AERO 2258A THIN AEROFOIL THEORY Lecture Notes

Author : Hadi Winarto

Two-dimensional, incompressible, inviscid and irrotational flow

This note is prepared as lecture material for the course AERO 2258A Fundamentals
of Aerodynamics for the topic of Thin Aerofoil Theory.
It begins with a discussion on the governing equations for 2-dimensional,
incompressible and inviscid flow, which is the Laplace equation. Discrete
singularities that are the elementary solutions of the Laplace equation are then
discussed. It is then followed by a discussion on the linearity property of the Laplace
equation, which leads to a discussion on the concept of continuous singularities such
as the vortex panel with constant and linear vortex strength per unit length
distribution. This is the basic concept behind the panel method, which is explained in
some detail, particularly for a second order vortex panel method. Even though the
panel method is described in some detail, this is not an article on panel method as
such. The concept is introduced so that students will have some familiarity with the
concept of continuously distributed singularity, vortex in particular, which is a
fundamental concept in Thin Aerofoil Theory (TAT). The basic concept in TAT is
that an aerofoil is replaced by a single vortex panel on which there is a continuously
distributed vortex singularities, the strength per unit length of which is unknown and
needs to be evaluated. A good understanding of this basic concept is not possible
without some rudimentary knowledge of a vortex panel. The concept of boundary
conditions, that must be satisfied by the sought for solution, is also discussed. The
mathematics involved is discussed in some detail, but from an engineering point of
view where mathematical knowledge is used as a tool to help in solving engineering
problems. Basic mathematical formulas are assumed as given without showing their
derivations. The manipulation and application of those mathematical formulas are,
however, shown in sufficient details so that students can gain understanding on how
the final engineering equations are obtained and not merely given as formulas to be
memorized blindly. Finally the engineering application of the derived formulas to
calculate the aerodynamic properties of aerofoils is discussed in relatively great
details. The application discussed includes aerofoils, which can be represented by a
flat plate or a cambered (curved) plate, as well as a control surface, which is
represented by a bent flat plate.
Please report any typographical or other errors to the author at the following address
hadi.winarto@rmit.edu.au
Bundoora, Melbourne, 15 March 2004
Last modified 25 May 2004
1. The governing equations
The simplest model of airflow is represented by 2-D, incompressible, inviscid and
irrotational flow. The governing equations for this type of flow consist of 2 partial
differential equations, each representing the conservation of mass or the Continuity
Equation, and the irrotationality condition.

Continuity Equation:

u v
+ =0
x y

(1)

Irrotationality condition:

u v
=0
y x

(2)

Two scalar functions (magnitude only, without direction) can be defined so that the
two components of velocity (vectors), u and v, in the above equations can be replaced
by the equivalent but simpler scalar functions, namely the stream function and the
velocity potential function.
Stream Function, , is defined to satisfy the continuity equation as follows

v=
u=
and
y
x
u v

Therefore
+ =
+

=0
x y x y
y
x

(3)

It can be seen that the stream function automatically satisfies the continuity equation.
Furthermore, the stream function must also satisfy the irrotationality conditions and
thus
u v

2 2
=

= 2 + 2 =0
y x y y
x
x
y
x

(4)

The potential function is defined to automatically satisfy the irrotationality condition

as follows

u=
and
v=
(5)
x
y
u v

thus
=

=0
y x y x
x y
Since the potential function must also satisfy the continuity equation, therefore
u v

2 2
+ =
+
= 2 + 2 =0
x y x x
y y
x
y

(6)

It can be seen that the two first order partial differential equations in 2 unknowns,
namely equations (1) and (2), can be replaced by a single second order elliptic partial
differential equation, namely the Laplace equation either in terms of stream function
(equation (4)) or in terms of the potential function (equation (6)).
Furthermore, it can be shown that the stream function and the potential function are
harmonic conjugate of each other, and thus we can define a complex potential
function as follows
( z ) = ( x, y ) + i. ( x, y )

(7)

where z is the complex variable

z = x + i. y

(8)

and i is the imaginary number

i = 1

(9)

Equations (4) and (6) can be combined as follows

2 ( z ) = 2 ( ( x, y ) + i. ( x, y ) ) = 2 ( x, y ) + i. 2 ( x, y ) = 0

(10)

2. Solution of the Laplace equation

From the theory of complex variable, it is known that the solution of the Laplace
Equation in terms of a complex function is any complex analytic function. We will
not discuss this further except to note that the theory helps us in obtaining a very large
number of elementary solutions of the Laplace equation. Among the very large
number of elementary solutions, also known as singularities, there are four (4) that are
especially useful in the study of aerodynamics, namely the source, sink, doublet and
vortex singularities. Source and sink are actually the same type of singularity except
that they have opposite signs for their strength.
The problem of a uniform airflow, which is disturbed by the presence of an aerofoil
located within the flow field, is modelled by assuming that the disturbance can be
represented mathematically by the singularities, solutions of the Laplace equation.
The Laplace equation is said to be linear, meaning that a linear combination of some
simpler or elementary solutions is also a solution.
Let us now have a quick look at the elementary solutions
Source / sink :

singularity located at ( x0 , y0 ) with a strength of

( x, y ) =

Stream function induced at (x,y):

Potential function induced at (x,y):

Doublet :

( x, y ) =

y y0

. tan 1
2
x x0

2
2
.ln ( x x0 ) + ( y y0 )
2

(11)

1
2

(12)

singularity located at ( x0 , y0 ) with a strength of

y y0

.
2 ( x x0 ) 2 + ( y y0 )2

Stream function induced at (x,y):

( x, y ) =

Potential function induced at (x,y):

( x, y ) =

x x0

.
2 ( x x0 )2 + ( y y0 )2

(13)

(14)

Vortex :

singularity located at ( x0 , y0 ) with a strength of

Stream function induced at (x,y): ( x, y ) =

2
2
.ln ( x x0 ) + ( y y0 )
2

( x, y ) =

Potential function induced at (x,y):

1
2

y y0

.tan 1
2
x x0

(15)

(16)

The velocity components induced at (x,y) by the presence of the singularity at

( x0 , y0 ) can be evaluated as follows.

x-component of velocity:

u=

=
x y

(17)

y-component of velocity:

v=

=
y
x

(18)

Source / sink :

singularity located at ( x0 , y0 ) with a strength of

x-component:

u ( x, y ) =

x x0

.
2 ( x x0 )2 + ( y y0 ) 2

(19)

y-component:

v ( x, y ) =

y y0

.
2 ( x x0 )2 + ( y y0 )2

(20)

Doublet :

singularity located at ( x0 , y0 ) with a strength of

x-component:

( x x0 ) ( y y0 )
u ( x, y ) =
.
2
2
2
( x x0 ) + ( y y0 )

y-component:

v ( x, y ) =

2 ( x x0 )( y y0 )

.
2
2
2
( x x0 ) + ( y y0 )

Vortex :

(21)

(22)

singularity located at ( x0 , y0 ) with a strength of

( y y0 )

.
2 ( x x0 )2 + ( y y0 )2

x-component:

u ( x, y ) =

y-component:

v ( x, y ) =

( x x0 )

.
2 ( x x0 )2 + ( y y0 )2
4

(23)

(24)

The mathematical expressions for the stream and potential functions of a uniform
flow, which is inclined at an angle to the horizontal axis x, are given as follows
Stream function value induced at (x,y):

( x, y ) = V . ( y.cos x.sin )

(25)

Potential function value induced at (x,y)

( x, y ) = V . ( x.cos + y.sin )

(26)

The x- component of velocity is:

u ( x, y ) =

=
= V .cos
x y

(27)

The y- component of velocity is:

v ( x, y ) =

=
= V .sin
y
x

(28)

The quantity V is the speed of the undisturbed air or the free-stream air.
3. Linearity property of the Laplace equation
It was previously mentioned that the Laplace Equation is a linear second order partial
differential equation. The linearity property of the equation implies that any linear
combination of some elementary solutions is also a solution. It was mentioned also
that the problem of air flow over an aerofoil (or in American terminology: airfoil) is
modelled as a uniform air flow which is disturbed by the presence of the aerofoil. We
shall now have a look whether the aerofoil can be represented by a single source, or a
single doublet or a single vortex. For simplicity it will be assumed that the value of
the angle is zero.
Linear combination of uniform flow and a source:

( x, y ) = V . y +

y y0

.tan 1
2
x x0

(29)

Linear combination of uniform flow and a doublet:

( x, y ) = V . y +

y y0

.
2 ( x x0 )2 + ( y y0 )2

(30)

Linear combination of uniform flow and a vortex:

( x, y ) = V . y +

2
2
.ln ( x x0 ) + ( y y0 )
2

1
2

(31)

The flow pattern of a flow field is defined by the streamlines of the flow.
For a steady flow, i.e. one that does not change with time, if we release a particle at a
point in the flow field and then follow the motion of the particle as it is swept
downstream, the path of the particle motion is known as a streamline. Another particle
released at another point would describe another streamline. By drawing a number of

streamlines, we can get a picture of the shape of the flow field. For an unsteady flow
or turbulent flow, obviously the shape of the streamlines will change continuously
with time. However, for a steady flow the pattern of the flow field is constant.
Another property of the streamline is the fact that at any given point on the streamline,
the direction of the velocity vector is tangent to the streamline at the given point.
Therefore no flow can cross a streamline. Furthermore, the stream function value at
any point on the streamline is a constant. In other words, we can define a streamline
as being the locus of points within the flow field where the values of the stream
function at all points on the streamline are the same. Thus a streamline can also be
called a constant stream function curve
Now let us consider what sort of a flow field we get from the combination of a
uniform flow, which is disturbed by a doublet.
To simplify the discussion let us assume that the doublet is located at the origin of the
system of axes, or at x0 = 0 and y0 = 0 . Furthermore, in order to get a meaningful
result it is assumed that the sign of the doublet strength is negative. With those
assumptions equation (30) can now be written as follows

