BMTECHRE Unlocked

TECHNICAL UNIVERSITY OF MOMBASA
SCHOOL OF ENGINEERING AND TECHNOLOGY
DEPARTMENT OF MECHANICAL AND AUTOMOTIVE ENGINEERING
Heat Transfer and Heat Exchanger TMC 4404/TMC 4415 CLASS: BTME/A Y4S1
LECTURE NOTES
Prepared by: Njeru Gatumu

UNIT: Heat Transfer and Heat Exchanger TMC
4404
Learning Outcomes for the Unit:
At the end of this course the student will be able to:

1. Explain the relative significance of heat transfer by various modes.
2. Describe heat exchangers designed, fabrication and use in industry.
3. Apply the laws of thermodynamics to Heat transfer processes;
4. Apply the laws of heat flow processes in Heat exachangers;
5. Make basic calculations of heat transfer in engineering systems.;
1
Chapter 1
Heat Transfer by conduction and
convection
Learning outcomes:
After completing solving of problems, and reading explanations and examples in this chapter,
the student should be able to:
1. Explain the terminologies used in heat transfer,
2. Describe heat transfer processes of conduction, and convection and radiation,
3. Derive the expressions heat transfer through common materials and shapes,
4. Apply derived equations of heat transfer to solve problems,
5. Relate given heat transfer processes to their practical applications in thermodynamics sys-
tems.
1.1 Introduction
The transfer of heat across the boundaries of a system, either to or from the system, has been
considered for non-flow and flow processes; the definition of heat as simply stated that; heat is
a form of energy which is transferred from one body to another body at a lower temperature by
virtue of the temperature difference between the bodies. When the mechanism of the transfer
of heat is considered a slightly different approach is necessary compared with the approach of
fundamental thermodynamics. For instance it becomes difficult to define a system. In order to
illustrate this point consider a bar of metal being heated at one end and cooled at the other. Now
a boundary may be put round the source of heat or round the sink for the rejection of heat, but
a boundary encircling the metal bar encloses a body the temperature of which varies throughout
its length. In order to apply the laws of thermodynamics to the system consisting of the metal
bar, a mean temperature must be assumed.
In solids, the heat is conducted by the following two mechanisms:
• By lattice vibration (The faster moving molecules or atoms in the hottest part of a body
transfer heat by impacts some of their energy to adjacent molecules).
• By transport of free electrons (Free electrons provide an energy flux in the direction of
decreasing temperature. For metals, especially good electrical conductors, the electronic
mechanism is responsible for the major portion of the heat flux except at low temperature).
2
In case of gases, the mechanism of heat conduction is simple. The kinetic energy of a molecule
is a function of temperature. These molecules are in a continuous random motion exchanging
energy and momentum. When a molecule from the high temperature region collides with a
molecule from the low temperature region, it loses energy by collisions.
In liquids, the mechanism of heat is nearer to that of gases. However, the molecules are more
closely spaced and intermolecular forces come into play.
In previous studies many problems have been considered in which a certain quantity of heat has
been transferred from one system to another. In this chapter we shall be concerned with the rate
at which heat is transferred. The rate of heat transfer may be constant or variable, depending on
whether conditions are such that the temperatures remain the same or change continually with
time. Most problems in practice are concerned with steady-state heat transfer, in which heat
flows continuously at a uniform rate, but there are many cases of transient heat transfer and some
of these will also be considered.
In general there are three ways in which heat may be transferred: Conduction, Convection and
Radiation.
1.1.1 Conduction
Conduction is the transfer of heat from one part of a substance to another part of the same
substance, or from one substance to another in physical contact with it, without appreciable
displacement of the molecules forming the substance. For example, the heat transfer in the metal
bar mentioned previously is by conduction.
1.1.2 Convection
Convection is the transfer of heat within a fluid by the mixing of one portion of the fluid with
another. The movement of the fluid may be caused by differences in density resulting from
the temperature differences as in natural convection (or free convection), or the motion may be
produced by mechanical means, as in forced convection. For example, the heat transferred from
a hot-plate to the atmosphere is by natural convection, whereas the heat transferred by a domestic
fan-heater, in which a fan blows air across an electric element, is by forced convection.
The transfer of heat through solid bodies is by conduction alone, whereasthe heat transfer from a
solid surface to a liquid or gas takes place partly byconduction and partly by convection. When-
ever there is an appreciablemovement of the gas or liquid, the heat transfer by conduction in the
gas or liquid becomes negligibly small compared with the heat transfer by convection. However,
there is always a thin boundary layer of fluid on a surface, and through this thin film the heat is
transferred by conduction.
• Convection is possible only in a fluid medium and is directly linked with the transport of
medium itself.
• Convection constitutes the macroform of the heat transfer since macroscopic particles of a
fluid moving in space cause the heat exchange.
• The effectiveness of heat transfer by convection depends largely upon the mixing motion
of the fluid.
3
This mode of heat transfer is met with in situations where energy is transferred as heat to a
flowing fluid at any surface over which flow occurs. This mode is basically conduction in a very
thin fluid layer at the surface and then mixing caused by the flow. The heat flow depends on the
properties of fluid and is independent of the properties of the material of the surface. However,
the shape of the surface will influence the flow and hence the heat transfer.
Free or natural convection. //
Free or natural convection: occurs where the fluid circulates by virtue of the natural differences
in densities of hot and cold fluids ; the denser portions of the fluid move downward because of
the greater force of gravity, as compared with the force on the less dense.
Forced convection: When the work is done to blow or pump the fluid, it is said to be forced
convection.
1.1.3 Radiation
All matter continuously emits electromagnetic radiation unless its temperature is absolute zero.
It is found that the higher the temperature then the greater is are so placed that the radiation
from each body is intercepted by the other, then the body at the lower temperature will receive
more energy than it is radiating, and hence its internal energy will increase; similarly the internal
energy of the body at the higher temperature will decrease. Thus there is a net transfer of energy
from the high-temperature body to the low-temperature body by virtue of the temperature differ-
ence between the bodies. This form of energy transfer satisfies the definition of heat given in the
previous study, and hence we may say that heat is transferred by radiation.
Note: Details of heat by radiation will be studied under a separate topics due to its distinct heat
transfer mechanism.
In any particular example in practice, heat may be transferred by a combination of conduction,
convection, and radiation, and it is usually possible to assess the effects of each mode of heat
transfer separately and then to sum up the results. There are two main groups of problems; first,
the desirable transfer of heat to or from a fluid as in a heat exchanger, boiler, or condenser, and
second, the prevention of heat losses from a fluid to its surroundings.
1.2 Heat Transfer by Conduction
1.2.1 Fourier’s Law of Heat Conduction

Fourier’s law of heat conduction is an empirical law based on observation and states as follows:
“The rate of flow of heat through a simple homogeneous solid is directly proportional to the area
of the section at right angles to the direction of heat flow, and to change of temperature with
respect to the length of the path of the heat flow”.
Mathematically, it can be represented by the equation:
dt
Q̇ A
dx
4
where, Q̇ Heat flow through a body per unit time (in watts), W,
2
A = Surface area of heat flow (perpendicular to the direction of flow), m ,
dt = Temperature difference of the faces of block (homogeneous solid) of thickness ‘dx’ through which h
dx = Thickness of body in the direction of flow, m.
The law is illustrated in Figure 1.1(a) in which a thin slab of material of thickness dx and surface
area A has one face at a temperature t and the other at a temperature (t + dt). Then applying
Fourier’s law we have for the rate of heat flow in the direction x,
dt
i.e., Rate of heat flow,Q̇ .A
dx
dt
Thus, Q̇ K A (1.1)
dx
where, k = Constant of proportionality and is known as thermal conductivity of the body.

The -ve sign of k [equation (1.1)] is to take care of the decreasing temperature along with the
dt
direction of increasing thickness or the direction of heat flow. The temperature gradient is
dx
always negative along positive x direction and therefore the value of Q becomes +ve.
Figure 1.1: Heat flow

through a thin slab of
material.
The rate of heat flow in the direction x is taken as positive, hence the negative sign in equation
(1.1) since dt is always negative. The term A is called the thermal conductivity of the material.
1.2.2 Thermal Conductivity of Materials

The thermal conductivity of a substance can be defined as the heat flow per unit area per unit
time when the temperature decreases by one degree in unit distance.
The units of A are usually written as W /m K or kW /m K.
Consider the transfer of heat through a slab of material as shown in Figure 1.1 (b). At section
X-X, using equation equation (1.1).
5
1
dt
Q̇ kA or Q̇dx kA dt
dx1
Integrating
E0 Q̇dx Et
x t2
kA dt
1
Et
t2
or Q̇x A k dt
1
This equation can be solved when the variation of thermal conductivity, k, with temperature, t,
is known. Now for most solids the value of the thermal conductivity is approximately constant
over a wide range of temperatures, and therefore k can be taken as constant,
Et
t2
i.e., Q̇x Ak dt
1
Ak Ak
or, Q̇
x t2 t1 x t1 t2 (1.2)
Note that in this case the area in the direction at right angles to the heat flow remains constant
through the slab. Cases will be considered later in which the area varies.
The thermal conductivities of some materials encountered in engineering are shown in Table
1.1. It follows from equation equation (1.1) that materials with high thermal conductivities are
good conductors of heat, whereas materials with low thermal conductivities are good thermal
insulators. Conduction of heat occurs most readily in pure metals, less so in alloys, and much
less readily in non-metals. The very low thermal conductivities of certain thermal insulators (e.g.
cork) are due to their porosity, the air trapped within the material acting as an insulator. Gases
and liquids are good insulators, but unless a completely stagnant layer of fluid is obtained, heat
is transferred by convection currents.
Assumptions:
The following are the assumptions on which Fourier’s law is based:
1. Conduction of heat takes place under steady state conditions.
2. The heat flow is unidirectional.
3. The temperatures gradient is constant and the temperature profile is linear.
4. There is no internal heat generation.
5. The bounding surfaces are isothermal in character.
6. The material is homogeneous and isotropic (i.e., the value of thermal conductivity is con-
stant in all directions).
Some essential features of Fourier’s Law:
Following are some essential features of Fourier’s law:
1. It is applicable to all matter (may be solid, liquid or gas).
2. It is based on experimental evidence and cannot be derived from first principle.
6
3. It is a vector expression indicating that heat flow rate is in the direction of decreasing
temperature and is normal to an isotherm.
4. It helps to define thermal conductivity ‘k’ (transport property) of the medium through
which heat is conducted.
Q̇ dx
From equation (1.1), we have, k =
A dt
dx
The value of k = 1 when Q̇ = 1, A = 1 and 1
dt
Q̇ dx 1 m `
Now k =
1 dt
; unit of k : W 2
m K (or` C
= W/mK or W/m C
Thermal conductivity (a property of material) depends essentially upon the following factors:
1. Material structure, 3. Density of the material,

4. Pressure and temperature (operating con-
2. Moisture content, ditions).
Table 1.1: Thermal conductivities (average values at normal pressure and temperature) of some
common materials.
S/No. Material Thermal conductivity (k) S/No. Material Thermal conductivity (k)
(W/mK) (W/mK)
1. Silver 410 8. Asbestos sheet 0.17
2. Copper 385 9. Ash 0.12
3. Aluminum 225 10. Cork, felt 0.05-0.10
4. Cast-iron 55-65 11. Saw dust 0.07
5. Steel 20-45 12. Glass wool 0.03
6. Concrete 1.20 13. Water 0.55-0.7
7. Glass (window) 0.75 14. Freon 0.0083
Following points regarding thermal conductivity-its variation for different materials and under
different conditions are worth noting:
1. Thermal conductivity of a material is due to flow of free electrons (in case of metals) and
lattice vibrational waves (in case of fluids).
`
2. Thermal conductivity in case of pure metals is the highest (k = 10 to 400 W/m C). It de-
creases with increase in impurity. The range of k for other materials is as follows:
`
Alloys: = k 12 to 120 W/m C
Heat insulating and building materials: k = 0.023 to 2.9 W/m C
`
`
Liquids: k 0.2 to 0.5 W/m C
`
Gases and vapours: k 0.006 to 0.05 W/m C
.
7
3. Thermal conductivity of a metal varies considerably when it (metal) is heat treated or
mechanically processed/formed.
4. Thermal conductivity of most metals decreases with the increase in temperature (alu-
minium and uranium being the exceptions).
• In most of liquids the value of thermal conductivity tends to decrease with tempera-
ture (water being an exception) due to decrease in density with increase in tempera-
ture.
• In case of gases the value of thermal conductivity increases with temperature. Gases
with higher molecular weights have smaller thermal conductivities than with lower
molecular weights. This is because the mean molecular path of gas molecules de-
creases with increase in density and k is directly proportional to the mean free path
of the molecule.
5. The dependence of thermal conductivity (k) on temperature, for most materials is almost
linear;
k = k0 1 βt (1.3)
`
where, k0 Thermal conductivity at 0 C, and
`
β Temperature coefficient of thermal conductivity, 1© C. It is usually
positive for non-metals and insulating materials (magnesite bricks being the
exception and negative for metallic conductors aluminium and certain non-ferrous
alloys are the exceptions.
6. In case of solids and liquids, thermal conductivity (k) is only very weakly dependent on
pressure; in case of gases the value of k is independent of pressure (near standard atmo-
spheric).
7. In case of non-metallic solids:
• Thermal conductivity of porous materials depends upon the type of gas or liquid
present in the voids.
• Thermal conductivity of a damp material is considerably higher than that of the dry
material and water taken individually.
• Thermal conductivity increases with increase in density.
8. The Wiedemann and Franz law (based on experiment results), regarding thermal and elec-
trical conductivities of a material, states as follows: “The ratio of the thermal and electri-
cal conductivities is the same for all metals at the same temperature; and that the ratio is
directly proportional to the absolute temperature of the metal.”
k
Mathematically, σ T
k
or C (1.4)
σT
8
where, k = Thermal conductivity of metal at temperature T(K),
σ Electrical conductivity of metal at temperature T(K), and
C = Constant (for all metals) is referred to as Lorenz number
8 2
2.45 10 WΩ©K ; Ωstands for ohms.
This law conveys that the materials which are good conductors of electricity are also good
conductors of heat.
Example 1.1
` `
The inner surface of a plane brick wall is at 40 C and the outer surface is at 20 C. Calculate
the rate of heat transfer per unit area of wall surface; the wall is 250 mm thick and the thermal
conductivity of the brick is 0.52 W/m K.
Solution:
From equation (1.2)
Ak
Q̇ x t1 t2
therefore,
3
Q̇ 10 0.52 2
q̇ 40 20 41.6 W /m
A 250
Note that the symbol q̇ is used for the rate of heat transfer per unit area.
1.2.3 Thermal Resistance (Rth )

When two physical systems are described by similar equations and have similar boundary con-
ditions, these are said to be analogous. The heat transfer processes may be compared by analogy
with the flow of electricity in an electrical resistance. As the flow of electric current in the electri-
cal resistance is directly proportional to potential difference (dV) ; similarly heat flow rate,Q, is
directly proportional to temperature difference (dt), the driving force for heat conduction through
a medium.
As per Ohm’s law (in electric-circuit theory), we have
Potential difference (dV)

Current (I) (1.5)
Electrical resistance (R)
By analogy, the heat flow equation (Fourier’s equation) may be written as
Temperature difference (dt)

Heat flow rate Q̇ (1.6)
dx

kA
9
By comparing equations (1.5) and (1.6), we find that I is analogus to, Q̇, dV is analogous to
dx dx
dt and R is analogous to the quantity
. The quantity
is called thermal conduction
kA kA
dx
resistance Rth cond. (Figure 1.2) i.e., Rth cond.
kA
The reciprocal of the thermal resistance is called thermal conduc-
tance.
It may be noted that rules for combining electrical resistances in
series and parallel apply equally well to thermal resistances.
The concept of thermal resistance is quite helpful white making
calculations for flow of heat.
Figure 1.2: Analogy of ther-
mal resistance to electrical
resistance.
1.2.4 Heat Conduction Through Plane and Com-
posite Walls
Conduction through a plane wall

Referring to Figure 1.3 (a). Consider a plane wall of homogeneous material through which heat
is flowing only in x-direction.
Let, L = Thickness of the plane wall,

A = Cross-sectional area of the wall,
k = Thermal conductivity of the wall material, and
t1 , t2 Temperatures maintained at the two faces 1 and 2 of the wall, respectively.
From equation (1.41), the general heat conduction equation in cartesian coordinates is given by:
2 2 2
∂ t ∂ t ∂ t qg 1 ∂t

∂ x2 ∂ y2 ∂ z2 k α ∂τ
If the heat conduction takes place under the conditions, steady state ∂∂τt 0, one-dimensional
2
∂ t
2
∂ t
2
∂ t
qg
2 2 2 0 and with no internal heat generation 0 then the above equation is
∂x ∂y ∂z k
reduced to:
2 2
∂ t dt
0 or 0 (1.7)
∂ x2 dx2
By integrating the above differential twice, we have
∂t
C1 and t C1 x C2 (1.8)
∂x
where C1 and C2 are the arbitrary constants. The values of these constants may be calculated
from the known boundary conditions as follows :
At x = 0 t = t1
10
Figure 1.3: Heat conduction through a plane wall.
At x = L t = t2
Substituting the values in the equation (1.8), we get t1 O C2 and t2 C1 L C2 After simpli-
t2 t1
fication, we have, C2 t1 and C1
L
Thus, the equation (1.8) reduces to:
t2 t1
t
x t1 (1.9)
L
The equation (1.9) indicates that temperature distribution across a wall is linear and is indepen-
dent of thermal conductivity. Now heat through the plane wall can be found by using Fourier’s
equation, eqn. (1.9) as follows:
dt dt
Q kA , temperature gradient
dx dx
dt d t2 t1 t2 t1
But,
x t1
dx dx L L
t2 t1 kA t1 t2

Q kA (1.10)
L L
Equation (1.10) can be written as:
t1 t2 t1 t2
Q (1.11)
L©kA RT cond.
11
where, (RT cond. = Thermal resistance to heat conduction. Figure 1.3 (b) shows the equivalent
thermal circuit for heat flow through the plane wall.
Let us now find out the condition when instead of space, weight is the main criterion for selection
of the insulation of a plane wall.
L
Thermal resistance (conduction) of the wall, RT cond. (i)
kA
Mass of the wall, m ρA L (ii)
Eliminating L from eqn.(i) and eqn.(ii), we get
2
m ρA RT cond. kA ρ.k)A . RT cond. (1.12)
The equation (1.12) stipulates the condition that, for a specified thermal resistance, the lightest
insulation will be one which has the smallest product of density (ρ) and thermal conductivity
(k).
Heat conduction through a composite wall

Referring Figure 1.4(a). Consider the transmission of heat through a composite wall consisting
of a number of slabs.
Let xA , xB , xC = Thicknesses of slabs A, B and C respectively (also called path lengths),
kA , kB , , kC , = Thermal conductivities of the slabs A, B and C respectively,
t1 , t4 t1 % t4 = Temperatures at the wall surfaces 1 and 4 respectively, and
t2 , t3 = Temperatures at the interfaces 2 and 3 respectively.
Figure 1.4: Steady state conduction through a composite wall.
12
Since the quantity of heat transmitted per unit time through each slab/layer is same, we have
kA A t1 t2 kB A t2 t3 kC A t3 t4
Q xA xB xC
(Assuming that there is a perfect contact between the layers and no temperature drop occurs
across the interface between the materials).
Rearranging the above expression, we get
Q xA
t1 t2 (i)
kA A
Q xB
t2 t3 (ii)
kB A
Q̇ xC
t3 t4 (iii)
kC A
Adding eqns.(i), (ii) and (iii), we have

xA xB xC
t1 t4 Q̇
kA A kB A kC A
A t1 t4
or, Q̇ (1.13)
xA xB xC

kA kB kC
t1 t4 t1 t4
and, Q̇ (1.14)
xA xB xC RA RB RC

kA A kB A kC A
If the composite wall consists of n slabs/layers, then
t1 tn1
Q̇ (1.15)
<n1 kAx
In order to solve more complex problems involving both series and parallel thermal resistances,
the electrical analogy may be used. A typical problem and its analogous electric circuit are
shown in Figure 1.5.
∆toverall
Q̇
< RT
(1.16)
Thermal contact resistance. In a composite (multi-layer) wall, the calculations of heat flow
are made on the assumptions: (i) The contact between the adjacent layers is perfect, (ii) At the
interface there is no fall of temperature, and (iii) At the interface the temperature is continuous,
although there is discontinuity in temperature gradient.
In real systems, however, due to surface roughness and void spaces (usually filled with air) the
contact surfaces touch only at discrete locations. Thus there is not a single plane of contact,
13
Figure 1.5: Series and parallel one-dimensional heat transfer through a composite wall and elec-
trical analog.
which means that the area available for the flow of heat at the interface will be small compared
to geometric face area. Due to this reduced area and presence of air voids, a large resistance to
heat flow at the interface occurs. This resistance is known as thermal contact resistance and it
causes temperature drop between two materials at the interface as shown in Figure 1.10.
Figure 1.6: Temperature drops

at the interfaces.
Referring Figure 1.10. The contact resistances are given by
t2 t3 t4 t5
RAB cond. and RBC cond.
Q©A Q̇©A
Example 1.2
` `
The inner surface of a plane brick wall is at 60 C and the outer surface is at 35 C.
2
Calculate the rate of heat transfer per m of surface area
of the wall, which is 220 mm thick. The thermal con-
`
ductivity of the brick is 0.51 W/m C.
Solution:
Referring to Figure 1.7
Temperature of the inner surface of the wall, t1 60 C.
`
Temperature of the outer surface of the wall, t2 35 C
`
The thickness of the wall, L = 220 mm = 0.22 m
14
Thermal conductivity of the brick, k = 0.51 W/m C
`
2
Rate of heat transfer per m , q̇
Rate of heat transfer per unit area,
Q̇ k t1 t2 0.51 60 35
q̇ x
A 0.22
2
57.95.W/m .
Example 1.3
A reactor’s wall 320 mm thick, is made up of an inner layer of fire brick (k = 0.84 W/m C)
`
`
covered with a layer of insulation (k = 0.16 W/m C). The reactor operates at a temperature of
` `
1325 C and the ambient temperature is 25 C.
(a) Determine the thickness of fire brick and insulation which gives minimum heat loss;
(b) Calculate the heat loss presuming that the insulating material has a maximum temperature
`
of 1200 C;
(c) If the calculated heat loss is not acceptable, then state whether addition of another layer of
insulation would provide a satisfactory solution.
Solution:
Given:
` `
t1 1325 C; t2 1200 C, t3 25 C;
`
xA xB x 320 mm or 0.32 m
or xB 0.32 xA
kA 0.84W/m C;
`
kB 0.16W/m C.
`
(a) Thickness of fire brick and insulation xA & xB :
The heat flux, under steady state conditions, is constant
throughout the wall and is same for each layer. Then for
Q̇
unit area of wall from eqn.(1.13) i.e., q̇
A
t1 t4
xA xB xC Figure 1.8: The heat transfer through
a composite wall for Example 1.3.
kA kB kC
t1 t3 t1 t2 t2 t3
i.e., q̇ xA xB
xA xB
kA kB
kA kB
1325 25
Considering first two quantities, we have xA xB

0.84 0.16
1325 1200
xA
0.84
1300 125
or, xA
xA 5.25 0.32 xA
15
1300 125
or, xA
1.68 4.25xA
or, 1300xA 125 1.68 4.25xA
and, xA 0.1146 m or 114.6 mm.
also, Thickness of insulation xB 320 114.6 205.4 mm.
(b) Heat loss per unit area, q:
t1 t2 1325 1200 2
Heat loss per unit area, q xA 916.23 W/m .
0.1146©0.84
kA
If another layer of insulating material is added, the heat loss from the wall will reduce ; conse-
quently the temperature drop across the fire brick lining will drop and the interface temperature
t2 will rise. As the interface temperature is already fixed. Therefore, a satisfactory solution will
not be available by adding layer of insulation.
Example 1.4
An exterior wall of a house may be approximated by a 0.1 m layer of common brick (k = 0.7
`
W/m C) followed by a 0.04 m layer of gypsum plaster (k = 0.48 W/m C).
`
`
Determine the thickness of loosely packed rock wool insulation (k = 0.065 W/m C) that should
be added to reduce the heat loss or (gain) through the wall by 80 per cent. Solution:
Given:
Thickness of common brick, xA = 0.1 m
Thickness of gypsum plaster,xB = 0.04 m
Thickness of rock wool, = xC (in m) = ?
Thermal conductivities:
Common brick, kA = 0.7 W/m C
`
Gypsum plaster, kB = 0.48 W/m C
`
Rock wool, kC = 0.065 W/m C
`
Q̇ t1 t4
From eqn.(1.13) i.e., q̇
A xA xB xC

kA kB kC
Taking temperature difference to be ∆t in case,

Case I. Rock wool insulation not used:
A ∆t A ∆t
Q̇1
xA xB 0.1 0.04

kA kB 0.7 0.48
Case II. Rock wool insulation used:
A ∆t
i.e., Q̇
xA xB xC

kA kB kC
16
Figure 1.9: The heat transfer through
A ∆t
0.1 0.04 xC

0.7 0.48 0.065
But Q̇2 1 0.8Q̇1 0.2Q̇1
A ∆t A ∆t
0.2
0.1 0.04 xC 0.1 0.04

0.7 0.48 0.065 0.7 0.48
0.1 0.04 0.1 0.04 xC

0.2
0.7 0.48 0.7 0.48 0.065
or, 0.1428 0.0833 0.20.1428 0.0833 15.385x

and, 0.2261 0.2 0.2261 15.385x
thus, x 0.0588 m or 58.8 mm
Therefore, the thickness of rock wool insulation should be 58.8 mm.
Example 1.5
A furnace wall consists of 200 mm layer of refractory bricks, 6 mm layer of steel plate and a 100
`
mm layer of insulation bricks. The maximum temperature of the wall is 1150 C on the furnace
`
side and the minimum temperature is 40 C on the outermost side of the wall. An accurate
2
energy balance over the furnace shows that the heat loss from the wall is 400 W/m . It is known
that there is a thin layer of air between the layers of refractory bricks and steel plate. Thermal
`
conductivities for the three layers are 1.52, 45 and 0.138 W/m C respectively. Evaluate:
(a) Number of millimetres of insulation brick equivalent to the air layer,
(b) the temperature of the outer surface of the steel plate
Solution:
Referring to Figure ??
Given:
Thickness of refractory bricks, xA = 200 mm = 0.2 m
Thickness of steel plate, xC = 6 mm = 0.006 m
Thickness of insulation bricks, xD = 100 mm = 0.1 m
Difference of temperature between the innermost and outermost side of the wall,
`
∆t 1150 40 1110 C
` `
Thermal conductivities: kA = 1.52 W/m C ; kB = kD = 0.138 W/m C ; kC = 45 W/m C
`
2
Heat loss from the wall, q = 400 W/m
17
Figure 1.10: Heat loss through
a composite wall for Example
1.5.
(a) The value of xB :

We know, from eqn.(1.15),
t1 tn1
i.e., Q̇
< kAx
n
A∆t Q̇ ∆t
or, Q̇ x or A q̇
< k
< x
k
1110 xA xB xC xD
i.e., 400
xA xB xC kA kB kC kD

kA kB kC
1110 1110
i.e., 400
xA xB xC xD 0.2 xB ©1000 0.006 0.1

kA kB kC kD 152 0.138 45 0.138
1110 1110
0.1316 0.0072xB 0.00013 0.7246 0.8563 0.0072xB
1110
and, 0.8563 0.0072xB 2.775
400
2.775 0.8563
i.e., xB 266.5 mm.
0.0072
(b) Temperature of the outer surface of the steel plate tso :
tso 40
q̇ 400 xD
kD
tso 40
or, 400 1.38 tso 40
0.1©0.138
400 `
or, tso 40 329.8 C.
1.38
18
Example 1.6
Calculate the heat flow rate through the composite wall shown in Figure 1.11. Assume one
dimensional flow.
` `
kA = 150 W/m C, kB = 30 W/m C, kC = 65 W/m C and
`
`
kD = 50 W/m C. Solution:
Given:
The thermal circuit for heat flow in the given composite
system (Figure 1.11) has been illustrated in Figure 1.12.
Thickness : xA = 3 cm = 0.03 m ; xB = xC = 8 cm = 0.08
m ; xD = 5 cm = 0.05 m
2
Areas: AA = 0.1 0.1 = 0.01 m ; AB = 0.1 0.03 = Figure 1.11: One dimensional heat
2
0.003 m
2 2 flow through a composite slab for Ex-
AC = 0.1 0.07 = 0.007 m ; AD = 0.1 0.1 = 0.01 m ample 1.6.
Heat flow rate, Q̇:
The thermal resistances are given by equation,
L
RT cond.
kA
xA 0.03
RT A 0.02
kA AA 150 0.01
xB 0.08
RT B 0.89
kB AB 30 0.003
xC 0.03
RT C 0.176
kC AC 65 0.007
xD 0.05
RT D 0.1
kD AD 50 0.001
The equivalent thermal resistance for the parallel ther-
mal resistances RT B and RT C is given by:
1 1 1 1 1

