Problem Solutions in Transport Phenomena: Heat Transfer Problems

For theory relevant to the heat transfer problems below, please refer to the book:
Bird, R. B., Stewart, W. E., and Lightfoot, E. N., "Transport Phenomena", 2nd edition, John Wiley, New York
(2002).

The solutions below will also help you solve some of the problems in BSL (an abbreviation often used for
this classic textbook based on the initials of its authors).

Maximum current in an electric wire

Problem
An electric wire with radius r0 of 0.50 mm is made of copper [electrical conductivity = 5.1 x 107
ohm-1 m-1 and thermal conductivity = 380 W/(m K)]. It is insulated (see figure) to an outer radius
r1 of 1.50 mm with plastic [thermal conductivity = 0.350 W/(m K)].

Figure: Heating of an insulated electric wire.

The ambient air is at 38.0oC and the heat transfer coefficient from the outer insulated surface to
the surrounding air is 8.500 W/(m2 K). Determine the maximum current in amperes that can flow
at steady-state in the wire without any portion of the insulation getting heated above its
maximum allowable temperature of 93.0oC.

Solution
Step. Thermal resistance representation for insulation and air

In general, the heat flow is given by Q = T/Rth, where T is the temperature driving force
(thermal potential difference). The thermal resistance for a cylindrical annulus is Rth = ln (r1/r0)/
(2  kL) and the thermal resistance for a fluid film at a solid-fluid interface is Rth = 1/(hA). Here,

1
k is the thermal conductivity, h is the heat transfer coefficient and A is the surface area for
convection.

The thermal resistances for the insulation and air film are in series as shown in the figure below.

Figure: Thermal resistance representation of insulation and air film

Based on the above thermal resistance representation, the heat flow is

T0  T2 ln(r1 / r0 ) 1
Q where Rth   (1)
Rth 2kL h 2r1 L

where k is the thermal conductivity of the plastic insulation.

Step. Heat flow due to current in wire

The flow of an electric current results in some electrical energy getting converted to thermal
energy irreversibly. The heat generation by electrical dissipation per unit volume is given by S =
I 2/ke where I is the current density (in amp/m2) and ke is the electrical conductivity (in ohm-1 m-1).

The total heat generated within the wire is simply the product of S and the volume of the wire. At
steady-state, all this heat generated within the wire by electrical dissipation must leave through
the wire surface and therefore the heat flow is given by

I2
Q r02 L (2)
ke
Step. Expression for current

On eliminating Q from the above two equations, the current density is

1/ 2
 2k e (T0  T2 )  1 
I     (3)
 ln(r1 / r0 ) / k  1 /( hr1 )   r0 

On multiplying the current density by the cross-sectional area of the wire, the current is obtained
from
1/ 2
 2k e (T0  T2 ) 
Current  Ir0  
2
 r0 (4)
 ln(r1 / r0 ) / k  1 /(hr1 ) 

For the maximum current, the temperature T0 must be maximized.

Step. Substitution of numerical values

2
On setting the temperature T0 to 93.0oC (i.e., the maximum allowable temperature for the
insulation), the maximum current that can flow through the wire may be calculated as 13.027
amp.

The numerical values substituted in the equation are given below.

The values below may be changed and the problem solution recalculated with the new values
provided in consistent units.

Variable name Symbol Value Unit

electrical conductivity ke 51000000 ohm-1 m-1
o
maximum temperature T0 93.0000 C
o
ambient temperature T2 38.0000 C
outer radius r1 0.001500 m
wire radius r0 0.000500 m
plastic thermal conductivity k 0.35000 W/(m K)
heat transfer coefficient h 8.50000 W/(m2 K)
Calculated Variable Symbol Value Unit
Maximum current 13.027 amp

Recalculate

(Reset to original values)

Minimum thickness for a composite furnace wall

Problem
The wall of a furnace comprises three layers as shown in the figure. The first layer is refractory
(whose maximum allowable temperature is 1400oC) while the second layer is insulation (whose
maximum allowable temperature is 1093oC). The third layer is a plate of 6.35 mm thickness of
steel [thermal conductivity = 45 W/(m K)]. Assume the layers to be in very good thermal contact.

3
 Figure: Layers in a composite furnace wall.

The temperature T0 on the inside of the refractory is 1370 oC, while the temperature T3 on the
outside of the steel plate is 37.8oC. The heat loss through the furnace wall is expected to be
15800 W/m2. Determine the thickness of refractory and insulation that results in the minimum
total thickness of the wall.

Given thermal conductivities in W/(m K):

Layer k at 37.8oC k at 1093oC

Refractory 3.12 6.23
Insulation 1.56 3.12

Solution
Step. Thermal resistance representation for composite furnace wall

In general, the heat flow is given by Q = T/Rth and the thermal resistance for a rectangular slab
is Rth = x/(kA), where T is the temperature driving force (thermal potential difference), x is
the slab thickness, k is the thermal conductivity, and A is the cross-sectional area of the slab.

The thermal resistances for the three layers are in series as shown in the figure below.

Figure: Thermal resistance representation of composite furnace wall

Based on the thermal resistance representation for the composite furnace wall, the heat flux q is
Q T  T1 T  T2 T  T3
q  0  1  2
A x1  x0 x 2  x1 x3  x 2 (1)
k 01 k12 k 23
In the refractory and insulation, the thermal conductivity k varies with temperature. If a linear
variation is assumed, then the arithmetic mean is to be used for the thermal conductivity.

