Lecture 16

Final Version
Contents
Combinations of Solutions: Solid
Bodies in a Potential Flow
(Rankine Oval etc.)
Cylinder in Uniform Flow
Cylinder with Circulation in a
Uniform Flow
Pressure Distribution Around the
Cylinder
Kutta-Joukowski Lift Theorem
Circulation and Lift for Aerofoil
Applications

Design Project

Instructions now on www as PDF file. (Instructions

should also appear as hardcopies via your pigeon hole
Deadline for submission extended until Friday Week
12 (= Fri. 19 Jan. 2007)
Submission sheet will appear on www soon-ish (and
also via pigeon holes)

COMBINATIONS OF SOLUTIONS: SOLID

BODIES IN A POTENTIAL FLOW
Recall: Can use PRINCIPLE OF SUPERPOSITION for velocity potential.
In addition, have shown that for incompressible, irrotational flow, stream function
also satisfies Laplace Eq. So can similarly construct flow solutions by combining
S.F. associated with uniform flow, source/sink flow and line-vortex flow.
In fact, we will almost exclusively use stream function here because we are interested in
pattern of streamlines; once we find stream function, we can use fact that it is constant
along streamlines to plot out streamlines.

Uniform Flow and Source: THE RANKINE BODY

What happens if we combine...

uniform flow source

y

Cartesian Coordinates:

x, y U y m tan 1

(1)

Polar Coordinates:

r , U r sin m

(2)

(1)/(2) represent complete descriptions of flow field. But what does it

look like?...
To graph lines of constant , first look for STAGNATION POINTS.
There ...

V 0

Thus, both velocity components must be zero Differentiate to get

expressions for velocity components ...

Continued...

Cartesian Coordinates: x, y U y m tan 1 y

x

x
r cos
U

cos
u
U m 2

2
2
r
y
x y
r
m

y
r sin
sin

v
2
2
2
r
x
x y
r

ur

m
U

m U , 0

and

u 0 need

Since m, r positive choose...

U sin
r

0, , 2 , 3 , ...
to get a solution for

m
m
ur U cos U
r
r

r , U r sin m

1
m
U cos
r
r

For

u 0

m
u U
x

STAGNATION POINT at

Polar Coordinates:

(4)

y0

For v=0 Eq. (4) requires ...

...with this Eq. (3) gives...

(3)

STAGNATION POINT at

ur 0

m
U

m U ,

ur 0

Continued...

In both cases same location for Stagnation Point ...

Cartesian Coordinates

m U , 0

Polar Coordinates

m U ,

We repeated ourselves to demonstrate that either coordinate system can be used.

In general choose the one that makes the analysis easiest.

Now use for distance between origin and stagnation point:

m
U

Find S.L. that arrives at stagnation point and divides there. Using ...

const.
along this S.L., use a known point - the stagnation point - to evaluate constant.
With Eq (2) from above ...

m
r , U r sin m s U
sin m m
U
where suffix s denotes along particular streamline through stagnation point.
Streamline found by equating Eq (2) to the constant and rearranging...

m U r sin m
m a
r

U sin
sin

PLOT

(5)

Continued...

Plotting stagnation streamline: r

2
3 2
0
with
r

a
sin

a 2
a 2
r sin a

etc. ...

Stagnation streamline defines shape of (imaginary) solid half-body

which may be fitted inside streamline boundary; remember flow does
not cross streamline or solid boundary. Call this special S.L.
SURFACE STREAMLINE.
Body shape named after Scottish engineer W.J.M. Rankine (18201872).

Now only have stagnation/surface. To get other S.L.s, choose point,

determine constant for S.L. through point and then sketch particular
S.L. through this point by compiling a table as above.

Continued...

How Does Flow Speed Vary Along Surface Streamline?

Recall velocity components in cartesian coord. from Eqs. (3)/(4) :
m
m
u U cos
v sin
r
r

U 2 v u 2 v2

Flow speed is:

U U m cos
r

2U m
m2
m2
2
2

sin U
cos 2 cos 2 sin 2
r
r
r
r

2U m
m2
U
cos 2 sin 2 cos2
r
r
2

2m
m2
U 1
cos 2 2
r
U
r U

2U m
m2

cos 2
r
r

NOW RECALL: a

a2 2 a

U U 1 2
cos
r
r

U 2

(6)

To find flow speed on body surface, VS , evaluate Eq. (6) subject to

a
sin

(5) for surface streamline. This gives ...

m
U

Continued...

