ONE-DIMENSIONAL,

STEADY-STATE
CONDUCTION

1
Temperature gradient exists along only a single
coordinate direction, and heat transfer occurs
exclusively in that direction.
T=T(x,t)
Steady-state implies that at any given location, the
temperature does not change with time.
T ( x , t )
0
t

The two conditions above will make T=T(x). Initially

heat generation will not be considered.
2
3.1 The Plane Wall
3.1.1 Temperature Distribution
The two surfaces of the plane wall are at temperatures
Ts,1 and Ts,2, where due to the existing temperature
gradient conduction heat transfer will occur as
shown in fig-chp3\fig.3.1.pptx. At the surfaces
convective heat transfer will also occur.
The appropriate equation to be used for one-
dimensional steady-state conduction with no heat
generation will be
d  dT 
k 0
dx  dx 
3
For constant k, the general solution is
T(x) = C1x +C2
Two boundary conditions will be required to
determine the two constants.
T(0) = Ts,1, and T(L) = Ts,2
The above will give
Ts,2  Ts,1
C1  and C2  Ts,1
L
Substitution will give
x
T ( x )  Ts,1  (Ts, 2  Ts,1 )
L
4
The above shows T varying linearly with x.
Heat transfer rate or flux can be determined as:
dT kA
q x   kA  (Ts,1  Ts,2 )
dx L
qx k
qx 
''
 (Ts,1  Ts,2 )
A L

3.1.2 Thermal Resistance

Thermal Ohm’s law can be formed as
Ts,1  Ts, 2 L L
qx  Rt  ; Re 
( L / kA ) kA A
5
ρ = σ-1 = resistivity of material (Ω.m)
For convection heat transfer
q = hA(Ts - T∞)
Corresponding thermal resistance
Ts  T 1
R t ,conv  
q hA
Since qx is constant, the thermal circuit diagram will
fig-chp3\fig.3.1.pptx give

T ,1  Ts,1 Ts,1  Ts, 2 Ts, 2  T , 2

qx   
1 / h1A L / kA 1 / h2A
6
For an overall temperature difference, (T∞,1-T∞,2 )
i.e. thermal potential difference
T ,1  T , 2 1 L 1
qx  R tot   
R tot h1A kA h 2 A

Also for radiation

qrad = hrA(Ts – Tsur) and

Ts  Tsur 1
R t ,rad  
q rad hrA

7
3.1.3 The Composite Wall
Composite walls can be analyzed as series and parallel
thermal network. The schematic is seen in fig-
chp3\fig.3.2.pptx where the series thermal circuit is
also shown for both conduction and convection heat
transfer.
T ,1  T , 4
qx 
 Ri
T ,1  T , 4
qx 
1 LA LB LC 1
   
h1A k A A k BA k C A h 4 A
8
For constant qx
T ,1  Ts,1 Ts,1  T2 T2  T3 T3  Ts, 4 Ts,4  T , 4
qx     
(1 / h1A) ( LA / k A A) ( LB / k BA) ( LC / k C A) (1 / h 4 A)

An overall heat transfer coefficient U is defined as

qx = UA ΔToverall and Rtot = 1/UA

1 LA LB LC 1 1
R tot      
h1A k A A k B A k C A h 4 A UA

9
In general
Toverall 1
R tot   Ri  
q UA

Series-parallel arrangement can also be analyzed as

shown in fig-chp3\fig.3.3.pptx. Although the heat
flow is now multidimensional, it is often reasonable
to assume one-dimensional conditions. Two
equivalent thermal circuits are presented. In (a)
surfaces normal to the x direction are isothermal and
in (b) being symmetric to the x-axis, the surfaces
can be considered adiabatic.

10
3.1.4 Contact Resistance
At the interface of composite materials, due to surface
roughness, there is what is called thermal contact
resistance. This results in a temperature drop across
the interface as shown in fig-chp3\fig.3.4.pptx.
This thermal contact resistance is designated by Rt,c
and expressed by
TA  TB
R t ,c 
''

q 'x'
The contact resistance may be viewed as two parallel
resistances:
• Due to contact spots
• Due to gaps
11
For rough surfaces, the contact area is typically small
and the major contribution is from the gaps.
By increasing the joint pressure the contact area can be
increased thus decreasing the contribution from the
gap.
Example 3.1
To reduce the heat loss rate, a person wears special
sporting clothes (insulation, k=0.014 W/m.K)
against an environment of air or water at 10oC. The
emissivity of the cloth is 0.95. What thickness of
insulation is needed to reduce the heat loss rate to
100 W ( a typical metabolic heat generation rate) in
air and water? What are the resulting skin 12
temperatures? Take surface area as 1.8 m2.

13
Solution
The thermal circuit showing conduction through the
skin fat and insulation, and convection and radiation
from the insulation surface

14
 The total thermal resistance will be
Toverall (35  10)
R tot    0.25 K / W
q 100
From the thermal circuit diagram
1
Lsf Lins  1 1  1  Lsf Lins 1 
R tot           
ksf A k insA  1 / hA 1 / h r A  A  ksf k ins h  h r 

Air
To determine hr, iteration will be required. Assume of
To=291K (Tins) and use
hr   (To  Tsur )(T  T ) 2
o
2
sur 15
Substitution of the values will give
hr = 0.95 x 5.67 x 10-8 (291 +283)(2912+2832) = 5.09
To check on the heat transfer from the outer surface,
we will use
To  Tsur To  Tsur 291  283
q    102W
 1

1 

1
A(h  h r ) 1
1.8{2  5.09} 
1

 1 / hA 1 / h r A 

It can further be refined to give q = 100 W.

