AIAA 99-3523

Unsteady Flows in Semi-Infinite

Expanding Channels with Wall Injection
E. C. Dauenhauer and J. Majdalani
Marquette University
Milwaukee, WI 53233
For permission to copy or to republish, contact the American Institute of Aeronautics and Astronautics,
1801 Alexander Bell Drive, Suite 500, Reston, VA, 20191-4344.
30th AIAA Fluid Dynamics Conference
28 June1 July 1999
Norfolk, VA
AIAA-99-3523
1
American Institute of Aeronautics and Astronautics
Unsteady Flows in Semi-Infinite Expanding
Channels with Wall Injection

E. C. Dauenhauer* and J. Majdalani

Marquette University, Milwaukee, WI 53233

The current article presents an exact similarity solution of the Navier-Stokes equations for
unsteady flows established in a semi-infinite rectangular channel. The case considered here
pertains to a channel that experiences injection through two opposing porous walls while
undergoing uniform expansion. In the analysis, the Stokes stream function and the vorticity
equation are used in concert to yield an ordinary differential equation that lends itself to a
similarity solution in space and time. Following this similarity transformation, the original
problem is reduced to solving a single fourth-order differential equation with four boundary
conditions. The mathematical procedure is somewhat reminiscent of the classic Berman
approach. It is found that, in rectangular chambers and ducts where wall regression and
injection occur simultaneously, the streamlines, pressure distributions, and shear stresses
are time and space dependent. They can be characterized by two dimensionless parameters:
the expansion ratio of the walls and the cross-flow Reynolds number R. Instances of
direct application include the modeling of pulsating diaphragms with permeable walls and
the grain regression during solid rocket motor combustion. In order to examine the
incompressible laminar flow character in the channel, a numerical scheme is devised. The
scheme is based on a shooting method that can be applied in conjunction with a fourth-order
Runge-Kutta technique. The algorithm requires solving a set of twelve first-order
differential equations with two unspecified start-up conditions. In seeking the two unknown
initial guesses, the inverse Jacobian method is used to accelerate the convergence process.

I. Introduction


HE current paper presents a procedure that leads to
an exact similarity solution of the Navier-Stokes
equation in semi-infinite rectangular channels with
porous and uniformly expanding walls. The solution to
be developed provides the means to fully characterize
the flow field established in such a physical setting.
One of the objectives will be, therefore, to extend our
current knowledge of flow analyses involving channels
with porous walls.
Laminar flow studies in permeable walls have
received considerable attention in the past due to their
relevance to a number of engineering applications.
These include binary gas diffusion, filtration, ablation
cooling, surface sublimation, grain regression (as in the
case of combustion in solid rocket motors), and the
modeling of air and blood circulation in the respiratory
system.

*Student Member AIAA.
Assistant Professor, Department of Mechanical and
Industrial Engineering. Member AIAA.
Copyright 1999 by E. C. Dauenhauer and J. Majdalani.
Published by the American Institute of Aeronautics and
Astronautics, Inc., with permission.
The earliest studies of steady flows through
permeable and stationary walls can be traced back to
Berman
1
who investigated the laminar, two-
dimensional flow of a viscous incompressible fluid
driven by uniform injection (or suction) in a rectangular
channel with porous walls. In his study, Berman
assumed that the transverse velocity component was
independent of the streamwise coordinate. He thus
reduced the Navier-Stokes equations to a single,
nonlinear, fourth-order, ordinary differential equation
(ODE) with four boundary conditions and a cross-flow
Reynolds number R . The latter was based on the
normal injection speed
w
v and channel half-spacing h .
For small R , he employed a regular perturbation
scheme to derive an asymptotic formulation. Numerous
studies of channel flows with permeable walls
followed. These included the works of Sellars,
2
Terrill
3

(who extended Bermans small R case), Proudman
4

(who investigated the large R case using an integral
approach), Morduchow
5
(who applied the method of
averages over the entire injection range), White et al.
6

(who provided for an arbitrary R a convergent power
series solution), Taylor,
7
and Yuan
8
(who provided
analytic solutions for the infinite and large injection
cases, respectively). Improvements to the latter were
provided by Terrill.
9