RT eq RT B RT C 0.89 0.176
1
RT eq 0.147 Now, the total ther-

6.805
mal resistance is given by RT total RT A RT eq
RT D 0.02 0.147 0.1 0.267
∆toverall 400 60
1273.4 W
RT total 0.267
Figure 1.12: One dimensional heat

flow through a composite slab for Ex-
ample 1.6.
1.2.5 The Overall Heat-transfer Coeffi-
cient
Newton’s law of cooling
19
In order to consider the rate at which heat is transferred
from one fluid to another through a plane wall it is nec-
essary to know something of the way in which heat is
transferred from a solid surface to a fluid and vice versa.
Newton’s law of cooling states that the heat transfer
from a solid surface of area A, at a temperature tw , to a fluid of temperature t, is given by
Q̇ αA tw t (1.17)
2
where α is called the heat transfer coefficient. The units of α are seen to be W /m K, or kW
2
/m K. The heat transfer coefficient, α, depends on the properties of the fluid and on the fluid
velocity; it is usually necessary to evaluate it by experiment. This will be discussed more fully
in later section studies.
Equation (1.17) does not include the heat loss from the surface by radiation. This effect can
be calculated separately, and in many cases is negligible compared with the heat transferred by
conduction and convection from the surface to the fluid. When the surface temperature is high,
or when the surface loses heat by natural convection, then the heat transfer due to radiation is of
a similar magnitude to that lost by convection.
Figure 1.13: Temperature variation for heat
transfer from one fluid to another through a
dividing wall.
Consider the transfer of heat from a fluid A to a fluid B through a dividing wall of thickness
x, and thermal conductivity, λ , as shown in Figure 1.13. The variation of temperature in the
direction of the heat transfer is also shown. In fluid A the temperature decreases rapidly from
tA to t1 in the region of the wall, and similarly in fluid B the temperature decreases rapidly from
t2 to tB in the region of the wall. In most practical cases the fluid temperature is approximately
constant throughout its bulk, apart from a thin film near the solid surface bounding the fluid. The
dotted lines drawn on Figure 1.13 show that the thickness of this film of fluid is given by δA for
fluid A and δB for fluid B. The heat transfer in these films is by conduction only, hence applying
equation (1.10) we have, considering unit surface area, from fluid A to the wall.
kA
q̇. t t (a)
δA A 1
from the wall to fluid B
kB
q̇. t t (b)
δB 2 B
20
Also from equation (1.17), from fluid A to the wall
q̇ αA tA t1 (c)
from the wall to fluid B
q̇ αB t2 tB (d)
Comparing equations (a) and (c), and equations (b) and (d), it can be seen that
kA kB
αA and αB
δA δB
k
In general, α , where δ is the thickness of the stagnant film of fluid on the surface.
δ
The heat flow through the wall in Figure 1.13 is given by equation (1.10).
k
i.e., for unit surface area,q̇. x t1 t2
For steady-state heat transfer, the heat flowing from fluid A to the wall is equal to the heat flowing
through the wall, which is also equal to the heat flowing from the wall to fluid B. If this were
not so, then the temperatures tA , t1 , t2 and tB would not remain constant but would change with
time.
We therefore have
k
q̇ αA tA t1 x t1 t2 αB t2 tB
Rewriting these equations in terms of the temperatures, then

q̇ q̇x q̇
tA t1 αA ; t1 t2 ; t2 tB
k αB
Hence adding the corresponding sides of the three equations

q̇ q̇x q̇
tA t1 t1 t2 t2 tB
αA λ αB
1 x 1

tA tB q̇ α
A k αB
tA tB
i.e., q̇ (1.18)
1©αA x/k 1©αB
By analogy with equation (1.17) this can be written as,
q̇ U tA tB (1.19)
Q̇ UA tA tB (1.20)
1 1 x 1

where,
U αA k αB
(1.21)
U is called the overall heat transfer coefficient, and it has the same units as α.
21
Example 1.7
`
A mild steel tank of wall thickness 10 mm contains water at 90 C when the atmospheric tem-
`
perature is 15 C. The thermal conductivity of mild steel is 50 W/mK, and the heat transfer
2
coefficients for the inside and outside of the tank are 2800 and 11 W /m K respectively. Calcu-
late:
(a) the rate of heat loss per unit area of tank surface;
(b) the temperature of the outside and inside surface of the tank.
Solution:
(a) the rate of heat loss;
The wall of the tank is shown diagrammatically in Fig-
ure 1.14.
From equation (1.21),
1 1 x 1 1 1 1
i.e.,
U α A k α B 2800 10 50 11
3
0.000357 0.0002 0.0909
1
0.0915
U
Then substituting in equation (1.19),
q̇ U tA tB , we have
90 15 2
q̇
820 W/m
0.0915
i.e., Rate of heat loss per square metre of surface area 0.82 kW
Figure 1.14: Tank wall for
(b) temperature of the outside surface; Example 1.7.
i.e., q̇ α t2 tB
or, 820 11 t2 15

where t2 is the temperature of the outside surface of the tank as shown in Figure 1.14.
820 `
and, 15 89.6 C
11
`
i.e., Temperature of outside surface of tank 89.6 C
3
k 820 10 10 `
Also, q̇ x t1 t2 or t1 11
89.6 90.3 C
`
i.e., 90.3 C
22
The composite wall with fluid film boundaries and the further electrical
analogy
There are many cases in practice when different materials are constructed in layers to form a
composite wall. An example of this is the wall of a building, which usually consists of a layer of
plaster, a row of bricks, an air gap, a second row of bricks, and perhaps a cement rendering on
the outside surface.
Consider the general case of a composite wall as shown in Figure 1.14. There are n layers of
material of thickness x1 , x2 , x3 , etc. and of thermal conductivity k1 , k2 , k3 , etc. On one side
of the composite wall there is a fluid A at temperature tA , and the heat transfer coefficient from
fluid to wall is αA ; on the other side of the composite wall there is a fluid B, and the heat transfer
coefficient from wall to fluid is αB . Let the temperature of the wall in contact with fluid A be t0
and the temperature of the wall in contact with fluid B be tB ; the interface temperatures are then
t1 , t2 , t3 , etc. as shown. The most convenient method of solving such a problem is by making
use of an electrical analogy as seen earlier. The flow of heat can be thought of as analogous to an
electric current. The heat flow is caused by a temperature difference whereas the current flow is
caused by a potential difference, hence it is possible to postulate a thermal resistance analogous
to an electrical resistance. From Ohm’s law we have,
V
V IR or I
R
where V is the potential difference, I the current, and R the resistance.
Figure 1.15: Heat transfer through

a composite wall.
Comparing this equation with equation (1.2), Q̇ kA/x t1 t2 , we have,
x
Thus, thermal resistance, R (1.22)
kA
where Q̇ is analogous to I, and t1 t2 is analogous to V.

The composite wall is analogous to a series of resistances, as shown in Figure 1.15, and resis-
tances in series can be added to give the total resistance. To find the resistance of a fluid film it
is necessary to compare Ohm’s law with equation (1.17) Q̇ αA tw t,
23
1
i.e., Thermal resistance of a fluid film, R (1.23)
αA
where, Q̇ is analogous to I and tw t is analogous to V. Note that the units of thermal resistance
are K/W or K/kW.
Referring to Figure 1.15, we therefore have,
1 x1 x1
RA , R1 , R1 , etc.
αA A k1 A k1 A
xn 1
Rn and RB
kn A αB A
The total resistance to heat flow is then,
1 x1 xn 1
RT RA R1 R2 Rn RB
αA A k1 A k n A αB A
Or for any number of layers of material,
Total resistance, RT
1
αA A
= kAx
1
αB A
(1.24)
It can be seen from equation (1.24) that in this case the surface area, A, remains constant through
the wall, and it is usual to calculate the total resistance for unit surface area in such problems.
Cases in which the area varies through the various layers are considered later.
Using the electrical analogy for the overall heat transfer we have,
A tA tB
Q̇ (1.25)
RT
(analogous to I = V/R ).
In equation (1.21) the overall heat transfer coefficient, U, is defined as
1 1 x 1

U αA k αB
For any number of walls we have
1
U

1
αA = xk
1
αB
It can be seen that the reciprocal of U is simply the thermal resistance for unit area,
1 1
i.e., RT A or U (1.26)
U RT A
If the inner and outer wall surface temperatures are known then the heat transfer can be found by
calculating the thermal resistance of the composite wall only,
24
i.e., R = kAx
The overall heat transfer coefficient from one wall surface to the other is given by,
1
U
= xk
It should be noted that there may be an additional thermal resistance at the various interfaces of
a composite wall, due to the small pockets of air trapped between the surfaces.
Example 1.8
The interior of a refrigerator having inside dimensions of 0.5 m 0.5 m base area and 1 m
`
height, is to be maintained at 6 C. The walls of the refrigerator are constructed of two mild steel
`
sheets 3 mm thick (k = 46.5 W/m C) with 50 mm of glass wool insulation (k = 0.046 W/m C)
`
between them. If the average heat transfer coefficients at the inner and outer surfaces are 11.6
2` 2`
W/m C and 14.5 W/m C respectively, calculate:
(a) The rate at which heat must be removed from the interior to maintain the specified temper-
`
ature in the kitchen at 25 C, and
(b) The temperature on the outer surface of the metal sheet.
Solution:
Referring to Figure 1.16 Given:
xA = xC = 3 mm = 0.003 m;
xB = 50 mm = 0.05 m;
`
kA = kC = 46.5 W/m C ; kB = 0.046 W/m C;
`
2` 2`
h0 = 11.6 W/m C ; hi = 14.5 W/m C ;
` `
t0 = 25 C ; ti = 6 C.
The total area through which heat is coming into the refrigerator
2
A 0.5 0.5 2 0.5 1 4 2.5 m
Figure 1.16: Heat leaks through re-

frigerator insulation for Example
1.8.
25
(a) The rate of removal of heat, Q̇:
From eqn. (1.18)
tA tB
i.e., q̇
1©αA x/K 1©αB
And for this case,
A t0 ti
Q̇
1©α0 xA ©KA xB ©KB xC ©KC 1©αi
2.5 25 6
38.2 W
1 0.003 0.05 0.003 1

11.6 46.5 0.046 46.5 14.5
(b) The temperature at the outer surface of the metal sheet, t1 :

From eqn. (1.17),
i.e., Q̇ αA tw t
Here,
Q̇ αA 25 t1
or, 38.2 11.6 2.5 25 t1
382 `
and, t1 25 23.68 C.
116 2.5
Example 1.9
A furnace wall consists of 125 mm wide refractory brick and 125 mm wide insulating firebrick
separated by an air gap. The outside wall is covered with a 12 mm thickness of plaster. The inner
` `
surface of the wall is at 1100 C and the room temperature is 25 C. The heat transfer coefficient
2
from the outside wall surface to the air in the room is 17 W/m K, and the resistance to heat flow
of the air gap is 0.16 K/W. The thermal conductivities of refractory brick, insulating firebrick,
and plaster are 1.6, 0.3, and 0.14 W/mK, respectively. Calculate:
(a) the rate of heat loss per unit area of wall surface;
(b) the temperature at each interface throughout the wall;
(c) the temperature at the outside surface of the wall.
Solution:
(a) rate of heat loss per unit area, q̇
2
The wall is shown in Figure 1.17. Consider 1 m of surface area. Then using equation (1.22),
x
i.e., Ṙ
kA
,
Thus,
125
Resistance of refractory brick 0.0781 K/W
103 1.6
26
Figure 1.17: Composite wall for Example
1.9.
125
Resistance of insulating firebrick 0.417 K/W
103 1.6
125
Resistance of plaster 0.0857 K/W
103 1.6
Also, using equation (1.23) for a fluid film,
1
i.e., R
αA
1
and, Resistance of air film on outside surface K/W
17
Hence,
1
Total resistance, RT 0.0781 0.417 0.0857 0.16
17
where the resistance of the air gap is 0.16 K/W,
i.e., RT 0.8 K/W
Then using equation (1.25)

tA tB
i.e., Q̇
RT
1100 25
1344 W
0.8
Then, Rate of heat loss per square metre of surface area 1.344kW
(b) temperature at each interface; Referring to Figure 1.17, the interface temperatures are t1
, t2 , and t3 ; the outside surface is at t4 . Applying the electrical analogy to each layer and using
the values of thermal resistance calculated above, we have,
1100 t1
Q̇ 1344
0.0781
`
i.e., t1 1100 1344 0.0781 995 C
27
t1 t2
Also, Q̇ 1344
0.16
`
i.e., t2 995 0.16 1344 780 C
t2 t3
Again, Q̇ 1344
0.417
`
i.e., t3 780 1344 0.417 220 C
t3 t4
And, Q̇ 1344
0.0857
`
i.e., t4 220 1344 0.0857 104 C
(c) temperature at the outside surface; The temperature t4 can also be found by considering
the air film,
t4 25
i.e., Q̇ 1344
1©17
1
or, t4 1344
25
17
`
and, t4 104.1 C
`

Temperature at outside surface of wall 104.1 C
Example 1.10
A furnace wall is made up of three layers of thicknesses 250 mm, 100 mm and 150 mm with
`
thermal conductivities of 1.65, k and 9.2 W/m C respectively. The inside is exposed to gases
` 2` `
at 1250 C with a convection coefficient of 25 W/m C and the inside surface is at 1100 C, the
` 2`
outside surface is exposed air at 25 C with convection coefficient of 12 W/m C. Determine:
(a) The unknown thermal conductivity ‘k’;
(b) The overall heat transfer coefficient;
(c) All surface temperatures.
Solution:
xA = 250 mm 0.25 m; xB = 100 mm = 0.1 m;
xC = 150 mm = 0.15 m ;
`
kA = 1.65 W/m C;
`
kC = 9.2 W/m C ;
` `
th = 1250 C ; t1 = 1100 C
2` 2`
hh = 25 W/m C; hc = 12 W/m C
(a) Thermal conductivity, k (= kB ):
28
Figure 1.18: Composite furnace wall for
Example 1.10.
The rate of heat transfer per unit area of the furnace wall is obtained from eqn.(1.17),
i.e., q̇ α t2 tB
here,
q̇ αh th t1
2
25 1250 1100 3750 W/m
Also, from equation (1.25)
tA tB
q̇
RT
In this case,
∆toverall
q̇
RT total
th tc
Rth conv.h RthA RthB RthC Rth conv.c
1250 25
i.e., 3750
1©αh xA ©KA xB ©KB xC ©KC 1©αc
1225
or, 3750
1 0.25 0.1 0.15 1

25 1.65 KB 9.2 12
1225 1225
0.1 0.1
0.04 0.1515 0.0163 0.0833 0.2911
KB KB
0.1
or, 3750 0.2911
1225
KB
0.1 1225
and, 0.2911 0.0355
KB 3750
0.1 2`
thus, KB K 2.817 W/m C.
0.0355
29
(b) The overall transfer coefficient, U:
1
The overall heat transfer coefficient, U
Rth total
1 0.25 0.1 0.15 1

Now, Rth total
25 1.65 2.817 9.2 12
` 2
0.04 0.1515 0.0355 0.0163 0.0833 0.3266 C m ©W
1 1 2`
U 3.06 W/m C.
Rth total 0.3266
(c) Surface temperatures ; t1 , t1 , t1 , t1 :
q qA qB qC
t1 t2 t2 t3 t3 t4
or, 3750
xA ©kA xB ©kB xC ©kC
1110 t2 0.25 `
i.e, 3750 or t2 1100 3750 531.8 C
0.25©1.65 1.65
531.8 t3 0.1 `
Similarly, 3750 or t3 531.8 3750 398.6 C
0.1©2.817 2.817
398.6 t4
and, 3750
0.15©9.2
0.15 `
or, t4 398.6 3750 337.5 C
9.2
337.5 25 337.5 25 2

Check using outside convection,q̇ 3750 W/m
1 1©12
hc
1.2.6 Heat flow through a cylinder and a sphere

One of the most commonly occurring problems in practice is the case of heat being transferred
through a pipe or cylinder. Less common is the case of heat being transferred through a spherical
wall, but both cases will now be considered.
The cylinder
Consider a cylinder of internal radius r1 , and external radius r2 as shown in Figure 1.19. Let the
inside and outside surface temperatures be t1 and t2 , respectively. Consider the heat flow through
30
a small element, thickness dr, at any radius r, where the temperature is t. Let the conductivity of
the material be k. Then applying equation (1.1), for unit length in the axial direction,
dt dt
i.e., Q̇ k A k 2πr 1
dx dr
dr
or, Q̇ r 2πkdt
Integrating between the inside and outside sur-
faces,
Q̇ Drr1
2 dr
r 2πk Dtt dt
1
2
where Q̇ and k are both constant

r2
Then, Q̇ ln r 2πk t2 t1 2πk t1 t2
1
2πk t1 t2
and, Q̇ (1.27)
ln r2 ©r1
Now from equation (1.2),

Ak
i.e., Q̇ x t1 t2
If we substitute a mean area Am in this equa-

tion, and substitute also for the thickness x = Figure 1.19: Cross-section through a cylinder.
r2 r1 , we have
kAm t1 t2
i.e., Q̇
r2 r1
Comparing this equation with equation (1.27), then
kAm t1 t2 2πk t1 t2
Q̇
r2 r1 ln r2 ©r1
Am 2π
or,
r2 r1 ln r2 ©r1
2π r2 r1 A2 A1
and, Am
ln r2 ©r1 ln r2 ©r1
Here Am is called the logarithmic mean area, and using this area in equation (1.2) an exact
solution is obtained. It can be seen from the above that there is also a logarithmic mean radius
given by,
2π r2 r1
rm
ln r2 ©r1
In the case of a composite cylinder (e.g. a metal pipe with several layers of lagging) the most
convenient approach is again that of the electrical analogy; by using equation (1.22)
31
x
i.e., thermal resistance, R
kA
where x is the thickness of a layer, and Am is the logarithmic mean area for that layer.
From equation (1.27), applying the electrical analogy (I= V/R), it can be seen that,
ln r2 ©r1
R (1.28)
2πk
The film of fluid on the inside and outside surfaces can be treated as before using equation (1.23),
1
i.e., Routside
α0 A0
where A0 is the outside surface area, 2πr2 , referring to Figure 1.19, and α0 is the heat transfer
coefficient for the outside surface.
1
Also, Rinside
αi Ai
where Ai is the inside surface area, 2πr1 and αi is the heat transfer coefficient for the inside
surface.
It can be seen from equation (1.27),
2πk t1 t2
i.e., Q̇
ln r2 ©r1
that the heat transfer rate depends on the ratio of the radii, r2 ©r1 , and not on the difference
r2 r1 . The smaller the ratio, r2 ©r1 , then the higher is the heat flow for the same temperature
difference. In many practical problems the ratio, r2 ©r1 , tends towards unity since the pipe-wall
thickness or lagging thickness is usually small compared with the mean radius. In these cases it
is a sufficiently close approximation to use the arithmetic mean radius,
r2 r1
i.e., Arithmetic mean radius
2
The error in the rate of heat transfer in using the arithmetic mean instead of the logarithmic mean
is just over 4% for a ratio r2 ©r1 2. Most heat transfer experiments in practice cannot give better
accuracy than about 4 or 5%, hence it is a good approximation to use the arithmetic mean area
when r2 ©r1 $ 2.
Heat conduction through a composite cylinder

Consider flow of heat through a composite cylinder as shown in Figure 1.20.
32
Let,th The temperature of the hot fluid flowing inside the cylinder,
tc The temperature of the cold fluid (atmospheric air),
kA Thermal conductivity of the inside layer A,
kB Thermal conductivity of the outside layer B,
t1 , t2 , t3 Temperature at the points 1, 2 and 3 (see Figure 1.20.),
l = Length of the composite cylinder, and
αh , αc Inside and outside heat transfer coefficients.
Figure 1.20: Cross-section of a

composite cylinder.
The rate of heat transfer is given by,

kA 2πl t1 t2
Q̇ αh 2πr1 l th t1
ln r2 ©r1
kB 2πl t2 t3
αc 2πr3 l t3 tc
ln r3 ©r2
Rearranging the above expression, we get
Q̇
th t1 (i)
hh 2πr1 l,
Q̇
t1 t2 (ii)
kA 2πl
ln r2 ©r1
Q̇
t2 t3 (iii)
kB 2πl
ln r3 ©r2
33
Q̇
t3 tc (iv)
hc 2πr3 l
Adding equations (i), (ii), (iii) and (iv), we have,
Z
^ [
_
^ _
Q̇ ^
^
^ 1 1 1 1 _ _
_
^
^ _
hc r3 _
th tc
2πl ^
^
^ hh r1 kA kB _
_
_
^
^ _
_
\ ln r2 ©r1 ln r3 ©r2 ]
2πl th tc
Q̇ Z [
^
^ _
_
^
^ 1 1 1 1 _ _
^
^ _
_
^
hc r3 _

^
^ hh r1 kA kB _
_
^
^ _
_
^ ln r2 ©r1 ln r3 ©r2 _
\ ]
2πl th tc
Q̇ (1.29)
1 ln r2 ©r1 ln r3 ©r2 1

hh r1 kA kB hc r3
If there are ‘n’ concentric cylinders, then
2πl th tc
Q̇ (1.30)
<
1 n n 1 1
ln rrn1 ©rn x
hh r1 n 1 kn hc r3
If inside the outside heat transfer coefficients are not considered then the above equation can be
written as,
2πx t1 t n1

Q̇
<
n n 1
ln rn1 ©rn
n 1 kn
Example 1.4
A steel pipe of 100 mm bore and 7 mm wall thickness, carrying steam at 260 C, is insulated
`
with 40 mm of a moulded high-temperature diatomaceous earth covering, this covering in turn
insulated with 60 mm of asbestos felt.
`
The atmospheric temperature is 15 C. The heat transfer coefficients for the inside and outside
2
surfaces are 550 and 15 W /m K respectively, and the thermal conductivities of steel, diatoma-
ceous earth, and asbestos felt are 50, 0.09, and 0.07 W/mK respectively.
Calculate:
(a) the rate of heat loss by the steam per unit length of pipe;
(b) the temperature of the outside surface.
34
Solution:
(a) the rate of heat loss;

A cross-section of the pipe is shown in Figure 1.21.
Consider 1 m length of the pipe. From equation (1.23),
1
i.e., R
αA
3
10
and, Resistance of steam film 0.00579 K/W
550 2π 50 1
From equation (1.28), for the steel pipe,
ln r2 ©r1 Figure 1.21: Cross-

i.e., R section through an in-
2πk
sulated cylinder for Ex-
ample 1.4.
ln 57©50

Resistance of pipe 0.000417 K/W
2π 50
ln 97©57
Similarly, Resistance of diatomaceous earth 0.94 K/W
2π 0.09
ln 157©97
and, Resistance of asbestos felt 1.095 K/W
2π 0.07
From equation (1.23), for the air film on the outside surface
3
1 10
Resistance of air film 0.0675 K/W
αA 15 2π 157 1
Hence, Total resistance, RT 0.00579 0.000417 0.94 1.095 0.0675 2.1087 K/W
Note that the resistance to heat flow of the pipe metal is very small; also in this case the resistance
of the film on the inside surface is very small because the heat transfer coefficient for steam is
high.
Then, using equation (1.25)
tA tB
i.e., Q̇
RT
260 15
116 W
2.1087
thus, Rate of heat loss per metre length of pipe 116 W
35
(b) the temperature of the outside surface;
Using the electrical analogy for the air film we have
t 15
also, Q̇ 116
0.0675
where, t is the temperature of the outside surface

`
or, t 116 0.0675 15 22.8 C
`
Hence, Temperature of outside surface 22.8 C
Example 1.5
A thick walled tube of stainless steel with 20 mm
inner diameter and 40 mm outer diameter is cov-
ered with a 30 mm layer of asbestos insulation (k =
`
0.2 W/m C). If the inside wall temperature of the
`
pipe is maintained at 600 C and the outside insu-
`
lation at 1000 C, calculate the heat loss per metre
of length.
Solution:
Given,
20
r1 = = 10 mm = 0.01 m
2
40
r2 = = 20 mm = 0.02 m
2
r3 = 20 + 30 = 50 mm = 0.05 m
` ` `
t1 = 600 C, t3 = 1000 C, kB = 0.2 W/m C
Heat transfer per metre of length, Q/l: Figure 1.22: Cross-section through an insu-
From equation, lated cylinder for Example 1.5.
2πl th tc
Q̇
ln r2 ©r1 ln r3 ©r2

kA kB
and since the thermal conductivity of stainless steel is not given, then resistance offered by
stainless steel to heat transfer across the tube is neglected,
Q̇ 2π th tc
i.e.,
l ln r3 ©r2
kB
Q̇ 2π 600 1000
i.e., 548.57 W/m
l ln 0.05©0.02
0.2
Negative sign indicates that the heat transfer takes place radially inward.
36
Example 1.6
`
Hot air at a temperature of 65 C is flowing through a steel pipe of 120 mm diameter. The pipe
is covered with two layers of different insulating materials of thickness 60 mm and 40 mm, and
`
their corresponding thermal conductivities are 0.24 and 0.4 W/m C. The inside and outside heat
` `
transfer coefficients are 60 and 12 W/m C. The atmosphere is at 20 C. Determine the rate of
heat loss from 60 m length of pipe.
Solution:
Figure 1.23: Cross-

section of a composite
cylinder for Example 1.6.
Given:
120
r1 = = 60 mm = 0.06 m
2
r2 = 60 + 60 mm = 120 mm = 0.12 m
r3 = 60 + 60 + 40= 160 mm = 0.16 m
`
th = 65 C; tc = 20 C
`
`
kA = 0.24 W/m C;
`
kB = 0.4 W/m C
2` 2`
hh = 60 W/m C; hc = 12 W/m C.
Length of pipe, l = 60 m
Rate of heat loss, Q:

Rate of heat loss as given by (1.29);
2πl th tc
Q̇
1 ln r2 ©r1 ln r3 ©r2 1

hh r1 kA kB hc r3
37
2π 60 65 20
1 ln 0.12©0.06 ln 0.16©0.12 1

60 0.06 0.24 0.4 12 0.16
16964.6
3850.5 W
0.2777 2.8881 0.7192 0.5208
i.e., Rate of heat loss 3850.5 W
Example 1.7
A 150 mm steam pipe has inside dimater of 120 mm and outside diameter of 160 mm. It is
`
insulated at the outside with asbestos. The steam temperature is 150 C and the air temperature is
` 2` 2`
20 C. h (steam side) = 100 W/m C, h (air side) = 30 W/m C, k (asbestos) = 0.8 W/m C and
`
`
k (steel) = 42 W/m C. Calculate the thickness of the asbestos to be provided in order to limit the
2
heat losses to 2.1 kW/m .
Solution:
Figure 1.24: Cross-section of a

composite cylinder for Exam-
ple 1.7.
Given:
120
r1 = = 60 mm = 0.06 m
2
160
r2 = = 80 mm = 0.08 m
2
`
th = 150 C; tc = 20 C
`
`
kA = 42 W/m C;
`
kB = 0.8 W/m C
2` 2 `
hh = 100 W/m C; hc = 30 W/m C.
2
Heat loss = 2.1 kW/m
Thickness of insulation (asbestos), (r3 - r2 ):
38
Area for heat transfer 2πrL where L = length of the pipe

Heat loss 2.1 2πrL kW
2.1 2π 0.075 L 0.989L kW
3
0.989L 10 watts
150
where r, mean radius 75 mm or 0.075 m Given
Heat transfer rate in such a case is given by (1.29);
2πL th tc
i.e., Q̇
1 ln r2 ©r1 ln r3 ©r2 1

hh r1 kA kB hc r3
Substituting,
3 2πL 150 20

0.989L 10
1 ln 0.08©0.06 ln r3 ©0.08 1

100 0.06 42 0.8 30 r3
3 816.81
also, 0.989 10
ln r3 ©0.08 1
0.16666 0.00685
0.8 30r3
ln r3 ©0.08 1 816.81
or, 0.16666 0.00685 0.6524
0.8 30r3 0.989 103
1
or, 1.25 ln r3 ©0.08 0.6524 0
30r3
Solving by hit and trial, we get
r3 0.105 m or 105 mm

Thickness of insulation r3 r2 105 80 25 mm.
The sphere
Consider a hollow sphere of internal radius r1 and external radius r2 , as shown in Figure 1.25.
Let the inside and outside surface temperatures be t1 and t2 , and let the thermal conductivity be
k.
39
Consider a small element of thickness dr at any
radius r. It can be shown that the surface area of
this spherical element is given by 4πr2 Then, using
equation (1.1),
Q̇ dx
i.e., k
A dt
dt 2 dt
or, Q̇ kA k4πr
dr dr
Separating variables and Integrating,
Er Et
r2 dr t2
Q̇ 4πk dt
1 r2 1
1 1
Q̇

,
r2 r1
4πk t2 t1 Figure 1.25: Steady state conduction
through a hollow sphere.
Q̇ r2 r1
or, r1 r2 4πk t1 t2
4πkr1 r2 t1 t2 t1 t2
and, Q̇ (a)
r2 r1 r2 r1
4πkr1 r2
Hence applying the electrical analogy, (I = V/R), we have,
r2 r1
R (1.31)
4πkr1 r2
If a mean area, Am , is introduced, then from equation (1.2),

Ak
i.e., Q̇ x t1 t2
kAm kAm t1 t2
then, Q̇ x t1 t2 r2 r1
(b)
Comparing the equations (a) and (b) above, we have

Am 4πr1 r2
A mean radius, rm , can be defined,
2
i.e., Am 4πrm 4πr1 r2
Õ

Mean radius, rm r1 r2
It can be seen that rm is a geometric mean radius.
Example 1.8
40
A spherical shaped vessel of 1.4 m diameter is 90 mm
thick. Establish the rate of heat leakage, if the temper-
ature difference between the inner and outer surfaces is
`
220 C.
Thermal conductivity of the material of the sphere is
0.083 W/m C.
`
Solution:
Given:
1.4
r1 = = 0.7 m
2
90
r2 = 0.7 - = 0.61 m
1000
` ` Figure 1.26: Steady state conduction
t1 - t2 = 200 C; kA = 0.083 W/m C; through a sphere for Example 1.8.
The rate of heat transfer/leakage is given by equation
(a),
4πkr1 r2 t1 t2
i.e., Q̇
r2 r1
4π 0.083 0.61 0.7 220

1088.67 W
0.7 0.61
thus, Rate of heat leakage 1088.67 W.
Heat conduction through a composite sphere;

Considering Figure 1.27 as cross-section of a composite sphere, the heat flow equation can be
written as follows:
2 4πkA r1 r2 t1 t2 4πkB r2 r3 t1 t3 2
i.e., Q̇ αh 4πr1 th t1 αc 4πr3 t3 tc
r2 r1 r3 r2
By rearranging the above equation, we have
Q̇
th t1 (i)
αh 4πr12
Q̇ r2 r1
t1 t2 (ii)
4πkA r1 r2
Q̇ r3 r2
t2 t3 (iii)
4πkB r2 r3
Q̇
t3 tc (iv)
αc 4πr32
41
Figure 1.27: Steady state conduction through
a composite sphere.
Adding equations (i), (ii), (iii) and (iv), we get,
Q̇ 1 r2 r1 r3 r2 1
th tc
4π αh r2 kA r1 r2 kB r2 r3 αc r2
1 3
4π th tc

Q̇
1 r2 r1 r3 r2 1

αh r12 kA r1 r2 kB r2 r3 αc r2
3
If there are n concentric spheres then the above equation can be written as follows :
4π th tc
Q̇ (1.32)
< rn1 rn
1 n n 1
v |
αh r21 n 1 kn rn rn1 αc r2n1
If inside the outside heat transfer coefficients are not considered, then the above equation can be
written as,
4π t1 tn1
Q̇ (1.33)
< rn1 rn
n n

n 1 kn rn rn1
Example 1.9
A small hemispherical oven is built of an inner layer of insulating firebrick 125 mm thick, and
`
an outer covering of 85% magnesia 40 mm thick. The inner surface of the oven is at 800 C and
2 `
the heat transfer coefficient for the outside surface is 10 W/m K; the room temperature is 20 C.
Calculate the rate of heat loss through the hemisphere if the inside radius is 0.6 m.
42
Take the thermal conductivities of firebrick and 85% magnesia as 0.31 and 0.05 W/mK respec-
tively.
Solution:
from equation (1.31), for a hemisphere,
r2 r1
i.e., R
4πkr1 r2
For the insulating firebrick:
0.125
Resistance of firebrick
2π 0.31 0.6 0.725
0.1478 K/W
For the 85% magnesia
0.04
Resistance of 85%magnesia
2π 0.05 0.725 0.765
0.2295 K/W
For the outside surface: from equation (1.23)
1
i.e., Thermal resistance of a fluid film, R
αA Figure 1.28: Cross-section through
a composite hemisphere for Example
1.8.
2
Thus, Resistance of outside air film 10 2π 0.765
0.0272 K/W
Hence, Total resistance, RT 0.1478 0.2295 0.0272