Using  Lsf 1 
L  k AR    h  h r 
ins ins tot
 k sf
 0.003 1 
 0.0141.8x 0.25   
 0.3 ( 2  5.09)  16
Lins = 0.0042 m = 4.2 mm
This will give a temperature of 34.6oC on the skin
which will be comfortable.
Water
The convection resistance is much lower than the
radiation resistance. So it will be assumed that all
the heat transfer from the skin will be by convection.
This will give
 Lsf 1   0.003 1 
Lins  k ins  AR tot     0.0141.8x 0.25   
 k sf h   0.3 200
 0.0061m  6.1 mm
17
This will give a skin temperature of 34.6oC which is
about the same as that due to the air stream the air.
Had it not been for the insulation the temperatures on
the skin surface would have been 18oC and almost
10oC for the air and water respectively. You can
imagine how cold you can feel in the case of water!
Example 3.2
The thermal conductivity of a D=14-nm carbon
nanotube is measured with an instrument that is
fabricated of a wafer of silicon nitride at a
temperature of T∞=300K. The 20-μm-long nanotube
rests on two 0.5-μm-thick, 10 μm x 10 μm square
islands that are separated by a distance
s=5 μm. A thin layer of platinum is used as an 18
electrical resistor on the heated island (at temperature
Th) to dissipate q=11.3 μW of electrical power. On
the sensing island, a similar layer of platinum is
used to determine its temperature Ts. The platinum’s
electrical resistance, R(Ts) = E/I, is found by
measuring the voltage drop and electrical current
across the platinum layer. The temperature of the
sensing island, Ts, is then determined from the
relationship of the platinum electrical resistance to
its temperature. Each island is suspended by two
Lsn=250-μm-long silicon nitride

19
beams that are wsn=3 μm wide and tsn=0.5 μm thick.
A platinum line of width wpt=1 μm and thickness
tpt=0.2 μm is deposited within each silicon nitride
beam to power the heated island. The entire
experiment is performed in a vacuum at a steady-
state temperature of Ts=308.4 K. Estimate the
thermal conductivity of the carbon nanotube.
kpt=71.6 W/m.K ksn=15.5 W/m.K

20
Figure for example 3.2 21
Solution
The thermal circuit diagram is shown below.

22
The cross-sectional areas of the materials in the
support beams are
Apt=wpt tpt = (1x10-6) x (0.2x10-6) =2 x 10-13m2
Asn=wsn tsn – Apt = (3x10-6) x (0.5x10-6)- 2 x 10-13=
= 1.3 x 10-12 m2
Cross-sectional area of the carbon nanotube is
Acn = πD2/4 = π(14x10-9)2/4=1.54 x 10-16 m2
The thermal resistance of each support is
1 1
 k pt A pt k sn Asn   71.6x 2x1013 12
15.5x1.3x10 
R t ,sup     6
 
 pt
L L sn   250 x10 250x106 
= 7.25 x 106 K/W
23
The combined heat loss through both sensing island
supports is
qs = 2 (Ts-T∞)/Rt,sup = 2 x (308.4 – 300) / (7.25 x
106)
= 2.32 x 10-6 W = 2.32 μW
qh = q – qs = 11.3 – 2.32 = 8.98 μW

And from
q h Th  T 1 8.98x10 6 x 7.25x10 6
 Th  T  q h R t ,sup  300 
2 R t ,sup 2 2

Th = 332.6 K 24
Using the circuit connecting Th and Ts
Th  Ts
qs 
s /(k cn Acn )
6 6
q ss 2.32x10 x5x10
k cn   16
 3113W / m.K
Acn (Th  Ts ) 1.54x10 x (332.6  308.4)

Approximate heat transfer due to radiation shows a

value of 0.047 μW which is negligible on a
comparative basis.

25
3.2 ALTERNATIVE CONDUCTION ANALYSIS
For steady-state, no heat generation and no heat loss
from the sides ( fig-chp3\fig.3.5.pptx) qx is constant
and independent of x while A(x),dT/dx, and k(T)
may vary.
T may be T(x,y), but the y-coordinate effect may be
neglected. This makes it a one dimensional
analysis.
Starting with Fourier’s law
dT dx
q x   kA or q x   k(T )dT
dx A( x )
x dx T
qx     k(T )dT
x o A( x ) To 26
For A(x) and k(T) known, the integration results in a
functional form of T(x).
3.3 RADIAL SYSTEMS
Cylindrical and spherical systems can be analyzed by
using the standard method and alternative method.
For the one dimensional case T=T(r)
3.3.1 The cylinder (standard method)
The appropriate equation for steady-state with no
generation is
1 d  dT  dT
 kr 0 or kr  cons tan t
r dr  dr  dr
27
Also
dT dT dT
q r   kA   k(2rL)  (2L)kr
dr dr dr
which shows qr to be constant (independent of r).
The cylindrical configuration is shown in fig-
chp3\fig.3.6.pptx
Integrating for constant k gives
T(r) = C1 ln r + C2
Two boundary conditions
T(r1)=Ts,1 and T(r2) = Ts,2

28
The final solution becomes
Ts ,1  Ts , 2 r
T (r )  ln  Ts,2
r1 r2
ln
r2
dT Ts ,1  Ts , 2 1 2Lk (Ts ,1  Ts , 2 )
qr   kA   k (2rL) 
dr r1 r r2
ln ln
This gives r2 r1
ln ( r2 / r1 )
Rt 
2Lk
The thermal circuit is also shown in the figure.
For composite cylinders shown in fig-
chp3\fig.3.7.pptx
29
Using the thermal circuits
T ,1  T , 4
qr 
1 ln( r2 / r1 ) ln( r3 / r2 ) ln( r4 / r3 ) 1
   
2r1Lh1 2k A L 2k B L 2k C L 2r4 Lh 4

Using the overall heat transfer coefficient U

T ,1  T , 4 1
qr   UA(T ,1  T , 4 ); U  ( ARtot )
Rtot
where U is defined based on arbitrary area A.
It could also be defined on the basis of other areas
such that
U1A1= U2A2 = U3A3 = U4A4 = (∑Rt)-1
30
If U is defined on A1=2πr1L, then
1
U1 
1 r1 r2 r1 r2 r1 r4 r1 1
 ln  ln  ln 
h1 k A r1 k B r1 k C r3 r4 h 4

Example 3.3
A thin-walled copper tube of radius ri is used to
transport a low-temperature refrigerant and is at a
temperature Ti that is less than that of the ambient
air at T∞ around the tube. Is there an optimum
thickness associated with application of insulation to
the tube? Use graph for justification. 31
Solution
The arrangement is shown in the figure below. Use
cellular glass for insulation k=0.055 W/m.K and h=
5 W/m2.K and ri=5 mm.