T
AIAA-99-3523
2
American Institute of Aeronautics and Astronautics
For unsymmetrical flows caused by different wall
velocities, Terrill & Shrestha
10
constructed a
generalized perturbation series for small R . For large
injection, Shrestha & Terrill
11
extended Proudmans
one-term expression using matched asymptotic
expansions. In the same vein, Cox
12
considered the
practical case of an impermeable wall opposing a
transpiring wall.
Treatments of moving boundaries can be attributed to
Brady & Acrivos
13
who presented an exact solution to
the Navier-Stokes equations for a flow driven by an
axially accelerating surface velocity and symmetric
boundary conditions. Their work was motivated by the
need to study long slender droplets trapped in
extensional flows. This was later generalized by
Watson et al.
14
who allowed the accelerating walls to
be permeable. Watson
15
also examined the case of
asymmetrically accelerating walls.
In validating the similarity solutions recounted
above, investigators have constantly relied on
numerical simulations. Experimental verifications have
also been accomplished. In fact, numerous laboratory
experiments on steady channel flow through porous
walls have been conducted by Taylor,
7
Varapaev &
Yagodkin,
16
Raithby & Knudsen,
17
and Sviridenkov &
Yagodkin.
18

The addition of time-dependent motion in a long
rectangular channel with porous walls was achieved
experimentally by Ma et al.
19,20
and Barron et al.
21
In
fact, both used the sublimation process of carbon
dioxide to simulate the injection process at the walls.
As a result, the walls of their channel expanded during
the sublimation process. Other physical settings that
are better modeled with expanding permeable walls
include the regression of the burning surface in solid
rocket motors and the pulsation of porous diaphragms.
For instance, in simulating the laminar flow field in
cylindrical solid rocket motors, Goto and Uchida
22
have
analyzed the laminar incompressible flow in a semi-
infinite porous pipe whose radius varied with time. The
present analysis extends their analysis to expanding
channels and ducts which are rectangular in constitution
and possess porous and transpiring walls.
II. Formulation
We consider a rectangular channel of semi-infinite
length which possesses a high aspect ratio of width to
height. In actuality, the channel is of finite length,
however, it is commonplace in the literature to assume
semi-infinite length in order to neglect the influence of
the opening at the end.
23
The other end is closed by a
solid membrane which is allowed to stretch with
channel expansion. The axial flow is zero at the closed
end ( ) 0 x = due to the fluids viscosity and thus, the
closed end can be employed as a symmetry plane. This
is illustrated in Fig. 1. With x indicating the axial
direction and y the transverse direction, the
corresponding axial and transverse velocity components
are defined as u and v , respectively. With this
notation, the fundamental equations governing the
unsteady flow of an incompressible fluid are the classic
continuity and momentum expressions:

0
u v
x y

+ =

(1)

2 2
2 2
1 u u u p u u
u v
t x y x x y

1


+ + = + +

( )
(2)

2 2
2 2
1 v v v p v v
u v
t x y y x y

1


+ + = + +

( )
. (3)

where , p , , and t are the density, pressure,
kinematic viscosity, and time. The channel walls of
interest regress only in the transverse direction and their
separation is a function of time only. Under the porous
wall stipulation, the fluid is injected or extorted
uniformly and orthogonally through the channel walls
at a velocity
w
v . The latter is proportional to the
expanding wall surface velocity. The boundary
conditions necessary for solving the continuity and
momentum equations are:

0
w
u v v Aa = = = ! ; ( ) y a t = (4)
0 0
u
v
y

= =

; 0 y = (5)
0 u = ; 0 x = . (6)

The injection coefficient A is a measure of the porosity
of the wall and can be utilized as a control parameter in
the final solution.
By applying mass conservation to a volume of fluid
extending from 0 x = to x , the mean flow velocity
( )
m
u x can be found to be proportional to x :


m
ax
u
a
=
!
. (7)

0
a ( t )
0 2 4 6 8
d a / d t
x
y

Fig. 1. Coordinate system and bulk fluid motion.
AIAA-99-3523
3
American Institute of Aeronautics and Astronautics
At this point, the Stokes stream function can be
introduced. This allows replacing the two velocity
components by one single function

u
y

=

, and v
x

=

. (8)

By definition of the stream function, the continuity
equation is automatically satisfied and thus no longer
useful in determining .
A new variable can now be removed at the expense
of introducing another. In fact, the vorticity transport
equation for the given geometry can be arrived at by
taking the curl of the momentum equation. This
operation eliminates the pressure from the momentum
equation and introduces, in its stead, the flow vorticity.
This is illustrated below. Starting with