4
Step. Temperature at insulation - steel interface

The temperature T2 at the interface between the insulation and steel layers is given by

q T  T3 q
 2  T2  ( x3  x 2 )  T3 (2)
k 23 x3  x 2 k 23

The above expression can be used to calculate temperature T2 (as done later).
Step. Thickness of refractory and insulation

The thickness of the refractory (x1 - x0) and insulation (x2 - x1) are

T0  T1 T1  T2
x1  x 0  k 01 and x 2  x1  k12 (3)
q q

On adding the above two equations, we get

T0 T T
x 2  x 0  k 01  k12 2  ( k12  k 01 ) 1 (4)
q q q

Since q, T0 and T2 are known in the above expression, the first two terms on the right-hand side
are nearly fixed. The last term is negative as k12 is less than k01 (i.e., thermal conductivity of
insulation is less than that of refractory). Since the aim is to minimize (x2 - x0), the temperature T1
must be maximized.

Step. Substitution of numerical values

The temperature at the insulation - steel interface is

(5)

Next, the temperature T1 is set to 1093oC (i.e., the maximum allowable temperature for the
insulation). By linear extrapolation, the thermal conductivity of the refractory at 1370oC is 7.05
W/(m K). Substitution of numerical values gives the thickness of refractory as

(6)

and the thickness of insulation as

5
Heat conduction from a sphere to a stagnant fluid

Problem
A heated sphere of diameter D is placed in a large amount of stagnant fluid. Consider the heat
conduction in the fluid surrounding the sphere in the absence of convection. The thermal
conductivity k of the fluid may be considered constant. The temperature at the sphere surface is
TR and the temperature far away from the sphere is Ta.

Figure: Heated sphere in a large amount of stagnant fluid.

a) Establish an expression for the temperature T in the surrounding fluid as a function of r, the
distance from the center of the sphere.

b) If h is the heat transfer coefficient, then show that the Nusselt number (dimensionless heat
transfer coefficient) is given by

hD
Nu  2
k

Hint: Equate the heat flux at the sphere surface to the heat flux given by Newton's law of
cooling.

Solution
a) Step. Differential equation from heat balance

From a heat balance over a thin spherical shell in the surrounding fluid,

d
(r 2 q r )  Sr 2 (1)
dr

where S is the rate of generation of heat per unit volume. In this case, S = 0 in the fluid.

6
Since the thermal conductivity k for the fluid is constant, on substituting Fourier's law (
dT
qr   k ) we get
dr
d  2 dT 
r 0 (2)
dr  dr 

Step. Temperature profile by solving differential equation

On integrating,

dT C
r2  C1 or T   1  C 2 (3)
dr r

The integration constants are determined using the boundary conditions:

BC 1: r   , T = Ta or C2 = Ta (4)

BC 2: r = R, T = TR or -C1 = (TR – Ta)R (5)

where R is the radius of the sphere.

On substituting the integration constants, the temperature profile is

R T  Ta R
T  (TR  Ta )  Ta or  (6)
r TR  Ta r

b) Step. Nusselt number from heat flux

Using Fourier's law and differentiating the temperature profile, the heat flux is

dT R
qr   k  k (TR  Ta ) 2 (7)
dr r

Equating the heat flux at the sphere surface (r = R) to the heat flux as per Newton's law of
cooling, we get
T  Ta hR
k R  h(TR  Ta ) or 1 (8)
R k

The Nusselt number (which is the dimensionless heat transfer coefficient) is

hD
Nu  2 (9)
k
where D is the diameter of the sphere.

Note:

7
  This is a well-known result that is worth remembering. It provides the limiting value of
the Nusselt number for heat transfer from a sphere in the presence of convection at low
Reynolds and Grashof numbers.
 The Nusselt number Nu must not be confused with the Biot number Bi. Though the two
dimensionless groups are similar-looking, they differ as given below.

hD convection
Nu   (10)
k fluid conduction in fluid
hD convection
Bi  or Bi  (11)
k solid (int ernal ) conduction in solid

Maximum temperature in lubricant by viscous heating

Problem
An oil (of viscosity  and thermal conductivity k) acts as a lubricant between two coaxial
cylinders. The inner cylinder is stationary and the outer cylinder of radius R rotates at an angular
velocity . The clearance between the cylinders is b, which is small compared to the radii of the
cylinders; so, curvature effects can be neglected and the cylindrical system can be approximated
by a plane narrow slit (to be solved in Cartesian coordinates) as shown in the figure. Derive an
expression for the maximum temperature in the lubricant if both cylinders are at temperature T0.
Neglect the temperature dependence of  and k, but explicitly take into account the heat
generated by viscous dissipation.

Figure: Temperature profile for viscous heat generation. The rectangular section in the flow
between two coaxial cylinders can be approximated by the plane narrow slit on neglecting the
curvature of the bounding surfaces

Solution
Step. Velocity profile in narrow slit

8
The clearance between the cylinders is b, which is small compared to the radii of the cylinders;
so, curvature effects are neglected and the cylindrical system is approximated by a plane narrow
slit. From a shell momentum balance in Cartesian coordinates over a thin rectangular slab of
thickness x, the momentum flux (shear stress) distribution is obtained as

d xx P
  0   xz  K 1 (1)
dx L

The flow is solely due to the movement of the top surface and there is no pressure gradient
imposed on the system.

 dv 
On substituting Newton's law of viscosity  xz    z  , the expression for the velocity
 dx 
profile is

dv z K K1
 1  vz   x  K2 (2)
dx  

The integration constants K1 and K2 are determined using the boundary conditions:

BC: x = 0, vz = 0  K2 = 0 (3)

BC: x = b, vz = V  V = -(K1b)/ (4)

On substituting the integration constants, the velocity profile is given by

vz x
 (5)
V b

Step. Rate of generation of heat by viscous dissipation

During viscous flow, fluid layers rub against adjacent layers of fluid and the internal friction
produces heat. In other words, mechanical energy is degraded into thermal energy irreversibly.
The heat generation by viscous dissipation per unit volume depends on the local velocity
gradient and is given by
2 2
dv z  dv  V 
S    xz   z     (6)
dx  dx  b
For the system under consideration, S is found to be constant.