( x, y ) = V . y

where

R2 =

y
R2
. 2
=
V
.
y
.
1

2 x + y 2
x2 + y 2

is a positive constant
2 V

(32)

(33)

Let us now have a look at a particular streamline with a value of stream function of
zero. If = 0 then equation (32) is simplified to become
y. 1

R2
=0
x2 + y 2

There are 2 solutions to the above equation, namely

y=0
and also
R2
1 2
=0
x + y2

(34)

(35)

The above equation can be simplified further as follows

x2 + y2 = R 2

(36)

Equation (36) is the equation of a circle whose centre is located at the origin and its
radius is R. Since fluid can not cross a streamline, therefore a streamline may be
replaced by an impermeable wall. Therefore equation (32) actually represents the flow
field of a uniform flow which is flowing over a circular cylinder with radius R. The
streamline with a stream function value of = 0 , which includes the x-axis or
equation (35) and the circle given by equation (36), is known as the dividing
streamline. This streamline divides the flow field into 2 parts, namely one that flows

over the upper part of the circle and another which flows over the lower part of the
same circle or circular cylinder.
All the other streamlines, which describe the flow over the upper part of the circle,
can be obtained by solving the following equation
y. 1

where

R2

=
2
2
x +y
V

(37)

is a constant and
> 0.
V
V
Similarly, all the other streamlines that describe the flow over the lower part of the

circle are solutions of equation (37) where

< 0.
V
If we evaluate the coordinates (x, y) of a large number of points within the flow field
for a particular value of /V , and all the neighbouring points are connected to each
other by short straight lines, then we will get the shape of the streamline for that
particular value of /V . The flow pattern that we wish to analyse is then given
visually as a collection of streamlines for various values of /V .
A doublet represents the disturbance of a circular cylinder immersed in a uniform
flow, whereas a source or a vortex represents another type of disturbance.. However
none of those would represent the disturbance of an aerofoil. The flow pattern of a
uniform wind being disturbed by a source or a vortex will not be discussed here, but
they can be readily found in most textbooks on aerodynamics.
It is important to note here that whilst the flow around a circular cylinder is not
particularly important in aerodynamics, however we can employ the theory of
complex variables to transform the flow field around a circular cylinder into that of
the flow around an aerofoil. In the simplest case, the transform function or mapping
function is assumed known and the shape of the aerofoil is obtained as a result.
Perhaps the most well known mapping function is the Joukowskis transformation,
which really is a special form of the more general Karman-Trefftz mapping function.
If we want to obtain the flow field around any arbitrary shape aerofoil, then we will
need to evaluate the mapping function. This is a far more difficult problem than for
Joukowski or Karman-Trefftz mapping, and will not be discussed here. It is sufficient
to note that one of the possible methods to use is the Laurent Series transformation.
The use of complex variable transformation is known as the conformal mapping
method (see appendix 4).
4. Discrete and Continuous Singularities
Let us now return to our original statement, which is that a more complex flow can be
obtained by adding together a number of simpler solutions and see if this can be used
to obtain a more direct solution to the problem of evaluating the flow field of a
uniform wind which is disturbed by an arbitrary shape aerofoil. This more direct
approach is known as the panel method. Here we will discuss the basics of the 2-D
panel method only, but the same approach may be applied to 3-D problems. This is in

contrast to the conformal mapping method, which is only applicable for 2-D
problems.
The stream function value at any point (x, y) in a uniform flow, which is disturbed by
2 vortices, each being located at ( x1 , y1 ) and ( x2 , y2 ) respectively, is given by the
following expression

( x, y ) = V . y +

1
2
2
.ln ( x x1 ) + ( y y1 )
2

1
2

2
2
2
.ln ( x x2 ) + ( y y2 )
2

1
2

(38)

If there are N, rather than 2, vortices then we have

( x, y ) = V . y +

N
n =1

n
2
2
.ln ( x xn ) + ( y yn )
2

1
2

(39)

Let us now consider the situation where there are 2 points, P and Q, with coordinates
of (XP, YP) and (XQ, YQ). The straight line from P to Q is divided up into a large
number of equal length interval of s = PQ / N , where PQ is the distance from P to
Q. Let us now imagine that at the mid point of each small interval there is a vortex of
strength G, which is the same for all intervals. The stream function value at any point
(x, y) for this case is given by

G
2
2
( x, y ) = V . y +
.ln ( x xn ) + ( y yn )
n =1 2
N

1
2

(40)

It should be noted that even though N can be made to be very large, approaching
infinity, we shall impose the condition that the total strength of the vortices is always
the same regardless of the actual value of N, and this total vortex strength is where
= N .G

Let us denote the mid point n as being the point Sn , such that the coordinates of the
point Sn is ( XS n , YS n ) where

X = (XQ XP) / N =

XPQ / N

Y = (YQ YP) / N =

YPQ / N

s = PQ / N = XPQ 2 + YPQ 2 / N
XS n = XP + n

1 XPQ
.
. s = XP + XPS n
2
PQ

YS n = YP + n

1 YPQ
.
. s = YP + YPS n
2
PQ

G is the strength of the discrete vortex located at the midpoint of an elemental length
of s . If N is chosen to be sufficiently large, then s is sufficiently small to be
replaced by a continuous differential, i.e. s ds . Furthermore we shall assume that
the strength of the vortex, G, is distributed evenly along ds and hence

G=

PQ
= .
= . s .ds
N
N

Here is the strength per unit length of the continuous vortex sheet ds. Since G is the
same for all elemental lengths, ds, therefore is also the same for all ds along the
vortex sheet or panel PQ.
Equation (40) can now be written as follows

( XT , YT ) = V .YT +

. s
2
2
.ln ( XT XS n ) + (YT YS n )
n =1 2
N

1
2

(41)

where (XT, YT) are the coordinates of the point T, at which the value of the stream
function is to be calculated.
Now it is to be noted that
XT XS n = XT XP XPS n = XPT XPS n
YT YS n = YT YP YPS n = YPT YPS n
Therefore,
S nT 2 = ( XT XS n ) + (YT YS n ) = PT 2 2. ( XPT .XPS n + YPT .YPS n ) + PS n2
2

where it is defined that

PT 2 = XPT 2 + YPT 2
PS n2 = XPS n2 + YPS n2

Equation (41) can now be written as follows

( XT , YT ) = V .YT +

N
.ln S nT
. s
.ln S nT = V .YT +
.s
2
n =1 2
n =1
N

Taking the limit of N approaching infinity, the above summation can be replaced by
the following integral

( XT , YT ) = V .YT +

PQ

ln r ( s ). ds
2

(42)

r 2 = PT 2 2. ( XPT . XPS + YPT . YPS ) + s 2

where r(s) is the distance from point S, anywhere along the line PQ, to the point T,
and s is the distance from point P to point S.
It can also be observed that
XPS =

s
.XPQ
PQ

YPS =

and

s
.YPQ
PQ

Therefore
r 2 = PT 2 2. PR . s + s 2 = TR 2 + ( PR s )

PR = ( XPT . XPQ + YPT . YPQ ) / PQ

TR = PT 2 PR 2

(43)
(44)
(45)

Since is a constant, therefore equation (42) can be simplified to become

( XT , YT ) = V .YT +
2

PQ

ln r ( s ). ds

(46)

It an be shown (see appendix 1) that

Q

ln r ( s ) .ds = ( PQ PR ) .ln QT + PR.ln PT PQ + TR. APTQ

(47)

where

APTQ = tan 1

PQ.TR
PT PQ.PR
2

(48)

It should be noted that APTQ is the angle subtended by the lines PT and TQ, or the
visible angle of PQ seen from T.
Equation (42) is the value of stream function at point T, which is immersed in the
flow field of a uniform flow with a free stream velocity of V , and is influenced by
the presence of a continuously distributed vortex. The vorticity is distributed on a line
PQ, which is actually the intersection of rectangle with the x-y plane or the plane of
the paper. The rectangle is perpendicular to the x-y plane and is infinitely long in the
z-direction. It is referred to as a panel and because vorticity is distributed on it,
therefore, it is called a vortex panel. The vortex strength per unit length distribution
along PQ is ( s ) , which is a function of the variable length representing the distance
from P to a point S located anywhere between P and Q. Generally speaking the
functional form of ( s ) is unknown and represents the problem that has to be solved.
The above discussion shows that the Laplace equation can have a discrete solution as
well as a continuous solution. In terms of the stream function equation (15) represents
10

the solution of the stream function value at any arbitrary point (x, y) induced by a
discrete vortex with strength of and located at ( x0 , y0 ) . The solution of the same
problem in terms of the potential function is given by equation (16), whereas for the
velocity components u and v the solutions are equations (23) and (24) respectively.
If the discrete vortex is replaced by a distributed vortex along panel PQ, with vorticity
strength per unit length of ( s ) , the solutions are given as follows

( x, y ) =

1
( s ).ln r ( s ) .ds
2 P
Q

1
( s ). ( s ) .ds
2 P

(50)

sin ( s )
1
. (s)
.ds
2 P
r (s)

(51)

( x, y ) =
u ( x, y ) =

(49)

Q
cos ( s )
1
v ( x, y ) =
( s ).
.ds
2 P
r (s)