32
Thermal circuit for example 3.3
33
Using the thermal circuit per unit length of the tube
ln( r / ri ) 1 T  Ti
R '
  and q 
'

2k 2rh
tot '
R tot

The above equation shows that with increase in r the

conduction resistance increases while that of
convection decreases. This indicates that there may
be an optimum thickness of insulation that will
minimize the heat transfer. This can be achieved by
using
'
dR 1 1
 0 which gives
tot
 0
dr 2kr 2r h
2 34
The above gives
k
r
h
Checking the second derivative gives
2 '
dR 1
tot
 0 at r  k / h
dr 2
2k / h
3 2

which suggests that the resistance determined is the

minimum or gives maximum heat transfer.
However this benchmark (r=rc) which is the
indicator of maximum heat transfer, can still be used
to get information on the heat transfer characteristics
with increase of insulation thickness. 35
The figure shows a plot of R 'tot against insulation
thickness (r-ri).
rc=k/h=0.055/5=0.011m rc-ri=(0.011-0.005)=0.006m
The graph shows that insulation thicknesses below this
critical radius results in increased heat transfer with
increase in insulation thickness. This is desirable for
electrical wire insulation (good cooling effect).
For insulation thicknesses above the critical radius, heat
transfer decreases with increase in thickness of
insulation. (But make a note that the bare tube has the
same resistance up to a certain insulation thickness)
This is the usual setup in practical insulation. The
thickness of the insulation will be optimized with
respect to the cost of the insulation material and the
saved energy cost. 36
Thermal resistance vs. insulation thickness
37
3.3.2 The Sphere
To diversify our approach, here we will use the
alternative method.
Considering fig-chp3\fig.3.8.pptx, steady-state
conduction with no generation gives qr=constant.

dT 2 dT
q r   kA   k ( 4r )
dr dr
q r r2 dr Ts , 2

4 r1 r 2
  k (T )dT
Ts ,1

For constant k

38
 4k (Ts,1  Ts, 2 )
qr 
1 1

r1 r2

The thermal resistance will be

1 1 1
R th    
4k  r1 r2 

39
Example 3.4
A spherical, thin-walled metallic container is used to
store liquid nitrogen at 77 K. The container has a
diameter of 0.5 m and is covered with an evacuated,
reflective insulation composed of silica powder.
The insulation is 25 mm thick, and its outer surface
is exposed to ambient air at 300 K. The convection
coefficient is known to be 20 W/m2.K. The latent
heat of vaporization and the density of liquid
nitrogen are 2 x 105 J/kg and 804 kg/m3,
respectively. (1) What is the rate of heat transfer to
the liquid nitrogen? (b) What is the rate of liquid
boil-off? 40
Figure for example
41
Solution
(1) The thermal circuit is shown below

1 1 1 1
R t ,cond     R t ,conv 
4k  r1 r2  h(4r2 )
2

Hence
T , 2  T ,1
q
(1 / 4k )[(1 / r1 )  (1 / r2 )]  [1 / h(4r2 )]
2

300  77

 1  1 1  1
 4 (0.0017)  0.25  0.275   [20(4 )(0.275)2 ]
   42
Thermal circuit for example 3.4
43
which gives

223
q  13.06 W
17.02  0.05

(2)Energy balance on the control volume (liquid N2)

E in  E out  0 q  m
 h fg  0
and the boil  off m
 is
q 13.06 5
 
m  5
 6.53x10 kg / s  5.64kg / day
h fg 2x10

m
5.64
V   0.007m / day (  10.8% of container)
3

 804 44
3.4 CONDUCTION WITH THERMAL ENERGY
GENERATION
Internal energy generation could be from current,
nuclear, or exothermic chemical reaction where the
generation will be assumed to be uniform. Typical
for electric current is
E g  I 2 R and the energy generation per unit
volume will be
E g 2
I R
q ' ' '  
V V

45
3.4.1 The Plane Wall
The applicable equation for constant k is
2
dT q  '''

2
 0
dx k
And the surfaces are maintained at Ts,1 and Ts,2 as
shown in fig-chp3\fig.3.9.pptx
The general solution is
q 2
' ''
T   x  C1x  C2
2k
The boundary conditions to determine C1 and C2 are
T(-L)=Ts,1 and T(L)=Ts,2
46
Applying the boundary conditions will give
Ts, 2  Ts,1 q 2 Ts, 2  Ts,1
' ''
C1  and C2  L 
2L 2k 2
Substitution will give the temperature distribution as
q L   x   (Ts, 2  Ts,1 ) x (Ts, 2  Ts,1 )
''' 2  2

T( x )  1    
2k   L   2 L 2
Unlike the previous equations, heat transfer is no more
independent of x as the derivative dT/dx shows.
For symmetric boundary conditions Ts,1=Ts,2=Ts
q ''' L2  x 2 
T( x )  1  2   Ts
2k  L  47
Maximum temperature exists at the midplane (x=0)
with a magnitude of
q ''' L2 q ''' L2
T(0)  To   Ts   To  Ts
2k 2k
This will give the dimensionless temperature
distribution as
T( x )  Ts T( x )  To  x 
2 2
x
 1     
To  Ts  L Ts  To  L

The midplane temperature distribution has (dT/dx)=0

at x = 0. This acts like an adiabatic surface.
Implication is that same equation applies to plane 48
walls that are perfectly insulated on one side (x=0) and
maintained at a fixed temperature Ts on the other
side (x=L).
It is better to relate to the adjoining fluid, since it may
not be convenient to measure Ts.
Neglecting radiation, energy balance at the surface,
x=L
dT
 kA  hA (Ts  T )
dx
dT   '' ' 2
q L  2x  
k   k   L2    h(Ts  T )
dx  2k xL
49
This will give
q ''' L
Ts  T 
h
The same can be derived from
E  E ; for a unit area q ''' L  h(T  T )
g out s 