2
1 DV
p V
Dt

1

= +

( )
"
"
(9)
gives
2 2
2 2
u v
t x y x y

1


+ + = +

( )
(10)
where
v u
x y

=

. (11)
III. Mathematical Procedure
A. Similar Solution in Space
A similar solution with respect to x can be
developed due to mass conservation, since the variation
of the channel height is uniform along the axial
direction. Following Bermans classic approach, and in
view of the boundary conditions represented by Eqs. (4)
-(5), a similar solution can be assumed to be of the form

( ) ( , ) xf a F t = , where
y
a
= (12)

and ( , ) F t is independent of the axial coordinate.
Placing Eq. (12) into Eq. (8), the axial and transverse
velocities can be expressed as a function of F :

( )
x
u f a F
a

= , and ( ) ( ) , v f a F t = (13)

where / F F

= . At this point, we notice that

0
xx x
u v = = . The association ( ) 1/ f a a = can then
be determined when 0
xx x
u v = = is inserted into the
vorticity equation and the dimensions of the inertial and
viscous terms are compared. With this realization, the
axial and transverse velocities can be written in their
final form, namely,


2
x
u F
a

= , and ( ) , v F t
a

= . (14)

Noting that the transverse velocity v is independent of
x , the vorticity equation simply reads


u
y

(15)

and Eq. (3) will then be 0
yx
p = . Upon substitution of
Eq. (15) and Eq. (1) into the vorticity transport
expression given by Eq. (10), one obtains


yt yx yy yyy
u uu vu u + + = . (16)

By inserting the velocity expression from Eq. (14) into
Eq. (16), we can develop a differential equation for our
function F . The result is

( )
2
2 0
t
a
F FF F F F F

l
l
+ + + =
l
l
(17)

where is the Wall Expansion Ratio defined as

( )
aa
t

=
!
. (18)

Allowing 0
xy
p = in Eq. (2), this will also lend itself as
a method of arriving at Eq. (17). The boundary
conditions given by Eq. (4)-(5) should then be updated
to account for the normalization:


0, 0; 0,
0, ; 1 .
F F
F F A R

= = =
= = = =
(19)

By equating the expression we had before in Eq. (4),
v Aa = ! , with the expression for the transverse
velocity in Eq. (14) and inserting the expansion ratio
where applicable, we recognize that our function F is a
product of the injection coefficient A and the
expansion ratio . Since this occurs at the permeable
wall, we can write /
w
R A av = = where R is
the cross-flow Reynolds number.
B. Similar Solution in Space and Time
A similar solution with respect to space and time can
now be developed. If our function F is assumed to be
dependent on only and that is constant, it follows
AIAA-99-3523
4
American Institute of Aeronautics and Astronautics
that 0
t
F

= . The value of the expansion ratio can

be specified by its initial value


o o
a a aa

= =
! !
(20)

with
o
a and /
o o
a da dt = ! denoting the initial channel
height and channel expansion ratio, respectively. By
integrating Eq. (20) with respect to time, a similar
solution for the channel height variation with time can
be obtained. This is given by

1
2
2
2
1
o o
a t
a a

1

= +

( )
. (21)

Subsequently, an expression for the injection velocity
variation can be deduced provided that the injection
coefficient A in Eq. (4) is held constant:


( )
1
2
2
0
2
1
w
o w o
t
v a t
a v a

=
1

= = +

( )
!
!
. (22)

Now, an ODE for the function F can be realized by
direct integration of Eq. (17)

( ) 2 F FF F F F K + + + = (23)

where a prime replaces / d d and K is the constant of
integration. The necessary boundary conditions
translate into:


0, 0; 0,
0, ; 1 .
F F
F F A R

= = =
= = = =
(24)
C. Numerical Solution
Since we have four boundary conditions in Eq. (24)
that are imposed on Eq. (23), it is convenient to
eliminate the integration constant K by differentiating
Eq. (23) with respect to . The resulting governing
ODE becomes:

( ) 3 0
IV
F FF F F F + + + = . (25)

This can be solved numerically by applying a
shooting method coupled with fourth order Runge-
Kutta integration. The procedure adopted in resolving
the function ( ) F is described next.
1. Shooting Method
The boundary conditions for Eq. (25) can be
redefined as


( ) ( )
( ) ( )
( ) ( )
0 0, 0 0,
0 , 0
1 0, 1
F F
F k F t
F F A R
= =
= =
= = =
(26)

where t and k are initial guesses introduced as
dummy variables and t should not be confused with
time. We can now express Eq. (25) in the following
fashion:


( )
( )
, , , ,
3
IV
y f x y y y y
y y yy xy
=
=
(27)

where y should not be confused with the transverse
location. Thus,

( ) 3 f y y yy xy = . (28)

The boundary conditions become


( ) ( )
( ) ( )
0 , 0 0,
0 , 0 0.
y k y
y t y
= =
= =
(29)

To satisfy the boundary conditions at 1 = , the
functions ( ) , g t k and ( ) , h t k are introduced such that


( ) ( )
( ) ( )
, 1, , 0
, 1, , 0 0 .
g t k y t k R
h t k y t k
= =
= =
(30)

Based on Newtons Method, an algorithm for
calculating the zeros * t and * k of Eq. (30) can be
developed. The algorithm requires updating initial
guesses following

[ ]
( )
( )
1
1
1, ,
1, ,
n n n
t t y t k R
J
k k y t k

+
l l l
l l l
=
l l l
l l l
(31)

where n is the iteration index and [ ]
1
J

is the inverse
of the Jacobian matrix.
At this point, we introduce functions Z and
W according to

( ) , ,
y
Z x t k
t

, and ( ) , ,
y
W x t k
k

. (32)

Following this transformation, the boundary conditions
can be expressed as illustrated in the following
example for ( ) 0, ,
x
Z t k
AIAA-99-3523
5
American Institute of Aeronautics and Astronautics

,
x
y y
Z
x t t x
1 1

= =


( ) ( )


then,

( ) ( ) [ ] 0, , 0 1
x
Z t k y t
t t

l = = =
l

. (33)

Likewise, the boundary conditions for Z and W in
Eq. (28) become


( ) ( )
( ) ( )
( ) ( )
( ) ( )
0, , 0 0, , 0
0, , 1 0, , 0
0, , 0 0, , 0
0, , 0 0, , 0 .
x x
xx xx
xxx xxx
Z t k W t k
Z t k W t k
Z t k W t k
Z t k W t k
= =
= =
= =
= =
(34)

Next, the order of differentiation in Eq. (27) can be
switched such as


( )
4
4
xxxx
y f y
t t x t
1

= =

( )
,

Since / Z y t = , a fourth-order differential equation
for Z can be generated in the form


IV
y y y y
Z f Z f Z f Z f Z

= + + + (35)

whose boundary conditions are given in Eq. (34).
Similarly, an expression involving W is found to be


IV
y
y y y
W f W f W f W f W

= + + + (36)

with the boundary conditions listed in Eq. (34).
Returning to Newtons Method that solves for the
correct initial guesses, the Jacobian matrix is defined by


1 1
2 2
1
2

;
f f
t k
f f
t k
f y R
J
f y




l
=
l
=
l
=
l
l
, (37)

which simplifies to:

( ) ( )
( ) ( )
1 1
1 1
x x
Z W
J
Z W
l
l
=
l
l
. (38)

Once the inverse of the Jacobian in Eq. (38) is
determined, Eq. (31) reduces to


1
.
x
x
n n n
t t W W y R
c
k k Z Z y
+
l l l l
l l l l
=
l l l l
l l l l
(39)

Where ( ) ( ) ( ) ( )
1
1/det( ) [ 1 1 1 1 ]
x x
c J Z W W Z

= = .
2. Runge-Kutta Solution Composition
The shooting method described above was chosen for
its ease of conversion into a Runge-Kutta solution set.
Twelve variables are chosen to represent y , Z , W ,
and their respective derivatives out to the third order.
Differentiating each of these functions and performing
the appropriate substitutions provides the solution set.
The Runge-Kutta method is an effective approach in
this application since it permits us to enter the
prescribed boundary conditions only once for all twelve
variables. Of course, our solution requires the input of
initial guesses for the boundary condition values of t
and k which are then iterated according to


( ) ( ) ( ) ( )
( ) ( ) ( ) ( )
( ) ( ) ( ) ( )
( ) ( ) ( ) ( )
1
1
1 1 1 1
1 1 1 1
1 1 1 1
1 1 1 1
x
n n
x x
x
n n
x x
W y W R y
t t
Z W W Z
Z y R Z y
k k
Z W W Z
+
+
l +
l
= +

l
l
= +

. (40)
IV. Flow-Field Characterization
A. Non-dimensional Parameters
Once we obtain a numerical solution for the function
( ) F , the dimensionless stream function and axial and
transverse velocities will be available. The stream
function can be normalized at 1 = (at the wall) by
considering Eq. (24), which then suggests letting
( ) 1 / /
w
xF a xR a = = . Consequently,


w
F
R

= . (41)

