Chapter 3

Unidirectional transport
In this chapter, we consider transport in which there is a variation in the
mass, momentum and temperature fields in only one dimension. The analysis is considerably simplified in this case, since there is only one spatial
coordinate to be considered. However, the examples solved here illustrate
the basic principles of the solution of more complex problems in multiple
dimensions, which involve shell balances to derive differential equations for
the concentration, velocity and temperature fields, and then an integration
procedure for determining the variations in the concentration, velocity and
temperature.

3.1

Solutions of the diffusion equation

3.1.1

Unsteady transport into an infinite fluid

Mass transfer
Consider a flat surface in the x y plane located at z = 0 through which the
solvent diffuses into the fluid. The plane surface is assumed to be of infinite
extent in the x y plane, and the height of the fluid supported by the plane
is also considered to be of infinite extent. Initially, the concentration of
the solute in the solvent at the surface is c . At time t = 0, the solute
concentration at the surface is instanteneously increased to c0 , and there
is diffusion of the solute into the solvent. We would like to determine the
variation of the concentration of the solute with time.
As the diffusion proceeds, there is transport of solute from the surface
1

CHAPTER 3. UNIDIRECTIONAL TRANSPORT

to the solvent, resulting in an increase in the concentration near the surface.

However, the concentration far from the surface z remains unchanged
at c = c . The conditions for the concentration field at the spatial boundaries and at the initial time are
c = c
c = c0
c = c

as z at all t
at z = 0 at all t > 0
at t = 0 for all z > 0

(3.1)

The problem is simplified by defining a non-dimensional concentration

c =
The conditions for c are
c = 0
c = 1
c = 0

c c
c0 c

as z at all t
at z = 0 at all t > 0
at t = 0 for all z > 0

(3.2)

(3.3)

Heat transfer
The configuration consists of a flat surface in the xy plane of infinite extent,
and a fluid of conductivity K and specific heat Cp in the half space z > 0.
Initially, the temperature of the fluid and the surface is T . At time t = 0,
the temperature of the surface is instanteneously increased to T0 . Determine
the temperature as a function of time.
As conduction proceeds, there is a transport of energy from the surface to
the fluid, resulting in an increase in the temperature of the fluid. However,
the temperature at a large distance from the surface, z , remains at T =
T . Therefore, the conditions for the temperature at the spatial boundaries
of the fluid and at initial time are
T = T
T = T0
T = T

as z at all t
at z = 0 at all t > 0
at t = 0 for all z > 0

(3.4)

The problem is simplified by defining a non-dimensional temperature

T =

T T
T0 T

(3.5)

3.1. SOLUTIONS OF THE DIFFUSION EQUATION

The conditions for T are

T = 0
T = 1
T = 0

as z at all t
at z = 0 at all t > 0
at t = 0 for all z > 0

(3.6)

Momentum transfer
The configuration consists of an infinite fluid in the z > 0 half space bounded
by an infinite flat surface in the x z plane. The fluid and the surface are
initially at rest. At time t = 0, the plane is instanteneously moved with
a constant velocity U in the x direction. Determine the fluid velocity as a
function of time.
As time proceeds, the momentum that is transported from the surface
diffuses through the fluid, resulting in fluid motion. However, the fluid at
a large distance from the surface z remains at rest. If ux is the fluid
velocity in the x direction, it is convenient to define a non-dimensional fluid
velocity ux = (ux /U ). The conditions for the non-dimensional fluid velocity
at the spatial boundaries of the flow and at initial time are
ux = 0
ux = 1
ux = 0

as z at all t
at z = 0 at all t > 0
at t = 0 for all z > 0

(3.7)

Shell balance
In all three cases, the boundary and initial conditions 3.3, 3.6 and 3.7 are
identical in form. Since the transport in all three cases involves diffusion in
the direction perpendicular to the surface, and no variation along the surface,
it is anticipated that the differential equation governing the transport will
also be identical in form. The differential equation for the concentration field
is first derived using a shell balance, and the analogous equations for heat
and momentum transfer are provided.
Consider a shell of thickness z in the z coordinate as shown in figure 3.1,
and of thickness x and y in the x y plane. There is a transport of mass
across the surfaces of the shell due to diffusion, which results in a change in
the concentration in the shell. We consider the variation in the concentration within this control volume over a time interval t. Mass conservation

CHAPTER 3. UNIDIRECTIONAL TRANSPORT

requires that
Accumulation of mass
in the shell

Input of
mass into shell

Output of
mass from shell
(3.8)

The accumulation of mass in a time t is given by

Accumulation of mass
in the shell

= (c(x, y, z, t+t)c(x, y, x, t))xyz (3.9)

The mass flux at the surface at z is given by



c 
jz = D 
z z

(3.10)

and the mass entering the shell through the surface at z in a time interval
t is given by the product of the mass flux, the area of transfer and the time
interval t,
Input of
mass into shell

c
= txy D
z

(3.11)

In a similar manner, the mass leaving the surface at z + z is given by

Output of
mass from shell

c
= txy D
z

(3.12)
z+z

Substituting equations 3.9, 3.11 and 3.12 into equation 3.8, and dividing by
xyzt, we obtain
c(x, y, z, t + t) c(x, y.z, t)
1
=
t
z

c 
c 
D 
D 
z z+z
z z

(3.13)

The limit t 0 and z 0 is taken, to obtain a partial differential

equaiton for the concentration field.
c
c

D
=
t
z
z

(3.14)

If the diffusion coefficient is independent of the spatial position, the differential equation 3.14 reduces to
c
2c
=D 2
t
z

(3.15)

3.1. SOLUTIONS OF THE DIFFUSION EQUATION

z>Infinity
c*=0
T*=0
u*=0
x
z+ z
z

z=0
c*=1
T*=1
u*=1
x

Figure 3.1: Configuration for similarity solution for unidirectional transport.