The temperature distribution, then, becomes

q L''' 2  x 2
 q ''' L
T( x )  1    T 
2k   L   h
 
50
Example 3.5
A plane wall is a composite of two materials, A and B.
The wall of material A has uniform heat generation
of 1.5 x 106 W/m3, kA = 75 W/m.K, and thickness LA
= 50 mm. The wall material B has no generation
with kB=150 W/m.K and thickness LB=20 mm. The
inner surface of material A is well insulated, while
the outer surface of material B is cooled by a water
stream with T∞ =30oC and h = 1000W/m2.K.
1. Sketch the temperature distribution that exists in
the composite under steady-state conditions.
2. Determine the temperature To of the insulated
surface and the temperature T2 of the cooled
surface. 51
Figure for the example
52
Solution
1. Features of temperature distribution
Parabolic in material A ; Zero slope at insulated
boundary; Linear in material B; Slope change =
kB/kA = 2 at interface; Large gradients near the
surface (water).

53
2. At the outer surface, x=LA +LB
q  h (T2 - T∞)
''

Also 
q '''
L A  q ''
Equating the two gives
q ''' LA 6
1.5x10 x 0.05
T2  T   30   105o C
h 1000
Since To is given in terms of T1, using the circuit
diagram shown below

''
q

54
Using
T1  T
q  ''
''
; T  T  ( R ''
 R ''
) q ''

R cond ,B  R conv
'' 1 cond , B conv , B

LB 1
R ''
cond , B  R ''
conv 
kB h
 0.02 1 
T1  30     x1 . 5x106
x 0 . 05  30  85  115o
C
 150 1000 
Substitution gives
1.5x106 x 0.05
To   115  25  115  140o C
2x 75

55
Thermal circuit for example 3.4
Temperature distribution for h = 200 W/m2.K
56
Thermal circuit for example 3.4
Temperatur distribution for h = 1000 W/m2.K
57
3.4.2 Radial Systems
For the cylinder shown in fig-chp3\fig.3.10.pptx , the
heat generated is convected from the surface of the
cylinder.
The appropriate equation for analysis is (constant k)
1 d  dT  q '''
r  0
r dr  dr  k
Integrating twice gives
q ''' 2
T (r )   r  C1 ln r  C2
4k
58
Boundary conditions are
dT (0)
0 and T ( ro )  Ts
dr
The first gives C1 = 0 and the second gives
q ''' 2
C2  Ts  ro This gives the temperature
4k distribution as
q ''' ro2  r 2 
T( r )   1  2   Ts
4k  ro 
Using the centerline temperature T(0) = To
2
T( r )  Ts r
 1   
To  Ts  ro  59
To relate Ts to T∞, use is made of the equality of heat
generated and heat convected from surface

q ''' (ro2 L)  h(2ro L)(Ts  T )

q ''' ro
Ts  T 
2h
One needs to make a note that the thermal resistance
approach has not been used as the heat transfer rate
with heat generation is not constant.

60
Example 3.6
Consider a long solid tube, insulated at the outer
radius r2 ( adiabatic surface with a prescribed
temperature Ts,2)and cooled at the inner radius r1,
with uniform heat generation within the solid.
1. Obtain the general and particular solution for the
temperature distribution in the tube.
2. Determine the heat removal rate per unit length of
tube.
3. If the coolant is available at a temperature T∞,
obtain an expression for the convection coefficient
that would have to be maintained at the inner
surface to allow for operation at prescribed 61
 T
values of s,2 and 
q '''
.

62
Solution
(1) The general solution is
q 2
' ''
T( r )   r  C1 ln r  C2
2k
dT ( r2 )
Boundary conditions 0
dr
T(r2) = Ts,2 and
These give
q ''' 2 q ''' 2 q ''' 2 r2
C1  r2 and C2  Ts, 2  r2  r2 ln
2k 4k 2k r
And the temperature distribution after substitution
q ''' 2 2 q ''' 2 r2
T( r )  Ts, 2  ( r2  r )  r2 ln 63
4k 2k r
(2) Heat removal rate is the conduction heat rate at r=r1
dT ( r1 )  q q r 
''' ' '' 2
q ( r1 )   k(2r1 )
'
r   k 2r1   r1   2
dr  2k 2k r1 
qr (r1 )  q  (r2  r1 )
' '' ' 2 2

qr  tot  q ''' (r22  r12 ) L

Alternatively, the same result could be found by equating
the heat generated with the removal at r1.

E g  q ''' (r22  r12 ) L  Eout

(3) At the inner surface

q '
cond q '
conv or
q ( r2  r1 )  h2r1 (Ts,1  T )
 ' '' 2 2
64
Hence
q ''' ( r22  r12 )
h
2r1 (Ts,1  T )
3.5 HEAT TRANSFER FROM EXTENDED
SURFACES
The characteristics of an extended surface is shown in
fig-chp3\fig.3.11.pptx . For T1>T2, heat transfer by
conduction occurs in the strut. For T1>T2>T∞, in
addition to heat conduction there will be convection
heat transfer from the surfaces of the strut thus
decreasing qx continuously and hence dT . When
compared with pure conduction the slope dx at the base
is increased thus increasing the conduction rate. 65
Thus extended surfaces are used to enhance the heat
transfer rate by increasing the surface area. The
term fin is used for extended surfaces.
Consider the heat transfer from surface given in fig-
chp3\fig.3.12.pptx (a). For fixed Ts, T∞ and A, heat
transfer can be maximized by increasing h, which
will require a higher velocity of fluid (need of
blower or pump)-becomes costly. The other option
is to increase the surface area for convection as
shown (b) where fins or extended surfaces are used.
• k has pronounced effect on the heat transfer effect
Infinite k-no temperature gradient, max q
fig-chp3\fig.3.13.pptx and fig-chp3\fig.3.14.pptx 66
Selection depends on space, weight, manufacturing
and cost reduction as well as the reduction of h and
increase of ∆P.
3.5.1 A General Conduction Analysis
Temperature distribution along the fin is to be
obtained. This will allow the heat transfer rate to be
determined from the surface and tip of fin shown in
fig-chp3\fig.3.15.pptx.
The heat transfer originates from the base and finally
transferred through the surface and tip of the fin.
Since the fins involved are very thin it can safely be
assumed the heat transfer is one-dimensional.
67
Energy balance on the differential CV gives
qx = qx+dx + dq conv
dT dqx
q x   kAc q x  dx  q x  dx
dx dx
dT d  dT 
q x  dx   kAc  k  Ac dx
dx dx  dx 
qconv  h dAs (T  T )
Substitution in the energy balance equation gives
d  dT  h dAs
 Ac  (T  T )  0
dx  dx  k dx 68
After performing the differentiation, the final general
fin equation becomes
d T  1 dAc  dT  1 h dAs 
2