Note: Since we know surface flow speed, we can evaluate static

pressure at any point on surface. Using Bernoulli equation along central
streamline that divides into surface streamline and, as usual, ignoring
gravitational term...

1
1
p U 2 pS VS2
2
2
Upstream
undisturbe d
flow

On surface

1
1
2
U
pS p VS2
2
2

1
2

V
2
S
p

p
V
pS p
S

1
2
2
1
1
1
2
2
2
U
U
U
U
2
2
2

pS p
VS2
1
2
1
2
U
U
2
Hence we get the non-dimensional pressure coefficient :

ps p
VS2
Cp
1
2
1
2
U
U
2

Continued...

Uniform Flow and Sink

(instead of Source)

Stream Function
Source

Sink

(previous case)

Cart. Coord.

x, y U y m tan 1

x, y U y m tan 1

Pol. Coord.

r , U r sin m

r , U r sin m

Only difference: plus sign(s) changed into minus sign(s)!

Hence, for sink expect a very similar analysis as above for source.

Source

Sink

S.P.

Note: In real world (inviscid) flow pattern for sink would not be observed! Flow
would initially follow body contour but (due to viscosity) detach at separation
points indicated by S.P. in sketch for sink. The phenomenon of SEPARATION will
be covered later. At this stage learn that...

Potential flows do NOT model ALL features of a real flow!!!

This has lead to potential flow often being termed IDEAL FLOW.

Obvious question now is what happens for ...

Uniform Flow + Source + Sink

We consider symmetric case where:
Strength
Source
Sink

Location
c, 0

c, 0

We cannot have both at origin now! What would happen if both at

origin?
By considering the two sketches on previous slide we can anticipate
shape of surface streamline and resulting body...
an oval.

Using superposition, can readily write stream function for this flow:

y
x, y U y m tan 1

Uniform
flow

Source at ( c , 0 )

y
m tan 1

c

Sink at ( c , 0 )

Second and third terms can be combined using:

tan 1 tan 1 tan 1

1

To give a more concise form for stream function

2c y

2
2
2
x y c

x, y U y m tan 1

(1)

Continued...

From either of the two forms of S.F. on previous slide, one can determine velocity
components

xc
xc
U m

2
2
y
x c 2 y 2
x c y

1
1
my

x
x c 2 y 2 x c 2 y 2

(2)
(3)

Now find stagnation points, where u=v=0. From Eq. (3) one sees that
when y=0 then v=0.
Substitute y=0 into Eq. (2) and then find value of x which gives that
u=0.
After some manipulation the solutions for x are:

2m
x c 1

cU

Hence, stagnation points at:

L, 0

and

L, 0

Now determine value of S.F. for surface streamline from Eq (1).

xc

x, y U y m tan 1

It can be seen that this is trivial and that

xc

m tan 1

0
S

(1) - repeated

Continued...

Rankine Oval then looks

like ...

We already determined value

of L. Can find points of
maximum
velocity
and
minimum
pressure
at
shoulders +/-h, of oval using
similar methods. All these
parameters are a function of
the...

basic dimensionless parameter

m
U c

In summary one obtains

L
2m

1
c
U c

umax
2 m U c
1
U
1 h2 c2

As one increases dimensionless

parameter d from zero to large values,
oval shape increases in size and
thickness from flat plate of length 2c
to huge, nearly circular cylinder.
Here think of increase when
c const. and U const

All Rankine ovals, except very thin

ones, have large adverse pressure
gradient on leeward surface. Thus,
boundary-layer will separate in rear,
broad wake flow develops, inviscid
pattern unrealistic in that region.

h
ha

cot
c
2m U c

Overall Strategy for Plotting Streamlines from

Stream Function was...
Write down stream function for flow by appropriately
combining individual solutions for source, sink and line
vortex as a sum:

x , y 1 x , y 2 x , y 3 x , y

Note: Huge choice as far as selction of parameters is concerned!

Souce strength, vortex direction of rotation, strength ...

Calculate expressions for vel. components u, v from

, v
y
x

Determine [Link] stagnation point(s) via u=0 , v =0.