(52)

where
s=

( XS XP ) + (YS YP )

r (s) =

(( x XS )

( s ) = tan 1

+ ( y YS )

)
)

y YS
x XS

1
2

1
2

(53)
(54)
(55)

It should be remembered that (XS, YS) are the coordinates of the point S, which is
located anywhere along the panel PQ. The coordinates of points P and Q are (XP, YP)
and (XQ, YQ) respectively.
Similar expressions can also be derived for the case where the singularity is source
and sink or doublet, rather than vortex. The final results are given below
Source and Sink:
Q

1
( x, y ) =
( s ). ( s ) .ds
2 P

11

(56)

1
( x, y ) =
( s ).ln r ( s ) .ds
2 P

(57)

cos ( s )
1
u ( x, y ) =
( s ).
.ds
2 P
r (s)

(58)

sin ( s )
1
v ( x, y ) =
( s ).
.ds
2 P
r (s)

(59)

sin ( s )
1
( x, y ) =
( s ).
.ds
2 P
r ( s)

(60)

Doublet :
Q

cos ( s )
1
.ds
( s).
2 P
r ( s)

(61)

u ( x, y ) =

cos 2 ( s )
1
( s ). 2
.ds
2 P
r ( s)

(62)

v ( x, y ) =

Q
sin 2 ( s )
1
.ds
( s ). 2
2 P
r ( s)

(63)

( x, y ) =

It should be noted that ( s ) and ( s ) are the strength per unit length of the source
and doublet distribution respectively.
It can be seen clearly that the integrals involved in the above equations are very
complex and cant be solved unless the strength per unit length distribution of the
chosen singularity (source, doublet or vortex) is given. Furthermore, even for the
simplest case where the strength per unit length distribution is just a constant it is
already quite difficult to evaluate the explicit expression for the integral. Another
difficulty is the geometry of the singularity panel. If the panel is not straight but
curved instead, then it is impossible to obtain an analytical solution for the integral.
On the other hand we know that wings or aerofoils are not flat panels, but are highly
curved. In the panel method this problem is overcome by replacing the continuously
smooth curve of the aerofoil with an approximate curve consisting of a large number
of panels or straight lines connecting adjacent points on the surface of the aerofoil.
The overall effect of the whole aerofoil on the value of stream function, or any of the
other functions that is chosen, at a point is then obtained as the sum of the effects of
all panels representing the aerofoil. Since each panel is a straight line therefore it is
possible to derive the expression for the required integral, provided that the strength
per unit length distribution of the singularity is kept simple. If the distribution is
assumed constant along the panel, then the method is known as the first order panel
method, since the distribution only requires the knowledge of one unknown constant.
A much better approximating distribution is given by a linear function of s, which
12

involves two unknown constants, hence it is referred to as a second order panel

method. Below we will show the derivation of the expression for the integral on the
right hand side of equation (49) for a second order panel method.
The vortex strength per unit length linear distribution is given by the following

( s) = P +

Q P
PQ

(64)

.s

where P and Q are the values of ( s ) at points P and Q respectively, and

represent the two unknown constants of the second order panel method.
The stream function value at (x, y) for this case is thus given by

( x, y ) =

1
2

P+

Q P

PQ

s .ln r ( s ) .ds

(65)

The above equation can be rewritten as follows

( x, y ) = CI1. P + CI 2 . ( Q P ) = ( CI1 CI 2 ) . P + CI 2 . Q

(66)

CI1 = ln r ( s ).ds

(67)

P
Q

PQ.CI 2 = s.ln r ( s ).ds

(68)

The expression for the integral in equation (67) has already been worked out in
appendix 1 with the folowing result
CI1 = ( PQ PR ) .ln QT + PR.ln PT PQ + TR. APTQ

(69)

Details for the evaluation of the second integral are given in appendix (2) the result of
which is
CI 2 =

1
2

PQ 2 ( PR 2 TR 2 )
PQ

ln QT +

PR 2 TR 2
ln PT ( PQ + PR )
PQ

(70)

PR.TR. APTQ
+
PQ
Our discussion so far only deals with the solution of the Laplace equation in general,
whereas the real problem that we want to solve is how to analyse the aerodynamic
properties of a 2-dimensional wing or an aerofoil. The mathematical model for our
simplified problem, i.e. confined to inviscid, incompressible flows only, is indeed the
Laplace equation, but we have not discussed about the boundary conditions that must
be satisfied. To obtain a unique solution we need to specify what boundary conditions
the solution must satisfy.

13

5. The boundary conditions

The real question we want to find the answer for here is that if the shape of a wing or
an aerofoil is given, and the wing is immersed in a given airflow, what is the lift and
drag acting on the aerofoil as a result of its interaction with the flowing air. It should
be remembered that in a Galilean transformation, it makes no difference if the body is
stationary and the air is moving (model aircraft in a wind tunnel) or whether the body
is moving in a stagnant atmosphere (aircraft moving in the atmosphere).
Drag and lift are the two orthogonal components of the resultant force acting on the
aerofoil, which is the summation of all the pressure acting on the surface of the
aerofoil. From Bernoulli equation we know that the total pressure in an isenthalpic
flow is constant, hence the static pressure decreases if the dynamic pressure increases
and vice versa. The dynamic pressure is of course half multiplied by air density
multiplied by velocity squared. Thus pressure is directly related to the fluid velocity
squared.
Imagine an airflow moving uniformly from the left to the right. Everywhere within
the flow field the velocity is the same, i.e. the free stream velocity, thus the static
pressure is also the same with a value of free stream static pressure. Now imagine that
suddenly an aerofoil is inserted into the airflow. Obviously the presence of the
aerofoil would disturb the uniform airflow, and the velocity field would change from
the previously uniform value, at least in the vicinity of the aerofoil surface.
On the other hand we have also seen that the singularities, which are the elementary
solutions of the Laplace equation, also has the effect of disturbing a uniform flow of
fluid, at least in the vicinity of the location of the singularity. From this observation
we can make the conclusion that perhaps the airflow around an aerofoil can be
simulated by placing singularities on the aerofoils surface, or on a surface that can be
assumed to be representative of the aerofoil surface.
It is an observed fact that air cant penetrate into the inside of the aerofoil, the surface
of which is made out of solid. Since the surface of the aerofoil represents part of the
boundary of the airflow, therefore the requirement that air cant penetrate into the
aerofoils surface is called a boundary condition. The aerofoils surface is called the
inner boundary since it represents the boundary of the inner part of the flow. The
other boundary is the outer boundary, which ideally is infinitely far away from the
aerofoils surface, but from a practical point of view may be defined as being
sufficiently far away from the surface. At the outer boundary the flow is undisturbed
by the presence of the aerofoil, therefore in our simulated flow the Laplacian
singularities placed on the surface also must not disturb the free stream uniform flow
far away from the singularities. This requirement is always automatically satisfied by
all of the singularities, namely source/sink, doublet and vortex. Therefore, it is only
the inner boundary condition that must be satisfied in our simulated flow, i.e. that the
flow must not penetrate the aerofoils surface. This requirement can be expressed in
either one of at least two different ways.
Firstly, it may be stated that the surface of the aerofoil is part of the dividing
streamline that splits the airflow into an upper half and a lower half of the flow field.
Since a streamline is a curve on which the value of the stream function is constant,
this type of boundary condition is called the Dirichlet Condition where the value of

14

the function (the stream function) is defined (as being of a constant value) at the
boundary of the flow field.
The other way of specifying the boundary condition is that the normal component of
the flow velocity at the boundary must be zero, since otherwise it would imply that air
is allowed to penetrate into the aerofoils surface. Since flow velocity is the derivative
of the function (i.e. the stream function) this is the same as specifying the values of
the functions derivative at the boundary and is known as the Neuman Condition.
Generally speaking, in some problems the function values are specified at some parts
of the boundary, while at the rest of the boundary the derivative values are specified.
This third type of boundary condition is known as the Mixed Boundary Condition,
which is also known as the Robin Condition.
Now we can begin to describe how our problem is to be simulated using a second
order vortex panel method.

Figure 1. Definition of parts of an aerofoil (from Georgia Institute of Technology web

site http://www.adl.gatech.edu/classes/lowspdaero/lospd5/lospd5.html)
The undisturbed flow is represented as a uniform flow the velocity vector of which is
inclined at an angle of relative to the horizontal or x-axis. The aerofoil is fixed in
space and is so located such that its chord line is along the x-axis, with the nose being
at the origin and the tail being to the right of the nose. The aerofoils surface is
represented by a large but finite number of points on it. The smoothly continuous
curved surface is approximated by a piecewise linear segmented continuous surface
consisting of small segments of straight lines or panels. Obviously the approximation
gets better as the number of points on the aerofoils surface is increased. However,
this has the implication of increasing amount of computational work to be done.
Since in reality all wings have tails with finite thickness, it will be assumed here that
the aerofoil has a blunt trailing edge, with distinct lower and upper tail points. The
lower trailing edge (tail) point is identified as the point P1 and the upper tail point is
the point PN , where N is the number of points representing the aerofoils surface. The
number of panels is obviously (N-1). Indexation of all surface points is done in a
clockwise direction, thus the point next to and to the left of the first point is identified
as the point P2 etc. The nth panel is the line connecting point Pn to Pn +1 , which is also
identified as the point Qn . Thus the nth panel is also referred to as the panel PQn .