0.4045 K/W
Then using equation (1.25)
tA tB
i.e., Q̇
RT
800 20
and, Q̇ 1930 W
0.4045
thus, Rate of heat loss from the oven 1.93 kW
1.2.7 General Heat Conduction Equation in Cartesian Coordinates

Consider an infinitesimal rectangular parallelepiped (volume element) of sides dx, dy and dz
parallel, respectively, to the three axes (X, Y, Z) in a medium in which temperature is varying
with location and time as shown in Figure 1.29.
43
Let, t = Temperature at the left face ABCD; this temperature may be assumed uniform over
the entire surface, since the area of this face can be made arbitrarily small.
dt
Temperature changes and rate of change along X-direction.
dx
∂t
Then,
dx Change of temperature through distance dx, and
∂x
∂t
t
dx Temperature on the right face EFGH
∂x
(at distance dx from the left face ABCD).
Further, let, kx , ky , kz Thermal conductivities

(direction characteristics of the material) along X, Y and Z axes.
Figure 1.29: Elemental volume for three-dimensional heat conduction analysis-Cartesian co-
ordinates.
If the directional characteristics of a material are equal/same, it is called an “Isotropic material”

and if unequal/different ‘Anisotropic material’. q̇g Heat generated per unit volume per unit time.
Inside the control volume there may be heat sources due to flow of electric current in electric mo-
tors and generators, nuclear fission etc.
[ Note: q̇g may be function of position or time, or both].
ρ = Mass density of material.
c = Specific heat of the material.
Energy balance/equation for volume element:
Net heat accumulated in the element due to conduction of heat from all the coordinate directions
considered (A) + heat generated within the element (B) = Energy stored in the element (C)
44
Let, Q = Rate of heat flow in a direction, and
Q̇’ Q̇dτ Total heat flow (flux) in that direction (in time dτ .
A. Net heat accumulated in the element due to conduction of heat from all the directions consid-
ered:
Quantity of heat flowing into the element from the left face ABCD during the time interval dτ in
X-direction is given by:
¬ ∂t
Heat influx Q̇x kx dy.dz dτ (i)
∂x
During the same time interval dτ the heat flowing out of the right face of control volume (EFGH)
will be:
¬ ¬ ∂ ¬
Heat efflux Q̇xdx Q̇x Q̇x dx (ii)
∂x
Heat accumulation in the element due to heat flow in X-direction,
¬ ¬ ¬ ∂ ¬
dQ̇x Q̇x Q̇x Q̇x dx substracting (ii) from (i)
∂x
∂ ¬
Q̇ dx
∂x x
∂ ∂t
kx dy.dz dτ dx
∂x ∂x
∂ ∂t
kx dx dy dz dτ (1.34)
∂x ∂x
Similarly, the heat accumulated due to heat flow by conduction along Y and Z directions in time
dτ will be:
∂ ∂t
ky dx dy dz dτ (1.35)
∂y ∂y
∂ ∂t
kz dx dy dz dτ (1.36)
∂z ∂z
Net heat accumulated in the element due to conduction of heat from all the co-ordinate
directions considered,
∂ ∂t ∂ ∂t ∂ ∂t
kx dx dy dz dτ ky dx dy dz dτ kz dx dy dz dτ
∂x ∂x ∂y ∂y ∂z ∂z
∂ ∂t ∂ ∂t ∂ ∂t
kx
ky
kz
dx dy dz dτ (1.37)
∂x ∂x ∂y ∂y ∂z ∂z
B. Total heat generated within the element (Q̇g ):

The total heat generated in the element is given by:
45
¬
Qg q̇g dx dy dzdτ (1.38)
C. Energy stored in the element:

The total heat accumulated in the element due to heat flow along coordinate axes (equation
(1.37)) and the heat generated within the element (equation (1.38)) together serve to increase the
thermal energy of the element/lattice. This increase in thermal energy is given by:
∂t
ρ dx dy dzc dτ (1.39)
∂τ

Heat stored in the body = Mass of the body
specific heat of the body material rise in the temperature of body.
Now, substituting equations (1.37), (1.38), (1.39), in the equation (1.1), we have
∂ ∂t ∂ ∂t ∂ ∂t
kx
ky
kz
dx dy dz dτ q̇g dx dy dzdτ
∂x ∂x ∂y ∂y ∂z ∂z
∂t
ρ dx dy dzc dτ
∂τ
Dividing both sides by dx.dy.dz.dτ, we have
∂ ∂t ∂ ∂t ∂ ∂t ∂t
kx
ky
kz
q̇g ρ c (1.40)
∂x ∂x ∂y ∂y ∂z ∂z ∂τ

or, using the vector operator V , we get

V
kVt q̇g ρ c
∂t
∂τ
This is known as the general heat conduction equation for ‘non-homogeneous material’,
‘self heat generating’ and ‘unsteady three-dimensional flow’. This equation establishes in
differential form the relationship between the time and space variation of temperature at any
point of solid through which heat flow by conduction takes place.
General heat conduction equation for constant thermal conductivity:
In case of homogeneous (in which properties e.g., specific heat, density, thermal conductivity
etc. are same everywhere in the material) and isotropic (in which properties are independent of
surface orientation) material, kx ky kz k and diffusion equation equation (1.40) becomes
2 2 2
∂ t ∂ t ∂ t q̇g ρ c ∂t 1 ∂t

α ∂τ τ (1.41)
∂ x2 ∂ y2 ∂ z2 k k ∂τ
k Thermal conductivity
where α ρ c Thermal capacity
k
The quantity, α ρ c is known as thermal diffusivity.
46
• The larger the value of α, the faster will the heat diffuse through the material and its
temperature will change with time. This will result either due to a high value of thermal
conductivity k or a low value of heat capacity ρ c. A low value of heat capacity means
the less amount of heat entering the element would be absorbed and used to raise its tem-
perature and more would be available for onward transmission. Metals and gases have
relatively high value of α and their response to temperature changes is quite rapid. The
non-metallic solids and liquids respond slowly to temperature changes because of their
relatively small value of thermal diffusivity.
• Thermal diffusivity is an important characteristic quantity for unsteady conduction situa-
tions.
2
Equation (1.41) by using Laplacian V , may be written as:
V t2 q̇g
k
1 ∂t
α ∂τ (1.41a)
Equation (1.41), governs the temperature distribution under unsteady heat flow through a mate-
rial which is homogeneous and isotropic.
Other simplified forms of heat conduction equation in cartesian co-ordinates:
(i) For the case when no internal source of heat generation is present. Equation (1.41) reduces
1
j 0 heat flow with no internal heat
2 2 2
∂t
∂ t
∂ x2

∂ t
∂ y2

∂ t
∂ z2 α ∂τ
. [Unsteady state ∂∂τt
generation]
(ii) For the case when no internal source of heat generation is present. Equation (1.41) reduces
1
j 0 heat flow with no internal heat
2 2 2
∂t
∂ t
∂ x2

∂ t
∂ y2

∂ t
∂ z2 α ∂τ
. [Unsteady state ∂∂τt
generation]
or 2
V t
1 ∂t
α ∂τ Fourier’s equation (1.42)
(iii) Under the situations when temperature does not depend on time, the conduction then takes
place in steady state i e., ∂∂τt 0 and the equation (1.41) reduces to
2 2 2
∂ t ∂ t ∂ t q̇g
0
∂ x2 ∂ y2 ∂ z2 k
or, 2
V t
q̇g
k
0 Poisson’s equation (1.43)
In the absence of internal heat generation, equation (1.43) reduces to

2 2 2
∂ t ∂ t ∂ t
0
∂ x2 ∂ y2 ∂ z2
or, 2
V t 0 (Laplace equation) (1.44)
47
(iv) Steady state and one-dimensional heat transfer
2
∂ t q̇g
0 (1.45)
∂ x2 k
(v) Steady state, one dimensional, without internal heat generation
2
∂ t
0 (1.46)
∂ x2
(vi) Steady state, two dimensional, without internal heat generation

2 2
∂ t ∂ t
0 (1.47)
∂ x2 ∂ y2
(vi) Unsteady state, one dimensional, without internal heat generation
2
∂ t 1 ∂t
(1.48)
∂ x2 α ∂τ
General heat conduction equation with cylindrical or spherical coordinates:

Equations using cylindrical or spherical coordinates may be derived in a similar way, or obtained
from equation (1.41) by transforming the coordinates. For one-dimensional problems ( e.g. an
infinitely long cylinder or a sphere), it is simpler to derive the equations directly as will be shown
ahead in heat transfer through cylinder and a sphere, and as below.
(a) Infinite slab (Figure 1.1(b)). From equation (1.41) we have,
Figure 1.30: Cross-section through an infinitely long cylinder.
2
∂ t qg 1 ∂t
(1.41)
∂ x2 k α ∂τ
(b) Infinitely long cylinder (Figure 1.30). Applying an energy balance to an element of thick-
ness, dr, we have, Rate of increase of energy of the element = mass specific heat rate
of change of temperature with time, or,
Heat generated - heat loss = rate of increase of energy of the element
48
∂ Q̇ ∂t
qg 2πr dr dr ρ2πr drc
∂r ∂τ
∂ ∂t ∂t
or, qg 2πr dr k2πr
dr ρ2πr drc
∂r ∂r ∂τ
2
∂ t ∂t ∂t
also, qg r kr 2
k ρcr
∂r ∂r ∂τ
2
∂ t 1 ∂ t qg 1 ∂t
also, (1.49)
∂ r2 r ∂ r k α ∂τ
(c) Sphere (Figure 1.31). Applying an energy balance,
Figure 1.31: A hollow sphere.
2 ∂ Q̇ 2 ∂t
qg 4πr dr dr ρ4πr drc
∂r ∂τ
2 ∂ 2 ∂t 2 ∂t
or, qg 4πr dr k4πr
dr ρ4πr drc
∂r ∂r ∂τ
2
∂ t 2 ∂ t q̇g 1 ∂t
and, (1.50)
∂ r2 r ∂ r k α ∂τ
For steady-state cases the right-hand side of equations (1.41), (1.49) and (1.50) becomes
zero, and the equations become ordinary differential equations.
Heat generated in electrical heating:
The heat generated per unit volume due to the current flowing is given by
2
I R
q̇g
AL
where I is the current, R the electrical resistance, A the cross-sectional area, and L the length.
Current density, J = I/A and resistance, R = sL/A where s is the electrical resistivity of the
conductor material,
49
2 2
J A sL 2
i.e., q̇g J s (1.51)
AL A
Example 1.10
A hollow cylindrical copper conductor of 30 mm outside diameter and 14 mm inside diameter has
2
a current density of 40 A/mm . The external surface is covered with a uniform layer of insulation
`
of thickness 10 mm, and the ambient temperature is 10 C. Neglecting axial conduction and
`
assuming that the temperature of the insulation must not exceed 135 C at any point, calculate:
(a) the heat required to be removed per unit time by forced cooling from the inside of the
conductor;
(b) the temperature at the inside surface of the conductor.
Given data: Thermal conductivity of copper= 380 W/mK; thermal conductivity of insulating
2
material= 0.3 W/mK; heat transfer coefficient at outside surface= 40 W/m K; electrical resistiv-
5
ity of copper= 2 10 Ω mm.
Solution:
From equation (1.49), for the steady state
2
∂ t 1 ∂ t q̇g 1 ∂t

∂ r2 r ∂ r k α ∂τ
Inspection of (1.49) shows the first terms as product of differentiation;
1 ∂ ∂t q̇g
i.e., r ∂ r r ∂ r

k
Hence separating variables and integrating,
2
∂t q̇g r
r C1
∂r 2k
∂t q̇g r C1

, (a)
∂r 2k r
Integrating further, now eqn.(a)
2
q̇g r
t C1 ln r C2 (b)
4k
where C1 and C2 are integration constants.
From equation (1.51)
2 2
J A sL 2
i.e., q̇g J s
AL A
50
2 5 3 6 3
q̇g 40 2 10 W/mm 32 10 W/m
(a) the heat required to be removed per unit time:

`
The maximum temperature of the insulation ( = 135 C) occurs at the interface between the
insulation and the copper tube. Hence for the insulation, using equation (1.27),
2πk t1 t2
i.e., Q̇
ln r2 ©r1
2π 0.3 135 t
and, Heat transfer to the outside, Q̇0
ln 50©30
where t is the temperature of the outside surface of the insulation,
or, 135 t 0.271Q̇0 (c)
For the heat transferred from the outside surface of the insulation by convection, from equation
(1.17)
i.e., Q̇ αA tw t
in this case,
Q̇0 αA t t f luid
i.e., Q̇0 40 2π 0.025 t 10

t 10 0.159 Q̇0 (d)
Adding equations (c) and (d)
135 10 0.43Q̇0
and, Q̇0 290.7 W
Total heat generated internally q̇g volume
6 π 2 2
32 10 0.03 0.014 17693.5 W
4
Hence, Heat removed from inside of conductor 17693.5 290.7 W
17.4 kW
(b) the temperature at the inside surface:

Two boundary conditions are required to find the constants C1 and C2 and hence to obtain the
solution of equation (b).
51
At the inside surface of the conductor
∂t
Hence, Heat supplied to conductor kA
17400 W
∂r r 0.007
∂t 17400
thus,
1041.1 K/m
∂r r 0.007 380 π 0.014
Substituting in equation (a)

6
32 10 0.007 C1
1041.1 w }
2 380 0.007

C1 9.351
`
At the outside surface of the conductor, t = 135 C, hence in equation (b)
6 2
32 10 0.015
135 w } 9.351 ln 0.015 C2
4 380
and, C2 179
Therefore the complete solution for the temperature distribution in the conductor is,
6
32 10 2
t w } r 9.351 ln r 179
4 380
Hence, at the inside surface, when r = 0.007

6 2
32 10 0.007
t w } 9.351 ln 0.007 179
4 380
`
131.6 C
1.2.8 Critical Thickness of Insulation
Insulation-General aspects
Definition: A material which retards the flow of heat with reasonable effectiveness is known as
‘Insulation’. Insulation serves the following two purposes:
(i) It prevents the heat flow from the system to the surroundings;
(ii) It prevents the heat flow from the surroundings to the system.
52
Applications:
The fields of application of insulations are:
(i) Boilers and steam pipes
(ii) Air-conditioning systems
(iii) Food preserving stores and refrigerators
(iv) Insulating bricks (employed in various types of furnaces)
(v) Preservation of liquid gases etc.
Factors affecting thermal conductivity Some of the important factors which affect thermal con-
ductivity (k) of the insulators (the value of k should be always low to reduce the rate of heat flow)
are as follows:
1. Temperature: For most of the insulating materials, the value of k increases with increase
in temperature.
2. Density: There is no mathematical relationship between k and ρ (density). The common
understanding that high density insulating materials will have higher values of k in not
always true.
3. Direction of heat flow: For most of the insulating materials (except few like wood) the
effect of direction of heat flow on the values of k is negligible.
4. Moisture: It is always considered necessary to prevent ingress of moisture in the insulating
materials during service, it is however difficult to find the effect of moisture on the values
of k of different insulating materials.
5. Air pressure: It has been found that the value of k decreases with decrease in pressure.
6. Convection in insulators: The value of k increases due to the phenomenon of convection
in insulators.
Critical Thickness of Insulation

The addition of insulation always increases the conductive thermal resistance. But when the total
thermal resistance is made of conductive thermal resistance [(Rth cond. ] and convective thermal
resistance [(Rth conv. ], the addition of insulation in some cases may reduce the convective thermal
resistance due to increase in surface area, as in the case of a cylinder and a sphere, and the total
thermal resistance may actually decrease resulting in increased heat flow. It may be shown that
the thermal resistance actually decreases and then increases in some cases.
“The thickness” upto which heat flow increases and after which heat flow decreases is termed as
Critical thickness. In case of cylinders and spheres it is called ‘Critical radius’.
A. Critical thickness of insulation for cylinder:
Consider a solid cylinder of radius r1 insulated with an insulation of thickness r2 r1 as shown
in Figure 1.32.
Let, x = Length of the cylinder,
t1 = Surface temperature of the cylinder,
tair = Temperature of air,
α0 = Heat transfer coefficient at the outer surface of the insulation, and
53
k = Thermal conductivity of insulating material.
Then the rate of heat transfer from the surface of the

solid cylinder to the surroundings is given by
2πx t1 tair
Q̇ (1.52)
ln r2 ©r1 1

k α0 r2
From eqn. (1.52) it is evident that as r2 increases,

ln r2 ©r1 1
the factor increases but the factor α r de-
k 0 2
creases.
Thus Q̇ becomes maximum when the denominator

ln r2 ©r1 1
k

α0 r2 becomes minimum.
Figure 1.32: Critical thickness of insu-
The required condition is, lation for a cylinder.
d ln r2 ©r1 1

dr2 k

α0 r2 0 r2 being the only variable
1 1 1 1
i.e., 2 0
r
k 2 α 0 r 2
1 1
or, 0 or α0 r2 k
k α0 r2
k
and, r2 rc α0 (1.53)
The above relation represents the condition for minimum resistance and consequently *maximum
heat flow rate. The insulation radius at which resistance to heat flow is minimum is called the
‘critical radius (rc ). The critical radius rc is dependent of the thermal quantities k and α0 and is
independent of r1 (i.e., cylinder radius).
*It may be noted that if the second derivative of the denominator is evaluated, it will come out
to be positive. This would verify that heat flow rate will be maximum, when r2 rc .
In eqn. (1.52),
2πx t1 tair
i.e., Q̇ ,
ln r2 ©r1 1

k α0 r2
ln r2 ©r1
is the conduction (insulation) thermal resistance which increases with increasing
k
r2 and 1/α0 .r2 is the convective thermal resistance which decreases with increasing r2 . At r2 =
rc the rate of increase of conductive resistance of insulation is equal to the rate of decrease of
convective resistance thus giving a minimum value for the sum of thermal resistances.
54
In the physical sense we may arrive at the following conclusions:
(i) For cylindrical bodies with r1 $ rc , the heat transfer increases by adding insulation till r2 =
rc as shown in Figure 1.33 (a)]. If insulation thickness is further increased, the rate of heat
loss will decrease from this peak value, but until a certain amount of insulation denoted
by r2 ’ at b is added, the heat loss rate is still greater for the solid cylinder. This happens
when r1 is small and rc is large, viz., the thermal conductivity of the insulation k is high
(poor insulating material) and α0 is low. A practical application would be the insulation of
electric cables which should be good insulator for current but poor for heat.
(ii) For cylindrical bodies with r1 % rc , the heat transfer decreases by adding insulation[Figure 1.33
(b)]. This happens when r1 is large and rc is small, viz., a good insulating material is used
with low k and α0 is high. In steam and refrigeration pipes heat insulation is the main
objective. For insulation to be properly effective in restricting heat transmission, the outer
radius must be greater than or equal to the critical radius.
Figure 1.33: Effect of insulation thickness on rate of heat loss.
B. Critical thickness of insulation for sphere:

Referring to Figure 1.34. The general equation of heat flow through a sphere with insulation is
as given in eqn.(1.32)
4π th tc
i.e., Q̇
< v knrnr1n
1 n n rn 1
|
hh r21 n 1 rn1 hc r2n1
4π th tc
Here, Q̇
r2 r1 1

k r1 r2 h0 r2 2
Since the only variable here is r2 , the denominator will be differentiated and equated to zero,
d r2 r1 1
i.e., 0
dr2 4πk r1 r2 4πr2 h0
2
d 1 1 1
and,
2
0
dr2 kr1 kr2 r h0
2
55
1 2
or, 0
kr22 r32 h0
3 2
or, r2 h0 2kr2
2k
and, r2 rc (1.54)
h0
Figure 1.34: Critical thickness of insulation for a sphere.
Example 1.11
A small electric heating application uses wire of 2 mm diameter with 0.8 mm thick insulation
`
(k = 0.12 W/m C). The heat transfer coefficient (h0 ) on the insulated surface is 35 W/m C.
2`
Determine the critical thickness of insulation in this case and the percentage change in the heat
transfer rate if the critical thickness is used, assuming the temperature difference between the
surface of the wire and surrounding air remains unchanged.
Solution:
Given:
2
r1 = = 1 mm = 0.001 m
2
r2 = 1 + 0.8 = 1.8 mm = 0.0018 m
`
k = 0.12 W/m C, h0 = 35 W/m C
2`
Critical thickness of insulation:

The critical radius of insulation is given by
k 0.12 3
rc 3.43 10 m or 3.43 mm.
h0 35

Critical thickness of insulation rc r1 3.43 1 2.43 mm.
Percentage change in heat transfer rate:

Case I : The heat flow through an insulated wire is given by eqn.(1.29)
56
2πx th tc
i.e., Q̇
1 ln r2 ©r1 ln r3 ©r2 1

hh r1 kA kB hc r3
For this case;

2πL t1 tair
Q̇1
ln r2 ©r1 1

k h0 r2
2πL t1 tair 2πL t1 tair

(i)
ln 0.0018©0.001 1 20.77

0.12 35 0.0018
Case II : The heat flow through an insulated wire when critical thickness is used is given by;
2πL t1 tair
Q̇2
ln rc ©r1 1

k h0 rc
2πL t1 tair 2πL t1 tair

(i)
ln 0.00343©0.001 1 18.6

0.12 35 0.00343

Percentage increases in heat flow by using critical thickness of insulation.
1 1

Q̇2 Q̇1 18.6 20.77
100 100 11.6%.
Q̇1 1
20.77
1.2.9 Transient conduction in one dimension

The equations for one-dimensional transient conduction, (1.41), (1.49), and (1.50), can be solved
using the separation of variables method. For example,from equation (1.41),
2
∂ t qg 1 ∂t
i.e.,
∂ x2 k α ∂τ
for the case when there is no internal heat generation is given by eqn. (1.48);
2
∂ t 1 ∂t
i.e.,
∂ x2 α ∂τ
solution to this differential equation it can be shown to be;

α pτ
t e rC1 sin px C2 cos pxx
57
where p, C1 and C2 are determined by the boundary conditions.
For an infinite slab of half-thickness, L, initially at a uniform temperature, ti , throughout, which
is suddenly exposed to a fluid at a constant temperature, tF , the temperature at any point, x, at
time, τ, is given by;

t tF
= e p,L sin pn L cos pn x
2
Fo
2 v | (1.55)
ti tF pn .L sin pn L cos pn L
n 1
and, pn L tan pn L Bi (1.56)
2 2
where Bi is the Biot number, αL/k and Fo the Fourier number, kτ ©ρc L or λ τ ©L .
Similar equations can be derived from the cases of the infinitely long cylinder and the sphere,
and graphs of non-dimensional temperature against Fourier number for various values of 1/ Bi
have been drawn (see for example reference 1.1).
1.2.10 Newtonian heating or cooling

This approach, which is sometimes known as lumped capacity, may be used when the temper-
ature within a body does not vary appreciably as the body’s average temperature changes with
time due to exposure of the body to a fluid at a different temperature. This is the case when the
surface thermal resistance is very much greater than the internal thermal resistance, and hence
the heat transfer from the surface is the controlling factor.
For a body of surface area, A, volume, V, specific heat, c, and density, ρ, with an average
temperature, t̄, at any time, τ, we have,
dt̄
α A t̄ tF ρVc
dτ
where α is the surface heat transfer coefficient and tF the temperature of the fluid surrounding
the body, assumed constant with time. Therefore,
Et E0 ρVc
t̄ d t̄ ταA
t tF dτ
i
where ti is the initial temperature of the body,
t̄ tF αAτ
i.e., ln t t

i F ρVc
t̄ tF αAτ ©ρV c
or ti tF e (1.57)
Equation (1.57) can also be written as,
t̄ tF AL©V BiFo
or ti tF e (1.58)
58
The dimension of length, L, may be the half-thickness of an infinite slab, or the radius of an
infinite cylinder, or the radius of a sphere. The term AL/ V for an infinite slab, cylinder or
sphere, may be shown to be 1, 2, or 3 respectively, e.g. for a sphere
2
AL 4πL L
V 4πL2 ©3
In the previously considered exact solution, equation (1.55), it can be shown that when Fo %
0.2 then only the first term of the summation need be considered within engineering accurancy.
Also, when Bi is small, then in equation (1.56) tan p1 L approximates to p1 L, and hence Bi
2
approximates to p1 L . Similarly, sin p1 L approximates to p1 L and cos p1 L approaches
unity. Therefore, substituting these approximations into equation (1.55), for the centre where x
= 0, and hence cos p1 x is 1, we have,
t tF BiFo p1 L BiFo
or 2e
e
ti tF p1 L p1 L
Comparing this with equation (1.58) it can be seen to be equivalent to Newtonian cooling of an
infinite slab, i.e. when Fo % 0.2 and Bi is very small the problem approximates to Newtonian
cooling.
Example 1.12
For transient conduction in a sphere when Fo % 0.2 it can be shown that the solution of equation
(1.50)
2
∂ t 2 ∂ t q̇g 1 ∂ t
i.e.,
∂ r2 r ∂ r k α ∂τ
for the temperature at the centre of the sphere, tc , when initially at ti , and plunged into a fluid at
tF , is given by,
tc tF p,L2 Fo sin p1 L p1 L cos p1 L
2e v |
ti tF p1 .L sin p1 L cos p1 L
and, 1 p1 L cot p1 L Bi
Using the data below determine the temperature at the centre of a sphere, initially at a uniform
`
temperature of 500 C, twenty minutes after it is plunged into a large bath ofliquid at a tempera-
`
ture of 20 C:
(a) from the above equation;
(b) assuming Newtonian cooling.
3
Data: Radius of sphere= 50 mm; density of sphere= 7600 kg/m ; thermal conductivity of sphere
= 40 W/m K; specific heat of sphere = 0.5 kJ /kg K; heat transfer coefficient from sphere surface
2
to liquid = 88.8 W /m K. It may be assumed that the heat transfer coefficient and the temperature
of the liquid remain constant over the time period.
Solution:
59
(a) Using transient equation for conduction;
Bi αL/k, , 88.8 0.05©40 0.111

, 1 p1 L cot p1 L 0.111
or, p1 L cot p1 L 0.889
This equation may be solved by trial and error,
p1 L 0.7 0.6 0.5

i.e
p1 L cot p1 L 0.831 0.877 0.915
By further trial and error, or by drawing a graph, it can be shown that p1 L = 0.57.
2 2 3 2
Also; Fo λ τ ©R kτ ©ρcR 40 20 60© 7600 0.5 10 0.05
5.053
tc tF 2
p,L Fo sin 0.57 0.57 cos 0.57 2
0.57 5,053
thus, e 2e
ti tF 0.57 sin 0.57 cos 0.57
tc tF
or, 0.5161 2 0.1936 0.2
ti tF
`
or, tc 20 0.2 500 20 l16 C
(b) assuming Newtonian cooling;

For Newtonian cooling of a sphere, from equation (1.58)
t̄ tF AL©V BiFo
i.e., ti tF e
t̄ tF 3BiFo 30.1115.053
and, e e AL/V 3
ti tF
`

t̄ 20 0.1859 500 20 109.2 C
60
1.2.11 Numerical methods for conduction
The most commonly used numerical method is the finite difference method in which a differential
equation is replaced by an approximate algebraic expression. The set of equations thus produced
can be solved using a computer. The reader is recommended to consult references 1.2 and 1.3
for a fuller treatment of numerical methods and their application in heat transfer.
A different method known as the finite element method is increasingly being used for heat trans-
fer applications, but is not considered in this book. In specialized texts the derivation of finite
difference expressions is given in detail, using for example the Taylor series, but in this book only
the following brief illustration will be given. Referring to a graph of t against x (Figure 1.35),
three approximations to the true tangent to the curve dt/dx are illustrated,
dt tx tx δ x
i.e., backward difference (1.59)
dx δx
dt txδ x tx
forward difference (1.60)
dx δx
dt txδ x txδ x
central difference (1.61)
dx δx
Figure 1.35: Diagrammatic defini-

tion of backward, forward, and cen-
tral difference approximations.
It can be seen from the Figure 1.35 that the central difference approximation, equation (1.61), is
a more accurate approximation to the true slope.
2 2
The second derivative, d t «dx , is the rate of change of slope at the point x. This may be
approximated as the change of dt/dx over the distance δ x. From Figure 1.35 it can be seen
that the slope at x + (δ x/2) is approximately txδ x tx ©δ x, and the slope at x δ x©2 is
approximately tx txδ x ©δ x. Hence the rate of change of slope over the distance δ x is given
by i.e.
2
dt txδ x tx tx txδ x 1
2
v |
dx δx δx δx
61
2
dt txδ x txδ x 2δ x
i.e., (1.62)
dx2 δ x2
To solve a conduction problem by the finite difference method the relevant partial differential
equation is replaced using expressions such as the above. The space and time dimensions are
divided into a number of increments of finite size and the approximate expression which replaces
the differential equation applies to every point in the grid of points, or nodes; separate equations
are derived for the boundary conditions. Hence the relevant differential equation is effectively
replaced by a large number of identical algebraic expressions for the temperature at each point
in the space at any time.
A set of simultaneous equations can be put in the form of a matrix and solved by matrix inversion
methods. However, in the case of conduction the matrix of temperature coefficients has a small
number of non-zero terms and hence matrix inversion is not recommended. It is better to solve
such a matrix by a direct method such as Gaussian elimination.
In the case of steady conduction in two dimensions the initial temperatures are unknown and
hence the set of equations is more conveniently solved by a relaxation method such as Gauss-
Siedel iteration.
Errors
Using a finite difference method the answer obtained converges towards the exact solution as
the size of the increments chosen approaches zero. Finite difference expressions must be chosen
such that the computer solution converges towards the exact solution; in certain cases the solution
will become unstable because errors generated are increasing in size as the solution proceeds, or
are growing at a faster rate than the rate of convergence.
There are basically two types of error: round-off error and discretization error. Round-off error
occurs when the answer is taken to a specific number of significant figures, and is cumulative;
fortunately in modern computers this error is not usually important. Discretization error is mainly
due to the inaccuracy of the finite difference expression, see Figure 1.35, and can be reduced by
reducing the size of the increments.
Notation
Referring to Figure 1.36, a two-dimensional space may be divided into a grid of nodes as shown.
The temperature at any point may then be designated as ti, j . Note that i increases from left to
right, and j from bottom to top, of the grid, following the normal x-direction and y-direction
respectively.
τ
For a problem in transient conduction the temperature at any time will be denoted by ti, j ; the
τ 1
next time is therefore τ + 1 and the temperature at that instant is ti, j . For transient problems in
one-dimension the j-direction will be omitted.
1.2.12 Two-dimensional steady conduction
62
Figure 1.36: Grid definition for
two-dimensional steady conduc-
tion.

