PROBLEM 26

A cylindrical artic hut is subjected to a crosswind as shown in the gure:

Ground
Crosswind
Semicylindrical Hut
Vent
The interior of the hut is ventilated to the outside through a small vent at a position, , as indicated. Hence
the pressure inside the hut (assumed uniform and constant) is the same as the pressure just outside the vent.
Assuming potential ow over the hut nd the angle, , at which the net vertical lift on the hut is zero. Neglect
the thickness of the wall of the hut and assume that the vent has no eect on the exterior ow. Also assume
that the air is incompressible.
SOLUTION 26
As derived in class, the pressure distribution on the surface of a cylinder in potential ow is given by p() where:
p() = p

+
1
2
U
2
(1 4 sin
2
)
where is the angle measured from the front or rear stagnation point. Note that as p() is symmetric about
= /2 and it does not matter whether the angle is measured from the front or rear stagnation point. The
above expression describes therefore the pressure acting on the outside or the arctic hut where p

is the pressure
far away, is the (incompressible) uid density and U is the velocity of the cross-wind. The pressure inside of
the hut is p
I
where
p
I
= p

+
1
2
U
2
(1 4 sin
2
)
where is the constant angle of the vent location. Any force on the hut will be given by integrating the dierence
in pressure between the inside and outside over the arc.
F =
_

0
(p p
I
) Rd
where
p p
I
= 2U
2
_
sin
2
sin
2

_
Since the hut is stationary, the lift on the hut must be zero so the forces in the y direction must balance.
F
y
=
_

0
2RU
2
_
sin
2
sin
2

_
sind
= 2RU
2
_
2sin
2

4
3
_
= 0
sin
2
=
2
3
hence = 54.7

or 125.3

3
PROBLEM 27
The incompressible, axisymmetric potential ow around a sphere can be generated by superposition of a uniform
stream ( = Ux) and a three-dimensional doublet whose potential is given by Acos /r
2
where A is a constant
representing the doublet strength. The coordinates r, are centered on the doublet and the direction x (x =
r cos ) is in the direction of the uniform stream:
Doublet
r

U x
On the basis of this information construct the velocity potential for potential ow around a sphere of radius R
in terms of U, R and the coordinates r, . What is the maximum velocity on the surface of the sphere?
SOLUTION 27
The velocity potential for ow around a sphere of radius R is created using the superposition of uniform ow
and a 3-D doublet:
= Ux
..
Uniform stream
+ A
cos
r
2
. .
3-D Doublet
=
_
Ur +
A
r
2
_
cos
The dierentiation of the velocity u to the radius r is given as:
u
r
=

r
=
_
U
2A
r
3
_
cos
The impenetrable condition (u
r
= 0) must hold at the surface of the sphere (r = R) such that:
u
r
|
r=R
=
_
U
2A
R
3
_
cos = 0
Therefore the constant A can be determined as:
A =
UR
3
2
The total velocity potential for ow over a sphere is given as:
= Ur
_
1 +
1
2
_
R
r
_
3
_
cos
The only component of the velocity along the surface of the sphere will be the one in the tangential direction,
u

. Therefore, the maximum velocity on the surface of the sphere will be the position of maximum velocity in
tangential direction u

:
u

=
1
r

= U
_
1 +
1
2
_
R
r
_
3
_
sin
4
on the surface of the sphere r = R
u

|
r=R
=
3
2
U sin
The magnitude of the velocity will be maximized when
| sin| = 1
such that:
=

2
,
3
2
Thus, the maximum velocity is given as
max (|u

|
r=R
) =
3
2
U
PROBLEM 28
For the purposes of estimating the drag force on a spherical body (radius, R) in a uniform stream (velocity,
U, and density, ) it is assumed that the pressure distribution over the upstream side (facing the oncoming
stream) is the same as in potential ow whereas the pressure on the downstream side is constant simulating the
conditions in a wake. Moreover the pressures match at the equator, = /2 (where is the angle measured
from the front stagnation point). Find the drag, F
D
, on the sphere as a function of , R and U. Evaluate the
drag coecient dened as C
D
= F
D
/(0.5U
2
R
2
).
SOLUTION 28
From the previous problem, the velocity potential for the ow past a sphere is given as
(r, ) = U cos
_
r +
1
2
R
3
r
2
_
The angular velocity component is:
u