15

In the following discussion we will describe a solution based on the method that
utilizes the Dirichlet boundary conditions. Thus the solution obtained will be in terms
of the stream function.
The stream function value at a point T with coordinates (XT,YT) due to T being
immersed in a uniform flow, which is inclined at an angle of to x-axis, is given as
follows

uw ( XT , YT ) = V . ( XT .sin + YT .cos )

(71)

The stream function value at T induced by the vortex panel PQn is given by equation
(65), where the vortex strength per unit length distribution is assumed given by a
linear distribution, i.e. we shall use a second order method. Thus along the panel PQn
we have the following distribution

n ( s ) = Pn +

Pn +1 Pn
PQn

(72)

.s

For this case equation (65) can be rewritten as follows

PQn =

Pn+1

Pn +

1
2

Pn+1 Pn
PQn

Pn

PQn =
PQn =

1
2

1
2

( CI

1n

.s .ln r ( s ) .ds

(73)

. Pn + CI 2 n . ( Pn +1 Pn ) )

( ( CI

1n

CI 2 n ) . Pn + CI 2 n . Pn +1 )

(74)

The expressions for CI1n and CI 2n are given by equations (69) and (70) as follows
CI1n = ( PQn PRn ) .ln PTn+1 + PRn .ln PTn PQn + TRn . APTQn
CI 2 n =

1
2

PQn2 ( PRn2 TRn2 )

PQn

PRn2 TRn2
ln PTn ( PQn + PRn )
PQn

ln PTn +1 +

(75)

(76)

PRn .TRn . APTQn

PQn

The stream function value at T induced by all the vortex panels making up the
complete aerofoil shape is given by

vs ( XT , YT ) =

N 1
n =1

PQn =

1
2

N 1
n =1

( ( CI

1n

16

CI 2 n ) . Pn + CI 2 n . Pn +1 )

(77)

If it is now defined that

CT ,1 = CI11 CI 21
CT ,n = CI1n CI 2 n + CI 2 n1

for n = 2,3,..., N 1

(78)

CT , N = CI 2 N 1
then equation (77) can be rewritten more compactly as follows

vs ( XT , YT ) =

1
2

N
n =1

CT ,n . Pn

(79)

The value of the stream function at point T is the sum of the stream function due to
being immersed in the uniform wind plus the stream function induced by all the
vortex panels making up the surface of the aerofoil

T = uw ( XT , YT ) + vs ( XT , YT )
(80)

T = V . ( XT .sin + YT .cos ) + 21

N
n =1

CT ,n . Pn

The above equation can be rewritten as follows

N
n =1

CT ,n . Pn 2 . T = 2 .V . ( XT .sin YT .cos )

(81)

If the coordinates of all points on the aerofoils surface ( XPn , YPn ) and the coordinates

of point T, i.e. ( XT , YT ) , are given then all the influence coefficients, CI1n and
CI 2n hence CT , n for all values of n from 1 to N can be evaluated.
Now let us define a dummy value of N +1 as follows

PN +1 = T

(82)

and also a dummy coefficient

CT , N +1 = 2

(83)

then equation (81) can be written more compactly as follows

N +1
n =1

CT ,n . Pn = 2 .V . ( XT .sin YT .cos )

17

(84)

All the values of CT , n and the right hand side of the above equation are known and the
only unknowns are Pn for n = 1 to n = N+1 provided that the control point T is
chosen to be on the aerofoils surface such that T is a constant. Since there are N+1
unknowns therefore we need N+1 equations to be solved simultaneously to calculate
the values of the unknowns, Pn . However there are only N points on the surface of
the aerofoil that can be selected to be the control points where equation (84) is
applied, thus we need one more equation to complete the system of equations to be
solved simultaneously.
The extra equation is obtained from the physical observation that the airflow leaving
the upper surface of the aerofoil must have exactly the same velocity as the airflow
leaving the lower surface. This means that the velocity at the two tail points (upper
and lower tail points) must be the same. It can be shown that the airflow velocity at an
aerofoils surface point is exactly the same as the value of the vortex strength per unit
length, P , at the point. This trailing edge flow condition is known as the Kutta
condition and can be represented by the following equation

P1 + PN = 0

(85)

For each of the control point, which is chosen to be the aerofoils surface point P, we
can write down an equation based on the general expression of equation (84). The
( N + 1)th must be derived differently, namely it is based on satisfying the Kutta
condition. Now we can define the following for the ( N + 1)th equation

CT ,1 = 1
CT ,n = 0 for n = 2,3,..., N 1
(86)
CT , N = 1
CT , N +1 = 0
With the above definitions we now have a system of equations consisting of (N+1)
equations involving (N+1) unknowns as follows
N +1
n =1

CT ,n . Pn = Dn

(87)

where
Dn = 2 .V . ( XTn .sin YTn .cos ) for n = 1, 2,...., N
(88)
DN +1 = 0
18

The system of equations (87) can be solved simultaneously to calculate the unknown
Pn and since the absolute value of Pn is the same as the airflow velocity at the
point Pn , therefore the distribution of flow velocity along the aerofoils surface is
known. Furthermore, it can be shown that for vortex panel method as described here,
the value of pressure coefficient Cp can be obtained from the Bernoulli equation and
the final result is
Cp = 1

(89)

The pressure coefficient at point on the surface of the aerofoil can thus be calculated
and plotted as desired. The pressure coefficients can also be integrated to give the
resultant force acting on the aerofoil. This resultant force can be resolved into 2
components, one being the lift force in the direction normal to the free stream
direction and the other is the drag force acting along the free stream direction. Due to
the inviscid flow assumption, it is expected that the drag force or drag coefficient
must have a value of zero. In practice the computed drag coefficient will be found to
have a non-zero value due to computational error such as round off error etc. The
moment acting on the aerofoil can also be obtained from the known pressure
distribution. A simple example of the application of the panel method is given in
appendix 3.
Even though the panel method is quite good for computing the aerodynamic
properties of an aerofoil, it is basically a numerical method and doesnt give analytical
insight into the aerodynamic behaviour of an aerofoil. To get such an insight we need
an analytical tool, even if it is very much simplified. This tool is known as the Thin
Aerofoil Theory, which is the next topic to be discussed.

6. Thin Aerofoil Theory

The camber line of an aerofoil is the curve midway between the lower and upper
surfaces of the aerofoil. For a symmetric aerofoil the camber line is simply a straight
line. If a line is drawn perpendicular to the camber line, then it must intersect both the
upper and the lower surfaces of the aerofoil. The distance between the two
intersection points is called the thickness of the aerofoil. Obviously this thickness
would vary along the chord of the aerofoil. The thickness is normally expressed as a
percentage of the chord length. The thickness of an aerofoil is defined as the
maximum thickness as described previously. A thin aerofoil is defined as any aerofoil
whose (maximum) thickness is a very small percentage of the chord length, such that
it is reasonable to model the aerofoil as a curved or flat plate of zero thickness. While
this restriction is quite severe, it enables us to get an analytical solution to the
problem. Here we sacrifice accuracy to get analytical insight.

19

6.1 Flat plate as an aerofoil

Let us now consider a very thin symmetric aerofoil, which is quite reasonable to be
approximated as a flat plate of zero thickness, with a chord length c. The aerofoil is
located along the x-axis with its nose being at the origin of the system of coordinates.
The aerofoil is immersed in a uniform wind whose velocity vector is at an angle
relative to the horizontal or x-axis. The disturbance to the uniform wind due to the
presence of the aerofoil is modelled as the disturbance caused by a vortex sheet
located along the camber line of the aerofoil. The vortex strength per unit length
distribution of the sheet ( x ) is unknown and must be evaluated.
The stream function at any point ( x* , y * ) within the flow field is given as follows
c

( x* , y* ) = V ( x*.sin + y * .cos ) + 21 ( x ).ln r ( x ) .dx

(90)

r ( x) =

(x

x ) + ( y* y )
2

(91)

The velocity components at any point are given by the following

c

sin ( x )
.dx
r ( x)

(92)

cos ( x )
.dx
r ( x)

(93)

u ( x* , y* ) = V .cos + 21 ( x ) .
0

v ( x* , y * ) = V .sin 21 ( x ) .
0

sin ( x ) =

y*
x* x
and cos ( x ) =
r ( x)
r ( x)

(94)

If the problem is specified as a Dirichlet boundary condition problem, then we must

find the analytical expression for ( x ) such that equation (90) is satisfied at all
control or boundary points. The boundary condition to be satisfied is that all the
points on the camber line, or part of x-axis from 0 to c, must have the same stream
function value because the camber line is a streamline. It is quite easily seen that this
problem is impossible to solve analytically.
Now let us recast the problem as a Neuman boundary condition problem. The
boundary condition to be satisfied is that the component of velocity normal to the
camber line, v ( x* , y* ) , must be zero at all points along the camber line. The equation
to be solved is thus equation (93), which can be rewritten as follows
c

V .sin 21 ( x ) .
0

cos ( x )
.dx = 0
r ( x)

20

(95)

Since the control point is on the camber line thus y* is always zero. Therefore
cos ( x ) = 1 and r ( x ) = x* x .
Equation (95) is then simplified to
c
1
2
0

( x ) .dx
x* x

= V .sin

(96)

It can be seen that this problem looks simpler than the Dirichlet formulation of the
same problem. Even so the problem is not quite so simple.
We know that any continuous function can always be approximated by a Fourier
series, even if the exact expression for the function is unknown. Therefore, it is
reasonable to replace the unknown function ( x ) by a Fourier series with unknown
coefficients. However, before we can do that it should be realized that the Fourier
series is expressed in terms of angles rather than variable such as x. Therefore, it is
necessary that we perform a coordinate transformation from x to . It is required that
when x = 0 then = 0 and when x = c we want = .
A suitable transformation function can be given as follows
x=

c
2

(1 cos )

(97)

dx = c2 sin .d
Equation (95) can now be rewritten as follows
1
2

( ) .sin
d = V .sin
cos cos *
0

(98)

We will not go into the mathematical details of how to solve the above equation. It is
sufficient to simply apply the known mathematical results to help find the solution.
For example it is known that

cos n .d
sin n *
=

cos cos *
sin *
0

for n = 0,1, 2,3....

