3.10 Oscillatory Boundary Layers: 3.10.1 Stokes Problem

You might also like

Download as pdf or txt
Download as pdf or txt
You are on page 1of 5

1

Notes on
1.63 Advanced Environmental Fluid Mechanics
Instructor: C. C. Mei, 2001
ccmei@mit.edu, 1 617 253 2994
December 1, 2002
3-10-oc-BL.tex
3.10 Oscillatory Boundary Layers
3.10.1 Stokes problem
If the viscosity is weak or the frequency is high (time scale is small) , there is a boundary
layer near the boundary of an oscillating body or the solid bottom under a wave.
Let the external ow have the tangential velocity u = <U(x)e
it
. If the inertia parame-
ter is small, then
uu
x
, vu
y
u
t
= O

U
o
L

1.
If furthermore
1
U
L


L
2
we keep keep just the most important viscous stress term and seek perturbation solution in
successive powers of the inertia parameter. Let us save the trouble of nondimensionalization
and assume
u = u
1
+u
2
+. . . (3.10.1)
with the understanding that the relative orders of magnitude is represented by the subscripts,
i.e., u
1
is the leading order and u
2
is smaller than u
1
by a factor O(
U
L


L
2
), etc. We get
from the Navier-Stokes equations the leading order approximation
u
1
t
=

t
Re

Ue
it

2
u
1
y
2
y > 0 (3.10.2)
subject to the boundary conditions that
u
1
Re Ue
it
y (3.10.3)
and
u
1
= 0 y = 0 (3.10.4)
Let
u
1
= Re
h
u
1
(x, y)e
it
+Ue
it
i
(3.10.5)
then
iU i u
1
= iU +
d
2
u
1
dy
2
2
Therefore,
d
2
u
1
dy
2
+
i

u
1
= 0 (3.10.6)
u
1
0, y (3.10.7)
u
1
= U
1
(x), y = 0 (3.10.8)
The solution is
u
1
= U(x) exp

(1 i)y
r

2

(3.10.9)
or,
u
1
= <

U(x)

1 exp

(1 i)y
r

2

e
it

(3.10.10)
The sign of

i is chosen so that (3.10.7) is satised. The boundary layer thickness is
=
s
2

(3.10.11)
It is easy to show that the vorticity at the wall is

u
1
y

y=0
= (1 i)U
r

2
= e
i/4
U (3.10.12)
which has the phase lag of /4 behind U, and diuses away from y = 0 like a propagating
wave exponentially attentuated after the Sokes boundary layer thickness O().
3.10.2 Induced Streaming
By considering the small eect of convective inertia, the second order improvement is phys-
ically even more interesting.
If the inviscid outer ow has tangential variation
dU
dx
6= 0, then by continuity there is
transverse ow v
1
in the boundary layer :
v
1
=
Z
y
0
u
1
x
dy = ie
it
dU
dx
Z
y
0
h
1 e
(1i)y/
i
dy (3.10.13)
= e
it
dU
dx
(
y

1 i
h
1 e
(1i)y/
i
)
which is valid in y O() only.
Let us examine the second order:
u
2
t

2
u
2
y
2
= U
dU
dx

u
1
u
1
x
+v
1
u
1
y
!
= U
dU
dx

"
(u
1
u
1
)
x
+
(u
1
v
1
)
y
#
3
Taking average over a period

2
u
2
y
2
= U
dU
dx

u
1
u
1
x
+
u
1
v
1
y
!
On the right-hand-side the last two terms u
1
u
1
, u
1
v
1
are wave-induced Reynolds stresses
u
1
u
1
= rate of transporting x-momentum in x-direction
u
1
v
1
= rate of transporting x-momentum in y-direction
Alternatively:

2
u
2
y
2
=
1
2

x
U
2

1
2

x
u
2
1
v
1
u
1
y
Let
= (1 i)/ (3.10.14)
Since
v
1
= ie
it
1

dU
dx

y 1 +e
y

u
1
y
= U(x)e
it
e
y
v
1
u
1
y
=
1
2
Re
"
U

dU
dx

y 1 +e
y

#
Thus

2
u
2
y
2
= G(y)
1
2
d|U|
2
dx
h
1

1 e
y

1 e

y
i
+Re U

dU
dx

y 1 +e
y

u
2
y
=
Z

y
G(y
0
)dy
0
u
2
=
Z
y
0
dy
00
Z

y
00
G(y
0
)dy
0
= y
Z

y
G(y
0
)dy
0
+
Z
y
0
y
00
G(y
00
)dy
00
One more integration gives
u
2
= Re F
4
U
dU

dx
. where
F
4
=
1
2
(1 3i)e
(1i)

i
2
e
(1+i)

1
4
(1 +i)e
2
+
1
2
(1 +i)e
(1i)
+
3
4
(1 i)
4
Note that as y , just outside the boundary layer,
u
2
=
1
4
Re
"
(3 i) U
dU

dx
#
Let U = Ae
i
U
dU

dx
= Ae
i
dAe
i
dx
=
dA
2
/2
dx
iA
2
d
dx
Hence
u
2
() =
1
4

3
2
dA
2
dx
3A
2
d
dx
!
(3.10.15)
Example: Progressive waves, U = U
o
e
ikx
, where U
o
, k are constants
u
2
() =
3k
4
U
2
o
(3.10.16)
3.10.3 Physics of the Induced Streaming
Take a progressive water waves as an example:
Outside the B.L. :
u

= A cos(t kx) (3.10.17)


Inside the B.L.
u = A
h
cos(t kx) e
y/
cos(t kx y/)
i
(3.10.18)
where the velocity amplitude A is related to the surface amplitude a by
A = a/ sin hkh (3.10.19)
Lets nd the induced transverse velocity v
u
x
= A sin(t kx) Ae
y/
sin(t kx y/)
v

=
Z
y
0
u
x
dy = y A sin(t kx)
1
2
Ak cos(t kx) +
1
2
Ak sin(t kx)
Now
u

=
1
4
A
2
k < 0
where the sin(tkx) terms in v

are out of phase with u

by /2, hence does not contribute


to the mean.
Now consider a slice of boundary layer one wavelength long. Because of periodicity, there
is no net transfer of momentum or forces at two ends x
0
and x
0
+2/k. But the momentum
transfer downwards is
A
2
4
k, causing a positive shear stress. To balance it there must be a
non-zero
u
y
at all levels y below the top. Hence, u 6= 0, resulting in induced streaming.
Homework: Find the induced streaming in Stokes boundary layer under a standing
wave with
U(x) = Acos kxe
it
.
5
Figure 3.10.1: Reynolds stress and Induced streaming in Stokes layer
.

You might also like