Determine value of stream function passing through

(stagnation) point by substituting coordinates of
(stagnation) point(s) into the stream function.

Set stream function equal to the value you have

determined for point in question.

Determine values of x, y (or r, ) that satisfy this

expression and plot to obtain streamline.

Choose new point x,y

From previous it should be obvious how one can find stream function for a
cylinder (circle) in a uniform flow ...

Cylinder in a Uniform Flow

Turn Rankine oval into circle by allowing ...

U c

c0

Achieved by moving source and sink closer to origin ...

Limit (c=0) would ultimately cancel pair!

Ensure their influence remains by allowing m to increase in size.

Necessary limit is...

c0

Recall that Rankine

oval had S.F. ...
Now let c 0

with

c m const.

2c y

2
2
2
x y c

x, y U y m tan 1

then the argument of tan-1 goes to zero...

Noting that tan 1 for small gives

x, y U y m

2c y
x2 y2 c2

2mc

Define DOUBLET STRENGTH:

Stream Function for Cylinder Flow

y
x, y U y 2
x y2

Uniform
flow

(1)

DOUBLET
at origin

Note: Merging of source and sink as above produces structure known as

DOUBLET.

Continued...

More convenient to work in polar coordinates! S.F. can be written ...

r , U r sin

sin
U sin
r

U r

From Eq. (1) can get velocity components in usual way...

(1)

Continued...

ur
u

1
R
U cos 1 2

r
r

U sin
r

R 2
r
2

(2)

Where
we
used...

(3)

CHECK that this flow really does represent a cylinder in uniform flow.

Stagnation points:
Eq. (3) :

u 0
u 0

Substitute these angles into Eq.(2) and set ur 0

For 0 :
For :

R2
0 U cos 0 1 2
r

R2
0 U cos 1 2
r

Hence,... Stagnation points:

R, 0

to get ...

R2
1 2
r

r R

R2
1 2
r

r R

R,

and

Surface S.L. VALUE by substituting one stag. point into Eq. (1)...

R, 0 U sin 0 R
U
R

Now get equation for Surface S.L. by equating Eq. (1) to zero ...

r2

or

As required surface S.L. is circle with radius R

rR

PLOT

Continued...

Uniform Flow +

Doublet

= Flow over a Cylinder

Velocity Components on cylinder surface are obtained, by setting

r=R, from...

Eq. (2) ur U cos

R2
1 2
r

R2
Eq. (3) u U sin 1 2
r

ur 0
u 2 U sin

Might have expected to find that radial flow component is zero on

surface - flow cannot pass through (solid) cylinder wall!
Note also that maximum flow speed occurs at

3
and
2
2
where it is 2 U and 2 U

respectively.

Means in both cases (top and bottom half of cyl.) flow is from left to right! On
top negative value as velocity points in clockwise (negative angle) direction. On
bottom in anti-clockwise (positive angle) direction.

Finally, note symmetry of flow about both the x- and y- axes. What does
this tell you about the pressure distribution on the cylinder surface
remember the Bernoulli Equation!

Continued...

Since we can now mathematically describe

...we can, in principle, also describe flow through an arbitrary array of

cylinders as, for instance, the flow shown in the photo below. We simply
need to put several doublets in our uniform flow.

Cylinder with Circulation in a Uniform Flow

Without performing calculation, can see in preceding flow no net lift or drag on
cylinder since pressure distribution on surface symmetric about x- and y-axis..
In order to generate lift need to break symmetry. Achieved by introducing line
vortex of strength, K, at origin which introduces circulation
.
2 K

Note that this does not violate the flow around cylinder: line vortex produces a u
component of velocity only. Hence, we are still adhering to condition that flow
cannot pass through cylinder boundary.
Working from S.F. for cylinder in uniform flow additional inclusion of line vortex
gives:

sin
r , U r sin
K ln r C
r

Uniform
flow

Use result that radius

R
of resulting cylinder is :
(1)

Line vortex Arbitrary

at origin constant

Doublet
at origin

And set :

r , U sin r
U
r

C K ln R

K ln r K ln R

U sin r
U r

K ln r ln R

R2

U sin r
r

Velocity
Components

ur
u

K ln

r
R

1
R
U cos 1 2
r
r

(1)

R2

U sin
1 2

r
r
r

Continued...