(99)

sin n .sin .d
= cos n *
*
cos

cos

for n = 0,1, 2, 3....

We shall now assume a solution and then substitute the solution into equation (96)
and verify that the trial solution indeed satisfies the equation or otherwise.
The trial solution is

( ) = 2V .sin .

1 + cos
sin

(100)

The left hand side of equation (96) can now be expanded as follows
21

1
2

( ) .sin
1 2V .sin . (1 + cos ) sin
d =
.d
cos cos *
2 0 sin ( cos cos * )
0

(101)
1
2

( ) .sin
V .sin
1 + cos
.d
d =
*
cos cos

cos cos *
0
0

The first equation in (99) for n = 0 and n = 1 gives the following results

d
=0
cos cos *
0

(102)

cos .d
=
*
0 cos cos

Therefore equation (98) can be simplified as follows

V .sin

1 + cos
V .sin
.d =
. ( 0 + ) = V .sin
*
cos cos

(103)

Thus it has been proven that the trial solution (100) indeed satisfies the governing
equation (96).
In the previous section it was stated that the flow solution must also satisfy the Kutta
condition at the trailing edge or at = (since x = c). It should be noted that the
Kutta condition for an aerofoil with a sharp trailing edge is that the trailing edge must
be a rear stagnation point, where = V = 0 . Substituting the value of = into (100)
we get the following result

( ) = 2V .sin .

0
(indeterminate value).
0

The above value is indeterminate and we should apply the LHospitals rule. The rule
states that if the ratio of 2 functions, say f(z)/g(z), becomes indeterminate as z
approaches 0, then the value can be calculated by replacing the functions with their
derivatives. Thus in the limit of z 0, the value of f ( 0 ) / g ( 0 ) is given
by f '( 0 ) / g '( 0 ) . In our example both f and g approach 0 as . Therefore the

required value should be calculated as f '( ) / g '( ) .

Now it is noted that
d (1 + cos )
= sin
d

and

d sin
= cos
d

22

Therefore

( ) = 2V .sin .

sin
=0
cos

It can be seen that the solution also satisfies the Kutta condition.
We can also obtain the solution as a function of the Cartesian variable, x, as follows.
From equation (97) we can get the following
1 + cos = 2 (1 x / c )
sin = 1 cos 2 = +2 x (1 x ) c 2
Therefore

( x ) = 2V .sin .

cx
x (c x)

= 2V .sin .

cx
x

(104)

At the trailing edge x = c, therefore ( c ) = 0 , thus satisfying the Kutta condition.

The values of * =

V .2sin

can be calculated as a function of x/c and the results are

tabulated below
x/c

0.1

0.2
2

0.3
1.528

0.4
1.225

0.5
1.000

0.6
0.816

0.7
0.655

0.8
0.500

0.9
0.333

1.0
0.00

The lift, L, acting on the flat plate can be calculated using the following KuttaJoukowski lift equation
L = V

(105)

The circulation around the aerofoil is equal to the total strength of the distributed
vortex, which can be obtained by integrating or summing up all the values of ( x ) .
c

= ( x ).dx =
0

c
( ).sin .d
20

(106)

= cV sin (1 + cos )d = cV sin

0

The lift coefficient is defined as follows

23

Cl =

1
2

L
= 2 sin
V2 .c

(107)

For small values of (in radian) the following approximation is quite accurate
sin

(108)

Therefore the lift coefficient can be rewritten as follows

Cl = 2 .

(109)

The force dF acting on an aerofoil elemental length dx is given by the KuttaJoukowski lift equation as follows
dF = V ( x ) .dx

(110)

The pitching moment about the leading edge due to the force dF is
c

M LE = x.dF = V ( x ) .x. dx

M LE = V .2V
0

M LE =

1 + cos c
c
. (1 cos ) . sin d
sin 2
2

c2

V2 = V2c 2
2
2
4

(111)

The pitching moment about any other point can be obtained quite easily by
remembering the definition of moment, which is simply force multiplied by the arm
length. The pressure distribution along the aerofoil surface is such that it can be
replaced by a resultant moment and a resultant force (lift force). There is a particular
location or value of x, where the resultant moment of the pressure distribution is zero.
This zero resultant moment point is known as the centre of pressure and its location
can be calculated as follows.
Let us assume that the centre of pressure is located at xcp . At this point the resultant
moment is zero and the resultant force is given by equation (105). Now imagine that
an equal but opposite force is located at the same point. To cancel this force we must
add another force, which is equal to the resultant force but located elsewhere, say at
the leading edge. Adding two equal but opposite forces doesnt change the resultant
force acting on the aerofoil, but those forces which is called a couple is equivalent to a
moment. The magnitude of this moment of the couple is simply xcp multiplied by the
lift force. However, to get the moment acting at the centre of pressure to remain zero
we must add a moment which is equal but in opposite direction of the moment due to
the couple. Now we have the lift force at the centre of pressure, a couple of forces the
magnitude of which is the same as the lift force and a moment, which has a magnitude
of xcp .L in the direction that the lift force would rotate the aerofoil about the leading

24

edge as the axis of rotation. This of course is in the counter clockwise direction. A
moment is defined as being positive if it tends to rotate the aerofoil nose up. Therefore
the moment acting at the leading edge is negative. A lift force acting upwards is
defined to be positive and thus the value of xcp can be calculated as follows

.V2 .c 2
c
M LE
xcp =
= 4 2
=
L
4
V c.

(112)

The coefficient of moment about the leading edge is defined as

Cm , LE

14 V2c 2
M LE
=1
= 1
= 12
2 2
2 2

V
c

V
c

2
2

(113)

It should be noted that xcp can also be obtained from the following equation
xcp
c

Cm, LE
Cl

12 1
=
2
4

(114)

6.2 Cambered thin aerofoil

In the previous section it has been shown how a simple analysis based on a very crude
model with severe restrictions can still be very useful in giving insight into the
aerodynamic properties of aerofoils. Whilst the flat plate model for an aerofoil is not
capable of giving detailed knowledge of velocity or pressure distribution around the
aerofoil surface with any accuracy, it is capable of predicting reasonably accurately
the value of the slope of the lift versus angle of attack curve, namely 2 . This value
is quite close to the value obtained from wind tunnel measurement or much more
sophisticated numerical modelling, which gives slightly lower value of the slope but
differing by a factor of not more than 10 percent or so.
The method is also capable of predicting the location of the centre of pressure, which
for this simplified flow model is the same as the aerodynamic centre. For the case of
aerofoils in actual subsonic flows, the aerodynamic centre is located near the quarter
chord point as predicted by the flat plate model. However, the centre of pressure
moves as a function of the angle of attack, .
In this section we shall relax the constraints slightly by allowing the aerofoil to have
curvature or camber. The thickness of the aerofoil must still be very small, and the
maximum camber must also be very small such that the aerofoil is still very much like
a flat plate. However, because the aerofoil is allowed to have camber the boundary
condition that must be satisfied is somewhat different from the flat plate model. At a
point on the camber line, which represents the aerofoils surface, the normal to the
camber line is not exactly in the y-direction (as in the case of the flat plate model) but
along a line that is slightly inclined to the y-axis. Let this angle be and is the same
as the inclination of the tangent (or slope) to the camber line at the control point. This
implies that
dy
dy
(115)
tan =
or
= tan 1
dx
dx

25

It is further assumed that the angle is sufficiently small such that the tangent of the
dy
angle is equal to the angle itself in unit of radian, i.e. tan =
.
dx
The equation to be satisfied is now slightly different from equation (95) as follows
c
1
2

( x ) .dx
x* x

= V .sin ( )

(116)

Since ( ) is quite small, therefore the above equation can be simplified as follows
c
1
2

( x ) .dx

x* x

= V . ( )

(117)

From the web site http://www.desktopaero.com/appliedaero/airfoils1/tatderivation.html

From the web site http://www.desktopaero.com/appliedaero/airfoils1/tatderivation.html

The solution of equation (116) has 2 components. The first one is the same solution as
for the flat plate situation or equation (98). The second component is to account for
the aerofoils camber. The trial solution is thus given as follows

( ) = 2V A0

1 + cos
+ An sin n
sin
n =1

(118)

Substituting (118) into (117) we get

1

A0
0

1 + cos
sin n .sin
+
An
.d = ( )
*
cos cos
cos cos *
n =1

(119)

From equation (99) it is known that

1 + cos
.d = 1
0 cos cos *
1

26

(120)

The second part of equation (99) gives the following results

sin n .sin
.d = .cos n * for n = 1, 2,..
*
cos

cos

(121)

Substituting (120) and (121) into (119) we get the following

A0

n =1

An cos n * = ( * )

(122)

Therefore our trial solution (118) satisfies the governing boundary condition (117)
provided the Fourier series coefficients satisfy the requirement of equation (122).
We know from basic calculus that

Therefore

cos ( n + m ) * = cos n * cos m * sin n * sin m *

(123)

cos ( n m ) * = cos n * cos m * + sin n * sin m *

(124)

1
2

( cos ( n + m )

+ cos ( n m ) * ) = cos n * cos m *

(125)

The results above can be used to get the following results

cos n cos m d =
*

1
2
0

( cos ( n + m )

+ cos ( n m ) * ) d *

if n = m or ( n + m ) = 0

2
=0

(126)

if n m

Integrating equation (122) from 0 to , we can get the following

A0

n =1

An cos n

.d =
*

( ( )).d
*

Since cos n *d * = 0 for n = 1, 2, 3,, therefore

0

A0 =

( * ) .d * =

( * ).d *

(127)

Multiplying equation (122) by cos m * and integrating from 0 to we get

n =1

A0 cos m *d *

An cos n * cos m *d * =

27

( ( ) ) cos m .d
*

The integrals on the left hand side of the equation above are all zero except if n = m ,
in which case it has the value of 2 . The first term on the right hand side of the
equation is always zero hence the equation can be simplified as follows
2

Am =

( * ).cos m *.d * for m = 1, 2, 3, ..