The concentration equation 3.15 can also be expressed in terms of the scaled
concentration variable, c , defined in 3.2,
2 c
c
=D 2
t
z

(3.16)

A similar shell balance procedure can be carried out for momentum transfer, and the equations for the temperature and velocity fields are

T
T
=
DT
t
z
z

(3.17)

Though the final result for the momentum transfer process is exactly
analogous to equations 3.16 and 3.17, the procedure is slightly different, and
so we provide a brief outline of the calculation. First, note that there are

CHAPTER 3. UNIDIRECTIONAL TRANSPORT

now two directions in the problem. Since momemtum is a vector, there is a

direction associated with the momentum itself, which is the x direction. The
second is the direction of diffusion, which is the z direction. The fundamental
balance relation, equivalent of equation 3.8, is

Rate of change of

x momentum

=
in the shell

Sum of forces
in x direction

(3.18)

The rate of change of momentum in the differential volume of thickness z

about z in a time interval t is given by,

Rate of change of
ux xyz

x momentum
=

t
in the shell

(3.19)

The forces acting on the two surfaces at z and z + z are the products of the
shear stress xz and the surface area (xy). It is important to keep track
of the directions of the forces in this case, since the force is a vector. The
shear stress xz is defined as the force per unit area in the x direction acting
at a surface whose outward unit normal is in the positive z direction. For the
surface at z +z, the outward unit normal is in the +z direction, as shown in
figure 3.1, and therefore the force per unit area at this surface is + xz |z+z .
For the surface at z, the outward unit normal is in the z direction, and
therefore the force per unit area at this surface is xz |z . Finally, the rate
of change of momentum, which is the change in momentum per unit time, is
given by (ux )(Az)/t. Therefore, the momentum balance equation is,
(Az)

ux
= A( xz |z+z xz |z )
t

(3.20)

Dividing throughout by Az, we obtain,

xz |z+z xz |z
ux
=
t
z

(3.21)

Taking the limit t 0 and z 0, we obtain the partial differential

equation,
xz
ux
=
(3.22)

t
z

3.1. SOLUTIONS OF THE DIFFUSION EQUATION

The shear stress is given by the product of the viscosity and the gradient of
the velocity,
ux
ux
xz = lim
=
(3.23)
z0 z
z
With this, the governing equation for the velocity field becomes,

ux
ux
=

t
z
z

(3.24)

Note that the differential equation derived above has the same form as the
concentration and energy diffusion equations 3.16 and 3.17, though it was
derived from a force balance. This indicates that the diffusion process is the
same for mass, momentum and energy. However, it should be noted that
momentum could be transmitted by pressure forces in addition to viscous
forces, and there is no analogue of pressure in mass and energy transport.
When the velocity ux is scaled by U , the velocity of the bottom surface,
the momentum equation becomes,

u
ux
=
x
t
z
z

(3.25)

where = (/rho) are the thermal and momentum diffusivities respectively.

Solution procedure
Since the conservation equations 3.16, 3.17 and 3.25 are identical in form, and
the boundary conditions 3.9, 3.11 and 3.12 are also identical, the same solution procedure can be used for all these, and the solutions for the concentration, velocity and temperature files, expressed in terms of the dimensionless
variables c , T and ux , turn out to be identical.
In order to solve the concentration equation 3.16 with the boundary condition 3.9, it is first important to realise that there no intrinsic length scale
in the problem, because the fluid and the flat plate are of infinite extent.
Since the concentration c is dimensionless, there are only three dimensional
variables z, t and D in the problem. These contain two dimensions, L and
T,
and it is possible to construct only one dimensionless number, = (z/ Dt).
Therefore, just from dimensional analysis, it can be concluded that the concentration field doesnot vary independently with z and t, but only on the
combination = (z/ Dt). If this inference is correct, it should be possible to

CHAPTER 3. UNIDIRECTIONAL TRANSPORT

express the conservation equation 3.16 in terms of the variable alone. When
z and t are expressed in terms of , the concentration equation becomes

z
2D 1/2 t3/2

D 2 c
c
=

Dt 2

(3.26)

After multiplying throughout by t, the equation for the concentration field

reduces to
c 2 c
+
=0
(3.27)
2
2
Equation 3.27 validates the inference that the non-dimensionalised concentration field is only a function of . It is also necessary to transform the
boundary conditions 3.9 into conditions for the coordinate. The transformed boundary conditions are
c = 0
as z at all t
as

c = 1 at z = 0 at all t > 0 at = 0
c = 0 at t = 0 for all z > 0 as

(3.28)

It is useful to note that the original conservation equation, 3.16, is a second

order differential equation in z and a first order differential equation in t, and
so this requires two boundary conditions in the z coordinate and one initial
condition. The conservation equation expressed in terms of is a second
order differential equation, which requires just two boundary conditions for
. Equation 3.28 shows that one of the boundary conditions for z and
the initial condition t = 0 turn out to be identical conditions for ,
and therefore the transformation form (z, t) produces no inconsistency in the
boundary and initial conditions.
Equation 3.27 can be easily solved to obtain

c () = C1 + C2

2
d exp
2
0

(3.29)

The constants C1 and C2 are determined from the conditions c = 1 at = 0,

and c = 0 for , to obtain

c () =

2
d
exp

2
(z/ Dt)

(3.30)

3.1. SOLUTIONS OF THE DIFFUSION EQUATION

From the solution 3.30, it can be inferred that there is no intrinsic length
scale in the system, but the
length scale for the z coordinate is a function of
time, and is proportional to Dt at time t. Thus, the length scale over which
diffusion has taken place increases proportional to t1/2 . Thus, equation 3.30
will be a solution for the concentration diffusion in a configuration bounded
by two plates,
so long as the distance between the two plates L is large
compared to Dt.

3.1.2

Steady diffusion into a falling film

This problem is a simplification of the actual diffusion in a falling film, which

involves a combination of convection and diffusion. A thin film of fluid flows
down a vertical surface. One side of the film is in contact with a gas which
is soluble in the liquid, and as the liquid flows down, the gas is dissolved
in the liquid. The concentration of gas in the liquid at the entrance is c ,
while the concentration of gas at the liquid-gas interface is c0 . The difference in concentration between the initial concentration in the liquid and the
concentration at the interface acts as a driving force for diffusion. The z
coordinate is perpendicular to the gas-liquid interface, which is located at
z = 0. We also assume that the penetration depth for the gas into the liquid
(to be determined a little later) is small compared to the thickness of the
liquid film, so that the boundary conditions in the z coordinate given by 3.9
are applicable. In addition, the boundary condition in the x coordinate is
c = c

3.1.3

at x = 0

(3.31)

Diffusion in bounded channels

Next we consider the problem of diffusion in a channel bounded by two walls

separated by a distance L, as shown in figure 3.2. Initially, the concentration
of the fluid in the channel is equal to c . At initial time, the concentration
of the solute on the wall at z = 0 is instanteneously increased to c0 , while
the concentration on the surface at z = L remains equal to c . The problem
is to find the concentration field as a function of time. Similar problems can
be formulated for heat and momentum transfer.
The concentration field is first expressed in terms of the scaled concentration field c by equation 3.2. The diffusion equation, obtained by a shell
balance as before, is given by equation 3.16. However, there is a modification

10

CHAPTER 3. UNIDIRECTIONAL TRANSPORT

z=L
c*=0
T*=0
u*=0
x

z+ z
z

z=0
c*=1
T*=1
u*=1
x

Figure 3.2: Configuration for similarity solution for unidirectional transport.