2 2 2
∂ t ∂ t ∂ t q̇g 1 ∂t
where,
∂ x2 ∂ y2 ∂ z2 k α ∂τ
But, for zero internal heat generation and for steady conduction in two dimensions, the equation
reduces to the Laplace equation (1.47),
2 2
∂ t ∂ t
i.e., 0
∂ x2 ∂ y2
This equation may be put into finite difference form using the central difference expression,
equation (1.62). Using the notation outlined in above (see Figure 1.36), we have,
ti1, j ti1, j 2ti, j ti, j1 ti, j1 2ti, j

0
δ x2 δ y2
The grid may be chosen such that δ x = δ x, then
ti, j ti, j1 ti1, j ti, j1 ti1, j ©4 (1.63)
All the internal points within the boundaries of the two-dimensional space are represented by
equation (1.63).
Conducting rod analogy

Equation (1.63) may be derived using the basic Fourier equation and the concept of heat flow
paths. In Figure 1.37 conducting paths of width, δ x, from each point towards the centre point,
are shown cross-hatched. Fourier’s law can be applied to each conducting path; for example, the
heat transferred from point (i + l,j) to point (i,j) is given by
Thermal conductivity area temperature gradient
63
Figure 1.37: Conducting rod anal-
ogy for two-dimensional steady
conduction.
ti1, j ti, j
k δ x
δx
Then a simple energy balance gives,
ti1, j ti, j ti, j1 ti, j ti1, j ti, j ti, j1 ti, j
k δ x k δ x k δ x k δ x 0
δx δx δx δx
This equation reduces to the same expression as in equation (1.63).

The conducting rod analogy can be used in more complex cases, including the case with internal
heat generation or at a boundary convecting to a fluid; it may be found easier to apply since it
relates to a simple physical model.
Boundary conditions
Surface convecting to a fluid For a point (i,j) on the surface (see Figure 1.38):
dt
k α tF ti, j
dx
Using a central difference expression for dt/dx, equation eq:HTTFR56
dt txδ x txδ x
i.e.,
dx δx
ti1, j ti1, j
then, k α tF ti, j (1.64)
2δ x
The point (i - 1,j) is fictitious and can be eliminated from the equation by assuming that it lies on
the extrapolated temperature distribution line (see Figure 1.39), i.e. from equation eq:HTTFR58:
64
Figure 1.38: Grid for a left-hand
surface convecting to a fluid in two-
dimensional steady conduction.
ti1, j 4ti, j ti1, j ti, j1 ti, j1
Substituting in equation (1.64)

2αδ x
ti1, j 4ti, j ti1, j ti, j1 ti, j1 ti, j tF
k
Figure 1.39: Fictitious point

for a left-hand surface in two-
dimensional steady conduction.
2ti1, j ti, j1 ti, j1 2αδ x©k

ti, j (1.65)
4 2αδ x/k
Similar expressions may be obtained for a surface with the fluid on the right, or at the top or
bottom. Equation (1.65) can be derived using the conducting rod analogy with a rod of half-
width, δ x/2, running from point (i,j + 1) to point (i,j), and from point (i,j - 1) to (i,j),
ti1, j ti, j δ x ti, j1 ti, j δ x ti, j1 ti, j

i.e., k δ x k k αδ x ti, j tF
δx 2 δx 2 δx
65
Simplifying this equation the same expression as in equation (1.65) is obtained.
Insulated surface
At an insulated surface, - k dt/dx = 0, or α = 0, and hence for a left-hand surface which is
insulated equation (1.65) reduces to,
ti1, j ti, j1 ti, j1

ti, j
4
It should be noted that a line of thermal symmetry within a two-dimensional space will act as an
insulated surface .
Corners
Expressions can be derived for the temperatures at outside corners (top left, bottom right, etc.)
and at inside corners. For example, for a top left outside corner
ti1, j ti, j1 2αδ xtF ©k

ti, j
2 2αδ x©k
For a bottom right inside corner
ti1, j ti, j1 ti1, j 2αδ xtF ©k

ti, j
6 2αδ x©k
The derivation of expressions for corner points such as the above is left as an exercise for the
reader; the conducting rod analogy is the best method, particularly for inside corners.
1.3 Heat Transfer by Convection

The rate equation for the convective heat transfer (regardless of particular nature) between a sur-
face and an adjacent fluid is prescribed by Newton’s law of cooling (Figure 1.39) and governed
by equation (1.17)
i.e., Q̇ αA ts t f
where, Q̇ Rate of conductive heat transfer,

A = Area exposed to heat transfer,
ts Surface temperature,
t f Fluid temperature, and
α Co-efficient of conductive heat transfer.
The coefficient of convective heat transfer ‘α’ (also known as film heat transfer coefficient) may
be defined as “the amount of heat transmitted for a unit temperature difference between the fluid
and unit area of surface in unit time.”
The value of ‘α’ depends on the following factors:
(i) Thermodynamic and transport properties (e.g., viscosity, density, specific heat etc.) ;
66
(ii) Nature of fluid flow ;
(iii) Geometry of the surface ;
(iv) Prevailing thermal conditions.
Since ‘α’ depends upon several factors, it is difficult to frame a single equation to satisfy all the
variations, however a dimensional analysis gives an equation for the purpose which is given as
under:
kD ρCD a cµ b D c
Z π

(1.66)
λ λ L
c
a b D
or,Nu Z Re Pr
kD
where, Nu = Nusselt number,
λ
ρ ūD
Re = Reynolds number, π
cµ
Pr = Prandtl number,
λ
D
Diameter to length ratio,
L
Z = A constant to be determined experimentally,
ρ Density,
µ Dynamic viscosity, and
ū mean Velocity
C = Velocity
c = specific heat capacity.
The mechanisms of convection in which

phase changes are involved lead to the
important fields of boiling and condensa-
tion.
Referring to Figure 1.40 (b).
1
The quantity,
kA
ts t f
From equation(1.17) i.e. Q̇ ,
1©αA Figure 1.40: Convective heat-transfer.
is called convection thermal resistance [ Rth conv. ]

to heat flow.
Other dimensionless numbers used in heat

transfer:
67
α Nu
Stanton number, St =
ρVc Re pr
uL
Peclet number, Pe = σ Re Pr
u is fluid velocity, L is a characteristic dimension, and σ is thermal diffusivity
πD
Graetz number, G = Pe
4
2 3
ρ β g∆tL
Grashoff number, Gr =
µ2
1.3.1 Forced convection

The study of forced convection is concerned with the transfer of heat between a moving fluid
and a solid surface. In order to apply Newton’s law of cooling, given by equation (1.17), it
is necessary to find a value for the heat transfer coefficient, α. It is stated that α is given by
k
, where k, is the thermal conductivity of the fluid and δ is the thickness of the fluid film on
δ
the surface. The problem is then to find a value for δ in terms of the fluid properties and the
fluid velocity; δ depends on the type of fluid flow across the surface and this is governed by the
Reynolds number.
The various kinds of forced convection, such as flow in a tube, flow across a tube, flow across
a flat plate, etc. can be solved mathematically when certain assumptions are made with regard
to the boundary conditions. It is exceedingly difficult to obtain an exact mathematical solution
to such problems, particularly in the case of turbulent flow, but approximate solutions can be
obtained by making suitable assumptions.
It is not within the scope of this study to approach the subject of forced convection fundamentally.
However, many of the results used in heat transfer are derived from experiment, and in fact for
many problems, no mathematical solution is available and empirical values are essential. These
empirical values can be generalized using dimensional analysis, which will now be considered.
Dimensional analysis
In order to apply dimensional analysis it is necessary to know from experience all the variables
upon which the desired function depends. The results must apply to geometrically similar bodies,
therefore one of the variables must always be a characteristic linear dimension.
Consider the dimensional analysis for forced convection, assuming that the effects of free con-
vection, due to differences in density, may be neglected. It is found that the heat transfer coef-
ficient, α, depends on the fluid viscosity, µ, the fluid density, ρ, the thermal conductivity of the
fluid, ,k, the specific heat capacity of the fluid, c, the temperature difference between the surface
and the fluid, ∆, and the fluid velocity, C. Therefore we have,
α f µ, ρ, k, c,∆t, C, l (1.67)
68
where l is a characteristic linear dimension and f is some function. Equation eq:HTTFR62 can
be written as follows:
a b c d e f g1 a2 b c d e f g2
α Aµ 1 , ρ 1 , k 1 , c 1 , ∆t 1 , C 1 , l Bµ , ρ 2 , k 2 , c 2 , ∆t 2 ,C 2 , l etc. (1.68)
where A and B are constants, and a1 , b1 , c1 , etc. are arbitrary indices.

Each term on the right-hand side of the equation must have the same dimensions as the dimen-
sions of α. Considering the first term only, we can write,
a b c d e f g
Dimensions ofα dimensions of (Aµ 1 , ρ 1 , k 1 , c 1 , ∆t 1 ,C 1 , l 1
Each of the properties in the equation can be expressed in terms of five fundamental dimensions;
these are mass, M, length, L, time, T, temperature, t, and heat, H.
W H
Forαthe units are , i.e.,
m2 K L2 Tt
kg M
Forµthe units are m s , i.e.,
LT
W H
For k the units are , i.e.,
mK LTt
kg M
Forρthe units are , i.e.,
m3 L3
kJ H
For c the units are , i.e.,
kg K Mt
For∆t the units are K, i.e., t
m L
For C the units are s , i.e.,
T
For l the units are m, i.e., L
Hence, substituting,
H M a M b H c H d e L
f
g

, 3
,
,
, t ,
, L
L2 Tt LT L LTt Mt T
H abd f ga3bc ac f ecd cd

i.e., M L T t H
L2 Tt
For the dimensions of each side of the equation to be the same, the power to which each funda-
mental dimension is raised must be the same on both sides of the equation. Therefore, equating
indices we have,
69
For H: l = c+d
For L: -2 = f + g - a - 3b - c
For T: -1 = -a - c - f
For t: -1 = e - c - d
For M: O=a+b-d
We have five equations and seven unknowns, therefore a solution can only be obtained in terms
of two of the indices. It is most useful to express a, b, c, e, and g in terms of d and f Then it can
be shown that,
a = (d - f); b = f; c = (1 - d); e = 0; g = (f - 1)
Substituting these values in equation (1.68), we have,
d1 f1 f 1d1 d 0 f f1 1 d2 f2 f 1d2 d 0 f f2 1

α Aµ , ρ 1, k , c 1 , ∆t ,C 1 , l Bµ , ρ 2, k , c 2 , ∆t ,C 2 , l etc.
k µc d1 ρCl f1 k µc d2 ρCl f2
i.e., α A µ
B µ
etc.
l k l k
Hence it can be seen that,
αl µc , ρCl
KF v
µ
|
k k
where K is a constant and F is some function.

Example 1.13
A wire 1.5 mm in diameter and 150 mm long is submerged in water at atmospheric pressure. An
`
electric current is passed through the wire and is increased until the water boils at 100 C. Under
2`
these condition the convective heat transfer coefficient is 4500 W/ C Evaluate the quantity of
electric power that must be supplied to the wire to maintain the wire surface at 120 C.
`
Solution:
Diameter of the wire, d = 1.5 mm = 0.0015 m,
Length of the wire, l = 150 mm = 0.15 m
Surface area of the wire (exposed to heat transfer),
4 2
A πdl π 0.0015 0.15 7.068 10 m
Wire surface temperature, ts = 120 C

`
`
Water temperature, tf = 100 C
2`
Convective heat transfer coefficient, h = 4500 W/m C
Electric power to be supplied:
Electric power which must be supplied = Total convection loss (Q̇)
4

Q̇ αA ts t f 4500 7.068 10 120 100 63.6 W.
Example 1.14
70
Calculate the heat transfer coefficient for water flowing through a 25 mm diameter tube at the
`
rate of 1.5 kg/s, when the mean bulk temperature is 40 C. For turbulent flow of a liquid take,
0.8 0.4
Nu 0.0243Re Pr
where the characteristic dimension of length is the tube diameter and all properties are evaluated
at mean bulk temperature.
Solution:
First it is necessary to ascertain whether the flow is turbulent or laminar. For flow through a tube
it can be assumed that the flow is turbulent when Re % 2100 approximately. The properties of
water can be taken from the steam tables.
Then,
3
Volume flow 1.5 v f 1.5 0.001 0.0015 m ©s
0.0015 4
i.e., Velocity in tube, C 3.06 m/s
π 0.0252
ρCd Cd 3.06 0.025

Re 117500
µ vf µ 0.001 651 106
The flow is therefore well into the turbulent region and the formula given for turbulent flow can
be applied.
From tables, Pr = 4.3, hence substituting,
0.8 0.4
Nu 0.0243 117500 4.3
0.0243 11377 1.792 495.5
αd
Nu 495.5
k
495.5 632 10
6
2

α 12.53 kW©m K
0.025
2
i.e., Heat transfer coefficient 12.53 kW /m K
For laminar flow in a tube an exact mathematical solution has been found; this gives Nu = 3.65.
It can be seen that, since Nu αd©k 3.65, the heat transfer coefficient, α, for any one tube,
depends only on the thermal conductivity of the fluid.
In the foregoing dimensional analysis five fundamental dimensions, heat H, length L, time T,
temperature t, and mass M, were chosen. The units of work, or energy in general, are given by,
Force distance mass acceleration distance

2
L ML
M 2L
T T2
71
Since heat is a form of energy it can be seen that there is no need to choose heat as one of the
fundamental dimensions. If the dimension, H, is omitted, and the units of heat are replaced by
2 2
ML /T whenever they occur, then four dimensionless groups are obtained from the dymensional
analysis done earlier,
2
C
i.e., Nu = KF w(Pr), (Re), }
c∆t
2
C
Now if the group is divided by (γ - 1), which is a constant for any one gas, and if ∆t is
c∆t
replaced by the absolute bulk temperature of the gas, T, then we have,
2 2
C C2 C 2
a Ma R = cv γ 1
cT γ 1 γRT
where a is the velocity of sound in the gas and Ma the Mach number (discussed in study of
nozzles)
2
Hence, Nu K’F Pr, Re, Ma
where K’ is another constant.

The influence of the Mach number, Ma, on the heat transfer is negligible for most problems.
For high-speed flow however, large amounts of kinetic energy are dissipated by friction in the
boundary layer near the surface, and the Mach number becomes an important parameter.
Reynolds analogy
Reynolds postulated that the heat transfer from a solid surface is similar to the transfer of fluid
momentum from the surface, and hence that it is possible to express the heat transfer in terms of
the frictional resistance to the flow.
Consider turbulent flow. It can be assumed that particles of mass, m, transport heat and momen-
tum to and from the surface, moving perpendicular to the surface. Then on the average,
Heat transferred per unit area,q̇ ṁc∆t
where c is the specific heat capacity of the fluid and ∆ t the temperature difference between the
surface and the bulk of the fluid. Also, the rate of change of momentum across the stream is
given by,
m C Cw ṁC
where C is the velocity of the bulk of the fluid and Cw the fluid velocity at the surface = 0. Then,
Force per unit area τw ṁC
where τw is the shear stress in the fluid at the wall.

Combining the equations for heat flow and momentum transfer,
72
q̇ τw
then,
c∆t C
τw c∆t
or, q̇ (1.69)
C
For turbulent flow in practice there is always a thin layer of fluid on the surface in which viscous
effects predominate. This film is known as the laminar sublayer. In this layer heat is transferred
purely by conduction. Therefore, from Fourier’s law, for unit area;
dt
q̇ k v |
dy y 0
where k is the thermal conductivity of the fluid and y the distance from the surface perpendicular
to the surface. Also, for viscous flow,
Shear stress,τ µ velocity gradient
Hence the shear stress at the wall is given by,
dC
τw µ
dy y 0
where µ is the fluid viscosity and C the fluid velocity.

Now since the laminar sublayer is very thin it may be assumed that the temperature and velocity
vary linearly with the distance from the wall, y,
∆t C
i.e., q̇ k and τw µ
δb δb
where δb is the thickness of the laminar sublayer.

Then eliminating δb , and neglecting the minus sign, we have
q̇ τw
k∆t µC
τw ∆t
i.e., q̇
µC
It can be seen that this equation is identical with equation (1.69) when,
k
c µ
i.e., when,
cµ
1 or Pr 1
k
73
Therefore for fluids whose Prandtl number is approximately unity the simple Reynolds analogy
can be applied, since .the heat transferred across the laminar sublayer can be considered in a
similar way to the heat transferred from the sublayer to the bulk of the fluid. For most gases, dry
vapours, and superheated vapours Pr lies between about 0.65 and 1.2.
For unit surface area, q̇ α∆t, therefore substituting in equation (1.69), we have,
α τw
c C
Dividing through by ρC, where ρ is the mean density of the fluid, we have,
α τw
ρcC ρC2
Both sides of this equation are dimensionless. The term on the left-hand side is called the Stanton
number, St,
α
i.e., St (1.70)
ρCc
A dimensionless friction factor, f, is defined by,
τw
f (1.71)
ρC2 ©2
Therefore we have for the Reynolds analogy,
f
St (1.72)
2
The Stanton number, St, can be written as,

α αl µ k Nu
St
ρCc k ρCl cµ RePr
Nu
i.e., St (1.73)
RePr
The friction factor, f, can be derived mathematically for some cases, but in other cases a practical
determination is necessary. For turbulent flow in a pipe a simple measurement of the pressure
drop gives f, and then, using equation (1.69) or equation (1.72), the approximate heat flow can
be found.
For flow in a pipe of diameter, d, the resistance to flow over unit length is given by,
π 2
Resistance τw πd ∆p d
4
where ∆p is the pressure drop in unit length.
∆pd
i.e., τw (1.74)
4
74
An important factor in heat exchanger design is the pumping power required. The pumping
power is the rate at which work is done in overcoming the frictional resistance, i.e. for flow in a
pipe,
Pumping power per unit length,Ẇ τw πdC
Also, from equation (1.69),
τw c∆t
Heat flow per umt area,q̇
C
τw c∆tπd
Heat flow per umt length,Q̇
C
Then the ratio of the pumping power, Ẇ, to the rate of heat flow, Q̇, can be expressed as,
2
Ẇ τw πdCC C
(1.75)
Q̇ τw c∆tπd c∆t
(for a heat exchanger, ∆t is the log mean temperature difference, ∆t̄in , - to be shown in later
study).
It can be seen from equation (1.75) that the power required for a given heat transfer rate can be
reduced by decreasing the velocity of flow, C. However, a reduction in fluid velocity means that
the required surface area must be increased, and hence a compromise must be made.
Example 1.15
Water flows inside a tube 45 mm in diameter and 3.2 m long at a velocity of 0.78 m/s. Given
` `
the following data: the mean water temperature = 50 C and the wall is isothermal at 70 C. For
` 6 2
water at 50 C, k = 0.66 W/mK, kinematic viscosity, ν 0.478 10 m ©s and Prandtl number
= 2.98. Determine;
(a) the heat transfer co-efficient, and
(b) the rate of heat transfer.
Solution:
Given:
Diameter of the tube, D = 45 mm = 0.045 m
Length of the tube, l = 3.2 m
Velocity of water, ū = 0.78 m/s
`
For water at 60 C, k = 0.66 W/mK
6 2
Kinematic viscosity, ν 0.478 10 m ©s
Pr = 2.98
(a) the heat transfer co-efficient;
Reynolds number (Re), is given by,
Dū 0.045 0.78

Re 73431
ν 0.478 106
75
From Dittus and Boelter equation, Nusselt number,
αD 0.8 0.4
Nu 0.023 Re Pr
k
0.8 0.4
0.023 73431 2.98
α 0045
and, 0.023 7810.9 1.547
0.66
2
4076 W/m K
2
thus, Heat transfer co-efficient 4076W/m K
(b) the rate of heat transfer,Q̇

Q̇ αA (tw t f
4076 π DL 70 50
4076 π 0.045 3.2 20 36878 or 36.878 kW
i.e., Rate of heat transfer 36.878 kW.
Example 1.16
When 0.5 kg of water per minute is passed through a tube of 20 mm diameter, it is found to be
` `
heated from 20 C to 50 C. The heating is accomplished by condensing steam on the surface of
`
the tube and subsequently the surface temperature of the tube is maintained at 85 C. Determine
the length of the tube required for developed flow.
` 2
Take the thermo-physical properties of water at 60 C as: Density, ρ = 983.2 kg/m , c p = 4.178
` 6 2
kJ/kg K, k = 0.659 W/m C, kinematic viscosity, ν 0.478 10 m ©s.
Solution:
Given:
` `
m = 0.5 kg/min, D = 20 mm = 0.02 m, ti = 20 C, t0 = 50 C.
Length of the tube required for fully developed flow, L:
1 20 50 `
The mean film temperature, t f 85
60 C
2 2
First, the type of the flow should be determined;
π 2 0.5
m ρAu̇ 983.2 0.02 u̇ kg/s
4 60
0.5 4 1
and, u̇ 0.0269 m/s
60 π 983.2 0.022
Dū 0.02 0.0269

Then, Reynolds number, Re Re 1125.5
ν 0.478 106
Since Re $ 2000, hence the flow is laminar. With constant wall temperature having fully devel-
oped flow,
76
αD
Nu 3.65
k
3.65k 3.65 0.659 2`

and, α 120.26 W/m C
D 0.02
The rate of heat transfer, Q As α ts t mc p to ti
20 50 `
Here, t 35 C tb
2
0.5

π 0.02 L 120.26 85 35 4.178 103 50 20
60
or, 377.8L 1044.5
1044.5
and, L 2.76 m.
377.8
Example 1.17
In a 25 mm diameter tube the pressure drop per metre length is 0.0002 bar at a section where
the mean velocity is 24 m/s, and the mean specific heat capacity of the gas is 1.13 kJ /kg K.
Calculate the heat transfer coefficient.
Solution:
For a 1 m length
∆p 0.0002 bar
∆pd
i.e., τw
4
5
10 0.0002 25 2
and, τw 0.125 N/m
4 103
Then from equation (1.71)

τw
i.e., f
ρC2 ©2
2 0.125
and, f (i)
ρC2
Also, from equation (1.72)

f α
i.e., St (ii)
2 ρCc
Substituting for f from eqn.(i) and equating the two terms in St. no. equation in eqn.(ii),
2 0.125 α
2ρC2 ρCc
77
0.125ρcC 0.125 1.13

α
ρC2 24
2
0.00588 kW/m K
2
and, Heat transfer coefficient 5.88 W/m K
Various modifications have been made to the simple Reynolds analogy in an attempt to obtain
an equation which will give a solution for turbulent heat transfer over a wide range of Prandtl
numbers. (For very viscous oil the Prandtl number is of the order of thousands, whereas for
liquid metals it may be as low as 0.01). Equations based on modern theories of turbulent flow
give the Stanton number as a function of the Reynolds number, the Prandtl number, and the
friction factor, and in general these equations reduce to St = f/2, when the Prandtl number is put
equal to unity (see for example references 1.1, 1.3, and 1.5). Colburn found experimentally that
for a wide range of Prandtl numbers,
2©3
StPr f©2
2©3
The term, StPr , is known as the Colburn j-factor.
Large temperature differences

When the temperature difference between the surface and the bulk of the fluid is very high, then
the property variations become large enough to be taken into consideration. It is then no longer
sufficient to use a mean film temperature to evaluate the properties, as given by equation (??),
tb tw
i.e., tf . The variation of each property with temperature across the stream
2
must be known; sometimes it is sufficiently accurate to use an equation of the form,
Ts
Nu= Kφ v Pr, Re, |
Tw
where Ts . and Tw are the absolute temperatures at the axis of the pipe and at the pipe wall
respectively, and fluid properties are taken at the mean film temperature.
Entry length
The equations for flow in a pipe do not usually allow for the effects of the entry length. At the
entry to a heated tube the hydrodynamic and thermal boundary layers start to build up on the wall,
gradually thickening until the flow becomes fully developed. In this initial region of the tube the
heat transfer coefficient is much larger since the resistance to heat flow of the boundary layer is
less, and hence an equation which neglects this effect will give a low value for the calculated
heat transfer. The effect is more marked for laminar flow than for turbulent flow, and is much
more important for fluids with high Prandtl numbers. In most heat exchange processes the flow
is turbulent and the tube length is sufficiently long to make the entry length effect negligibly
78
small. In the case of oil coolers the flow is laminar, the Prandtl number is high, and hence the
entry effect may be appreciable.
When flow across a flat plate is considered, the characteristic dimension of length is taken as the
distance from the leading edge, and the heat transfer coefficient obtained is then the local value
at that section of the plate. The average value of the heat transfer coefficient over the whole
plate is the value to be used in calculating the heat transfer to or from the plate. It can be shown
that the average heat transfer coefficient for a heated plate over a length l is twice the local heat
transfer coefficient at a distance l from the leading edge.
Example 1.18
`
Air at 20 C, flowing at 25 m/s, passes over a flat plate, the surface of which is maintained at
`
270 C. If the heat transfer from a flat plate is governed by,
1©3 1©2
Nu 0.332 Pr Re
where the characteristic linear dimension is the distance from the leading edge, and all properties
are evaluated at mean film temperature, calculate the rate at which heat is transferred per metre
width from both sides of the plate over a distance of 0.25 m from the leading edge.
Solution:
20 270 `
Mean film temperature 145 C 418 K
2
Taking the values from tables of properties of air, we have
5
Cl 25 0.25 10
Pr 0.687 and Re 223214
ν 2.8
1©3 1©2
Then, Nu 0.332 0.687 223214
0.332 0.882 472.5 138.4
αl
but,
k
138.4 3.49 2
and, α 0.0193 kW/m K
0.25 105
Hence the average heat transfer coefficient is,
3 2
0.0193 10 2 38.6 W/m K
Then the rate of heat transfer from both sides of the plate over the length of 0.25 m for 1 m width
is given by
Q̇ αA∆t 38.6 0.25 1 2 270 20 4825 W
and, Rate of heat transfer 4.825 kW
79
The friction loss for the initial length of a flat plate where the boundary layer is still laminar is
given by
1©2
f 0.664 Re
Hence it can be seen that the simple Reynolds analogy, given by equation (1.72), i.e., St
f
, gives for the initial length of a flat plate,
2
1©2
St 0.332 Re
Nu 1©2
Also, 0.332 Re
PrRe
i.e., Nu 0.332 Pr Re1©2
This is the same as the immediate equation given in example given above if the Prandtl number
is unity.
1.3.2 Natural convection

As stated previously, heat transfer by free or natural convection is due to differences in density in
the fluid causing a natural circulation, and hence a transfer of heat. For the majority of problems
in which a fluid flows across a surface, the superimposed effect of natural convection is small
enough to be neglected. When there is no forced velocity of the fluid then the heat is trans-
ferred entirely by natural convection (when radiation is negligible). The heat transfer in this case
depends on the coefficient of cubical expansion, β , which is given by,
ρ1 ρ2 l β ∆t or ρ1 ρ2 ρ2 β ∆t
where ∆t is the temperature difference between the two parts of the fluid of density ρ1 and ρ2
The upthrust per unit volume of fluid is ρ1 ρ2 g, and the velocity of the convection current is
dependent on the upthrust, i.e. natural convection depends on,
ρ1 ρ2 g ρ2 β ∆tg
The heat transfer also depends on the fluid viscosity, the thermal conductivity of the fluid, and a
characteristic dimension of length. Since the coefficient of cubical expansion, β , and the local
acceleration due to gravity, g, do not have a separate effect on the heat transfer, then only the
product, β g, need be considered. For a dimensional analysis we then have,
a b c d e f g1 a2 b c d e f g2
α Aµ 1 , ρ 1 , k 1 , c 1 , ∆t 1 , β g 1 , l Bµ , ρ 2 , k 2 , c 2 , ∆t 2 , β g 2 , l etc.
Then by the same procedure of dimension analysis as in then previous work, it can be shown that
2 3
µc β gρ l ∆t
Nu KF w , }
k µ2
80
.
or, Nu KF r Pr , Grx
2 3 3
β gρ l ∆t β gl ∆t
where, Gr
µ2 ν2
Gr is the Grashof number.