(128)

Since in this last result m is just a dummy index, thus we can change it to n or k or
any other symbol if we so wish.
The total circulation around the aerofoil is given by equation (106), which for
cambered aerofoil can be written as follows

n =1

= V .c A0 (1 + cos ) .d +

An sin n .sin .d

(129)

From basic mathematics we know that

sin n .sin = 12 ( cos ( n 1) cos ( n + 1) )
sin n .sin 2 =
Therefore

1
2

( cos ( n 2 ) cos ( n + 2 ) )

sin n .sin .d =

2
=0

sin n .sin 2 .d =

(131)

for n = 1

(132)

if n 1

(133)

for n = 2
2
= 0 if n 2

(130)

(134)
(135)

Equation (129) can now be simplified as follows

= V .c. A0 + A1.

= V .c.

( 2 A0 + A1 )

(136)

The lift acting on the aerofoil is given by the Kutta-Joukowski lift equation (120),
hence the lift coefficient can be calculated as follows

Cl =

1
2

V2 .c. ( 2 A0 + A1 )
1
= 2 A0 + A1
2
1
2
2 V .c

(137)

The lift curve slope can be calculated knowing that A0 is given by equation (127)

Cl , =

Cl
= 2

(138)

28

The pitching moment about the leading edge can also be evaluated as follows
c

M LE = x.dF = V ( x ) .x. dx
where ( x ) = ( ) is given by equation (96) and x and dx are given by equation
(112), hence the above equation can be simplified further as follows

M LE = 12 V2 c 2

M LE = 12 V2 c 2

A0 (1 cos 2 ) +

n =1

A0 (1 cos 2 ) +

1
2
0

An sin n .sin (1 cos ) d

n =1

An ( sin n .sin 12 sin n .sin 2 ) .d

Substituting equations (149), (150), (151) and (152) into the above, we finally get

M LE = 12 V2 c 2 A0

+ A1

A2

= 12 V2 c 2

[ A0 + A1 12 A2 ]

The leading edge moment coefficient is then

Cm , LE =

1
2

M LE

= [ A0 + A1 12 A2 ]
2 2
V c
2

(139)

The location of the centre of pressure is given by

xcp
c

Cm, LE 1 A0 + A1 12 A2
=
Cl
4 A0 + 12 A1

(140)

The moment coefficient at the quarter chord point is

Cm ,c / 4 = Cm, LE + 14 Cl = 2 [ A0 + A1 12 A2 ] + 2 ( A0 + 12 A1 ) = 4 ( A2 A1 )

(141)

The equations for the camber lines for NACA aerofoils can be obtained from the book
by I.H.Abbott and A.E.von Doenhoff : Theory of Wing Sections or it can be obtained
from the following web address
http://www.desktopaero.com/appliedaero/appliedaero.html

As an example of the application of TAT for a cambered aerofoil we will consider a

x
x
simple aerofoil whose camber is given by y ( x) = 4h. . 1
where the maximum
c
c
camber h is a small positive number and c is the chord
hence

dy
h
x
h
h
= 4 1 2 = 4 (1 (1 cos ) ) = 4 cos
dx
c
c
c
c

29

(142)

The Fourier coefficients for this aerofoil can be calculated as follows

A0 =

1 4h
cos d =
0 c

2 4h
8h 4h
A1 =
cos .cos .d =
. =
c 0
c 2 c
A2 =

2 4h
8h
cos 2 .cos .d =
.0 = 0
c 0
c

Note that equation (141) has been used to evaluate the integrals for A1 and A2 .
The value of the lift coefficient is
2h
(143)
Cl = 2 ( A0 + 12 A1 ) = 2 +
c
The leading edge and quarter chord moment coefficients, and the centre of pressure
location are

4h
(144)
Cm, LE = +
2
c
Cm ,c / 4 =

xcp
c

h
c

(145)

1 + 4h / c 1
h/c
= 1+ 2
4 + 2h / c 4
+ 2h / c

(146)

Let us now study a similar aerofoil, except that the maximum camber is now located
at 0.25c, rather than at the mid point 0.5c.
The equation for the camber of the aerofoil is now given by the following
8x ( c 2 x )
dy 8h
= (c 4x )
and
2
c
dx c 2
dy 8h
8h
=
1 4 ( 12 (1 cos ) ) = ( 2 cos 1)
Therefore
dx c
c

For 0 x 14 c : y ( x ) = h.

For 14 c x c : y ( x ) = 8h.
Therefore

( 2 x + c )( c x )
9c

and

dy 8h
=
(c 4x)
dx 9c 2

dy 8h
8h
=
1 4 ( 12 (1 cos ) ) = ( 2 cos 1)
dx 9c
9c

It should be noted that x/c =1/4 is equivalent to cos = 0.5 or = / 3

The Fourier coefficients can now be calculated as follows

30

(147)
(148)

(149)

(150)

 /3

8h
A0 =
c
A0 =

8h
c

A0 =

8h
1
h
0.6849 x 3.8265 = 0.6613
c
9
c

A1 =

16h
c

16h
A1 =
c

A1 =

16h
c

/3

( 2sin ) 0

/3
0

1
( 2sin ) / 3
9

( 2 cos2 cos ) .d +

1
+ sin 2 sin
2

/3

1
( 2 cos2 cos ) .d
9 /3

1
1
+ + sin 2 sin
9
2

/3

16h 1
2
1 2 1
2

+ sin
sin +
sin
+ sin
c 3 2
3
3 9 3 2
3
3
A1 =

A2 =

1
( 2 cos 1) .d +
( 2 cos 1) .d
9 /3

/3

16h
A2 =
c

16h
2.5274
h
0.6142 +
= 4.5583
c
9
c

( 2 cos 2 .cos cos 2 ) .d +

1
1
sin 3 + sin sin 2
3
2

/3

1
( 2 cos 2 .cos cos 2 ) .d
9 /3

1 1
1
+
sin 3 + sin sin 2
9 3
2

A2 =

16h
1
2 1
1
2
sin sin
+ sin + sin
c
3 2
3 9
3 2
3

A2 =

16h
1
h
0.4330 x 0.4330 = 1.9602
c
9
c

/3

The lift coefficient can now be calculated as follows

Cl = 2 ( A0 + 12 A1 ) = 2

h
h
( 0.6613 2.2792 ) = 2 + 1.6179
c
c

(151)

The pitching moment coefficients and the centre of pressure location are
Cm , LE =

h
c

0.6613 + ( 4.5583 0.9801)

31

h
= + 2.9169
c
2
c

(152)

Cm ,c / 4 = 0.6495
xcp
c

h
c

(153)

Cm, LE 1 + 2.9169h / c 1
h/c
=
= 1 + 1.299
Cl
4 + 1.6179h / c 4
+ 1.6179h / c

(154)

Now let us consider another situation where the maximum camber location is pushed
back even further. For the cases where the maximum camber is located at x/c = , the
results are as follows
For 0 x 34 c : y ( x ) =
Therefore

dy 8h
8h
=
3 4 ( 12 (1 cos ) ) = ( 2 cos + 1)
dx 9c
9c

For 34 c x c : y ( x ) =
Therefore

8h x ( 3c 2 x )
dy 8h ( 3c 4 x )
.
and
=
2
9
c
dx 9c
c

(157)

8h ( 2 x c )( c x )
dy 8h ( 3c 4 x )
.
and
=
c
c
dx c
c

dy 8h
8h
=
1 4 ( 12 (1 cos ) ) = ( 2 cos + 1)
dx c
c

A0 =

8h
9 c
8h
9 c

2 / 3

16h
9 c

16h
A1 =
9 c

2 / 3

2 / 3

( 2sin + ) 0

( 2 cos + 1) .d

+ 9. ( 2sin + ) 2 / 3

8h
h
[3.8264 9 x 0.6849] = + 2.078
9 c
c

( 2 cos2 + cos ) .d + 9.

2 / 3

1
+ sin 2 + sin
2

A1 =

2 / 3

A0 =

A1 =

( 2 cos + 1) .d + 9.

2 / 3

( 2 cos + cos ) .d
2

1
+ 9. + sin 2 + sin
2

16h
h
( 3.3934 + 9 x0.9023) = 6.5156
9 c
c

32

(158)

(159)

It should be noted that x/c =3/4 is equivalent to cos = 0.5 or = 2 / 3

The Fourier coefficients can now be calculated as follows

A0 =

(156)

2 / 3

16h
A2 =
9 c

A2 =

2 / 3

( 2 cos 2 .cos + cos 2 ) .d + 9.