(R, ) = U sin
_
1 +
1
2
R
3
r
3
_

r=R
=
3
2
U sin
By using Bernoullis equation on a streamline between the surface and innity, it can be determined that:
p

+
1
2
U
2
= p
s
+
1
2

3
2
U sin
_
2
p

p
s
=
1
2
U
2
_
1
9
4
sin
2

_
The pressure should match at the equator = /2,
p

p
w
= (p
s
p

)|
=/2
=
5
8
U
2
The drag is dened as the force acting on the surface parallel to the oncoming stream such that:
D =
_
dA
(p

p
s
) cos dA
=
_
2
0
_

0
(p

p
s
)R
2
sin cos dd
= 2R
2
_

0
(p

p
s
) sin cos d
5
= 2R
2
_
2
0
_
5
8
U
2
_
sin cos d + 2R
2
_

2
_

1
2
U
2
_
1
9
4
sin
2

__
sin cos d
=
5
8
U
2
R
2

1
16
U
2
R
2
=
9
16
U
2
R
2
By dening a non-dimensional drag, the drag coecient C
D
as drag divided by the frontal projected area of the
body A
p
= R
2
and by the term
1
2
U
2

, the drag coecient is:

C
D
=
9
8
PROBLEM 29
A nite dierence method is to be used with a mesh having a uniform node spacing, h, in the x and y directions
to solve for the quantity, f, which is governed by the following partial dierential equation:
f

2
f
x
2
= 4
_
f
y
_
2
(1)
Determine the nite dierence form of this equation at the node 0 that utilizes values of f at the nodes 0, 1, 2,
3 and 4 as shown below:
0
4
1 3
2
h
x
y
h
SOLUTION 29
First nd an expression for

2
f
x
2
at the point 0. Find the Taylor series expansion (around 0) for f
1
and f
3
:
f
1
= f
0
+
h
1!
_
f
x
_
0
+
h
2
2!
_

2
f
x
2
_
0
+
h
3
3!
_

3
f
x
3
_
0
+O(h
4
)
f
3
= f
0

h
1!
_
f
x
_
0
+
h
2
2!
_

2
f
x
2
_
0

h
3
3!
_

3
f
x
3
_
0
+O(h
4
)
Add the expressions for f
1
and f
3
, and solve for
_

2
f
x
2
_
0
:
6
f
1
+f
3
= 2f
0
+h
2
_

2
f
x
2
_
0
+O(h
4
)
_

2
f
x
2
_
0
=
f
1
+f
3
2f
0
O(h
4
)
h
2
_

2
f
x
2
_
0
=
f
1
+f
3
2f
0
h
2
+O(h
2
)
Now nd an expression for
f
y
at the point 0 using the same method, except this time subtract f
4
from f
2
because we want to keep the
_
f
y
_
0
term.
f
2
= f
0
+
h
1!
_
f
y
_
0
+
h
2
2!
_

2
f
y
2
_
0
+
h
3
3!
_

3
f
y
3
_
0
+O(h
4
)
f
4
= f
0

h
1!
_
f
y
_
0
+
h
2
2!
_

2
f
y
2
_
0

h
3
3!
_

3
f
y
3
_
0
+O(h
4
)
f
2
f
4
= 2h
_
f
y
_
0
+O(h
3
)
_
f
y
_
0
=
f
2
f
4
+O(h
3
)
2h
_
f
y
_
0
=
f
2
f
4
2h
+O(h
2
)
Square the expression for
f
y
at the point 0, and be sure to keep track of the error terms:
_
f
y
_
2
0
=
_
f
2
f
4
2h
+O(h
2
)
_
2
_
f
y
_
2
0
=
_
f
2
f
4
2h
_
2
+O(h
4
) +
f
2
f
4
h
O(h
2
)
_
f
y
_
2
0
=
_
f
2
f
4
2h
_
2
+O(h)
Putting everything together, and dropping our error terms O(h) and higher (the terms we keep are O(h
2
)):
f
0
_

2
f
x
2
_
0
= 4
_
f
y
_
2
0
f
0
f
1
+f
3
2f
0
h
2
= 4
_
f
2
f
4
2h
_
2
f
0
(f
1
+f
3
2f
0
) = (f
2
f
4
)
2
7