In many cases of natural convection it is possible to use an approximate equation to evaluate the
heat transfer coefficient, α. For example, for natural convection from a horizontal pipe,
∆t 1©4
α 1.32
d

when 10
4
$ Gr $ 109
and α 1.25∆t
1©3
when 10
9
$ Gr $ 1012
2
where α is in W/m K, ∆t is in K, and d is in m.
Example 1.19
A horizontal pipe of 150 mm diameter has a surface of which is at 277 C. The room temperature
`
`
is 17 C. It is known that for a horizontal cylinder,
1©4
Nu 0.53 GrPr
where the properties are evaluated at the mean film temperature. Taking the coefficient of cubical
expansion, β , as 1/T, where T is the absolute temperature of the air, calculate:
(a) the rate of heat loss by natural convection per unit length;
(b) the heat transfer coefficient using the approximate equation,
∆t 1©4
α 1.32
d

for 10
4
$ Gr $ 109
2
Solution:
(a) the rate of heat loss by natural convection:
From tables, at a mean film temperature of (550 + 290)/2 = 420 K, we have, Pr = 0.686 and,
3
β gd ∆t
Gr
ν2
3
9.81 277 17 0.15 6
37.27 10
290 2.822 105 2
2 1 1 1
Note, that g 9.81 m/s , and β K
17 273 290
81
Substituting,
6 1©4
Nu 0.53 37.27 10 0.686
37.7
αd
also, Nu 37.7
k
37.7 3.635 10
5
2
and, α 0.00913 kW/m K
0.15
Then from equation eq:HTTFR31
i.e., Q̇ αA tw t
then, Q̇ 0.00913 π 0.15 1 277 17 1.119 kW

Heat loss per metre length 1.119 W
(b) the heat transfer coefficient:
277 17 1©4 1©4 2

α 1.32
1.32 1733 1.32 6.45 8.52 W/m K
0.15
2
thus, Heat transfer coefficient 8.52W/m K
2
( compared with the more accurate value, 9.13 W/m K).
For natural convection from a vertical wall the air in rising due to the convection currents builds
up a boundary layer, starting from the bottom and thickening gradually up the wall. The heat
transfer coefficient varies up the wall, and the formulae for heat transfer from a vertical wall
give the local heat transfer coefficient at a distance, l, from the bottom of the wall, where the
characteristic linear dimension to be used in the Grashof number is the length, l.
It can be shown that the average heat transfer coefficient for the wall from the bottom up to the
distance, l, is given by,
4
αav α
3
whereαav is the average heat transfer coefficient, and α the heat transfer coefficient at the section
distance, l, from the bottom of the wall.
Example 1.20
82
`
A vertical surface 1 m high is at a temperature of 627 C, and the atmospheric temperature is
`
27 C. Calculate the rate at which heat is lost by convection from the surface per metre width.
For natural convection from a vertical surface take,
∆t 1©4
α 1.42
l
for 10
4
$ Gr $ 109
or α 1.31∆t
1©3
for 10
9
$ Gr $ 1012
where all properties are at the mean film temperature, and β = 1/T, where T is the absolute air
2
temperature, α is in W/m K, ∆t in K, and l in m.
Solution:
The Grashof number in such problems has the same limiting function as the Reynolds number
in fluid flow. For the lower range of Grashof numbers the flow of air due to natural convection
remains laminar on the wall surface, whereas for the larger Grashof numbers the boundary layer
1©3
on the wall is turbulent. It can be seen from the second equation above, α 1.31∆t , that when
the boundary layer is turbulent the heat transfer coefficient is assumed to be the same at all parts
of the wall, since α no longer depends on,the distance, l.
Mean film temperature 900 300©2 600 K
Taking properties from tables, we have,
3 3
β gl ∆t 1 9.81 l 327 30 9
Gr 3.65 10
ν2 303 5.128 105 2
1 1
where, β
30 273 303
1©3 1©3
Hence α 1.31∆t 1.31 327 30
2
1.31 6.67 8.75 W/m K
and, Q̇ αA∆t 8.75 1 1 627 27 5250 W
thus, Rate of heat loss per metre width 5.25 kW
Note: expressions for α as above give average values.
ASSIGNMENT QUESTIONS
Question ONE
A steam main of 150 mm outside diameter containing wet steam at 28 bar is insulated with an
inner layer of diatomaceous earth, 40 mm thick, and an outer layer of 85% magnesia, 25 mm
thick. The inside surface of the pipe is at the steam temperature, and the heat transfer coefficient
2
for the outside surface of the lagging is 17 W /m K. The thermal conductivities of diatomaceous
earth and 85% magnesia are 0.09, and 0.06 W/mK respectively.
Neglecting radiation, and the thermal resistance of the pipe wall, calculate
(a) the rate of heat loss per unit length of the pipe, and
83
(b) the temperature of the outside surface of the lagging, when the room temperature is 20 C.
`
[Answers: (a) 0.156 kW /m; (b) 30.5 C]

`
Question TWO
The temperature-time history of the centre of a large slab of material initially at a constant tem-
perature which is suddenly plunged into a fluid at a different temperature can be shown to be
given by

∆te
∆t1
=
2 e n
p L2 Fo sin pn L
pn L sin pn Lcos pn L
n 1
and pn L tan pn L Bi. Where ∆te is the temperature difference at time τ, between the centre
of the slab thickness and the surrounding fluid; ∆t1 the temperature difference between slab
2
and fluid initially; Fo the Fourier number, Kτ ©L ; σ the thermal diffusivity of slab; L the half-
thickness of slab; Bi the Biot number, αL©λ ; α the heat transfer coefficient on slab surfaces;
and k the thermal conductivity of slab material.
Show that for a case where the temperature of the slab surfaces is approximately equal to the
fluid temperature (i.e. α ), and for a reasonably long time period
∆te 4 2
π ©4Fo
∆t1 πe
Tutorial
1. 16.1 A furnace wall consists of 250 mm firebrick, 125 mm insulating brick, and 250 mm
`
building brick. The inside wall is at a temperature of 600 C and the atmospheric temper-
` 2
ature is 20 C. The heat transfer coefficient for the outside surface is 10 W/m K, and the
thermal conductivities of the firebrick, insulating brick, and building brick are 1.4, 0.2, and
0.7 W/m K, respectively. Neglecting radiation, Calculate:
(a) the rate of heat loss per unit wall surface area, and
(b) the temperature of the outside wall surface of the furnace.
2
[Answers:(a) 460W/m ; (b) 66 C]
`
`
2. An electric hot-plate is maintained at a temperature of 350 C and is used to keep a solution
`
just boiling at 95 C. The solution is contained in an enamelled cast-iron vessel of wall
thickness 25 mm and enamel thickness 0.8 mm. The heat transfer coefficient for the boiling
2
solution is 5.5 kW /m K, and the thermal conductivities of cast iron and enamel are 50 and
1.05 W/m K respectively. Calculate:
(a) the resistance to the heat transfer for unit area, and the rate of heat transfer per unit
area.
(b) the rate of heat transfer per unit area if the base of the cast-iron vessel is not perfectly
2
flat, and the resistance of the resultant air film is 35m K/kW.
2 2
[Answers:(a) 1.444 m K/kW; 176.6 kW; (b) 7 kW/m ]
3. The wall of a house consists of two 125 mm thick brick walls with an inner cavity. The
inside wall has a 10 mm coating of plaster, and there is a cement rendering of 5 mm on
the outside wall. In one room of the house the external wall is 4 m by 2.5 m, and contains
a window of 1.8 m by 1.2 m of 1.5 mm thick glass. The heat transfer coefficients for the
84
2
inside and outside surfaces of the wall and window are 8.5 and 31 W /m K respectively.
The thermal conductivities of brick, plaster, cement, and glass are 0.43, 0.14, 0.86, and
2
0.76 W/m K respectively. Assuming that the resistance of the air cavity is 0.16 m K/W,
neglecting all end effects, and neglecting radiation, calculate the proportion of the total
heat transfer which is due to the heat loss through the window. [Answer:63.8%]
`
4. Water at 80 C flows through a 50 mm bore steel pipe of 6 mm thickness, and the atmo-
`
spheric temperature is 15 C. The thermal conductivity of steel is 48 W/m K and the inside
2
and outside heat transfer coefficients are 2800 and 17 W /m K respectively.
Neglecting radiation, calculate:
(a) the rate of heat loss per unit length of pipe.
(b) the percentage reduction in heat loss for the pipe when a layer of hair felt 12 mm
thick, of thermal conductivity 0.03 W/mK, is wrapped round the outside surface.
Assume that the heat transfer coefficient for the outside surface remains unchanged.
[Answers: (a) 0.213 kW/m; (b) 85.1%]
5. A spherical pressure vessel of 1 m inside diameter is made of 20 mm steel plate. The

vessel is lagged with a 25 mm thickness of vermiculite held in position by 10 mm thick
2
asbestos. The heat transfer coefficient for the outside surface is 20 W /m K, and the
thermal conductivities of steel, vermiculite, and asbestos are 48, 0.047, and 0.21 W/m K,
respectively.
Neglecting radiation, calculate the rate of heat loss from the sphere when the inside surface
`
is at 500 C, and the room temperature is 20 C.
`
[Answer:2.744kW]
2
6. A solid copper conductor of 13 mm diameter carries a current density of 5 A/mm . The
conductor is electrically insulated with a thickness of rubber insulation such that the wire
temperature is kept to the minimum possible. Assuming that the surrounding air is at 30 C
`
and given the following data: Data: Heat transfer coefficient for outside surface of rubber
2
or copper (assumed constant), 20 W/m K; thermal conductivities of copper and rubber,
5
380 and 0.2 W/m K; electrical resistivity of copper, 2 10 Ω mm,
calculate:
(i) the thickness of insulation;
(ii) the wire temperature at the axis;
(iii) the temperature of the outside surface of the insulation;
(iv) the wire temperature at the axis with the insulation removed and the new steady state
reached.
Give a physical explanation why any larger or smaller thickness of insulation will lead to
a higher wire temperature.
` ` `
[Answers: (a) 3.5 mm; (b) 105.6 C; (c) 82.8 C; (d) 111.3 C ]
7. (a) A gas-cooled nuclear reactor has solid fuel rods of radius, rF , thermal conductivity,
,kF , sheathed with zirconium of thickness, t, thermal conductivity, kz . The heat trans-
fer coefficient from the sheath to the gas in the surrounding annulus is α.
Assuming that the uniform heat generation rate per unit volume within the fuel is q̇g ,
show from first principles that the temperature difference between the axis of the fuel
85
rod and the bulk of the coolant at any cross-section is given by,
2
q̇rF 2kF 2kF
v1 ln 1 t©rF |
4kF kz α rF t
(b) At a particular cross-section in the reactor the gas mean bulk temperature is 220 C
`
3
when the internal heat generation rate in the fuel rod is 30 MW/m
Using the data below, calculate:
(i) the temperature at the axis of the fuel rod;
(ii) the temperature at the inner and outer surfaces of the sheath. Data Thermal con-
ductivities of reactor fuel and zirconium, 33.5 and 18.7 W/m K; radius of fuel
rod, 12 mm; thickness of sheath, 3.6 mm; heat transfer coefficient from sheath
2
to cooling gas, 500 W/m K.
` `
[Answers: (a) 559.5 C; (b) 527.3 C; 496.9 C]
`
8. The concrete biological shield of a nuclear reactor is 2 m thick and can be considered to
be an infinite flat plate of uniform thermal conductivity, ,k. = 2.0 W/m K. The rate of heat
generation per unit volume due to the incident gamma radiation is given by
8.5x 3
qg He W/m
where x is the distance in metres measured from the inside surface, and H is a constant
dependent on the gamma radiation. The maximum temperature difference in the concrete
is to be limited to 4 K, and it may be assumed that the outer surface is well insulated.
Calculate the maximum allowable value of the gamma radiation on the inside surface of
the concrete in watts per square metre.
2
[Answer:68W/m ]
9. A large slab of rubber of thickness 40 mm is vulcanized by heating the faces using steam
` `
at 330 C. The required temperature at the centre is 120 C and the rubber is initially at
`
20 C. Calculate the time required for the process:
(a) using the method of (a) above;
(b) using a numerical method.
[Take σ for rubber as 64.5 10

9 m2 /s.] [Answer:26.4 min]
`
10. A large metal plate of thickness 200 mm is initially at a uniform temperature of 20 C.
`
One surface of the plate is in contact with ambient air at a constant temperature of 20 C,
2
while the other surface may be exposed to a constant net radiant heat flux of 100kW/m
from an electric element.
3
Data: Thermal conductivity of plate, 45 W/m K; density of plate, 7800 kg/m ; specific
heat capacity of plate, 0.5 kJ /kg K; heat transfer coefficient from plate to air, 200 W
2
/m K, Using the above data:
(a) and assuming as an approximation that the plate temperature is uniform throughout
`
at any instant, calculate the time taken for the plate to reach 70 C from the instant
the electric element is switched on;
86
(b) estimate the temperature distribution through the plate thickness 9 min after the el-
ement is switched on using a numerical method with four space increments and a
Fourier number of 0.5.
` ` ` `
[Answers:(a) 6.85 min; (b) 48 C; 54 C; 91 C; 162 C; 274 C]
`
`
11. The wall of a large vessel is 50 mm thick and is initially at 12.5 C throughout. A hot
`
fluid at a constant temperature of 500 C is suddenly pumped across the inside surface; the
outside surface may be assumed to be perfectly insulated. Given the following information:
Data: Thermal conductivity of wall, 22 W/mK; thermal diffusivity of wall, 6.22 10

6
2
m /s; heat transfer coefficient from fluid to wall, 110 W/mK.
Using a numerical method, determine the time taken for the junction of wall and insulation
`
to reach 110 C. [Answer: 455 s]
12. A thick fin ofrectangular cross-section, 1 m x 1 m, projects from a flat surface at 200 C
`
`
into a fluid at 20 C. Using the data below, estimate the temperature distribution in the
steady state assuming two-dimensional conduction, and hence calculate the rate of heat
loss from the fin surface per unit length.
Data Thermal conductivity of fin material, 25 W /m K; heat transfer coefficient for all parts
2
of the fin surface, 10 W / m K. [Answer:2.5 kW /m]
13. (a) In an oil cooler the oil enters 10 mm diameter tubes at 160 C and is cooled to
`
`
40 C; the mean velocity of the oil in the tubes is 1.5 m/s. Calculate the heat transfer
coefficient.
For turbulent flow of a liquid being cooled take:
0.8 0.3
Nu 0.0265 Re Pr
and for laminar flow take Nu = 3.65. Take all properties at the mean bulk temperature
and use the properties of engine oil given in Table 1.2.
Table 1.2: Properties of engine oil.
` 3
t/( C) ρ/(kg/m ) ν/(cSt)) λ /(W/m K) c/(kJ/kg K)
40 878 251.0 0.l44 1.96
100 839 20.4 0.137 2.22
160 806 5.7 0.31 2.48
6 2
1 centistoke (cSt) = 10 m /s
(b) In (a) above, the length of each tube is 1.2 m and the tube wall temperature is 20 C.
`
Allowing for the entry length effect a more accurate expression for the mean Nusselt
number over a length, L, for laminar flow is given by
µ 0.14 1©3
Nu 1.86 µ
r d/L RePrx
w
where properties are at mean bulk temperature with the viscosity of the oil at the tube
surface, µw = 0.8 kg/m s. Calculate the mean heat transfer coefficient allowing for
2 2
the entry length. [Answers:(a) 50 W /m K; (b) 177.5 W/m K]
87
` `
14. Air at l5 C and 1 bar is to be heated to 285 C while flowing at 34.2m /h through a 25 mm
3
`
diameter tube which is maintained at 455 C. Assuming that the simple Reynolds anal-
1©4
ogy is valid, taking f 0.0791© Re , and all properties at the mean bulk temperature,
calculate the length of the tube required. [Answer:1.88 m]
15. Air flows through a 20 mm diameter tube 2 m long with a mean velocity of 40 m/s. The
` `
tube wall temperature is 150 C and the air temperature increases from 15 to 100 C.
Using the simple Reynolds analogy with all properties at the mean bulk temperature, esti-
mate:
(a) the pressure loss in millimetres of water in the tube due to friction, and
(b) the pumping power required. Take the mean air pressure as 1 atm.
[Answers: (a) 173 mm water; (b) 21.3 W]
`
16. Air at a temperature of 15 C is blown across a flat plate with a surface temperature of
`
550 C at a mean velocity of 6 m/s. Neglecting radiation, calculate the rate of heat transfer
per metre width from both sides of the plate over the first 150 mm of the plate. For heat
transfer from a flat plate with a large temperature difference between the plate and the
fluid, the local Nusselt number is given by
1©3 1©2 0.117
Nu 0.332 Pr Re Tw ©Ts
where all properties are at the mean film temperature, Re is based on the distance from the
leading edge of the plate, and Tw and T. are the absolute temperatures of the plate and the
free stream of the air. [Answer:4.39 kW]
`
17. A wall 0.6 m high by 3 m wide is maintained at 79 C in an atmosphere at 15 C.
`
Neglecting end effects and radiation, calculate the rate of heat loss by natural convection.
For natural convection from a vertical flat surface take, at any distance, x:
1©2 1©4 1©4
Nux 0.509 Pr Pr 0.952 Grx
where all properties are at the mean film temperature, and /3 = 1 / T, where Tis the absolute
temperature of the bulk of the air.
[Answer:528 W]
`
18. A wall 0.6 m high by 3 m wide is maintained at 79 C in an atmosphere at 15 C.
`
Using the approximations:
α 1.42 ∆t©l
1©4
f or10
4
$ Gr $ 109
or α. 1.31 ∆t
1©3
f or10
9
$ Gr $ 1012
2
where α is in W /m K, ∆t is in K, and l is in m.
[Answer: 604 W]
LABORATORY EXERCISE
1. To find out the overall thermal conductance and plot the temperature distribution in case
of a composite wall
2. Determination of Thermal Conductivity of Insulating Materia
3. To determine the thermal conductivity of a liquid.
88
4. To find out the temp. Distribution along the length of a Pin Fin under free convection
5. To find out the temp. Distribution along the length of a Pin Fin under forced convection
6. To find out the Heat Transfer Coefficient of vertical cylinder in natural convection.
7. Determination of Thermal Conductivity of Insulating Material
8. Determination of Critical Heat Flux.
References
1.1: EASTOP T. D. and McCONKEY A. 1993 Applied Thermodynamics For Engineering Tech-
th
nologists, 5 edn. Pearson.
1.2 WELTY JR 1984 Fundamentals of Momentum, Heat and Mass Transfer 3rd edn John Wiley;
1.3 CROFT D R and LILLEY D G 1986 Heat Transfer Calculations Using Finite Difference
Equations Pavic Publications;
1.4 INCROPERA F P and DE WITT DP 1990 Fundamentals of Heat and Mass Transfer 3rd edn
John Wiley;
1.5 ROGERS G F C and MAYHEW YR 1987 Thermodynamic and Transport Properties of Flu-
ids 4th edn Basil Blackwell;
1.6 ECKERT ER and DRAKE RM 1971 Analysis of Heat and Mass Transfer Taylor and Francis;
1.7 KERN D Q 1950 Process Heat Transfer McGraw-Hill;
1.8 EASTOP TD and CROFT D R 1990 Energy Efficiency Longman;
1.9 McADAMS w H 1954 Heat Transmission 3rd edn McGraw-Hill;
1.10 EASTOP TD and WATSON w E 1992 Mechanical Services for Buildings Longman.
89
Chapter 2
Heat Exchangers
Learning outcomes:
1. Explain the terminologies used in heat exchangers (HE),
2. Describe construction and working the variuos type of heat exchangers,
3. Express heat transfer equations applicable to heat exchangers,
4. Apply identified equations to solve heat exchangersproblems,
5. Explain practical applications of HE in different thermodynamics systems.
2.1 Introduction
One of the most important processes in engineering is the heat exchange between flowing fluids.
In heat exchangers the temperature of each fluid changes as it passes through the exchanger, and
hence the temperature of the dividing wall between the fluids also changes along the length of the
exchanger. Examples in practice in which flowing fluids exchange heat are air intercoolers and
preheaters, condensers and boilers in steam plant, condensers and evaporators in refrigeration
units, and many other industrial processes in which a liquid or gas is required to be either cooled
or heated.
There are three main types of heat exchanger: the most important type is the recuperator in
which the flowing fluids exchanging heat are on either side of a dividing wall; the second type
is the regenerator in which the hot and cold fluids pass alternately through a space containing a
matrix of material that provides alternately a sink and a source for heat flow; the third type is the
evaporative type in which a liquid is cooled evaporatively and continuously in the same space as
the coolant. An example of the latter type is the cooling tower. Very often when the term ‘heat
exchanger’ is used it refers to the recuperative type, which is by far the most commonly used in
engineering practice. This section will deal almost entirely with the recuperative type.
2.2 Classification of Heat Exchangers

1. Based on principle of operation heat exchangers;
(a) Recuperative
90
(b) Regenerative
(c) Direct contact
In recuperative type of heat exchangers, cold and hot fluid flow through the unit without
mixing with each other. The transfer of heat occurs through the metal wall. Examples of
recuperative heat exchangers are boilers, heaters, coolers, vaporizers, condensers etc.
Regenerative type of heat exchangers same heating surface is alternately exposed to hot
and cold fluid. Heat associated with hot fluid is stored or absorbed by pickings or solids.
The hot fluid supply is then shut off and cold fluid is passed over pickings or solids to
regenerate the heat. Example of such type of heat exchangers is re-generators of open
hearth furnace, glass melting furnace etc.
Recuperative and regenerative units can also be called as surface condensers,
In direct contact type of heat exchangers hot and cold fluids are in direct contact and
mixing Occurs among them during the process of heat transfer. Mass transfer also occurs
simultaneously. Examples of direct contact type of heat exchangers are spray columns
cooling towers, scrubbers etc.
2. Based on flow pattern the heat exchangers;
(a) Co-current
(b) Counter-current
(c) Crossflow
In co-current flow-arrangement hot and cold fluid flow in the same direction. It is also
called as parallel flow arrangement. In counter-current flow arrangement hot and cold
fluid moves in opposite directions. In cross flow arrangement hot and cold fluids move at
right angles to each other.
3. Based on function of heat exchangers;
(a) Chiller (e) Partial condense (i) Steam generator

(b) Heater (f) Vaporizer (j) Waste heat boiler
(c) Evaporator (g) Re-boiler
(d) Cooler (h) Condenser
4. Based on type of construction;
(a) Shell and tube type (d) Air cooled

(b) Bayonet type (e) Spiral type
(c) Double pipe (f) Plate type
2.2.1 Recuperative type of HE

By far the most common type of HE are the recuperative and regenerative heat exchangers. The
difference between recuperative and regenerative heat exchanger systems is that in recuperative
91
heat exchangers (commonly called recuperators), each fluid simultaneously flows through its
own channel within the heat exchanger. On the other hand, regenerative heat exchangers, also
referred to as capacitive heat exchangers or regenerators, alternately allow warmer and cooler
fluids to flow through the same channel. Both recuperators and regenerators can be further sep-
arated into different categories of exchangers, such as direct or indirect and static or dynamic,
respectively. Of the two types indicated, recuperative heat exchangers are more commonly em-
ployed throughout industry.
Parallel-flow and counter-flow recuperators

Consider the simple case of a fluid flowing through a pipe and exchanging heat with a second
fluid flowing through an annulus surrounding the pipe. When the fluids flow in the same direc-
tion along the pipe the system is known as parallel-flow, and when the fluids flow in opposite
directions to each other the system is known as counter-flow. Parallel-flow is shown in Fig-
ure 2.1(a) and counter-flow is shown in Figure 2.1(b). Let the mean inlet and outlet temperatures
of fluid A be tA1 , and tA2 , respectively, and let the mean temperatures of fluid B at sections 1
and 2 be tB1 , and tB2 , respectively. Let the mass flow rates of fluid A and fluid B be mA and
mB respectively, and let the specific heats of fluid A and fluid B be cA and cB respectively. The
temperature difference at section 1 is tA1 tB1 ∆t1 , and the temperature difference at section
2 is tA2 tB2 ∆t2 .
Figure 2.1: Parallel-flow and

counter-flow heat exchangers and
the temperature distributions with
length.
Since the tube wall separating the fluids is thin it is possible to use an overall heat transfer
coefficient, U, based on equation (1.26),
1 1
i.e., RT A or U
U RT A
where A is the mean surface area of the tube.
Since the resistance of the tube wall is negligibly small we can write,
1 1 1
(2.1)
U αA αB
In practice the values of αA and αB will vary along the length of the tube, but suitable mean
values can be found. A mean value of U along the tube will be assumed.
Consider any section X-X where fluid A is at tA and fluid B is at B . The temperature difference
at this section is tA tB M, and a small amount of heat, d Q̇, is transferred across an element
of length dl. Using equation (1.20),
i.e., Q̇ UA tA tB , we have
92
dQ̇ πD dl U∆t (2.2)
where D is the mean diameter of the tube.

Fluid A increases in temperature by dtA along element dl, and fluid B increases in temperature
by dtB along element dl. Also since ∆t tA tB , we have,
d ∆t dtA dtB (2.3)
In the case of parallel-flow (Figure 2.1), temperature tA decreases with the length l, while tem-
perature tB increases with the length l. The heat given up by fluid A must equal the heat received
by fluid B, i.e. for parallel flow,
dQ̇ ṁA CA dtA ṁB CB dtB (2.4)
dQ̇ dQ̇

dtA and dtB
ṁA CA ṁB CB
dQ̇ dQ̇ 1 1
d ∆t dQ̇
(2.5)
ṁA CA ṁB CB ṁA CA ṁB CB
Integrating equation (2.5) between sections 1 and 2
1 1
∆t2 ∆t1 Q̇
ṁA CA ṁB CB
1 1
or, ∆t1 ∆t2 Q̇
(2.6)
ṁA CA ṁB CB
Also, from equation (2.5),
d∆t
dQ̇
1©ṁA CA 1©ṁB CB

therefore,
∆t
d
πD dl U∆t
1©ṁA CA 1©ṁB CB
d ∆t 1 1
and,, πDU
dl
∆t ṁA CA ṁB CB
Integrating between sections 1 and 2,
∆t2 1 1
ln
πDlU
(2.7)
∆t1 ṁA CA ṁB CB
93
where l is the total length of the tube. Now from equation (2.6)
1 1 ∆t1 ∆t2

ṁA CA ṁB CB Q̇
Hence substituting in equation (2.7),
∆t2 πDlU ∆t1 ∆t2

ln
∆t1 Q̇
πDlU ∆t1 ∆t2

or, Q̇ (2.8)
ln ∆t1 ©∆t2
In the case of counter-flow (Figure 2.1), both temperature tA and temperature tA decrease in the
direction of the length l. In place of equation (2.4) we therefore have,
dQ̇ ṁA CA dtA ṁB CB dtB
When the same procedure as for parallel-flow is carried out equation (2.8) is again obtained; this
procedure is left as an exercise for the student.
Comparing equation (2.8) with equation (1.20), i.e., Q̇ UA tA tB , it can be seen that the
mean temperature difference, ∆t̄, is given by,
∆t1 ∆t2
∆t̄ln (2.9)
ln ∆t1 ©∆t2
∆t̄ln is known as the logarithmic mean temperature difference. Then we have,
Q̇ UA∆t̄ln (2.10)
where A is the mean surface area of the tube, πDl.

There are several important points that should be mentioned here:
1. When one of the fluids is a wet vapour or a boiling liquid then its temperature remains
constant. Assuming fluid A to be a wet vapour, then the temperature variations are as
shown in Figure 2.2. It follows that under these circumstances the variation in temperature
of fluid B is the same whether the flow is parallel-flow or counter-flow.
2. In counter-flow the temperature range possible is greater, since, in theory, the fluid being
heated can be raised to a higher temperature than that of the heating fluid at exit. In
parallel-flow the final temperatures of the fluids must be somewhere between the initial
values of each fluid. This should be clear from Figures 2.1 (a) and 2.1(b).
3. When the product ṁA CA is equal to ṁB CB then the temperature difference in counter-flow
is the same all along the length of the tube. This must be the case since the heat given up
by fluid A is equal to the heat received by fluid B.
Referring to Figure 2.1 (b ), we have
ṁA CA tA1 , tA2 ṁB CB tB1 , tB2
94
Figure 2.2: Temperature variations
with length with one fluid condens-
ing or boiling.

, tA1 , tA2 tB1 , tB2 or tA1 , tB1 tA2 tB2
i.e., ∆t̄ ∆t1 ∆t2
Note that if we attempt to substitute in equation eq:HTTFR80, under these circumstances,

then the result is indeterminate,
∆t1 ∆t2 0 0
i.e., ∆t̄ln
ln ∆t1 ©∆t2 ln1 0
The proof of the logarithmic mean temperature difference given previously is not valid
when ∆t1 is equal to ∆t2 , since d ∆t is then zero.
4. From equation (2.10), Q̇ UA∆t̄ln , it can be seen that, for a given surface area, A, and
a given mean value of the overall heat transfer coefficient, U, then the logarithmic mean
temperature difference, ∆t̄ln , must be made as large as possible. It is found that for given
temperature changes, ∆t̄ln is always greater for counter-flow than it is for parallel-flow.
The initial temperature difference, ∆t1 , is greater for parallel-flow, but the value of ∆t̄ln is
always less. It follows that for given rates of mass flow of the two fluids, and for given
temperature changes, the surface area required is less for counter-flow.
Example 2.1
Exhaust gases flowing through a tubular heat exchanger at the rate of 0.3 kg/s are cooled from
` `
400 to 120 C by water initially at 10 C. The specific heat capacities of exhaust gases and water
may be taken as 1.13 and 4.19 kJ/kg K respectively, and the overall heat transfer coefficient from
2
gases to water is 140 W/m K. Calculate the surface area required when the cooling water flow is
0.4 kg/s,
(a) for parallel-flow,
(b) for counter-flow.
Solution:
Referring to Figures 2.3 and 2.4
95
(a) Surface area required in parallel-flow;
The heat exchanger is shown diagrammatically in Figure 2.3. The heat given up by the exhaust
gases is equal to the heat taken up by the water,
Q̇ 0.3 1.13 400 120 0.4 4.19 t 10 95 kW
95 `
i.e., t 10 66.6 C
0.4 4.19
∆t1 ∆t2
From equation (2.9) i.e., ∆t̄ln
ln ∆t1 ©∆t2 Figure 2.3: Parallel-flow heat ex-
changer for Example 2.1 (a).
400 10 120 66.6 336.6
169 K
ln r 400 10© 120 66.6x 1.99
i.e., Q̇ UA∆t̄ln
3
and, 95 10 140 A 169
3
95 10 2
i.e., A 4.01 m
140 169
2
thus, Surface area required 4.01 m
(b) Surface area required in counter-flow;
The heat exchanger is shown diagrammatically in Fig-

`
ure 2.4. The water temperature at outlet is 66.6 C and Q
= 95 kW as calculated in part(a). Again, from equation
∆t1 ∆t2
(2.9) i.e., ∆t̄ln
ln ∆t1 ©∆t2
120 10 400 66.6 223.4

201 K Figure 2.4: Counterflow heat ex-
ln r 120 10© 400 66.6x 1.11
changer for Example 2.1(b).
i.e., Q̇ UA∆t̄ln
3
and, 95 10 140 A 201
96
3
95 10 2
i.e., A 3.37 m
140 201
2
thus, Surface area required 3.37 m
Cross-flow recuperator
A simple cross-flow recuperator is shown in Figure ??. The calculation of the mean temperature
difference is much more difficult in this case. The true mean temperature difference depends on
the ratio of the product of the mass flow and specific heat capacities of fluids A and B, as well
as on the ratio of the temperature difference between the fluids at inlet and outlet. Tables are
available of a correction factor for various values of the ratios,
tB2 , tA2 ṁA CA

tB1 , tA1 and ṁB CB
Figure 2.5: Simple cross-flow heat

exchanger Heat exchangers.
The correction factor is multiplied by the arithmetic mean temperature difference to give the true
value of ∆t̄. When the temperature differences at inlet and outlet are not substantially different,
it is a sufficiently good approximation to use the arithmetic mean temperature difference,
tA1 , tA2 tB1 , tB2

i.e., ∆t̄a (2.11)
2 2
It has been shown in Example 2.1 that the surface area required for a given heat flow is smaller
with counter-flow than with parallel-flow. For cross-flow the required surface area is between
that for parallel-flow and counter-flow. As with counter-flow, the outlet temperature of the heated
fluid in cross-flow can be raised to a higher temperature than the outlet temperature of the cooled
fluid (e.g. in Figure 2.5, tA2 can be higher than tB2 ); this is not possible in parallel-flow.
97
Multipass and mixed-flow recuperators
The simple parallel-flow and counter-flow heat exchangers discussed above occur very rarely in
practice. To obtain the necessary surface area with a simple tube and annulus arrangement the
length of the tube may be too large for practical purposes. For instance, in Example 2.1 (b), if
the tube diameter were 150 mm, the length required would be,
A 3.37
l 7.15 m
πD π 0.15
In order to make the heat exchanger more compact, which is desirable from space considerations,
and also to reduce the heat loss from the outside surface, it is necessary to have several tubes and
perhaps several passes or bundles of tubes. The flow can be either cross-flow, or a mixture
of parallel-flow, counter-flow, and cross-flow. The latter case is called mixed-flow. A typical
example of a shell-type mixed-flow heat exchanger is shown in Figure 2.6. The analysis of a
mixed-flow heat exchanger is complex and correction factors have been plotted, which must be
used to evaluate the mean temperature difference. The logarithmic mean temperature difference
in counter-flow is evaluated and then multiplied by the correction factor.
Figure 2.6: Multipass shell-and-

tube heat exchanger.
Correction factors for most types of mixed-flow heat exchangers are given in ref. 2.6. Note that
when one of the fluids is a condensing vapour or a boiling liquid then the mean temperature
difference is the same whether the heat exchanger is parallel-flow, counter-flow, cross-flow, or
mixed-flow.
In certain heat exchangers of the multipass type, the mean temperature difference for counter-
flow or parallel-flow can still be used as a reasonable approximation. For example, the heat
exchanger in Figure 2.7 is essentially a counter-flow type. The larger the number of passes made
by fluid B then the nearer the heat exchanger is to pure counter-flow.
Example 2.2
A two-pass surface condenser is required to handle the exhaust from a turbine developing 15
MW with specific steam consumption of 5 kg/kWh. The condenser vacuum is 660 mm of Hg
when the barometer reads 760 mm of Hg. The mean velocity of water is 3 m/s, water inlet
`
temperature is 24 C. The condensate is saturated water and outlet temperature of cooling water
`
in 4 C less than the condensate temperature. The quality of exhaust steam is 0.9 dry. The overall
2
heat transfer coefficient based on outer area of tubes is 4000 W/m `C. The water tubes are 38.4
mm in outer diameter and 29.6 mm in inner diameter. Calculate the following :
(i) Mass of cooling water circulated in kg/min,
98
Figure 2.7: Shell-and-tube heat ex-
changer.
(ii) Condenser surface area,