2
     (T  T )  0
dx  Ac dx  dx  Ac k dx 

This equation will be used on specific fin geometries

to get the particular fin equations.
3.5.2 Fins of Uniform Cross-Sectional Area
The rectangular and the pin fin are shown in fig-
chp3\fig.3.16.pptx .
Ac=constant, dAc/dx =0 As=Px dAs/dx=P
69
Substitution in the general equation gives
d 2T hP
2
 (T  T )  0
dx kAc
The above equation can be further simplified if we
define θ(x) = T(x) – T∞
dθ/dx=dT/dx θ is also called excess temperature
This will give
d2
hp
2
m  0
2
where m 2

dx kAc
The above is a linear, homogeneous, second order
differential equation with constant coefficients
70
General solution
d
 ( x )  C1e  C2 e
mx  mx
 C1me  C2 me
mx  mx

dx
Boundary Conditions
(1)Common θ(0) = Tb – T∞ = θb
(2) There are four possible cases
I. Convection heat transfer from the tip (TL=T(L))
fig-chp3\fig.3.17.pptx
dT
hAc [T ( L )  T )  kAc or
dx @ x  L
d
h ( L )  k
dx @ x  L 71
Substitution will give
0
 b  C1  C2
h(C1e mL  C2 e  mL )  km(C2 e  mL  C1e mL )
Final solution ( details.docx )
 cosh m( L  x )  ( h / mk ) sinh m( L  x )

b cosh mL  ( h / mk ) sinh mL
Particular interest is the heat transfer rate from such
fins. (a) from base (b) from surface of fins
(a) is easier
dT (0) d (0)
q f  qb  kAc  kAc
dx dx 72
This will give
sinh mL  ( h / mk ) cosh mL
q f  hPkAc  b
cosh mL  ( h / mk ) sinh mL
II. Convection heat transfer from the tip is
negligible in which case the tip may be treated
as adiabatic.
(dT/dx)x=L=0 or (dθ/dx)x=L=0
This will give the second equation as
C1emL – C2e-mL =0 C1 +C2 = θb (common)
The above will give the solution as
 cosh m( L  x )

b cosh mL 73
The heat transfer in this case is determined as
q f  hPkAc  b tanh mL  M tanh mL; M  hPkAc  b
This is the most popular form
III. Temperature is prescribed at the tip,
θ(L)=θL=TL-T∞
Solution will be
 ( L /  b ) sinh mx  sinh m(l  x )
 and
b sinh mL
cosh mL  ( L /  b )
qf  M
sinh mL

74
IV.The fin is very long such that as
L→∞, T→T∞ or θL→0
The solution
  mx
e qf  M
b
table3.1.docx summarizes the solutions for different
BC.
Example 3.7
A very long rod 5 mm in diameter has one end
maintained at 100oC. The surface of the rod is
exposed to ambient air at 25oC with a convection
heat transfer coefficient of 100 W/m2.K. 75
1. Determine the temperature distributions along rods
constructed from Cu, Al and SS with k=398, 180,
14 W/m.K, respectively. What are the
corresponding heat losses from the rods?
2. Estimate how long the rods must be for the
assumption of infinite length to yield an accurate
estimate of the heat loss.

76
Solution
1. θ/θb=e-mx or T=T∞+(Tb – T∞)e-mx
where m=(hP/kAc)1/2 = (4h/kD)1/2= 14.2, 21.2,
75.6 m-1. T(x) shown below

77
qf = (hPkAc)1/2θb =(100π x 0.005 x 398 x (π/4) x
0.0052)1/2 (75)= 8.3W
For Al and SS qf = 5.6 W and 1.6 W respectively.
2. Since no heat transfer at the tip, comparing cases II
and IV one can approximate with
tanh mL  0.99 or mL  2.65
Hence a rod may be assumed to be infinitely long if
1/ 2
2.65  kAc 
L  L   2.65 
m  hP 
For copper
 398( / 4)(0.005) 2 
L     0.19 m
 100( x 0.005)  78
Corresponding values for Al and SS are 0.13 and 0.04
m respectively.
If the approximation is to accurately predict the
temperature distribution one may use
θ(L)/θb = exp(-mL)<0.01
This gives mL>4.6 in which case L∞ ≈ 0.33, 0.23, and
0.07 for the copper, aluminum, and stainless steel.
This is consistent with the graphical representation

79
3.5.3 Fin Performance
As the fin itself does have conduction resistance, there
may be no assurance that the heat transfer will be
increased by using fins. The evaluation uses fin
effectiveness, εf defined by
qf
εf 
hA c,bθb

where Ac,b is the fin cross-sectional area at the base.