in the boundary conditions,
c = 0
c = 1
c = 0

at z = L at all t
at z = 0 at all t > 0
at t = 0 for all z > 0

(3.32)

In this case, it is not possible to effect a reduction to a similarity form,

because there is an additional length scale L in the problem, and so the z
coordinate can be scaled by L. A scaled z coordinate is defined as z = (z/L),
and the diffusion equation in terms of this coordinate is
c
D 2 c
= 2 2
(3.33)
t
L z
The above equation suggests that it is appropriate to define a scaled time
coordinate t = (Dt/L2 ), and the conservation equation in terms of this
scaled time coordinate is
2 c
c
=
(3.34)
t
z 2
The boundary conditions, in terms of the scaled coordinates z and t , are
c = 0

at z = 1 at all t

3.1. SOLUTIONS OF THE DIFFUSION EQUATION

c = 1
c = 0

at z = 0 at all t > 0
at t = 0 for all z > 0

11

(3.35)

It is expected that in the long time limit, t , the concentration field

will attain a steady state value cs which is independent of time. This steady
state concentration field is obtained by solving 3.16 with the time derivative
set equal to zero, and the solution for the steady concentration field is a
linear concentration profile,
c = C 1 z + C 2

(3.36)

The constants C1 and C2 are obtained from the boundary conditions at z = 0

and z = 1,
cs = (1 z )
(3.37)
The variation of concentration with time can be determined by dividing
the concentration into a steady and an unsteady part,
c = cs + cu ,

(3.38)

where cu is the difference between the actual concentration and the concentration at steady state. The reason for this decomposition will become clear
a little later. The conservation equation for the unsteady concentration field
is identical to that for the original concentration field, because (cs /t ) = 0
and ( 2 cs /z 2 ) = 0,
cu
2 cu
=
(3.39)
t
z 2
However, the boundary condition for cu is different from that for the c , and
is obtained by subtracting cs from c at the boundaries,
cu = 0
cu = 0
cu = cs

at z = 1 at all t
at z = 0 at all t > 0
at t = 0 for all z > 0

(3.40)

Equation 3.16 can be solved by the method of separation of variables,

where the unsteady concentration field is expressed as two functions, one of
which is only a function of t , while the other is only a function of z .
cu = (t )Z(z )

(3.41)

12

CHAPTER 3. UNIDIRECTIONAL TRANSPORT

This is inserted into the conservation equation 3.16, and the equation is
divided by the production Z, to obtain
1 d2 Z
1 d
=
dt
Z dz 2

(3.42)

In equation 3.42, the left side is only a function of t , while the right side
is only a function of z . From this, it can be inferred, as follows, that these
two functions have to be constants independent of z and t . To infer this,
assume that these two functions are not constants, and that the left side
varies as t is varied, and the right side varies as z is varied. In this case,
if we keep z a constant and vary t , then the left side of 3.42 varies, while
the right side remains a constant, and so the equality is destroyed. The only
way for the equality to hold, if the left side is only a function of t and the
right side is only a function of z , is if the two sides are constants.
The solution for Z is first obtained by solving
1 d2 Z
= 2
Z dz 2

(3.43)

where is a positive constant. The reason for choosing the right side of
3.43 to be negative will become apparent a little later. The solution for this
equation is
Z = C1 sin (z ) + C2 cos (z )
(3.44)
where C1 and C2 are constants to be determined from the boundary conditions. The boundary condition cu = 0 (Z = 0) at z = 0 is satisfied for
C2 = 0. The boundary condition cau = 0 at z = 1 is satisfied if = (n).
Therefore, the solution for Z which satisfied the boundary conditions in the
z coordinate is
Z = C1 sin (nz )
(3.45)
where n is an integer.
The solution for can now be obtained from the equation
1 d
= 2 = n2 2
dt

(3.46)

This equation is solved to obtain

= C3 exp (n2 2 t )

(3.47)

3.1. SOLUTIONS OF THE DIFFUSION EQUATION

13

The final solution for cu = (t )Z(z ) is

cu = C exp (n2 2 t ) sin (nz )

(3.48)

The solution 3.48 contains the integer n which is as yet unspecified, and it is
verified that the solution 3.48 satisfies the equation 3.46 for any value of n.
The most general solution is one which contains a linear combination of the
solution 3.48 for different values of n,
cu =

Cn exp (n2 2 t ) sin (nz )

(3.49)

n=1

The values of the coefficients Cn have to be determined from the boundary

condition that has not been used so far,

n=1

cu = (1 z )

for t = 0

Cn sin (nz ) = (1 za)

(3.50)

The coefficients can be determined because the functions sin (nz ) satisfy
orthogonality conditions,
Z

1
0

dz sin (nz ) sin (mz ) = 0 for n 6= m

= (1/2) for n = m

(3.51)

To use this condition, the left and right sides of 3.50 are multiplied by
sin (mz ), and integrated over the interval 0 z 1, to obtain
Cm = 2

1
0

dz sin (mz )(1 z )

1
m

(3.52)

The reason for choosing the right side of 3.46 to be negative is now apparent. If we had chosen the right side of 3.46 to be positive, the solution for
Z consists of an exponentially growing and an exponentially decaying function. In this case, it can easily be verified that it is not possible to satisfy
the boundary condtions c = 0 at z = 0 and at z = 1 simultaneously. In
addition, the solution for in equation 3.47 would have been a function that
is exponentially increasing in time, and therefore, there is no steady solution
in this case.

14

CHAPTER 3. UNIDIRECTIONAL TRANSPORT

3.1.4

Unsteady diffusion in a cylinder

A cylindrical container of radius R containing fluid with temperature T is

dipped into a bath of fluid containing solute with temperature T0 . Assume
that the bath is large enough that the conduction of heat into the fluid does
not appreciably change the temperature of the bath, and that there is no
resistance to heat transfer in the walls of the container. The temperature in
the cylinder as a function of time and position is to be determined.
A cylindrical coordinate system is used, as shown in figure 3.3, where z
is the axial coordinate and r is the radial coordinate. There is no variation
of temperature in the z coordinate, since the cylinder is considered to be of
infinite extent. In order to obtain a differential equation for the temperature
variation in the cylinder, a shell of thickness r and height z at radius r
is considered. The terms in the balance equation 3.8 for the energy in the
cylindrical shell are as follows.
Accumulation of energy
in the shell

= Cp (T (x, y, z, t+t)T (x, y, x, t))2rrz

(3.53)
The total energy entering the shell at r is the product of the heat flux, the
surface area, and the time interval t,
Input of
energy into shell

T
= K
r




(2rzt)


(3.54)
r

Similarly, the total energy leaving the shell at r + r is

Output of
energy from shell

T
= K
r




(2rzt)