Step. Differential equation for heat flux from shell balance

Now,
Rate of Heat In - Out + Generation = Accumulation
At steady-state, the accumulation term is zero. From a thermal energy balance over a thin
rectangular slab of thickness x in the fluid, we get

9
  Aq x  x   Aq x  x  x
 S ( A x )  0 (7)

where A is the cross-sectional area of the narrow slit and qx is the heat flux for conduction in the
x-direction.

Dividing by A x and taking the limit as x tends to zero,

 q x  x  x   q x  x
lim x 0 S (8)
x
dq x
S (9)
dx

Step. Temperature profile by solving differential equation

On integrating, the heat flux is given by

qx = Sx + C1 (10)

 dT 
On substituting Fourier's law  q x   k  we get
 dx 

dT S C
 x 1 (11)
dx k k

On integrating, the temperature profile is given by

S 2 C1
T  x  x  C2 (12)
2k k

The integration constants C1 and C2 are determined using the boundary conditions:

BC 1: x = 0, T = T0  C2 = T0 (13)

BC 2: x = b, T = T0  -C1/k = Sb/2k (14)

On substituting the integration constants, the temperature profile is

Sb 2  x  x 
2

T  T0      (15)
2k  b  b  

Step. Maximum temperature in lubricant

The maximum temperature occurs when

10
 dT Sb 2  1 2 x 
  0  x b / 2 (16)
dx 2k  b b 2 

Thus, the maximum temperature occurs at the mid-plane of the slit and is obtained as

Sb 2 V 2  ( R ) 2
Tmax  T0   T0   T0  (17)
8k 8k 8k

Radial temperature distribution in annular chemical reactor

Problem
An annular chemical reactor consists of a packed bed of catalyst between two coaxial cylinders.
The inner and outer cylinders have radii of r0 and r1, respectively. It is reasonable to assume that
there is no heat transfer through the surface of the inner cylinder, which is at a constant
temperature T0. The catalytic reaction releases heat at a uniform volumetric rate S throughout the
reactor, whose effective thermal conductivity k may be considered constant. Neglect the
temperature gradients in the axial direction.

a) Derive a second-order differential equation to describe the radial temperature distribution in

the annular reactor starting with a shell thermal energy balance.

b) Establish the radial temperature distribution by solving the differential equation.

c) What viscous flow problem is analogous to this heat conduction problem?

d) Derive an expression for the volumetric average temperature in the reactor.

e) Develop an expression for the temperature at the outer cylindrical wall of the reactor. What
will be the outer wall temperature if both the inner and outer radii are tripled?

Solution
a) Step. Differential equation from thermal energy balance
From a thermal energy balance over a thin cylindrical shell of thickness r in the annular reactor,
we get

Rate of Heat: In - Out + Generation = Accumulation

At steady-state, the accumulation term will be zero. So,

 2rLq r  r   2rLq r  r  r
 S  2rrL   0 (1)

where S is the rate of generation of heat by chemical reaction per unit volume and qr is the heat
flux in the radial direction.

11
Dividing by 2 r L and taking the limit as r tends to zero,

 rq r  r  r   rq r  r
lim r 0  Sr (2)
r
d
( rq r )  Sr (3)
dr

Since the effective thermal conductivity k of the reactor bed may be considered constant, on
 dT 
substituting Fourier's law  q r   k  we get
 dr 

d dT
k (r )  Sr (4)
dr dr

b) Step. Radial temperature profile by solving differential equation

On integrating,

dT Sr 2 Sr 2
r   C1 or T   C1 ln r  C 2 (5)
dr 2k 4k

The integration constants are determined using the boundary conditions:

dT Sr02
BC 1 : r  r0 , q r   0 or C1  (6)
dr 2k

Sr02 Sr02
BC 2 : r  r0 , T  T0 or C 2  T0   ln r0 (7)
4k 2k
The first boundary condition suggests no heat transfer through the inner cylindrical wall of the
annulus.

On substituting the integration constants, the temperature profile is

Sr02   r  
2
 r
T  T0  1     2 ln  (8)
4k   r0   r0 


c) Step. Analogous problem in fluid mechanics

12
 Figure: Velocity profile in falling film on circular tube is analogous to temperature profile in
annular chemical reactor.