Usage of fins may not be justified unless  f  2 .
As an example for case IV
1/2
hPkAc θb  kP 
εf    
hA c,bθb  hA c  80
since Ac = Ac,b for uniform cross-section. The above
relation gives sufficient information for the
important parameters
• εf increases with k-with respect to Al and Cu, Al
preferred due to low weight and low cost
• εf increases with P/Ac- Use of thin fins
recommended
• εf increases with small h-the need for fins is
stronger on the gas side than on the liquid side or the
natural convection side than the forced convection
side.

81
Fin performance may also be quantified in terms of
thermal resistance.
Thermal resistance due to convection at the exposed
base
1
R t,b 
hA c,b

The conduction/convection resistance of a fin is given

by
θb
R t,f  Performing division gives
qf
R t,b (1/hAc,b ) qf
   εf
R t,f θb /qf hA c,bθb
82
Need low Rt,f to increase εf. Also Rt,f < Rt,b ( a must)
A commonly used measure of fin performance is by
using fin efficiency, ηf. This compares qf with the
maximum possible heat transfer (when all the fin
surface is at base temperature
T = Tb or θ = θb = (Tb-T∞)
Based on this
qf qf
f  
q max hA f  b
For a straight fin of uniform cross-section and
adiabatic tip
83
 M tanh mL tanh mL
f = = ( M = hPkAc  b )
hPL b mL

ηf: 1 → 0 as L: 0 → ∞
For a fin with an active tip, the length can be slightly
increased (corrected) so that we have an adiabatic
tip, and use the simplified relation of case II. The
corrected lengths are given as
Lc = L + (t/2) - rectangular fin
Lc = L + (D/4) - pin fin
Errors associated with these approximations are
negligible if ht hD
or  0.0625
k 2k 84
For w >> t which is usually the case P ≈ 2w
And (P/Ac) = (2w/wt) = 2/t
1/ 2
 hP 
1/ 2
 2h 
mLc    Lc    Lc
 kAc   kt 
Multiplying numerator and denominator by (Lc)1/2 and
introducing a corrected fin profile area Ap=Lct, it
follows that
1/ 2
 2h 
mLc    3/ 2
Lc
 kA 
 p 1/ 2
 2h  3 / 2
ηf is plotted as a function of   Lc in
 kA 
 p
85
 fig-chp3\fig.3.18.pptx for rectangular, triangular, and
parabolic profiles.
3.5.4 Fins of Nonuniform Cross-Sectional Area
Consider the annular fin shown in fig-
chp3\fig.3.19.pptx . While the thickness is
constant, the cross-sectional area Ac varies with r.
Ac = 2πrt and As = 2π(r2 - r12)
Substituting in the general fin equation will give
2
d T 1 dT 2h
2
  (T  T )  0
dr r dr kT
With m2 = (2h/kt) and θ =T - T∞
86
The equation becomes
d  1 d
2
2 d 
2
d
2
 - m 2
  0 r 2
 r - m r  0
2 2

dr r dr dr dr
This is a modified Bessel equation of order zero whose
general solution is given by (Hyperbolic Functions
Error Function.docx
θ(r) = C1Io(mr) + C2Ko(mr)
Where Io and Ko are modified, zero order Bessel
functions of the first and second kind respectively.
B.C. used are  ( r )   and adiabatic tip
1 b

d ( r2 )
0
dr 87
will finally give the solution for the temperature
distribution as
 I o ( mr ) K1 ( mr2 )  K o ( mr ) I1 ( mr2 )

 b I o ( mr1 ) K1 ( mr2 )  K o ( mr1 ) I1 ( mr2 )
Where I1(mr) and K1(mr) are modified first order
Bessel functions of the first and second kinds.
The heat transfer by conduction from the base
dT ( r1 ) d ( r1 ) gives
q f   kA c ,b   k( 2r1t )
dr dr
K1 ( mr1 ) I1 ( mr2 )  I1 ( mr1 ) K1 ( mr2 )
q f  2r1t b m
K o ( mr1 ) I1 ( mr2 )  I o ( mr1 ) K1 ( mr2 )
88
The fin efficiency is given by
qf
f 
h[2 ( r2  r1 )] b
2 2

2 r1 K1 ( mr1 ) I1 ( mr2 )  I1 ( mr1 ) K1 ( mr2 )


m( r2  r1 ) K o ( mr1 ) I1 ( mr2 )  I o ( mr1 ) K1 ( mr2 )
2 2

For an active convection tip, use r2c = r2 + t/2

The fin efficiency is given graphically in fig-
chp3\fig.3.20.pptx .
For determination of fin effectiveness, εf , the
conduction/convection fin resistance is given by
Rt,f = 1/(hAfηf) 89
For equivalent heat transfer the triangular fin requires
much less volume than a rectangular profile.
Slightly larger (q/V) is parabolic. But when
considering the manufacturing cost triangular is the
more preferred one. The annular fin of rectangular
profile is commonly used to enhance heat transfer in
circular tubes.
Expressions for the efficiency and surface area of
several common fin geometries are summarized in
table3.2.docx

90
3.5.5 Overall Surface Efficiency
Unlike ηf which considers a single fin, the overall
surface efficiency, ηo considers the array of fins and
the base surface. fig-chp3\fig.3.21.pptx shows
typical arrays of fins.
In each case the overall efficiency is defined as
qt qt
o  
q max hA t b
qt = total heat transfer rate = qf + qb
At = total surface area = NAf + Ab

91
The total heat transfer rate from fins and bare surface
by convection is given by and then substitution of
Ab =At - NAf
q t = N η f h A f θb + h A b θb
 NA f 
q t  h[ Nf Af  ( A t  NA f )] b  hA t 1  (1  f ) b
 At 
Substitution in the expression for ηo gives
NA f
o  1  (1  f )
At

92
Using thermal fin resistance, the overall thermal
resistance can be expressed as

b 1
R t ,o  
q t o hA t
For fins machined as an integral part of the wall, the total
resistance can be seen to be one of parallel circuit
shown in fig-chp3\fig.3.22(a).pptx . This circuit will
give the same overall efficiency as determined earlier.
But if the fins are manufactured separately and are
attached (metallurgical, adhesive joint) or press fit to
the wall, ( fig-chp3\fig.3.22(b).pptx ) there will be a
thermal resistance, Rt,c, to be considered.