(3.55)
r+r

When these are inserted into the conservation equation 3.8, and divided by
rt, the net energy balance for the shell is


(T (x, y, z, t + t) T (x, y, x, t))

1
Cp r
=
t
r

T 
T 
Kr


Kr

r r+r
r r
(3.56)
Taking the limit r 0 and t 0, the partial differential equation for
the temperature field is
DT
T
T
=
r
t
r r
r

(3.57)

3.1. SOLUTIONS OF THE DIFFUSION EQUATION

15

R
r r

T*=0
z

z
r
Figure 3.3: Heat diffusion into a cylinder.
It is important to note that there is a variation in the surface area of the
shell as r varies, and there is a variation in the total energy transported
through the shell due to the variation in this surface area. This leads to a
more complicated form for the diffusion term in 3.57, in comparison to the
second derivative encountered in the diffusion from a flat plane, 3.17.
The conservation equation 3.57 can be scaled as follows. The scaled
temperature field is defined by equation 3.5, while the scaled distance and
time coordinates are defined as r = (r/R) and t = (tR2 /DT ), since R
is the length scale in the radial direction. The conservation equation 3.57,
expressed in these coordinates, is

1
T
T
=
r
t
r r
r

(3.58)

16

CHAPTER 3. UNIDIRECTIONAL TRANSPORT

The boundary conditions are

T = 0
T = 1
T
=0
r

at r = 1 for all t
at t = 0 for all r
at r = 0 for all t

(3.59)

Equation 3.58 is solved using the method of separation of variables. The

final steady state solution is one in which the temperature is uniform, T = 0,
throughout the cylinder. The unsteady solution is solved using the substitution
T = R(r )(t )
(3.60)
where R is a function of the radial coordinate, and is a function of time.
The form 3.60 is inserted into the temperature equation, 3.58, and the resulting equation is divided by R(r )(t ) to obtain
1 1
R
1
=
r

t
R r r
r

(3.61)

The left side of the above equation is only a function of time, while the right
side is only a function of r . Therefore, the equality can only be satisfied if
both sides are equal to constants. The right side of the above equation is
first solved,
!
1 dR
1 d2 R
= 2
(3.62)
+
R dr 2 r dr
A special function solution for equation 3.62 can be obtained by recasting
the equation as
d2 R
dR
r 2 2 + r + 2 r R = 0
(3.63)
dr
dr
The solution of the above equation is in the form of Bessel functions,
R = C1 J0 (r ) + C2 Y0 (r )

(3.64)

The values of the constants C1 and C2 are determined from the boundary
conditions at r = 0 and r = 1. At r = 0, the radial derivative of J0 (r ) is
identically zero, whereas the radial derivative of Y0 (r ) approaches infinity
for r 0. Therefore the boundary condition at r = 0 can be satisfied only

3.1. SOLUTIONS OF THE DIFFUSION EQUATION

17

if C2 = 0. The second boundary condition, T = 0 at r = 1, is used to

determine the value of in equation 3.62,
J0 () = 0

(3.65)

There are multiple solutions for this equation, the first few of which are given
by. The equation for ,
1 d
= n2
(3.66)
dt
can be solved to obtain
= exp (n2 t )
(3.67)
With this, the solution for the temperature field is

T =

Cn J0 (n r ) exp (n2 t )

(3.68)

n=1

The coefficients Cn should satisfy the initial condition at t = 0,

Cn J0 (n r ) = 1

(3.69)

n=1

The values of the coefficients can be determined using orthogonality conditions for the Bessel functions, which in this case are
Z

1
0

1
(J1 (n ))2 for m = n
2
= 0 for m 6= n

r dr J0 (n r )J0 (m r ) =

(3.70)

In order to determine the coefficients, the right and left sides of 3.70 are
multiplied by r J0 (m r ), and integrated from r = 0 to r = 1, to obtain
Z 1
1
(J1 (n ))2 Cn =
r dr J0 (n r )
2
0

3.1.5

(3.71)

Oscillatory flow

This example is used to illustrate the use of complex variables in problems

where the forcing on the fluid is oscillatory in time. Consider the flat plate
bounding a fluid in the half space z > 0, shown in figure 3.4, with the

18

CHAPTER 3. UNIDIRECTIONAL TRANSPORT

modification that the plate has an oscillatory velocity U = U0 cos (t). The
differential equation for the velocity field is given by equation 3.25,
ux
2 ux
= 2
t
z

(3.72)

The boundary conditions in this case, analogous to 3.7, are

ux = U cos (t)
ux = 0

at z = 0
at z = L

(3.73)

The solution procedure is simplified if the boundary condition is recast as

follows. The complex velocity field ux is defined as a velocity field that
satisfies the same differential equation as 3.72,
ux
2 u
= 2x
t
z

(3.74)

but which satisfies the boundary conditions

ux = U exp (t)
ux = 0

at z = 0
at z = L

(3.75)

where is the square root of 1. It is apparent that the solution to the

differential equation 3.74, with the boundary condition 3.75, is the real part
of the solution to the differential equation 3.72 with the boundary condition
3.73. Since dealing with exponential functions is easier than dealing with
sines and cosines, it is more convenient to solve equation 3.74 with boundary
conditions 3.75 for ux , and then take the real part of the solution to obtain
the solution ux of equation 3.72.

The non-dimensional velocity u

x is defined as ux = (ux /U ), while the
scaled z and time coordinates are defined as z = (z/L) and t = (t/L2 ).
The momentum conservation equation and boundary conditions in terms of
this scaled velocity are
2 u
u
x
x
=
(3.76)
t
z 2

u
x = exp ( t )
u
x = 0

at z = 0
at z = L

(3.77)

19

3.1. SOLUTIONS OF THE DIFFUSION EQUATION

z>Infinity
~
u*=0
x
z+ z
z

z=0
~
u*=1
x

ux=U cos( t)
Figure 3.4: Oscillatory flow at a flat surface.

where the scaled frequency is given by = (L2 /).