The velocity profile for the falling film on the outside of a circular tube (see Figure) is given by:

gR 2   r   r 
2

vz  1     2a ln  
2
(9)
4    R   R  

Substituting aR = r0 and R = r1,

gr12   r   r 
2 2
r 
vz  1     2 0  ln  (10)
4    r1   r1   r1 

The maximum velocity (which occurs at r = r0) is

gr12   r0   r 
2 2
r 
v z , max  1     2 0  ln  0   (11)
4    r1   r1   r1  


The difference between the above two equations yields

gr12  r 
2
r 
2
r 
2
 r 
v z  v z , max   0      2 0  ln 
4  r1   r1   r1   r0 
(12)
gr02   r  
2
 r
 1     2 ln 
4    r0   r0 


13
Equations (8) and (12) are identical in form. Thus, the analogous viscous flow problem is the
laminar flow of a falling film on the inside of a circular tube. The equivalent quantities are

T  T0  v z  v z , max , S  g , and k   (13)

d) Step. Expression for volumetric average temperature

The volumetric average temperature in the reactor may be defined as

r1 r1
2 r0 T rdr 2 r0 T rdr
T   (14)
r1
  r12  r02 
2 r0
rdr

On substituting the temperature profile in the above expression and integrating [using
 r ln r dr  (r / 2) ln r  (r / 4) ], we get
2 2

Sr02  2r12  r1  r12  r02 

T  T0   2 ln    (15)
4k r
 1  r0
2
 r0  2r02 

e) Step. Expression for outer wall temperature

The temperature at the outer cylindrical wall (r = r1) of the reactor is given by

Sr02   r1  
2
r
Touter wall  T0  1     2 ln 1   (16)
4k   r0   r0  


When both the inner and outer radii are n times their original values, the term in square brackets
gets multiplied by n2 and the outer wall temperature is thus given by
Sr02   r1  
2
r
Touter wall  T0  n 2
1     2 ln  1   (17)
4k   r0   r0  


For the case when both the radii are tripled, n = 3 in the above expression.

Heat transfer from a radial circular fin

Problem.
A pipe of radius R0 has a circular fin of radius R1 and thickness 2B on it (as shown in the figure
below). The outside wall temperature of the pipe is Tw and the ambient air temperature is Ta.
Neglect the heat loss from the edge of the fin (of thickness 2B). Assume heat is transferred to the
ambient air by surface convection with a constant heat transfer coefficient h.

14
 Figure: Radial circular fin on heated pipe.

a) Starting with a shell thermal energy balance, derive the differential equation that describes the
radial temperature distribution in the fin.

b) Obtain the radial temperature distribution in the circular fin.

c) Develop an expression for the total heat loss from the fin.

Solution
A circular fin is also called a radial fin or circumferential fin.

a) Step. Differential equation from thermal energy balance

From a thermal energy balance over a thin cylindrical ring of width r in the circular fin, we get

Rate of Heat: In - Out + Generation = Accumulation

The accumulation term (at steady-state) and the generation term will be zero. So,

 2r 2 B q r  r   2r  2 B q r r  r
 2 2r r  h  T  Ta   0 (1)

where h is the (constant) heat transfer coefficient for surface convection to the ambient air and qr
is the heat flux for conduction in the radial direction.

Dividing by 4 B r and taking the limit as r tends to zero,

 rq r  r  r   rq r  h
lim r 0 r
 r  T  Ta  (2)
r B
d
 rq r    h r  T  Ta  (3)
dr B

15
If the thermal conductivity k of the fin material is considered constant, on substituting Fourier's
 dT 
law  q r   k  we get
 dr 
d  dT  h
r  r  T  Ta  (4)
dr  dr  kB

Let the dimensionless excess temperature be denoted by  = (T - Ta)/(Tw - Ta). Then,

d  d  h
r  r (5)
dr  dr  kB

Using the chain rule on the left-hand side of the above equation and dividing throughout by r,

d 2 1 d h
2
   0 (6)
dr r dr kB

Alternatively, starting with the general expression for the cooling fin, we have

d  d 
 kA   hP (7)
dz  dz 

For the radial fin, the cross-sectional area (for conduction) is A = 2 r 2B and the perimeter (for
surface convection) is P = 4 r. Also, r = R0 + z, where z is the coordinate measuring distance
from the outside wall of the pipe. Substituting these expressions for A and P with dr = dz, we get

d  d 
k  2r 2 B   h 4r (8)
dr  dr 
The above equation on simplifying is identical to equation (5).

b) Step. Radial temperature profile by solving differential equation

Equation (6) is a modified Bessel equation of order zero. Its solution is

  C1 I 0 (cr )  C 2 K 0 (cr ) (9)

where c2 = h/(kB). Note that I0 and K0 are modified Bessel functions (of order zero) of first and
second kind, respectively.

The integration constants C1 and C2 are determined using the boundary conditions:

BC 1 : r  R0 ,   1  1  C1 I 0 (cR0 )  C 2 K 0 (cR0 ) (10)

d
BC 2 : r  R1 ,  0  0  C1 c I 1 (cR1 )  C 2 c K 1 (cR1 ) (11)
dr

16
The second boundary condition suggests no heat loss through the edge of the circular fin (of
thickness 2B), and requires the evaluation of derivatives of Bessel functions as given below:

d d
I 0 (cr )   I 1 (cr ) c  cI1 (cr ) and K 0 (cr )    K 1 (cr ) c   cK1 (cr )
dr dr
(12)

Equations (10) and (11) may be solved to yield C1 and C2. Thus,

K 1 (cR1 ) I 1 (cR1 )
C1  and C 2 
I 0 (cR0 ) K1 (cR1 )  I 1 (cR1 ) K 0 (cR0 ) I 0 (cR0 ) K1 (cR1 )  I 1 (cR1 ) K 0 (cR0 )
(13)

On substituting the integration constants, the dimensionless temperature profile is

T  Ta K 1 (cR1 ) I 0 (cr )  I 1 (cR1 ) K 0 (cr )