93
Fig.3.22(b) Fin array and thermal circuit-fins
attached to base
94
The effective circuit resistance will be
qt qt b 1
o( c )      R t.o( c )
q max hA t b q t o( c ) hA t
Using the parallel/series circuit shown the effective
overall efficiency can be determined as
NA f  f 
o ( c )  1 1  
At  C1 
where
C1  1  f hA f ( R 't',c / Ac,b )
While manufacturing make sure Rtc<<Rtf

95
Example 3.8
The engine cylinder of a motor cycle is constructed of
2024-T6 aluminum alloy and is of height H = 0.15
m and outside diameter D = 50 mm. Under typical
operating conditions the outer surface of the cylinder
is at a temperature of 500 K and is exposed to
ambient air at 300 K, with a convection coefficient
of 50 W/m2.K. Annular fins are integrally cast with
the cylinder to increase heat transfer to the
surroundings. Consider five such fins, which are of
thickness t = 6 mm. length L = 20 mm, and equally
spaced. What is the increase in heat transfer due to
use of the fins? 96
Solution

Figure for example 3.8

97
  NA f 
qt  hAt 1  (1   f ) b
 At 
A f  2 (r2 c  r1 )  2 (0.048  0.025 )  0.0105m
2 2 2 2 2

At  NA f  2r1 ( H  Nt )
 5 x0.0105  2 (0.025)[0.15  5 x0.006]  0.0716m 2
r2 c
 1.92, Lc  0.023m, Ap  1.380x10 4 m 2 gives
r1
 h  3/ 2
  Lc  0.15
 kA 
 p

From Fig.3.19, ηf ≈ 0.95

98
With the fins the total heat transfer rate becomes
 0.0527 
q t  50x 0.07161  (1  0.95) ( 200)  690 W
 0.0716 
Without fins
qwo = h(2πr1H)θb = 50(2π x 0.025 x 0.15)(200)
=236W
Increase
Δq = qt – qwo = 454 W (nearly 300% increase)
The increase could even be much higher if number of
fins are increased for the same fin thickness or fin
thicknesses are decreased to accommodate more
number of fins in a given space. (see figures) 99
t = 6 mm No of fins are increased
100
In the given H for a fin thickness of 2mm, 25 fins
could be accommodated
101
Example 3.9
In the example on PEM on chapter 1, we saw that to
generate an electrical power P = 9 W, the
temperature of the PEM fuel cell had to be
maintained at Tc≈56.4oC, which required total
removal of 11.25 W from the fuel cell and a cooling
air velocity of V = 9.4 m/s for T∞=25oC. To provide
these convective conditions, the fuel cell is centered
in a 50 mm x 26 mm rectangular duct, with 10 mm
gaps between the exterior of the 50 mm x 50 mm x 6
mm fuel cell and the top and bottom of the well-
insulated duct wall. A small fan, powered by the
fuel cell, is used to circulate the cooling air.
Inspection of a particular fan vendor’s data
102
Figure for example 3.9
103
 sheets suggest that the ratio of the fan power
consumption to the fan’s volumetric flow rate is
 = C =1000 W/(m3/s) for the range
Pf/ V f
104  V   102 m3 / s .
f
1. Determine the net electric power produced by the
fuel-cell-fan system, Pnet = P-Pf.
2. Consider the effect of attaching an aluminum
(k=200 W/m.K) finned heat sink, of identical top
and bottom sections, onto the fuel cell body. The
contact joint has a thermal resistance of
R 't',c  103 m2 .K / W, and the base of the heat sink is
of thickness tb = 2 mm. Each of the N rectangular
fins is of length Lf = 8 mm and thickness
104
of 1 mm, and spans the entire length of the fuel cell,
Lc = 50 mm. With the heat sink in place, the radiation
losses are negligible and the convective heat transfer
coefficient may be related to the size and geometry
of a typical air channel by an expression of the form
h = 1.78 kair (Lf + a )/(Lf . a)
where a is the distance between fins. Draw an
equivalent thermal circuit for part 2 and determine
the total number of fins needed to reduce the fan
power consumption to half the value found in part 1.

105
Solution
1. Volumetric flow rate air for cooling
V  VA and A  W( H  t )
c c c
Substitution gives
  V[W( H  t )  9.4[0.05(0.026  0.006)]
V c
3
 9.4x10 m / s 3

This enables us to use the formula for power

consumption by the fan
  1000x 9.4 x103  9.4 W
Pfan  CV
Pnet  Pfc  Pfan  9  9.4  0.4 W
106
As the fan consumes more power than is generated,
the system cannot produce net power.
2. To reduce the fan power by 50%, the volumetric
flow rate of air must be reduced to V   4.7x103 m3 / s
If we consider half of the fins the thermal
resistances can be written as follows:
Contact resistance of the fin base:-
R t ,c  R 't',c / Lc Wc  103 /(0.05x0.05)  0.4K / W
Fin base resistance:-
R t ,base  t b / kLcWc  0.002 /(200x0.05x0.05)  0.004 K / W

107
108
Thermal circuit for example 3.9
109
Bare surface of the base:-
R t ,b  1 /[h(Wc  Nt f )Lc  1 /[h(0.05  Nx0.001)0.05]
The above cannot be evaluated until N and h are
determined
Fin resistance:-
For a single fin (insulated tip)
q f  hPkA c  b tanh mL R t ,f   b / q f  1 / hPkA c tanh mLf
P = 2(Lc +tf) =2(0.05+0.001)=0.102 m
Ac = Lctf = 0.05x0.001=0.00005 m2
m  hP / kA c  hx 0.102 /(200x 0.00005)