The differential equation 3.76 for the velocity field is a linear differential
equation, since all terms in the equation contain only the first power of u
x .
This first order differential equation is driven by a wall which is oscillatory
wall velocity with scaled frequency . When a linear system is driven by
wall motion of frequency , the response of the system also has the same
frequency . (This is not true if the system is non-linear, since forcing of a
certain frequency will generate response at different harmonics of this base
frequency). Therefore, the time dependence of the velocity field in the fluid
can be considered to be of the form
u
x (z ) exp ( t )
x = u

(3.78)

When this form is inserted into the differential equation 3.76, and divided by
exp ( t ), the resulting equation is an ordinary differential equation for u
x .
u
x =

2u
x
z 2

The boundary conditions for u

x (3.77) become
u
x = 1

at z = 0

(3.79)

20

CHAPTER 3. UNIDIRECTIONAL TRANSPORT

u
x = 0

at z = L

This equation is easily solved to obtain

u
x (z ) = C1 exp ( z ) + C2 exp ( z )

(3.80)

(3.81)

where C1 and C2 are determined from the boundary conditions at z = 0

and z = 1,

exp ( z ) exp ( (2 z ))

u
x (z ) =
(3.82)
1 exp (2 )
The physical velocity field, which is the real part of the product of ux and
exp ( t ), is

#
"
exp ( z ) exp ( (2 z ))

exp ( t )
(3.83)
ux (z ) = Real
1 exp (2 )

Before solving the above equation, it is useful to examine some limiting

cases for the frequency of oscillation. In the limit 0, the solution for
the fluid velocity, 3.83, is
ux (z ) = (1 z ) exp ( t )

(3.84)

The physical reason for this result is as follows. The scaled frequency of
oscillation is obtained by dividing the frequency, , by the inverst of the
time required for the momentum to diffuse across the length of the channel,
(L2 /). If the scaled frequency is small, then the time period of variation
of the velocity of the bottom plate is long compared to the time required for
momentum to diffuse across the channel. In this case, it is expected, that
the velocity profile at any time instant along the oscillation cycle is identical
to the velocity for a flow driven by a steady plate velocity which is equal to
the instanteneous value of the plate velocity at that instant in the cycle. In
the limit  1, the fluid velocity field is given by

(3.85)
ux (z ) = exp ( z )

In this case, the velocity field decreases over a distance z (1/ ) from
the surface. This is because the frequency of oscillation is large compared
to the time required for diffusion of momentum across the
channel, and the
momentum diffuses only to a distance comparable to (L/ ). Beyond this
distance, the momentum generated during the positive and negative parts of
a cycle cancel out, and the fluid velocity approaches zero.

3.2. EFFECT OF BULK FLOW AND REACTION IN MASS TRANSFER21

Dry air
z=L

z
z=0
































































Water

















Figure 3.5: Diffusion with bulk flow.

3.2

Effect of bulk flow and reaction in mass

transfer

In this section, the special effects of bulk flow and reactions on the solutions
for unidirection mass transfer problems are examined.

3.2.1

Diffusion in a stagnant film

Water evaporates from a container through a stagnant air film through a

glass tube into dry air flowing at the top of the tube, as shown in figure 3.5.
If the mole fraction at the surface of the liquid surface is the saturation mole
fraction xW s , and the dry air flowing past the tube does not contain any
water, what is the concentration profile of water in the glass tube?
Though the air in the tube is stationary, the mean velocity across any
horizontal surface in the tube is not zero, because of the flow of water vapour

22

CHAPTER 3. UNIDIRECTIONAL TRANSPORT

across the surface. The flux of water across a surface, jW z , contains a component due to the bulk flow, as well as a component due to the diffusion of
water across the surface.
jW z = cDW A

dxW
+ xW (jW z + jAz )
dz

(3.86)

The last term on the right side of equation 3.86 is the flux of water due to the
bulk flow, where the total molar flow rate is the sum of the fluxes of water
(jW z ) and air (jAz ). In this particular case, the flux of air is identically zero,
and so the flux of water vapour across a surface is given by
jW z =

dxW
c
1 xW dz

(3.87)

The mass balance equation is obtained by writing a flux balance across a

section of thickness z of the tube, which at steady state provides,
jW z |z+z jW z |z = 0

(3.88)

If the above equation is divided by z the differential equation for the flux
in the limit z 0 is
djW z
d
=
dz
dz

1 dxW
1 xW dz

=0

(3.89)

This equation is solved to obtain

log (1 xW ) = A1 z + A2

(3.90)

The constants A1 and A2 are determined from the boundary conditions xW =

xW s at z = 0 and x = 0 at z = l,
(1 xW )
=
(1 xW s )

3.2.2

(1 xW f )
(1 xW s )

!z/l

(3.91)

Diffusion with homogeneous reaction

A gaseous reactant A dissolves in a liquid B, and undergoes a first order

reaction A+B AB, in a tank of height L, as shown in figure 3.6. The mass
balance equation, 3.14 has to be modified in this case due to the presence of

23

3.3. EFFECT OF PRESSURE ON MOMENTUM TRANSPORT

a consumption term due to chemical reaction. The mass balance equation

at steady state takes the form
jAz |z jAz |z+z kcA Sz = 0

(3.92)

This can be reduced to a differential equation by dividing throughout by

Sz, and taking the limit z 0,
djAz
kCA = 0
dz

(3.93)

If the concentration of A is small, the flux of A is given by

jAz = DAB

dcA
dz

(3.94)

Inserting this into the concentration equation 3.93, we get

DAB

d 2 cA
+ kCA = 0
dz 2

(3.95)

The boundary conditions are

cA = cA0 at z = 0
jAz = 0 at z = L

(3.96)

The solution that satisfied both these conditions is

CA
cosh ((kL2 /DAB )1/2 (1 (z/L)))
=
CA0
cosh (kL2 /DAB )1/2

3.3

(3.97)

Effect of pressure on momentum transport

The momentum balance condition states that the rate of change of momentum is equal to the applied force.
"

Rate of
momentum in

"

Rate of
momentum out

"

Sum of forces
acting on the system

=0
(3.98)

24

CHAPTER 3. UNIDIRECTIONAL TRANSPORT

c=c at z=0

z=0
z
z+ z

z=L

jz=0 at z=L
Figure 3.6: Diffison with homogeneous chemical reaction.

3.3. EFFECT OF PRESSURE ON MOMENTUM TRANSPORT

25

This balance equation is written for a control volume of fluid which is in the
form of a thin shell. Momentum enters or leaves the control volume due to
fluid flow into or out of the control volume, or due to the stresses acting on
the surface due to viscosity. The forces acting on the system are usually the
gravitational force or the centrifugal force. Momentum balances are usually
easy to apply only if the streamlines are straight. Applying momentum
balances to systems with curved streamlines is more difficult, as we shall see
in the last example of this section.
The procedure for solving problems with momentum balance is to write
the momentum balance equation for a shell of finite thickness, and then
let the thickness go to zero. In this limit, the difference equations for the
velocity field across a finite shell reduces to differential equations for the
velocity fields. These can then be solved, subject to appropriate boundary
conditions, in order to determine the velocity fields.
The boundary conditions generally involve specifying the velocity or stress
fields at the boundaries of the flow. These conditions depend on the surface
adjoining the liquid at its boundaries.
1. If there is a solid surface adjacent to the fluid, the appropriate boundary
condition is the no slip condition which states that the velocity of the
fluid at the surface is equal to the velocity of the surface itself.
2. At a liquid - gas interface, the momentum flux, and consequently the
velocity gradient, in the liquid side is assumed to be zero, because the
viscosity of the gas is small compared to that of the liquid.
3. At a liquid - liquid interface, the momentum flux and the velocity are
continuous across the interface.
Falling film
We first consider the flow of a falling film along an inclined plane, as shown
in figure 3.7. This type of flow is often encountered in cooling towers, evaporation chambers and gas absorption equipment. The plane is inclined at
an angle to the vertical, and we consider a section L of the film which is
sufficiently far removed from the entrance and exit that the velocity profile
is independent of variation in the z direction along the flow. The only non zero component of the velocity, uz , is a function of the coordinate x alone.

