 Results

The function subroutine has been shown below. Here the function value has been
denoted as co-efficient of lift.

Figure 37: Function evaluation subroutine

3 AERODYNAMIC OPTIMIZATION 62

Figure 38: First Generation

3 AERODYNAMIC OPTIMIZATION 63

The best individual occurs at 450th generation. It has been indicated here below.
We can see that the optimized individual occurs suddenly during the evolution. Even
after the optimized individual has been arised, the variability exists as seen in the
himmelbleau function.

Figure 39: List of best individuals

3 AERODYNAMIC OPTIMIZATION 64

Figure 40: Optimized Offspring

3 AERODYNAMIC OPTIMIZATION 65

P arameterstandard valuestandard
Rleu 0.0216mm
Rlel 0.008mm
Xup 0.3445mm
Yup 0.07912mm
Y XXup -0.6448mm
Xlo 0.16912mm
Ylo -0.03379mm
Y XXlo 0.6748mm
Tte 0.000mm
Tof f 0.000mm
te -4.785 deg
te 15.082 deg

P arameteroptimized valueoptimized
Rleu 0.02271mm
Rlel 0.008823mm
Xup 0.309306mm
Yup 0.078452mm
Y XXup -0.641529mm
Xlo 0.169961mm
Ylo -0.032196mm
Y XXlo 0.663843mm
Tte 0.000mm
Tof f 0.000mm
te -4.733071 deg
te 15.021260 deg
3 AERODYNAMIC OPTIMIZATION 66

Figure 41: Optimized NACA 2411 airfoil geometry

Figure 42: Standard airfoil Vs Optimized airfoil

3 AERODYNAMIC OPTIMIZATION 67

Figure 43: Fitness Vs Generation

3 AERODYNAMIC OPTIMIZATION 68

xupper yupper
0.0000000 0.0000000
1.01010101E-02 2.11484861E-02
2.02020202E-02 2.95333397E-02
3.03030312E-02 3.57201658E-02
4.04040404E-02 4.07349691E-02
5.05050495E-02 4.49809320E-02
6.06060587E-02 4.86675054E-02
7.07070678E-02 5.19204997E-02
8.08080807E-02 5.48226871E-02
9.09090936E-02 5.74319921E-02
0.10101011 5.97908497E-02
0.11111112 6.19314127E-02
0.12121213 6.38787299E-02
0.13131315 6.56527430E-02
0.14141415 6.72695786E-02
0.15151516 6.87425360E-02
0.16161616 7.00826868E-02
0.17171717 7.12993741E-02
0.18181817 7.24005476E-02
0.19191918 7.33930543E-02
0.20202018 7.42828175E-02
0.21212119 7.50750527E-02
0.22222219 7.57743344E-02
0.23232320 7.63847455E-02
0.24242420 7.69099593E-02
0.25252521 7.73532763E-02
0.26262623 7.77177066E-02
0.27272725 7.80060440E-02
0.28282827 7.82208368E-02
0.29292929 7.83644989E-02
0.30303031 7.84392878E-02
0.31313133 7.84473196E-02
0.32323235 7.83906579E-02
0.33333337 7.82712623E-02
0.34343439 7.80910179E-02
0.35353541 7.78517872E-02
0.36363643 7.75553435E-02
0.37373745 7.72034824E-02
0.38383847 7.67979324E-02
0.39393950 7.63404146E-02
0.40404052 7.58326426E-02
0.41414154 7.52763152E-02
0.42424256 7.46731013E-02
0.43434358 7.40246996E-02
0.44444460 7.33327791E-02
0.45454562 7.25990087E-02
0.46464664 7.18250573E-02
0.47474766 7.10125789E-02
0.48484868 7.01632425E-02
0.49494970 6.92786872E-02
0.50505072 6.83605522E-02
0.51515174 6.74104542E-02
0.52525276 6.64300099E-02
0.53535378 6.54208139E-02
0.54545480 6.43844306E-02
0.55555582 6.33224025E-02
3 AERODYNAMIC OPTIMIZATION 69

xupper yupper
0.56565684 6.22362457E-02
0.57575786 6.11274429E-02
0.58585888 5.99974282E-02
0.59595990 5.88476025E-02
0.60606092 5.76793030E-02
0.61616194 5.64938337E-02
0.62626296 5.52924089E-02
0.63636398 5.40762059E-02
0.64646500 5.28463125E-02
0.65656602 5.16037308E-02
0.66666704 5.03493957E-02
0.67676806 4.90841344E-02
0.68686908 4.78086844E-02
0.69697011 4.65236641E-02
0.70707113 4.52295914E-02
0.71717215 4.39268500E-02
0.72727317 4.26157042E-02
0.73737419 4.12962772E-02
0.74747521 3.99685465E-02
0.75757623 3.86323333E-02
0.76767725 3.72873060E-02
0.77777827 3.59329470E-02
0.78787929 3.45685817E-02
0.79798031 3.31933275E-02
0.80808133 3.18061151E-02
0.81818235 3.04056741E-02
0.82828337 2.89905090E-02
0.83838439 2.75589060E-02
0.84848541 2.61089187E-02
0.85858643 2.46383548E-02
0.86868745 2.31447723E-02
0.87878847 2.16254648E-02
0.88888949 2.00774558E-02
0.89899051 1.84974819E-02
0.90909153 1.68819930E-02
0.91919255 1.52271325E-02
0.92929357 1.35287307E-02
0.93939459 1.17822951E-02
0.94949561 9.98299848E-03
0.95959663 8.12566932E-03
0.96969765 6.20477693E-03
0.97979867 4.21442790E-03
0.98989969 2.14834814E-03
1.0000007 -1.25412868E-07
3 AERODYNAMIC OPTIMIZATION 70