Additionally, the mean axial velocity in Eq. (7)
becomes


m
u F
u R

= (42)

for any axial position. Finally, the velocity at the wall,
which is expressed as /
w
v Aa R a = = ! , becomes

AIAA-99-3523
6
American Institute of Aeronautics and Astronautics

w w
v F
v R

= = . (43)
The normalized stream function and velocity
components are shown in Figs. 2 and 3.
B. Pressure and Shear Distributions
To complete our flow-field analysis, the resulting
pressure gradients and shear stress distributions are
considered. An expression for the transverse pressure
gradient can be developed by substituting the velocity
components of Eq. (14) into Eq. (3) with F acting as a
function of only. The result is

( )
2 2
p a F FF F F

l
= + + +
l
l
. (44)
The transverse pressure distribution is found by
manipulating Eq. (44) before integration and
application of the boundary conditions of Eq. (24).
Proceeding by writing


( )
( )
0
2
2
c
p
p a
dp F FF F F

=
=
l
= + + +
l
l

,(45)

the pressure distribution in the transverse direction can
be non-dimensionalized and written as

( )
2 1
2
0
2
2 2 2
/
t
c
o o
p p
a
a a
p F F F F



=

= = + + + . (46)

Since F behaves as a function of only and
0
t
F

= in the axial direction, substitution of the

velocity components from Eq. (14) into the axial
momentum Eq. (2) yields the axial pressure gradient:

( )
2
4
2
x
x
p F FF F F yF
a

l
= + + +
l
l
. (47)

In light of Eq. (23), Eq. (47) can be simplified into


2
4 x
x
a
p K

= . (48)

This can be readily integrated and non-dimensionalized
to obtain the axial pressure distribution at any
transverse position:


4 2
2 2 4 2
2
a
o
o o o
p p
a K x
a a a
p

= = . (49)

Additionally, the shear stress due to flow past the
porous surface can be determined. Substituting the
velocity components from Eq. (14) into the shear stress
equation and non-dimensionalizing yields

( )
1
3
2 2 3
/
w
w
o
o o
a x
a
a a
F

=
= = . (50)
V. Results
To further expound the results of the presented flow,
a variety of values for the expansion ratio are
calculated for the example case scenario of 5 R = .
The stream function for flow under the given
parameters is illustrated in Fig. 2. Note that from a
comparison of Eq. (42) and (44), the relationships for
the normalized stream function and transverse
velocities are identical for the entire channel height
0 1 . Thus, Fig. 2 also serves to display the
transverse velocity distribution. The axial velocity
profiles for a range of expansion ratio values are
depicted in Fig. 3. As the expansion ratio increases, the
axial velocity in the center of the channel also
increases. However, at higher values of , as can be
observed at 20 = , the velocity near the half range
point of the channel height demonstrates reverse flow at
lower values of the cross-flow Reynolds number R .
0 . 0 0 . 2 0 . 4 0 . 6 0 . 8 1 . 0
0 . 0
0 . 2
0 . 4
0 . 6
0 . 8
1 . 0
= 2 0
= 1 0
= 5
= 1
= 0

w

=

v
/
v
w

Fig. 2. Stream function and transverse velocity

component (R = 5).