26

CHAPTER 3. UNIDIRECTIONAL TRANSPORT

z=0
vz (x)





































































































































































































































































z 


























































z+




z












g 
































































































































































































































































































































z=L

Figure 3.7: Flow down an inclined plane.

x
z

3.3. EFFECT OF PRESSURE ON MOMENTUM TRANSPORT

27

A momentum balance is set up for a thin shell of fluid of thickness x

bounded by the planes z = 0 and z = L. The various components of the
momentum balance in the z direction are
Rate of z
momentum in
across surface
at x

(LW ) xz |x

Rate of z
momentum out
across surface
at x + x

(LW ) xz |x+x

Rate of z
momentum in
across surface
at z = 0

(W xvz ) vz |z=0

Rate of z
momentum out
across surface
at z = L

(W xvz ) vz |z=L

Gravity force
acting on
the fluid

(W Lx)(g cos ()

(3.99)

When these terms are substituted into the momentum balance equation, we
get



LW xz |x + LW xz |x+x + W x vz2 

W x vz2 

+LW xW g cos () = 0
(3.100)
Dividing throughout by LW x and taking the limit as x 0, we get
lim

x0

z=0

xz |x+x xz |x
x

z=L

+ g cos () = 0

(3.101)

The first term on the left of the above equation is the derivative of the shear
stress, and so this can be written as
dxz
+ g cos () = 0
dx

(3.102)

28

CHAPTER 3. UNIDIRECTIONAL TRANSPORT

This can be integrated to give

xz = g cos ()x + C1

(3.103)

where C1 is a constant of integration. The constant of integration can be determined by making use of the boundary condition at the liquid gas interface
xz = 0 at x = 0

(3.104)

With this, the equation for the shear stress becomes

xz = g cos ()x

(3.105)

If the fluid is Newtonian, the momentum flux is related to the velocity

gradient by
dvz
xz =
(3.106)
dx
This is easily integrated to give an expression for the velocity
!

g cos () 2
x + C2
vz =
2

(3.107)

The constant of integration can be determined using the condition that at

x = h, vz = 0, since the surface at x = h is a solid surface and a no
slip boundary condition is applicable here. With this, the equation for the
velocity profile is
"
 2 #
x
gh2 cos ()
vz =
1
(3.108)
2
h
Hence the velocity profile is a parabolic profile.
Various parameters can be determined once this velocity profile is known.
1. the maximum velocity, vzm , is clearly at x = 0
vzm =

gh2 cos ()
2

(3.109)

2. The total flow rate is determined from

Q =
=

h
0

dx

dyvz

gW h3 cos ()
3

(3.110)

29

3.3. EFFECT OF PRESSURE ON MOMENTUM TRANSPORT

3. The mean velocity can be calculated from
Q
h
gW h2 cos ()
=
3

hvz i =

(3.111)

4. The film thickness can be expressed in terms of the flow rate as

3Q
gW cos ()

h=

!1/3

(3.112)

5. The total force on the inclined surface in the z direction is given by

F =

L
0

dz

W
0

dy xz |x=h

= ghLW cos ()

(3.113)
(3.114)

This is just equal to the weight of the fluid in the z direction under
steady flow conditions.
The momentum balance equation in the x direction can also be written
in an analogous fashion.
Rate of x
momentum in
across surface
at x
Rate of x
momentum out
across surface
at x + x
Gravity force
acting on
the fluid

(LW ) xx |x

(LW ) xx |x+x

(W Lx)(g sin ()

(3.115)

The momentum balance equation obtained from the above equation is

LW xx |x + LW xx |x+x + LW xW g sin () = 0

(3.116)

30

CHAPTER 3. UNIDIRECTIONAL TRANSPORT

x
R

Figure 3.8: Flow in a circular tube.

Since xx = p for a Newtonian fluid, the above equation reduces to
dp
= g sin ()
dx

(3.117)

The variation in the pressure in the film is then given by

p = gx sin ()

(3.118)

The above analytical results are valid only if the flow is laminar and
the streamlines are smooth, so that the flow can be considered steady. These
conditions are satisfied for the slow viscous flow of a thin film. As the velocity
increases or the film thickness increases, it has been found that there is a
transition from a laminar flow ith straight streamlines to a laminar flow with
rippling and then to a turbulent flow. The conditions under which these
transitions occur is determined by the Reynolds number, Re = 4hhvz i/.
A laminar flow without rippling is observed for Re < 10, while there is
rippling for 10 < Re < 1000. The flow becomes turbulent when the Reynolds
number increases beyond about 1000.
Example: Flow through a circular tube.
The flow through a circular tube is often encountered in engineering applications, and the laminar flow can be analysed using shell momentum balances. The only new feature here is the cylindrical coordinate system that is
used for the analysis, as shown in figure 3.8.
Consider a cylindrical shell of thickness r and length x. The balance

3.3. EFFECT OF PRESSURE ON MOMENTUM TRANSPORT

31

equation for the x component of the momentum can be written as,

Rate of change
= Sum of forces l
of x momentum

(3.119)

The rate of change of momentum in a time interval t within the differential

volume under consideration can be written as,
(ux )
Rate of change
(2rrz
=
of x momentum
t

(3.120)

where (ux ) is the change in the momentum per unit volume in the time
interval t.
There are four bounding surfaces for the differential volume under consideration, two or which are at perpendicular to the x axis and located at
x and x + x, and two of which are perpendicular to the radial co-ordinate
and are located at r and r + r. The forces acting on the surfaces at x and
x + x can be separated into two parts, the first due to the pressure acting
on the surfaces, and the second due to the flux of momentum due to fluid
motion. The forces due to fluid pressure can be written as,
Force due to pressure
on surface at x

= p(r, , x)(2rr)

Force due to pressure

on surface at x + x

= p(r, , x + x)(2rr)

(3.121)