( 2 cos 2 .cos + cos 2 ) .d

2 / 3

16h
9 c

1
1
sin 3 + sin + sin 2
3
2

A2 =

2 / 3

+ 9.
0

1
1
sin 3 + sin + sin 2
3
2

2 / 3

16h
2 1
4
2 1
4
0 + sin
+ sin
+ 9. 0 sin
sin
9 c
3 2
3
3 2
3
A2 =

16h 1

h
sin 9 x sin
= 1.9603
9 c 2
3
2
3
c

The lift coefficient can now be calculated as follows

Cl = 2 ( A0 + 12 A1 ) = 2 +

h
h
( 2.078 + 3.2578 ) = 2 + 5.3358
c
c

(160)

The leading edge pitching moment and centre of pressure location are
Cm , LE =

h
c

+ 2.078 + ( 6.5156 + 0.9801)

h
= + 9.5738
c
2
c

Cm ,c / 4 = 4 ( 1.9603 6.5156 )( h / c ) = 2.119 . ( h / c )

xcp
c

Cm, LE
Cl

(h / c)
1 + 9.5738 ( h / c ) 1
= .
= 1 + 4.238
4 + 5.3358 ( h / c ) 4
+ 5.3358 ( h / c )

(161)
(162)
(163)

The various results can be tabulated to see the effects of shifting the location of the
maximum camber point as follows.
Max. camber
location
Cl / 2

+1.6179 ( h / c )

+2.0 ( h / c )

+5.3358 ( h / c )

( 2 / ) Cm. LE

+2.9169 ( h / c )

+4.0 ( h / c )

+9.5738 ( h / c )

xcp
c/4

Cm ,c / 4 /

0.25 c

0.50 c

( h / c)
(h / c)
2
+ 1.6179 ( h / c )
+ 2.0 ( h / c )
0.6495 ( h / c )
( h / c)

1.299

0.75 c

(h / c)
+ 5.3358 ( h / c )
2.119 ( h / c )
4.238

From the table it can be seen that as the location of the maximum camber is moved
backward towards the trailing edge, for a given value of maximum camber, h/c, and
angle of attack, , then the lift coefficient increases and the leading edge pitching
moment also increases or becomes more negative. The quarter chord moment also
becomes more negative, but the location of the centre of pressure moves further
33

backward by an amount, which depends on both h/c and as well as on the location
of the maximum camber.
If the maximum camber location is fixed, then for a given angle of attack the lift
coefficient and the moment coefficients all increase with increasing value of the
maximum camber. However the location of the centre of pressure behaves in a more
complex manner and depends on the magnitude of the angle of attack as well.
If the maximum camber, h/c, and its location are kept constant, then the lift
coefficient and the leading edge moment coefficient both increase linearly with
increasing angle of attack. The quarter chord moment remains constant and is
independent of the angle of attack. However, the location of the centre of pressure
again behaves in a complex manner depending on the relative values of the maximum
camber, its location and the angle of attack.
If the angle of attack is zero, the above table can be simplified as follows
Max. camber
location
Cl
2
2
Cm.LE

xcp
c

0.25 c

h
c
h
2.9169
c

0.50 c

h
c
h
4.0
c

0.75 c

h
c
h
9.5738
c

1.6179

2.0

5.3358

0.4508

0.500

0.4485

As can be seen from the above table, an aerofoil seems to have a positive value of lift
coefficient even if the angle of attack is zero. This implies that at a particular negative
value of angle of attack, the lift acting on an aerofoil is zero. The angle of attack for a
cambered aerofoil when the lift produced is zero is called the zero lift angle of attack
and is denoted by the symbol of 0 . The lift coefficient for a cambered aerofoil is
usually given as
Cl = 2 ( 0 )
(164)
From the above table it can be seen that the magnitude of zero lift angle of attack for a
given camber, h/c, increases monotonically as the location of that maximum camber is
moved further and further backward towards the trailing edge. If the location of the
maximum camber is fixed, both the lift and moment coefficients increase linearly
with increasing camber, h/c. Furthermore, the constant of proportionality increases
rapidly with increasing value of the location of the maximum camber (or the further
rearwards the maximum camber point is located on the aerofoil). The magnitude of
0 is obviously determined by the curvature of the aerofoil, which is dependant on
both the magnitude and location of the maximum camber h/c.
Max. camber location
0 (for h/c = 0.01)

0.25c
-0.93 0

0.50c
-1.15 0

0.75c
-3.06 0

0 (for h/c = 0.02)

0 (for h/c = 0.03)

-1.85 0

-2.29 0

-6.11 0

-2.78 0

-3.44 0

-9.17 0

34

A rather interesting point to note is the complex behaviour of the location of the
centre of pressure. From previous result we know that the centre of pressure for a
symmetric aerofoil (or flat plate) is located at the quarter chord point. However, from
the above table it can be seen that for zero angle of attack, the centre of pressure
location for a cambered aerofoil is actually near the mid point of the aerofoil rather
than the quarter chord point. Furthermore, that location is independent of the value of
the camber, but is slightly dependent on the location of the maximum camber point.
For an aerofoil with its maximum camber located at the chords mid point, the centre
of pressure is also located at the mid point. However, if the location of the maximum
camber is moved either forward or backward of the mid point, the centre of pressure
always moves slightly forward of the mid point.
The centre of pressure is a concept in Mechanics, which is useful in presenting data
about distributed stresses or pressure in the most efficient way. All the relevant
information about the distributed pressure can be presented in terms of only 2
quantities, namely the resultant force and the location of the centre of pressure. The
centre of pressure is defined as the point about which the moment of the distributed
pressure is zero. It can also be thought of as being the point at which the resultant
force acts. The moment at any other point can be calculated simply as the product of
the resultant force multiplied by the distance of the point from the centre of pressure.
The concept of the centre of pressure is, however, sometimes rather useless in the
field of aerodynamics. In mechanics the resultant force is normally quite large in
magnitude. On the other hand, in aerodynamics the resultant force is often quite small
and sometimes even zero. When the lift, or total force, is zero then the definition of
the centre of pressure is quite meaningless. When lift is zero the resultant of the
distributed pressure is a pure couple, which of course has a constant moment value at
any point along the chord of the aerofoil. There is no point where the moment acting
on the aerofoil is zero, so there is no centre of pressure. To get out of this difficulty, in
aerodynamics we define the aerodynamic centre as the point about which the moment
acting on the aerofoil is constant, independent of angle of attack or the lift produce.
The information about the distributed pressure can then be represented by 3
quantities, namely the resultant force, the moment about the aerodynamic centre and
the location of the aerodynamic centre itself. The result of TAT analysis suggests that
the aerodynamic centre is located at the quarter chord point, since as has been pointed
out earlier the moment at the quarter chord is independent of angle of attack.
Experimental results suggest that real aerofoils in real flows do have their
aerodynamic centres located very near the quarter chord point, at least for subsonic
flows. (Aerofoils in supersonic flows have their aerodynamic centres located closer to
the midpoint of the chord).
Because of this rather fortunate situation, it is not necessary to present the data of the
location of the aerodynamic centre. Thus it can be seen that the aerodynamic centre is
a very useful concept in aerodynamics, similar to the usefulness of the concept of the
centre of pressure in Mechanics. It is for that reason that we will be discussing more
about the moment about the aerodynamic centre (or the quarter chord point) rather
than the moment about the centre of pressure.

35

6.3 Flapped Aerofoil and Control Surfaces.

Let us now apply our knowledge to study the behaviour of control surfaces such as
the rudder, elevator or aileron. The control surfaces are normally symmetric aerofoils,
and here we will impose the constraint that they are very thin so that the results of
TAT are applicable for them.
The elevator is the control surface to create pitching moment, and is attached to the
horizontal tail plane via a hinge line. Let the horizontal tail plane, including the
elevator, be represented by a portion of the x-axis, with its nose located at origin of
the coordinates system and its chord length is c. Let the hinge line of the elevator be
located at xh = (1-F).c, where F is the length of the elevator as a fraction of the total
chord length c and normally has a value of around 0.3 or there about. The elevator
may be deflected up or down. To create a nose up or positive pitching moment the
elevator must be deflected upwards, and downwards deflection will create a negative
pitching moment. A horizontal tail plane with its elevator deflected is modelled very
simply as a horizontal straight line from the nose at the origin to the hinge point H at
xh and another straight line representing the elevator from the hinge line to the trailing
edge. This elevator is at a negative deflection angle of when deflected downward.
Let us also assume that the airflow is moving uniformly at an angle of relative to
the horizontal line.
Transformation of coordinate from Cartesian to polar coordinate gives us the
following results. The nose is at = 0 and the trailing edge is at = , while the
hinge line is at h , where xh = (1 F ) c = (1 cos h ) .c / 2 or h = cos 1 ( 2 F 1) .
The Fourier series coefficients of the TAT solution for this case can be calculated as
follows

dy
.d
(165)
0
0 dx
Now it should be remembered that for 0 h we have = 0 since the line
representing the main tail plane is horizontal, whereas for h , i.e. from the
hinge line to the trailing edge, we have a constant negative value of .
This means that the coefficient can be calculated as follows
A0 =

.d

A0 =

. d = 1
h

(166)

The other coefficients, for n = 1, 2, 3, , are given as follows

An =

cos n .d =

2sin n. h
.
n.

(167)

A1 = 2 sin ( h ) .