- 1 . 0 - 0 . 5 0 . 0 0 . 5 1 . 0
- 1
0
1
2
3
4
5
= 0
= 1
= 5
= 1 0
= 2 0
u

/

u
m

Fig. 3. Axial velocity profiles over range of

expansion ratios (R = 5).
AIAA-99-3523
7
American Institute of Aeronautics and Astronautics
The streamlines for the cases of 0 = and 20 =
are illustrated in Figs. 4a and 4b, respectively. For the
20 = plot, the streamlines enter at a steeper angle
than the stationary wall case since the former possesses
regressing walls which serve to increase the transverse
velocity near the walls. We also witness in Fig. 4b that
the streamlines rebound rapidly from their nearly co-
axial approach. This may ascribed to the regressing
walls which act to widen the channel and alleviate the
minor throttling effect that would occur for situations
experiencing high injection rates relative to the
expansion rate.
The transverse pressure distribution computed from
Eq. (46) is illustrated in Fig. 5. Note that the pressure
drop away from the walls increases as the expansion
ratio increases. This may be directly attributed the
increased widening of the channel which allows for less
restricted flow and, thus, the increased pressure drop.
Figure 6 illustrates the axial pressure distribution.
Here, the pressure drop decreases with increasing .
Such trends are in agreement with observations made
by Weissberg.
24
Figure 7 demonstrates the calculated
results of the shear stress at the walls. Clearly, the
shear stress increases as the flow travels further down
the channel. We also witness a rise in the shear stress
as the expansion ratio decreases. For large expansion
ratios, the negative axial velocities near the wall are
responsible for reversing the shearing stress direction.
VI. Conclusions
The current analysis presents a procedure that
establishes an exact similarity solution of the Navier-
Stokes equations in semi-infinite rectangular channels
with porous and uniformly expanding walls. It is
found that in rectangular chambers and ducts where
wall regression and injection occur simultaneously,
the streamlines, pressure distributions, and shear
stresses are time and space dependent. They can be
characterized by two dimensionless parameters: the
expansion ratio of the walls and the cross-flow
Reynolds number R. A shooting method, coupled
with a Runge-Kutta integration scheme, were utilized
to numerically solve the fourth order differential
0 . 0 0 . 5 1 . 0 1 . 5 2 . 0
0 . 0
0 . 5
1 . 0
a )
= 0
2 . 5
2 . 0
1 . 5
1 . 0
0 . 5
/ a
0
R

x / a
0
0 . 0 0 . 5 1 . 0 1 . 5 2 . 0
0 . 0
0 . 5
1 . 0
b )
= 2 0
2 . 5
2 . 0
1 . 5
1 . 0
0 . 5
/ a
0
R

x / a
0
Fig. 4. Streamline patterns for a) = 0 and b)
= 20 (R = 5).

0 . 0 0 . 2 0 . 4 0 . 6 0 . 8 1 . 0
- 8 0
- 6 0
- 4 0
- 2 0
0
= 0
= 1
= 5
= 1 0
= 2 0
p
t

Fig. 5. Normalized transverse pressure distribution

over a range of wall expansion ratios (R = 5).
0 . 0 0 . 5 1 . 0 1 . 5 2 . 0
- 8 0
- 6 0
- 4 0
- 2 0
0
= 0
= 1
= 5
= 1 0
= 2 0
p
a
x / a
0
Fig. 6. Normalized axial pressure distribution over a
range of wall expansion ratios (R = 5).

0 . 0 0 . 5 1 . 0 1 . 5 2 . 0
- 8 0
- 6 0
- 4 0
- 2 0
0
2 0
= 0
= 1
= 5
= 1 0
= 2 0

w
x / a
0
Fig. 7. Wall shearing stress distribution over a range
of wall expansion ratios (R = 5).
AIAA-99-3523
8
American Institute of Aeronautics and Astronautics
equation. The results helped explain the flow-field
character.