Note that there is a negative sign for the force at x+x, because the pressure
always acts along the inward unit normal at the surface, and the inward unit
normal at x + x is in the negative x direction. There is an additional force
due to the flow of momentum into the differential volume through the surface
at x and the flow of momentum out of the differential volume through the
surface at x + x. This force is given by the product of the momentum flux
(momentum transported per unit area per unit time) and the surface area.
The momentum flux is the product of the momentum density (ux ) (per unit
volume) and the normal velocity to the surface ux . This is analogous to the
mass flux, which is the product of the concentration (mass density) c and
the normal velocity. Therefore, the force due to the flow of momentum is,
Force due to momentum flow
on surface at x

= ((ux )ux (2rr)

Force due to momentum flow

on surface at x + x

= p(r, , x + x)(2rr) (3.122)

32

CHAPTER 3. UNIDIRECTIONAL TRANSPORT

Note that the force at (x + x) is negative, since momentum leaves the

differential volume at this surface. We can now add up all the contributions
to get



2rLrz |r + 2rLrz |r+r + 2rrvz2 

2rrvz2 

+2rrLg+2rr(p0 pL ) = 0
(3.123)
Since the fluid is considered to be incompressible, vz is equal at z = 0 and
z = L. Therefore, the momentum flux due to fluid motion cancels out, and
we can divide by 2Lr and take the limit r 0 to get
z=0

z=L

P0 P L
1 d(rrz )
+
r dr
L


=0

(3.124)

where P = p gz. This equation can be integrated once to give

rz

C1
P0 P L
r+
=
2L
r


(3.125)

If the momentum flux is not infinite at r = 0, the constant C1 is zero and so

we get


P0 P L
r
(3.126)
rz =
2L
The velocity profile is calculated from the relation
rz =

dvz
dr

(3.127)

This is used to obtain

vz =

P0 P L 2
r + C2
4L


(3.128)

The constant C2 is determined from the condition that the fluid velocity is
zero at the wall, vz = 0 at r = R.
vz =

(P0 PL )R2
4L

r2
1 2
R

(3.129)

This is the familiar parabolic velocity profile for the flow in a tube. The
velocity is a maximum at the center, and decreases to zero at the walls. The
shear stress ia a maximum at the walls and decreases to zero at the center.

3.3. EFFECT OF PRESSURE ON MOMENTUM TRANSPORT

33

The volumetric flow rate is determined by integrating the velocity profile

over the radius of the tube.
Q=

(P0 PL )R4
8L

(3.130)

This shell balance analysis is also valid only for steady flows, where the
streamlines are straight. This occurs in a fully developed flow, away from
the entrance or exit of the pipe, and also in a the Reynolds number is less
than 2100, where the Reynolds number is defined as Re = vmax R/, where
vmax is the maximum velocity and R is the radius of the pipe.
Oscillatory flow in a pipe
An oscillatory pressure gradient (p/L) = k cos (t) is applied across a pipe
of length L. Determine the velocity profile in the pipe.
The differential equation for the velocity profile is, The momentum equation in the flow x direction is
1
ux

t
Re

2 ux 2 ux
+
x2
r 2

p
x

(3.131)

Setting ur = 0 for a unidirectional flow, and neglecting variations in the

streamwise direction, the equation for the velocity profile is
ux
1

t
Re

1 d dux
r
r dr dr

= cos (t)

(3.132)

where Re = (R2 /). Use the solution for an oscillatory flow,

ux = ux exp (t)

(3.133)

The equation for u

x becomes
1

ux
Re

ux
1 d d
r
r dr dr

=1

(3.134)

The solution for u

x is divided into a homogeneous and a particular solution.
The homogeneous solution is given by

(3.135)
u
xh = CJ0 ( Rer)

34

CHAPTER 3. UNIDIRECTIONAL TRANSPORT

x
Figure 3.9: Flow of immiscible fluids.
where J0 is the Bessel function. The particular solution is given by
u
xp =

(3.136)

The complete solution is

u
x =

J ( r)
0

J0 ( )

(3.137)

Adjacent flow of two immiscible liquids.

This is an example which illustrates the application of boundary conditions
between two immiscible liquids. Two immiscible liquids flow through a channel of length L and width W under the influence of a pressure gradient, as
shown in figure 3.9. The flow rates are adjusted such that the channel is half
filled with fluid I (more dense phase) and half filled with fluid II (less dense
phase). It is necessary to find the distribution or velocity in this case.
The momentum balance is similar to that for the flow in a pressure gradient shown in the last example. The pressure balance reduces to
dxz p0 pL
+
=0
(3.138)
dx
L
This equation is valid in either phase I or phase II. Integration gives two
relations in the two phases.
(I)
xz
(II)
xz

p0 p L
=
x + C (I)
L


p0 p L
x + C (II)
=
L


(3.139)

3.3. EFFECT OF PRESSURE ON MOMENTUM TRANSPORT

35

We can make use of one of the boundary conditions, that the stress is equal
at the interface, to relate C (I) and C (II) .
Atx = 0

(I)
(II)
xz
= xz
C (I) = C (II)

(3.140)

Using Newtons law of viscosity to relate the stress to the strain rate, we get

(I)
(I) dvz

p0 p L
x + C (I)
=
L


p0 p L
x + C (II)
=
L


dx
(II)
dv
(II) z
dx
This can be integrted to give
vz(I)

p0 p L
C (I) x
(I)
2
=
+ C2
x
+
(I)
(I)
2 L

vz(II)

C (I) x
p0 p L
(II)
2
x + (I) + C2
=
(II)
2 L

(3.141)

(3.142)
There are three constants in the above equations, which are determined using
the three available boundary conditions.
Atx = 0
Atx = b
Atx = b

vz(I) = vz(II)
vz(I) = 0
vz(II) = 0

(3.143)

These boundary conditions provide three relationships between the constants

(II)
(I)
C (I) , C2 and C2 .
(I)

C2

(II)

= C2

p0 pL 2 C (I) b
(I)
b (I) + C2
0 =
2(I) L

p0 pL 2 C (I) b
(I)
0 =
b + (I) + C2
(II)
2 L

These can be solved to obtain

(p0 pL )b
C1 =
2L
(I)
C2

(I) (II)
(I) + (II)

(3.144)

(p0 pL )b2
(I)
=
2
2(I) L
(I) + (II)

(3.145)

36

CHAPTER 3. UNIDIRECTIONAL TRANSPORT

Therefore, the velocity profiles in each layer are

vz(I)

(p0 pL )b2
=
2(I) L

"

2(I)
(I) + (II)

(I) (II)
+
(I) + (II)

x x2
2
b
b

vz(I)

(p0 pL )b2
=
2(II) L

"

2(II)
(I) + (II)