(iii) Number of tubes required per pass, and
(iv) Tube length.
Solution:
Referring to Figures 2.8
Given : di = 29.6 mm = 0.0296 m ; do = 38.4 mm = 0.0384 m ;
2 `
U = 4000 W/m `C ; V = 3 m/s ; tc1 = 24 C ; x (dryness fraction) = 0.9.
The pressure of the steam in the condenser,
760 660
ps 1.0133 0.133 bar
760
`
The properties of steam at ps 0.133 bar, from steam table, are : tsat = 51 C ; h f g = 2592 kJ/kg,
`
tc1 = 51 - 4 = 47 C
The steam condensed per minute,
15 1000 5
ṁs ṁh 1250 kg/min
60
(i) Mass of cooling water circulated per minute,ṁw ṁs

Heat lost by steam = Heat gained by water,
ṁh x h f g ṁc ċ pc tc2 tc1
99
1250 0.9 2592 ṁc 4.187 47 24

ṁc ṁw 30280 kg/min
Figure 2.8: A
two-pass sur-
face condenser
for Q(2)(b).
(ii) Condenser surface area, A:
ṁs x h f g
Q̇ U A∆t̄ln (i)
60
∆t1 ∆t2
where, ∆t̄ln
ln ∆t1 ©∆t2
th1 tc1 th2 tc2

ln th1 tc1 © th2 tc2
51 24 51 47 27 4 `
12.04 C
ln 51 24© 51 47 ln 27©4
Substituting the values in equation (i), we get
1250 3
0.9 2592 10 4000 A 12.04
60
100
2
i.e., A 1009.1 m
(iii) Number of tubes required per pass, Np :
π 2
mw di V ρ Np
4
30280 π 2
0.0296 3 1000 N p
60 4
30280 4
or, Np 244.46 say, 245
60 π 0.02962 3 1000
i.e., Total number of tubes required, N 2N p 2 245 490
(iv) Tube length, L:

A πdo L 2N p
1009.1 π 0.0384 L 2 245
1009.1
L
π 0.0384 2 245
or, 17.1 m.
Fouling resistance
In most heat exchangers the fluid flowing is not completely free from dirt, oil, grease, and chem-
ical deposits, and a coating tends to collect on all metal surfaces. This increases the resistance to
heat transfer and must be allowed for in design calculations. It is usual to allow for the effect of
this coating of dirt by adding a fouling resistance to the total thermal resistance. Typical values
2
of fouling resistance for 1 m of surface area are: 1.8 K/kW for fuel oil; 0.6 K/kW for river
water; and 0.2 K/kW for boiler feedwater which has been treated. Facility must be provided for
easy periodic cleaning of the tubes. For a comprehensive treatment of process heat exchangers
see references 2.6 and 2.7.
Extended surface recuperators

Another form of recuperator which should be mentioned briefly is the extended surface type.
The metal wall containing a fluid to be cooled can be extended on the outside in the form of fins,
studs, or ribs. Examples of this type are the finned hot-water space-heater (sometimes misnamed
‘radiator’) and the air-cooled cylinders of small air compressors and IC engines.
The fins on the surface give a larger outside surface area for the same internal surface area, and
hence increase the cooling effect for a given volume. Details of compact heat exchangers are
101
Figure 2.9: Compact plate-fin heat
exchanger.
given in reference 2.8, two examples are shown in Figure 2.9 and 2.10. A simple analysis of
run-around coils used for energy recovery is given in reference 2.9.
Figure 2.10: Compact plate heat ex-

changer.
2.2.2 Regenerators
In the various types of recuperator described above, the hot and cold fluids are separated at all
times by a metal wall. The characteristic feature of a regenerator is that the fluids occupy the
same space in turn or are in contact with the same matrix in turn. The fluids used in regenerators
are nearly always gaseous. When the hot gas occupies the space, it gives up heat to the walls, or
to solid matter distributed throughout the space, called a matrix. The hot gas is then withdrawn,
the cold gas enters the space and is heated by the walls and the matrix; the process is a cyclic
one and analysis is complex.
In order to have continuous operation it is usual to use two regenerators with hot gas flowing
through one, and cold gas flowing through the other at any instant. When the flows are switched
from one regenerator to the other, the hot gas is cooled while the cold gas is heated. The period
of time between the switching of the flows must be chosen to give the required heat transfer
between the two gases. This type of generator is shown diagrammatically in Figure 2.11 (a).
Another method used is the rotating matrix type, in which a cylinder containing solid inserts is
rotated so that it passes alternately through cold and hot gas streams which are sealed from each
other. An example of this type of regenerator is the Ljungström air pre-heater for boiler furnaces,
shown diagrammatically in Figure 2.11 (b ).
There are many applications for regenerators, from air pre-heaters in blast furnaces to heat ex-
changers in plants for gas liquefaction. One application of the rotating matrix type which is
becoming important is in energy conservation.
102
Figure 2.11: Two types of regenera-
tive heat exchangers: stationary (a)
and rotating (b) matrices.
2.3 Heat exchanger effectiveness

In heat exchanger design the efficiency of the heat transfer process is very important. A method
due to Nusselt and developed by Kays and London for compact heat exchangers (see reference
2.8) is described in this section.
The effectiveness, ε, of a heat exchanger is defined as the ratio of the actual heat transferred
to the maximum possible heat transfer. For any heat exchanger with mass flow rates of hot and
cold fluids, ṁH and ṁC , with specific heat capacities, cH and cC , let the overall temperature
changes of each fluid be tHi tHe and tCi tCe .; where subscripts i and e refer to inlet and
exit. Neglecting heat losses to the surroundings:
Q̇ ṁH cH tHi tHe ṁC cC tCi tCe
or, Q̇ CH tHi tHe CC tCi tCe (2.12)
where, CH ṁH cH and CC ṁC cC are the thermal capacities of the hot and cold fluids. From
equation (2.12) it can be seen that the fluid with the smaller thermal capacity, C, has the greater
temperature change. The maximum possible temperature change of one of the fluids is tHmax.
tHmin. , and this ideal temperature change can only be aspired to by the fluid with the minimum
thermal capacity,
Q̇
i.e., ε (2.13)
Cmin. tHmax. tHmin.
The object of a well-designed heat exchanger is to obtain the maximum possible change of
temperature of a fluid for a given driving force, that is for a given logarithmic mean temperature
difference, ∆t̄ln . Hence another useful measure of the efficiency of a heat exchanger is the number
of transfer units, Ntu , defined as
Temperature change of one fluid
Ntu
∆t̄ln
Hence we can write for the hot fluid,

tHi tHe
NtuH
∆t̄ln
103
tHi tHe
NtuH
∆t̄ln
and for the cold fluid,
tCe tCi
NtuC
∆t̄ln
From equations (2.12) and (2.10), i.e.,
Q̇ UA∆t̄ln
also, Q̇ CH tHi tHe CC tCi tCe
i.e., Q̇ UA∆t̄ln CH tHi tHe CC tCi tCe
tHi tHe UA tC tCe UA

then, and i
∆t̄ln CH ∆t̄ln CC
UA UA
hence, NtuH and NtuC
CH CC
A more general definition of Ntu is as follows:
Greater of the two fluid temperature differences

Ntu
∆t̄ln
UA
i.e., Ntu (2.14)
Cmin.
The greater the number of transfer units the more effective is the heat exchanger.
The ratio of the minimum to the maximum thermal capacity is usually given the symbol R,
Cmin.
i.e., R (2.15)
Cmax.
Note that R may vary between 1 (when both fluids have the same thermal capacity) and O (when
one of the fluids has an infinite thermal capacity, e.g. a condensing vapour or a boiling liquid).
Figure 2.12 shows a typical example of a graph of effectiveness, ε, against Ntu for various values
of the thermal capacity ratio, R.
Consider a counter-flow heat exchanger as shown in Figure 2.13. From the figure it can be seen
that CC Cmin , since ∆tC % ∆tH , i.e., R CC ©CH or, using equation (2.12), i.e.,CH tHi
tHe CC tCi tCe
CC tCi tCe
or, R
CH tHi tHe
104
Figure 2.12: Effectiveness against
number of transfer units for various
values of thermal capacity ratio.
Figure 2.13: Temperature varia-

tions in a counter-flow heat ex-
changer.
tH1 tH2
Here, R tC1 tC2 (2.16)

Q̇ Cmin. tC1 tC2 tC1 tC2
i.e., ε (2.17)
Cmin. tHmax. tHmin. Cmin. tH1 tC2 tH1 tC2
From equation eq:HTTFR85

UA
i.e., Ntu
Cmin.
UA tC1 tC2
or, Ntu
Cmin. ∆t̄ln
From equation (2.9)
∆t1 ∆t2 tH1 tC1 tH2 tC2
∆t̄ln
ln ∆t1 ©∆t2 ln s tH1 tC1 © tH2 tC2 y
105
tC1 tC2 tH tC1
i.e., Ntu ln v 1 |
tH1 tC1 tH2 tC2 tH2 tC2
tH1 tH2 tC1 tC2 tH1 tC2 tC1 tC2

or, Ntu ln v |
tC1 tC2 tH1 tC2 tH1 tH2
tH tH tC1 tC2 tH1 tC2 tC1 tC2

or, Ntu t 1 t 2
tC1 tC2 ln v |
C1 C2 tH1 tC2 tH1 tH2
tH1 tH2
Introducing thermal ratio, R, and effectiveness, ε, into equation using equation (2.16),R
tC1 tC2
tC tC
and equation (2.17),ε t 1 t 2
H1 C2
tH1 tH2 tC1 tC2 tC1 tC2 ©ε tC1 tC2

i.e., Ntu v | ln v |
tC1 tC2 tC1 tC2 tC1 tC2 ©ε R tC1 tC2
1©ε 1
and, Ntu R 1 ln
1©ε R
1ε
then, Ntu R 1 ln
1 Rε
1e
Ntu 1R
or, (2.18)
1 ReNtu 1R
ε
Note that for a counter-flow heat exchanger when CH CC , i.e. R = 1 (say for a gas turbine heat
exchanger), then the expression for effectiveness cannot be obtained by substituting R = 1 in
equation (2.18). For this case the temperature change of each fluid is the same, since CH CC ,
and hence ∆t̄ln is equal to the temperature difference between the hot and cold fluids which
remains constant throughout the heat exchanger. Equation (2.14) is therefore written as, Ntu
tC1 tC2 © tH1 tC1 , and the derivation then proceeds as above, giving,
Ntu
ε (2.19)
1 Ntu
For a parallel-flow heat exchanger it can be shown that,
N
1 e tu
1R
ε (2.20)
1R
When R = 0 in the case of a condenser say, then it can be seen from equation (2.17) or equation
(2.19) that the effectiveness is
N
ε 1 e tu .
Example 2.3
106
A single-pass shell and tube counter-flow heat exchanger uses waste gas on the shell side to heat
`
a liquid in the tubes. The waste gas enters at a temperature of 400 C at a mass flow rate of 40
`
kg/s; the liquid enters at 100 C at a mass flow rate of 3 kg/s.
Assuming that the velocity of the liquid is not to exceed 1 m/s, using the data below: Neglect
fouling factors and the thermal resistance of the tube wall. Tube inside diameter= 10 mm; tube
outside diameter = 12.7 mm; tube length = 4 m; specific heat capacity of waste gas = 1.04 kJ/kg
3
K; specific heat capacity of liquid = 1.5 kJ/kg K; density of liquid = 500 kg/m ; heat transfer
2 2
coefficient on the shell side = 260 W/m K; heat transfer coefficient on the tube side= 580 W /m
K;
Calculate:
(a) the required number of tubes;
(b) the effectiveness of the heat exchanger;
(c) the exit temperature of the liquid.
Neglect fouling factors and the thermal resistance of the tube wall. Tube inside diameter= 10
mm; tube outside diameter= 12.7 mm; tube length = 4 m; specific heat capacity of waste gas =
1.04 kJ/kg K; specific heat capacity of liquid = 1.5 kJ/kg K; density of liquid = 500 kg/m 3; heat
2
transfer coefficient on the shell side = 260 W/m K; heat transfer coefficient on the tube side=
2
580 W /m K.
Solution:
Figure 2.14: Temperature varia-

tions in a counter-flow heat ex-
changer.
(a) Number of tubes;

ṁ 3 3
Volume flow rate of liquid,V̇ 0.006 m
ρ 500
V̇ 0.006 2

Total cross-sectional area for a velocity, c, of 1 m/s c 0.006 m
1
107
0.006 4
i.e., Number of tubes 76.39, say 77
π 0.012
Note: the velocity in the tubes is then less than 1 m/s as required.
2.3.1 (b) effectiveness of the HE;

1 1 1
From equation (2.1) αA αB , taking into account the area difference,

U
1 1 1
i.e.,
UAo αo Ao αi Ai
where subscripts O and i refer to the outside and inside of the tube. The overall heat transfer
coefficient, U, is referred to the outside area which is the usual practice in heat exchanger design,
1 1 Ao 1 12.7 2

i.e.,
U αo αi Ai

260 580 10
0.00604 m K/W
Ao
since Do ©Di
Ai
2
and, U 165.68 W/m K
UA
Then from equation (2.14) i.e., Ntu .
Cmin.
165.68 π 0.0127 4 77
and, O.452
3 1.5 1000
3 l.5
also, R 0.1082
40 1.04
Then from equation (2.18),
1e
Ntu 1R
i.e.,
1 ReNtu 1R
ε
1e
0.452 1 0.1082
and,
1 0.1082e0.452 10.1082
ε
or, ε 0.358
(c) Exit temperature

tC1 tC2
Also from equation (2.17), i.e., ε ,
tH1 tC2
tLe 100
i.e., ε
400 100
,
where tLe is the exit temperature of the liquid.
`
i.e., tLe 300 0.358 100 207.4 C
108
Example 2.4
` `
In a counter-flow double pipe heat exchanger, water is heated from 25 C to 65 C by an oil with
`
a specific heat of 1.45 kJ/kg K and mass flow rate of 0.9 kg/s. The oil is cooled from 230 C to
` 2`
160 C. If the overall heat transfer coefficient is 420 W/m C, calculate the following :
(a) The rate of heat transfer,
(b) The mass flow rate of water, and
(c) The surface area of the heat exchanger
Solution:
` `
Given: tc1 = 25 C; tc2 = 65 C, c ph = 1.45 kJ/kg K ; ṁh= 0.9 kg/s;
` ` 2`
th1 = 230 C ; th2 = 160 C, U = 420 W/m C. (i) The rate of heat transfer, Q̇:
Q̇ ṁh c ph th1 th2
or, Q̇ 0.9 1.45 230 160 91.35 kJ/s
(ii) The mass flow rate of water, ṁc

Heat lost by oil (hot fluid) = Heat gained by water (cold fluid)
ṁh c ph th1 th2 ṁc c pc tc2 tc1
91.35 ṁc 4.187 65 25
91.35

ṁc 0.545 kg/s
4.187 65 25
(iii) The surface area of heat exchanger, A:

Logarithmic mean temperature difference (LMTD), ∆t̄ln is given by;
∆t1 ∆t2
∆t̄ln
ln ∆t1 ©∆t2
th1 tc2 th2 tc1 230 65 160 25

ln th1 tc2 © th2 tc1 ln 230 65© 160 25
165 135 `
i.e., ∆t̄ln 149.5 C
ln 165©135
Also, Q̇ UA∆t̄ln
109
Figure 2.15: Counter-flow heat exchanger for Example 2.4.
3
Q̇ 91.35 10 2
And, A 1.45 m .
U∆t̄ln 420 149.5
2.4 Extended surfaces - Fins

From equation eq:HTTFR81, (Q̇ UA∆t̄ln ), it can be seen that for a given heat transfer coef-
ficient and given fluid temperatures the heat transfer can be increased by increasing the heat
transfer area. One way of doing this is to increase the area on one side of the heat exchanger by
adding fins or studs which project into the fluid; the effective heat transfer area is thus increased.
The thermal resistance on either side of a heat exchanger is 1/αA, therefore when α is very large
the resistance to heat transfer is low and hence there is no advantage in increasing the area. One
of the fluids usually has a much lower value of α than the other and hence the resistance on this
side of the heat exchanger controls the heat transfer. It is therefore on this side that the area can
be extended with advantage. (The resistance of the separating wall is usually small compared
with the resistance of the fluid film on either side.) Extended surfaces or fins are generally used
to enhance convective heat transfer rate between a solid and the surrounding fluid.
A fin extends the surface area of heat transfer. The fin material generally has a high thermal
conductivity which is exposed to a flowing fluid.
110
Heat transfer coefficients are very high for condensing and boiling fluids; they are generally
higher for liquids than for gases, and generally higher for forced convection than for natural
convection. A typical application for an extended surface would therefore be natural convection
to air.
Fins are often seen in electrical appliances and electronics such as on computer processors and
power supplies. It is also used in industrial applications such as heat exchangers and engine
cooling. Type of Fins.
2.4.1 Type of Fins

Fins are designed and selected for a particular purpose and environment. Rectangular fins are
often used in computers for cooling, Annular fins are often used in heat exchangers.
Regardless of the shape, the heat transfer occurring in a fin is governed by the same laws. (Con-
duction, Convection, Radiation) Figures 2.16 shows the various types of fins in use.
Figure 2.16: Types of fins.
2.4.2 Deriving the Heat Equation for a Fin:

Consider an extended surface which has a constant cross-sectional area, A, which is small com-
pared to its length, L, so that heat transfer along it is one-dimensional (see Figures 2.17).
Figure 2.17: Extended surface of

small cross-section in a fluid of con-
stant temperature.
In order to setup the problem, several assumptions need to be made:

• Steady State
– Temperature is not a function of time.
• Material properties are constant ( and )
– Independent of temperature
• No internal heat generation
• Uniform convective heat transfer coefficient over surface
• One dimensional (1-D) conduction
111
– Temperature at each location is uniform over
• Radiation is negligible
Heat Transfer for a Differential Element

The Non-flow energy equation (Figures ??):
∆E = Q - W
∆E Internal energy is constant with time.

W No work is done on the system.
Generally,
Q̇in Q̇out
For the differential element,
Q̇x Q̇x∆x Q̇Conv. (2.21)
For a perimeter, p, and a heat transfer coefficient, α, as-

sumed constant with temperature and uniform over the
surface, then for the steady state,
Figure 2.18: Heat flow in differential
element
Heat loss from surface,Q̇Conv. α P∆x t tF
where tF is the temperature of the surrounding fluid, as-

sumed uniform and constant.
If we rearrange equation (2.21) and divide by ∆x, we
obtain:
Q̇x∆x Q̇x
αP t tF 0
∆x
Take the limit as ∆x 0:
dQ̇
αp t tF 0
dx
where Q̇ is from Fourier’s Law of Thermal Conduction.

i.e., the rate of heat transfer at any section, distance x from the primary surface, where the
temperature is t, is given by
dt
Q̇ kAc
dx
Thus, the differential equation governing heat transfer in fins is:
d dt
kAc
αp t tF 0 (2.22)
dx dx
112
NOTE: Both the cross sectional area, Ac , and perimeter,p, are functions of x, differentiating the
first term;
2
dAc dt d t
k kAc αp t tF 0
dx dx dx2
Dividing through by kAc and re-arranging to get;
2
d t 1 dAc dt αp
t tF 0 (2.23)
dx 2 Ac dx dx kAc
If Ac = constant, than the equation can be reduced to:
2
d t αp
2
t tF 0 (2.24)
dx kAc
nd
The temperature profile in a fin is governed by a 2 order, Ordinary Differential Equation
(ODE).
In order to obtain a solution to the homogeneous equation we must have two boundary conditions
(at the fin base and tip).
×
αp
Let, θ t tF , and, m (i)
kAC
2
d θ 2
m θ 0
dx2
The soln. is of the form,
mx mx
or, θ x C1 e C2 e
mx mx
t x tF C1 e C2 e (2.25)
In order to obtain an exact solution to the homogeneous equation, two boundary conditions (at
the fin base and tip) need to be known (Figures 2.19 and Figures 2.20).
One METHOD of solving equation (2.25):
Boundary Condition 2
m
θ 0 C1 e
and c1 0
Boundary Condition 1
m0
θ 0 θB tB t C2 e
i.e., C2 θB tB t
113
Figure 2.19: Tip boundary condi-
tions.
Figure 2.20: Boundary conditions

for infinitely long fin.
Solution is:
mx
θ x θB C1 e
mx
t x t tB t e
mx
Alternative METHOD: The solution is found by putting t tF e then the differentiating
the resulting eqtn. twice,
2
d t tF 2 mx
i.e., 2
m e
dx
2 mx αp mx

, m e e 0
kA
×
αp
or, m (ii)
kA
Again, takingC1 and C2 are constants determined from the boundary conditions:
i.e.,
(i) at x = 0, tx tF t1 tF ; therefore in equation (2.25)
m0 m0
t1 tF C1 e C2 e
114
and, t1 tF C1 C2 (2.26)
(ii) at the end of the extended surface the heat convected from the end is equal to the heat
conducted at the section x = L.
Assuming the same heat transfer coefficient, α, for the end then,
dt1 tF
kA
αA t1 tF x L (2.27)
dx x L
dt1 tF αA mL mL

C1 e C2 e x L
dx x L kA
From equation (ii), ×

αp
m
kA
dt1 tF mL mL

mC1 e mC2 e (iii)
dx x L
Integrating,
mL mL
i.e., t1 tF x L C1 e C2 e (iv)
Substituting equations (iii) and (iv) in (2.27);
mL mL mL mL
kmC1 e kmC2 e αC1 e αC2 e (2.28)
From equations (2.26) and (2.28) the values of C1 and C2 can be found, and hence the solution
to equation (2.24) is,
m Lx m Lx α em Lx em Lx

t tF e e
mk
emL emL
t1 tF α mL mL
e e
mk
α
t tF cosh m L x sinh m L x
or, mk (2.29)
t1 tF α
cosh mL sinh mL
mk
since, cosh mL e
mL
e
mL ©2 and sinh mL mL
e e
mL ©2. At the end of the extended
surface, at x = L,
t2 tF 1
t1 tF α (2.30)
cosh mL sinh mL
mk
115
The heat transfer from the extended surface, which is the same as the heat leaving the primary
surface to the fin, is given by,
dt1 tF
Q̇1 kA
(v)
dx x 0
Substituting equation (2.29) in equation (v) and differentiating and putting x = 0, it can be shown
that,
tanh mL
1
α ©mk
Q̇1 αA t1 tF (2.31)
1 α ©mk tanh mL
From equation (2.31) it can be seen that when α/ml= 1 then, Q̇1 αA t1 tF , which is the heat
loss from the primary surface with no extended surface, i.e. when α = ml an extended surface of
whatever length, L, will not increase the heat transfer from the primary surface.
For α ©ml % 1 then Q̇1 $ αA t1 tF and hence adding a secondary surface reduces the heat
transfer; the surface added acts as an insulation. For α ©ml $ 1, then Q̇1 % αA t1 tF and the
extended surface will increase the heat transfer. This is illustrated in Figure 2.22.
Figure 2.21: Heat transfer against

length for various values of α ©mk.
1©2
αA 1©2
2
α α kA
Note, 2
mk k αp kp
Hence assuming α ©ml $ 1 the heat transfer becomes more effective when α/k is low for a given
geometry.
2.4.3 Approximate end condition

A simplified approach is possible by making the approximation that the loss of heat from the end
of the extended surface is negligible, i.e. at x = L
dt1 tF
kA
0 (vi)
dx x L
116
Therefore substituting equation (2.25) in equation(vi) and differentiating, we obtain,
mL mL
kAm C1 e C2 e 0 (2.32)
Solving for C1 and C1 from equations (2.32) and (2.26) we have,
t2 tF cosh m L - x
(2.33)
t1 tF cosh mL
Then at x = L,
t2 tF 1
(2.34)
t1 tF cosh mL
The heat transfer from the extended surface is obtain by substituting equation (2.33) in,
dt1 tF
Q̇1 kA
dx x L
and differentiating as required. The result obtained is;
Q̇1 mkA t1 tF tanh mL (2.35)
In most cases in practice the approximate expressions given by equations (2.33) - (2.35) can be
used instead of the more accurate expressions given by equations (2.28) - (2.31). In the important
case of compact plate-fin heat exchangers where corrugated plates are sandwiched between flat
plates (see for example Figure 2.9 , then the expressions given in equations (2.33) - (2.35) are
the accurate ones. In this case the fin bridges the two hot surfaces which may be assumed to be
at the same temperature; at the mid-point of the fin the change of temperature with fin length is
zero which corresponds to the condition of zero heat transfer. In the equations (2.28) - (2.31) the
half-width, w/2, is substituted for the length, L.
2.4.4 Rectangular section fins

For a fin of rectangular cross-section on a plane surface as shown in Figure ??, the perimeter, P,
is given by (2 + 2b) per unit length in the z-direction, and the cross-sectional area, A, is given by
b per unit length in the z-direction,
1©2
αp 1©2 α 2 2b 2α 2α 1©2
i.e., m
1 b
(2.36)
kA kb kb kb
α
Also, by obtaining and multiplying both side of the inverse of equation (2.36) with it, we
k
obtain,
1©2
αb©2 1©2
2
α α kb 1©2
2
Bi
mk k 2α k
where the Biot number, Bi, is based on the half-thickness of the fin, b/2.
117
Figure 2.22: Rectangular cross-
section fin on a plane surface.
2.4.5 Fin efficiency

Fin efficiency, ηF , is defined as the ratio of the heat loss from the fin surface to the heat loss from
the fin surface if it were everywhere at the temperature of the primary surface.
Using the approximate expression, equation (2.35), for Q̇1 , we have
Q̇1 mkA t1 tF tanh mL,
mk t1 tF tanh mL mk tanh mL tanh mL

i.e., ηF (2.37)
α 2L t1 tF 2αL mL
For a finned surface with unfinned area, Ab , and total fin surface area, AF , then for unit length in
the z-direction;
Heatloss α t1 tF Ab ηF AF (2.38)
Example 2.5
For a point in one of the hot fluid channels for a typical plate-fin cross-flow heat exchanger where
`
the fluid mean temperature within the channel is 200 C and the separating plates on either side
`
are at 100 C, calculate:
(a) the mean temperature of a fin at that point in the heat exchanger;
(b) the fin efficiency.
Data Height of flow channel= 11.78 mm; thickness of fin= 0.203 mm; heat transfer coefficient
2
between the hot fluid and all surfaces= 137 W /m K; thermal conductivity of fin material = 168
W/m K. Solution:
Referring to Figures ??
t2 tF cosh m L x
t1 tF cosh mL
118
For this example;
tm tF D0w 2 t
©
tF dx sinhmw/2 - x0
w©2

t1 tF t1 tF mw©2 cosh mw©2
sinh mw©2 tanh mw©2

mw©2 cosh mw©2 mw©2
It can be seen by reference to equation (2.37) and the above equation,
mk t1 tF tanh mL mk tanh mL tanh mL

i.e., ηF ,
α 2L t1 tF 2αL mL
tm tF
that the ratio, is an alternative expression for fin efficiency.
t1 tF
(a) mean temperature of fin
2α 1©2
i.e., m
kb
1©2
2 137
and, m
89.63
168 0.203 103

mw©2 89.63 11.78©2 103 0.5279
tm tF tanh 0.5279
and, t1 tF 0.916
0.5279
`
tm 200 0.916 200 100 108.4 C
i.e. Mean temperature of fin = 108.4 C
`
(b) Fin efficiency, η Then,

tm tF 200 108.4
η t1 tF 200 100
0.916 or 91.6%
Question ONE
An oil engine develops 300 kW and the specific fuel consumption is 0.21 kg/kWh. The exhaust
from the engine is used in a tubular water heater, flowing through 25 mm diameter tubes, entering
` ` `
with a velocity of 12 m/s, at 340 C and leaving at 90 C. The water enters the heater at 10 C and
`
leaves at 90 C, flowing in counter-flow to the hot gases. The air-fuel ratio of the engine is 20,
and the exhaust pressure is 1.01 bar. The overall heat transfer coefficient of the heat exchanger
2
when designed is found to be 56 W /m K, but after running for some time a fouling factor of 0.5
2
m K/kW must be assumed. Taking the specific heat capacity and the gas constant for the gases
as 1.11 kJ/kg K and 0.29 kJ/kg K, and the specific heat capacity for the water as 4.19 kJ /kg K,
calculate:
119
(a) the mass flow rate of water;
(b) the number of tubes required;
(c) the required tube length.
[Answers: (a) 1096 kg/h; (b) 110; (c) 1.457 m]
Question TWO
`
A flat surface at a temperature of 300 C has rectangular section cooling fins perpendicular to
`
the surface projecting into a fluid at 20 C. There are 12.5 fins per 100 mm and the fins have a
thickness of 3 mm and a length of 30 mm.
The thermal conductivity of the fin material is 26 W/mK and the heat transfer coefficient for all
2
surfaces may be taken as 40 W/m K.
Neglecting the heat loss at the tip of each fin, calculate:
(a) the fin efficiency;
(b) the rate of heat loss from unit area of the flat surface;
(c) the temperature at the tip of each fin.
2
[Answers: (a) 77.5%; (b) 72.1 kW/m ; (c) 206.9°C]
Tutorial
`
1. A pipe containing dry saturated steam at 177 C is 150 mm bore and has a 50 mm thickness
of 85% magnesia covering. The steam velocity is 6 m/s and the heat transfer coefficient
may be found from
0.8 0.4
Nu 0.023 Re Pr
where all properties are at the mean bulk temperature. The atmospheric temperature is
`
17 C and the heat transfer coefficient from a horizontal cylinder is given approximately
by
1©4
α l.32 ∆t/d
2
The pipe wall is 7 mm thick and the thermal conductivity of the pipe metal is 50 W/m K;
the thermal conductivity of the 85% magnesia insulation is 0.06 W/m K.
Neglecting radiation, taking arithmetic mean areas for the pipe wall and lagging, and using
a trial-and-error method, calculate:
(a) the temperature of the outside surface of the lagging;
(b) the rate of heat loss from the pipe per unit length.
`
[Answers: (a) 46.3 C; (b) 104 W/m]
2. An exhaust pipe of 75 mm outside diameter is cooled by surrounding it by an annular
`
space containing water. The exhaust gas enters the exhaust pipe at 350 C, and the water
`
enters from the mains at 10 C. The heat transfer coefficients for the gases and water may
2
be taken as 0.3 and 1.5 kW/m K, and the pipe thickness may be taken to be negligible. The
`
gases are required to be cooled to 100 C and the mean specific heat capacity at constant
pressure is 1.13 kJ/kg K. The gas flow rate is 200 kg/h and the water flow rate is 1400 kg/h.
Taking the specific heat capacity of water as 4.19 kJ/kg K, calculate:
(a) the required pipe length for parallel-flow;
120
(b) the required pipe length for counter-flow.
[Answers: (a) 1.48 m; (b) 1.44 m]
3
3. In a chemical plant a solution of density 1100 kg/m and specific heat capacity 4.6 kJ/kg
` `
K is to be heated from 65 C to 100 C; the required flow rate of solution is 11.8 kg/s. It
is desired to use a tubular heat exchanger, the solution flowing at about 1.2 m/s in 25 mm
`
bore iron tubes, and being heated by wet steam at 115 C. The length of the tubes must
not exceed 3.5 m. Taking the inside and outside heat transfer coefficients as 5 and 10 kW
2
/m K, and neglecting the thermal resistance of the tube wall, estimate:
(a) the number of tubes, and
(b) the number of tube passes required.
[Answers: (a) 18; (b) 4]
4. In an air cooler the air is blown across a bank of tubes at the rate of 240 kg/h at a velocity
` `
of 24 m/s, the air entering at 97 C and leaving at 27 C. The cooling water enters the tubes
` `
at 10 C and leaves at 20 C, at a mean velocity of 0.6 m/s. The tubes are 6 mm diameter
and the wall thickness may be neglected. The heat transfer coefficient from the air to the
tubes may be calculated from
0.6 0.33
Nu 0.33 Re Pr
with properties at the mean bulk temperature.