(168)

A2 = 1 sin ( 2 h ) . = A1.cos h

(169)

Let us now define the following functions of h

36

k1 ( h ) = 1
k2 ( h ) =

1
4

(170)

( 2sin h sin 2 h )

k3 ( h ) = 1

(171)

h sin h
sin h
+
= k1 ( h ) +

(172)

The lift and quarter chord moment coefficients of the tail plane with the elevator
deflected are then given by the following equations.
Cl = 2 ( A0 + 12 A1 ) = 2 ( k3 ( h ) . )

(173)

Cm ,c / 4 = 14 ( A1 A2 ) = k2 ( h )

(174)

The pressure that acts on the elevator would produce a turning moment about the
hinge axis. If the pilot wishes to change the elevator deflection angle setting, she or he
has to exert a force on the control stick to overcome the elevators hinge moment. It is
useful, therefore, to be able to predict how the hinge moment would vary as a function
of the 3 parameters, namely the angle of attack, the deflection angle and the fractional
length.
The resultant force of the pressure that acts only on the elevator can be calculated as
follows:
Cl ,el =

Liftel
=
A
+

+
An .sin n .sin d
2
1
cos
(
)
0
2
1
n =1
2 V c
h

The moment of the distributed pressure, on the elevator only, about the leading edge
is given by the following equation
CmLE ,el =

A0 (1 + cos ) +

CmLE ,el =

n =1

An .sin n .sin

(1 cos ).d

A0
An
sin n .sin 2 .d
(1 cos 2 ) d + An sin n .sin .d
2 h
n =1
n =1 2 h
h

The integrals can be evaluated as follows

(1 + cos ).d = h sin h

(1 cos 2 ).d = h + 12 sin 2 h

37

For n = 1

sin .sin .d =

For n 1

sin n .sin .d =

1
2

For n = 1

1
2

sin 2 .sin 2 .d =

sin ( ( n + 1) h ) sin ( ( n 1) h )

n +1
n 1

sin n .sin 2 .d = 12

For n>2

( cos ( ( n 1) ) cos ( ( n + 1) ) ).d

1
2

= 12

(1 cos 2 ).d = 12 ( h + 12 sin 2 h )

sin ( 3 h ) sin ( h )

3
1

sin 2 .sin .d = 12

For n = 2

1
2

(1 cos 4 ).d = 12 ( h + 14 sin 4 h )

( cos ( ( n 2 ) ) cos ( ( n + 2 ) )).d

sin ( ( n + 2 ) h ) sin ( ( n 2 ) h )

n+2
n2

Putting the above results together to evaluate the integrals, we can get the following

n =1

An . sin n .sin .d = 12 A1 ( h + 12 sin 2 h )

h

+ 12

n=2

n=2

An

sin ( ( n + 1) h )

n +1

sin ( n + 1) h sin ( n 1) h
2

=
n +1
n 1

An

Therefore

n =1

An . sin n .sin .d =
h

1
2

sin ( ( n 1) h )

n 1

sin n h sin ( n 1) h sin ( n + 1) h

.

n
n 1
n +1
n=2

sin h.sin 2 h = 12 A1 sin 2 h

A1 ( h + 12 sin 2 h ) 12 A1 sin 2 h = 12 A1 ( h )

For the coefficient of moment we need the following

n =3

An

sin ( n + 2 ) h sin ( n 2 ) h
2

=
n+2
n2

sin n h sin ( n 2 ) h sin ( n + 2 ) h

.

n
n2
n+2
n =3

sin 3 h sin 2 h sin 4 h

+

3
2
4
sin 3 h
sin 4 h
= A1
A2
3
4

38

sin h.

Therefore

n =1

An . sin n .sin 2 .d = 12 A1
h

sin 3 h sin h
sin 4 h

+ 12 A2 h +
3
1
4

sin 3 h 1 sin 4 h
2 A2
3
4
1
1
= 2 A1 sin h + 2 A2 ( h )
12 A1

The coefficient of moment of pressure distribution on the elevator, about the leading
edge of the aerofoil is given as follows
CmLE ,el =

A0
sin 2 h
A
A
A
h +
1 ( h ) 1 sin h 2 ( h )
2
2
2
4
4

CmLE ,el =

A0
sin 2 h
A
sin h
A
1 h +
+ 2 ( h )
h +
2
2
2
2
4

CmLE ,el =

1
sin 2 h
h +
2
2

CmLE ,el =

CmLE ,el =

sin h

h +

h
.

sin h
sin 2 h
. +
( h ) .
2
4

1
sin 2 h
1
sin h
sin h
2
h +
( h ) .
h +
.
2
2
2

CmLE ,el

h sin 2 h

2
2

2sin h

h sin h
+

h sin 2 h

h sin h
= 1
+
1 +
2

2
2

Let us now define the following

h sin h
sin h
k4 ( h ) = 1

= k1 ( h )

h
k5 ( h ) = 1

k6 ( h ) = 1

= k12 ( h )

(175)

(176)
(177)

h sin 2 h
sin 2 h
+
= k1 ( h ) +

h sin h
k7 ( h ) = 1
+

= k32 ( h )

39

(178)

(179)

The lift coefficient and the leading edge moment coefficients of the elevator are now
given by
Cl ,el = 2 ( k4 ( h ) . k5 ( h ) . )
CmLE ,el =

(180)

( k ( h ) . k ( h ) . )
6

(181)

The hinge moment of the elevator is then given by the following

CmH ,el = CmLE ,el + (1 F ) .Cl ,el

(182)

The lift and hinge moment coefficients of the elevator can also be rewritten as follows

where

Cl ,el ( , ) = a1 ( F ) . b1 ( F ) .

(183)

CmH ,el ( , ) = a2 ( F ) . b2 ( F ) .

(184)

a1 ( F ) = 2 .k4 ( F ( h ) ) and b1 ( F ) = 2 .k5 ( F ( h ) )

a2 ( F ) = 2 . k6 ( F ) 4 (1 F ) k4 ( F )

and

b2 ( F ) = 2 . k7 ( F ) 4 (1 F ) .k5

The values of the above coefficients for some particular values of F are shown in the
table below
F

0.10
0.08700

0.20
0.2546

0.30
0.4855

0.40
0.7793

0.50
1.1416

b1 ( F )

0.2636

0.5474

0.8556

1.1939

1.5708

- 0.2365

-0.2200

- 0.1268

0.0212

0.2146

- 0.0088

0.0369

0.0869

0.2665

0.2665

a1 ( F )
a2 ( F )

b2 ( F )

It should be noted that the hinge moment coefficient (see equation (184)) is dependent
on the angle of attack as well as on the angle of deflection. This is undesirable, since
the hinge moment for a given deflection angle is different for different value of angle
of attack. As far as the pilot is concerned, it is desirable that when s/he wants to
change the setting of the angle of deflection, the force that s/he has to exert should be
the same regardless of the angle of attack (which is not associated with trying to
control the aircraft as such). However, for a particular value of F between 0.30 and
0.40 the value of a2 ( F ) is zero. This means that for the particular value of F, the
hinge moment is independent of angle of attack, thus that value of F should be chosen
as the fractional length of the elevator (or the movable part of any control surface,
such as the rudder on the vertical fin etc).

40

Additional information:
Airfoil Geometry
http://www.desktopaero.com/appliedaero/appliedaero.html
Airfoil geometry can be characterized by the coordinates of the upper and lower
surface. It is often summarized by a few parameters such as: maximum thickness,
maximum camber, position of max thickness, position of max camber, and nose
radius. One can generate a reasonable airfoil section given these parameters. This was
done by Eastman Jacobs in the early 1930'
s to create a family of airfoils known as the
NACA Sections.

The NACA 4 digit and 5 digit airfoils were created by superimposing a simple
meanline shape with a thickness distribution that was obtained by fitting a couple of
popular airfoils of the time:
y = (t/0.2) * (.2969*x0.5 - .126*x - .3537*x2 + .2843*x3 - .1015*x4)
The camberline of 4-digit sections was defined as a parabola from the leading edge to
the position of maximum camber, then another parabola back to the trailing edge.

NACA 4-Digit Series:

4
4
1
2
max camber position
max thickness
in % chord of max camber in % of chord
in 1/10 of c
After the 4-digit sections came the 5-digit sections such as the famous NACA 23012.
These sections had the same thickness distribution, but used a camberline with more
curvature near the nose. A cubic was faired into a straight line for the 5-digit sections.
NACA 5-Digit Series:
2
approx max
camber
in % chord

3
0
1 2
position
max thickness
of max camber in % of chord
in 2/100 of c
41

The 6-series of NACA airfoils departed from this simply-defined family. These
sections were generated from a more or less prescribed pressure distribution and were
meant to achieve some laminar flow.
NACA 6-Digit Series:
6
3,
Six- location
Series of min Cp
in 1/10 chord

2
2
1 2
half width
ideal Cl max thickness
of low drag
in tenths in % of chord
bucket in 1/10 of Cl

After the six-series sections, airfoil design became much more specialized for the
particular application. Airfoils with good transonic performance, good maximum lift
capability, very thick sections, very low drag sections are now designed for each use.
Often a wing design begins with the definition of several airfoil sections and then the
entire geometry is modified based on its 3-dimensional characteristics.

The following web sites give the coordinates of various NACA aerofoils
Appendix I- Profiles http://www.pdas.com/profiles.htm
Appendix II- Mean Lines http://www.pdas.com/meanline.htm
Appendix III- 4 and 5 Digit Sections http://www.pdas.com/sections45.htm
Appendix III- 6-Series Sections http://www.pdas.com/sections6.htm
Appendix III- 6A-Series Sections http://www.pdas.com/sections6a.htm
See also
http://www.adl.gatech.edu/classes/lowspdaero/lospd5/lospd5.html
http://www.desktopaero.com/appliedaero/airfoils1/tatderivation.html
http://adg.stanford.edu/aa208/fundamentals/TATResults.html
Created by Hadi Winarto
Created on Monday, March 15, 2004
Modified: 18 March 2004
Last updated: 23 March 2004
Major revision 25 May 2004

42