  (14)
Tw  Ta K 1 (cR1 ) I 0 (cR0 )  I 1 (cR1 ) K 0 (cR0 )

c) Step. Total heat loss from the fin

The heat flux at the base of the fin (r = R0) is given by

 dT  K (cR1 )cI1 (cR0 )  I 1 (cR1 )cK1 (cR0 )

qr r  R0
  k    k  Tw  Ta  1
 dr  r  R K 1 (cR1 ) I 0 (cR0 )  I 1 (cR1 ) K 0 (cR0 )
0
(15)

On multiplying the heat flux by the cross-sectional area for heat conduction, the total heat loss
from the fin is obtained as
K (cR0 ) I 1 (cR1 )  I 1 (cR0 ) K 1 (cR1 )
Total heat loss from fin  4R0 Bk  Tw  Ta  c 1
K 1 (cR1 ) I 0 (cR0 )  I 1 (cR1 ) K 0 (cR0 )
(16)

Heat conduction in a conical solid

Problem.
A solid is formed from the conical section of a sphere of radius R as shown in the figure. The
spherical surface at r = R is insulated, while the two conical surfaces at  = 1 and  = 2 are held

17
at temperatures T1 and T2, respectively. The thermal conductivity k of the solid material may be
assumed constant.

Figure: Heated conduction in a solid bounded by two conical surfaces and a spherical surface.

a) Establish an expression for the temperature T () in the solid object at steady state.

b) Find the total rate of heat flow across each of the conical surfaces.

Solution:
a) Step. Differential equation by simplifying the equation of energy

The problem involves pure heat conduction in a solid and therefore the equation of energy
simplifies to .q = 0 (where q is the heat flux). On substituting Fourier's law (q = k T), we
get 2T = 0 (because the thermal conductivity k is constant). Since the temperature in the conical
solid is a function of  only, the equation in spherical coordinates is

d  dT 
 sin  q   0 or d  sin  0 (1)
d d  d 
The above differential equation in dimensionless form is

d  d  T  T2
 sin    0 where   (2)
d  d  T1  T2

where  is the dimensionless temperature.

Step. Temperature profile by solving differential equation

The ordinary differential equation may be solved by integrating twice. Thus,

d C  1 
 1 or   C1 ln tan    C 2 (3)
d sin   2 

18
The integration constants C1 and C2 are determined using the boundary conditions:

 1 
BC 1 :    2 , T  T2 or C 2   C1 ln tan  2 
0  (4)
 2 
 1 
 tan 1 
BC 2 :   1 , T  T1 or  1  1  C1 ln 2  (5)
 tan 1  
 2 
 2 
On substituting the integration constants, the temperature profile is

 tan 1  
 
ln 2

1
 tan  
T  T2
   2 2 
(6)
T1  T2  tan 1  
 
ln 2 1

 tan 1  
 2 2 

b) Step. Total rate of heat flow across conical surface

Using Fourier's law and differentiating the temperature profile, the heat flux is

1 1
sec 2 
k dT k T1  T2 2 2 k T1  T2 1
q    
r d r  1  1 r  1  sin 
 tan 1  tan   tan 1  (7)
ln  2 2 2
 ln 
1
 tan   1
 tan  
 2 2   2 2 
The total rate of heat flow is obtained by integrating the product of the heat flux and the area.
Thus,

R 2k  T1  T2  R 2Rk  T1  T2 
Q  2 0 q sin  rdr 
 tan 1   0
dr 
 tan 1  
 2   
ln 2 2 2
 ln 
 tan 1    tan 1  
 2 1   2 1 
(8)
As expected, the total rate of heat flow across each of the conical surfaces is the same (since
there is no generation of heat within the solid).

Forced convection heat transfer for plug flow in circular tube

- constant wall heat flux in thermally fully developed flow region

Problem

19
A very thick slurry flows in a circular tube of radius R. Since it flows nearly as a solid plug, the
velocity profile is approximately flat over the pipe cross-section and vz = V (constant) may be
assumed. For z < 0, the fluid temperature is uniform at the inlet temperature T1. For z > 0, heat is
added at a uniform constant radial flux q0 through the tube wall. The axial heat conduction and
viscous dissipation effect may be neglected. The thermal conductivity k and thermal diffusivity 
may be assumed constant.

Figure: Plug flow in circular tube being heated by a uniform heat flux.

a) Show that the temperature profile T(r, z) far downstream in the thermally fully developed
region (i.e., for large z) is given by the following dimensionless asymptotic solution:

1 2 1
  ,    2   
2 4
where

T  T1 r z
 ,   , and  
q0 R / k R VR 2

b) Show that the limiting local Nusselt number far downstream for plug flow in a circular tube
with constant wall heat flux is Nu = 8.

Solution
Step. Differential equation and boundary conditions

For forced convection heat transfer, the equation of energy (on neglecting the viscous dissipation
effect) simplifies to

C p v  T  k 2T (1)

In cylindrical coordinates for flow in a circular tube, the above equation gives

T  1   T 
vz   r  (2)
z  2 r  r 

20
 k
 
where   C is the thermal diffusivity. Note that the heat conduction in the axial direction
p
is neglected because it is typically small compared to the heat convection in the axial direction.
The solution of the above partial differential equation gives the temperature profile T(r, z).