110
Substitution will give
( hx 0.102x 200x 0.000005) 1 / 2
R t ,f 
tanh( mx0.008)
And for N fins, qf(N) =Nqf this will give
Rt,f(N) = θb/Nqf = Rt,f/N
a = (Wc – Ntf)/N = (0.05 – Nx0.001)/N
1 1 1
R tot  R t ,c  R t ,base  R equiv and R equiv  [ R t ,b R t ,f ( N )]
Requiv can be determined from
Tc  T Tc  T
q 
R tot R t ,c  R t ,base  R equiv
111
In which case
Tc  T
R equiv   ( R t ,c  R t ,base )
q
=(56.4-25)/5.625-(0.4+0.004)=5.2 K/W
The solution requires iteration which can be initiated
by assuming N. Let N=11
Then this will give a=0.0035 m, h=19.1 W/m2.K,
m=13.9 m-1, Rt,f(N)=5.88 K/W, Rt,b=26.8 K/W
This will give Requiv = 4.82 K/W and Rtot= 5.224
resulting in a fuel cell temperature of 54.4oC.
N=10 and N=12 give fuel cell temperatures of 58.9oC
and 50.7oC which are far from the given fuel cell 112
temperature.
So the total number of fins required will be 22.

Pnet = P – Pf = 9.0 - 4.7 = 4.3 W

In actual cases h is not sensitive to the velocity of the

air when the flow is confined to passages. This will
be seen in detail in convection heat transfer topic.

113
3.6 THE BIOHEAT EQUATION
As blood flows through arteries and veins, there is a
heat exchange between the two. Also the flow of
blood through tissues results in heat transfer process
between the capillary and the surrounding tissue.
This process is what is called perfusion. What
makes it more complicated is the additional heat
generation due to metabolic effect given by q m. The
perfusion term given by q p is an energy source or
sink depending on whether heat transfer is from or
to the blood, respectively.
Assuming a one-dimensional heat conduction, with
114
modification, it can be written as
d T qm  qp
 
2 ' '' '' '

2
 0
dx k
Pennes assumed that within any small volume of
tissue, the blood flowing in the small capillaries
enters at an arterial temperature, Ta, and exits at the
local tissue temperature, T. The rate at which heat is
gained by the tissue is the rate at which heat is lost
from the blood. If the perfusion rate is ω (m3/s of
volumetric blood flow per m3 of tissue), the heat lost
from the blood can be calculated as
q   bcb Ta  T 
'''
p 115
Substitution in the equation gives
d T q   b cb (Ta  T )
2 '''

2
 m
0
dx k
Defining   T  Ta  qm / bcb
 '''

Then the above equation can be written as

d 2T ~ 2 ~ 2   c / k
2
 m   0 , m b b
dx
This is the same as the fin equation for uniform cross
section. Using the appropriate boundary conditions
the temperature of the tissue can be determined.

116
Example 3.10
Consider a region of muscle with a skin/fat layer over
it. At a depth Lm = 30 mm into the muscle, the
temperature can be assumed to be at the core body
temperature of Tc = 37oC. The muscle thermal
conductivity is km = 0.5 W/m.K. The metabolic heat
generation rate within the muscle is
q 'm''  700 W / m3 (t). The perfusion rate is
ω = 0.0005m3(b)/s. m3(t); ρb = 1000 kg/m3 and cb =
3600 J/kg.K, and the arterial temperature, Ta, is the
same as the core body temperature. The thickness,
emissivity, and thermal conductivity of the skin/fat
layer as given in the previous examples; perfusion
117
Solution

118
and metabolic heat generation within this layer can be
neglected. We wish to predict the heat loss rate
from the body and the temperature at the inner
surface of the skin/fat layer for air and water
environments of T∞=297 K.
Solution
q=(Ti – T∞)/Rtot
1
Lsf  1 1  1  Lsf 1 
R tot         
ksf A  1 / hA 1 / h r A  A  ksf h  h r 
Using the previous example values for air will give
1  0.003 1 
R tot      0.076 K / W
1 .8  0 .3 2  5 .9  119
For water, with hr = 0 and h = 200 W/m2.K
Rtot = 0.0083 K/W
Boundary conditions for the governing equation are
give in terms of Tc and Ti (unknown). In terms of θ,
the boundary conditions are
q 'm'' q 'm''
 (0)  Tc  Ta   c and  ( L m )  Ti  Ta   i
 bc b  bc b
Using the solution of the fin equation for known tip
temperature (case III), the solution becomes
~ ~
 ( i /  c ) sinh mx  sinh m( L m  x )
 ~L
c sinh m m
120
Our interest is to determine the heat transfer to the
skin/fat which is at x=Lm or
dT d
q @ x  Lm   k m A  kmA
dx @ x  Lm dx @ x  Lm
( /  ) cosh ~L 1 T  T
m
~
  k m Amc i c m
 i 
~
sinh mL m R tot

The above equation can be solved for Ti (since Ti also

appears in θi) as
~ ~   
q ' ''
 ~ 
T sinh mL m  k m AmR tot  c   Ta  m
 cosh mL m 
   bc b  
Ti  ~ L  k Am ~ R cosh m ~L
sinh m m m tot m 121
where
~
m   b cb / k  0.0005x1000x3600 / 0.5  60m 1

~ L )  sinh(60x0.03)  2.94
sinh( m m

cosh(m~ L )  cosh(60x0.03)  3.11

m

qm''' qm''' 700

 c  Tc  Ta   
 b cb  b cb 0.0005x1000x3600
 0.389K
Substitution will give
24x 2.94  0.5x1.8x 60x 0.076[0.389  (37  0.389)x 3.11]
Ti 
2.94  0.5x1.8x 60x 0.076x 3.11
 34.8o C
122
For the previous example the assumed value was 35oC
which agrees well with the result.
The heat loss rate can be determined as
Ti  T 34.8  24
q   142W
R tot 0.076

Ts=Ti – q(Lsf/ksfA)=34.8-142(0.003/0.3x1.8)=34oC
Repeating the calculation for water, we find
Ti = 28.2oC q = 514 W
Ts = Ti – q(Lsf/ksfA)=28.2 -514(0.003/0.3x1.8)=25.3oC

123