The heat transfer coefficient from the water to the tubes is given by
f©2
St
1 Pr1©6 Re1©8 Pr 1
1©4
where f = 0.0791/Re and properties are at the mean bulk temperature. Assuming that
the tubes are arranged in six passes, and that the logarithmic mean temperature difference
for counter-flow can be assumed, calculate:
(a) the number of tubes required in each pass;
(b) the necessary tube length.
[Answers: (a) 7; (b) 0.528 m]
5. A two-pass shell-and-tube heat exchanger is used to condense a chemical on the shell side
`
at a rate of 50 kg/s at a saturation temperature of 80 C. The chemical enters as a dry sat-
`
urated vapour and is not undercooled during the process. Water at 10 C and a mass flow
rate of 100 kg/s is available as coolant; the velocity of the water is to be approximately 1.5
m/s. Using the data below and taking a nominal tube diameter of 25 mm, Data: Specific
enthalpy of vaporization of chemical, 417.8 kJ /kg; heat transfer coefficient for shell side,
2 2
10 kW /m K; fouling factor for shell side, 0.1 m K/kW; fouling factor for tube side, 0.2
2
m K/kW.
For turbulent flow in a pipe take,
0.8 0.4
Nu 0.023 Re Pr with properties at the mean bulk temperature.
Neglecting tube wall thickness, calculate:
(a) the number of tubes required;
121
(b) the tube length;
(c) the number of transfer units;
(d) the effectiveness of the heat exchanger.
[Answers: (a) 274; (b) 13.55m; (c) 1.253; (d) 71.4%]
6. An oil cooler consists of a single-pass, counter-flow shell-and-tube heat exchanger with
300 tubes of internal diameter 7.3 mm and length 8 m. The oil flows in the tube side
`
entering at a mass flow rate of 8 kg/s at a temperature of 70 C. Cooling water in the
`
shell side enters at a mass flow rate of 12 kg/sat a temperature of 15 C. Using the data
2
below: Data: Shell side heat transfer coefficient, 1000 W/m K; heat transfer coefficient
0.8 0.4
for the tube side given by Nu 0.023 Re Pr with properties as follows: specific
3
heat capacity of oil, 3.42 kJ/kg K; density of oil, 900 kg/m ; dynamic viscosity of oil, 1.5
10
3 kg/ms; thermal conductivity of oil, 0.15 W/m K.
Calculate:
(a) the number of transfer units;
(b) the effectiveness of the heat exchanger;
(c) the outlet temperature of the oil.
[Answers: (a) 1.1; (b) 58.8%; (c) 37.7 C]
`
7. A double pipe heat exchanger has an effectiveness of 0.5 when the flow is counter-current
and the thermal capacity of one fluid is twice that of the other fluid. Calculate the effec-
tiveness of the heat exchanger if the direction of flow of one of the fluids is reversed with
the same mass flow rates as before.
[Answer: 0.469]
`
8. 500 kg/h of oil at 120 C is to be cooled in the annulus of a double pipe counter-flow heat
`
exchanger by water which enters the inside pipe at 10 C. The inner pipe has an inside
diameter of 25 mm and a wall thickness of 2 mm, and the inside diameter of the outer pipe
is 50 mm; the effective length is 12 m. Using the data below:
Data: Oil; take Nu = 30, based on an equivalent diameter, de , given by de = 4 (flow
area)/(heat transfer area per unit length); specific heat capacity, 2.31 kJ/kg K; thermal con-
2
ductivity, 0.135 W/m K; fouling factor, 0.001 m K/W. Water; assume the simple Reynolds
analogy holds true, taking the velocity as 1 m/s and the friction factor, f, as 0.0002; specific
3 2
heat capacity, 4.18 kJ/kg K; density, 1000 kg/m ; fouling factor, 0.0002 m K/W.
Neglecting the thermal resistance of the pipe wall, calculate the exit temperature of the oil.
`
[Answer: 98.8 C]
9. A condenser contains four tube passes with tubes 3 m long, 25 mm internal diameter, each
`
pass containing 100 tubes. Cooling water enters the tubes at 20 C at the rate of 80 kg/s
`
when the shell side vapour is at 50 C. Before cleaning, the fouling factor on the water
2
side is 0.0005 m K/W; the outside of the tubes may be taken to be clean. Neglecting the
thermal resistance of the fluid film on the outside of the tubes and the thermal resistance
of the tube wall, calculate, using the data below:
Data: Specific enthalpy of vaporization for shell side fluid, 300 kJ/kg; mean properties of
3
water for the temperature range considered: density, 1000 kg/m ; specific heat capacity,
3
4.19 kJ/kg K; thermal conductivity, 0.6 W/mK; dynamic viscosity, 0.9 10 kg/m s.
0.8 1©3
For heat transfer in the tubes: Nu 0.023Re Pr .
Calculate:
122
(a) the effectiveness of the heat exchanger;
(b) the condensation rate;
(c) the fouling factor required on the water side if the effectiveness is to be increased to
0.7 for the same mass flow rate of water.
2
[Answers: (a) 0.337; (b) 11.3 kg/s; (c) 0.000049 m K/W]
10. In a closed-cycle gas turbine plant air from the compressor enters one side of a compact
`
heat exchanger at 150 C at a mass flow rate of 10 kg/s. The air leaving the turbine enters
`
the heat exchanger at 504 C and flows in counter-flow to the air. The heat exchanger has
2 2
a flow area of 0.144 m and an effective heat transfer area of 115.2 m per unit length in
the direction of flow on both the hot and cold sides of the heat exchanger. Calculate the
required length of the heat exchanger to obtain an effectiveness of 0.7.
[Assume that the heat exchanger surfaces are clean and neglect the thermal resistance
of the separating plates. For flow of air in the heat exchanger passages assume Nu
0.8 0.3
0.023Re Pr based on an equivalent diameter given by 4 (flow area)/(heated surface
area per unit length); take the properties at the mean temperature between the cold air inlet
and the hot air inlet. [Answer: 1.257 m]
11. Circular cross-section studs of radius 10 mm, length 100 m, thermal conductivity 24
W/mK are attached to a flat surface with their axes perpendicular to the surface on a square
pitch of 30 mm. The primary surface is at 300 C.
`
`
A fluid at 50 C is forced across the surface such that the mean heat transfer coefficient is
2
100 W/m K.
Calculate the rate of heat loss per unit area of studded surface.
[Assume that the heat transfer coefficient is the same for the primary surface and for the
rod surfaces.]
2
[Answer: 77.66 kW/m ]
LABORATORY EXERCISE
1. To calculate the overall heat transfer coefficient for parallel flow heat exchanger.
2. To calculate the overall heat transfer coefficient for counter current flow heat exchanger.
3. To find the heat transfer co-efficient for Drop-wise condensation.
4. To find the heat transfer co-efficient for Film-wise condensation process.
5. Determination of LMDT and Effectiveness in a Parallel Flow and
6. Counter Flow Heat Exchangers. Determination of Effectiveness on a Metallic fin.
References
th
John Wiley;
123
ids 4th edn Basil Blackwell; 2.6 ECKERT ER and DRAKE RM 1971 Analysis of Heat and Mass
Transfer Taylor and Francis; 2.7 KERN D Q 1950 Process Heat Transfer McGraw-Hill;
2.8 WALKER G 1990 Industrial Heat Exchangers 2nd edn McGraw-Hill;
2.9 KAYS w Mand LONDON AL 1984 Compact Heat Exchangers 3rd edn McGraw-Hill;
2.12 EASTOP TD and WATSON w E 1992 Mechanical Services for Buildings Longman.
124
Chapter 3
Heat transfer by radiation
Learning outcomes:
1. Explain the terminologies used in the study of radiation,
2. Describe heat transfer processes of radiation,
3. Derive the expressions heat transfer from and to different surfaces,
4. Apply derived equations of radiation to solve problems,
5. Relate given radiation mechanism and processes to their practical applications in thermo-
dynamics systems.
3.1 Introduction
‘Radiation’ heat transfer is defined as “the transfer of energy across a system boundary by means
of an electromagnetic mechanism which is caused solely by a temperature difference.” Whereas
the heat transfer by conduction and convection takes place only in the presence of medium,
radiation heat transfer does not require a medium.Both the amount of radiation and the quality
of radiation depend upon temperature. The dissipation from the filament of a vacuum tube or the
heat leakage through the evacuated walls of a thermos flask are some familiar examples of heat
transfer by radiation.
The contribution of radiation to heat transfer is very significant at high absolute temperature
levels such as those prevailing in furnaces,combustion chambers, nuclear explosions and in space
applications. The solar energy incident upon the earth is also governed by the laws of radiation.
Radiant energy (being electromagnetic radiation) requires no medium for propagation and will
pass through a vacuum. Heat transfer by radiation is most frequent between solid surfaces,
although radiation from gases also occurs. The energy which a radiating surface releases is not
continuous but is in the form of successive and separate (discrete) packet or quanta of energy
called photons. The photons are propagated through space as rays; the movement of swarm of
photons is described as electromagnetic waves. The photons travel (with speed equal to that
of light) in straight paths with unchanged frequency; when they approach the receiving surface,
there occurs reconversion of wave motion into thermal energy which is partly absorbed, reflected
or transmitted through the receiving surface (the magnitude of each fraction depends, upon the
nature of the surface that receives the thermal radiation).Certain gases emit and absorb radiation
on certain wavelengths only, whereas most solids radiate over a wide range of wavelengths.
125
All types of electromagnetic waves are classified in terms of wavelength and are propagated at
8
the speed of light (c) i.e., 3 × 10 m/s. The electromagnetic spectrum is shown in Figure 3.1.
The distinction between one form of radiation and another lies only in its frequency (f) and
wavelength (γ) which are related by
c γ f
7
The emission of thermal radiation (range lies between wavelength of 10 m and 10 m) de-
4
pends upon the nature, temperature and state of the emitting surface. However, with gases the
dependence is also upon the thickness of the emitting layer and the gas pressure.
Note: The rapidly oscillating molecules of the hot body produce electromagnetic waves in hy-
pothetical medium called ether. These waves are identical with light waves, radio waves and
8
X-rays, differ from them only in wavelength and travel with an approximate velocity of 3 10
m/s. These waves carry energy with them and transfer it to the relatively slow-moving molecules
of the cold body on which they happen to fall. The molecular energy of the later increases and
results in a rise of its temperature. Heat traveling by radiation is known as radiant heat.
Figure 3.1: Spectrum of electro-

magnetic radiation.
Thermal radiations exhibit characteristics similar to those of visible light, and follow optical
laws. These can be reflected, refracted and are subject to scattering and absorption when they
pass through a media. They get polarised and weakened in strength with inverse square of radial
distance from the radiating surface.
Some of the properties are:
• It does not require the presence of a material medium for its transmission.
• Radiant heat can be reflected from the surfaces and obeys the ordinary laws of reflection.
• It travels with velocity of light.
• Like light, it shows interference, diffraction and polarisation etc.
• It follows the law of inverse square.
The wavelength of heat radiations is longer than that of light waves, hence they are invisi-
ble to the eye.
3.2 Surface Emission Properties

The rate of emission of radiation by a body depends upon the following factors :
126
(i) The temperature of the surface,
(ii) The nature of the surface, and
(iii) The wavelength or frequency of radiation.
The parameters which deal with the surface emission properties are given below :
(a) Total emissive power (Ė).

The emissive power is defined as the total amount of radiation emitted by a body per unit
2
area per time. It is expressed in W/m . The emissive power of a black body, according to
Stefan - Boltzmann, is proportional to absolute temperature to the fourth power.
4 2
Ėb σ T W/m (3.1)
4
Ėb σ AT W (3.1a)
where, σ = Stefan-Boltzmann constant = 5.67 10 W/m K .

8 2 4
(b) Monochromatic (spectral) emissive power (Ė1 ). It is often necessary to determine the
spectral distribution of the energy radiated by a surface. At any given temperature the
amount of radiation emitted per unit wavelength varies at different wavelengths. For this
purpose the monochromatic emissive power Ėγ of the surface is used. It is defined as the
rate of energy radiated per unit area of the surface per unit wavelength.
The total emissive power is given by;
E0
in f ty
Ė Eγ dγ (3.2)
(c) Emission from real surface-emissivity. The emissive power from a real surface is given by
4
Ė εσ AT W (3.3)
where, ε = emissivity of the material.
Emissivity(ε). It is defined as the ability of the surface of a body to radiate heat. It is also
defined as the ratio of the emissive power of any body to the emissive power of a black
Ė
body of equal temperature i e.,ε
. Its values varies for different substances ranging
Ėb
from 0 to 1. For a black body ε = 1, for a white body surface ε = 0 and for grey bodies it
lies between 0 and 1. It may vary with temperature or wavelength.
(d) Intensity of radiation.
(e) Radiation density and pressure.
(f) Radiosity (J). It refers to all of the radiant energy leaving a surface.
(g) Interrelationship between surface emission and irradiation properties
127
3.3 Absorptivity, Reflectivity and Transmissivity
When incident radiation also called irradiation (defined as the total incident radiation on a
2
surface from all directions per unit time and per unit area of surface), expressed in W/m and
denoted by (G) impinges on a surface, three things happens ; a part is reflected back (Gr), a part is
transmitted through (Gt), and the remainder is absorbed (Ga) depending upon the characteristics
of the body, as shown in Figure 3.2.
Figure 3.2: Absorption, reflection

and transmission of radiation.
By the conservation of energy principle,
Ga Gr Gt G
Dividing both sides by G, we get

Ga Gr Gt G

G G G G
α ρ τ 1 (3.4)
where, α Absorptivity(or fraction of incident radiation absorbed),

ρ Reflectivity(or fraction of incident radiation reflected), and
τ Transmittivity(or fraction of incident radiation transmitted).
When the incident radiation is absorbed, it is converted into internal energy.
For most solids and liquids encountered in engineering the amount of radiation transmitted
through the substance is negligible, and it is possible to write
α ρ 1
Black body: For perfectly absorbing body, α = 1, ρ = 0, τ = 0. Such a body is called a ‘black
body’ (i.e., a black body is one which neither reflects nor transmits any part of the incident
radiation but absorbs all of it). In practice, a perfect black body (α = 1) does not exist. However
128
its concept is very important. Opaque body: It is a body on whose there is no incident radiation
is transmitted through it.
For the opaque body τ = 0, and equation reduces to
α ρ 1 (3.5)
Solids generally do not transmit unless the material is of very thin section. Metals absorb radi-
ation within a fraction of a micrometre, and insulators within a fraction of millimetre. Glasses
and liquids are, therefore, generally considered as opaque.
White body: If all the incident radiation falling on the body are reflected, it is called a ‘white
body’. For a white body, ρ = 1, α = 0 and τ = 0.
Gases such as hydrogen, oxygen and nitrogen (and their mixture such as air) have a transmissivity
of practically unity. Reflections are of two types (Figure 3.3): 1. Regular (specular) reflection 2.
Diffuse reflection.
Figure 3.3: Regular and diffuse re-
flections.
Specular reflection - implies that angle between the reflected beam and the normal to the surface
equals the angle made by the incident radiation with the same normal. Reflection from highly
polished and smooth surfaces approaches specular characteristics.
Diffused reflection - the incident beam is reflected in all directions. Most of the engineering
materials have rough surfaces, and they give diffused reflections.
Grey body: If the radiative properties, α, ρ, τ of a body are assumed to be uniform over the
entire wavelength spectrum, then such a body is called grey body. A grey body is also defined
as one whose absorptivity of a surface does not vary with temperature and wavelength of the
incident radiation [α α γ = constant.] A coloured body is one whose absorptivity of a surface
varies with the wavelength of radiation [[α j α γ ].
3.4 Black-body radiation

As already stated an ideal body which absorbs all the radiation which falls upon it is called a
black body. For a black body, α = 1 and ρ = 0. It should be noted that the term ‘black’ in this
context does not necessarily imply black to the eye. A surface which is black to the eye is one
which absorbs all the light incident upon it, but a surface can absorb all the thermal radiation
incident upon it without necessarily absorbing all the light ( e.g. snow is almost ‘black’ to
thermal radiation, α = 0.985). Although no totally black body exists in practice, many surfaces
approximate to the definition. For example, consider a small object radiating energy in a large
space, as shown in Figure 3.4. The energy striking the surface surrounding the body is reflected
129
and absorbed many times by the surface, and the fraction of energy reflected back and intercepted
by the body is exceedingly small.
Therefore, when a body is placed in large surroundings
the surroundings are approximately black to thermal ra-
diation.
As a better example of a black body, consider a small
hole in the surface of a wall, as shown in Figure 3.5.
The hole leads into a small chamber as shown. Rays
of thermal radiation entering the hole are successively
absorbed by the walls of the chamber such that only a
negligible amount of radiation is emitted from the hole.
Thus the hole acts as a black body. This is the closest
approximation to a black body which can be devised in Figure 3.4: Radiation from a small
practice; the inside surfaces of the chamber can be made body in large surroundings.
of a material with a high absorptivity ( e.g. lampblack).
Another example of a black body is a hollow enclosure
with a very small hole for the passage of incident radi-
ation as shown in Figure 3.6. Again, a small fraction of
incident radiant energy is reflected as it strike the sur-
face. After a number of such reflections the amount un-
absorbed is exceedingly small and very little of the original incident energy is reflected back out
of the opening. A small hole leading into a cavity (Hohlraum) thus acts very nearly as a black
body because all the radiant energy entering through it gets absorbed. Isothermal furnaces, with
small apertures, approximate a black body and are frequently used to calibrate heat flux gauges,
thermometers and other radiometric devices.
Figure 3.6: Concept of a black

body.
It can be shown that a black body, as well as being the

best possible absorber of radiation, is also the best pos-
sible emitter. Consider an enclosure at a uniform tem-
perature, and let a black body be placed in the enclo-
sure as shown in Figure 3.7. If the body is at the same
temperature as the enclosure then it follows that all the
energy radiated by the body and absorbed by the walls
130
of the enclosure, must exactly equal the energy radiated
by the enclosure and absorbed by the body. If this were
not so then the body would gain or lose energy, and this
is not possible in an isolated system, by the laws of ther-
modynamics. Let the emissive power of the black body
be EB. Therefore the rate at which energy impinges on
unit surface of the black body is also ĖB .
Now replace the black body by any other body at the
same temperature, and of the same shape and size. This
body must receive exactly the same amount of energy
from the enclosure as the black body received when it
was in the same position in the enclosure. However, this body is not black and hence will only
absorb a fraction of the energy it receives,
i.e., Rate of energy absprotion α ĖB

where α is the absorptivity of the body.
Now as before the energy absorbed must be equal to the
energy emitted, therefore if the body has an emissive
power of Ė, we have
Ė α E˙B
Ė Figure 3.7: Radiation from a body to a
or, α (3.6)
E˙B surrounding enclosure.
Since α $ 1 then Ė $ ĖB , and hence the black body is

the best possible emitter of radiation. The ratio of the
emissive power of a body to the emissive power of a
black body is called the emissivity, ε.
3.5 Kirchhoff’s Law

The law states that at any temperature the ratio of total
emissive power Ė to the total absorptivity α is a con-
stant for all substances which are in thermal equilib-
rium with their environment.
Let us consider a large radiating body of surface area A
which encloses a small body (1) of surface area A1 (as
shown in Figure 3.8. Let the energy fall on the unit sur-
face of the body at the rate Ėb of this energy, generally,
a fraction α, will be absorbed by the small body. Thus
this energy absorbed by the small body (1) is α1 A1 Ėb ,
in which α1 is the absorptivity of the body. When ther-
mal equilibrium is attained, the energy absorbed by the
body (1) must be equal to the energy emitted, say, E1
per unit surface. Thus, at equilibrium, we may,
131 Figure 3.8: Derivation of Kirchhoff’s

law..
A1 Ė1 al pha1 A1 Ėb (3.7)
Now we remove body (1) and replace it by body (2)

having absorptivity α2 . The radiative energy impinging
on the surface of this body is again Ėb . In this case, we
may write,
A2 Ė2 al pha2 A2 Ėb (3.8)

By considering generality of bodies, we obtain
Ė1 Ė2 Ė
Ėb α1 α2 α (3.9)
Also, as per definition of emissivity ε, we have
Ė
ε
Ėb
Ė
or, Ėb ε (3.10)
By comparing equations (3.9) and (3.10), we obtain
ε α (3.11)
(α is always smaller than 1. Therefore, the emissive power Ė is always smaller than the emissive
power of a black body at equal temperature). Thus, Kirchhoff’s law also states that the emissivity
of a body radiating energy at a temperature, T, is equal to the absorptivity of the body when
receiving energy from a source at a temperature, T.
3.6 The grey body

In the foregoing sections it has been assumed that the energy emitted by thermal radiation is the
same for all wavelengths of the radiation. In fact this is not the case, and Figure 3.9 shows the
emissive power per unit wavelength plotted against wavelength, λ , in micrometres for a black
body at any one temperature. A corresponding curve at the same temperature is shown for a
non-black body. The ratio of an ordinate of each curve at any wavelength gives the emissivity,
and hence the absorptivity, at that wavelength. For example, at a wavelength of 4.5 micrometres
(µm), we have
AB
ελ αλ
AC
The terms ελ ; and αλ , are called the monochromatic emissivity and the monochromatic absorp-
tivity respectively.
It can be seen from Figure 3.9 that the monochromatic emissivity varies with wavelength. The
variation is greater for some materials than for others, and there are certain materials for which
132
Figure 3.9: Emissive power against
wavelength for a black body and a
non-black body.
the emissivity is practically constant over the entire waveband (e.g. slate rock). To simplify
calculations, surfaces in practice are very often assumed to have a constant emissivity over all
wavelengths and for all temperatures. Such an ideal surface is called a grey body. Then, for
a grey body,α and ε at all temperatures, where α ε are the total absorptivity and the total
emissivity over all wavelengths.
It is an experimental fact that the emissive power of a body increases as the temperature of the
body is increased. This is illustrated in Figure 3.10 in which the emissive power of a black body
per unit wavelength is plotted against the wavelength in micrometres, for several temperatures.
It can be seen that the wavelength which gives maximum emissive power becomes smaller as
the temperature is increased, and hence more and more of the energy emitted is radiated over
the shorter wavelengths as the temperature increases. The value of the wavelength for maximum
emissive power is given by Wien’s law
2900
i.e., λmax (3.12)
T
where λmax is in micrometres and T is in K.
Figure 3.10: Emissive power
against wavelength for a black
body and a non-black body.
The limits of the visible spectrum are λ = 0.4 micrometre at the blue end and λ = 0.8 micrometre
at the red end. Now the sun has a temperature of approximately 6000 K, hence, using equation
(3.12), the maximum wavelength of the radiation is
2900
λmax 0.483 micrometres
6000
133
Therefore most of the thermal radiation from the sun is in the visible waveband. The waveband
for light is shown shaded in Figure 3.10. At a temperature of 800 K a very small amount of the
energy emitted is just within the red end of the visible spectrum. A surface at 800 K will appear
as a dull red colour. At about 1250 K more of the energy emitted is in the visible range and the
surface is then said to be red-hot. The temperature of the filament of an electric light bulb is
approximately 2800 K, and even at this temperature only about 10% of the energy emitted is in
the visible region, which shows the inefficiency of such a bulb as a light source.
For a grey body a set of curves exactly similar to those of Figure 3.10 can be drawn, with
each ordinate only a fraction, ε, of the corresponding ordinate of the curves of Figure 3.10. In
practice, although a suitable total value of the absorptivity may be taken for a large number of
industrial surfaces over a wide range of wavelengths, nevertheless there is still a variation of total
absorptivity with temperature. This is illustrated in Figure 3.11. When the temperature range is
small the approximation that α ε = constant, for a grey body, is still sufficiently accurate
for most calculations. Materials or surfaces for which the emissivity varies considerably and
irregularly with wavelengths and temperature are called selective emitters. Some values of total
emissivity over all wavelengths but for different temperatures are shown in Table 3.1.
Figure 3.11: Absorptivity against

surface temperature for various ma-
terials.
Surface finish plays a large part in determining the emissivity of a material. When the surface is
very smooth it reflects radiation specularly; when the surface is rough, as in most practical cases,
it reflects diffusely. Rough surfaces are much better absorbers and hence much better emitters -
of radiation than smooth surfaces.
Table 3.1: Emissivities of some surfaces at various temperatures
Enmissivity
Surface
`
0-40 C
`
120 C
`
260 C
`
540 C
White paint 0.95 0.94 0.88 0.70
Black glossy paint 0.95 0.94 0.90 0.85
Lampblack 0.97 0.97 0.97 0.97
Building brick 0.93 0.93 0.79 0.74
Concrete 0.85 0.84 0.69 0.69
Polished steel 0.07 0.09 0.11 0.14
134
For mild steel, rough turned, ε at 15`C is 0.87; for mild steel, well finished on a lathe, ε at 15`C
is 0.39; and it can be seen from Table 3.1 that when the steel is polished the emissivity, εe, is
reduced to 0.07.
3.7 The Stefan-Boltzmann law

It was found experimentally by Stefan, and proved theoretically by Boltzmann, that the emissive
power of a black body is directly proportional to the fourth power of its absolute temperature,
and this is known as the Stefan-Boltzmann law,
4
i.e., ĖB σT (3.13)
the value of σ is 5.67 10
8 W ©m2 K 4 . The rate of energy emitted by a non-black body is then
given by
4
ĖB εσ T (3.14)
where ε is the emissivity of the body.
Consider a body 1 of emissivity ε1 at a temperature T1 , completely surrounded by black sur-
roundings at a lower temperature T2 . The energy leaving body 1 is completely absorbed by the
surroundings, and from equation (3.14).
4
Rate of energy emission ĖB ε1 σ T1
The rate of energy emitted by the black surroundings is given by equation (3.13)
4
i.e., ĖB σ T2
Now the fraction of this energy which is absorbed by body 1 depends on the absorptivity of body
1. For a grey body α ε at all temperatures and hence,
4 4
Rate of energy absorption εσ T1 ĖB εαT1
Then the rate of heat transferred from the body to its surroundings per square metre of the body
is
4 4
q̇ εσ T1 εσ T2
4 4
i.e., q̇ εσ T1 T2 (3.15)
If the emissivity of the body at T1 is largely different from the emissivity of the body at T2 then
the approximation of the grey body may not be sufficiently accurate. In that case it is a good
approximation to take the absorptivity of the body 1 when receiving radiation from a source at
T2 as being equal to the emissivity of body 1 when emitting radiation at T1 .
4 4
Then, q̇ εT1 σ T1 εT1 σ T2 (3.16)
135
The absorptivity, while depending mainly on the temperature of the source of radiation, also
depends on the temperature of the surface itself. For most metals this factor can be important
and it has been shown that the absorptivity of a metal surface at T1 for radiation from a source at
T2 is approximately equal to the emissivity of the surface when at a temperature, T3 , given by
Ô
T3 T1 T2 (3.17)
Example 3.1
` `
A body at 1000 C in black surroundings at 500 C has an emissivity of 0.42 at 1000 C and an
`
` 2
emissivity of 0.72 at 500 C. Calculate the rate of heat loss by radiation per m .
(a) When the body is assumed to be grey with ε = 0.42.
(b) When the body is not grey.
Assume that the absorptivity is independent of the surface temperature.
Solution:
(a) Grey body:
`
Given: T1 = 1000 +273 = 1273, ε at 1000 C = 0.42
`
T2 = 500 +273 = 773, ε at 500 C = 0.72
8
σ 5.67 10
Heat loss per m2 by radiation using equation (3.15),
4 4
i.e., q̇ εσ T1 T2 ,
8 4 4
and, q̇ 0.42 5.67 10 1273 773 54893 W
i.e., Rate of heat loss per square metre by radiation 54.893kW

(b) the body is not grey:
` `
Absorptivity when the source is at 500 C is = the emissivity when the body is at 500 C,
i.e., absorptivity,α 0.72

4 8 4
Then,energy emitted εσ T1 0.42 5.67 10 1273
4 8 4
and, energy absorbed εσ T1 0.72 5.67 10 773
i.e., q̇ Energy emitted - Energy absorbed
8 4 8 4
0.42 5.67 10 1238 0.72 5.67 10 773
62538 14576 47962 W
2
i.e., Heat loss per m by radiation 47.962 kW.
Rate of heat loss per square metre by radiation 47.962 kW