For plug flow, vz = V (constant) and the above equation in dimensionless form is

 1    
   (3)
     

where the dimensionless quantities are defined as

T  T1 r z
 ,   , and   (4)
q0 R / k R VR 2

The dimensionless quantities are typically chosen such that the number of parameters in the
problem is minimized. In this problem, the choice of dimensionless radial coordinate  and the
dimensionless axial coordinate  naturally follow from the differential equation, while the choice
of the dimensionless temperature  truly follows from the boundary conditions (BC 2 and BC 3)
given below.

The boundary conditions are given by

BC 1: r = 0, T is infinite   = 0,  is infinite (5)

BC 2: r = R, k (T/r) = q0   = 1, / = 1 (6)

BC 3: z = 0, T = T1   = 0,  = 0 (7)

The second boundary condition states that heat is added at a uniform constant radial flux through
the tube wall, i.e., qr = q0 at r = R with qr in accordance with Fourier's law.
Step. Temperature profile in thermally fully developed region

The asymptotic solution far downstream in the thermally fully developed region (i.e., large ) is
postulated to be of the form:

(,) = C0 + () (8)

where C0 is a constant to be determined. The above form suggests that the constant wall heat flux
will cause the fluid temperature to rise linearly with  once the fluid is far downstream from the
start of the heated section. Furthermore, the shape of the temperature profile T(r) will eventually
remain the same for increasing values of z.

On substituting the above form for  in the partial differential equation (3), the following
ordinary differential equation is obtained.

21
 1 d  d 
    C 0 (9)
 d  d 
On integrating twice, we get

d C 0 2 C
    C1 or   0  2  C1 ln   C 2 (10)
d 2 4

Thus, the dimensionless temperature profile is of the form:

C0 2
( ,  )  C 0    C1 ln   C 2 (11)
4

On imposing the boundary conditions, BC 1 gives C1 = 0 and BC 2 gives C0 = 2. However, it is

not possible to satisfy BC 3 because the postulated function is the asymptotic solution (and not
the complete solution); therefore, BC 3 is replaced by the following condition:

BC 3 : 2Rzq 0   C p  T  T1  2r v z dr      ,   d
R 1
 (12)
0 0

The above condition states the heat entering through the wall over the distance z equals the heat
leaving with the fluid at z (with T1 as the reference temperature at z = 0). On substituting the
temperature profile and integrating, the condition gives
1

1 2    4 C 2 2  1 C2 1
    2   C 2   d   2        C2  
0
 2   8 2 0 8 2 4
(13)

On substituting C0 = 2, C1 = 0 and C2 = 1/4, the asymptotic solution for the dimensionless

temperature profile in the thermally fully developed region is finally obtained as

1 2 1
  ,    2    (14)
2 4

Step. Calculation of bulk temperature

The bulk temperature is defined as the temperature obtained if the fluid flowing at z were
collected in a vessel and completely mixed. This average temperature is therefore also called the
flow-average temperature or the cup-mixing temperature. Thus, the bulk temperature is
mathematically defined by
R

Tb 
 T 2r v dr
0
z
(15)
R
 2r v dr
0
z

22
On substituting the temperature and velocity distributions, integration gives the bulk temperature
as
1
2R 2
V  T d  q0 R   2
1 
T1   2
     d
1
Tb  0
 2  k  2 4 

R 2V 0

1
(16)
 T1
 2  2

q0 R 
  2

 4

 2

  T1   2  q0 R
 
 2 k  8 8  0
k

Step. Expression for Nusselt number

The heat flux is the product of the heat transfer coefficient h and the local heat transfer driving
force (which is the difference between the wall temperature and the bulk temperature at an axial
distance z). Thus,

q0  h T  1
   1q R q R q R
 Tb  h T1   2   0  T1  (2 ) 0   h 0
 4 k k  4k
(17)


Thus the Nusselt number for plug flow in a circular tube far downstream for constant wall heat
flux is
hD
Nu  8 (18)
k

Note that the characteristic length used above is the tube diameter D (= 2R).

Forced convection heat transfer for plug flow in plane slit

- constant wall heat flux in thermally fully developed flow region

Problem
A very thick paste flows in a plane narrow slit formed by two parallel walls a distance 2B apart.
The length L and width W of the slit are such that B << W << L, so end effects can be neglected.
Since the paste flows nearly as a solid plug, the velocity profile is approximately flat over the slit
cross-section and vz = V (constant) may be assumed. For z < 0, the fluid temperature is uniform at
the inlet temperature T1. For z > 0, heat is added at a uniform constant flux q0 through both the
slit walls. The heat conduction in the z-direction and the viscous dissipation effect may be
neglected. The thermal conductivity k and thermal diffusivity  may be assumed constant.

Figure: Plug flow in plane narrow slit being heated by a uniform heat flux.

23
 a) Show that the temperature profile T(x, z) far downstream in the thermally fully developed
region (i.e., for large z) is given by the following dimensionless asymptotic solution:

1 2 1
  ,       
2 6
where

T  T1 x z
 ,   , and  
q0 B / k B VB 2

b) Show that the limiting local Nusselt number far downstream for plug flow in a plane slit with
constant wall heat flux is Nu = 12.

Solution
Step. Differential equation and boundary conditions

For forced convection heat transfer, the equation of energy (on neglecting the viscous dissipation
effect) simplifies to

C p v  T  k 2T (1)

In Cartesian coordinates for flow in a plane narrow slit, the above equation gives

T  2T
vz  (2)
z x 2

k
 
where   C is the thermal diffusivity. Note that the heat conduction in the z-direction is
p
neglected because it is typically small compared to the heat convection in the z-direction. The
solution of the above partial differential equation gives the temperature profile T(x, z).