References
1
Berman, A. S., Laminar Flow in Channels with
Porous Walls, Journal of Applied Physics, Vol. 24,
No. 9, 1953, pp. 1232-1235.
2
Sellars, J. R., Laminar Flow in Channels with
Porous Walls at High Suction Reynolds Numbers,
Journal of Applied Physics, Vol. 26, No. 4, 1955, pp.
489-490.
3
Terrill, R. M., Laminar Flow in a Uniformly
Porous Channel, The Aeronautical Quarterly, Vol. 15,
1964, pp. 299-310.
4
Proudman, I., An Example of Steady Laminar
Flow at Large Reynolds Number, Journal of Fluid
Mechanics, Vol. 9, No. 4, 1960, pp. 593-612.
5
Morduchow, M., On Laminar Flow through a
Channel or Tube with Injection: Application of Method
of Averages, Quarterly Journal of Applied
Mathematics, Vol. 14, No. 4, 1957, pp. 361-368.
6
White, F. M., Jr., Barfield, B. F., and Goglia, M. J.,
Laminar Flow in a Uniformly Porous Channel,
Transactions of the American Society of Mechanical
Engineers: Journal of Applied Mechanics, Series E,
Vol. 25, 1958, pp. 613-617.
7
Taylor, G. I., Fluid Flow in Regions Bounded by
Porous Surfaces, Proceedings of the Royal Society,
London, Series A, Vol. 234, No. 1199, 1956, pp. 456-
475.
8
Yuan, S. W., Further Investigation of Laminar
Flow in Channels with Porous Walls, Journal of
Applied Physics, Vol. 27, No. 3, 1956, pp. 267-269.
9
Terrill, R. M., Laminar Flow in a Uniformly
Porous Channel with Large Injection, The
Aeronautical Quarterly, Vol. 16, 1965, pp. 323-332.
10
Terrill, R. M., and Shrestha, G. M., Laminar Flow
Through Parallel and Uniformly Porous Walls of
Different Permeability, Journal of Applied
Mathematics and Physics (ZAMP), Vol. 16, 1965, pp.
470-482.
11
Shrestha, G. M., and Terrill, R. M., Laminar Flow
with Large Injection Through Parallel and Uniformly
Porous Walls of Different Permeability, Quarterly
Journal of Mechanics and Applied Mathematics, Vol.
21, No. 4, 1968, pp. 413-432.
12
Cox, S. M., Two-dimensional Flow of a Viscous
Fluid in a Channel with Porous Walls, Journal of
Fluid Mechanics, Vol. 227, 1991, pp. 1-33.
13
Brady, J. F., and Acrivos, A., Steady Flow in a
Channel or Tube with an Accelerating Surface
Velocity. An Exact Solution to the Navier-Stokes
Equations with Reverse Flow, Journal of Fluid
Mechanics, Vol. 112, 1981, pp. 127-150.
14
Watson, E. B. B., Banks, W. H. H., Zaturska, M.
B., and Drazin, P. G., On Transition to Chaos in Two-
dimensional Channel Flow Symmetrically Driven by
Accelerating Walls, Journal of Fluid Mechanics, Vol.
212, 1990, pp. 451-485.
15
Watson, P., Banks, W. H. H., Zaturska, M. B., and
Drazin, P. G., Laminar Channel Flow Driven by
Accelerating Walls, European Journal of Applied
Mathematics, Vol. 2, 1991, pp. 359-385.
16
Varapaev, V. N., and Yagodkin, V. I., Flow
Stability in a Channel with Porous Walls, Fluid
Dynamics (Izvestiya Akademii Nauk SSSR, Meckanika
Zhidkosti i Gaza), Vol. 4, No. 5, 1969, pp. 91-95.
17
Raithby, G. D., and Knudsen, D. C.,
Hydrodynamic Development in a Duct with Suction
and Blowing, Transactions of the American Society of
Mechanical Engineers: Journal of Applied Mechanics,
Series E, Vol. 41, 1974, pp. 896-902.
18
Sviridenkov, A. A., and Yagodkin, V. I., Flow in
the Initial Sections of Channels with Permeable Walls,
Fluid Dynamics (Izvestiya Akademii Nauk SSSR,
Meckanika Zhidkosti i Gaza), Vol. 11, No. 5, 1976, pp.
689-693.
19
Ma, Y., Van Moorhem, W. K., and Shorthill, R.
W., Innovative Method of Investigating the Role of
Turbulence in the Velocity Coupling Phenomenon,
Journal of Vibration and Acoustics-Transactions of the
ASME, Vol. 112, No. 4, 1990, pp. 550-555.
20
Ma, Y., Van Moorhem, W. K., and Shorthill, R.
W., Experimental Investigation of Velocity Coupling
in Combustion Instability, Journal of Propulsion and
Power, Vol. 7, No. 5, 1991, pp. 692-699.
21
Barron, J., Majdalani, J., and Van Moorhem, W.
K., A Novel Investigation of the Oscillatory Field over
a Transpiring Surface, AIAA Paper 98-2694, June
1998.
22
Goto, M., and Uchida, S., Unsteady Flows in a
Semi-infinite Expanding Pipe with Injection Through
Wall, Transactions of the Japan Society for
Aeronautical and Space Sciences, Vol. 33, No. 9, 1990,
pp. 14-27.
23
Uchida, S., and Aoki, H., Unsteady Flows in a
Semi-infinite Contracting or Expanding Pipe, Journal
of Fluid Mechanics, Vol. 82, No. 2, 1977, pp. 371-387.
24
Weissberg, H. L., Laminar Flow in the Entrance
Region of a Porous Pipe, The Physics of Fluids, Vol.
2, No. 5, 1959, pp. 510-516.