xlower ylower
0.0000000 0.0000000
1.01010101E-02 -1.28683383E-02
2.02020202E-02 -1.75581221E-02
3.03030312E-02 -2.07668394E-02
4.04040404E-02 -2.31777728E-02
5.05050495E-02 -2.50679981E-02
6.06060587E-02 -2.65850220E-02
7.07070678E-02 -2.78189518E-02
8.08080807E-02 -2.88295541E-02
9.09090936E-02 -2.96586212E-02
0.10101011 -3.03364489E-02
0.11111112 -3.08855269E-02
0.12121213 -3.13228369E-02
0.13131315 -3.16613354E-02
0.14141415 -3.19109745E-02
0.15151516 -3.20794173E-02
0.16161616 -3.21725719E-02
0.17171717 -3.21949795E-02
0.18181817 -3.21501233E-02
0.19191918 -3.20406705E-02
0.20202018 -3.18686515E-02
0.21212119 -3.16356272E-02
0.22222219 -3.13427970E-02
0.23232320 -3.09911165E-02
0.24242420 -3.05813812E-02
0.25252521 -3.01142931E-02
0.26262623 -2.95905136E-02
0.27272725 -2.90107373E-02
0.28282827 -2.83757132E-02
0.29292929 -2.76862904E-02
0.30303031 -2.69434396E-02
0.31313133 -2.61482857E-02
0.32323235 -2.53021251E-02
0.33333337 -2.44064424E-02
0.34343439 -2.34629177E-02
0.35353541 -2.24734452E-02
0.36363643 -2.14401335E-02
0.37373745 -2.03653071E-02
0.38383847 -1.92515180E-02
0.39393950 -1.81015320E-02
0.40404052 -1.69183426E-02
0.41414154 -1.57051459E-02
0.42424256 -1.44653516E-02
0.43434358 -1.32025648E-02
0.44444460 -1.19205834E-02
0.45454562 -1.06233824E-02
0.46464664 -9.31510143E-03
0.47474766 -8.00003577E-03
0.48484868 -6.68261619E-03
0.49494970 -5.36739454E-03
0.50505072 -4.05902648E-03
0.51515174 -2.76225433E-03
0.52525276 -1.48188532E-03
0.53535378 -2.22772083E-04
0.54545480 1.01020862E-03
0.55555582 2.21217889E-03
3 AERODYNAMIC OPTIMIZATION 71

xlower ylower
0.56565684 3.37828347E-03
0.57575786 4.50371299E-03
0.58585888 5.58373006E-03
0.59595990 6.61369599E-03
0.60606092 7.58909574E-03
0.61616194 8.50556791E-03
0.62626296 9.35893133E-03
0.63636398 1.01452153E-02
0.64646500 1.08606918E-02
0.65656602 1.15019009E-02
0.66666704 1.20656881E-02
0.67676806 1.25492327E-02
0.68686908 1.29500832E-02
0.69697011 1.32661881E-02
0.70707113 1.34959351E-02
0.71717215 1.36381788E-02
0.72727317 1.36922849E-02
0.73737419 1.36581575E-02
0.74747521 1.35362856E-02
0.75757623 1.33277718E-02
0.76767725 1.30343745E-02
0.77777827 1.26585476E-02
0.78787929 1.22034792E-02
0.79798031 1.16731301E-02
0.80808133 1.10722724E-02
0.81818235 1.04065379E-02
0.82828337 9.68245231E-03
0.83838439 8.90747923E-03
0.84848541 8.09006579E-03
0.85858643 7.23968307E-03
0.86868745 6.36687083E-03
0.87878847 5.48328133E-03
0.88888949 4.60172445E-03
0.89899051 3.73621355E-03
0.90909153 2.90201069E-03
0.91919255 2.11567339E-03
0.92929357 1.39510108E-03
0.93939459 7.59582617E-04
0.94949561 2.29844591E-04
0.95959663 -1.71900872E-04
0.96969765 -4.21907025E-04
0.97979867 -4.94843291E-04
0.98989969 -3.63745436E-04
1.0000007 3.47035680E-08
3 AERODYNAMIC OPTIMIZATION 72

The pressure distribution on the surface of the optimized airfoil has been calculated
using surface panel theory. The plot of the pressure distribution has been indicated
in the following figure.

Figure 44: Pressure Distribution on the optimized airfoil

3 AERODYNAMIC OPTIMIZATION 73

Figure 45: Wing geometry of optimized NACA 2411 airfoil

4 CONCLUSION 74

4 Conclusion
A fortran code has been developed to generate general airfoil geometries using PARa-
metric SECtion method. Different kinds of airfoil geometries has been generated using
this code. Pressure distribution and co-efficient of lift of airfoil geometries has been
calculated using a code for panel method. A fortran code for genetic algorithm has
been used for optimization. NACA 2411 airfoil has been generated using PARSEC
and optimized for 5.0 deg angle of attack using genetic algorithm. The optimized
airfoil has the improved co-efficient of lift of 0.9681 as compared to the original one
which has 0.8420. From this results of the project work it is concluded that,

. The usage of PARametric SECtion method has been demonstrated and seems
to be suitable for airfoil geometries.
. An additional parameter lower leading edge radius has been added to the general
PARSEC.
. It shows very high effectiveness in controlling the aerodynamic characteristics of
airfoil.
. Panel method gives reasonable accuracy over the prediction of co-efficient of lift
for low speed, subsonic incompressible flows.
. The effectiveness of the genetic algorithm in finding the best solution among
many possible solutions has been demonstrated.