For plug flow, vz = V (constant) and the above equation in dimensionless form is

  2 
 (3)
  2

where the dimensionless quantities are defined as

T  T1 x z
 ,   , and   (4)
q0 B / k B VB 2
The dimensionless quantities are typically chosen such that the number of parameters in the
problem is minimized. In this problem, the choice of dimensionless x-coordinate  and the

24
dimensionless z-coordinate  naturally follow from the differential equation, while the choice of
the dimensionless temperature  logically follows from the boundary conditions given below.

The boundary conditions are given by

T 
BC 1 : x   B,  k   q 0    1, 1 (5)
x 
T 
BC 2 : x   B,  k  q 0     1,  1 (6)
x 
BC 3 : z  0, T  T1    0,   0 (7)

The first two boundary conditions state that heat is added at a uniform constant flux through the
two walls, i.e., qx = q0 at x = +B and qx = q0 at x = B with qx in accordance with Fourier's law.
Step. Temperature profile in thermally fully developed region

The asymptotic solution far downstream in the thermally fully developed region (i.e., large ) is
postulated to be of the form:

  ,    C 0   ( ) (8)

where C0 is a constant to be determined. The above form suggests that the constant wall heat flux
will cause the fluid temperature to rise linearly with  once the fluid is far downstream from the
start of the heated section. Furthermore, the shape of the temperature profile T(x) will eventually
remain the same for increasing values of z.

On substituting the above form for  in the partial differential equation (3), the following
ordinary differential equation is obtained.

d 2
 C0 (9)
d 2

On integrating twice, we get

d C0 2
 C 0  C1 or    C1  C 2 (10)
d 2

Thus, the dimensionless temperature profile is of the form:

C0 2
  ,    C 0     C1  C 2 (11)
2
On imposing the boundary conditions, BC 1 gives C0 + C1 = 1 and BC 2 gives C0 + C1 = 1.
Thus, C0 = 1 and C1 = 0. However, it is not possible to satisfy BC 3 because the postulated
function is the asymptotic solution (and not the complete solution); therefore, BC 3 is replaced
by the following condition:

25
 C p  T  T1 W v z dx  2     ,   d
B 1
BC 3': 2Wzq0   (12)
B 1

The above condition states the heat entering through both the walls over the distance z equals the
heat leaving with the fluid at z (with T1 as the reference temperature at z = 0). On substituting the
temperature profile and integrating, the condition gives
1
1  2   3  1 1
2       C 2  d     C 2   2   2C 2  C 2   (13)
1
 2   6  1 3 6

On substituting C0 = 1, C1 = 0 and C2 = 1/6, the asymptotic solution for the dimensionless

temperature profile in the thermally fully developed region is finally obtained as

1 2 1
  ,        (14)
2 6

Step. Calculation of bulk temperature

The bulk temperature is defined as the temperature obtained if the fluid flowing at z were
collected in a vessel and completely mixed. This average temperature is therefore also called the
flow-average temperature or the cup-mixing temperature. Thus, on noting that the velocity and
temperature profiles are symmetric about the mid-plane of the slit, the bulk temperature is
mathematically defined by
B

Tb
 TW v
 0
z dx
B
(15)
0 W v z dx

On substituting the temperature and velocity distributions, integration gives the bulk temperature
as
1
WBV  T d 1
 q B  2 1 
Tb  0
  T1  o       d
WBV 0 
k  2 6 
1
 q B 3  q B
 T1  0        T1   0
 k 6 6 0 k
(16)

Step. Expression for Nusselt number

The heat flux is the product of the heat transfer coefficient h and the local heat transfer driving
force (which is the difference between the wall temperature and the bulk temperature at an axial
distance z). Thus,

26
 
q0  h T  1
   1 q B q B q B
 Tb  h T1      0  T1   0   h 0
3 k k  3k
(17)
 

Thus the Nusselt number for plug flow in a plane slit far downstream for constant wall heat flux
is

4hB
Nu   12 (18)
k
Note that the characteristic length used above for noncircular conduits is the equivalent diameter
= 4 (mean hydraulic radius), where the mean hydraulic radius Rh is the ratio of the flow cross-
sectional area to the wetted perimeter. For the plane slit, Rh = 2BW / (4B + 2W). Since B << W,
the hydraulic radius is approximately given by Rh = 2BW / (2W) = B. Thus, the equivalent
diameter for the plane slit is 4B.

Forced convection heat transfer for laminar Newtonian flow in plane

slit
- constant wall heat flux in thermally fully developed flow region

Problem
A Newtonian fluid is in fully developed laminar incompressible flow in a plane narrow slit
formed by two flat parallel walls a distance 2B apart. The length L and width W of the slit are
such that B << W << L, so end effects can be neglected. For z < 0, the fluid temperature is
uniform at the inlet temperature T1. For z > 0, heat is added at a uniform constant flux q0 through
both the slit walls. The heat conduction in the z-direction and the viscous dissipation effect may
be neglected. The thermal conductivity k and thermal diffusivity  may be assumed constant.

Figure: Laminar Newtonian flow in plane narrow slit being heated by a uniform heat flux.

a) Show that the temperature profile T(x, z) far downstream in the thermally fully developed
region (i.e., for large z) is given by the following dimensionless asymptotic solution:

3 3 1 39
  ,      2 4
2 4 8 280
where

27
 T  T1 x z
 ,   , and  
q0 B / k B v z , max B 2

and vz,max is the maximum velocity in the slit.

b) Determine the limiting local Nusselt number far downstream for laminar Newtonian flow in a
plane slit with constant wall heat flux.