It can be seen that the grey body assumption of part (a) overestimates by
136
54.893 47.962
100 14.5%
47.962
Example 3.2
Calculate the rate of heat loss by radiation from unit surface area of a body at 1100 C in black
`
` ` `
surroundings at 40 C, when the emissivity at 40 C is 0.9, and the emissivity at 1100 C is as in
Example 3.1:
(a) when the body is grey with ε = 0.4;
(b) when the body is not grey.
Assume that the absorptivity is independent of the surface temperature. Solution:
(a) Grey body:
`
Given: T1 = 1100 +273 = 1373, ε at 1100 C = 0.4
`
T2 = 40 +273 = 313, ε at 40 C = 0.9
8
σ 5.67 10
Heat loss per m2 by radiation using equation (3.15),
4 4
i.e., q̇ εσ T1 T2 ,
8 4 4
and, q̇ 0.4 5.67 10 1373 313 80380 W
i.e., Rate of heat loss per square metre by radiation 80.38kW

(b) the body is not grey:
` `
Absorptivity when the source is at 500 C is = the emissivity when the body is at 500 C,
i.e., absorptivity,α 0.72

4 8 4
Then,energy emitted εσ T1 0.4 5.67 10 1373
4 8 4
and, energy absorbed εσ T1 0.9 5.67 10 313
i.e., q̇ Energy emitted - Energy absorbed
8 4 8 4
0.4 5.67 10 1373 0.9 5.67 10 313
80598 490 80108 W
2
i.e., Heat loss per m by radiation 80.108 kW.
Rate of heat loss per square metre by radiation 80.108 kW

It can be seen that the grey body assumption of part (a) overestimates by
80.38 80.11
100 0.337%
80.11
137
It can be seen from Examples 3.1 and 3.2 that the grey body assumption gives a very accurate
approximation when one of the temperatures is small compared with the other. The assumption
also gives a very accurate approximation when both temperatures are small.
3.8 Planck’s Law

In 1900 Max Planck showed by quantum arguments that the spectral distribution of the radiation
intensity of a black body is given by;
2
2πc h 1
Ėλ &b (3.18)
λ5 hc

e λ kT 1
where, Ėλ b Monochromatic (single wavelength) emissive power of a black body,

c Velocity of light in vacuum,2.998 10
8
3 108 m/s,
34
h Planck’s constant 6.625 10 js,
λ Wavelength,µm,
23
k Boltzmann constant 1.3805 10 J/K, and
T Absolute temperature, K.
2
Hence the unit of E˙λ b is W/m µm
Quite often the Planck’s law is written as,
C1 1
Ėλ b (3.19)
lambda5
C2

e λT 1
2 8 4 2
where, C1 2πc h 3.742 10 W µm ©m ;
ch 4
C2 1.438810 µmK
k
Equation (3.18) is of great importance as it provides quantitative results for the radiation from a
black body.
3.9 Wien’s Displacement Law

In 1893 Wien established a relationship between the temperature of a black body and the wave-
length at which the maximum value of monochromatic emissive power occurs. A peak monochro-
matic emissive power occurs at a particular wavelength. Wien’s displacement law states that the
product of λmax and T is constant, i.e.,
138
λmax T constant (3.20)
C1 1
and, Ėλ b
λ 5 e C2 ©λ T 1
Ėλ b becomes maximum (if T remains constant) when
d Ėλ b
0
dλ
d Ėλ b d C1 1
i.e., 0
dλ dλ λ e 2 T 1
5 C ©λ
e
C2 ©λ T
1 5C1 λ
6 C λ 5 ve C2 ©λ T C2
2
|
1
1
T λ
and, 0
e C2 ©λ T 12
6 C2 ©λ T 6 5 1 C2 ©λ T
or, 5C1 λ e 5C1 λ C1 C2 λ e 0
λ 2T
Dividing both side by 5C1 λ

6 , we get
C2 ©λ T 1 1 C2 ©λ T
e 1 C2 e 0
5 λT
Solving this equation by trial and error method, we get
C2 C2
or 4.965
λ T λmax T
4
C2 1.439 10

λmax T
4.965 4.965
µmk 2898 µmk 2900 µmk
i.e., λmax T 2898 µmk (3.21)
This law holds true for more real substances ; there is however some deviation in the case of a
metallic conductor where the product λmax T is found to vary with absolute temperature. It is
used in predicting a very high temperature through measurement of wavelength. A combination
of Planck’s law and Wien’s displacement law yields the condition for the maximum monochro-
matic emissive power for a black body.
C1 λmax
5
Ėλ b
eC2 ©λmax T 1
3 5

0.374 10
15 2.898 10

T
2 ©2.898103
e1.438810 1
139
5 5 2
Ėλ bmax 1.285 10 T W/m per metre wavelength (3.22)
3.10 Intensity of radiation, Lambert’s law and the geometric

factor
When a surface element emits radiation, all of it will be intercepted by a hemispherical surface
placed over the element. The intensity of radiation (I) is defined as the rate of energy leaving
a surface in a given direction per unit solid angle per unit area of the emitting surface normal
to the mean direction in space. A solid angle is defined as a portion of the space inside a
sphere enclosed by a conical surface with the vertex of the cone at the centre of the sphere. It is
measured by the ratio of the spherical surface enclosed by the cone to the square of the radius
of the sphere ; it unit is steradian (sr). The solid angle subtended by the complete hemisphere is
2
2πr
given by: 2 2π Figure 3.12 (a) shows a small black surface of area dA (emitter) emitting
r
radiation in different directions. A black body radiation collector through which the radiation
pass is located at an angular position characterised by zenith angle θ towards the surface normal
and angle φ of a spherical coordinate system. Further the collector subtends a solid angle dω
when viewed from a point on the emitter. Let us now consider radiation from the elementary
area dA1 at the centre of a sphere as shown in Figure 3.12. Suppose this radiation is absorbed by
a second elemental area dA2 , a portion of the hemispherical surface.
The projected area of dA1 on a plane perpendicular to the line joining dA1 and dA2 = dA1 cos θ .
dA2
Then, dA2
r2
(a) Special distribution of radi-(b) Illustration for evaluating

ations emitted from a surface area dA2
Figure 3.12: Radiation from an elementary surface.
Most surfaces do not emit radiation strongly in all directions; the greater part of the energy emit-
ted is in a direction normal to the surface. Before considering the interchange of energy between
two bodies which receive only a part of the radiation emitted by each other, it is necessary to find
out how the radiation is distributed in the various directions from the two surfaces. The rate of
energy emission from unit surface area through unit solid angle, along a normal to the surface,
is called the intensity of normal radiation, iN . The intensity of radiation in any other direction
at any angle φ to the normal is denoted by iφ , (Note: a surface subtends a solid angle at a point
2
distance r from all points on the surface, equal to the surface area divided by r . The surface of
140
2
a sphere is 4πr and hence the solid angle subtended by the surface of the sphere at its centre is
4π.)
The variation in the intensity of radiation is given by Lambert’s cosine law,
i.e., iφ iN cos φ (3.23)
The rate of energy emission from a surface of area dA is then given by,
E iφ dw dA
where dw is a small solid angle.
Consider a small area dA, and consider the radia-
tion from dA which passes through a small element
of the surface area of a hemisphere with dA at its
centre, as shown in Figure 3.13. The element sub-
tends an angle φ at the centre of the hemisphere
and the small increase in angle over the width of
the element is then dφ . The width of the element is
the length of the arc, of angle dφ , and radius r (i.e.
AB in Figure 3.13). Therefore,
Width of element, AB r dφ
The radius of the element is CA = r sin φ . Hence Figure 3.13: Radiation from a small ele-
the surface area of the element is given by, ment to a hemisphere for Q4(a).
Surface area width circumference r dφ 2πr sin φ
the circumference of circle 2π Radius
2
2πr sin φ dφ
i.e., Solid angle, dw, subtended at dA
r2
2
area subtended by 1 steradian of a solid angle = r
and, dw 2π sin φ dφ
Hence the rate of total energy emission, Ė from dA is given by,
E0 E0
π ©2 π ©2
ĖdA iφ dw dA dAiφ 2π sin φ dφ
Substituting for iφ from equation (3.23), iφ iN cos φ , then
E0
π ©2
ĖdA 2πdAiN cos φ sin φ dφ
141
E0
π ©2 sin 2φ
2πdAiN dφ πiN dA
2
4
Now from equation, Ė εσ T
4
Thus, εσ T dA πiN dA
4
εσ T
i.e., iN π (3.24)
Consider two small black surfaces of area dA1 and dA2

at temperatures T1 and T2 , and distance x apart. The
angles of inclination of surfaces are as shown in Fig-
ure 3.14. This is a case where neither body receives all
the radiation from the other. Let the surface dA2 sub-
tend a solid angle dw1 at the centre of the surface dA1 .
Then we have,
Rate of energy emission from dA1 incident on dA2
iN1 cos φ1 dw1 dA1
4
σT Figure 3.14: Radiation interchange be-
From equation (3.24), iN π , for a black surface (ε tween two small surfaces in large sur-
= 1), therefore,
roundings.
4
σ T1 cos φ1 dw1 dA1
i.e., rate of energy incident on dA2 π
Also, from the definition of solid angle (solid angle, Ω,

in steradian, if A is the area of a part of the spherical
surface, and r is the radius of the sphere, then the solid
2
angle is given as, Ω A©r )
dA2 cos φ2
dw1
x2
4 dA2 cos φ2
σ T1 cos φ1 dA1
Hence, rate of energy incident on dA2 x2
π
4
Now the rate of total energy emission from dA1 is σ dA1 T1 . The ratio of the energy incident on
the second body to the energy emitted by the first is called the geometric factor, F12 ,
4
σ T1 cos φ1 dA2 cos φ2 dA1
i.e., F12
πx2 σ dA1 T41
cos φ1 cos φ2 dA2

or, F12 (3.25)
πx2
142
In the same way it can be shown that the geometric factor for radiation from surface 2 to surface
1 is given by,
cos φ1 cos φ2 dA1

or, F21 (3.26)
πx2
The net rate of energy interchange between the surfaces is given by,
σ cos φ1 cos φ2 dA1 dA2 4 4

Q̇12 2
T1 T2
πx
This can be written as,

4 4
Q̇12 F12 dA1 σ T1 T2
4 4
or, Q̇12 F21 dA1 σ T1 T2
For a larger area made up of small areas dA1 and dA2 , average geometric factors can be defined
in the same way as above,
4 4
i.e., Q̇12 F12 A1 σ T1 T2 (3.27)
4 4
and, Q̇12 F21 A1 σ T1 T2 (3.28)
From equations (3.27) and (3.28) it can be seen that,
A1 F12 A2 F21 (3.29)
This is known as the reciprocal relationship or theorem of reciprocity.

In practice calculating F can be a long and difficult process except for simple shapes; charts are
available for some of the more common configurations (see, for example, references 3.1 and
3.10). When a body, 1, is completely enclosed by other surfaces then
F1sur f aces 1
If the surfaces have separate elements, 2, 3, etc. it follows that,
F1sur f aces F11 F12 F13 etc. 1 (3.30)
The term F11 is necessary in cases where the body 1 can ’see’ parts of itself, e.g. a concave
body.
In Table 3.2 values of geometric factor, F12 , are given for some common configurations; for
more complex geometries see references 3.1 and 3.10.
143
Table 3.2: Emissivities of some surfaces at various temperatures
Configuration Geometric factor, F12

(a) Body 1 complete enclosed 1 by body 2 1
(b) Parallel circular discs, radii
2 2 2 Ó 2 2 2 2 2 2
x r1 r2 ux1 r1 r2 4r1 r2 {
r1 and r2 , distance x apart
2r21
on a common-axis
(c) Small disc opposite a parallel
2
R
circular plate of radius R
R2 L2
at a perpendicular distance L
(d) Small sphere opposite a circular
1 L
plate of radius R at a v1 Ó 2 |
2 R L2
perpendicular distance L
(e) Small sphere at the centre
L
of the axis of a cylinder Ó
R2 L2
of radius R and length 2L
Example 3.3
Solution:
A hemispherical cavity of 0.6 m radius is covered by a plate with a hole of 0.2 m diameter drilled
`
in its centre. The inner surface of the plate is maintained at 250 C by a heater embedded in the
surface. The surfaces may be assumed to be black and the hemisphere may be assumed to be
well insulated. Calculate:
(a) the temperature of the surface of the hemisphere;
(b) the power input to the heater.
State any other assumption made.
Solution:
(a) the temperature of the surface:

Let the inner surface of the plate be 1, the hemisphere surface 2, and the hole projected surface
3, as shown. Then, since surface 1 is completely surrounded, we have
F12 F13 1
or, F12 1 since surface 1 cannot ‘see’ surface 3.
144
From equation, A1 F12 A2 F21
2 2
A1 F12 π 0.6 0.1 35
F21
A2 2π 0.62 72
Similarly,
F32 1 and A2 F23 A3 F32
2
π 0.1 1 1

F23 Figure 3.15: Hemispherical cavity for
2π 0.62 72
Q4(b).
4 4
Then, Rate of energy emission from surface 2 A2 F23 σ T2 A2 F21 σ T2
4 1 35
A2 σ T2
72 72
4
A2 σ T2 0.5
The rate of energy incident on surface 2 may be taken as the rate of energy emission from surface
1, since the rate of energy entering the hole from outside will be negligible if the surroundings
are large and at normal temperature,
4 4
i.e., Rate of energy incident on surface 2 A1 F12 σ T1 A1 σ T1
4 4
Then, A1 σ T1 A2 σ T2 0.5for the steady state
4 2 2
4 T1 π 0.6 0.1 4 35
i.e., T2 T1
2π 0.62 0.5 36
35 1©4 `
i.e., T2 250 273
519.3 K 246.3 C
36
(b) Rate of heat input from heater:

4 4 2 2 5.67 4 35
A1 F12 σ T1 T2 π 0.6 0.1 8
523 1
10 36
129.6 W
145
3.11 Radiant interchange between grey bodies:
Radiosity, J̇ - is the total radiant energy leaving a body per unit area per unit time, and
Irradiation, Ġ - Is the total radiant energy incident on a body per unit area per unit time. Hence
Net heat transfer from body, Q̇ J̇ ĠA (i)
where A is the area of the body surface. For a black body,

4
J̇ J̇T ;
where, J̇ is the Stefan-Boltzmann constant.

For a grey body, the radiosity must include the fraction of energy which is reflected from the
surface,
4
i.e., J̇ εσ T ρ Ġ
where, the fractions of the radiation is emitted (represented by emissivity, ε) is absorbed, the
fraction reflected, and fraction transmitted are called the absorptivity, α, the reflectivity, ρ, and
the transmissivity, τ, respectively.
Also, for a grey body, equation (3.6), ε α 1 ρ, neglecting transmissivity.
4

J̇ εσ T 1 ε Ġ
4
J̇ εσ T
or, Ġ
1ε
Substituting for Ġ in equation (i),

4
Q̇ J̇ εσ T
i.e., J̇ Ġ J̇
A 1ε
εA 4
or, Q̇ σ T J̇ (3.31)
1ε
For any two bodies 1 and 2, the geometric factor, F12 , is the fraction of radiation A1 J̇1 which is
intercepted by body 2,
i.e., Q̇12 A1 F12 J̇1 A2 F21 J̇2
Using equation (3.29), A1 F12 A2 F21
i.e., Q̇12 A1 F12 J̇1 J̇2 (3.32)
An electrical analogy can be used based on Ohm’s law. For example, from equation (3.31)
1ε
Resistance due to emissivity of surface (3.33)
εA
4
where Q̇ is analogous to current and σ T J̇ is analogous to potential difference. Similarly,
from equation (3.32),
146
1
Resistance due to geometry (3.34)
A1 F12
Take the simple case of a body 1, completely enclosed by a body 2. Figure 3.16 shows the
electrical analogy.
1 ε1 1 1 ε2
Total resistance, RT
A1 ε1 A1 F12 A2 ε2
Also in this case, F12 = 1, therefore
1 1 A1 1 1 1 A1 1
RT 11 v 1|
v 1|
A1 ε1 A2 ε2 A1 ε1 A2 ε2
4 4 4 4
σ T1 T2 A1 σ T1 T2
Q̇12 (3.35)
RT 1©ε1 A1 ©A2 r 1©ε2 1x
Figure 3.16: Electrical analogy for

radiation from body 1 enclosedby
body 2.
When the bodies are very close together then A1 A2,
4 4
A1 σ T1 T2
i.e., Q̇12
1©ε1 r 1©ε2 1x
The latter expression for the heat transfer also applies to the case of two large flat parallel surfaces
where the size of the surfaces is large compared with their distance apart, i.e. the radiant energy
escaping to the surroundings is negligible.
3.12 Gas radiation

In the problems considered in the previous sections on radiation the effect of the transmission of
radiation through the gaseous atmosphere has been neglected; some radiation will be absorbed
by the surrounding gases in such cases, but this is normally so small that it can be neglected.
Certain types of gases are transparent to thermal radiation; these include inert gases (e.g. argon)
and gases with symmetric diatomic molecules (e.g. oxygen and nitrogen). Hence for radiation
between surfaces in the normal atmospheric environment the effect of the surrounding gas can be
ignored. For gases with certain types of asymmetric molecular structures (e.g. carbon dioxide,
carbon monoxide, sulphur dioxide, nitrous oxide, and water vapour), radiation is absorbed from,
and emitted to, surrounding surfaces. Due to the rotational and vibrational motions within any
gas molecule, the radiation absorbed is dependent on the frequency of the radiation striking the
molecule. Absorption and emission of radiation in gases is therefore selective, occurring in only
certain bands of wavelengths.
When considering radiation within a furnace, or any enclosure containing combustion gases, it is
necessary to allow for the absorption of radiation due to the presence of CO2 , H2 O, and perhaps
147
CO and S02 . A simplified procedure for the calculation of absorption and emission of radiation
in such gases was first suggested by Hottel; this is summarized in the chapter on radiation written
by Hottel in reference 3.10. Also included is the radiation from flames made luminous by the
thermal decomposition of hydrocarbons.
3.13 The greenhouse effect

The so-called greenhouse effect on the earth is caused by the absorption of the sun’s rays by
gases, mainly carbon dioxide, in the atmosphere. The greenhouse effect is essential for the
survival of life on the planet since without it the earth’s surface would be rapidly cooled to a
`
temperature of about -20 C. Gases are selective absorbers of radiation and the major greenhouse
gases, such as carbon dioxide and methane, absorb the long-wavelength radiation from the sur-
face of the earth much more readily than the low-wavelength, high-temperature radiation from
the sun to the earth’s surface. Hence there is a net transfer of heat from the sun to the earth,
maintaining the temperatures that sustain life as we know it.
As long as there has been life on earth there has been a build-up of carbon dioxide in the at-
mosphere; one of our main constituents, carbon, oxidizes to carbon dioxide as we breathe, and
whenever wood, vegetable matter, or fossil fuel is burned. Methane is continuously released by
sheep and cattle, from any shallow wetland or marsh, and by certain insects such as termites.
The reason why there is now international concern is the realization that a dangerously high
level of greenhouse gases has built up over the last two centuries, and particularly over the last
50 years or so. This is mainly due to an acceleration in the world-wide combustion of fossil
fuel; it is estimated that 19 000 million tonnes per year of carbon dioxide are released into the
atmosphere from combustion of coal, oil, and gas. Methane concentration has also increased
markedly due to the rapid growth in intensive farming of livestock and of crops such as rice.
Other gases that are identified as causing the increase in the greenhouse effect are CFCs (Refer
to studies in Refrigeration and Air conditioning ), nitrous oxide from vehicle exhausts and power
station chimneys, and ozone from photochemical smog; the Montreal protocol should solve the
problem of CFCs by the end of this century, although CFCs are estimated to cause only 15% of
the present increase in the greenhouse effect.
The average temperature of the earth’s surface is known to be rising although the exact increase
due to the increased greenhouse effect is masked by natural cyclic fluctuations due to, for exam-
ple, changes to the sun itself. The main concern about the relatively small temperature increases
occurring in the earth due to the increased greenhouse effect is in the increase in the melting of
the polar ice-caps. This will cause increased sea-levels leading to catastrophic flooding; more
areas of the world would turn into desert, but other areas would in turn become warmer and able
to sustain crops more abundantly. The main solution to this problem is to effect a reduction in the
consumption of fossil fuels by first applying the principles of the efficient use of existing energy,
and secondly by vigorously introducing and improving renewable energy-conversion methods
using solar, wind, tidal, wave, and sea devices.
Question ONE
A thermos flask consists of an inner cylindrical vessel of 60 mm outside diameter and an outer
cylindrical vessel of 65 mm inside diameter. Both surfaces are of polished silver, emissivity 0.02.
Calculate the rate of heat loss per millimetre length of the flask when it contains boiling water
`
and the temperature of the outside surface is 17 C. Neglect the thermal resistance of the metal
walls of the flask. (NB Polished surfaces reflect specularly and hence in this case the surfaces
act as large parallel planes.)
148
[Answers: 0.00133 W]
Question TWO
In a muffle furnace the floor, 4.5 m by 4.5 m, is constructed of refractory material (emissivity =
0.7). Two rows of oxidized steel tubes are placed 3 m above and parallel to the floor, but for the
purpose of analysis these can be replaced by a 4.5 m by 4.5 m plane having an effective emissivity
`
of 0.9. The average temperatures for the floor and tubes are 900 and 270 C respectively.
Taking the geometric factor for radiation from floor to tubes as 0.32, calculate:
(a) the net rate of heat transfer to the tubes;
(b) the mean temperature of the refractory walls of the furnace, assuming that these are well
insulated.
[Answers: (a) 1009kW; (b) 687 C ]
`
Tutorial
1. A cylindrical electrode of radius, r, length, L, is immersed in a liquid which remains at
a constant temperature when the current density in the electrode is J. The heat transfer
coefficient, α, from the outside surface of the electrode may be assumed to be constant
over the entire surface.
Assuming steady-state conditions derive the differential equation
2
d ∆t 2α J s
∆t 0
dx rλ λ
where ∆t is the temperature difference between the electrode and the fluid at any distance
x from the end of the electrode, ,λ , the thermal conductivity of the electrode material, and
s the electrical resistivity of the electrode material.
Hence show that for the case when the rate of heat loss through the lead and support at
each end is a fraction, y, of the total electrical input, then
2
J s mly cosh m x L©2
∆t v1 |
2
m λ sinh mL©2
1©2
where m = 2α ©rλ
2. A cylindrical storage tank, 1 m diameter by 1.1 m long, has an outside surface temperature
`
of 60 C, and an emissivity of 0.9. Assuming that the tank is a grey body, calculate:
(a) the rate of heat loss by radiation when the tank is in a large room, the walls of which
`
are at 15 C;
(b) also the reduction in the rate of heat loss by radiation if the tank is painted with
aluminium paint of emissivity 0.4.
[Answers:1474 W; 819 W]
` `
3. A copper pipe at 260 C is in a large room at 15 C. Calculate the rate of heat loss per unit
`
area of pipe surface by radiation, taking the emissivity of copper as 0.61 at 260 C, and as
`
0.56 at 15 C. Assume that the absorptivity of a surface depends only on the temperature
of the source of radiation.
2
[Answer:2571.5 W/m ]
4. Calculate the rate of heat transfer per unit surface area by radiation between two brick
` `
walls a short distance apart, when the temperatures of the surfaces are 30 C and 15 C.
149
The emissivity of brick may be taken as 0.93, and the surfaces may be assumed to be grey.
2
[Answers:76.2 W/m ]
5. A gas turbine can-type combustion chamber of 0.3 m diameter reaches a temperature of
` `
500 C when undergoing a test in large surroundings at 15 C. The emissivity of the steel
surface is 0.79. Calculate the percentage reduction in the rate of radiant heat loss by
enclosing the combustion chamber with a cylindrical screen of 0.6 m diameter, the inside
and outside surfaces of which are painted with aluminium paint of emissivity 0.4.
[Answers: 61.3%]
`
6. A circular plate of radius 0.1 m is at a temperature of 500 C in a large room, the walls of
` `
which are at 10 C. The air in the room is at a mean temperature of 15 C. A small, spherical
thermocouple junction is placed at a distance of 0.1 m from the centre of the plate. Show
that the temperature recorded by the thermocouple is approximately 100 C.
`
2
The heat transfer coefficient from the thermocouple to the air is 25.6 W/m K, and the plate
surface may be assumed to be black for thermal radiation. Neglect conduction through the
thermocouple leads. The geometric factor is given in Table ??.
7. An electric heater 25 mm diameter and 0.3 m long is used to heat a room. Calculate the
`
electrical input to the heater when the bulk of the air in the room is at 20 C, the walls are
` `
at 15 C, and the surface of the heater is at 540 C. For convective heat transfer from the
heater, assume that
1©4
Nu 0.4 Gr
where all properties are at mean film temperature and β = 1/T, where T K is the bulk
temperature of the air.
[Take the emissivity of the heater surface as 0.55 and assume that the surroundings are
black.]
[Answer:481 W]
8. (a) Calculate the radiation heat transfer coefficient for the flat plate with a surface tem-
` `
perature of 550 C and air at temperature 15 C and a mean velocity of 6 m/s blown
across it, assuming that the surroundings are large and are at the air temperature, and
(b) compare this with the heat transfer coefficient for convection. Take the emissivity of
the plate surface as 0.6.
2
[Answer(a) 28.75 W/m K; (b) αT ©α = 1.05]
9. (a) Calculate the radiation heat transfer coefficient for the vertical wall of wall 0.6 m
` `
high by 3 m wide and maintained at 79 C in an atmosphere at 15 C., assuming that
`
the wall radiates into black surroundings at 15 C, and
(b) the emissivity of the wall surface is 0.93. Compare this value with the heat transfer
coefficient for convection.
2
[Answers:(a) 6.98 W /m K; (b) αr ©α 1.52]
10. A hot-water heater 150 mm wide by 1.2 m long by 1 m high is at a surface temperature of
` ` `
50 C in surroundings at 20 C. The walls of the room are at 13 C. The surface area of the
2
heater is 7 m and the heat transfer coefficient for convection is given by
1©3
α 1.31 ∆t
2
where α is in W /m K, and ∆ t in K.
Calculate the rate of heat transfer from the heater. Take the emissivity of the heater as 0.95
150
and assume that it is completely surrounded by black surroundings.
[Answers:1.545 kW]
LABORATORY EXERCISE(S)
1. Determination of combined convection and radiation heat transfer from a horizontal cylin-
der in natural and forced convection.
2. To find out the Stefan Boltzmann constant.
3. Determination of Emissivity of a Surface.
References
th
John Wiley;
ids 4th edn Basil Blackwell;
3.6 ECKERT ER and DRAKE RM 1971 Analysis of Heat and Mass Transfer Taylor and Francis;
3.6 KERN D Q 1950 Process Heat Transfer McGraw-Hill;
3.9 EASTOP TD and WATSON w E 1992 Mechanical Services for Buildings Longman. 3.10
EASTOP T. D. and McCONKEY A. 1993 Applied Thermodynamics For Engineering Technol-
th
ogists, 5 edn. Pearson.
151

BMTECHRE Unlocked

Uploaded by

Copyright:

Available Formats

You might also like

BMTECHRE Unlocked

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

BMTECHRE Unlocked

Uploaded by

Copyright:

Available Formats

TECHNICAL UNIVERSITY OF MOMBASA

SCHOOL OF ENGINEERING AND TECHNOLOGY

DEPARTMENT OF MECHANICAL AND AUTOMOTIVE ENGINEERING

Prepared by: Njeru Gatumu

At the end of this course the student will be able to:

1.2 Heat Transfer by Conduction

1.2.1 Fourier’s Law of Heat Conduction

where, k = Constant of proportionality and is known as thermal conductivity of the body.

Figure 1.1: Heat flow

1.2.2 Thermal Conductivity of Materials

1. Material structure, 3. Density of the material,

1.2.3 Thermal Resistance (Rth )

Potential difference (dV)

By analogy, the heat flow equation (Fourier’s equation) may be written as

Temperature difference (dt)

Conduction through a plane wall

Let, L = Thickness of the plane wall,

Mass of the wall, m ρA L (ii)

Eliminating L from eqn.(i) and eqn.(ii), we get

Heat conduction through a composite wall

Figure 1.4: Steady state conduction through a composite wall.

Adding eqns.(i), (ii) and (iii), we have

If the composite wall consists of n slabs/layers, then

Figure 1.6: Temperature drops

Referring Figure 1.10. The contact resistances are given by

and, xA 0.1146 m or 114.6 mm.

also, Thickness of insulation xB 320  114.6 205.4 mm.

(b) Heat loss per unit area, q:

Taking temperature difference to be ∆t in case,

Case II. Rock wool insulation used:

But Q̇2 1  0.8Q̇1 0.2Q̇1

0.1 0.04 0.1 0.04 xC

or, 0.1428  0.0833 0.20.1428  0.0833  15.385x

thus, x 0.0588 m or 58.8 mm

Therefore, the thickness of rock wool insulation should be 58.8 mm.

(a) The value of xB :

RT eq 0.147 Now, the total ther-

Figure 1.12: One dimensional heat

Newton’s law of cooling

Rewriting these equations in terms of the temperatures, then

Hence adding the corresponding sides of the three equations

or, 820 11  t2  15

Figure 1.15: Heat transfer through

Comparing this equation with equation (1.2), Q̇ kA/x t1  t2 , we have,

where Q̇ is analogous to I, and t1  t2  is analogous to V.

The total resistance to heat flow is then,

Or for any number of layers of material,

For any number of walls we have

Figure 1.16: Heat leaks through re-

And for this case,

11.6 46.5 0.046 46.5 14.5

(b) The temperature at the outer surface of the metal sheet, t1 :

i.e., RT 0.8 K/W

Then using equation (1.25)

(a) Thermal conductivity, k (= kB ):

1 0.25 0.1 0.15 1

337.5  25 337.5  25 2

1.2.6 Heat flow through a cylinder and a sphere

also, Thickness of insulation xB 320 114.6 205.4 mm.

But Q̇2 1 0.8Q̇1 0.2Q̇1

or, 0.1428 0.0833 0.20.1428 0.0833 15.385x

RT eq 0.147 Now, the total ther-

or, 820 11 t2 15

Comparing this equation with equation (1.2), Q̇ kA/x t1 t2 , we have,

where Q̇ is analogous to I, and t1 t2 is analogous to V.

337.5 25 337.5 25 2

2πx t1 t n1

ln r2 ©r1 Figure 1.21: Cross-

3 2πL 150 20

4π 0.083 0.61 0.7 220

Hence, Total resistance, RT 0.1478 0.2295 0.0272

or, 135 t 0.271Q̇0 (c)

Total heat generated internally q̇g volume

Hence, Heat removed from inside of conductor 17693.5 290.7 W