So, on surface (r=R), velocity components are:

ur 0

u 2U sin

Surface Stagnation points also need:

K
R

u 0
sin

K
2 RU

Note: By setting vortex strength zero (K=0), recover flow over cylinder in
uniform flow with stagnation points at 0,
Plotting, Choose value for K, Now first get value of S.F. for r=R,... then
set S.F. equal to that value, then compile table r vs. angle This gives
particular streamline through stagnation points.
Then choose any other point in flow field not on stagnation streamline,
determine value of S.F. for this point, set S.F. equal to that value, then
compile table r vs. angle This gives streamline through the chosen
particular points Then choose another point in flow field etc (compare
flow chart from beginning of lecture). For various values of K the following,
flow fields emerge...

K 0

K 1

K 2

K 3

Continued...

Can now also describe flow through an arbitrary array of cylinders when
each of them is rotating! (Note: In photo below cylinders are not
rotating)

Pressure Distribution Around the Cylinder

To evaluate press. on cyl. surface use Bernoulli Eq. along S.L. that
originates far upstream where flow is undisturbed. Ignoring grav. forces:

1
1
2
p U
pS U S2
2
2
Upstream
undisturbe d
flow

Re-arranging...

pS p

On cylinder
surface

1
U 2
2

1 US

U 2

Substituting for flow speed ( u R 0 , u 2U sin

1
U 2 1 2U sin K / R 2 U 2
2

1
U 2
2

1
U 2
2

pS p

2
2
1 u R u

U 2

K
)
R

gives...

2
1 u

U 2

difference in pressures between surface and undisturbed free stream

pS p

1
U 2
2

K
4K

4
sin

sin

R
U
R
U

In particular for non-rotating cylinder

where K=0:

p S p

1
U 2 1 4 sin 2
2

Def.: Pressure
Coefficient

Cp

pS p
1 4 sin 2
1
U 2
2

(1)

(2)

Only top half of cyl. shown.

Continued...

Qualitative behaviour of
for various values of

K
p p
4K

Cp S
1 4 sin 2
sin
1
R
U
R
U
2

K RU

( 0 : Rear of cyl. , 2 1.57 : Top of cyl.,

3.14 : Front of cyl. , 3 2 4.71 : Bott. of cyl. )

Best way of interpreting above graphs is to think of flow velocity and radius being constant
while vortex strength is increasing from one plot to next.
When plotting graphs I did not explicitly specify velocity or radius! I simply used different
numeric values for K RU in order to illustrate behaviour of graph. I have not considered
if any of these cases may not be realizable in reality or not!.

Continued...

Equation (1)

p S p

1
U 2
2

K
4K

sin
1 4 sin
R
U
R
U

can be used to calculate net lift and drag acting on cylinder!

Sketch (A)

Sketch (B)

In Sketch (b) ...

L p sin pS p sin

D p cos pS p cos

Hence, integration around cylinder surface yields total L and D ...

p
0

p sin b R d

p
0

p cos b R d

where b is width (into paper) of cylinder. Substituting for pressure using

Eq. (1), and integrating (most terms drop out), leads to following results:

1
L U 2
2

4K

b R
R
U

2 U K b
Or, lift per unit width:

D0

L
U 2 K U
b

Thus, drag zero

a remarkable result!

Kutta Joukowski
Lift Theorem

d' Alembert' s
Paradox

Continued...

Net lift is indicated in sketch below. ... Note that if a line vortex is used which
rotates in mathematically positive sense (anti-clockwise) then resulting lift is
negative, i.e. downwards.

L
U
b

L
Final notes: How is lift generated? ... From sketch above and from
pressure profiles plotted earlier it is evident how this is physically
achieved Breaking of the flow symmetry in x-axis means that flow
round lower part of cylinder is faster than round top - this means that
pressure is lower round bottom and so a net downward force results.
Notice that symmetry in y-axis is retained symmetry of pressure
on left-hand and right-hand faces is retained and so there is no net
drag force. Keep in mind that our analysis was for an ideal fluid (i.e.
there is no viscosity). In a real flow would fore-aft symmetry be
retained?

Lastly, since lift is proportional to circulation, we wish to make

circulation large to generate a large lifting force. In applications of
above flow this is achieved by spinning cylinder to produce large
vorticity but is there a limit to how much circulation we should
produce?