Solution
Step. Differential equation and boundary conditions

For forced convection heat transfer, the equation of energy (on neglecting the viscous dissipation
effect) simplifies to

C p v  T  k 2T (1)

In Cartesian coordinates for flow in a plane narrow slit, the above equation gives

T  2T   x  2  T  2T
vz  or v z . max 1       (2)
z x 2  B 
  z x 2

k
 
where   C is the thermal diffusivity and vz,max is the maximum velocity in the slit. Note
p
that the velocity profile for laminar Newtonian flow in a narrow slit is parabolic. Also, the heat
conduction in the z-direction is neglected because it is typically small compared to the heat
convection in the z-direction.
The solution of the above partial differential equation gives the temperature profile T(x, z). The
above equation in dimensionless form is

1      
2
2
 (3)
  2

where the dimensionless quantities are defined as

T  T1 x z
 ,   , and   (4)
q0 B / k B v x , max B 2

The dimensionless quantities are typically chosen such that the number of parameters in the
problem is minimized. In this problem, the choice of dimensionless x-coordinate  and the
dimensionless z-coordinate  naturally follow from the differential equation, while the choice of
the dimensionless temperature  logically follows from the boundary conditions given below.
The boundary conditions are given by

28
 T 
BC 1 : x   B,  k   q0    1, 1 (5)
x 

T 
BC 2 : x   B,  k  q0     1,  1 (6)
x 

BC 3 : z  0, T  T1    0,   0 (7)

The first two boundary conditions state that heat is added at a uniform constant flux through the
two walls, i.e., qx = q0 at x = +B and qx = q0 at x = B with qx in accordance with Fourier's law.
Step. Temperature profile in thermally fully developed region

The asymptotic solution far downstream in the thermally fully developed region (i.e., large ) is
postulated to be of the form:

  ,    C 0       (8)

where C0 is a constant to be determined. The above form suggests that the constant wall heat flux
will cause the fluid temperature to rise linearly with  once the fluid is far downstream from the
start of the heated section. Furthermore, the shape of the temperature profile T(x) will eventually
remain the same for increasing values of z.
On substituting the above form for  in the partial differential equation (3), the following
ordinary differential equation is obtained.

d 2

 C0 1  2  (9)
d 2

On integrating twice, we get

d  3   2  4 
 C 0      C1 or   C 0     C1  C 2 (10)
d  3   2 12 

Thus, the dimensionless temperature profile is of the form:

 2  4 
  ,    C 0  C 0     C1  C 2 (11)
 2 12 

On imposing the boundary conditions, BC 1 gives (2/3)C0 + C1 = 1 and BC 2 gives (2/3)C0 + C1

= 1. Thus, C0 = 3/2 and C1 = 0. However, it is not possible to satisfy BC 3 because the
postulated function is the asymptotic solution (and not the complete solution); therefore, BC 3 is
replaced by the following condition:

2     ,   1    d
B 1
BC 3' : 2Wzq 0   C p  T  T1 W v z dx  2
(12)
B 1

29
The above condition states the heat entering through both the walls over the distance z equals the
heat leaving with the fluid at z (with T1 as the reference temperature at z = 0). On substituting the
temperature profile and integrating, the condition gives

1  3 4 
1 2
3
2       2 
4 8

 C 2  1   2 d  2 
13 4

 C2
70 3
 C2  
39
280
 
(13)

On substituting C0 = 3/2, C1 = 0 and C2 = 39/280, the asymptotic solution for the dimensionless
temperature profile in the thermally fully developed region is finally obtained as

3 3 1 39
  ,      2 4 (14)
2 4 8 280

Step. Calculation of bulk temperature

The bulk temperature is defined as the temperature obtained if the fluid flowing at z were
collected in a vessel and completely mixed. This average temperature is therefore also called the
flow-average temperature or the cup-mixing temperature. Thus, on noting that the velocity and
temperature profiles are symmetric about the mid-plane of the slit, the bulk temperature is
mathematically defined by
B

Tb
 TW v
 0
z dx
B
(15)
0 W v z dx
On substituting the temperature and velocity distributions, integration gives the bulk temperature
as
 T 1    d  3
1
2
 39 
T1  k  2   4   8   280  1    d
1 q0 B  3 3 2 1 4
 
0 2
Tb
 1    d 2
1
2 0
   
0

3 q B
 T1     0
2  k
(16)

Step. Expression for Nusselt number

The heat flux is the product of the heat transfer coefficient h and the local heat transfer driving
force (which is the difference between the wall temperature and the bulk temperature at an axial
distance z). Thus,
  3  q B
q 0  hT  1  Tb   h T1    
3 17  q 0 B 17q 0 B
  T1     0   h
 2 35  k 2  k  35k
(17)

Thus the Nusselt number for laminar Newtonian flow in a plane slit far downstream for constant
wall heat flux is
4hB 140
Nu   (18)
k 17

30
Note that the characteristic length used above for noncircular conduits is the equivalent diameter
= 4 (mean hydraulic radius), where the mean hydraulic radius Rh is the ratio of the flow cross-
sectional area to the wetted perimeter. For the plane slit, Rh = 2BW / (4B + 2W). Since B << W,
the hydraulic radius is approximately given by Rh = 2BW / (2W) = B. Thus, the equivalent
diameter for the plane slit is 4B.

31