(I) (II)
+
(I) + (II)

x x2
2 (3.146)
b
b

#
#

Exercises
1. Consider a long and narrow channel two - dimensional of length L and
height H, where H  L. The ends of the channel are closed so that
no fluid can enter or leave the channel. The bottom and side walls of
the channel are stationary, while the top wall moves with a velocity
V (t). Since the length of the channel is large compared to the height,
the flow near the center can be considered as one dimensional. Near
the ends, there will be some circulation due to the presence of the side
walls, but this can be neglected far from the sides. For the flow far
from the walls of the channel,
(a) Write the equations for the unidirectional flow. What are the
boundary conditions? What restriction is placed on the velocity
profile due to the fact that the ends are closed and fluid cannot
enter or leave the channel?
(b) If the wall is given a steady velocity V which is independent of
time, solve the equations (neglecting the time derivative term).
Calculate the gradient of the pressure.
(c) If the wall is given an oscillating velocity V cos (t), obtain an
ordinary differential equation to obtain the velocity profile. Get
an analytical solution for this which involves the constants of itegration. Use the boundary conditions to determine all unknown
constants.
2. Wire-coating of dies Consider a cylinder in a thin annular region,
as shown in figure 3.10. The cylinder is pulled with a constant velocity
V. The pressure is equal on both sides of the cylinder. Determine the
fluid velocity, and the flow rate.
3. A fluid is contained in the annular region between two concentric cylinders of radius R1 and R2 moving with angular velocities 1 and 2 .

3.3. EFFECT OF PRESSURE ON MOMENTUM TRANSPORT

37

Outer radius R
Pressure p


























































































































Pressure p

Inner radius k R
Figure 3.10: Wire coating of dies.
The gravitational field acts along the axis of the cylinders as shown in
figure. The vessel is tall enough that the flow can be considered unidirectional when the distance from the bottom is large compared to the
gap width (R1 R2 ). In this case, choose a coordinate system and write
down the mass and momentum conservation equations. Solve these for
the pressure and velocity fields. Can you find the equation for the free
surface?
4. A resistance heating appratus for a fluid consists of a thin wire immersed in a fluid. In order to design the appratus, it is necessary to
determine the temperature in the fluid as a function of he heat flux
from the wire. For the purposes of the calculation, the wire can be
considered of infinite length so that the heat conduction problem is
effectively a two dimensional problem. In addition, the thickness of
the wire is considered small compared to any other length scales in the
problem, so that the wire is a line source of heat. The wire and the
fluid are initially at a temperature T0 . At time t = 0, the current is
switched on so that the wire acts as a source of heat, and the heat
transmitted per unit length of the wire is Q. The heat conduction in
the fluid is determined by the unsteady state heat conduction equation
t T = K2 T

(3.147)

and the heat flux (heat conducted per unit area) is

KT

(3.148)

38

CHAPTER 3. UNIDIRECTIONAL TRANSPORT

where K is the thermal conductivity of the fluid.
(a) Choose an appropriate coordinate system, and write down the
unsteady heat conduction equation.
(b) What are the boundary conditions? Give special attention o the
heat flux condition at the wire, and note that the wire is considered
to be of infinitesimal radius.
(c) Solve the heat conduction equation using the simplest method,
and determine the temperature field in the fluid.
(d) Use the boundary conditions to determine the constants in the
expression for the temperature field.
5. Consider a two dimensional incompressible flow in a diverging channel
with subtended angle bounded by solid walls as shown in the figure
below. The flux of fluid through the channel is Q. The two dimensional
(r, ) polar coordinate system used for the analysis is also shown in the
figure. Assume a one dimensional velocity field ur 6= 0 and u = 0. For
this case,
(a) From the mass conservation equation, determine the form of the
velocity in the radial direction.
(b) Write down the momentum conservation equations in the r and
directions and simplify assuming that the Reynolds number is
small, so that inertial terms can be neglected. Simplify the equations. What are the boundary conditions?
(c) Eliminate the pressure from the momentum equations to obtain
an equation for the velocity field.
(d) Solve the equation to obtain an expression for the velocity. How
many constants are there, and how are they determined? Do not
determine the constants.
(e) If we assume that the Reynolds number is high, so that potential flow conditions apply, what are the governing equations and
boundary conditions?
(f) What are the velocity and pressure fields in this case?

3.3. EFFECT OF PRESSURE ON MOMENTUM TRANSPORT

39

Wall
r

Wall

6. An ideal vortex is a flow with circular streamlines where the particle

motion is incompressible and irrotational. The velocity profile obeys
the equation in cylindrical coordinates:
v =

2r

(3.149)

with vr = vz = 0. At the origin, the above equation indicates that the

velocity becomes infinite. But this is prohibited because viscous forces
become important and the flow is rotational in a small region near the
core.
Consider an ideal vortex in which the velocity is given by the above
equation for t < 0, and the core velocity is constrained to be zero at
t = 0. Find the velocity profile for t > 0. Assume that v is the only
non - zero velocity component.
7. A cubic solid of side a is initially held at a temperature T0 . At times
t 0, its lateral faces are held at temperatures TA , TB , T0 and T0 as illustrated in figure 3.11. The top and bottom faces are insulated so that
no heat is transferred through them. The cube has heat conductivity
C, density and thermal conductivity K.
(a) Solve for the steady state temperature in the cube.
(b) Show how the transient problem may be set up in a form to which
separation of variables can be applied.

40

CHAPTER 3. UNIDIRECTIONAL TRANSPORT

q=0

TB

T0

T0

TA

q=0
Figure 3.11: Conduction from a cube.
8. A rotating cylinder geometry consists of a cylinder of radius R and
height H, filled with fluid, with two end caps. The cylinder rotates with
an angular velocity , while the end caps are stationary. Determine
the fluid velocity field using separation of variables as follows.
(a) Choose a coordinate system for the problem. Clearly, the only
non-zero component of the velocity is u . Determine the boundary
conditions for this component of the velocity.
(b) Write down the mass balance condition for an incompressible fluid.
For a uni-directional flow in which the density is a constant, what
does this reduce to?
(c) Use a shell balance to determine the conservation equation for the
velocity. Can you eliminate the pressure term using a result from
the mass balance condition?
(d) Solve the conservation equation at steady state using the method
of separation of variables. Frame the orthogonality conditions
which would be required to solve the problem.
Data:

3.3. EFFECT OF PRESSURE ON MOMENTUM TRANSPORT

41

(a) Bessel equation:

x2

d2 y
dy
+ x + (x2 n2 )y = 0
2
dx
dx

Solution:
y = A1 Jn (x) + A2 Yn (x)
where Jn (x) is bounded for x 0, and Yn (x) is bounded for
x . tem Modified Bessel equation:
x2

d2 y
dy
+ x (x2 + n2 )y = 0
2
dx
dx

Solution:
y = A1 In (x) + A2 Kn (x)
where In (x) is bounded for x 0, and Kn (x) is bounded for
x .